Isak Karabegović - New Technologies, Development and Application-Springer International Publishing (2019)

Lecture Notes in Networks and Systems 42
Isak Karabegović Editor
New
Technologies,
Development
and
Application
Lecture Notes in Networks and Systems
Volume 42
Series editor
Janusz Kacprzyk, Polish Academy of Sciences, Warsaw, Poland
e-mail: kacprzyk@ibspan.waw.pl
The series “Lecture Notes in Networks and Systems” publishes the latest
developments in Networks and Systems—quickly, informally and with high quality.
Original research reported in proceedings and post-proceedings represents the core
of LNNS.
Volumes published in LNNS embrace all aspects and subfields of, as well as
new challenges in, Networks and Systems.
The series contains proceedings and edited volumes in systems and networks,
spanning the areas of Cyber-Physical Systems, Autonomous Systems, Sensor
Networks, Control Systems, Energy Systems, Automotive Systems, Biological
Systems, Vehicular Networking and Connected Vehicles, Aerospace Systems,
Automation, Manufacturing, Smart Grids, Nonlinear Systems, Power Systems,
Robotics, Social Systems, Economic Systems and other. Of particular value to both
the contributors and the readership are the short publication timeframe and the
world-wide distribution and exposure which enable both a wide and rapid
dissemination of research output.
The series covers the theory, applications, and perspectives on the state of the art
and future developments relevant to systems and networks, decision making, control,
complex processes and related areas, as embedded in the fields of interdisciplinary
and applied sciences, engineering, computer science, physics, economics, social, and
life sciences, as well as the paradigms and methodologies behind them.
Advisory Board
Fernando Gomide, Department of Computer Engineering and Automation—DCA, School of
Electrical and Computer Engineering—FEEC, University of Campinas—UNICAMP,
São Paulo, Brazil
e-mail: gomide@dca.fee.unicamp.br
Okyay Kaynak, Department of Electrical and Electronic Engineering, Bogazici University,
Istanbul, Turkey
e-mail: okyay.kaynak@boun.edu.tr
Derong Liu, Department of Electrical and Computer Engineering, University of Illinois
at Chicago, Chicago, USA and Institute of Automation, Chinese Academy of Sciences,
Beijing, China
e-mail: derong@uic.edu
Witold Pedrycz, Department of Electrical and Computer Engineering, University of Alberta,
Alberta, Canada and Systems Research Institute, Polish Academy of Sciences, Warsaw,
Poland
e-mail: wpedrycz@ualberta.ca
Marios M. Polycarpou, KIOS Research Center for Intelligent Systems and Networks,
Department of Electrical and Computer Engineering, University of Cyprus, Nicosia, Cyprus
e-mail: mpolycar@ucy.ac.cy
Imre J. Rudas, Óbuda University, Budapest Hungary
e-mail: rudas@uni-obuda.hu
Jun Wang, Department of Computer Science, City University of Hong Kong
Kowloon, Hong Kong
e-mail: jwang.cs@cityu.edu.hk
More information about this series at http://www.springer.com/series/15179

Isak Karabegović
Editor
New Technologies,
Development
and Application
123
Editor
Isak Karabegović
Technical Faculty Bihać
University of Bihać
Bihać
Bosnia and Herzegovina
ISSN 2367-3370 ISSN 2367-3389 (electronic)

Lecture Notes in Networks and Systems
ISBN 978-3-319-90892-2 ISBN 978-3-319-90893-9 (eBook)
https://doi.org/10.1007/978-3-319-90893-9
Library of Congress Control Number: 2018942170
© Springer International Publishing AG, part of Springer Nature 2019

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made. The publisher remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
Printed on acid-free paper
This Springer imprint is published by the registered company Springer International Publishing AG
part of Springer Nature
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Interdisciplinary Research of New Technologies,
their Development and Application
The content of this book is very interesting and important as it covers a wide range
of technologies and technical disciplines including complex systems such as:
robotics, mechatronics systems, automation, manufacturing, cyber-physical sys-
tems, autonomous systems, sensor, networks, control systems, energy systems,
automotive systems, biological systems, vehicular networking and connected
vehicles, effectiveness and logistics systems, smart grids, nonlinear systems, power
systems, social systems, economic systems and other. The papers included in this
content have been presented at the international conference New Technologies,
Development and Application, held in Sarajevo, Bosnia and Herzegovina, on
28–30 June 2018. Majority of organized conferences are usually focusing on a
narrow part of the issues within a certain discipline while conferences such these are
rare. There is a need to hold such conferences. The value of this conference is that
various researchers, programmers, engineers and practitioners come to the same
place where ideas and latest technology achievements are exchanged. Such events
lead to the creation of new ideas, solutions and applications in the manufacturing
processes of various technologies. New coexistence is emerging, horizons are
expanding, and unexpected changes and analogies arise. Best solutions and
applications in technologies are critically evaluated.
The first chapter begins with robots, robotic systems, modelling of compressor
systems, mechatronic systems, automation of manufacturing processes and
advanced production. The first article offers an analysis of automation of welding
processes using industrial robots. The following article in this chapter analyses the
modelling of multiphase twin screw machines, commonly used for pumping fluids
which often contain gas, liquid and solid particles, and are of exceptional impor-
tance to industry and ecology. The following article in this chapter analyses the
influence of injection moulding process parameters on part quality. The last article
offers a power and control system of knee and ankle powered above knee
prosthesis.
The second chapter is intended to innovative and interdisciplinary applications of
advanced technologies (IATs). It is based on the analysis (IoT) of technological tools
as an opportunity to use new technologies in the development of society as a whole.
v
vi Interdisciplinary Research of New Technologies, their Development and Application
The first article is devoted to the cybersecurity capacity building planning within
organizations. The final article offers application of weighted particle swarm opti-
mization to imbalanced data in software defect prediction.
The third chapter is intended to transport systems, logistics and intelligent
transport systems. The first article gives an analysis of cooperative control in traffic
and transportation technology. The second article provides a solution to traffic
control in urban areas, while the final article offers the use of unmanned aerial
vehicles in logistics processes.
The fourth chapter is intended to electric power systems with different spectrum
of topics from turbulence analysis of wind power plants, pico power plants, high
energy efficiency to analysis of combustion technologies with the aim of achieving
ecological standards.
The fifth chapter is intended to new methods in agricultural culture of a broad
spectrum of topics: modelling the extraction process of sage, effects of sage extract,
from occurrence of apple powdery mildew to application of multivariate statistic to
classify blueberry fruits, and in addition the detection of heavy metals in hair
colours by the atomic absorption spectrophotometry, the content of heavy metals in
“PET” bottles of drinking water and its electrical conductivity, microbiological
analysis of surface waters and research of antimicrobial resistance of clinical
important multiresistant gram-negative bacterial isolates.
The sixth chapter focuses on new technology in civil engineering, education,
control quality and other. The first article focuses on nanotechnology in civil
engineering. In next article, information about parametric vector analysis of
available resources for minimization of project duration is given. The last article in
this chapter considers education at universities.
The whole content of this book is intended to a wide range of technical systems;
different technical disciplines in order to apply the latest solutions and achievements
in technologies and to improve manufacturing processes in all disciplines where
systemic thinking has a very important role in the successful understanding and
building of human, natural and social systems. We hope this content will be the first
in a series of publications that are intended to the development and implementation
of new technologies in all industries.
Isak Karabegović
Contents
New Technologies in Mechanical Engineering, Metallurgy,

Mechatronics, Robotics and Embedded Systems
Automation of the Welding Process by Use of Industrial Robots . . . . . . 3
Isak Karabegović and Riaz Mirza
Modelling of Multiphase Twin Screw Machines . . . . . . . . . . . . . . . . . . . 18
Ahmed Kovacevic, Sham Rane, and Nikola Stosic
Influence of Injection Molding Process Parameters on Part Quality . . . 33
Janez Gotlih, Miran Brezočnik, Igor Drstvenšek, Timi Karner,
and Karl Gotlih
Numerical Analysis of Material Fatigue Impact on Bicycle Frame
Safety in Accordance with EN 14764 . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
Nermina Zaimović-Uzunović, Ernad Bešlagić, and Almir Porča
The Influence of the Tool Geometry on the Quality of the Weld
in FSW Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
Aleksandra Koprivica, Nikola Šibalić, Milan Vukčević,
and Mirjana Jokanović
Dimension Measurement of Injection Moulded Toybricks . . . . . . . . . . . 57
Samir Lemeš and Anel Baručija
Wire and Arc Additive Manufacturing (WAAM) – A New Advance
in Manufacturing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
Nikola Knezović and Angela Topić
Analysis of the Type and Chemical Content of the Inclusion on SEM
of the Stainless Steel With and Without the Addition of Zr and Te . . . . 72
Derviš Mujagić, Aida Imamović, Mirsada Oruč,
and Sulejman Muhamedagić
vii
viii Contents
Strength and Deformation Calculation of Flat O-Springs . . . . . . . . . . . . 80

Ivan Balashev, Mariel Penev, Ivan Stoyanov, and Radoslav Balashev
Prototype Apparatus for Calibration Contact Sensors for Measuring
the Temperature of a Solid Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
Edin Terzić, Raif Seferović, and Narcisa Jarović-Bajramović
Automatic Control of Tube Hydroforming Process
in Experimental Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
Edina Karabegović, Edin Šemić, and Safet Isić
Analysis and Determination of Friction in Hydroforming Process
of Cross Tube . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
Mehmed Mahmić, Edina Karabegović, and Ermin Husak
Science of Metals Through Lens of Microscope . . . . . . . . . . . . . . . . . . . 113
Belma Fakić, Adisa Burić, and Edib Horoz
Increase of Performance of Grinding by Plate Circles . . . . . . . . . . . . . . 121
Tonkonogyi Vladimir, Yakimov Alexey, and Bovnegra Liubov
Analysis of Torsional Vibration of the Engine Connected
with Propeller Through Pair of Gears . . . . . . . . . . . . . . . . . . . . . . . . . . 128
Ermin Husak and Erzad Haskić
FEM Model of Misaligned Rotational System
with Rotating Looseness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
Emir Nezirić, Safet Isić, Isak Karabegović, and Avdo Voloder
Application of Explosives in Metal Forming . . . . . . . . . . . . . . . . . . . . . . 144
Darko Šunjić and Stipo Buljan
Application of Iterative Methods to Solve Inverse Kinematics
Problem of Robot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
Avdo Voloder
Parameter Fitting for Soft Dielectric Elastomer Actuator . . . . . . . . . . . 156
Timi Karner, Janez Gotlih, Boštjan Razboršek, and Karl Gotlih
Timber Construction and Robots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
Salah-Eldien Omer
Conceptual Solution of the Robotic Arm/Plotter . . . . . . . . . . . . . . . . . . 170
Milena Djukanovic, Rade Grujicic, Luka Radunovic, and Vuk Boskovic
Robot for Cleaning Ventilation Ducts . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
Milos Bubanja, Marina Mijanovic Markus, Milena Djukanovic,
and Mihailo Vujovic
Cloud Robotics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
Samir Vojić
Contents ix
Vibration Analysis of Motorcycle Handles . . . . . . . . . . . . . . . . . . . . . . . 196

Zlata Jelačić and Boran Pikula
Acoustic Diagnostics of Lever Mechanisms with Subsequent
Processing of Data on Neural Networks . . . . . . . . . . . . . . . . . . . . . . . . . 202
Sergiy Kovalevskyy, Olena Kovalevska, and Raul Turmanidze
Power and Control System of Knee and Ankle Powered
Above-Knee Prosthesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
Miljan Rupar, Zlata Jelačić, Remzo Dedić, and Adisa Vučina
Computer Science, Information and Communication

Technologies, e-Business
Cyber Security Capacity Building Planning Within Organisations . . . . . 219
Ramo Šendelj and Ivana Ognjanović
New Method of Sequences Spiral Hybrid Using Machine Learning
Systems and Its Application to Engineering . . . . . . . . . . . . . . . . . . . . . . 227
Matej Babič, Isak Karabegović, Sanda Ipšič Martinčič, and Gyula Varga
A Multifunctional Platform for Elders’ Assisting to Live Alone . . . . . . . 238
Blerina Zanaj, Fatjon Shaba, Majlinnda Belegu, and Gerti Boshnjaku
Economic Aspects of the Application of Cloud Computing . . . . . . . . . . 247
Mirha Bičo Ćar, Savo Stupar, and Elvir Šahić
The Role of Hadoop Technology in the Implementation
of Big Data Concept . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 254
Savo Stupar, Mirha Bičo Ćar, and Elvir Šahić
Cybernetization of Industrial Product-Service Systems
in Network Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 262
Elvis Hozdić and Zoran Jurković
Technology-Enhanced Systems in Idiopathic Scoliosis 3D Diagnosis
and Screening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
Saša M. Ćuković, William R. Taylor, and Ionuţ G. Ghionea
Contributions to Improve the Sustainability in Services Based
Organizations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279
Mihail Aurel Titu, Bianca Alina Pop, and Stefan Titu
Applying Weighted Particle Swarm Optimization to Imbalanced
Data in Software Defect Prediction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289
Lucija Brezočnik and Vili Podgorelec
x Contents
Intelligent Transport Systems, Logistics, Traffic Control

Cooperative Systems in Traffic Technology and Transport . . . . . . . . . . 299
Sadko Mandžuka
Wireless Sensor Network Based Alarm Detection and Monitoring
of Cyber-Physical System with Mobile Robot Inspection . . . . . . . . . . . . 309
Lejla Banjanovic-Mehmedovic and Mirzet Zukic
Review of Simulation Based Comparison of VANET Protocols . . . . . . . 317
Zlatan Jukic and Muhammad Arshad
Ramp Metering on Urban Motorways . . . . . . . . . . . . . . . . . . . . . . . . . . 332
Martin Gregurić, Sadko Mandžuka, and Edouard Ivanjko
Terminology Extraction to Build an Ontology of Intelligent
Transport Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 338
Pero Škorput, Sadko Mandžuka, and Markus Schatten
Cooperative Vehicle Actuated Traffic Control in Urban Areas . . . . . . . 345
Miroslav Vujić, Sadko Mandžuka, and Luka Dedić
Application of Mini Computers and RFID Technology
in Automation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 354
Malik Čabaravdić, Sanela Čančar, and Anel Husaković
Application of Unmanned Aerial Vehicles in Logistic Processes . . . . . . . 359
Jasmina Pašagić Škrinjar, Pero Škorput, and Martina Furdić
New Technologies in the Field Energy: Renewable Energy,

Power Quality, Advanced Electrical Power Systems
The Turbulence Intensity of the Wind Bora . . . . . . . . . . . . . . . . . . . . . . 369
Blago Pehar, Elvir Zlomušica, and Suad Zalihić
Lab-Scale Tests as Support to Selection of Sustainable Coal
Combustion Technology - Case Study: Support to Design
of TPP Kakanj Unit 8 - . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 377
Nihad Hodžić, Sadjit Metović, and Anes Kazagić
Influence of New Technologies on Higher Energy Efficiency
of Hydrostatic Devices and Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 386
Milutin Živković, Predrag Dašić, and Zvonko Petrović
Design of Pico Hydropower Plants for Rural Electrification . . . . . . . . . 397
Krsto Batinić, Dušan Golubović, Stojan Simić, and Goran Orašanin
Development of Construction of Mini Hydro Power Plant Model
Based on Pelton Turbine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 405
Radoslav Tomović, Aleksandar Tomović, Marko Mumović,
and Vuk Vujošević
Contents xi
Gas Escape from Combustion Chamber to Crankcase, Analysis

of a Set of Parameters Affecting the Blow by . . . . . . . . . . . . . . . . . . . . . 413
Cristiana Delprete, Arian Bisha, and Erjon Selmani
Comb-Based Decimator for Multiples-of-Five Decimation Factors . . . . . 423
Gordana Jovanovic Dolecek and Isak Karabegovic
The Filter-Compensation Device Applications to the AC 25 kV 50 Hz
AC of Serbian Railways . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 429
Branislav Gavrilovic and Zoran Bundalo
Optimization of Water Supply System Using Software EPANET 2.0 . . . 443
Aleksandar Košarac, Dejan Romić, Goran Orašanin,
and Jovana Blagojević
The Possibilities of the Cadastral Land Use Assessment
by the Methods of Remote Sensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 452
Admir Mulahusić, Jusuf Topoljak, Nedim Tuno, and Karlo Ajvazović
Analysis of the Energy Potential of Organic Bioradable Part
of Municipal Waste . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 459
Mahmut Jukić and Ifet Šišić
New Technologies in Agriculture and Ecology, Chemical Processes

Modeling the Extraction Process of Sage (Salvia Officinalis L.)
with Supercritical CO2 at Different Temperatures . . . . . . . . . . . . . . . . . 469
Sabina Begić, Vladan Mićić, and Darko Manjenčić
The Effect of Concentration of Methanol as a Solvent on the
Antioxidative Activity of Sage Extract . . . . . . . . . . . . . . . . . . . . . . . . . . 480
Selma Osmić, Sabina Begić, and Vladan Mićić
Occurrence of Apple Powdery Mildew, Podosphaera Leucotricha
(Ellis. & Everh.) E. S. Salmon in North-Western Region of Bosnia
and Herzegovina . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 491
Zemira Delalić
Application of Multivariate Statistic to Classify Blueberry Fruits . . . . . 498
Vildana Alibabić, Azra Skender, Melisa Orašćanin, and Ibrahim Mujić
The Effect of Technological Process on Physico-Chemical
and Nutritional Properties of Sour Cherries Products . . . . . . . . . . . . . . 507
Ramzija Cvrk, Azra Begović, Snježana Marić, and Nils V. Juul
Application of New Technologies in Meat Processing Industry
in the Function of Improvement of Total Quality of Products
and Consumer Protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 513
Amir Ganić, Munevera Begić, and Enver Karahmet
xii Contents
The Impact of Water on PVC Floor Coverings . . . . . . . . . . . . . . . . . . . 522

Mario Krzyk and Darko Drev
Exposure to PM10 Aerosol Particles and Other Aerial Pollutants
in the Capital City of Montenegro . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 529
Jovana Jovanovic and Svetlana Stevovic
New Technological Procedures for Production of Thioncarbamates
as a Selective Flotation Reagents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 542
Milutin M. Milosavljević, Milan M. Milosavljević, Milutin Živković,
and Ljiljana Pecić
The Content of Heavy Metals in “PET” Bottles of Drinking Water
and Its Electrical Conductivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 552
Ekrem Pehlić, Aida Šapčanin, Husein Nanić, and Adnan Ćehajić
Determination of Heavy Metals in Hair Dyes by the Atomic
Absorption Spectrophotometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 561
Ekrem Pehlić, Husein Nanić, Huska Jukić, and Aldina Aldžić
Microbiological Analysis of Surface Waters in the Area
of National Park “Una” . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 568
Melisa Zulić, Huska Jukić, Asmir Aldžić, and Adnan Ćehajić
Research of Antimicrobial Resistance of Clinical Important
Multi-resistent Gram Negative Bacterial Isolates
in the Una-Sana Canton Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 575
Asmir Aldžić, Huska Jukić, Kanita Dedić, and Amela Dubinović-Rekić
New Technologies in Civil Engineering, Education, Control Quality

Nanotechnology in Civil Engineering . . . . . . . . . . . . . . . . . . . . . . . . . . . 585
Merima Šahinagić-Isović, Marko Ćećez, and Fuad Ćatović
Support Parametric Vector Analysis of Available Resources for
Minimisation of Project Duration–Four Varieties of Conditions . . . . . . 590
Omer Kurtanović and Lejla Dacić
New Technologies in Education at “Džemal Bijedić” University
in Mostar . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 597
Alim Abazović, Dragi Tiro, and Fuad Ćatović
Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 605
Contributors
Alim Abazović “Dzemal Bijedic” University of Mostar, Mostar, Bosnia and

Herzegovina
Karlo Ajvazović University of Sarajevo, Sarajevo, Bosnia and Herzegovina
Aldina Aldžić University of Bihac, Bihać, Bosnia and Herzegovina
Asmir Aldžić University of Bihać, Bihać, Bosnia and Herzegovina
Yakimov Alexey Institute of Industrial Technologies, Design and Management,
Odessa National Polytechnic University, Odessa, Ukraine
Vildana Alibabić University of Bihać, Bihać, Bosnia and Herzegovina
Muhammad Arshad University of Engineering and Technology Lahore, Lahore,
Pakistan
Matej Babič Jožef Stefan Institute, Ljubljana, Slovenia
Ivan Balashev Technical University of Gabrovo, Gabrovo, Bulgaria
Radoslav Balashev PRONY Engineering Ltd., Gabrovo, Bulgaria
Lejla Banjanovic-Mehmedovic University of Tuzla, Tuzla, Bosnia and
Herzegovina
Anel Baručija University of Zenica, Zenica, Bosnia and Herzegovina
Krsto Batinić University of East Sarajevo, East Sarajevo, Bosnia and Herzegovina
Munevera Begić University of Sarajevo, Sarajevo, Bosnia and Herzegovina
Sabina Begić University of Tuzla, Tuzla, Bosnia and Herzegovina; University of
East Sarajevo, Zvornik, Bosnia and Herzegovina
Azra Begović Faculty of Technology, University of Tuzla, Tuzla, Bosnia and
Herzegovina
xiii
xiv Contributors
Majlinnda Belegu Agricultural University of Tirana, Tirana, Albania

Ernad Bešlagić University of Zenica, Zenica, Bosnia and Herzegovina
Arian Bisha Universiteti Politeknik i Tiranes, Tirana, Albania
Mirha Bičo Ćar University of Sarajevo, Sarajevo, Bosnia and Herzegovina
Jovana Blagojević University of East Sarajevo, East Sarajevo, Bosnia and
Herzegovina
Gerti Boshnjaku New Media Communications, Tirana, Albania
Vuk Boskovic University of Montenegro, Podgorica, Montenegro
Lucija Brezočnik Faculty of Electrical Engineering and Computer Science,
University of Maribor, Maribor, Slovenia
Miran Brezočnik University of Maribor, Maribor, Slovenia
Milos Bubanja University of Montenegro, Podgorica, Montenegro
Stipo Buljan Federal Ministry of Energy, Mostar, Bosnia and Herzegovina
Zoran Bundalo Railway College of Vocational Studies, Belgrade, Serbia
Adisa Burić University of Zenica, Metallurgical Institute “Kemal Kapetanović”,
Zenica, Bosnia and Herzegovina
Malik Čabaravdić University of Zenica, Zenica, Bosnia and Herzegovina
Sanela Čančar University of Zenica, Zenica, Bosnia and Herzegovina
Fuad Ćatović “Džemal Bijedić” University of Mostar, Mostar, Bosnia and
Herzegovina
Marko Ćećez “Džemal Bijedić” University of Mostar, Mostar, Bosnia and
Herzegovina
Adnan Ćehajić University of Bihać, Bihać, Bosnia and Herzegovina
Saša M. Ćuković Department of Health Sciences and Technology, Institute for
Biomechanics, Swiss Federal Institute of Technology – ETH Zurich, Zurich,
Switzerland
Ramzija Cvrk Faculty of Technology, University of Tuzla, Tuzla, Bosnia
and Herzegovina
Lejla Dacić University of Travnik, Travnik, Bosnia and Herzegovina
Predrag Dašić High Technical Mechanical School of Professional Studies,
Trstenik, Serbia
Kanita Dedić Cantonal Hospital, “Dr. Irfan Ljubijankić” Bihac, Bihac, Bosnia and
Herzegovina
Contributors xv
Luka Dedić University of Zagreb, Zagreb, Croatia

Remzo Dedić University of Mostar, Mostar, Bosnia and Herzegovina
Zemira Delalić University of Bihać, Bihać, Bosnia and Herzegovina
Cristiana Delprete Politecnico di Torino, Turin, Italy
Milena Djukanovic University of Montenegro, Podgorica, Montenegro
Darko Drev University of Ljubljana, Ljubljana, Slovenia
Igor Drstvenšek University of Maribor, Maribor, Slovenia
Amela Dubinović-Rekić Cantonal Hospital, “Dr. Irfan Ljubijankić” Bihac, Bihac,
Belma Fakić University of Zenica, Metallurgical Institute “Kemal Kapetanović”,
Martina Furdić University of Zagreb, Zagreb, Croatia
Amir Ganić University of Sarajevo, Sarajevo, Bosnia and Herzegovina
Branislav Gavrilovic Railway College of Vocational Studies, Belgrade, Serbia
Ionuţ G. Ghionea Faculty of Engineering and Management of Technological
Systems, University Politehnica of Bucharest, Bucharest, Romania
Dušan Golubović University of East Sarajevo, East Sarajevo, Bosnia and
Herzegovina
Janez Gotlih University of Maribor, Maribor, Slovenia
Karl Gotlih University of Maribor, Maribor, Slovenia
Martin Gregurić University of Zagreb, Zagreb, Croatia
Rade Grujicic University of Montenegro, Podgorica, Montenegro
Erzad Haskić University of Bihać, Bihać, Bosnia and Herzegovina
Nihad Hodžić University of Sarajevo, Sarajevo, Bosnia and Herzegovina
Edib Horoz University of Zenica, Metallurgical Institute “Kemal Kapetanović”,
Elvis Hozdić University of Ljubljana, Ljubljana, Slovenia
Ermin Husak University of Bihać, Bihać, Bosnia and Herzegovina
Anel Husaković University of Zenica, Zenica, Bosnia and Herzegovina
Aida Imamović Institute “Kemal Kapetanović”, University of Zenica, Zenica,
xvi Contributors
Safet Isić Faculty of Mechanical Engineering, University “Džemal Bijedić”

Mostar, Mostar, Bosnia and Herzegovina
Edouard Ivanjko University of Zagreb, Zagreb, Croatia
Narcisa Jarović-Bajramović Metallurgical Institute “Kemal Kapetanović”,
University of Zenica, Zenica, Bosnia and Herzegovina
Zlata Jelačić University of Sarajevo, Sarajevo, Bosnia and Herzegovina
Mirjana Jokanović University of East Sarajevo, East Sarajevo, Bosnia and
Herzegovina
Gordana Jovanovic Dolecek Institute INAOE, Puebla, Mexico
Jovana Jovanovic University Union Nikola Tesla, Belgrade, Serbia
Zlatan Jukic HTL Rankweil & Vienna University of Technology, Vienna, Austria
Huska Jukić University of Bihać, Bihać, Bosnia and Herzegovina
Mahmut Jukić University of Bihac, Bihać, Bosnia and Herzegovina
Zoran Jurković University of Rijeka, Rijeka, Croatia
Nils V. Juul Sør-Trøndelag University College (HiST), Høgskolen i
Sør-Trøndelag, Trondheim, Norway
Edina Karabegović Faculty of Technical Engineering Bihać, University of Bihać,
Bihać, Bosnia and Herzegovina
Isak Karabegović Technical Faculty, University of Bihać, Bihać, Bosnia and
Herzegovina
Enver Karahmet University of Sarajevo, Sarajevo, Bosnia and Herzegovina
Timi Karner University of Maribor, Maribor, Slovenia
Anes Kazagić University of Sarajevo, Sarajevo, Bosnia and Herzegovina
Nikola Knezović University of Mostar, Mostar, Bosnia and Herzegovina
Aleksandra Koprivica University of East Sarajevo, East Sarajevo, Bosnia and
Herzegovina
Ahmed Kovacevic Centre for Compressor Technology, City, University of
London, London, UK
Olena Kovalevska Department of Machine Building Technology, Donbass State
Engineering Academy, Kramatorsk, Ukraine
Sergiy Kovalevskyy Department of Machine Building Technology, Donbass State
Contributors xvii
Aleksandar Košarac University of East Sarajevo, East Sarajevo, Bosnia and

Herzegovina
Mario Krzyk University of Ljubljana, Ljubljana, Slovenia
Omer Kurtanović University of Bihac, Bihać, Bosnia and Herzegovina
Samir Lemeš University of Zenica, Zenica, Bosnia and Herzegovina
Bovnegra Liubov Institute of Industrial Technologies, Design and Management,
Odessa National Polytechnic University, Odessa, Ukraine
Mehmed Mahmić University of Bihać, Bihać, Bosnia and Herzegovina
Sadko Mandžuka Faculty of Traffic and Transport Sciences, University of
Zagreb, Zagreb, Croatia
Darko Manjenčić University of Novi Sad, Novi Sad, Serbia
Snježana Marić Faculty of Technology, University of Tuzla, Tuzla, Bosnia and
Herzegovina
Marina Mijanovic Markus University of Montenegro, Podgorica, Montenegro
Sanda Ipšič Martinčič University of Rijeka, Rijeka, Croatia
Sadjit Metović University of Sarajevo, Sarajevo, Bosnia and Herzegovina
Milan M. Milosavljević University of Priština, Kosovska Mitrovica, Serbia
Milutin M. Milosavljević University of Priština, Kosovska Mitrovica, Serbia
Riaz Mirza University of Engineering & Technology, Lahore, Pakistan
Vladan Mićić University of Tuzla, Tuzla, Bosnia and Herzegovina; University of
East Sarajevo, Zvornik, Bosnia and Herzegovina
Sulejman Muhamedagić Institute “Kemal Kapetanović”, University of Zenica,
Derviš Mujagić Institute “Kemal Kapetanović”, University of Zenica, Zenica,
Ibrahim Mujić University of Bihać, Bihać, Bosnia and Herzegovina; Colegium
Fluminense Polytechnic of Rijeka, Rijeka, Croatia
Admir Mulahusić University of Sarajevo, Sarajevo, Bosnia and Herzegovina
Marko Mumović University of Montenegro, Podgorica, Montenegro
Husein Nanić University of Bihac, Bihać, Bosnia and Herzegovina
Emir Nezirić ‘‘Džemal Bijedić’’ University of Mostar, Mostar, Bosnia and
Herzegovina
xviii Contributors
Ivana Ognjanović University of Donja Gorica, Podgorica, Montenegro

Salah-Eldien Omer SAG CONSULTING d.o.o., Zagreb, Croatia
Goran Orašanin University of East Sarajevo, East Sarajevo, Bosnia and
Herzegovina
Melisa Orašćanin University of Bihać, Bihać, Bosnia and Herzegovina
Mirsada Oruč Institute “Kemal Kapetanović”, University of Zenica, Zenica,
Selma Osmić University of Tuzla, Tuzla, Bosnia and Herzegovina
Ljiljana Pecić Bachelor School for Professional Technical Studies, Trstenik,
Serbia
Blago Pehar ‘‘Džemal Bijedić’’ University of Mostar, Mostar, Bosnia and
Herzegovina
Ekrem Pehlić University of Bihac, Bihać, Bosnia and Herzegovina
Mariel Penev Technical University of Gabrovo, Gabrovo, Bulgaria
Zvonko Petrović High Technical Mechanical School of Professional Studies,
Trstenik, Serbia
Boran Pikula University of Sarajevo, Sarajevo, Bosnia and Herzegovina
Vili Podgorelec Faculty of Electrical Engineering and Computer Science,
University of Maribor, Maribor, Slovenia
Bianca Alina Pop SC TEHNOCAD SA, Baia Mare, Romania
Almir Porča University of Zenica, Zenica, Bosnia and Herzegovina
Luka Radunovic University of Montenegro, Podgorica, Montenegro
Sham Rane Department of Engineering Science, University of Oxford, Oxford,
UK
Boštjan Razboršek University of Maribor, Maribor, Slovenia
Dejan Romić University of East Sarajevo, East Sarajevo, Bosnia and Herzegovina
Miljan Rupar University of Mostar, Mostar, Bosnia and Herzegovina
Markus Schatten University of Zagreb, Zagreb, Croatia
Raif Seferović Metallurgical Institute “Kemal Kapetanović”, University of Zenica,
Erjon Selmani Universiteti Politeknik i Tiranes, Tirana, Albania
Fatjon Shaba New Media Communications, Tirana, Albania
Contributors xix
Stojan Simić University of East Sarajevo, East Sarajevo, Bosnia and Herzegovina
Azra Skender University of Bihać, Bihać, Bosnia and Herzegovina
Svetlana Stevovic University of Beograd, Belgrade, Serbia
Nikola Stosic Centre for Compressor Technology, City, University of London,
London, UK
Ivan Stoyanov Podem Gabrovo Ltd., Gabrovo, Bulgaria
Savo Stupar University of Sarajevo, Sarajevo, Bosnia and Herzegovina
William R. Taylor Department of Health Sciences and Technology, Institute
for Biomechanics, Swiss Federal Institute of Technology – ETH Zurich,
Zurich, Switzerland
Edin Terzić Metallurgical Institute “Kemal Kapetanović”, University of Zenica,
Dragi Tiro “Dzemal Bijedic” University of Mostar, Mostar, Bosnia and
Herzegovina
Mihail Aurel Titu Lucian Blaga University of Sibiu, Sibiu, Romania
Stefan Titu The Oncology Institute “Prof. dr. Ion Chiricuță” Cluj Napoca,
Cluj-Napoca, Romania
Aleksandar Tomović University of Montenegro, Podgorica, Montenegro
Radoslav Tomović University of Montenegro, Podgorica, Montenegro
Angela Topić University of Mostar, Mostar, Bosnia and Herzegovina
Jusuf Topoljak University of Sarajevo, Sarajevo, Bosnia and Herzegovina
Nedim Tuno University of Sarajevo, Sarajevo, Bosnia and Herzegovina
Raul Turmanidze Department of Machine Building Technology, Donbass State
Gyula Varga University of Miskolc, Miskolc, Hungary
Tonkonogyi Vladimir Institute of Industrial Technologies, Design and
Management, Odessa National Polytechnic University, Odessa, Ukraine
Samir Vojić Technical Faculty Bihać, University of Bihać, Bihać, Bosnia and
Herzegovina
Avdo Voloder Faculty of Mechanical Engineering, University of Sarajevo,
Sarajevo, Bosnia and Herzegovina
Adisa Vučina University of Mostar, Mostar, Bosnia and Herzegovina
Miroslav Vujić University of Zagreb, Zagreb, Croatia
xx Contributors
Mihailo Vujovic University of Montenegro, Podgorica, Montenegro

Vuk Vujošević University of Montenegro, Podgorica, Montenegro
Milan Vukčević University of Montenegro, Podgorica, Montenegro
Nermina Zaimović-Uzunović University of Zenica, Zenica, Bosnia and
Herzegovina
Suad Zalihić ‘‘Džemal Bijedić’’ University of Mostar, Mostar, Bosnia and
Herzegovina
Blerina Zanaj Agricultural University of Tirana, Tirana, Albania
Elvir Zlomušica ‘‘Džemal Bijedić’’ University of Mostar, Mostar, Bosnia and
Herzegovina
Mirzet Zukic University of Tuzla, Tuzla, Bosnia and Herzegovina
Melisa Zulić University of Bihać, Bihać, Bosnia and Herzegovina
Merima Šahinagić-Isović “Džemal Bijedić” University of Mostar, Mostar,
Elvir Šahić University of Sarajevo, Sarajevo, Bosnia and Herzegovina
Aida Šapčanin Faculty of Pharmacy, University of Sarajevo, Sarajevo, Bosnia
and Herzegovina
Edin Šemić Faculty of Mechanical Engineering, University “Džemal Bijedić”
Mostar, Mostar, Bosnia and Herzegovina
Ramo Šendelj University of Donja Gorica, Podgorica, Montenegro
Nikola Šibalić University of Montenegro, Podgorica, Montenegro
Ifet Šišić University of Bihac, Bihać, Bosnia and Herzegovina
Pero Škorput University of Zagreb, Zagreb, Croatia
Jasmina Pašagić Škrinjar University of Zagreb, Zagreb, Croatia
Darko Šunjić University of Mostar, Mostar, Bosnia and Herzegovina
Milutin Živković High Technical Mechanical School of Professional Studies,
Trstenik, Serbia
Milutin Živković Bachelor School for Professional Technical Studies, Trstenik,
Serbia
New Technologies in Mechanical
Engineering, Metallurgy, Mechatronics,
Robotics and Embedded Systems
Automation of the Welding Process
by Use of Industrial Robots
Isak Karabegović1 ✉ and Riaz Mirza2

( )
1
University of Bihać, 77000 Bihać, Bosnia and Herzegovina
isak1910@hotmail.com
2
University of Engineering & Technology, Lahore, Pakistan
mriazmirza@uet.edu.pk
Abstract. The development of robotic technology, owing to the advancement

of digital technology, is evolving each year, resulting in increased representation
of industrial robots. We are currently in the fourth industrial revolution, referred
to as “Industry 4.0” by the Germans. The implementation of the fourth techno‐
logical revolution depends on a series of new and innovative technological
achievements, most of which are applied in robotic technology. Automation of
production processes, including automation of the welding process, must include
industrial robots. This paper demonstrates the representation of industrial robots
in the world continents, and four top countries: China, Japan, North America and
Germany. An analysis was conducted of the annual representation of industrial
robots in the welding process worldwide and on continents of Asia/Australia,
Europe, and Americas, for the period 2010–2016. As it is known that industrial
robots are most represented in the automotive industry, a tendency of their repre‐
sentation was depicted in the automotive industry for the same period, as well as
the percentage in all industrial branches by 2016. Industrial robots are most widely
used in two welding processes: arc welding and spot welding, so an analysis of
their representation in these two welding processes for the period 2010–2016 was
conducted. A comparative analysis of the annual production of vehicles in four
countries was made: China, Japan, USA and Germany, as well as the presence of
robots in these countries in the welding processes. The paper also includes the
analysis and possibilities of future industrial robot representation in this area.
Keywords: Industry · Process · Robot · Welding · Arc welding · Spot welding

Application
1 Introduction
The production process in any industry branch is unconceivable without the use of
industrial robots. The automation process in the industry started in the 1960s when
industrial robots were introduced, and it continues to this day. The process of automation
of production process with application of the first-generation industrial robots at that
time period was positive because they replaced people in performing difficult and
dangerous jobs. This was a rigid and non-flexible automation, because in order to initiate
the production of another product in the same manufacturing process, it was necessary

I. Karabegović (Ed.): NT 2018, LNNS 42, pp. 3–17, 2019.
https://doi.org/10.1007/978-3-319-90893-9_1
4 I. Karabegović and R. Mirza
to reprogram each robot with its grips, changing tools, etc., which in return caused long-
lasting delays in the production and presented additional production costs. We must also
state the fact that industrial robots were enclosed with fences for reasons not to hurt the
workers engaged in the production process. With the continuous automation of produc‐
tion processes and their flexibility, as well as demands for constant change in production
lines, the function of industrial robot becomes increasingly demanding and complex
with the tendency of increasing the use of industrial robots. Every day we have techno‐
logical improvements in terms of flexibility, accuracy, security and simplification of the
use of industrial robots. Medium and small businesses will start using flexible automa‐
tion in order to be competitive in the market. The presence of industrial robots is still
the largest in the automotive industry in the welding process, but other industries do not
lag behind in increasing the use of industrial robots. Industrial robots are ideal for jobs
that are considered to be difficult and disadvantageous for people, and jobs that are
hazardous to their health, particularly welding. They are used for repetitive jobs that are
considered monotonous, as well as for products which require high quality and high
productivity, such as automotive industry. Various industrial robots have been designed
precisely for a specific type of task. The application of robotic systems in the industry
also presents the humanization of work, and the best example is the welding processes
in any industry [1–14, 22–24, 29–35]. Industrial robots are used for arc welding, spot
welding, laser welding, soldering and other types of welding. The paper presents the
analysis of the use of industrial robots in the world, by continents, as well as in the four
countries where the automotive industry is the most represented, i.e. in the countries
where the most vehicles are produced. In addition, the tendency of representation of
industrial robots is presented in two welding processes: arc welding and spot welding,
since these two welding processes mostly use industrial robots. When it comes to the
transformation of production processes, we are referring to their modernization, with
the aim of achieving intelligent production processes. This process is unimaginable
without the presence of both industrial and service robots of the new generation. The
new generation of robots must be intelligent and autonomous, i.e. to make independent
decisions and communicate with people and machines. The application of such industrial
and service robots will increase the reliability of the manufacturing process, reduce the
time to create the finished product, and enable adapting and precision in performing
tasks that exceed human capabilities.
2 The Representation of Industrial Robots
In order to depict the representation of industrial robots in production processes world‐

wide, the data for statistical analysis were taken from the International Federation of
Robotics (IFR), the UN Economic Commission for Europe (UNECE) and the Organi‐
zation for Economic Co-operation and Development (OECD) [15–25, 29], as are shown
in Fig. 1.
Worldwide representation of industrial robots at an annual level in the period 2009–
2016 is continually increasing, as shown in Fig. 1(a), with the recorded decrease in
representation in 2012. In early 2008, there was an economic and industrial crisis in the
Automation of the Welding Process by Use of Industrial Robots 5
Number of units Number of units

259.000 1.828.024
x 1000
300.000 2.000
1.800
250.000
1.600
1.400 1.020.731
200.000
1.200
150.000 1.000
60.011 800
100.000
600
400
50.000
200
0 0
2009 2010 2011 2012 2013 2014 2015 2016 2009 2010 2011 2012 2013 2014 2015 2016
Fig. 1. The representation of industrial robots in the automation of production processes in the
world on annual and total level for the period 2009–2016
world that reflected the annual presence of industrial robots, so that the lowest repre‐
sentation of only 60.011 industrial robot units was marked in 2009. We can point out
that the increase in the number of industrial robots in the world is growing from year to
year, so that in 2016 the tendency of use has raised to 259.000 industrial robot units.
The reasons for this tendency of representation of industrial robots are many, including
the fact that companies want to be competitive in the market and introduce automation
into their production processes, which reflects the presence of industrial robots. Other
reasons include the development of robotic technology and the fourth digital revolution
that simplifies the use of robots in production processes, decreasing price of industrial
robots, etc. In order to depict the actual representation of industrial robots in the world,
we have to conduct an analysis of their representation on annual and total level by
continents, as shown in Fig. 2 [15–29].
250.000 1.200.000
1.024.897
191.000 Asia/Australia
1.000.000
200.000
Asia/Australia 800.000
150.000
600.000 459.937
Europa Europa
100.000 56.000
400.000
Americas
50.000
200.000
Americas 299.503
0 41.000 0
2010 2011 2012 2013 2014 2015 2016 2010 2011 2012 2013 2014 2015 2016
Fig. 2. The representation of industrial robots in the automation of production processes in the
world on annual and total level for the period 2010–2016
The tendency of representation of industrial robots at the annual level, Fig. 2.a, indi‐
cates that Asia/Australia holds the first place by the presence of industrial robots in the
production processes of the industry. We see the continuous increase in the application
of robots in production processes from year to year in the period 2010–2016, so that the
representation reached about 191.000 industrial robot units in 2016. The second place
is held by Europe, that demonstrates a growing tendency, but unlike Asia/Australia, it
is a slight increase from year to year. In 2016 it reached the value of about 56.000
industrial robot units, which compared to Asia/Australia is almost four times less. The
third place is held by America with a slight increasing tendency, with about 41.000
industrial robot units used in 2016, which is around five times less than Asia/Australia
in the same year. In regard to the representation of industrial robots at the overall level,
Fig. 2.b, we can see that the tendency in all three continents has a growing character. It
is more distinct in Asia/Australia, which holds the first place, and in 2016 the represen‐
tation reached about 1 million industrial robot units. Europe has a slower tendency of
representation, which in 2016 reached about 460.000 units, whereas America marked
around 300.000 industrial robot units in the same year. There is a far less difference in
the representation of industrial robots in Europe and America compared to Asia/
Australia at the overall level than at the annual level. We have to note that the analysis
did not include the continent of Africa, because the representation is very small and
cannot be compared with other three continents. We have conducted the analysis of the
representation of industrial robots in the four developed countries in the world where
industrial robots are most widely used. The tendency of presence of industrial robots in
these countries is shown in Fig. 3 [15–21].
Number of units
87.000
China
31.405
38.586
Japan US A
Germany 20.039
Fig. 3. Annual tendency of representation of industrial robots in the period 2010–2016 in four
top countries: China, Japan, USA and Germany
Four countries have been selected: China, Japan, USA and Germany, since Japan,
USA and Germany are among the most developed countries in the world, while China
with implementation of government strategy called “Made in China 2025” aims to
become one of the technologically most developed countries in the world by 2025.
Another reason why these countries are selected for the analysis is that these countries
produce the largest number of vehicles in the world, area in which industrial robots are
most represented. Based on Fig. 3 we can conclude that the first place by the represen‐
tation of industrial robots in production processes is held by China, with an increasing
tendency each year. In period from 2010 to 2016, China has increased the representation
of industrial robots six times, so that in 2016 about 87.000 robots were applied, which
is twice as high as Japan, almost three times more than the USA and four times more
than Germany, and it confirms the fact that their strategy is giving positive results when
industrial production is concerned. The second place by the presence of industrial robots
is held by Japan, with slight increase in the representation that reached 38.586 robot
units in 2016. The third place is held by USA, with the similar increasing trend of use
of industrial robots in the recent years, which in 2016 amounted to about 31.400 robot
units. The fourth place is reserved for Germany, where tendency is constant over the
last few years, and in 2016 it reached about 20.000 industrial robot units. Industrial
robots are installed on those jobs where they can protect the health of workers, assist
them with heavy and monotonous jobs, and gain greater accuracy. The first group of
tasks includes all welding processes, because they are harmful for health, and are largely
used in the automotive industry where automation of welding processes is advanced
[22–28]. For these reasons we have conducted the analysis of the representation of
industrial robots in the welding processes.
3 Automation of Welding Production Processes by Use of Industrial

Robots
Welding procedures: spot welding, arc welding, laser welding, and soldering are part of
procedures that are dangerous for workers to perform because of the dangerous gases
that can damage their health. In addition, these are boring, difficult and monotonous jobs
that should be avoided by the workers. Industrial robots must be used instead, except in
specific tasks when it is not possible to apply robots. In order to obtain the real illustration
of the automation of these jobs in the world, we need to conduct an analysis of the
representation of industrial robots in welding processes. The representation of industrial
robots in welding processes is shown in Fig. 4. The statistical data on the tendency of
representation of industrial robots were obtained from the International Federation of
Robotics (IFR), the UN Economic Commission for Europe (UNECE) and the Organi‐
zation for Economic Co-operation development (OECD) [3, 4, 15–21].
Number of units
665.004
W
World
41.439
Asia/Australia
Americas 8.186
113.154
Europa
Fig. 4. Annual tendency of representation of industrial robots in the welding process worldwide
and on the continents of Asia/Australia, Europe and North America for the period 2010–2016 [15–
21]
As we can see from Fig. 4, the tendency of industrial robot representation in welding
processes in the world is growing in the period 2010–2016, disregarding slight devia‐
tions in 2012. In 2016 about 65.000 industrial robot units were used in the welding
processes. Compared to the total use of industrial robots in the world in the same year,
it can be seen that the representation in welding processes is 25%, representing one
quarter of used industrial robots in all processes worldwide. The first place by the pres‐
ence of industrial robots in welding processes is held by Asia/Australia with growing
tendency each year. In 2016 about 41.400 industrial robot units were used in Asia/
Australia, representing 2/3 of the total industrial robot representation in welding
processes in the world. In the period from 2010 to 2016 the continent of Asia/Australia
has increased the representation of robots in welding processes with total 23.000 robotic
units, i.e. almost 100%. Representation of industrial robots in welding processes is far
behind on the continents of America and Europe, while the continent of Africa was
eliminated in this analysis due to the fact that the representation is negligible in relation
to other three continents. The representation of industrial robots in welding processes
in the North America can be considered as constant each year, and in 2016 it was about
13.154 industrial robots units. On the continent of Europe, the illustration of industrial
robot representation in welding processes is somewhat different, as the industrial robots’
representation in 2010 and 2011 was higher than in the last five years. It can be said that
the representation is constant from year to year (period 2012–2016), and in 2016 there
were about 8.186 robot units. Therefore, we can conclude that the tendency of industrial
robot application in welding processes is to be expected, when compared to the overall
tendency of industrial robot application on the continents, as shown in Fig. 2, and the
application of industrial robots by countries, as illustrated in Fig. 3. The reason for this
conclusion lies in the fact that China, which is located on the continent of Asia/Australia,
is the country with the largest representation of industrial robots in the world, with a
tendency of increasing application in the following years. Higher number of industrial
robots is installed in the automotive industry where welding processes are represented
[29–35]. In order to obtain the real illustration of the tendency of the representation of
industrial robots in automotive industry and how many industrial robots were installed
in welding production processes, we have conducted an analysis of their representation
in this industry in the period 2010–2016, as shown in Fig. 5 [15–21].
Number of units
259.000
World
103.323
Automotive
Welding 65.004
Fig. 5. Annual tendency of the representation of industrial robots worldwide in the automotive
industry and welding processes for the period 2010–2016
The tendency of representation of industrial robots in automotive industry is growing

every year and is accompanied by a growing tendency of the representation of industrial
robots in welding processes. If we compare the first two curves of the diagram in Fig. 5,
we come to the conclusion that about 40% of industrial robots are represented in the
automotive industry each year in relation to the total representation of industrial robots
in all other production processes. In 2016, around 103.323 robot units were used in the
automotive industry, which presents a very high level of representation, while in the
same year the representation of robot units in welding processes amounted to about
65.000 units, accounting for 25% of the total number of industrial robots used in other
processes worldwide. We can analyze the representation of industrial robots by industry
for 2016.
The accurate illustration of the representation of industrial robots in industrial
branches can be seen in Fig. 6(a), which indicates the percentage of the representation
of industrial robots in 2016. We come to the conclusion that the automotive industry
holds the first place with usage of about 39% of industrial robots in 2016, which is
reasonable conclusion, followed by the electric/electronic industry with about 34% of
industrial robots, metal industry with about 11% of industrial robots, plastic and rubber
industry (i.e. chemical) with about 7% of industrial robots, and at the end food industry
with about 3% of industrial robot units. Figure 6(b) shows the representation of industrial
robots in welding processes in relation to the total representation of industrial robots in
the automotive industry in 2016, which is about 103.323 robot units. It is notable that
63% of industrial robots in the automotive industry are used in welding processes,
whereas 37% of industrial robots are used in other processes of the automotive industry.
The production processes of the automotive industry in which car body products are
manufactured use the following welding methods: arc welding, spot welding and laser
welding. The first two processes are far more representative in car body production in
the automotive industry, and therefore the analysis was conducted of the presence of
industrial robots in these processes. So far, the analysis of the representation of industrial
robots in laser welding was omitted, due to the fact that industrial robots are far less
used than in the other two processes.
a – Representation by industry b – Representation in welding compared to

representation in automotive industry
Fig. 6. Percentage of representation of industrial robots in 2016 in industrial branches and

welding processes compared to other processes in the automotive industry in 2016 [15]
The annual representation of industrial robots in the arc welding process and spot
welding process in the period 2010–2016 is shown in Fig. 7. Based on the diagram shown
above, we conclude that the trend of industrial robots’ representation in spot welding is
growing, noticeably in the period 2014–2016, and there is far more robot application in
this welding process compared to the arc welding process. The presence of industrial
robots in the welding process reached about 34.056 units in 2016 in the world. When
we analyze the presence of industrial robots in the arc welding process, we can see that
the representation of robots is smaller compared to the spot welding process. In 2016
the representation of industrial robots in this welding process was around 2.415 units in
the world. In order to obtain a ratio of representation of industrial robots in all welding
processes in the world, we have analyzed the percentage of industrial robots in 2016, as
shown in Fig. 8.
Number of units 65.004
Welding process
34.056
Spot weelding
25.415
Arc welding
Fig. 7. The tendency of annual representation of industrial robots in welding processes in the
world and representation of arc welding and spot welding in the period 2010–2016 [7, 15–21]
Fig. 8. Percentage of representation of industrial robots in 2016 in all welding processes [15]
Based on data [x] in 2016, around 65.004 units of industrial robots were used world‐
wide in various welding processes. Of this number, the largest representation of indus‐
trial robots is in the process of spot welding with about 52%, which presents more than
half of the installed industrial robots. The second place in the representation of use of
industrial robots in welding processes is held by arc welding with about 39%, followed
by the soldering process with about 4%, whereas the smallest representation of industrial
robots is in laser welding process with about 1%. This is reasonable because this is the
latest technology in relation to the processes of arc welding and spot welding, including
more demanding conditions of the application of industrial robots in this welding
process. Finally, there are about 4% of industrial robots in other welding processes. The
welding processes are used in many constructive solutions in metal industry, and
soldering process is used in electronics industry, whereas automotive industry has the
highest representation of industrial robots. In order to prove this conclusion, we have
conducted an analysis of vehicle production in the world, and in the countries, that are
the largest vehicle producers in the world, as shown in Fig. 9.
Numbeer of units Numbeer of units 28.4331.000

x10000 94.997.000
x1000 China
77.584.000
18.2993.000
USA
A
Jaapan
13.2337.000
Germanyy 6.0662.000
a – Vehiclle production worldwide b – Vehicle production by country
Fig. 9. The tendency of annual vehicle production worldwide and in four countries with the
largest vehicle production in the period 2010–2016 [11–14]
The tendency of vehicle production in the world (Fig. 9.a) is growing on annual basis,
and in 2016 94.997.000 units of vehicles were produced. The analysis of vehicle produc‐
tion in the countries with highest vehicles production in the world (Fig. 9.b), namely:
China, USA, Japan and Germany (these countries were selected because of the highest
representation of industrial robots, Fig. 3), has established that China is the world’s
leader in vehicle production with around 28.431.000 cars units produced in 2016. China
is in the first place in the world by application of industrial robots. The second place,
according to the vehicles production in the world, is held by the USA, with annual
increasing tendency in production that in 2016 reached 18.293.000 produced vehicle
units. Japan’s vehicle production positions it in the third place with annual growth, which
in 2016 amounted to 13.237.000 vehicle units produced. Based on Fig. 3, we have noted
that Japan is the second in the world by the representation of industrial robots, followed
the USA in the third place, whereas in the vehicles production the situation is reverse.
The reason is that more industrial robots are installed in the electronic and electric
industry compared to the automotive industry in Japan. The fourth place in world
production of vehicles is held by Germany with continuous tendency, so that in 2016
around 6.062.000 units of industrial robots were produced. As we have stated, the largest
representation of industrial robots is in two welding processes, namely arc welding and
spot welding processes. In order to obtain an overview of the presence of industrial
robots in these two welding processes in the world, an analysis was conducted of their
representation on the continents, as shown in Fig. 10 [15–21].
The analysis of the representation of industrial robots in the arc welding by continents
(Fig. 10.a) indicates that the continent of Asia/Australia is the first place in the world
with increasing tendency of representation. In 2016 there were about 16.805 industrial
robot units applied in Asia/Australia, which is about four times more robot used than in
Europe or North America in the same year. The second place is held by Europe with
much smaller representation than Asia/Australia and a slight annual increasing tendency,
which in 2016 amounted to 4.349 industrial robot units. The third place is held by North
America with similar mild annual growth, which in 2016 reached 4.074 industrial robot
units. The illustration of the representation of industrial robots in the process of spot
welding is different in comparison to arc welding, as shown in Fig. 10(b). We can see
Numberr of units 6.805

16 Numbeer of units
Asia/Au
ustralia 200.621
Asia/Au
ustralia
8..932
Europaa 4.349
North
h America
1.564
North
N America 4..074 Euro
opa
a – Arc weldin
ng b – Spot weldin
ng
Fig. 10. The tendency of annual representation of industrial robots in two welding processes: arc
welding and spot welding by continents Asia/Australia, North America and Europe in the period
2010–2016
that the sudden trend of annual increase of application of industrial robots in Asia/
Australia in 2016 reached about 20.621 robot units, which is extremely high in relation
to other countries when it comes to industrial robots and vehicle production. This is a
consequence of Chinese technological development strategy named “Made in China
2025”. The second place is held by North America with increasing representation of
industrial robots in the spot welding process, that in 2016 amounted to about 8.932
industrial robot units, which is two times more than in arc welding process. Europe is
in the third place with the declining tendency of representation of industrial robots in
the spot welding process, which in 2016 reached 1.564 industrial robot units. The reason
for this reduced representation of industrial robots in spot welding process are manifold,
one of which being that the automotive companies are transferring their production
processes outside of Europe. The representation of industrial robots in arc welding and
spot welding processes in the countries of China, USA, Japan and Germany is shown in
Fig. 11 [8–21].
Numberr of units Numbeer of units 10

10.772
8.088
8.0
C
China China
USA 8.362
USA 3.9
3.948
Jappan
Japan 2.
2.854
Germanyy 3.6
3.687
Germany
1.1
1.157 673
A welding
a – Arc b – Spot weelding
Fig. 11. The tendency of annual representation of industrial robots in the automation of two
welding processes: arc welding and spot welding in China, North America, Japan and Germany
in the period 2010–2016
As shown in Fig. 11(a), China demonstrates growing tendency of use of industrial

robots in the process of arc welding during the period 2010–2016, with a decline in the
representation in the last two years compared to 2014. In 2016 the representation
amounted to 8.088 industrial robot units. With the far less presence of industrial robots
compared to China, USA holds the second place with a slight rising trend that in 2016
reached 3.948 units of industrial robots, which is two times less than in China in the
same year. The third place is held by Japan with almost the same trend as the USA,
which in 2016 reached value of 3.687 industrial robot units. The fourth place by the
representation of industrial robots in the arc welding process is held by Germany, whose
tendency is almost constant in the period 2010–2016, and in 2016 amounted to 1.157
units. Trends of the representation of industrial robots in the process of spot welding are
different in relation to the trends of representation of industrial robots in arc welding
process, as shown in Fig. 11(b). By 2013, the order of application of industrial robots
per country was as follows: USA, China, Japan, and finally Germany. Since 2013, China
holds the first place in the application of robots in the spot welding process with a
growing tendency in the period 2010–2016, so that in 2016 representation of industrial
robots in China reached 10.772 units of industrial robots, which is almost 2,500 more
robots used than in the arc welding process in the same year. USA holds the second
place with increasing annual trend of representation, that in 2016 amounted to 8.362
robot units, which is two times more than in the arc welding process. Japan is holding
the third place with a slight oscillating annual tendency of representation which in 2016
reached about 2.854 industrial robot units in the process of spot welding. The last place
is held by Germany that in 2010 and 2011 held the first place. The representation of
industrial robots in Germany has begun to decline since 2011, so that in 2016 it amounted
to about 673 industrial robot units. In 2010 and 2011, Germany was in the first place in
the tendency of representation of industrial robots in the process of spot welding, but in
the last five years the trend has declined, so that in 2016 it reached its lowest value. There
are many reasons for this tendency [28–35]. One reason is that vehicle manufacturers
and vehicle suppliers in Germany are increasingly investing in electric and hybrid vehi‐
cles, another reason is that automotive companies are organizing production in other
countries around the world, etc.
4 Conclusion
Based on the analysis of the automation of the welding production processes by use of
industrial robots worldwide, and countries with the largest vehicle production, we can
make the following conclusions:
• The tendency of application of industrial robots in the automation of production
processes in the world is continually increasing on an annual and total level, so that
in 2016 annual level reached about 259.000 industrial robot units, whereas overall
level reached about 1.8 million industrial robot units, as shown in the diagrams in
Fig. 1(a), and (b).
• Based on Fig. 2, we can see that the tendency of the automation of production
processes by continents worldwide, tracked by the representation of industrial robots,
is growing from year to year in the period 2010–2016. The first place is held by Asia/
Australia, followed by America and lastly Europe. However, when considering the
tendency of automation of production processes at the overall level of installed
industrial robots, it can be noted that the order is somewhat different. Again, the first
place is held by Asia/Australia, followed by Europe and finally America. Continent
of Asia/Australia has a far greater automation of production processes than Europe
and America.
• In regard to the representation of industrial robots in the automation of production
processes in four top countries, we can conclude that the tendency is constantly
increasing, with China in the first place, followed by Japan, USA and Germany.
Compared to other three countries, China is far more advanced, since it installed
around 82.000 industrial robot units in 2016, like all three countries together.
• The tendency of the automation of welding production processes worldwide is
increasing annually, with continent of Asia/Australia being in the first place, as indi‐
cated by the diagram in Fig. 4. The second place is held by America, followed by
Europe. We have to state that around 65.000 robot units were used in welding
processes in 2016 worldwide, of which 2/3 are related to Asia/Australia.
• The highest representation and application of industrial robots in the world is in the
automotive industry, particularly in the welding processes of the automotive industry,
as indicated in the Fig. 5.
• In regard to the automation of production process in 2016, we can see that the auto‐
motive industry is in the first place with 39% of industrial robots, followed by elec‐
trical/electronic industry with 34%, metal industry with 11%, rubber and plastic
industry with 7% and finally food industry with 3% of overall representation of
industrial robots in the world.
• The highest number of industrial robots is installed in production processes of arc
welding and spot welding, as shown in Fig. 7. The first place is held by spot welding,
followed by arc welding, both indicating growing tendency in the representation of
industrial robots.
• Analysis of Fig. 8 shows that in 2016 the representation of industrial robots in spot
welding process was 52%, followed by arc welding with 39%, soldering process with
4% and finally laser welding with 1%.
• Knowing that the automation of welding processes is mostly used in the automotive
industry, an analysis was conducted on the vehicle production in the world and in
the four top countries where industrial robots were represented the most: China, USA;
Japan and Germany, as indicated in Fig. 9. The tendency of vehicle production has
a growing nature. The first place is held by China with the highest vehicle production
in the past six years. China is also in the first place in terms of number of industrial
robots installed in the automation processes in the industry. The second place is held
by USA, followed by Japan and Germany.
• The analysis of the representation of industrial robots in the automation of arc welding
process by continent has shown that Asia/Australia holds the first place, followed by
Europe and North America. In regard to automation process of spot welding, Asia/
Australia also holds the first place, followed by North America, and decreasing
tendency in Europe, as shown in Fig. 10.
• The tendency of annual representation of industrial robots in the automation of two

welding processes: arc welding and spot welding in China, North America, Japan
and Germany for the period 2010–2016 is shown in Fig. 11. Based on this Figure we
see that China holds the first place and differs from other three countries. In addition,
it was shows that the tendency of representation of industrial robots in spot welding
is decreasing the Germany.
• Finally, we can conclude that the tendency of automation of welding production
processes will have a growing character in the future, in all manufacturing processes.
Such a conclusion is driven by the fact that the world at the moment is in the fourth
industrial revolution that the Germans refer to as “Industry 4.0”. Its implementation
in production processes and in the robotic technology itself will enable new gener‐
ation of industrial robots that can securely work together with man and machine,
without setting up the fences between them. Robots will gain or adapt new skills
through learning processes, and become smarter robots using a great deal of data and
collective learning. They will have simplified applications. Continuous quality
improvement requires sophisticated high-tech robotic systems. Robots improve the
quality of work by taking dangerous, painful, and dirty jobs that are not possible or
safe for people to perform.
References
1. Doleček, V., Karabegović, I.: Robots in the industry. Technical Faculty of Bihac, Bihac,
Bosnia and Herzegovina (2008)
2. Doleček, V., Karabegović, I: Robotics. Technical Faculty of Bihac, Bihac, Bosnia and
Herzegovina (2002)
3. Karabegović, I.: Role of industrial robots in the development of automotive industry in China.
Int. J. Eng. Works 3(12), 92–97 (2016). ISSN 2349-6495, www.kwpublisher.com/?
paper=1-114-The-Role-of-Industrial-Robots-in-the-Development-of-Automotive-Industry-
in-China#Author
4. Karabegović, I., Husak, E.: China as a leading country in the world in automation of
automotive industry manufacturing processes. In: IV International Congress Motor Vehicles
& Motors 2016, MVM-2016, 06–08 October 2016, Kragujevac, Serbia (2016)
5. Karabegović, I., Doleček, V.: The tendency of application of industrial robots in the
automotive, electrical engineering and metal industries worldwide. In: 3rd International
Scientific Conference, COMETa 2016, 07–09 December 2016. University of East Sarajevo,
Bosnia and Herzegovina (2016)
6. Kalpakjian, S., Schmid, S.: Manufacturing Processes for Engineering Materials, 5th edn., pp.
125–150. Amazon, USA (2007). ISBN-13:978-0132272711
7. Sulavik, C., Portnoy, M., Waller, T.: How a new generation of robots is transforming
manufacturing. Manufacturing Institute USA, September 2014, Gaithersburg, USA, pp. 1–
13 (2014)
8. de Luca, A.: Industrijal Robotics. Sapienze, University of Roma, Roma, Italia (2016)
9. McKerrow, P.J.: Introduction to Robotics. Addison-Wesley Publishing Company, Sydney
(1991)
10. Wolka, D.W.: Roboter sisteme. Technishe Universität des Saarlandes im Stadtwald,
Deutschland (1992)
11. Good Jobs in the Age of Automation. BSR-The Business af a Better World, New York, USA,
pp. 1–28 (2015)
12. Verband Deutscher Verkehrsunternehmen VDV: Jahresbericht 2015/2016, Koln, Deutschland
13. Friis, D.: Editorial, Universal Robots, Editorial_WR_Industrial_Robots_2016. www.ifr.org/
news/ifr-press-release/world-robotics-report-2016-832/
14. Bunse, B., Kagermann, H., Wahlster, W.: Industrie 4.0, Smart Manufacturing for the Future,
Germany Trade and Invest, 2016, Berlin, Germany (2016). www.gtai.com
15. World Robotics 2016 (2016). United Nations, New York and Geneva
22. Karabegović, I., Husak, E.: Significance of industrial robots in development of automobile
industry in Europe and the World. J. Mob. Veh. 40(1), 7–16 (2014). University of Kragujevac,
Faculty of Engineering, Kragujevac, Serbia (2014). ISSN 1450-5304
23. Makowieckaja, O.: Industrial robots on the market of means of production automation, pp.
22–27. Biuletyn Instytute Spawalnictwa, No. 2. Gliwice, Czechoslovak (2015). ISSN
0867-583X
24. Karabegović, I., Karabegović, E., Husak, E.: Application analyses of industrial robot in World
automobile industry in 2010. J. Int. Sci. Pub. Mater. Methods Technol. 5(2), 336–345 (2011).
ISSN 1313-2539
25. Institute for Public Policy Research: Technology, Globalization, and the Future of Work in
Europe: Essays on Employment in a Digitized Economy (2015). http://
www.oxfordmartin.ox.ac.uk/downloads/academic/technology-globalisation-future-
ofwork_Mar2015.pdf
26. Mannan, B., Khurana, S.: Enablers and barriers for introduction of robotics as an AMT in the
Indian industrie. Int. J. Comput. Appl. 19–24 (2012). https://www.researchgate.net/
publication/267450380, https://doi.org/10.13140/2.1.2625.5368
27. Ray, J., Atha, K., Francis, E., Dependahl, C., Mulvenon, J., Alderman, D.: China’s industrial
and military robotics development. Research Report Prepared on Behalf of the U.S.-China
Economic and Security Review, Commission, Center for Intelligence Research and Analysis
(CIRA), October 2016, Vienna, Austria, pp. 78–84 (2016)
28. Holdren, J.: National Science and Technology Council Washington, D.C. 20502, 15 February
2012, Washington, USA, pp. 4–27 (2012)
29. Kozul-Wright, R.: Robots and industrialization in developing countries, UNCTAD/
PRESS/PB/2016/6 (No. 50), United Nations Conference on Trade And Development
UNCATD, Geneva, Switzerland, pp. 1–4 (2015). www.unctad.org
30. Paper, W.: Robots fuel the next wave of U.S. productivity and job growth. Association for
Advancing Automation, October 2015, Michigen, USA, pp. 11–19 (2015).
www.a3automate.org
31. Robotics 2020 Strategic Research Agenda for Robotics in Europe, Produced by euRobotics
aisbl, Draft 0v42 11/10/2013, pp. 25–43. http://www.eurobotics-project.eu
32. Yang, G.-Z.: The Next Robotic Industrial Revolution, Manufacturing Robotics, Robotics and
Autonomous Systems (RAS). UK-RAS Manufacturing Review 2015–2016, pp. 2–17 (2016).
ISSN 2398-4422, www.ukras.org
33. Struijk, B.: Robot production volume date trend and analysis, Debreceni Muszaki
Kozlemenyek, vol. 1, pp. 1–10, Debrecen, Hungary (2012). ISSN 2060-6869
34. Asia in Tomorrows World: Technology Unveilsa New Paradigm, 30th Anniversary Issue, Asia
Today International, pp. 1–4, October/November 2013. www.asiatodayinternational.com
35. Roehricht, K.: Study on emerging markets, with special focus on Asia. Robotics Coordination
Action for Europe, KU Leuven, pp. 1–14 (2016)
Modelling of Multiphase Twin Screw Machines
Ahmed Kovacevic1(&), Sham Rane2, and Nikola Stosic1

1
Centre for Compressor Technology,
City, University of London, London EC1V 0HB, UK
a.kovacevic@city.ac.uk
2
Department of Engineering Science, University of Oxford, Oxford, UK
Abstract. Twin screw compressors used in refrigeration, gas and air com-
pression represent approximately 80% of the millions of industrial positive
displacement compressors produced globally each year. More than 95% of these
are oil injected. Oil injected twin screw compressor is just one type of multi-
phase screw machines. Other include twin screw expanders, multiphase pumps
and motors. Multiphase twin screw machines are traditionally analysed and
designed by use of chamber thermodynamic models. However for further
improvement of efficiency and reliability it is necessary to use more advanced
modelling techniques such as 3D Computational Fluid Dynamics (CFD). In
order to obtain fast and accurate solution of multiphase screw machines using
CFD, it is important that a numerical grid of the highest quality is generated
quickly and reliably. For that purpose, a deforming grid of a twin screw machine
is generated using algebraic transfinite interpolation upon which an elliptic
partial differential equations (PDE) of the Poisson’s form are solved numerically
to produce smooth final computational mesh. This paper gives a review of the
current state of art in the application of CFD in modelling of multiphase twin
screw machines including compressors, expanders and pumps. Future chal-
lenges and development trends in application of CFD for multiphase twin screw
machines are also shown in the paper.
Keywords: Screw compressor Computational Fluid Dynamics

Grid generation Conformal grid
1 Introduction
Compressors are used to compress air, refrigerants and various gases in many different
industrial sectors. It is estimated that they consume more than 17% energy produced in
developed countries [19]. This pollutes the environment with more than 3000 MtCO2
per year, while energy costs exceed €275 billion per year [4]. Based on the same
source, it is forecasted that the global CO2 emission will increase by 28% from 2015 to
2030. On the other side, the latest EU targets [5] for 2020 are to reduce the CO2
emissions by 20% from the levels recorded in 1990. This requires at least 20% of
energy to be produced by renewable sources and the increase in energy efficiency of

https://doi.org/10.1007/978-3-319-90893-9_2
Modelling of Multiphase Twin Screw Machines 19
industrial applications by 20% from the levels recorded in 2007. Currently these targets
may not be achieved despite large efforts by both industry and academia to investigate
new and innovative machines.
Industrial initiatives, like waste heat recovery, CO2 capture and storage technolo-
gies, district heat pump heating, the use of clean natural refrigerants, and use of
hydrogen rich gas resources, are recognised to play a major role in achieving the
targets. Oil injected screw machines are used for all these applications and their ability
to efficiently compress or expand gas at high pressures is a pre-requisite. Figure 1
shows typical screw compressor system used for high pressure applications on the left
and a typical oil injected screw compressor on the right.
Fig. 1. Typical screw compressor package for gas industry (left); model of an oil injected screw
compressor (right)
Twin screw compressors are positive displacement machines which make around
80% or all new industrial compressors. It is estimated that since early 1950’s to date
about 11 million screw compressors were put in operation. It is also estimated that in
2016, production of screw compressors was about 700,000 of which large industrial
compressors are about 10% in numbers but they make a large proportion of the
$7.99 billion compressor sales in the same year [12]. Furthermore, it is predicted that
the compound annual growth rate of these machines will be 6.62% in the next few
years and that the screw compressor market will reach the value of over $11 billion in
2021. The highest annual growth of over 8% is expected in the Asia-Pacific region
driven by China, India, and Southeast Asian countries. Although the great majority of
screw compressors in the market are oil injected, the demand for oil free compressed
gas is increasing due to the rise in industries like food, textile, electronics and phar-
maceutical. Therefore, oil injected compressors and other multiphase fluid handling
machines have great potential for improvements in efficiency, reduction in oil charge
and greater contribution to reduce CO2 emission.
Oil is injected into the working chamber of twin screw compressors for three rea-
sons, to cool the gas, to seal the clearances and to lubricate rotors and bearings [16].
Despite the benefits and necessity of oil injection, it’s presence has negative effects. The
shear of oil in clearances contributes to increase of mechanical losses during oil injection
and transport through the chambers. Deipenwisch and Kauder [3] used a mathematical
20 A. Kovacevic et al.
model for predicting the oil

induced losses in twin screw
compressors. Figure 2 shows
typical losses in clearances of an
oil injected screw compressor at
an optimal speed. Therefore it is
useful to control oil injection in
order to optimise oil quantity and
maximise efficiency. 90% of the
oil is injected through the oil
ports that are carefully positioned
in respect to the compression
cycle to allow the desired mass Fig. 2. Power losses due to oil in individual leakage
of oil to enter the compression gaps (Reproduced from Deipenwisch and Kauder [3])
chamber.
In a similar way other screw machines process multiphase fluids. For example an
ORC expander may process two phase refrigerant which changes the phase during the
process but can also be mixed with oil which may add complexity to modelling. Screw
pumps are also often used in multiphase regime to transport both gaseous and liquid
phase simultaneously. In addition to that, screw machines used in oil and gas industry
often admit solid particles which may be detrimental to the performance.
Figure 3 shows an example of the position of oil injection ports in a screw com-
pressor. Commonly, the compressor discharge pressure is utilized to transport and inject
the oil in the working chamber which requires the ports to be adequately positioned in
the system as shown in the figure. However in compressors with variable capacity and
variable built in volume index, oil is often injected under the pressure maintained by an
oil pump. This improves reliability of the machine at extreme working conditions but
comes at the cost of additional power loss. It is essential for the correct operation of a
machine to ensure sufficient oil injection at different operating conditions such as
variable rotor speed or discharge pressure. Additionally, the oil temperature at the point
of injection is important in achieving effective heat transfer. Further understanding of the
Fig. 3. Oil injection ports and injection angle on a compression volume curve
oil distribution in a screw machine can help with minimising oil charge and oil supply
which may be beneficial to the performance of the machine and therefore to saving
energy and environment [17].
Computational Fluid Dynamics (CFD) can provide a useful means of modelling
distribution of oil in a compressor and study the physical phenomenon in detail.
A prerequisite for 3D modelling of multiphase screw machines is the ability to generate
a good structured numerical mesh which could efficiently and accurately utilise
capabilities of general CFD solvers. The pioneering work on development of a
numerical grid for analysis of screw machines was by Kovačević (2002, 2005) who
successfully demonstrated the use of an algebraic grid generation method with
boundary adaptation and transfinite interpolation for analysis of Screw machines. This
method which utilises rotor-to-casing boundary distribution has been also implemented
in the program SCORG, generic grid generator for screw machines. Rotor-to-casing
generation method in which numerical mesh rotates with the rotation of the rotors,
allows excellent representation of the rotor geometry but requires special attention in
CFD solver to handling of the interface between the two rotor meshes. Kovačević et al.
(2007) have reported on CFD simulations of twin screw machines to predict flow, heat
transfer, fluid-structure interaction, etc. In the same work, a test case study of an oil
injected compressor using source terms in transport equations with a segregated
pressure based solver was reported. VandeVoorde et al. [18] used an algorithm for
generating block structured mesh from the solution of the Laplace equation for twin
screw compressors and pumps using differential methods. Reports on analysis of dry
air screw compressors and twin screw expanders with real gas models are available in
literature using these techniques (Papes et al. [13]). Recently, Arjeneh et al. [1] have
presented the analysis of flow through the suction port of a screw compressor with
water injection. It was reported that it was difficult to stabilize the solver in a full 3D
analysis with both deforming rotor domains and multiphase models. Although several
attempts have been made in the recent past to extend the CFD technology to oil injected
compressors, it has proven to be difficult to achieve the desired grid structure and the
modelling conditions that can provide stability to the numerical solvers. Rane [9] in his
thesis proposed a new analytical approach for grid generation that can independently
refine the interlobe region of the screw rotors. It was demonstrated that such grid
refinement improves the prediction of mass flow rates. The same algorithm has been
extended to produce a rotor grid that eliminates the interface between two rotors and
thereby to provide a desirable grid for oil injected or multiphase modelling. This recent
developments has been reported by Rane and Kovacevic in [9, 14]. Such mesh is called
casing-to-rotors conformal mesh as it remains fixed on the casing and slides on the
rotors. This enables much more conservative solution in variety of different CFD
solvers as reported in [9]. However, there is a possibility to improve the quality of such
mesh and further improve the speed and accuracy of a CFD solution by implementing
elliptic numerical smoothing and orthogonalisation of the mesh generated by use of
analytical grid generation. The principles of numerical smoothing and the results of
modelling of an oil injected twin screw compressor are reported in this paper.
2 Multi-phase Modelling in CFD
The complexity of the flow inside a screw compressors makes modelling challenging
since variety of flow regimes can occur within the relatively small working domain.
The oil injection itself starts as a spray. Once it hits the rotor surface a part of oil will
form a film. The film grows and spreads until the injection nozzle is cut-off from the
compression chamber by the rotors trailing lobes. Further rotation causes flooding on
the chamber whose volume is continuously reducing. In order to capture such complex
mechanism, a top level model is required that can account for large changes in oil
volume fraction. The applicable multiphase modelling treatment depends on the phase
coupling effects.
Crowe et al. [2] have presented a detailed characterization of multiphase flow
regimes based on the various coupling effects between the continuous and the dispersed
phases. There are three main methods used in multiphase modelling in CFD to cover
this variety of multiphase flow regimes.
The first model is called Euler-Lagrange which could be applied if the quantity of
the dispersed phase is very low and the phase particles are very fine with negligible
momentum. In that case the compressed gas can be treated as the continuous phase
while oil droplets are treated as particles in the Lagrangian frame which have just one
way coupling. This means, the oil will move with the flow and only heat transfer
between phases will occur. Such a computational model has five transport equations for
the gas phase and one momentum (motion kinematics) equation for the liquid phase
that needs to be conserved separately.
The second approach is the condition which would occur in heavily oil flooded
operation. In addition to the oil droplets, oil film on the rotor and housing will occur
resulting in oil being dragged along with the compressed gas. Under such conditions,
Eulerian – Eulerian treatment for both, the gas and oil phases is required; both can be
continuous phases with energy and momentum interphase transfer effects. Such a
computational model requires significant computational resources and has five trans-
port equations for each phase including momentum, energy and volume fraction. The
pressure field solution obtained from the common continuity equation is shared by the
two phases. The relative slip and shear between the gas and oil is calculated as the
difference between the individual velocities obtained by separate momentum equations.
In describing multiphase flow the concept of phase volume fractions is utilized. The
phase volume fraction represents the space occupied by each phase in a computational
cell in which case it is required that conservation of mass, momentum and energy is
satisfied by each phase individually. This model is called Eulerian – Eulerian inho-
mogeneous model. This is regarded to be the most accurate treatment for any multi-
phase flow regime and is most likely required for oil injected screw compressors.
The third model is a simplification of the Eulerian – Eulerian approach. It can be
applied to flows which have negligible slip between two phases, for example a mul-
tiphase flow with fluids of similar physical properties. This Euler – Euler homogeneous
multiphase treatment shares a common pressure solution as well as velocity field for
both phases. Such model requires five equations for the continuous phase and an
additional concentration equation for distributed phases. More on this could be found in
[8]. There are several sub variants of this method which could be used in modelling of
screw machines in different applications providing the interaction of phases can be
modelled appropriately.
In the presented work an inhomogeneous Eulerian – Eulerian treatment is used.
3 Grid Generation for Oil Injected Screw Rotors

3.1 Algebraic Grid Generation of an Initial Mesh
The analytical grid generation of screw machine working domain is explained in
Kovačević et al. [8]. It includes separating domains of a screw machine rotors by a rack
curve (Stošić [17]) and forming independent flow domains around each of the rotors.
After the grid points are distributed on boundaries, an initial grid is obtained by
analytical transfinite interpolation. This process applies to both rotor-to-casing and
casing-to-rotor grid generation approach. In order to achieve a conformal single domain
mesh, a new approach of background blocking was introduces by Rane and Kovacevic
[10, 15]. In this procedure, the outer boundary in each background block of a course
analytically generated mesh is defined as a combination of the rack segment and the
casing circle segment. The rack segment stretches between the bottom and top cusp
points and is closed by the casing. The distribution obtained on the outer boundaries of
the two blocks is used as a reference for rotor profile distribution as shown in Fig. 4a.
Fig. 4. Background blocking for regular boundary discretization and conformal algebraic mesh.
Figure 4b shows how the rack curve is used to partition the two rotor domains and
the boundary distribution so obtained is used to generate the 2D mesh using transfinite
interpolation. Using the blocking approach allows both conformal and non-conformal
boundary map to produce fully hexahedral 3D grid. Figure 4c shows the nodes on the
rack segment between the main and the gate rotor grids with conformal boundary
map. The 3D mesh generated from such 2D cross sections allows the rotor domains of
the male and female rotors to be combined into a single rotor mesh. This avoids inac-
curacies and instabilities that may arise due to the interface mismatch in a non-conformal
boundary map. The resultant grids are recommended for oil injected multiphase flow
modelling and have been described in more detail by Rane and Kovacevic in [10, 15].
However, even with this approach, during the operation as the rotors rotate, the rack
curve comes to a position when it transitions with a relatively large deformation between
the two consecutive steps. The current algebraic method results in a transition as a step
change at certain positions as shown in Fig. 6 left. One of the objectives of the elliptic
PDE mesh generation presented in this paper was to improve the time transition of the
partitioning rack curve between the two rotor domains.
3.2 Applying the Elliptic PDE Solver to Improve Algebraic Grid

Techniques based on solutions of Partial Differential Equations (PDE) to define
coordinate transformation are widely used in grid generation. The idea of using elliptic
PDE like Laplace equation or Poisson equation is based on the work of Crowley and
Winslow, and it is in detail described in Knupp and Steinberg [11]. Elliptic PDE’s have
certain beneficial properties in their solution that make them preferable for body fitted
curvilinear grids. These are less prone to folding of mesh lines, introduce inherent
smoothness so that the discontinuities over the boundaries are not propagated into the
interior of the domains and the physical boundaries can be exactly used as boundary
conditions in the computational space.
But the numerical grid generation is computationally expensive without an initial
grid. If an initial grid based on algebraic method could be used, the required time for
the solution of PDE’s is significantly reduced.
This treatment requires the solution of the coupled PDE equations in which
schemes like Tri-diagonal Matrix Algorithm can be used for the solution [11]. The
boundary conditions are specified as grid coordinates at computational boundaries.
This means that for the generation of grid in twin screw rotor domain the coordinates
(x, y) of the boundary nodes need to be used as boundary condition. At this stage the
initial grid generated by the algebraic transfinite interpolation is used for both initial
condition and boundary conditions. In the presented work, an O grid topology pro-
duced by algebraic transfinite interpolation shown in Fig. 4 has been used by the
differential solver with successive-over-relaxation procedures as described in Knupp
and Steinberg [11]. In addition, the convenient input parameters are used to control
intensity of smoothing and inflation layer formation of the elliptic solver. The final O
grids generated separately for the two rotors by the differential solver are merged to
produce a single domain mesh for the two rotors as shown in Fig. 5 – Final PDE Mesh.
3.3 Smoothing the Transition of the Boundary Between the Rotor

Domains
The PDE solver described in Sect. 3.2 is also used in the interlobe area bounded by the
cusp radial nodes (Fig. 4b) to convert the rack curve into a smooth transitioning curve
across the specified number of angular positions of the rotor. This is obtained over the
four steps to gradually change the partition between the two O grids into a smooth one
as shown in Fig. 5.
O Grid TFI Mesh

Conformal Algebraic Conformal Distribution II
Interlobe PDE Mesh Final PDE Mesh
Boundary Distribution I on the Smooth Rack
Smooth Rack
Fig. 5. Four-step procedure for improving time transition of the rack curve.
The smooth rack obtained

by this procedure is supplied
back to a second stage of
boundary distribution calcula-
tion resulting in a new con-
formal distribution. This
conformal distribution is fur-
ther used as a boundary con-
dition for final differential
Fig. 6. Improvement in the 3D rack surface. mesh generation. As a result a
significant improvement in the
mesh quality is achieved.
This four-step procedure is very quick and the time required for calculation of a
representative 2D mesh with 5000 nodes, across 50 rotor positions is less than 15 s.
Figure 6 left shows the original interface surface between the two rotors. Figure 6 right
shows the smooth rack surface generated using the new procedure with differential
smoothing. The sharp sections observed with the original surface are eliminated and a
completely smooth surface on the boundary between the two rotors was achieved.
3.4 Improvements in Grid Quality

The described method contributes to the significant improvement in mesh quality.
Figure 7 shows the comparison of the cell orthogonal quality between the algebraic
meshes and the elliptic meshes for the compressor with 4 lobes on the male and 5 lobes
on the female rotor. Figure 7(a–d) are algebraic meshes and Fig. 7(e–h) are elliptic
meshes in the respective rotor positions.
The values for the minimum orthogonal angle of the algebraic meshes (top) drops
to about 8° after the rack curve transitions from position b to c. The majority of cells are
in the range of 40–60° orthogonality. Low orthogonality values are also noticed in the
position d.
In case of an elliptic smoothing, the overall orthogonality has greatly improved so that
the minimum orthogonal angle is 25°. Most of the cells are in the range of 75–90°
Fig. 7. Comparison of cell orthogonal quality between Algebraic (a–d) and Elliptic (e–f)
meshing.
orthogonality. There is one cell at the bottom cusp which shows low orthogonality of
about 15° in both sets of meshes which is the consequence of the discontinuity at the cusp
point and it cannot be avoided. However the overall mesh quality is greatly improved.
4 Case Study CFD Model of Oil Injected Compressor
In order to test the new mesh, the test compressor used for this study is an oil flooded twin
screw compressor of 4/5 lobe combination, ‘N’ rotor profile, the combined axial/radial
suction port and an axial discharge ports. The male rotor diameter is 127 mm and the wrap
angle is 306°. The built-in volume index Vi of the compressor is 3.6. The nominal
interlobe, radial and axial leakage gaps are 60 µm. The end clearances are not included in
the CFD model. The operating speed of the machine is between 3000 rpm and 6000 rpm,
discharge pressure can vary between 4.0 bar and 12.0 bar. The compressor is driven by a
75 kW motor with a variable frequency drive. The discharge pressure is controlled by the
discharge valve and the gas flow rate is measured using an orifice in the discharge line. In
the current setup oil is injected using the discharge line pressure and oil temperature is
controlled using a water cooled heat exchanger.
Measurements at the specified operating conditions were carried out at the air
compressor test rig at City, University of London shown in Fig. 8 right and the results
have been used in this paper for comparison with the predictions from CFD model.
Figure 8 left shows the CFD domain that consists of the rotors, suction and discharge
ports. Both ports have an extended pipe to provide boundary pressure conditions far
from the rotor domains. The oil is injected through four injection holes that are con-
nected to the rotor domain by the fluid-fluid interfaces. The rotor domain itself is
represented by the deforming structured conformal grid, generated using the procedure
described in the Sect. 3. Three operating conditions have been calculated, listed in
Table 1.
Fig. 8. CFD Model (left) and the test rig (right) of the oil injected compressor
Table 1. Operating conditions for comparrison of CFD and measurements

Case Speed Suction Suction gas Discharge Oil injection Oil injection
(rpm) pressure (bar) temperature (K) pressure pressure (bar) temperature (K)
(bar)
1 3000 1.00 298.0 6.0 5.5 323.0
2 3000 1.00 298.0 8.0 7.5 323.0
3 6000 1.00 298.0 8.0 7.5 323.0
The case is solved using commercial Ansys CFX solver with Eulerian – Eulerian
multiphase model as described in Sect. 2. The continuous phase is air defined as an
ideal gas with the molar mass of 28.96 kg/kmol, specific heat capacity 1004.4 J/kg K,
dynamic viscosity 1.831e-05 kg/ms and thermal conductivity 0.0261 W/m K. Oil has a
constant density of 800 kg/m3, specific heat capacity 1800 J/kg K, dynamic viscosity
8.8e-05 kg/ms and thermal conductivity 0.18 W/m K. Dynamic viscosity of oil is
normally function of temperature but in this case it was assumed constant. Interphase
heat transfer has been defined by assuming an interfacial length scale of 5.0 µm and Nu
equal to 50. The same value of the interfacial length scale is also used for calculation of
interphase slip. The solver was set with the first order upwind advection scheme, first
order transient scheme and relaxation factors of the order of 15%. The turbulence
model SST – k Omega was used.
The discharge and oil injection pressures have been gradually increased over more
than 4 full rotations of the main rotor to enable stable further calculation. Calculations
were continued until the flow and pressure pulsations showed cyclic repetitions. Mass
imbalance between the discharge and suction for air was below 9% and for oil it was
about 11% in all three cases. This imbalance could be due to the still short time to
analyse the system starting from the ‘zero’ values of all properties set at the beginning
of the process, it can be due to the grid quality, lower order discretization schemes used
and the convergence level achieved which was in the range of 5e−4. Each case took
around 350 h to calculate 15 full rotations on a 6 node PC with 3 GHz processor and
32 GB RAM. The model mesh statistics is presented in Table 2 along with the cell
Table 2. Mesh statistics and quality report

Domain Cell Node Cell Orthogonality Expansion Aspect
structure count count angle (Min) factor ratio
Rotor Hexahedral 468677 406368 7.4 646 488
Suction port Tetra + Hex 119058 203255 30.2 279 9
Discharge port Tetra + Hex 98521 253095 19.6 53 28
Oil injection port Hexahedral 28340 25144 55.7 4 4
quality report. The orthogonality factor below value of 10.0 in less than 1% cells. Both,
the expansion factor and the aspect ratio were in good control due to the fewer divi-
sions in the radial direction of the rotor as well as the absence of a boundary layer mesh
in the ports with tetrahedral grids.
4.1 Integral Parameters

Figure 9 shows a comparison of the normalized mass flow rate of air and the nor-
malized indicated power between measurements and CFD results for the three cases.
Measured/flow and power at Case 1 conditions of 3000 rpm, 6.0 bar discharge pressure
have been used here for normalization.
Fig. 9. Comparison of flow and indicated power between experimental and CFD results
The gas flow predictions are generally few percent lower than measured values.
However, indicated power is few percent higher than the measured value. This is most
likely due to the constant multiplication factor which was used for estimating
mechanical losses and due to use of a relative coarse mesh.
4.2 Oil Distribution

The pulsating nature of the oil injection is shown in Fig. 10 right. The instantaneous oil
flow rate is aggregated for all four injection ports. As expected, the oil flow rate at
6.0 bar is lower than at 8.0 bar discharge but it does not vary much with the rotational
speed. Figure 10 left shows the distribution of oil and its interaction in the discharge
Fig. 10. Oil flow distribution near the discharge port and instantaneous flow rates of oil injection
port. Iso-surface of the oil volume fraction 0.1 shows an accumulation of oil in the rotor
tips and also high concentration on the discharge port walls. Similarly oil accumulates
in the interlobe gaps and is transported by the rotors. The oil distribution in the
discharge port is unsteady but cyclically repeating.
4.3 Temperature Distribution

Figure 11 shows the instantaneous discharge air temperature for three analysed cases.
Without oil injection, the adiabatic discharge temperature would be in the region of 500 K
but due to leakages and recompression effects, the temperature would be much higher.
With the injected oil at 323 K, the gas temperature is retained low and it peaks around
340 K. As expected, the temperature at 3000 rpm and 8 bar discharge, the peak tem-
perature is some 10° lower than that at 6000 rpm. Similarly there is only a small difference
in the gas temperature at the same speed and different discharge pressures. Figure 11 also
shows the distribution of gas temperature at 8.0 bar discharge pressure, 6000 rpm with an
iso-surface of oil volume fraction 0.1. The local maximum temperature in the chamber
reaches 360 K due to over compression. The gas temperature is not uniform across the
working domain due to the distribution of oil in the compression chamber.
Fig. 11. Discharge gas temperature cycles for three cases and distribution at 6000 rpm, 8.0 bar
pressure.
5 Development Trends and Future Work
This methodology to generate numerical mesh for screw machines originates from the
pioneering work of Kovacevic [7] in the late 1990. At the beginning of 2000 it allowed
the use of CFD in screw machines for the first time. Since then this methodology of
grid generation and CFD calculations has been further developed to account for the
phenomena like multiphase flows, large pressure gradients, fluid-solid interaction and
small clearances in the machine. It could be applied to most of currently available CFD
solvers like ANSYS CFX, ANSYS Fluent, Star CCM+, Pumplinx etc. As a conse-
quence this methodology is today often used in industry.
Apart of the authors of this paper, few more researchers [14, 18] have developed
similar methods with some variations in the procedure of generating numerical mesh
for CFD analysis of screw machines. All these methods are based on the same
decomposition of the flow domain and similar methodology of grid generation. The
major difference is that they do not use fast and reliable analytical methods as the
starting point for final grid generation. Therefore, they are faced with the issues related
to speed and ease of calculation, limitations of the solvers that could be used and the
applicability of this model to some types of screw machines such as single screw or
twin screw with large helix angles.
These two issues are currently being addressed by the authors in the series of Short
courses and forums held at City, University of London since 2011 [6]. There are
several possibilities to allow improvements in the speed of calculation and making it
more suitable for industry. Using faster CFD solvers could help, for example segre-
gated pressure based solvers which are in general faster than coupled solvers. However,
even more promising solution is to use initial and boundary conditions for CFD from
integrated chamber model. This could be applied both for CFD and for structural
analysis and it could significantly reduce time for obtaining converged solution. This
method can also be extended to single screw machines and machines with large helix
angles like vacuum pumps, but novel approach is required.
One of the currently investigated modifications by the authors is the use of normal
rack for separation of domains of the rotors.
And finally, validation of CFD results for multiphase flows is a big challenge due to
difficulties in obtaining experimental results. This will certainly be activity that will be
reported in publications in near future.
6 Conclusions
This paper presents developments of the grid generation techniques which allow fast and
accurate CFD modelling of multiphase screw machines. A single domain structured
numerical mesh of the flow domain was generated by SCORGTM using recently
developed boundary blocking, analytical grid generation and elliptical smoothing. An oil
injected screw compressor was measured and integral parameters such as flow rate of gas
and indicated power were used for the comparison with CFD results.
• Test analysis showed a close match in the prediction of the mass flow rates of gas.
• The indicated power obtained by CFD predictions matched well with the measured
shaft power.
• The mixing of the phases, distribution of oil, heat transfer between gas and oil and
also effects on sealing due to high oil concentration in leakage gaps were well
captured.
• Interaction of oil injection with the rotors, deposition on the rotor surface, discharge
port walls and heavy accumulation in the interlobe region was observed.
This demonstrates enhances capabilities of the newly developed mesh. Future work
will aim to full validation of the model for other multiphase screw machines. Also, this
will be used to improve oil injection systems in screw compressors with the aim for
reducing oil charge and improving efficiency.
References
1. Arjeneh, M., Kovačević, A., Gavaises, M., Rane, S.: Study of multiphase flow at the suction
of screw compressor. In: Proceedings of the International Compressor Conference at Purdue,
Paper 1353 (2014)
2. Crowe, T.C.: Multiphase Flow Handbook. Taylor and Francis, CRC Press (2006). ISBN
0-8493-1280-9
3. Deipenwisch, R., Kauder, K.: Oil as a design parameter in screw type compressors: oil
distribution and power losses caused by oil in the working chamber of a screw type
compressor. In: IMECHE Transactions, International Conference on Compressors and Their
Systems, London, vol. 6, pp. 49–58 (1999)
4. Enerdata: Global Energy Statistical Yearbook (2017). http://yearbook.enerdata.net/. Accessed
19 Feb 2018
5. European Commission: 2020 Climate and Energy Package policy (2014). http://ec.europa.
eu/clima/policies/strategies/2020/index_en.htm. Accessed 19 Feb 2018
6. Kovacevic (2017). https://www.city.ac.uk/compressorsconference/2017-conference/short-
course-on-compressors. Accessed 5 Mar 2018
7. Kovačević, A.: Three-dimensional numerical analysis for flow prediction in positive
displacement screw machines. Thesis, City University London (2002)
8. Kovačević, A.: Boundary adaptation in grid generation for CFD analysis of screw
compressors. Int. J. Numer. Methods Eng. 64, 401–426 (2005)
9. Kovačević, A., Stošić, N., Smith, I.K.: Screw Compressors - Three Dimensional
Computational Fluid Dynamics and Solid Fluid Interaction. Springer, Berlin (2007). ISBN
3-540-36302-5
10. Kovacevic, A., Rane, S.: Algebraic generation of single domain computational grid for twin
screw machines Part II – validation. Adv. Eng. Softw. 107 (2017). https://doi.org/10.1016/j.
advengsoft.2017.03.001
11. Knupp, P., Steinberg, S.: The Fundamentals of Grid Generation. CRC Press (2002). ISBN
9780849389870
12. Markets and Markets: Screw Compressor Marker (2017). https://www.marketsandmarkets.
com/Market-Reports/screw-compressor-market-122896617.html. Accessed 19 Feb 2018
13. Papes, I., Degroote, J., Vierendeels, J.: 3D CFD analysis of an oil injected twin screw
expander. In: Proceedings of the ASME 2013 International Mechanical Engineering
Congress and Exposition, ASME IMECE 2013, San Diego, USA (2013)
14. Rane, S.: Grid generation and CFD analysis of variable geometry screw machines. Thesis,
City University London (2015)
15. Rane, S., Kovacevic, A.: Algebraic generation of single domain computational grid for twin
screw machines. Part I. Implementation. Adv. Eng. Softw. 107, 38–50 (2017). https://doi.
org/10.1016/j.advengsoft.2017.02.003
16. Stošić, N., Milutinovic, L., Hanjalić, K., Kovačević, A.: Investigation of the influence of oil
injection upon the screw compressor working process. Int. J. Refrig. 15(4), 206–220 (1992)
17. Stošić, N., Smith, I.K., Kovačević, A.: Screw Compressors: Mathematical Modeling and
Performance Calculation, Monograph. Springer, Berlin (2005). ISBN 3-540-24275-9
18. VoordeVande, J., Vierendeels, J.: A grid manipulation algorithm for ALE calculations in
screw compressors. In: 17th AIAA Computational Fluid Dynamics Conference, Canada,
AIAA 2005-4701 (2005)
19. Wide, P., Brunner, C.: Energy efficiency policy opportunities for electric motor driven
systems. International Energy Agency (2011). http://www.iea.org/publications/
freepublications/publication/EE_for_ElectricSystems.pdf. Accessed 19 Feb 2018
Influence of Injection Molding Process
Parameters on Part Quality
Janez Gotlih ✉ , Miran Brezočnik, Igor Drstvenšek,

( )
Timi Karner, and Karl Gotlih

University of Maribor, 2000 Maribor, Slovenia
janez.gotlih@um.si
Abstract. A case study to determine the influence of injection molding process

parameters on part quality is performed. Part quality is evaluated based on deflec‐
tion rate, volumetric shrinkage, and sink mark depth results. Considered process
parameters are mold temperature, melt temperature, injection time, pack time,
pack pressure and cooling time. A thin walled butterfly valve flap and a thick
walled butterfly valve shaft are analyzed separately. Polypropylene is used as
molding material. A design of experiment based on the Taguchi method is gener‐
ated for both parts and a computer simulation to solve the case matrixes is
executed. Relative process parameter influence on each individual quality criteria
is determined by analysis of variance. Importance of process temperatures as well
as parameter impact dependence on characteristic part shape is reveled.
Keywords: Injection molding · Quality prediction · Parameter study

Design of experiment · Taguchi method · ANOVA
1 Introduction
To investigate the influence of injection molding (IM) process parameters on part

quality, considering characteristic shape features, a design of experiment case study
analysis was performed. For design of experiment studies of IM the Taguchi method
was proposed by many authors. IM part quality based on dimensional criteria such as
volumetric shrinkage, warpage and sink mark depth was improved by Taguchi method
and analysis of variance in [1, 2]. The relationship between cooling setup and part quality
was considered in [3] and a characteristic thin wall shaped part was studied by adapting
the cooling system layout [4]. Influence of conformal cooling channel design on part
quality was studied in [5, 6]. Additive manufacturing technology for conformal cooling
channel mold production was proposed and a computer simulated flow analysis to
predict cycle time reduction and part quality improvement was performed in [7]. A cheap
spiral cooling channel production method was proposed and an algorithm for spiral
channel design was developed in [8]. Micro IM process parameter tuning in order to
understand their characteristics was performed in [9]. Influence of IM process parame‐
ters on part shrinkage and warpage was measured by a 3D scanner and the molding
process was observed through acoustic emission and in-cavity pressure in [10]. Complex
part quality control with 3D scanning technique was proposed in [11]. Part thickness

https://doi.org/10.1007/978-3-319-90893-9_3
34 J. Gotlih et al.
was proven to have an important impact on micro surface structure replication quality
at IM [12]. The goal of this study was to determine which process parameter has the
greatest impact on part quality, independent of characteristic shape features, to help
decide on future optimization targets.
2 Methods
A thin walled butterfly valve flap and a thick walled butterfly valve shaft, both part of
the same butterfly valve assembly were chosen for this study. Selection of two different
geometries has the purpose to learn whether and to what extent characteristic geometry
features such as thin walls and thick walls influence injection molding process parameter
impact on final part quality. Each of the two chosen parts was considered in a separate
case study as process parameters like injection time and cooling time strongly depend
on part shape and size.
The study was performed with the help of an IM computer simulation. Two basic
elements were required for a successful calculation, a 3D mesh model of each part geom‐
etry and a finite element method (FEM) model for each considered geometry. In our case,
a butterfly valve computer aided design geometry was downloaded from a free online
database. Flap part and shaft part of the butterfly valve were separately imported into
Autodesk Moldflow Synergy 2016, where both meshes and all FEM calculation cases were
prepared. For case study calculation, Autodesk Moldflow Insight 2016 was used.
A 3D tetrahedral mesh was created for both CAD part geometries with minimum six
elements through part thickness and a 1.2 ratio of maximum tetra edge length in thickness
direction relative to global edge length. Global edge length on surface was set to 0.98 mm
for flap and to 0.53 mm for shaft. A higher mesh density is required for the shaft part
because of a more complex shape.
Fig. 1. Flap part (a) and shaft part (b) 3D tetrahedral mesh with injection location
Influence of Injection Molding Process Parameters 35
The 3D tetrahedral mesh for flap part geometry consisting of 122579 elements with
visible injection location, scale and coordinate system in isometric view is shown in
Fig. 1a and the 3D tetrahedral mesh for shaft part geometry consisting of 1009430
elements with visible injection location, scale and coordinate system in isometric view
is shown in Fig. 1b.
To set up a design of experiment (DOE) study based on the Taguchi method an
orthogonal case matrix for each part was required. In our case, six independent process
parameters were used as variables with two variation levels each, resulting in two eight
case study matrixes. The selected process parameters were mold temperature, melt
temperature, injection time, pack time, cooling time and pack pressure.
To generate the orthogonal Taguchi matrix an initial value for each process param‐
eter was determined and a variation multiplier was selected to symmetrically raise and
lower the initial value. Allowed mold and melt temperature variation range depends on
molding material. In our case, the selected material was polypropylene (PP). The recom‐
mended processing temperature value and the allowed maximum and minimum
processing temperature values were gathered from material manufacturer data sheet. In
addition, the criteria for cooling time calculation was determined based on material data
sheet. For PP, minimum ejection temperature is 124 °C.
Table 1. Case matrix for flap part

Case Mol. T. Mel. T. Inj. T. P. T. P. P. C. T.
[°C] [°C] [s] [s] [MPa] [s]
1 30 200 0.8 8 3.04 16
2 30 240 0.8 12 4.56 16
3 30 200 1.2 8 4.56 24
4 30 240 1.2 12 3.04 24
5 70 200 1.2 12 3.04 16
6 70 240 1.2 8 4.56 16
7 70 200 0.8 12 4.56 24
8 70 240 0.8 8 3.04 24
Based on recommended processing data for PP an initial mold temperature of 50 °C

was selected. 20 °C were added to and subtracted from the initial value to generate two
case matrix parameter values and to remain inside the recommended temperature range.
For melt temperature initial value of 220 °C was selected and a 20 °C variation constant
was applied in the same way as in case of mold temperature. For both parts, identical
initial temperature values were used, as recommended processing values for tempera‐
tures are only material dependent not geometry dependent. A preliminary fill sequence
simulation for each part determined the initial value for injection time separately. For
flap part calculated injection time was 0.1 s and for shaft part 0.6 s. A 20% variation
multiplier was applied in both cases. Initial value for pack time was selected as 10 s
according to industry best practices and a 20% multiplier was used for both parts. Initial
pack pressure values were calculated as 80% of the maximum fill pressure and varied
for 20% to generate case matrix values. Maximum fill pressure was extracted from a
36 J. Gotlih et al.
preliminary flow sequence simulation as 4.75 MPa for flap and 3.29 MPa for shaft
geometry. Initial value for cooling time was extracted from a preliminary cooling anal‐
ysis. For cooling time criteria to be fulfilled, the complete part geometry needs to reach
ejection temperature. In case of flap part 20 s and in case of shaft part 320 s was the
minimum required cooling time calculated by the simulation. The initial values for
cooling time were varied for 20% for both part geometries to generate case matrix values.
Table 2. Case matrix for shaft part

Case Mol. T. Mel. T. Inj. T. P. T. P. P. C. T.
[°C] [°C] [s] [s] [MPa] [s]
1 30 200 0.8 8 2.11 256
2 30 240 0.8 12 3.16 256
3 30 200 1.2 8 3.16 384
4 30 240 1.2 12 2.11 384
5 70 200 1.2 12 2.11 256
6 70 240 1.2 8 3.16 256
7 70 200 0.8 12 3.16 384
8 70 240 0.8 8 2.11 384
The orthogonal case matrix for flap is presented in Table 1 and for shaft in Table 2.
Mol. T. stands for mold temperature, Mel. T. stands for melt temperature, Inj. T. stands
for injection time, P. T. stands for pack time, P. P. stands for pack pressure and C. T.
stands for cooling time.
Each case in the flap and the shaft matrix was calculated individually using fill, pack
and warp analysis sequence. Fill sequence enables variation of mold temperature, melt
temperature, injection and cooling time and generates sink mark depth and volumetric
shrinkage results. Pack sequence enables variation of pack time and pack pressure. Warp
calculation sequence was used to generate results for deflection effects.
Table 3. Results of case matrix for flap part

Case S. M. Vol. S. D. E. X D. E. Y D. E. Z
[mm] [%] [mm] [mm] [mm]
1 0.0160 1.89 0.145 0.156 0.0134
2 0.0151 1.69 0.137 0.147 0.0126
3 0.0155 1.66 0.136 0.146 0.0123
4 0.0153 2.07 0.149 0.161 0.0141
5 0.0219 1.27 0.176 0.188 0.0121
6 0.0227 1.90 0.209 0.223 0.0146
7 0.0215 1.16 0.169 0.181 0.0114
8 0.0223 2.36 0.216 0.232 0.0153
Part quality was evaluated based on five quality criteria, sink mark depth, volumetric
shrinkage and deflection effects in x, y and z direction. A part is considered highest
quality when the maximum calculated sink mark depth is smallest. This means that the
case with the smallest calculated value for maximum sink mark depth is considered best.
The same is true for volumetric shrinkage and for deflection rate results, where smallest
values represented highest quality. Analysis of variance was used to determine relative
variable impact on quality criteria. Based on simulation results, every process variable
was analyzed for each quality criteria.
Simulation results of all calculated cases for flap part are presented in Table 3, where
S. M. stands for sink mark depth, Vol. S. stands for volumetric shrinkage and D. E. X,
Y, Z stand for deflection effects in x, y and z direction.
Results presented in Table 3 reveal, that in case of flap part, mold temperature has
a dominant effect on sink mark depth. Higher mold temperatures increase maximum
sink mark depth. The difference between lowest and highest observed value is 33.48%.
In case of volumetric shrinkage the difference between lowest and highest observed
value is 50.85%. A combination of high melt temperature, low pack time and low pack
pressure has a positive impact on volumetric shrinkage. Mold temperature has a close
to dominant impact on deflection in x and y direction. All results are similar for deflection
in x and y direction, which is the result of an almost symmetrical flap shape in x and y
direction. On deflection in z direction, melt temperature and pack time have dominant
impact, which is the result of flat flap shape in z direction. Deflection values in z direction
are also more than ten times smaller than in x or y direction. The greatest difference
between the lowest and the highest deflection in x and y direction is 37.07% and in z
direction 25.49%.
Relative influences of IM process parameters on each quality criteria for flap part
geometry are presented in Table 4. Sink mark depth has a linear dependency on mold
temperature with 82.10% impact, followed by pack time with 8.87% impact and cooling
time with 3.58% impact. A nonlinear behavior is observed for volumetric shrinkage
where the most influential variable is melt temperature with 33.12% impact, followed
by pack time with 26.40% and pack pressure with 19.05%. Mold temperature, cooling
time and injection time have less than 10% impact on volumetric shrinkage. Deflection
effects in x and y direction are strongly influenced by mold temperature, with over 49%
impact, followed by melt temperature with over 21% impact and pack time with over
18% impact. Deflection effects in z direction depend on melt temperature with 39.67%
impact, followed by pack time with 29.06% and pack pressure with 21.35% impact.
Table 4. Influence of variables on flap part quality

S. M. Vol. S. D. E. X D. E. Y D. E. Z
[%] [%] [%] [%] [%]
Mol. T. 82.10 9.80 49.74 49.13 5.32
Mel. T. 1.53 33.12 21.07 21.25 39.67
Inj. T. 1.74 3.35 0.67 0.57 1.98
P. T. 8.87 26.40 18.77 18.34 29.06
C. T. 3.58 8.28 0.83 1.40 2.63
P. P. 2.19 19.05 8.91 9.32 21.35
Results presented in Table 5 reveal, that in case of shaft part, melt temperature has
a dominant effect on sink mark depth. Higher melt temperature increases sink mark
38 J. Gotlih et al.
depth. The difference between lowest and highest observed value is 17.51%. In case of
volumetric shrinkage the difference between lowest and highest observed value is
18.32%. High mold temperature and low melt temperature have a positive impact on
volumetric shrinkage. Melt temperature has a dominant impact on deflection effects in
x, y and z direction. There are no close similarities between deflection values in x, y and
z direction as observed in case of flap part. The greatest difference between the lowest
and the highest deflection in x direction is 24.46%, in y direction 18.27% and in z direc‐
tion 18.06%. Differences between highest and lowest values for all quality criteria are
lower for shaft part than for flap part.
Table 5. Results of case matrix for shaft

Case S. M. Vol. S. D. E. X D. E. Y D. E. Z
[mm] [%] [mm] [mm] [mm]
1 1.51 6.65 0.952 0.166 0.201
2 1.75 7.26 1.150 0.191 0.216
3 1.46 6.66 0.914 0.161 0.186
4 1.76 7.25 1.150 0.194 0.227
5 1.52 5.93 1.010 0.169 0.200
6 1.76 6.71 1.150 0.192 0.211
7 1.52 6.00 0.987 0.168 0.196
8 1.77 6.53 1.210 0.197 0.212
Table 6. Influence of variables on shaft quality

S. M. Vol. S. D. E. X D. E. Y D. E. Z
[%] [%] [%] [%] [%]
Mol. T. 6.57 46.41 15.90 8.63 7.21
Mel. T. 79.34 43.61 63.24 71.96 51.39
Inj. T. 2.90 1.84 4.76 4.64 0.37
P. T. 2.99 1.61 5.57 4.22 17.82
C. T. 5.82 4.59 10.38 10.23 18.20
P. P. 2.38 1.94 0.15 0.32 5.00
Relative influences of IM process parameters on each quality criteria for shaft part
geometry are presented in Table 6. Sink mark depth has a linear dependency on melt
temperature with 79.34% impact, followed by mold temperature with 6.57% impact and
cooling time with 5.82% impact. For volumetric shrinkage the most influential variables
are mold temperature with 46.41% impact, followed by melt temperature with 43.61%
impact and cooling time with 4.59% impact. Deflection effects are strongly influenced
by melt temperature with 63.24% impact in x, 71.96% in y and 51.39% in z direction.
In x direction the second most influential parameter is mold temperature with 15.90%
impact followed by cooling time with 10.38% impact. In y direction the second most
influential parameter is cooling time with 10.23% impact followed by mold temperature
with 8.63% impact. In z direction the second most influential parameter is cooling time
with 18.20% impact followed by pack time with 17.82% impact. Pack pressure, injection
time and pack time have least impact on deflection effects in case of shaft part.
For flap and for shaft part all variable influences are presented in Table 7. For flap,
mold temperature variable has the greatest impact on part quality with 40.87% share,
followed by melt temperature with 22.96%, pack time with 19.94%, pack pressure with
11.71%, cooling time with 3.03% and injection time with 1.50% share. In case of shaft,
melt temperature variation has the greatest impact on part quality with 61.94% share,
followed by mold temperature with 16.94%, pack pressure with 9.84%, pack time with
6.44% injection time with 2.90% and cooling time with 1.96% share. Injection time and
cooling time have the weakest impact on part quality regardless of part shape.
Table 7. Overall variables influence

Overall variable impact [%]
Flap Shaft
Mold temperature 39.22 16.94
Melt temperature 23.33 61.91
Injection time 1.66 2.90
Pack time 20.29 6.44
Pack pressure 12.16 9.84
Cooling time 3.34 1.96
3 Conclusion
Based on simulation results a shape independent guideline for IM part quality optimi‐
zation can be made. A general conclusion is that the two dominant parameters for both
parts are mold temperature and melt temperature, followed by pack time and pack pres‐
sure. To achieve better IM part quality, these four parameters should be considered as
primary optimization parameters.
Our future research will concentrate on injection molding process optimization with
focus on the cooling phase, as mold temperature was found to be a crucial process
parameter considering part quality.
References
1. Kc, B., Faruk, O., Agnelli, J.A.M., Leao, A.L., Tjong, J., Sain, M.: Sisal-glass fiber hybrid
biocomposite: optimization of injection molding parameters using Taguchi method for
reducing shrinkage. Compos. A 83, 152–159 (2016)
2. Xie, X.C., Lin, C.G., Jia, C.C., Cao, R.J.: Effects of process parameters on quality of ultrafine
WC/12Co injection molded compacts. Int. J. Refract Metal Hard Mater. 48, 305–311 (2015)
3. Sánchez, R., Aisa, J., Martinez, A., Mercado, D.: On the relationship between cooling setup
and warpage in injection molding. Measurement 45, 1051–1056 (2012)
4. Nian, S.-C., Wu, C.-Y., Huang, M.-S.: Warpage control of thin-walled injection molding
using local mold temperatures. Int. Commun. Heat Mass Transf. 61, 102–110 (2015)
40 J. Gotlih et al.
5. Shayfull, Z., Sharif, S., Zain, A.M., Saad, R.M., Fairuz, M.A.: Warpage analysis with straight
drilled and conformal cooling channels on front panel housing by using Taguchi method. Key
Eng. Mater. 594/595, 593 (2014)
6. Rahim, S.Z.A., Sharif, S., Mohd Saad, R., Nasir, S.M., Zain, A.M.: Improving the quality
and productivity of molded parts with a new design of conformal cooling channels for the
injection molding process. Adv. Polym. Technol. 35 (2016)
7. Hsu, F.H., Wang, K., Huang, C.T., Chang, R.Y.: Investigation on conformal cooling system
design in injection molding. Adv. Prod. Eng. Manage. 8, 107–115 (2013)
8. Wang, Y., Yu, K.-M., Wang, C.C.L.: Spiral and conformal cooling in plastic injection
molding. Comput. Aided Des. 63, 1–11 (2015)
9. Packianather, M., Griffiths, C., Kadir, W.: Micro injection moulding process parameter
tuning. Procedia CIRP 33, 400–405 (2015)
10. Kusić, D., Kek, T., Slabe, J.M., Svečko, R., Grum, J.: Test method: the impact of process
parameters on test specimen deviations and their correlation with AE signals captured during
the injection moulding cycle. Polym. Testing 32, 583–593 (2013)
11. Brajlih, T., Tasic, T., Drstvensek, I., Valentan, B., Hadzistevic, M., Pogacar, V., et al.:
Possibilities of using three-dimensional optical scanning in complex geometrical inspection.
Strojniski Vestnik/J. Mech. Eng. 57, 826–833 (2011)
12. Masato, D., Sorgato, M., Lucchetta, G.: Analysis of the influence of part thickness on the
replication of micro-structured surfaces by injection molding. Mater. Des. 95, 219–224 (2016)
Numerical Analysis of Material Fatigue Impact
on Bicycle Frame Safety in Accordance
with EN 14764
Nermina Zaimović-Uzunović ✉ , Ernad Bešlagić, and Almir Porča

( )
University of Zenica, 72000 Zenica, Bosnia and Herzegovina

{nzaimovic,ebeslagic}@mf.unze.ba, porca.almir@gmail.com
Abstract. Bicycles represent the most widespread means of transport today. The
basis for a large bike usage is its relatively simple construction. There are many
different types and shapes of bicycles on the market. In order to protect buyers
and users from low quality and unstable bicycles, standards have been developed
that prescribe minimum safety requirements and test methods before placing the
bicycle on the market.
The European standard for urban and recreational bicycles has the designation
EN 14764 and prescribes, inter alia, the procedure for testing the impact of mate‐
rial fatigue on the safety of the bicycle frame.
The study of how the impact of material fatigue affects the appearance of
bicycle crack is carried out using numerical simulations. Simulations were
performed on a concrete sample of BASMAN’s bicycle manufacturer GS-TMT
Travnik.
The research carried out has shown that numerical simulations can help to
determine the direction in which the designer creates a solution meeting the
requirements of the standards at the very beginning of the bicycle design.
The results of the numerical simulations have shown what needs to be done
to make the analysed bicycle frame fulfilled the test requirements prescribed by
EN 14764. Analysis also showed the concentration of stresses in certain zones on
the examined bicycle frame, which coincide with the point at which a cracking
occurred at one used bicycle.
Keywords: Bicycle frame · Fatigue · EN 14764 · Numerical simulation
1 Introduction
Bicycles represent the most widespread means of transportation today. The basis of the large
use of the bike is its relatively simple construction. There are many different types and
shapes on the market. In order to protect customers and users from poor quality and unsafe
bicycles, standards that prescribe minimum safety requirements and test methods before
placing a bicycle on the market have been developed.
The company “GS-TvornicaMašinaTravnik”, among other products, produces frames
for the so-called cruiser bicycle (Fig. 1.a). The specific shape of these frames leads to the
occurrence of a concentration of stress on certain parts of it, which could cause frame

https://doi.org/10.1007/978-3-319-90893-9_4
42 N. Zaimović-Uzunović et al.
failure. The appearance of a crack on a single frame (Fig. 1.b) is what led to perform numer‐
ical security test in compliance with the standards pertaining to this area.
Fig. 1. (a) Bike frame BASMAN; (b) the crack appeared on the frame
2 Standards for Cycles
The International Organization for Standardization (ISO) has a special technical committee
for cycles, TC149. The scope of this committee is “Standardization in the field of cycles,
their components and accessories with particular reference to terminology, testing methods
and requirements for performance and safety, and interchangeability”. The European
Committee for Standardization (CEN) also has a specific Technical Committee, TC333, that
defines European standards for cycles. One of the main tasks of TC333 is to harmonize CEN
standards for bicycles with ISO standards. Some CEN cycle standards were developed
before ISO published their standards, leading to strong European influence in this area.
European cycle standards tend to describe minimum safety requirements, while ISO stand‐
ards have harmonized parts geometry [5].
2.1 Standard EN 14764
A bicycle whose frame was analysed in this paper belong to a group of city bicycles, so tests
are performed according to the standard EN 14764-City and trekking bicycles-Safety
requirements and test methods. This standard has been revised several times since its first
publication (2005 and 2009).After alignment with ISO standard which treats the same issue,
the standard was named “EN ISO 4210 – Bicycles-Safety requirements for bicycles”. The
standard EN ISO 4210 actually unifies standards: EN 14764, EN 14765, EN 14766 and EN
14781. The most recent version is EN ISO 4210:2016. The Institute for Standardization of
Bosnia and Herzegovina has adopted this standard and currently valid version in our country
is BAS EN ISO 4210:2016.
Numerical Analysis of Material Fatigue Impact on Bicycle Frame Safety 43
3 Bicycle Frame Fatigue Test According to Standard EN 14764
The frame strength test is performed as static load test, shock load test and fatigue test. The
frame is tested either as a single part or assembly comprise of frame and front-fork.
All types of frame have to be imposed to fatigue test. During test according to EN 14764
cracks or fractures must not be visible in any part of the frame and suspension system [2].
3.1 Test Method
A new frame/fork assembly fitted with standard head-tube bearings should be used for the
test. The front fork may be replaced by a dummy fork of the same length and at least the
same stiffness as the original fork. Before testing the frame, the assembly should be fixed to
the base as shown in Fig. 2 with the fork or dummy fork secured by its axle to a rigid mount
of height Rw (the radius of the wheel/tire assembly ±30 mm) and with the hub free to swivel
on the axle. The rear drop-outs should be secured by means of the axle to a stiff, vertical link
of the same height as that of the front, rigid mount. The upper connection of the link should
be free to swivel about the axis of the axle but must provide rigidity in a lateral plane. The
lower end of the link should be fitted with a ball-joint.
Fig. 2. Frame: fatigue test with pedalling forces [2] (Key: Rw-height of rigid mount and
vertical link, Rc-length of vertical arm (75 mm), L-length of crank replacement (150 mm), 1-
Rigid mount, 2-Vertical link, 3-Ball-joint, 4-Adaptor assembly, 5-Vertical arm, 6-Tie rod, 7-
Center-line of tie-rod)
If an adaptor assembly is used (Fig. 2), it should be ensured that the assembly is free to
swivel about the axis of the bottom-bracket and that both replacement arms are 175 mm long
(L) and that they are both inclined forwards and downwards at an angle 45° (accurate ±0,5°)
to the horizontal. The position of crank replacement arms is defined by a vertical arm (which
replaces the chain-wheel) and a tie-rod which has ball-joints at both ends and which is
attached to the rear axle perpendicular to the axis of the rear axle. The length of vertical arm
(Rc) shall be 75 mm and the axis of the tie-rod should be parallel to on 50 mm distance from
the vertical plane through the centre-line of the frame. Each pedal-spindle (or equivalent
adaptor component) is subjected to a repeated downward force of 1.000 N at a position
150 mm from the centre-line of the frame in a vertical, transverse plane and inclined at 7,5°
(accurate to within ±0,5°) to the fore/aft plane of the frame. During application of these test
forces, it should be ensured that the force on a “pedal-spindle” falls to 5% or less of the peak
force before commencing application of the test force to the other “pedal-spindle” [2].
During the test, 100.000 test cycles should be applied, where one test cycle consists of the
application and removal of the two test forces.
4 Creating of 3D Models for Frame Fatigue Test
To perform the test in accordance with the described standard procedure, SolidWorks soft‐
ware was used to create 3D model of the bicycle frame and required parts. The 3D model
of the bicycle frame is created based on the STEP file obtained from the producer, while the
other parts are modelled according to Fig. 3.
Fig. 3. Created 3D models of: (a) bicycle frame; (b) frames testing facilities
Certain simplifications have been made in the modelling of bike frame, which will
not affect the analysis results.
5 Static Analysis of Bicycle Frame
SolidWorks is a 3D CAD Software that represents all-around solution for product develop‐
ment, design and creating 3D documentation. SolidWorks Simulation package is a tool used
for analysing model through numerical method of finite elements and analysis.
In the pre-processing phase, the material from which the frame is made and the way in
which it is fixed, are defined. The force is applied and an appropriate mesh is created. The
frame is made of construction steel S235JR. The model is fixed according to the standard.
Although the standard prescribes otherwise, contacts between parts of testing facilities were
defined as “bonded”. The exception is only made with contacts between vertical arm and
frame. Contacts has been defined as “no penetration”. Performed analysis confirmed that
these simplifications do not affect the results, while the calculation time has been shortened.
5.1 Meshing–Creating Finite Element Model

Bicycle frame is quite demanding from the aspect of adequate mesh choosing. Shell
elements for all tubes and solid elements for three plates have been used on the first place.
Such combination of elements requires that frame is modelled as multibody part. The
results have shown that this combination does not give good results because the occur‐
rence of stress concentration at the tips of the plate (Fig. 4).
Fig. 4. The occurrence of stress concentration at the tips of the plate
Stress concentration can be avoided by creation rounded pass from the plate to the tube.
Since SolidWorks does not allow object rounding between separated particles of one part,
the down frame tube and the plate were connected, creating a round pass between them.
The solid element usage for down tube has increased the total number of finite elements
and extended the simulation. Because of that, the mesh was refined in the contact zone of
the plate and the tube, while the rest of the model was mashed with coarse mesh (Fig. 5).
The total element number for this type of mesh is around 700.000.
Fig. 5. Looks of adopted mesh and overview of the results obtained
The results of the analysis with static force on left and right side of the frame have
shown that maximal stress significantly exceeds the yield strength of material from which
the frame was made. Performed analyses used thickness of tube wall from 1,5 to 4,0 mm.
The results are shown in Table 1.
Table 1. The results of the static analysis frame for the different wall thickness of the lower tube
Wall thickness Name of study Number of Stress (von Deformation
[mm] element Mises) [N/m2] [mm]
1,5 Static_Left 707.410 6,790·108 9,982
Static_Right 5,419·108 10,210
2,0 Static_Left 722.612 8 7,930
4,228·10
Static_Right 8 7,918
3,564·10
2,5 Static_Left 772.022 8 6,668
2,963·10
2,712·10
3,0 Static_Left 809.968 8 5,885
2,439·10
2,410·10
3,5 Static_Left 821.265 8 5,537
2,455·10
2,439·10
4,0 Static_Left 842.048 8 4,984
2,471·10
2,449·10
Even after extending the thickness, the stress value did not go under 235 MPa, which
is yield strength of construction steel S235JR. The tube breaking in the zone around
plate can be solved by choosing another material with better mechanical characteristics.
Assuming that stresses and deformations under the load depend on elastic module E and
Poisson’s ratio ν selection of construction steel E360 as replacement material did not
require new numerical calculations. If the down tube of frame is made of steel E360 and
wall thickness is 2,5 mm then the frame will not break because its yield strength is
360 MPa (Fig. 6).
Fig. 6. Stress plot when force was applied on the right side of the frame: (a) complete bike frame;
(b) zone with maximum stress
6 Material Fatigue Impact on Bicycle Frame Safety
SolidWorks Simulation provides a fast, reliable, and cost-effective way to predict and
resolve fatigue-related problems before they happen. Within fatigue analysis, it’s also
possible to carry out analysis for case when more loads are applied to the structure.
Loads may be dependent or independent of each other. Force can be considered as
dependent because of never act at the same time. Limiting factor for application SW
Simulation in analysing case is deficiency of SN-Curves for the materials from which
the frame was made. SolidWorks derive the SN-curve automatically based on ASME
SN curves and the modulus of elasticity of the reference material and the active material.
The SN curve is derived by dividing each stress value of the reference SN curve by the
modulus of elasticity of the reference ASME material and multiplying it by the modulus
of elasticity of the current material. The associated numbers of cycles remain unchanged.
Since there is no SN-Curve for the steel E360 in the material database, simulations
were performed with SN-curves derived from the ASME austenitic and ASME carbon
steels. The following figures show the results obtained.
If the SN-Curve is derived from the ASME carbon steel, obtained results indicate
that the frame will not withstand 100.000 cycles. The most loaded zone of the frame can
withstand only 15.560 cycles before the fracture occurs. This is 6,42 times less than the
required 100.000 cycles. This means that the desired load cycles are 642% higher than
the maximum possible number of cycles which the frame can withstand before fracture
occurs (Fig. 7). The situation is different if the SN-Curve is derived from the ASME
austenitic steel. In this case the most loaded zone of the frame can withstand 285.292
cycles which far exceeds the desired 100.000 cycles. More precisely, the desired number
of cycles is only 35% of the maximum number of cycles which the weakest zone of
frame can withstand (Fig. 8).
Fig. 7. SN-Curves based on the ASME carbon steel: (a) damage percent for complete frame; (b)
the number of cycles that can withstand the frame before failure
Fig. 8. SN-Curves based on the ASME austenitic steel: (a) damage percent for complete frame;
(b) the number of cycles that can withstand the frame before failure
7 Conclusion
The frame of the BASMAN bicycle has been tested to the causes and consequences and
results are specified in this paper. Due to forces acting on the pedal axle, a stress concen‐
tration in the connection zone of the down tube frame and the plate exceeds the yield
stress of the material from which the bicycle frame was made. Such high stress values
are results both the shape of the plate and the shape of the down tube.
By increasing the wall thickness of the down tube, the desired effect is not achieved.
Although the maximum stress has decreased, it has not yet reached a value that is less
than the yield strength of the tube material. According to that, increasing only the wall
thickness of the tube, does not lead to the solution of the problem.
The maximum stress value at a tube of 1.5 mm thick is so much so that the problem
cannot be solved by choosing better quality material. There are no steels to meet these
requirements to be found in an acceptable price range. Therefore, a compromise solution
was made: the tube thickness was increased to 2.5 mm and the chosen steel with better
mechanical properties. The performed analyses showed the allowed stresses in the crit‐
ical zone of the frame would not be exceeded. The negative effects are increased total
weight of a bicycle by about 5% (which is acceptable) and the total price of a bicycle.
Further testing to obtain more convenient weight/price combination can be made after
the bicycle producer declares which of these two factors are more important.
After the fatigue test has been conducted no cracks in the material appeared. Material
and thickness of the wall tube were correctly selected. However, these results should be
taken with a reserve, because the test was not performed based on the SN-curve of the
adopted material, but on the replacement SN-curve available in the SolidWorks material
database. By using an adequate SN-curve results could be more reliable. Therefore, in
further research, it is necessary to determine the SN-Curves for materials S235JR and
E360, and check the effect of the fatigue to the frame safety.
References
1. Zaimović-Uzunović, N., Lemeš, S.: Metodkonačnihelemenata “Dom štampe” Zenica (2002).

ISBN 9958-42-079-1
2. European standard EN 14764:2005 - City and trekking bicycles – Safety requirements and test
methods, European committee for standardization (2005)
3. Dassault systems, SolidWorks Simulation (2011)
4. http://help.solidworks.com/2017/english/solidworks/cworks/c_Element_Types.htm.
Accessed 10 Feb 2018
5. https://www.iso.org/committee/53030.html. Accessed 10 Feb 2018
The Influence of the Tool Geometry
on the Quality of the Weld in FSW Process
Aleksandra Koprivica1 ✉ , Nikola Šibalić2, Milan Vukčević2,

( )
and Mirjana Jokanović1

1
University of East Sarajevo, East Sarajevo, Bosnia and Herzegovina
aleksandra.koprivica@ymail.com,
jokanovic.mirjana1@gmail.com
2
University of Montenegro, Podgorica, Montenegro
{nikola,milanvu}@ac.me
Abstract. There are several friction welding procedures that have, to a greater or
lesser extent, a commercial application, among which can be distinguished FSW -
Friction Stir Welding. Firstly, using the FSW procedure, straight-line compounds
were obtained, and soon afterwards also curvilinear, including circular. In addition
to the great interest in the FSW process, many aspects of this process have not been
sufficiently explored. This applies, inter alia, to the geometry of the tool. In this
paper, the application of FSW welding to aluminium alloys is investigated, by
varying different types of tool pin used during welding. Also, microstructure testing
was performed, which was used for comparison and gradation of the quality of
welds, welded with tools of different geometric shapes.
Keywords: FSW Method · Microstructure · Pin of tool
1 Introduction
FSW - Friction Stir Welding or friction welding is considered as one of the most successful
unconventional material mixing technologies used in recent decades. By means of this
method, welding of sheet metal of larger thicknesses, from different materials, was facili‐
tated [1].
The application of this welding technology is very wide and dynamic. It is now used
in shipbuilding, airline industry, railway wagon production, automotive industry, etc., or
where welding of plate or profiled elements of large lengths is welded, so that it has been
forced out of use by conventional welding processes.
This technology is ecological, of high energy efficiency and is not harmful to the envi‐
ronment. Unlike conventional technologies, FSW consumes significantly less energy and
does not use the atmosphere of a protective gas. The connection is done without the use of
additional material, and therefore the connection of a wide range of various alloys can be
achieved without the fear that the problem of inhomogeneity will occur. One of the most
important features of the FSW process, which distinguishes this technology from conven‐
tional methods, is to achieve better mechanical properties of the welded joint.

https://doi.org/10.1007/978-3-319-90893-9_5
The Influence of the Tool Geometry on the Quality of the Weld 51
For this technological process, it can rightly be said that it is one of the most environ‐
mentally friendly cleaner processes and is increasingly used for welding of same kind and
various materials, and most of all for the welding of aluminum alloys. The asymmetric
nature of the metal seams and the unusual shape of the tools have caused that, as a result,
the welded joint has a very characteristic microstructure. In this connection, it can be
noticed that various shapes of the tool pin affect the microstructure of the welded joint [2].
2 Geometry of Tools
Tool geometry is the most influential factor in the FSW welding process. The special tool
used for FSW welding, consists of two parts, from the body and workpiece of the tool. The
working part of the tool consists of a shoulder and a pin, which can have different design
shapes. The forehead of the tool can have concentric recesses on its surface, while the tool
trunking is usually cylindrical or conical, profiled with different coils and grooves. The tool
body serves to attach the tool to the jaw of the machine on which the weld is made and its
shape does not affect the welding process [3]. According to the appearance of the pin, the
friction welding tools differ from a simple cylindrical flat-bottom spindle to a complex cone
pinned shaped with coils and grooves on the side surfaces. Accordingly, the pin of the tool
can be cylindrical or conical, with a coil or without a coil, with flat or profile grooves on the
sides or without grooves. Thus, the cross section of the mandrel can be a circle, triangle,
square, pentagon, hexagon, or other shape. The shaft allows the material to be shifted from
the front of the tool to the rear, while mixing the welding material, where the heat is gener‐
ated. The height of the tool pin corresponds to the thickness of the welded material, but is
always less than the total thickness of the welding material [3]. In Fig. 1, a schema of the
tool intended for FSW welding is shown.
Fig. 1. Schema of tool for FSW [4]
The shape of the tool affects the movement of soft material, and various tools can
result in different mixing of the soft material between the working pieces in the joint.
Tool dimensions influence the amount of heat generated, the axial and longitudinal
force of the tool on the base material, the moment and the thermomechanical load acting
52 A. Koprivica et al.
on the tool. More important dimensions in tool making are the diameter of the shoulder, and
the diameter and height of the tool pin [3].
3 Microstructural Characteristics
In literary sources, there are several terms describing microstructural changes in the FSW
process. The first attempts at the microstructural division of the welded joint were made in
1997 for aluminum alloys only. The names of microstructural zones, adopted by
consensus, and which include microstructural divisions of other metals, came very soon.
The microstructural division consisting of the four zones shown in Fig. 2, was carried out
at the TWI Institute in the United Kingdom [4].
Fig. 2. Structure of the welded joint in the FSW process A - Zone of the base material, B - Heat
Affect Zone - HAZ, C - Thermo-Mechanicaly Affect Zone - TMAZ, D - Stirred Zoneor “Nugget
Zone” - NZ [4]
In the base material zone there is no plastic deformation of the material nor the influ‐
ence of heat, and hence, there are no microstructural changes.
Heat affect zone, as the name implies, it is exposed to the effect of heat, but it does not
deform during welding, but passes through the appropriate temperature cycle which leads
to changes in the microstructure. The size of this zone depends primarily on the heat input.
In the Thermo-Mechanicaly Affect Zone, the material is exposed to plastic deforma‐
tions and temperature problems, by tools, which occur on both sides. With aluminum in
this zone, the areas can be spotted: the so-called the “nugget” zone, the area outside the
“nugget” zone and the area of the “nugget” zone of the gloves. The area outside the
“nugget” zone is to a lesser degree deformed and, depending on the alloy, does not have to
be recrystallized. The zone of mixing (the so-called “nugget” dynamically recrystallized
zone) is the zone of large material deformation. Grains in this zone are roughly mixed and
often of a size smaller than the size of the grain in the base material area. The size of the
grain depends on the type of alloy and the welding conditions, ranging from 1 μm to
10 μm [4].
4 Experimental Research
FSW welding, aluminum plates, the chemical composition of which is shown in

Table 1, dimensions 200 mm × 100 mm × 10 mm, was carried out at the Laboratory of
the Faculty of Mechanical Engineering in Podgorica. The experiment was performed

on a vertical milling machine, under laboratory conditions of welding, which are similar
to production conditions.
Table 1. Chemical composition of Al – plate

Chemical composition
%Al %Fe %Si %Ti %Cu %Zn %V %Cr %Mn %Mg %Na
99,55 0,33 0,08 0,004 0,001 0,003 0,009 0,003 0,003 0,015 0,0001
For experimental research, welding by the FSW procedure, a set of different tools
has been adopted. The basic tool used in the research is given in Fig. 3. This tool is also
made of two parts, a working part and a tool body, with the tool body adapted to the
jaws of the machine on which welding is performed. The tool material used in the
experiment is tool steel 90MnV8 (Č.3840), which after processing, was subjected to
thermal treatment to obtain a hardness of 59 HRC.
Fig. 3. The basic type of welding tool with conical shape of the pin
On order to test the impact of the geometry of the tool pin on the quality of the welded
surface, during the welding, five types of tools (from each tool of two copies) were used,
the basic tool and four tools in which the working part of the tool (pin) was modified.
In Fig. 4, the tools used in the experiment are shown.
Fig. 4. Tools designed to perform the experiment

On Fig. 5, the pattern of cutting samples for metallographic testing was adopted.
Samples are 40 mm × 10 mm × 10 mm. The procedure for taking samples from a welded
joint is given in Table 2.
Fig. 5. Adoption scheme for sample cutting
Table 2. Scheme of sampling from welded joints

Tool I II III IV V
Join 1 2 1 2 1 2 1 2 1 2
Sample a b a b a b a b a b a b a b a b a b a b
The cutting of samples was done on a modern CNC machine (Wather Jet). Since, after
cutting, the sample has a rough and uneven surface, the alignment of the roughness is made
using a file. After rough treatment, the samples were finely processed in the “Swisslion”
Industrija alata a.d. Trebinje. In order to obtain fine, smooth surfaces, the samples were
brushed on a waterproofing grinder, 400, 600, and 1200 sanding papers, and then polished
on the polishing device, thereby achieving the effect of mirrors. The chemical treatment
was carried out by treating the polished surface for 20 min with a solution of the following
composition: 2 ml of 40% HF (Flavored Water Acid), 200 ml of distilled water.
To record the microstructure of the welded joint, the microscope “Lecia EZ4 HD” with
the software package “Lecia LAS EZ”, located at the Faculty of Medicine in Foča, was
used. This microscope offers magnification of 5, 10, 20, 30, 50, and 100 times images
illuminated by LEDs and also features an integral digital camera offering fast data transfer.
5 The Results of Experimental Researches
The microstructure was recorded on twenty samples, obtained using the five tools, with
which welding was performed twice (10 welds), and two tubes (a, b) were separated from
each weld (a total of twenty samples).
Microstructural analysis was carried out in certain characteristic positions. Figure 6
shows the macrostructure of individual samples, and Fig. 6a shows a sample of the micro‐
structure of sample III1b, where in the position 1, the boundary between the “nugget” zone
and the TMAZ zone is observed. Also, the zones of the basic material, marked with the
BM mark, are clearly visible. In Figs. 6b–d, microsystems of the FSW process for
individual aluminum alloy samples are presented. In these samples, different microstruc‐
tural zones have been identified: the base material (BM), the Heat Affect Zone (HAZ), the
thermo-mechanical zone (TMAZ) and the “nugget” zone (NZ - Nugget Zone). In Figs. 6c
and d, on the right, a “flash” is also seen, or a place of lower hardness value, while in
Figs. 6b and c, a “tunnel” effect is clearly visible, expressed to a greater extent. The
“tunnel” effect is characterized as a negative phenomenon and is created as a result of
insufficient material flow through the tool pin to the DC side, that is, it occurs when the
tool is progressing at a high speed.
Fig. 6. Recordings of microstructure of samples (a) III1b, (b) II2a, (c) V1a, (d) IV2b
On the samples obtained by welding with the tool III (the pin is profiled with a coil and
two sides of the cut parallel axis of the cone) and the tool IV (the pin is profiled with a coil
and three straight sides engraved in the parallel axis of the cone), no “tunnel” effect. On all
other samples, cut from the welded joint of all other tools, the appearance of the “tunnel”
effect is clearly visible, which is less pronounced on the sample II2a, i.e. a tool with a pin
shaped with a coil. The more pronounced occurrence of this effect is on the welds obtained
by the first tool, i.e. basic tool. The most pronounced phenomenon of “tunnel” effect is in
the fifth tool, i.e. a tool with a trapezoidal thread with three semicircular grooves cut
parallel to the cone axis. In Fig. 7, samples are shown showing no “tunnel” effects and
patterns in which this phenomenon is most pronounced.
Fig. 7. Samples without and with the appearance of a “tunnel” effect

6 Conclusion
The largest number of studies of the application of the FSW procedure, as well as the
microstructural changes of the welded joint, were done on welding tools of the basic
conical shape, due to the sample, easy design and construction. The FSW process can
be significantly improved by designing tools of different shapes.
Based on experimental research and observation of the welded surface of the tested
samples, different indicators were obtained regarding the welded microstructure.
Analyzing, it can be concluded that the best quality welded surface in terms of the
microstructure is given by the tools in which the pin is in the form of a coil with two
and three straight grooves engraved in the parallel axis of the cone. There are no visible
traces of the “tunnel” effect in these compounds.
On all other samples, including those samples that were not shown in the previous
images, the effect appeared to a lesser or greater extent. This negative phenomenon can
be avoided only by correctly selecting the tool geometry and the kinematic parameters
of the welding process.
References
1. Koprivica, A., Vukčević, M., Spaić, O., Jokanović, M.: The influence of tool geometry on the
microstructure of aluminum welded by FSW method. In: KODIP 2017, Budva, pp. 93–99
(2017). ISBN 978-9940-527-51-8
2. Vukčević, M., Šibalić, N., Spaić, O., Koprivica, A.: Tool geometry for the FSW process. In:
KODIP 2015, Budva, pp. 169–175 (2015). ISBN 978-9940-669-01-0
3. Šibalić, N., Vukčević, M., Damjanović, M., Koprivica, A.: Examination of microstructure of
welded joint of similar materials, obtained by FSW method. In: International Conference of
Materials Processing Technology, MAPT 2017, Bangkok, Thailand, pp. 164–169 (2017).
ISBN 978-974-456-785-7
4. Koprivica, A.: Welding aluminum alloys using the FSW procedure. Final Work on the Second
Cycle of Studies, University of East Sarajevo, Faculty of Production and Management
Trebinje, Trebinje (2017)
Dimension Measurement of Injection
Moulded Toybricks
Samir Lemeš ✉ and Anel Baručija

( )

slemes@unze.ba, anel_barucija@hotmail.com
Abstract. Polymer toybricks are available in different shapes and sizes, and their
price varies significantly. The aim of this paper is to test whether high price of
some brands of toybricks corresponds to their quality, in terms of dimension
accuracy. For this purpose, two kinds of toybricks were tested: one made by the
renowned LEGO® company, the other model made by anonymous Chinese
manufacturer. In addition to visual differences, such as colour stability and
finishing quality, the difference is in quality and accuracy of shapes and dimen‐
sions that can affect the characteristics of the bricks, such as: quality and stable
connection with other bricks in the assembly, lifetime etc. The main characteristic
tested here is the dimension tolerance of specific dimensions of toybricks.
Keywords: Quality · Tolerance · Coordinate measuring machine

LEGO® toybricks
1 Introduction
In order to estimate the relation between the production costs and the geometrical toler‐
ances, a number of researches were performed. Some of them analysed the manufac‐
turing aspects, some investigated the material properties, and some were focused on
reverse engineering technologies in order to test the geometry of manufactured parts.
Ebro and Howard in [1] investigated variation reduction principles (VRP) that can
be used to increase the robustness of a product and reduce its variation in functional
performance. Singh and Bettig in [2] evaluated and compared different schemes for
capturing the attributes of assembly interfaces and appending that information to solid
CAD models. Bici et al. in [3] used Reverse Engineering and Computer Aided technol‐
ogies to improve the inspection of injection moulded electro-mechanical parts. They
used a laser scanner installed on a CMM machine in order to measure components with
lengths in the range of 5–250 mm, and special issue they had to deal with were the
measurement algorithms for cylindrical features, in order to have the same level of reli‐
ability obtained in the planar case. Bici et al. in [4] used Computer Aided Tolerancing
and Inspection (CAT&I) methods based on Reverse Engineering techniques to test an
aeronautical component made by beta-forging of Ti6Al4V Titanium Alloy powders in
order to test the implementation oriented to non-planar surface recognition. Turc et al.
in [5] presented a methodology to estimate the mould cost, taking into consideration the
part geometry and its dimensional prescriptions.

https://doi.org/10.1007/978-3-319-90893-9_6
58 S. Lemeš and A. Baručija
However, in addition to real product quality, the price of the final product largely
depends on market status and implementation of copyright protection. In other words,
the brand name sells the product, and our intention is to test whether the key factor for
market price of injection moulded toy bricks is branding and copyright, or is it the real
product quality.
The product quality in this research was observed through dimension tolerances of
toybricks manufactured by the famous LEGO® company and by no-name imitation
made in China.
2 LEGO® Toybricks
LEGO® is famous Danish creator and manufacturer of children’s toys. The name
LEGO® is a combination of the Danish words Leg and Godt, meaning “play nice”. In
the beginning, LEGO® produced only wooden toys, but for the past 40 years, their major
product is based on injection-moulded toybricks. The toybricks can simply be merged
and separated (glue, shears and the like are not used), which contributes to the assembly
speed. The first LEGO® toybricks were made in 1949, when the company’s manage‐
ment made the decision to invest in the development of injection moulding equipment
and began research related to the production of toys [6] (Fig. 1).
Fig. 1. The first original LEGO® toybricks 2 × 4 and 2 × 2 presented in 1949. [6]
Fig. 2. The sixth, final version of LEGO® toybricks. [6]
The modern version of LEGO® toybricks is the sixth iteration of that product, which
is in use since 1960’s. Before that, the original cavity on the inside of the 3 mm toy-
brick was too large and the tubular connection wall was a trick to fit, which was not
good for a brand that strives for perfection. Sometimes they could have broken dock
Dimension Measurement of Injection Moulded Toybricks 59
connectors, and sometimes it happened that the dice collapsed due to the rough game
with the same. These problems were solved in 1960. when new injection moulds were
designed. The cavity inside the cube was reduced to about 2 mm and the improvement
was clearly visible compared to previous versions (Fig. 2).
2.1 The Characteristics of Tested Samples

The LEGO® toybricks material is Acrylonitrile/Butadiene/Styrene (ABS). It is a
polymer material that belongs to the group of plastomers and is characterized by good
mechanical properties: considerable strength, hardness and rigidity, good dimensional
stability, good abrasion resistance, UV-sensitivity [7]. The nominal dimensions are
integer numbers, when measured in millimetres.
Fig. 3. Basic dimensions of LEGO® toybricks. [8]
Fig. 4. The diameter of the cylinder used for geometry measurement on CMM.
There are five main dimensions:

1. Horizontal span or the distance between cylinders: 8 mm
2. Vertical span or height of classical toy-brick: 9.6 mm
3. Horizontal tolerance: 0.1 mm
4. Cylinder diameter: 4.8 mm
5. Cylinder height: 1.8 mm [8].
The horizontal tolerance is one half of the gap between bricks in the horizontal plane.
Horizontal tolerance prevents friction between bricks during assembly. Figure 3 shows
the basic nominal dimensions of LEGO® toybricks [8]. To measure the coordinates of
characteristic points on toybricks, it was necessary to calculate the diameter of the
cylinder as shown in Fig. 4.
√
d =8⋅ 2 − 4.8 = 6.51371 (1)
There are no official data publicly available about tolerances of LEGO® toybricks.
The usual deviations from the nominal dimensions of injection-moulded products range
between 0.01 mm and 0.1 mm. Moulds for plastic injection are subject to strict dimen‐
sional controls. Internal/external length, width and height, and geometric characteristics
such as parallelism, flatness, etc. are usually examined. Plastic injection moulding toler‐
ances range from 0.005 mm to 0.05 mm [9].
3 Experimental Setup
A coordinate measuring machine “ZEISS CONTURA G2” was used (Fig. 5). The
measurement uncertainty of the machine is ±1.8 μm.
Fig. 5. Coordinate measuring machine “ZEISS CONTURA G2”
3.1 Measurement of Deviations of Internal Dimensions of LEGO® Bricks
In order to measure the dimensions relevant for fitting the assembly, there was a need
to develop a special measurement strategy. If one looks at the inside of LEGO® brick
2 × 2, a hollow cylinder located in the middle of the brick can be seen, as well as ribbed
reinforcements placed on each side of the brick (Fig. 6). These characteristic brick
elements are very important because they ensure the proper functioning of the brick
because the small cylinders of the adjoining bricks located on the upper outer surface
are fitted in the cavities between the ribbed reinforcements and the central cylinder of
the brick. These elements make contact with each other and ensure proper and good
fitting of the bricks. The ribbed reinforcements serve, in addition to making contact with
another brick, to increase the rigidity and strength of the brick, and to avoid the appear‐
ance of cavities, strains and stress concentrations during injection moulding.
The dimensions to be measured are the diameter of the central cylinder of the brick
and the deviation from the cylindricity, as well as the deviation of the diameter of the
small cylinders (nominal size Ø4.8 mm) that fit into the brick, to determine whether
there will be an overlap or gap between the elements in the contact. To measure the
diameter of the central cylinder and the deviation from cylindricity, a measurement
Fig. 6. The inside of the LEGO® brick 2 × 2 and the measurement strategy for central cylinder
strategy was designed so that the machine made a measuring tool in the form of a spiral
around the cylinder, in height, and thus locates approximately 1000 points on the
cylinder (Fig. 6). Figure 7 sows how virtual cylinders D1 to D4 in CAD model were
used to define the characteristics dimensions of the inside reinforcements in the brick.
For each cylinder, a set of 3 vertical lines (100 points each) was defined to determine
the appropriate dimension deviations.
Fig. 7. The measurement strategy for virtual cylinders D1-D4
The measurements were performed for 10 different LEGO® bricks, chosen with
different colours, which means that they were not part of the same production lot, and
to assure that the results are not mixed with each other. The measurement results are
shown in the Table 1.
The measurements showed that the deviations of all measured dimensions are small
and are within the prescribed tolerances. In the case of measuring the dimensions of ten
LEGO® bricks, the average deviation from the nominal diameter of the central cylinder
is 22 μm, and the cylindricity deviation is 62 μm. The average deviations from the
nominal dimensions of virtual cylinders D1–D4 are 15 μm, 12 μm, 10 μm, 13 μm,
respectively.
Table 1. Measurement results: deviations from nominal dimensions of LEGO® bricks 2 × 2

Brick colour Deviations from nominal dimensions (mm)
Central Cylinder D1 Cylinder D2 Cylinder D3 Cylinder D4 Cylindricity Average
cylinder
Grey 0,000 0,045 0,047 0,031 0,051 0,047 0,044
Blue 0,011 0,059 0,070 0,056 0,056 0,057 0,060
Yellow 0,009 0,062 0,054 0,031 0,065 0,052 0,053
Red 0,043 0,027 0,056 0,045 0,040 0,090 0,042
Cyan 0,013 0,034 0,043 0,029 0,036 0,042 0,036
Black 0,027 0,012 0,035 0,031 0,022 0,060 0,025
White 0,042 0,054 0,057 0,052 0,061 0,075 0,056
Orange 0,024 0,053 0,053 0,036 0,053 0,070 0,049
Green 0,040 0,043 0,032 0,041 0,056 0,073 0,043
Brown 0,015 0,048 0,061 0,046 0,041 0,055 0,049
Average 0,022 0,044 0,051 0,040 0,048 0,062 0,046
St. dev. 0,015 0,015 0,012 0,010 0,013 0,015 0,010
It was a surprise that these deviations are much smaller than values usually found in
the literature for this kind of manufacturing technology. The results obtained lead to
conclusion that LEGO® bricks are produced in extremely precise injection moulds,
strictly controlled and quality conditions, and subjected to excellent quality control.
Therefore, it would be interesting to compare these results with similar products avail‐
able in the market, usually with significantly lower price.
3.2 Measurement of Deviations of Internal Dimensions of No-Name Bricks
Although no-name bricks are similar in shape to LEGO® bricks, they differ in a way
that there is no central cylinder, and the previous results are not comparable. However,
there are reinforcements in the inside of the brick walls, and they could be measured by
the CMM. The measurement strategy is shown in Fig. 8.
Fig. 8. The measurement strategy for no-name toy-brick

The measurement results: X, Y coordinates of reinforcements of no-name plastic

bricks are shown in Table 2, and Table 3 shows the corresponding coordinates of
LEGO® bricks.
Table 2. Measurement results: deviations from nominal dimensions of no-name bricks 2 × 2

Deviations from nominal coordinates (mm)
X1 X2 X3 X4 Y1 Y2 Y3 Y4 Mean X Mean Y
Brick 1 0,048 0,045 0,040 0,089 0,087 0,170 0,011 0,056 0,056 0,081
Brick 2 0,042 0,073 0,011 0,020 0,109 0,003 0,000 0,033 0,037 0,036
Brick 3 0,224 0,062 0,317 0,063 0,090 0,111 0,043 0,055 0,167 0,075
Brick 4 0,274 0,252 0,182 0,286 0,126 0,218 0,050 0,019 0,249 0,103
Brick 5 0,174 0,172 0,256 0,176 0,002 0,098 0,108 0,133 0,195 0,085
Average 0,152 0,121 0,161 0,127 0,083 0,120 0,042 0,059 0,140 0,076
St.dev. 0,104 0,088 0,133 0,106 0,048 0,081 0,042 0,044 0,108 0,054
Table 3. Measurement results: deviations from nominal dimensions of LEGO® bricks 2 × 2

Deviations from nominal coordinates (mm)
X1 X2 X3 X4 Y1 Y2 Y3 Y4 Mean X Mean Y
LEGO® 1 0,032 0,035 0,018 0,022 0,000 0,051 0,043 0,001 0,027 0,024
LEGO® 2 0,041 0,087 0,041 0,034 0,074 0,094 0,022 0,066 0,051 0,064
LEGO® 3 0,059 0,064 0,055 0,065 0,089 0,058 0,082 0,063 0,061 0,073
LEGO® 4 0,009 0,008 0,002 0,005 0,039 0,041 0,034 0,034 0,006 0,037
LEGO® 5 0,019 0,003 0,001 0,029 0,000 0,004 0,023 0,032 0,013 0,015
Average 0,032 0,039 0,023 0,031 0,040 0,050 0,041 0,039 0,031 0,043
St.dev. 0,019 0,036 0,024 0,022 0,041 0,032 0,025 0,027 0,025 0,031
When results are compared, it can easily be seen that the scattering of results is much
smaller at LEGO® bricks. Also, the deviations from the nominal dimensions are far
smaller in LEGO® bricks and their values range, on average, 31 μm for X and 43 μm
for Y coordinates, whereas these values are much larger in bricks made by unknown
producer and they are 140 μm for X, and 76 μm for Y coordinate. Not only that real
dimensions are much more stable in LEGO® bricks, but also their scattering (expressed
through standard deviation) is significantly smaller. That means that the production
process is much more stable, the dimensions do not vary much in different axes, and the
high price of such products is obviously justified.
4 Conclusion
When comparing two or more different products, which should have identical dimen‐
sions, shape and purpose, while they have different market prices, it is generally known
that more expensive products are usually of better quality. The reasons for this are usually
due to the fact that more expensive and better-quality raw materials are used for the
production of more expensive products, the production technology, labour force, more
expensive machinery and production processes, better quality control, etc. are used. It is
very important to point out that these statements are in general, for the vast majority of
cases, but there are always exceptions.
The aim of this paper was to analyse and to determine the justification of significantly
higher prices of LEGO® bricks on the market compared to other manufacturers. In other
words, the purpose of the work was to determine whether a more expensive product
implies better quality.
The results of the measurements carried out showed that deviations from all meas‐
ured values in cheaper toybricks were higher compared to LEGO® bricks. When the
identical dimensions for each brick were measured, it was found that cheaper products
had large differences in the obtained results, while in the case of LEGO® bricks these
differences were insignificant.
In addition to visible differences (finishing quality, colour stability), also the cheaper
bricks expressed the poorer fitting characteristics. This was confirmed by the results of
the measurements carried out, which showed that the deviations from the nominal
dimensions of LEGO® bricks are very small and are within the range of prescribed
tolerances.
In this term, it can be said that LEGO® has reasonably high price in the market and
that, in terms of precision and quality control, they are unlikely to be competed in the
production of injection-moulding toybricks.
References
1. Ebro, M., Howard, T.J.: Robust design principles for reducing variation in functional performance.
J. Eng. Des. 27(1–3), 75–92 (2016). https://doi.org/10.1080/09544828.2015.1103844
2. Singh, P., Bettig, B.: Port-compatibility and connectability based assembly design. In: ASME
2003 International Design Engineering Technical Conferences and Computers and
Information in Engineering Conference: 29th Design Automation Conference, Parts A and B,
Chicago, Illinois, USA, 2–6 September 2003, vol. 2, pp. 775–784 (2003). https://doi.org/
10.1115/detc2003/dac-48783
3. Bici, M., Broggiato, G.B., Campana, F., Dughiero, A.: Computer aided inspection procedures
to support smart manufacturing of injection moulded components. Procedia Manuf. 11, 1184–
1192 (2017). https://doi.org/10.1016/j.promfg.2017.07.243. ISSN 2351-9789
4. Bici, M., Campana, F., Testani, C., Trifirò, A.: Development of automatic tolerance inspection
through Reverse Engineering. In: 2014 IEEE International Workshop on Metrology for
Aerospace, Benevento, Italy (2014). https://doi.org/10.1109/metroaerospace.2014.6865903
5. Turc, C.G., Carausu, C., Belgiu, G.: Cost analysis in injection moulded plastic parts designing.
In: 2017 IOP Conference Series: Materials Science and Engineering, vol. 227, no. 1, p. 012132
(2017). https://doi.org/10.1088/1757-899X/227/1/01213
6. Mimoso, J.M.: A timeline of early LEGO® development (2009). Taken from https://
www.inverso.pt/legos/Textos/timeline.htm. Accessed 21 Jan 2018
7. Šercer, M., Križan, B., Basan, R.: Konstruisanje polimernih proizvoda. Sveučilište u Zagrebu
(2009)
8. Cailliau, R.: https://www.cailliau.org/Lego/Dimensions/zMeasurements-en.xhtml. Accessed
21 Jan 2018
9. Angel, J.A.B., De Chiffre, L.: Inter laboratory comparison on Industrial Computed
Tomography: CIA-CT comparison. Final Report 01, DTU Mechanical Engineering (2013).
http://orbit.dtu.dk/files/58959939/CIA_CT_Final_Report_01.pdf. Accessed 21 Jan 2018
Wire and Arc Additive Manufacturing
(WAAM) – A New Advance in Manufacturing
Nikola Knezović ✉ and Angela Topić

( )
University of Mostar, 88 000 Mostar, Bosnia and Herzegovina

{nikola.knezovic,angela}@sve-mo.ba
Abstract. Additive Manufacturing (AM) is a technique where structures are

produced by adding and depositing material in a layer upon layer manner. WAAM
(Wire and Arc Additive Manufacturing) is a technology which has been investi‐
gated in last 30 years, although the first patent dates from almost 100 years ago.
It became interesting for scientists and manufacturers due to its ability to produce
fully dense metal parts and large near-net-shape products. WAAM is mostly used
in modern industries, like aerospace industry. It uses existing welding equipment,
electric arc as energy source and welding wire as feedstock. Because of this, it is
cheaper than other AM technologies, which usually need specific equipment and
materials. The process consists of few steps (designing CAD model, slicing into
layers, tool-path generating, choosing welding parameters, material deposition
and post-processing). The aim of this paper is to explain WAAM in detail, give
are view about researches so far in this area and to give suggestions for new
advances.
Keywords: WAAM · Welding · Additive manufacturing
1 Introduction
Modern industry always develops and continuously looking for new investigations and
findings in new technologies. Great example for that is aerospace industry, which will
need about twenty million tonnes of raw material in next two decades [1]. That is mostly
because it uses materials like titanium, which has high BTF (buy-to-fly) ratio. That
means it has low material utilisation because it is necessary to buy greater amounts of
forgings or ingots to produce a low amount of finished parts, due to its poor machinability
[2]. One of the possible solutions are additive manufacturing (AM) technologies. The
bases of AM system are motion system, feedstock and energy (heat) source. Their
combination enables moving, heating and deposition of material. Mostly, AM technol‐
ogies use polymer materials, while the use of metals is possible only in its powder form,
so products are often porous and not fully functional. WAAM (Wire and Arc Additive
Manufacturing) is technology which can offer a solution for these problems.
It has been investigated in last 30 years, although the first patent dates from 1925 [3].
WAAM uses metal welding wire as the feedstock and electric arc as the heat source,
which makes it a combination of welding and AM technology. It uses ordinary welding
equipment (power source, welding torch, wire and protective gas feeding systems),

https://doi.org/10.1007/978-3-319-90893-9_7
66 N. Knezović and A. Topić
combining it with robotic systems or CNC machines which provide moving of the torch,
and that is its first advantage – it can be considerably cheaper than conventional AM
technologies, which use specific equipment and materials. Also, it can produce fully
functional, near-net-shape products, almost unlimited by size.
Achievable deposition rates are usually from 50–130 g/min, which is higher than
any other AM method, so it provides faster manufacturing and reduced time to market
[4]. Considering the fact it builds the product on vacant baseplate, design freedom is
greater than in conventional technologies. Of course, there still are some problems and
challenges to investigate – deformations and residual stresses due to enormous heat
input, relatively low dimensional accuracy and rough surface finish of the part (post-
processing is needed) [5]. Some of these problems are already reduced and researchers
are looking forward to find a solution for other challenges.
2 Process Steps
Figure 1 shows simple WAAM system with its main features (wire feeder and power
source, robot, welding torch and WAAM part) [6]. WAAM starts like any other AM
technology – with designing of 3D CAD model using special software or reverse engi‐
neering processes, like 3D scanning. That model is then saved in some standard format
(usually .stl), which represents model’s geometry. It is the basis for slicing part per its
height into 2D layers, whose contour is then used for generating tool (torch) path. This
step is followed by choosing suitable welding parameters (travel speed, current, voltage,
etc.) and bead modelling. With generated tool path and chosen welding parameters, the
product then can be made by additive manufacturing – the first layer is deposited on the
base plate, torch goes up for the specified layer’s height and deposits second layer onto
it, and the process continues until whole part is made). Some post-process machining
can be done simultaneously [7], but usually, post-processing is done independently,
often with heat treatment if it is needed. Schematic process diagram is shown in the
Fig. 2 [4].
Fig. 1. Simple WAAM system [6]

Wire and Arc Additive Manufacturing (WAAM) 67
Fig. 2. WAAM schematic diagram [4]
This maybe seems simple and easy, but all of this main steps contain numerous
smaller sub-steps which have to be calculated and implemented carefully in order to
produce the high-quality part. Slicing algorithms have been developed to provide layers
of better quality [8], as well as different bead modelling methods [9–11] and various
kinds of path patterns [12–14]. Also, there has been a lot of researches regarding opti‐
mization of welding parameters [15–17] and other welding-related problems like
changes in microstructure [18–20] and problems with heat input [21, 22]. It should be
clear this is a complex process where every mistake in some of the steps can cause major
defects, affecting the part surface finish, microstructure, mechanical properties and
overall quality of the part. If everything is done correctly, parts can look like the ones
shown in Fig. 3. Mostly, MIG is welding process used in WAAM. The wire is coaxial
with the welding torch, which provides easily generated path. While it is suitable for
steel and aluminium, there are problems with arc wandering when it comes to titanium,
and part’s surface is often rough. For producing titanium parts, TIG is a better choice,
but it needs external wire feeding, and to obtain a product of good quality, the wire has
to be fed from the same side. It means torch has to be rotated and it makes robot
programming and tool path generation much more complicated [3].
Fig. 3. Different WAAM products [5]

3 Important Characteristics
Most important WAAM characteristics are its advantages over other conventional
manufacturing technologies and its disadvantages and limitations, which are main
reasons why WAAM is still not used as much as it could be. Advantages are numerous.
Investments and material costs are lower compared with other AM technologies.
Also, use of welding wire as feedstock means the choice of materials is wide. High
deposition rates, near-net-shape processing, short lead time producing and almost
unlimited part size are already mentioned. There is no need for a lot of machining which
means there is less waste.
Some researchers found out that particular mechanical properties of WAAM prod‐
ucts can match or even exceed for castings and wrought products made of same material.
Strength and ductility, according to [23], can reach appropriate values, while fatigue
resistance can surpass them [24].
WAAM also offers a solution for producing functionally graded materials, designed
for particular applications, thanks to their specific properties. One of most known is iron
aluminide, Fe3Al, usually produced using metal powder, which results in porosity prob‐
lems [25]. WAAM can reduce or eliminate this kind of problems and it can make mate‐
rials like that simply by adding another wire into the process. Other materials and alloys
produced this way are TiAl [26], Al-6Mg [27] and NiAl bronze [28].
Unfortunately, there are still some problems and disadvantages which limit WAAM
application. First of all, WAAM is more suitable to make large and less complex parts,
rather than smaller parts with complex geometry and that is a constraint which clearly
defines its application.
Some researchers found anisotropic properties, with higher strength and lower
ductility in horizontal (deposition) direction [23, 24, 29].
Generating tool path is related to defects like excessive camber (peak development),
sag (deposition failure), cavities and porosity. Camber and sag usually occur in crossing
features, while cavities and porosity are common in thick parts. Solutions for mitigating
or at least reducing both of the problems have been found in some papers. A methodology
called MAT (Medial Axis Transformation) is used to solve problems with porosity and
cavities by filling the part from inside towards its boundaries [12–14], while [29, 30]
investigated path patterns for producing crossing features. They were trying to actually
avoid crossing by making path looks like two “L” letters touching each other at the
intersection.
Typical welding issues which also affect WAAM products are residual stresses,
distortions and deformations. WAAM part is usually made on one side of the base plate,
and due to great heat input, deposited layers tend to “pull” base plate towards themselves,
in building direction. Base plate bends together with the wall and deforms it. The base
plate can be clamped, but it produces residual stresses after unclamping base plate. The
solution called Balanced Building Strategy or Symmetrical Building is proposed in [1,
3] and shown in Fig. 4. Layers are deposited on both sides of the plate, and stresses
produced on one side are balanced with those produced on the other side. Sometimes
two parts are needed anyway, so this is an additional benefit, but sometimes redesign is
necessary.
Fig. 4. WAAM part produced using Balanced Building Strategy [1]
4 Conclusion
WAAM is interdisciplinary technology with plenty of space and different areas for
improving and introducing new ideas. Some investigations suggest interpass rolling as
a solution for issues with anisotropic properties [19], other researchers propose including
non-destructive testing in process, which can find porosities early [3], while some papers
presented the use of new materials [31]. Further investigations should include work on
these ideas, alongside improvement of parameters optimization, monitoring, process
control, part design and heat treatment, which will all together lead to better under‐
standing and implementation of WAAM technology. With improvements in these areas,
WAAM could become a replacement for conventional production methods, like casting
and forging, in particular applications. Present issues still limit the industrial use and
market’s approval of WAAM, but different researches are already suggested to reduce
and remove them, so WAAM can become equal to traditional methods.
References
1. Mehnen, J., Ding, J., Lockett, H., Kazanas, P.: Design study for wire and arc additive
manufacture. Int. J. Prod. Dev. 19(1/2/3), 2 (2014)
2. Martina, F.: Wire + arc additive vs. from solid (2015)
3. Williams, S.W., Martina, F., Addison, A.C., Ding, J., Pardal, G., Colegrove, P.: Wire + arc
additive manufacturing. Mater. Sci. Technol. 32(7), 641–647 (2016). https://doi.org/
10.1179/1743284715Y.0000000073
4. Ding, D.-H., Pan, Z.-X., Dominic, C., Li, H.-J.: Process planning strategy for wire and arc
additive manufacturing. In: Tarn, T.-J., Chen, S.-B., Chen, X.-Q. (eds.) Robotic Welding,
Intelligence and Automation: RWIA 2014, pp. 437–450. Springer International Publishing,
Cham (2015)
5. Ding, D., Pan, Z., Cuiuri, D., Li, H.: Wire-feed additive manufacturing of metal components:
technologies, developments and future interests. Int. J. Adv. Manuf. Technol. 81(1–4), 465–
481 (2015)
6. Ding, J., Martina, F., Williams, S.: Production of large metallic components by additive
manufacture – issues and achievements. In: 1st Metallic Materials and Processes: Industrial
Challenges, June 2015
7. Xiong, X., Zhang, H., Wang, G.: Metal direct prototyping by using hybrid plasma deposition
and milling. J. Mater. Process. Technol. 209(1), 124–130 (2009)
8. Ding, D., Pan, Z., Cuiuri, D., Li, H., Larkin, N., Van Duin, S.: Automatic multi-direction
slicing algorithms for wire based additive manufacturing. Robot. Comput. Integr. Manuf.
37, 130–150 (2016)
9. Ding, D., Pan, Z., Cuiuri, D., Li, H., Van Duin, S., Larkin, N.: Bead modelling and
implementation of adaptive MAT path in wire and arc additive manufacturing. Robot.
Comput. Integr. Manuf. 39, 32–42 (2016)
10. Li, Y., Sun, Y., Han, Q., Zhang, G., Horváth, I.: Enhanced beads overlapping model for wire
and arc additive manufacturing of multi-layer multi-bead metallic parts. J. Mater. Process.
Technol. 252, 838–848 (2018)
11. Suryakumar, S., Karunakaran, K.P., Bernard, A., Chandrasekhar, U., Raghavender, N.,
Sharma, D.: Weld bead modeling and process optimization in hybrid layered manufacturing.
Comput. Des. 43(4), 331–344 (2011)
12. Ding, D., Pan, Z., Cuiuri, D., Li, H., Larkin, N.: Adaptive path planning for wire-feed additive
manufacturing using medial axis transformation. J. Clean. Prod. 133, 942–952 (2016)
13. Ding, D., Pan, Z., Cuiuri, D., Li, H.: A practical path planning methodology for wire and arc
additive manufacturing of thin-walled structures. Robot. Comput. Integr. Manuf. 34, 8–19
(2015)
14. Ding, D., Pan, Z., Cuiuri, D., Li, H.: A tool-path generation strategy for wire and arc additive
manufacturing. Int. J. Adv. Manuf. Technol. 73(1–4), 173–183 (2014)
15. Xiong, J., Li, Y., Li, R., Yin, Z.: Influences of process parameters on surface roughness of
multi-layer single-pass thin-walled parts in GMAW-based additive manufacturing. J. Mater.
Process. Technol. 252, 128–136 (2018)
16. Geng, H., Li, J., Xiong, J., Lin, X., Zhang, F.: Optimization of wire feed for GTAW based
additive manufacturing. J. Mater. Process. Technol. 243, 40–47 (2017)
17. Adebayo, A., Mehnen, J., Tonnellier, X.: Limiting travel speed in additive layer
manufacturing. In: Proceedings of the 9th International Conference on Trends in Welding
Research, vol. 3, pp. 1038–1044 (2013)
18. Gu, J., Ding, J., Williams, S.W., Gu, H., Ma, P., Zhai, Y.: The effect of inter-layer cold
working and post-deposition heat treatment on porosity in additively manufactured aluminum
alloys. J. Mater. Process. Technol. 230, 26–34 (2016)
19. Donoghue, J., Antonysamy, A.A., Martina, F., Colegrove, P.A., Williams, S.W., Prangnell,
P.B.: The effectiveness of combining rolling deformation with Wire-Arc Additive
Manufacture on β-grain refinement and texture modification in Ti-6Al-4V. Mater. Charact.
114, 103–114 (2016)
20. Martina, F., Colegrove, P.A., Williams, S.W., Meyer, J.: Microstructure of interpass rolled
Wire + Arc additive manufacturing Ti-6Al-4V components. Metall. Mater. Trans. A Phys.
Metall. Mater. Sci. 46(12), 6103–6118 (2015)
21. Wu, B., et al.: Effects of heat accumulation on the arc characteristics and metal transfer
behavior in wire arc additive manufacturing of Ti6Al4V. J. Mater. Process. Tech. 250, 304–
312 (2017)
22. Jandric, Z., Labudovic, M., Kovacevic, R.Ã.: Effect of heat sink on microstructure of three-
dimensional parts built by welding-based deposition. Int. J. Mach. Tools Manuf 44, 785–796
(2004)
23. Brandl, E., Baufeld, B., Leyens, C., Gault, R.: Additive manufactured Ti-6A1-4V using
welding wire: comparison of laser and arc beam deposition and evaluation with respect to
aerospace material specifications. Phys. Procedia 5(Part 2), 595–606 (2010)
24. Wang, F., Williams, S., Colegrove, P., Antonysamy, A.A.: Microstructure and mechanical
properties of wire and arc additive manufactured Ti-6Al-4V. Metall. Mater. Trans. A Phys.
Metall. Mater. Sci. 44(2), 968–977 (2013)
25. Shen, C., Pan, Z., Ma, Y., Cuiuri, D., Li, H.: Fabrication of iron-rich Fe-Al intermetallics
using the wire-arc additive manufacturing process. Addit. Manuf. 7, 20–26 (2015)
26. Ma, Y., Cuiuri, D., Li, H., Pan, Z., Shen, C.: The effect of postproduction heat treatment on
γ-TiAl alloys produced by the GTAW-based additive manufacturing process. Mater. Sci. Eng.
A 657, 86–95 (2016)
27. Zhang, C., Li, Y., Gao, M., Zeng, X.: Wire arc additive manufacturing of Al-6Mg alloy using
variable polarity cold metal transfer arc as power source. Mater. Sci. Eng. A 711, 415–423
(2018)
28. Ding, D., Pan, Z., van Duin, S., Li, H., Shen, C.: Fabricating superior NiAl bronze components
through wire arc additive manufacturing. Materials (Basel) 9(8), 652 (2016)
29. Baufeld, B., Brandl, E., Van Der Biest, O.: Wire based additive layer manufacturing:
comparison of microstructure and mechanical properties of Ti-6Al-4V components fabricated
by laser-beam deposition and shaped metal deposition. J. Mater. Process. Technol. 211(6),
1146–1158 (2011)
30. Venturini, G., Montevecchi, F., Scippa, A., Campatelli, G.: Optimization of WAAM
deposition patterns for T-crossing features. Procedia CIRP 55, 95–100 (2016)
31. Guo, J., Zhou, Y., Liu, C., Wu, Q., Chen, X., Lu, J.: Wire arc additive manufacturing of AZ31
magnesium alloy: grain refinement by adjusting pulse frequency. Materials (Basel) 9(10),
823 (2016)
Analysis of the Type and Chemical Content
of the Inclusion on SEM of the Stainless Steel
With and Without the Addition of Zr and Te
Derviš Mujagić ✉ , Aida Imamović, Mirsada Oruč,

( )
and Sulejman Muhamedagić

Institute “Kemal Kapetanović”, University of Zenica, 72000
dervismujagic@yahoo.com, mirada.oruc@unze.ba
Abstract. Control of the formation of non-metallic inclusions and their charac‐

terization is presenting the foundation for improving of the qualities of steel
products, as well as designing of the new types of steel.
In order to produce steel with improved machinability, such as austenitic
stainless steel X8CrNiS18-9 modification of the chemical composition of non-
metallic inclusions is made. In this paper the SEM studies of the presence and
type of inclusions in austenitic stainless steel with the addition of sulphur of the
label X8CrNiS18-9 without addition and addition of alloying elements Zr and Te,
were carried out individually in order to investigate its influence on the shape and
chemical composition of the inclusions.
Keywords: Stainless steel · Alloying · Non-metallic inclusions

Testing on SEM
1 Introduction
The research was carried out on an austenitic stainless steel labelled X8CrNiS18-9, as
defined in standard EN 10088-3: 2005, which was produced at the semi-industrial plants
of the Metallurgical Institute “Kemal Kapetanović”, University of Zenica, in order to
determine the influence of zirconium and tellurium in austenitic stainless steel especially
from the aspect of mechanical workability.
In accordance with the program of testing at the Department for melting and casting
metals of the Metallurgical Institute “Kemal Kapetanović”, several melt of different
chemical composition was made [1]. The different chemical composition is reflected in
various contents of zirconium and tellurium. The first melt was made without the addi‐
tion of alloying elements, while in the remaining melts; the alloying element was added
individually. In this paper, the tests were performed for three variants (three different
melts), with the mentioned alloying elements being added within the boundaries
according to the literature [1]. The samples from variant 1 (excluding alloying elements)
were tested, followed by variants 2 with the addition of Zr, and variant 3 with the addition

https://doi.org/10.1007/978-3-319-90893-9_8
Analysis of the Type and Chemical Content of the Inclusion on SEM 73
of Te. For each of the above samples, line and point analysis were done, and also
mapping.
Tests of inclusions in the produced melt were carried out on samples in the valid
state, as final grade of processing: Scanning Electron Microscope (SEM) type JEOL
JSM 5610 at the Faculty of Natural Sciences and Engineering of the University of
Ljubljana [1, 2].
2 Testing on SEM
Examination of the inclusions on the scanning electron microscope gives the highest
data on the type of inclusions and their composition. Those inclusions effect on the
mechanical properties of steel depends of the addition applied elements [3]. Figure 1
shows a linear analysis of characteristic non-metallic inclusions for a sample of auste‐
nitic stainless steel X8CrNiS18-9 without additives of alloying elements (variant 1).
Fig. 1. Linear analysis of the sample without additives of alloying elements (variant 1): (a) SEM
image; (b) a distribution diagram of S and Mn
Figure 1, i.e. the distribution diagram of the contents of S and Mn in the line shown
(Fig. 1b), shows how the inclusions are given in the SEM picture (Fig. 1a), in its chemical
composition predominantly manganese sulphides [2]. Figure 2 shows the spotted anal‐
ysis of individual inclusions presented by the SEM image. For each of the mentioned
points in the SEM picture (Fig. 2a), diagrams of the contents of individual elements are
given, on which chemical analysis of the detected elements in mass percent is also given
(Fig. 2b–e). Based on the analysis of the above diagrams, it can be concluded that the
inclusions formed in the above sample are mainly the inclusion of manganese sulphides,
but also complex inclusions, which, in addition to a certain content of manganese
sulphide, contain other elements such as silicon, aluminium and oxygen (point 3 in SEM
picture and Fig. 2d) or molybdenum and nickel (points 2 and 4 in SEM, Fig. 2c and e).
74 D. Mujagić et al.
Fig. 2. Spot analysis of sample inclusions without alloying supplements (variant 1): (a) SEM
image; (b)–(e) diagrams of detected elements with chemical analysis in mass percentages
Figure 3 gives a mapping showing where each of the elements represented by

mapping is in relation to the SEM image, showing the inclusions of the sample without
the addition of alloying elements (variant 1). It is clear from the picture that the inclusions
from the SEM image (Fig. 3a) are almost identical to those in the images showing the
position of manganese and sulphur (Fig. 3c and f), whereas the black prints of these
inclusions in the image showing the distribution of iron, in fact, denote the impoverish‐
ment of iron of the mentioned areas (Fig. 3g). Also, with the pictures it can be seen that
the inclusion, which is positioned on the SEM picture extreme left, is a complex inclu‐
sion which contains, in addition to certain content of manganese and sulphur, certain
quantities of silicon, oxygen and aluminium (Fig. 3b, d and i).
Figure 4 shows a linear analysis of the zirconium supplement (variant 2). Figure 4,
i.e. the distribution diagram of the contents of S, Mn and Zr in the line (Fig. 4b) shows
how the inclusions are presented in the SEM figure (Fig. 4a), not only involving manga‐
nese sulphides, but complex inclusions, which contain a certain amount of zirconium.
The figures also show, and the diagram confirms that the zirconium is within the inclu‐
sions of manganese sulphide, as their integral part [2].
Fig. 3. Mapping of the sample without additions of alloying elements (variant 1): (a) SEM image;
(b)–(i) the elements Si, S, O, Ni, Mn, Fe, Cr and Al
Fig. 4. Line analysis of the zirconium supplement (variant 2): (a) SEM image; (b) the distribution
diagram of S, Mn and Zr
Figure 5 shows a spot analysis of one of the inclusions presented by the SEM image.
For each of the above points, marked with numbers 1 to 4 in the SEM picture
(Fig. 5a), diagrams of the contents of individual elements are given, on which chemical
analysis of the detected elements in mass percent (Fig. 5b–e) is also given. Based on the
Fig. 5. Spot analysis of the inclusion of the sample with the addition of zirconium (variant 2):
(a) SEM image; (b)–(e) diagrams of detected elements with chemical analysis in mass percentages
analysis of the above diagrams, it can be concluded that the shown inclusion is a complex
inclusion, which, in addition to manganese sulphide, contains other elements, in this
case zirconium and oxygen (point 1 and Fig. 5b), which are probably formed in part by
zirconium oxide, and molybdenum and nickel (points 2, 3 and 4 in the SEM figure,
respectively, Fig. 5c, d, and e).
Figure 6 shows a mapping where each of the represented elements is in relation to
the SEM image, showing the inclusions of the sample with the addition of zirconium
(variant 2) [4]. It is clear from the picture that the inclusions from the SEM image
(Fig. 6a) are almost identical to those in the images showing the position of the manga‐
nese and sulphur (Fig. 6d and e), while the black prints of these inclusions in the figure
showing the distribution of iron, in fact, denote the impoverishment of the iron of the
mentioned areas (Fig. 6c). Also, on the basis of the comparison of the SEM image with
an image indicating the zirconium distribution, it can be clearly concluded that the
zirconium is at its maximum in three different locations within the large inclusion
(Fig. 6f).
Figure 7 shows a linear analysis of the pattern with the addition of tellurium
(variant 3). Figure 7, i.e. the distribution diagram of the contents of S, Mn and Te
according to the shown line (Fig. 7b), shows that the inclusion shown in the SEM
figure (Fig. 7a), in addition to the already conventional manganese sulphide, contains
to a certain extent tellurium. The picture also shows, and the diagram confirms that
the tellurium practically surrounds the inclusion of manganese sulphide, and is
located at its ends.
Fig. 6. Mapping a sample with the addition of zirconium (variant 2): (a) SEM image; (b)–(f) Cr,
Fe, Mn, S, and Zr elements [4]
Fig. 7. Line analysis of the sample with the addition of tellurium (variant 3): (a) SEM image; (b)
distribution diagram of S, Mn and Te according to the displayed line
Figure 8 shows a spot analysis of the inclusions presented by the SEM image.
For each of the above points, marked with numbers 1 to 4 in the SEM picture
(Fig. 8a), diagrams of the contents of individual elements are given, which also contain
chemical analysis of the detected elements in mass percentages (Fig. 8b–e). Based on
the analysis of the above diagrams it can be concluded that a larger complex inclusion,
which, in addition to manganese sulphides, contains other elements, in this case a tellu‐
rium, and in certain amounts both molybdenum and nickel (points 1, 2 and 3 and Fig. 8b,
c and d). A smaller inclusion from the image, also in addition to manganese sulphide
and a certain amount of tellurium, contains significant molybdenum and nickel contents
(Point 4 and Fig. 8e).
Fig. 8. Spot analysis of inclusion of the sample with addition of tellurium (variant 3): (a) SEM
image; (b)–(e) diagrams of detected elements with chemical analysis in mass percentages
Fig. 9. Mapping a sample with addition of tellurium (variant 3): (a) SEM image; (b)–(f) Cr, Fe,
Mn, S, and Te elements
Figure 9 shows a mapping where each of the elements represented by mapping is in

relation to the SEM image, which shows the inclusions of the sample with the addition
of tellurium (variant 3). It is clear from the picture that the inclusions from the SEM
image (Fig. 9a) are almost identical to those in the images showing the position of
manganese and sulphur (Fig. 9d and e), whereas the black prints of these inclusions in
the image showing the distribution of iron, in fact, denote the impoverishment of iron
of the mentioned areas (Fig. 9c). Also, based on the comparison of the SEM image with
the image indicating the tellurium distribution, it can be clearly concluded that the tellu‐
rium is practically in all the inclusions shown in the SEM picture, mainly at their ends
(Fig. 9f).
The steel-variant 3 with the addition of tellurium has characteristically manganese
sulphides combined with a tellurium which are a typical globular shape and specifically
improve the machinability of this steel.
3 Conclusions
The aim of the research was to determine the effects of zirconium and tellurium in
austenitic stainless steel with addition of X8CrNiS18-9 sulphur on types of non-metallic
inclusions, as well as mechanical workability.
Metallographic tests on the SEM were performed and the influence of the chemical
elements of the zirconium and the tellurium on the change in the type and chemical
composition of non-metallic inclusions was monitored.
It has been found that the basic types of inclusions in this steel are manganese
sulphides and that the addition of zirconium and tellurium can be modified by the non-
metal inclusions of austenitic stainless steel X8CrNiS18-9, and that especially manga‐
nese sulphides in combination with a tellurium can be translated into a suitable form,
and on one side they will provide good machinability and, on the other hand, will not
adversely affect the properties of austenitic stainless steel X8CrNiS18-9.
References
1. Mujagić, D.: Doprinos istraživanju uticaja mikrolegiranja sa borom, cirkonijem i telurom na

osobine austenitnog nehrđajućeg čelika sa dodatkom sumpora X8CrNiS18-9, Doktorska
disertacija, Fakultet za metalurgiju i materijale Univerziteta u Zenici, Zenica (2017)
2. Mahmutović, A., Nagode, A., Rimac, M., Mujagić, D.: Modification of inclusions in austenitic
stainless steel by adding tellurium and zirconium. Mater. Technol. 51(3), 523–528 (2017).
MTAE9, ISSN:1580-2949
3. Oruč, M., Sunulahpašić, R.: Savremeni metalni materijali. Univerzitet u Zenici, Fakultet za
metalurgiju i materijale, Zenica (2005)
4. Mujagić, D., Mahmutović, A., Hadžalić, M., Karić, A.: The influence of zirconium on the
properties of austenitic stainless steel. In: XI International Scientific-Expert Conference
“Metallic and Nonmetalic Materials”, Zenica, 21–22 April 2016
Strength and Deformation Calculation
of Flat O-Springs
Ivan Balashev1, Mariel Penev1, Ivan Stoyanov2(&),

and Radoslav Balashev3
1
Technical University of Gabrovo, Gabrovo, Bulgaria
{balashev,mpenev}@tugab.bg
2
Podem Gabrovo Ltd., Gabrovo, Bulgaria
creative_studio1979@abv.bg
3
PRONY Engineering Ltd., Gabrovo, Bulgaria
rudib@abv.bg
Abstract. A range of flat O-springs has been developed. 3D computer mod-

eling has been performed with strength and deformation calculation using the
software products Autodesk Inventor Professional and SolidWorks. A com-
puter-planned experiment has been made and regression equations have been
obtained for determination of the maximum stresses and deformations of the
O-springs. The results are shown graphically.
Keywords: Flat springs Computer modeling Maximum stresses

Deformations
1 Introduction
Springs are elastic elements intended to create tensile and compression forces on
machine elements and to cushion impacts during dynamic loading of machinery and
equipment. Most used are the cylindrical and conical screw springs, the automotive
springs, the disk and ring springs. This article reviews the developed range of special
shaped flat O-springs, intended to operate under tension-compression.
Of the developed range of special shaped flat O-springs Fig. 1 features a spring
with one element, and Fig. 2 – a flat O-spring with n elements; the adopted size ratios
are shown in Table 1.
1.1 Computer Modeling and Calculation

Computer modeling has been made of an element of all sizes of flat O-springs (Table 1)
in the software product environment Solid Work. The stresses and deformations have
been determined for different loading and thicknesses d = 6, 8 and 10 mm. The results
of one model of springs are shown in Figs. 3 and 4. Computer modeling has been
performed of a packet with four element O-springs for various loadings; the results are
given in Figs. 5 and 6.

https://doi.org/10.1007/978-3-319-90893-9_9
Strength and Deformation Calculation of Flat O-Springs 81
L2
L1
2
R
1
b1
b 5 R1
5
1
L
Fig. 1. An element of a flat O – spring

L
H=n.b1
Fig. 2. Flat O–spring of n elements
Table 1. Sizes of the range ‘‘O”– springs

Spring Sizes, mm
L L1 L2 b b1 R1 R2 d
P60(6, 8, 10) 60 46, 8 48 12 13 6 3,6 6, 8, 10
P80(6, 8, 10) 80 60, 8 64 16 18 8 4,8 6, 8, 10
P100(6, 8, 10) 100 76 80 20 22 10 6 6, 8, 10
2 Determination of Spring Stiffness
The results obtained with the software product Solid Work are shown in Tables 2, 3
and 4.
A program has been created in the MathCAD-environment and the parametric
constants (stiffness) of the range of flat O-springs have been determined.
82 I. Balashev et al.
Fig. 3. Stresses von Mises, N/m2

Fig. 4. Deformations URES [mm]
Fig. 6. Deformations URES [mm]

Fig. 5. Stresses von Mises, N/m2 of a
of a 4 element-packet
4 element-packet
Qi
Ci ¼ ; N=m ð1Þ
Di
where: Qi, N is the force acting on the spring; Di, m - the deformation of the spring in
result of the relevant force.
The results are shown in Figs. 7, 8 and 9.
The deformation Dn, mm and the parametric constant (stiffness) Cn, N/m of a packet
of n element flat O-springs with equal sizes are determined [1]:
Ci
Cn ¼ ; N=m Dn ¼ n:D1 ; mm ð2Þ
n
Table 2. Calculation results for flat spring P60

Thickness Axis loading Q, N
d = 6 mm 1200 900 600 300
Deformations D, mm
0,6002 0,4502 0,3001 0,1501
d = 8 mm Axis loading Q, N
1600 1200 800 400
Deformations D, mm
0,5967 0,4475 0,2987 0,1491
2000 1500 1000 500
Deformations D, mm
0,5929 0,4447 0,2965 0,1482

d = 6 mm 1200 900 600 300
Deformations D, mm
0,6002 0,4502 0,3001 0,1501
1600 1200 800 400
Deformations D, mm
0,5967 0,4475 0,2987 0,1491
2000 1500 1000 500
Deformations D, mm
0,5929 0,4447 0,2965 0,1482

d = 6 mm 1200 900 600 300
Deformations D, mm
0,6002 0,4502 0,3001 0,1501
1600 1200 800 400
Deformations D, mm
0,5967 0,4475 0,2987 0,1491
2000 1500 1000 500
Deformations D, mm
0,5929 0,4447 0,2965 0,1482
84 I. Balashev et al.
0.8 3500
0.6 3
3.373×10
Δ1i 0.6 3000

C1
Δ2 i 0.4 C2 2500
Δ 3i C3
0.2 2000
3
0.148 0 1.999×10 1500
1 1.75 2.5 3.25 4 1 1.75 2.5 3.25 4
1 i 4 1 i 4
Fig. 7. Deformation Di, mm, and stiffness C, N/m of spring P60
1 33500
0.876 3.088×10
0.75 3000
Δ 1i C1
Δ2 i C2 2500
0.5
Δ3i C3
0.25 2000
3
0.234 1.705×10 1500
0
1 1.75 2.5 3.25 4 1 1.75 2.5 3.25 4
1 i 4 1 i 4
1 33500
0.921 3.258×10
3250
Δ 1i 0.75 C1 3000
Δ 2i 0.5
C2 2750
Δ3i C3 2500
0.25
2250
3
0.154 0 2.007×10 2000
1 2 3 4 5 1 2 3 4 5
1 i 5 1 i 5
3 Conclusion
The researches performed on a range of flat O-springs show that they have linear
characteristics and through packets of n elements the desired deformation and stiffness
can be obtained.
In the praxis, the study results may be used by technicians and engineers for the
design and development of equipment and machinery.
References
1. Balashev, I.: Machine Elements, TU Gabrovo (2002)
2. Efremov, P., Genchev, G.: Spring and Leaf Springs. Technica, Sofia (1963)
3. Autodesk Inventor Professional (2015)
4. SolidWorks
5. MathCAD
Prototype Apparatus for Calibration Contact
Sensors for Measuring the Temperature
of a Solid Surface
Edin Terzić(&), Raif Seferović, and Narcisa Jarović-Bajramović
Metallurgical Institute “Kemal Kapetanović”, University of Zenica,

72000 Zenica, Bosnia and Herzegovina
{edin.terzic,raif.seferovic}@miz.unze.ba,
72narcisa@gmail.com
Abstract. The paper is about calibration of contact sensors for measuring

temperature of solids surface. The paper describes process of idea development,
design and make of prototype apparatus for calibration of contact sensors in
temperature range from +50 °C to +600 °C is described. The apparatus consists
of two separate and integrated systems for heating. The test results are given in
the form gained by measurement determined characterization of the relevant
parameters of the apparatus, according to the protocol for the validation of the
method. The significant improvements and innovative approach are confirmed
and extending of the temperature range up to 600 °C is provided. These inno-
vations are related to the way of heating the reference body to a set temperature
calibration point. The innovations are also seen in a way to ensure long-term
uniformity of temperature profile and stability of the temperature of the refer-
ence surface with standard deviations less than 0,01 °C to 0,05 °C, based on
values read-out during 180 min in the entire calibration range. The specified
individual and summary contributions of measurement uncertainty of the
apparatus for all calibration points are specified in the end.
Keywords: Calibration of contact temperature sensors Surface temperature

Prototype apparatus Extrapolation Measurement uncertainty of apparatus
1 Introduction
The paper is dealing with calibration of contact sensors for measuring surface tem-
perature of solid materials. The prototype – apparatus, for calibration of this type of
sensors, has been developed and made to operate in temperature range from 20 °C to
600 °C. Apparatus consists of two separate and integrated systems for heating. Each
system has its separate temperature circle for measuring and regulation that can assure
precise and accurate leading of temperature regime. Homogeneity and uniformity of
temperature profile is assured by choosing type and positions of the heating bodies, and
control of uniformity of thermal field within the reference body is assured by instal-
lation of 5 sensitive temperature sensors along the hight of the specimen. The sensors
are replaceable and it is assured slight transition from one to another heating regime.
The elements which can be installed and uninstalled are enabling simulation and

https://doi.org/10.1007/978-3-319-90893-9_10
Prototype Apparatus for Calibration Contact Sensors 87
varying of different practical conditions in which the sensors are used. It is possible to
measure the heat load in and around the specimens as well as uniaxial independent
movement of the horizontal axis of the apparatus and the sensor/contact thermometer
holder.
This apparatus allows the control and varying of influence of ambient conditions at
the surface by exchange of heat meaning by convection and radiation between the
upper surface of the specimen and the surrounding atmosphere.
The paper presents the results of analysis of the impact of different temperature
gradients in the heat exchange of the heated or cooled specimens with the environment
and impact on the homogeneity of the temperature field specimens.
The results of the test apparatus are given in the form, established characterization
measurements, the important parameters of the apparatus. The significant improve-
ments and innovative approach are confirmed and provided extending of the temper-
ature range up to 600 °C, and for the first time enabled to quantify the measurement
uncertainty contribution of the method itself by measurements. So far this has been
done only by usage of computational evaluation, with so far unreached measurement
uncertainties.
2 Current State of the Art
In thermal metrology one of the most divergent areas of calibration is calibration of

surface temperature sensors. Some of the NMI-s have developed apparatus for cali-
bration. There are present also commercial products. However, due to significantly
different operational principles, it is questionable the comparability of such apparatus.
In the EUROMET 635 project, [5], it has been proven that difference among different
NMI apparatuses is in order of 2 to 5 °C at temperature of 300 °C.
At present, surface temperature sensors are generally calibrated by means of a
temperature-controlled hot plate, [1–3]. There are some limitations to such systems.
The reference surface temperature is determined as result of linear or quadratic
extrapolation from temperatures within the hot plate. However, the true surface tem-
perature of the hot plate surface largely depends of thermal resistance, thermal con-
ductivity of the referent plate, heat transfer, homogeneity and stability of the thermal
field inside referent plate etc. The most important thing is that apparatus should have
possibilities to quantify all the uncertainty contributions. One of the approaches is
possibility to change thermal conditions by control of surrounding temperature, which
is not a case in majority of current hot plate systems.
The basic idea of the newly developed prototype apparatus for the calibration of
surface temperature sensors is to try to eliminate (minimize) dynamical thermal
unbalance within and on the top of the reference hot plate, before and during the
calibration. This feature is the most important for higher temperatures over (>300 °C).
By identifying sources which are causing dynamic thermal unbalance and non-steady
state conditions during the calibration, we have concluded that the new approach in
construction of the calibration apparatus is needed. With new apparatus we are able to
achieve long-term stability of the surface temperature of the reference body during
calibration and to simulate different operating conditions.
88 E. Terzić et al.
In construction of this prototype apparatus we started with conditional hypothesis

that will be possible with this apparatus to carry out complete calibration with possi-
bility to eliminate or significantly to reduce the influence of all uncertainty contribu-
tions of apparatus which are existing.
3 Technical and Technological Characteristics of Prototype
Accurate contact measurements of surface temperature of solid bodies and mathe-

matical interpretations of thermal conditions on the referent plate with or without
embedded sensors are described in available literature [8–12]. Also, many mathemat-
ical models for extrapolation of embedded temperature sensors to the surface tem-
perature are developed, [9, 10]. Majority of them are based on the same design of the
apparatus. According to the available literature, deviation between the “true” surface
temperature, i.e., the undisturbed temperature without the sensor, and the actual sensor
output can be expressed as the sum of three contributions [8, 9]: (i) error resulting from
the change in the temperature field within the measurand, caused by the attached
temperature sensor; (ii) error due to the contact resistance between the reference surface
and the contact surface of the sensor; and (iii) error related to the intrinsic distance
between the sensing element in the probe and the external surface of the probe itself. In
order to reduce such errors, all of laboratories use operating procedures for calibration
which take in account various influence parameters.
In construction of the prototype apparatus we have started with hypothesis that it will
be possible, with a new design apparatus, to carry out complete characterization of the
device under calibration with different possibilities to estimate different uncertainty
sources. Our goal was also to develop calibration system, which will be able to provide
conditions without heat loss and distortion of the surface temperature of the reference
body.
The prototype apparatus has been developed by TM - Laboratory for technical
measurements in Metallurgical Institute Zenica, University of Zenica, Bosnia and
Herzegovina. There are number of original construction solutions [6].
General operating principle of the prototype apparatus is shown in Fig. 1.
According to the original design it is defined that apparatus has 5(five) key mechanical
parts which are main parts of the apparatus. These parts are: (i) Referent and inter-
changeable hot plate, temperature equalizing block with carrying structure of the
apparatus, sensors for surface temperature extrapolation and sensors for heating reg-
ulations; (ii) Mechanical construction of the moving part of apparatus for vertical
translations; (iii) Mechanical construction for end movement apparatus-angle adjust-
ment; (iv) Mechanical construction for clamping probe under calibration in total ver-
tical position related to referent hot plate; (v) Mechanical assembly for vertical
translations and probe attachment to referent surface with control thrust force.
As it is shown in Fig. 1, by combination and careful selection of construction
materials and choosing the best measurement–regulation equipment we can expect very
high homogeneity and steady state conditions on referent surface of the hot plate in
addition to possibility of identification the best thermal situation in the whole system
for reading calibration results. With precise control of the heating power of the heating
Fig. 1. Principal scheme of apparatus for calibration of contact temperature sensors on solid
surfaces
parts, density of thermal flux in equalizing block and referent plate can be determined
and changed appropriately. As the end result, we can assure minimal heat exchange
with ambient within the entire system.
In Fig. 1, which shows elements of calibrating apparatus, there is a referent body
with holes for micro-thermocouples (1), so-called equalizing cylindrical block (2) made
of electrolytic Cu purity of (99,999%), Watlow, high temperature ceramic knuckle
band heaters (3, 4) which are designed to provide high-performance heating at tem-
perature up to 760 °C. This level of performance is achieved from the ceramic knuckles
that provide excellent insulation, bottom Watlow spiral heater (5), different insulation
laminated-ceramic wool materials (6, 8, 10 and 12), carrying unmovable steel structure
(7), inside moving ceramic tube is a carrier of heaters and joined with bigger ceramic
tube (11) within a moving part of apparatus, housing unit (13) made of steel sheet
thickness 3,0 mm, mechanism for vertical movement (15) for operating adjustment.
Figure 2 shows global thermal situation in calibration system when it is placed in
the position (a), which shows the influence of constructive characteristics to the thermal
equilibrium in the system.
Figure 3, shows plan views of the respective elements in connection with mea-
surement regulation circles. Embedded temperature sensors are connected with nec-
essary electronic and other means for measuring and transferring the desired measuring
signals to external processing equipment which are described in the schematic.
Fig. 2. Thermal situation in equalizing block, referent plate and ambient air, over plate in
apparatus for calibration of contacting thermometer for temperature measurements of solid
surface;
Fig. 3. General Schema of heating elements placement, measurements and temperature

regulation in apparatus for calibration of contact thermometer for temperature measurements
of solid surface
4 Development and Testing of the Prototype Apparatus
After creating a project basis, sizing apparatus and main assemblies, conceptual designs
for individual components of the apparatus, a selection of equipment to be purchased
and drawings of individual parts for each assembly, it was made dynamic schedule for
Table 1. Construction, work and operational characteristics of the prototype apparatus.

1. Design features
The dimensions of the apparatus:
Outer dimensions: 500 500 1000 mm
The outer steel cylinder: ɸ 316/302 500 mm
The inner steel cylinder: ɸ 300/286 500 mm
The ceramic tube – Al2O3 99,999%: ∅ 109/97 360 mm
Thermal insulation: ceramic fibre blanket, rolled cloth, insulation bricks
Thickness: ∅ 90 mm 500 mm of ceramic tube
The ring heaters 2 pcs.: WATLOW, power 1250 W ∅ 140/110 75 mm
Spiral heater with ceramic holder 1 pcs.: ∅ 75 25 mm; ∅ 100/75 63 mm
Vertical Stroke: 150 mm
Shaft with coupling and screw: ∅ 21 390 mm
Block for temperature equalizing 1 pcs.– Cu 999: ∅ 96h6 75 mm
The reference body 3 pcs., Al, Cu, Stainless Steel: ∅ 50h5 50 mm
Motor for apparatus rotation: “STROJNA” Maribor, tip:GP1 B3-ZP.B; ser.br.236502
Bearing: SNL 206–305
2. Operational performance
Calibration temperature range: 50 do 600 °C
Turning apparatus around the axial axis: 90°
Compressive strength tests on the reference surface, torque wrench: from 0 to 5 KN/0,01 KN
Positions of the reference body: 2 working and 1 service
The total number of thermocouples: 3 pcs. for control of heaters and 4 pcs. for the extrapolation
of the surface temperature
Measuring range of thermocouples: from −50 °C to +1100 °C
Thermocouple type: K (NiCr – Ni)
Thermocouples dimensions: ∅ 1 mm 1000 mm
Dimensions of compensating cables: ∅ 3 mm 1,00 m
Temperature controller: “JUMO” type IMAGIO 500, 3 channel, resolution 0,01 °C
Type of temperature control: thyristorial
Thyristor switches: “JUMO” type: TYA −110/3, 3 pcs.
Digital Thermometer Readout with access to the computer: “FLUKE” CHUB E-4 type: 1529/A;
measurement range: from −200 °C to 1300 °C, 4 Channels
Resolution of Readout: 0,0001 °C
Multi-channel Data Acquisition Software: “FLUKE” LogWare II
Printer of the temperature: digital ENDRESS & HAUSER – Software ReadWin 2000
Electric cabinet for electrical supply and connectivity, signalling and security of electrical
installations
Dimensions of electrical cabinets: 70 cm 50 cm 26 cm
Thermal imager: FLUKE – Ti32 IR Fusion Technology
Control of ambient conditions: Digital thermo-hygrometer T&D Corp. Japan – TR-72U
Ambient conditions: 23 ± 3 °C
making apparatus, purchase of equipment and parts that cannot be made at the Institute
or close specialized companies in the field of precise machining. In Table 1 are given
structural, working and operational characteristics of the prototype apparatus.
Based on work drawings and mentioned specification within the Institute and other
companies the certain parts of apparatus construction have been made and the purchase
of specified equipment and materials that are built in the apparatus is done.
At Figs. 4, 5, 6 and 7 are presented the key parts of apparatus [7, 8].
Fig. 4. IMAGO 500, Fluke 1529 CHUB-4 in Fig. 5. Electric cabinet with built-in measur-
working conditions at the calibration point of ing devices and thyristor switches for electrical
600 °C. supply and security of electrical installations
Fig. 6. The probe and readout of contact Fig. 7. Prototype apparatus for calibrating of
thermometer at calibration point of 600 °C contact sensors for measuring temperature of a
solid surface
4.1 The Results of Testing Prototype Apparatus

In order to gain relevant and valid results of planed testing the program was made,
specified parameters and variables of measuring system and prepared proceeding
according to which the test is going to be performed in aim to prove measuring
capabilities and characteristics of prototype apparatus. All conditional, process and
outgoing information are taken into consideration and those have assured in the end to
express measurement uncertainty of apparatus with the reliability of at least 95%.
Test run of the apparatus, conducted before a test procedure. The corrections and
optimization of working and regulating parameters were done by which the best results
have been achieved and enabled by usage of the available instrumentation and control
technology incorporated in the apparatus. This additional optimization of system has

enabled excellent results of homogeneity and long-term stability of the reference sur-
face temperature, which is a key condition for the implementation of the calibration of
contact thermometers.
Due to thermal resistance, for each temperature measurement of a solid surface,
which includes the calibration of these criteria, it is needed much more time than in
other methods. This necessarily requires the stability of the temperature of the reference
surface, as well as probes for this type of calibration, at the time of temperature
readings. If we take into account the obligation of determining the reproducibility of
the measurement results under the same terms for each calibration point, it is clear how
important is the elimination of all the factors that cause distortions in the calibration
stability of the reference surface temperature [10]. This distortion increases exponen-
tially at higher temperature calibration points.
By controlling the air temperature above the reference surface and by preheating
the probe to an approximate surface temperature as well as eliminating the influence of
the operator on the value of temperature reading, the contributions to the measurement
uncertainty that appear in the budget are significantly eliminated or reduced.
By performing testing in 7 calibration points it was confirmed: long-term stability
of the temperature of the reference surface, stationarity and homogeneity of the tem-
perature field of the reference body, the ability to achieve higher temperature of
ambient air directly above the reference surface which is approximate to the temper-
ature of the surface, the ability to preheat the probe before contact to the approximate
value of the surface temperature and the lack of disruption of surface temperature after
contact with the probe.
The above technical improvements allow increasement of the reliability and
accuracy of the calibration results. The following test program, before turning on and
beginning of the process of heating the reference body to the set calibration point, the
control of ambient conditions in laboratories and test and adjustment by procedure
defined parameters of measuring and regulating parts of the apparatus, is made.
By entering into the software for data acquisition FLUKE (LogWare II), the
parameters showing the results of loaded measurement results were set. By activating
the Start Log Session begins the process of recording and displaying the values in the
selected time interval on all activated channels. According to procedure, the test pro-
gram is performed in two parts: The first part relates to heating program to set up
calibration point and second to program of stabilisation at set up of calibration points
which depends on time needed to reach and keep minimal temperature gradient
between extrapolating values.
On the following pictures and diagrams are presented the results achieved from the
test apparatus and for the highest calibration point of 600 °C which was reached by
testing. Figure 8 presents the diagram of heating to the set calibration point of 600 °C
with shown parameters of statistic on Channel 1 which shows the extrapolation value
of the thermocouple that is the closest to the surface of the reference body (1 mm under
the surface). From the calculated statistics data of the last 10 readings shown on the
diagram it can be seen that the average temperature reached 599,705 °C and that the
standard deviation for the same number of data is 0,087 °C.
Fig. 8. Diagram of heating up to temperature of calibration point 600 °C
As seen from the given diagram the heating period up to the calibration point of
600 °C took about 4 h. This completed the first Log Session- “Warming up to a
temperature calibration point of 600 °C” and launched other Log Session- “Homog-
enization and stabilization of temperature.” The results achieved in this period are
shown in the diagram in Fig. 9.
This Figure shows a diagram of temperature stabilization at a given calibration
point of 600 °C where are 350 temperature readings. The interval between two read-
ings is 1 min. In this case, the assessment of stationarity of the system is shown on
Channel 4 which shows the extrapolated value of temperature from thermocouple that
is the most distant from the surface of the reference body. From the calculated data
statistics of all 348 readings shown on the diagram it can be seen that the average
temperature reached 600,455 °C and that the standard deviation for the same number of
data was r = 0,078 °C. Also during the “Session Log” phase of stabilization in the
Fig. 9. Diagram of phase of stabilisation at the achieved calibration point 600 °C

heating zone, 1 mm above the surface of the reference body was placed probe contact
thermometer. In the moment before putting the probe against the reference surface, the
probe has been preheated to an approximate value of calibration points from 595,9 to
596,4 °C. The diagram on (Fig. 9) shows that even at the highest calibration point is
possible to establish equilibrium of extrapolated temperatures and with that also the
stabile temperature of the surface in long time period (of 348 min) needed to perform
calibration under the same conditions (standard deviation r = 0,078 °C) for needed
number of repeated measurements predicted with calibration procedure, with satisfying
requirements during time between two repeating.
4.2 Statistical Analysis of Test Results of Prototype Apparatus

For each selected calibration point individually, throughout the calibration range of 50
to 600 °C, the statistical analysis of data is made in order to do the budget of mea-
surement uncertainty for prototype apparatus [10]. This suggests that a proper budget
of measurement uncertainty of calibration is calculated separately for each calibration
point, during usage of the apparatus for calibration of contact sensors because the
contributions of apparatus measurement uncertainty are different on all calibration
points. The basis for the implementation of the statistical analysis of the results stored
in the database system for data acquisition in the software FLUKE Log Ware II from
which they were extrapolated by known equations for calculation of surface
temperatures.
The article presents the results of a statistical analysis of all the calibration points
while the diagram (Fig. 10) shows only for calibration point of 600 °C.
Fig. 10. Diagram of stability of temperature T1, T2 i Tsurf. gained by statistical data processing
for calibration point of 600 °C
Table 2 shows the overall results of the statistical analysis of the data and calcu-
lations for all calibration points.
Table 2. The results of statistical analysis of data obtained by the tests carried out and the
budget, for all calibration points
Parameter Calibration points (°C)
50 100 200 300 400 500 600
1 Average value Tsurf.av. 49,46 99,50 200,03 299,26 398,25 498,16 599,11
2 Average value T1 49,45 99,50 200,03 299,27 398,26 498,17 599,12
3 Average value T2 49,48 99,54 200,07 299,34 398,37 498,37 599,23
4 Temperature Gradient (G) 0,002 0,003 0,004 0,007 0,010 0,017 0,020
5 StD (r) 0,01 0,012 0,013 0,024 0,047 0,058 0,078
By analysing the data, it can be concluded that the differences between the mean
values of extrapolated surface temperature Tsurf and the mean temperature T1 are neg-
ligible, due to the very small temperature gradient at all calibration points. These good
results are coming from the narrow range of standard deviation as a key parameter for the
evaluation of measurement uncertainty of Type A for apparatus.
It follows that without long-term proven stability of the surface temperature of the
reference body is not possible to reduce the measurement uncertainty of the method for
thermal resistance that occurs in the contact zone. Therefore, for accurate reading of
temperature from the indicator of contact thermometer that is calibrated need much
longer time under condition of temperature stability of the reference surface.
Based on work drawings and mentioned specification within the Institute and other
companies the certain parts of apparatus construction have been made and the purchase
of specified equipment and materials that are built in the apparatus is done.
At Figs. 5, 6, 7 and 8 are presented the key parts of apparatus.
4.3 The Budget of Measurement Uncertainty of Apparatus Prototype

Type A (contributions (ua1) and (ua2)) are applied at the repeated measurements results
of stabilized extrapolation temperatures T1 and T2 in the reference body and on the
basis of the calculated StD for each calibration point. These values are defined through
practical tests on a large number of readings whose results were obtained in real
operating conditions and detailed and fully analyzed in the report of the test performed
on apparatus, and as such will be treated as a type A contribution to the apparatus
measurement uncertainty.
Type B (contributions (ub1) up to (ub10))
(ub1) – Contribution due to approximation of surface temperature.
(ub2) – Contribution due to unhomogeneinity of temperature field
(ub3) and (ub4) – Contributions of measurement uncertainty extrapolated from TC
(ub5) – Contributions due to usage of compensation cables
(ub6) – Contribution of measurement uncertainty of a temperature digital

Thermometer
(ub7) – Uncertainty contribution of temperature calculation by statistical tool
(ub8) – Uncertainty contribution of measuring thickness of the reference body (e)
(ub9) – Contribution due to position of the opening in the axis of ref. body (hinf)
(ub10) – Contribution due to position of the opening in the axis of ref. body (hsup)
In the following Table 3, based on the tests performed, specified and quantified the
contribution, which define the uncertainty of measurement apparatus, an example of the
budget for calibration point of 600 °C is given. The same approach making budget of
measurement uncertainty is applied to all other defined calibration points in the process
of testing apparatus. The equations for extrapolation are used to calculate the surface
temperature and the built-in statistical tools [8, 9].
Table 3. The budget of measurement uncertainty of apparatus at calibration point 600 °C

Value Xi Evaluation Distribution St. Sensitivity Contribution
xi uncertainty. coefficient │ci│ │ciu(xi)│
u(xi)
T1 599,11 °C N 0,007 °C 0,330 0,0023
T2 599,23 °C N 0,006 °C 0,330 0,0020
a1-stab 0 °C P 0,0782 °C 0,330 0,0258
a2-stab 0 °C P 0,0708 °C 0,330 0,0234
b1-apr 0 °C N 0,001 °C 1,330 0,0013
b2-hom 0 °C N 0 °C – –
b3-kal-1 0,84 °C P 0,42 °C 0,330 0,1386
b4-kal-2 0,84 °C P 0,42 °C 0,330 0,1386
b5-komp. 0 °C N 0,056 °C 0,330 0,0180
b6-kal-chub 0,17 °C P 0,010 °C 0,330 0,0033
b7-Log 0,0244 °C P 0,0344 °C 0,330 0,011
b8 - e 50,0 mm N 0,10 mm 0,002 °C/mm 0,0002
b9 - hinf 1,0 mm N 0,10 mm 0,002 °C/mm 0,0002
b10 - hsup 12 mm N 0,10 mm 0,002 °C/mm 0,0002
Tsurf. 599,109 °C / / / 0,2587 °C
k=2 U = – 0,52 °C
4.4 Analysis of Result of Testing Apparatus

The best indicator for the analysis can be a comparison of the results of some of the
laboratories that participated in the interlaboratory comparison within EUROMET
Project 635, where the specified measurement uncertainty of apparatus is the one that
significantly affect the measurement uncertainty of calibration as a whole.
The Fig. 11 shows the diagram where the picturesque were compared the results
related to the measurement uncertainty associated to apparatuses in mentioned labo-
ratories and expressed measurement uncertainty of prototype apparatus.
As it can be seen for a prototype of the apparatus is expressed an expanded

uncertainty in the entire range up to 600 °C while the other apparatus are limited at data
up to 300 °C. This diagram is a key for understanding the innovation achievements
realized through the development of this apparatus.
As previously stated the first and so far only interlaboratory comparison on the
European level of metrology in the field of calibration of contact temperature sensor is
carried out through the project [5, 8]. The results were published in a report from which
data were used and presented in this diagram (Fig. 11).
Fig. 11. The uncertainty of measurement of apparatuses participating in the interlaboratory

Comparison project EURAMET 635 and prototype apparatus in the TM – laboratory of technical
measurements. Legend: PTB – Physikalish-Technische Bundesansalt, Germany, LNE –
Laboratoire National d’Essais, France, OHM – Orszagos Meresugyi Hivatal, Hungary
The estimated values of the measurement uncertainty can be compared primarily in

function of getting conclusion about very large dispersion in relation to the values
among laboratories and secondary in relation to the temperature of calibration point. As
it can be seen the maximum temperature at which the inter-laboratory comparison is
done is 300 °C. Concerning the results reported at that temperature, the question is
which values of measurement uncertainty would be expressed at temperatures up to
600 °C, as it is shown for prototype apparatus.
5 Conclusions
The process of development and the creation of a prototype apparatus for the cali-
bration of contact thermometers for temperature measurement of solid surface required
a study of various thermal processes that occur during the process of calibration of this
type of thermometer.
Special attention is devoted to explaining and defining static and dynamic errors
which can occur in these types of measurements for various reasons, whether it can be
on the non-stationary system or heat resistance that occurs between the surface whose
temperature is measured and the surface of leaning probe.
As an introduction to the design of the apparatus heating modes of reference body
were analysed, type of heating body, insulating materials with their thermo physical
properties, and also were analysed temperature control systems which are commonly
used in the calibration of the apparatus.
In preparation to make the prototyping apparatus there were analysed the differ-
ences between the existing apparatus in; the dimensions of the reference body, dis-
tances between isotherms neighbouring sensors, distance of the closest sensor from the
reference surface, types of sensors that are built into the reference body, the diameter of
the opening in which extrapolation sensors are placed in etc. Each of the above
mentioned parameters significantly affect the accuracy of the calculated temperature
and contributes to the measurement uncertainty of calculated reference surface
temperature.
It was used a systematic approach to the development process and the creation of a
prototype apparatus for the calibration of contact thermometers, from the project idea to
the functional testing of the finished product, with the definition of all steps in its
development. The program for testing prototype apparatus is made in which is
described the test procedure at each calibration point individually and the results are
given in the calibration point of 600 °C. Diagram shows the results of statistical
processing of data collected in test, based on which a comparative statistical analysis of
test results is done and the results are presented in the form of tables and diagrams.
In order to obtain a more reliable results, in a longer period of stable conditions for
performance of potential calibration, the reporting period has been extended to more
than two hours or more than 150 readings of which were statistically processed, which
is a sufficient time interval to perform calibration of measuring instruments with a
minimum measurement uncertainty. The results on all of the selected temperatures of
calibration range are extremely good.
The first time they give clear evidence of the stability of the temperature of the
reference surface for a long period of time, which is expressed through statistical
fluctuations in temperature of the reference surface during calibration.
There were specified and quantified contributes to the apparatus measurement
uncertainty, stated expanded uncertainty of apparatus on all calibration points in the
specified temperature range and made a comparative analysis of the results of the
prototype with the results of some of the NMI laboratory presented in the report on
interlaboratory comparison Thermometry 635. The expanded measurement uncertainty
for the prototype apparatus is expressed throughout the whole range up to 600 °C while
the other apparatus are limited to data up to 300 °C.
References
1. Carslaw, H.S., Jaeger, J.C.: Conduction of Heat in Solids, 2nd edn. Oxford University Press,
London (1973)
2. Kovacs, T.: Qualification of Temperature Field by Temperature Gradient. VDI Berichte,
pp. 275–279 (1998)
3. Gray, W.T., Finch, D.J.: Accuracy of temperature measurement. In: Temperature: Its
Measurement and Control in Science and Industry, vol. 4(2), pp. 1381–1392. Instrument
Society of America, Pittsburgh (1972)
4. Morice, R., Andras, E., Devin, E., Kovacs, T.: Proceedings of TEMPMEKO 2001, vol. 2,
pp. 1111–1116. VDE Verlag GmbH, Berlin (2002)
5. EUROMET Project No. 635 (Thermometry) – Comparison of the reference surface
temperature apparatus at NMIs by comparison of transfer surface temperature standards,
Final report, Emese Andras; OMH, National Office of Measures, Hungary, November 2003
6. Report of collaboration on surface temperature sensors calibration: bilateral comparison of
measurement between BNM-LNE (F) and OMH (HG); Morice Ronan, Andras Emese, April
2001
7. Edin, T.: Doctoral dissertation: Prototype Development Apparatus for Calibration Contact
Sensors for measuring the surface temperature, University of Zenica, Mechanical
Engineering faculty, December 2015
8. Terzić, E., Zaimović-Uzunović, N., Jarović-Bajramović, N.: Measuring the surface and the
internal temperature of the solid bodies. In: 3rd International Conference “New Technologies
NT-2016” Development and Application, Mostar (2016)
9. Arpino, F., Fernicola, V., Frattolillo, A., Rosso, L.: A CFD study on a calibration system for
contact temperature probes. Int. J. Thermophys. 30, 306–315 (2009). https://doi.org/10.
1007/s10765-008-0451-8
10. Morice, R., Andras, E., Devin, E., Kovacs, T.: Contribution to the calibration and the use of
surface temperature sensors. In: Proceedings of TEMPMEKO 2001, pp. 1111–1116 (2001)
11. Michaelski, L., Eckersdorf, K., McGhee, J.: Temperature Measurement. Wiley, New York
(1991)
12. Andras, E., Kemper, M., Morice, R., Wagner, M., Garcia, C., Chimenti, V., Weckstrom, T.,
De Groot, M., Ivarsson, J., Falk, R., Bojkovski, J., Derelioglu, A., Kalemci, M., Liedberg,
H.: Proceedings of TEMPMEKO 2004, vol. 2, pp. 1017–1023. LPM, FSB, Zagreb (2005)
13. Kirkup, L., Frenkel, B.: An Introduction to Uncertainty in Measurement. Cambridge
University Press, Cambridge (2006). ISBN 13 978-0-521-60579-3
Automatic Control of Tube Hydroforming
Process in Experimental Conditions
Edina Karabegović1 ✉ , Edin Šemić2 ✉ , and Safet Isić2 ✉

( ) ( ) ( )
1
Faculty of Technical Engineering Bihać, University of Bihać, Dr. Irfana Ljubijankića bb,
Bihać, Bosnia and Herzegovina
edina-karabeg@hotmail.com
2
Faculty of Mechanical Engineering, University “Džemal Bijedić” Mostar,
Maršala Tita bb, Mostar, Bosnia and Herzegovina
edin.semic@bih.net.ba, safet.isic@unmo.ba
Abstract. Automatic control of tube hydroforming process enables control of

the process by application of appropriate sensors for measuring fluid pressure,
axial force of the press and axial displacement. In addition, it also provides control
of the operation system of electric motor, hydraulic system for achieving fluid
working pressure (pump, multiplier, proportional valve…), hydraulic drive for
opening/closing hydroforming tools and axial punches. The paper presents an
example of introducing automatic control of the tube hydroforming process in
experimental conditions.
Keywords: Automatic control · Sensors · Hydroforming · Fluid pressure

Multiplier
1 Introduction
Tube hydroforming process in experimental conditions requires a longer time for

preparatory and final operations. In order to reduce the time, the design of tube hydro‐
forming system with automated process control was performed. The introduction of
automatic process control requires the analysis of the process course, as well as the
method of the tube hydroforming system. This involves selection of process control
devices (sensors, actuators) and design of control unit (electronic components) for auto‐
matic process control according to the real conditions defined by the program. The paper
provides an example of the automation of tube hydroforming process for a T-shaped
tubes. Forming tubes of another form includes the introduction of additional components
in the hydroforming system (e.g. number of side supports, etc.) [1].
2 Tube Hydroforming in Experimental Conditions
The hydraulic system for tube hydroforming is designed to work in high fluid pressure
(oil) conditions. The application of the hydroforming system in experimental conditions
implies carrying out experiments for the purpose of scientific research (e.g. fluid pressure

https://doi.org/10.1007/978-3-319-90893-9_11
102 E. Karabegović et al.
control for different materials, different tube construction forms, etc.). Research in the
field of tube hydroformingis mainly conducted in laboratory conditions with manual
control of the hydroforming system, which requires adequate time for preparatory and
final operations during experimentation and difficult process control. The introduction
of automatic control would enable stability of the process, and simple and repeatable
experiments during hydroforming process [2]. The hydraulic system for tube hydro‐
forming in experimental conditions, Fig. 1, consists of:
• system for conducting the tube hydroforming process (tool, tube, axial punch, radial
holder, and fluid),
• process control system: electronic process control device, computer, measuring
amplifier device and parameter value measuring device (force sensors – 7 pcs, fluid
pressure sensor – 1 pcs, sensor axial feed – 2 pcs and sensor radial feed –1 pcs) [3, 4],
• electric and hydraulic components and installations: electric motor (3 pcs), pumps
(3 pcs), multiplier (1 pcs), hydraulic cylinders (4 pcs), control valves (4 pcs), safety
valves (3 pcs), suction filters (3 pcs), oil tank, hydraulic pipes and tubes.
closing force sensors
upper die
process
control
sensor sensor system
axial radial holder
fluid radial holder radial axial
punch force sensor feed
pressure force
sensor tube
bottom die
sensor axial feed
Electrical and hydraulic

components with installations
Fig. 1. Scheme of tube hydroforming device
Measurement and control of process parameters using sensor is performed by appli‐

cation of measuring amplifier device “SPIDER 8” (manufacturer: Hottinger Baldwin
Messtechnick-HBM Germany), Fig. 2.
Automatic Control of Tube Hydroforming Process 103
Fig. 2. Measuring amplifier device “SPIDER 8” [5]
The device is optimized for computer operation and has 8 independent measuring
channels.
2.1 Experiment Flow
The tube hydroforming process implies the action of pressure fluid inside the tube,
synchronized with the action of axial presses, until forming of workpiece is finished.
The experimental flow for tube hydroforming is given in Fig. 3 [5].
START
System start
(activation of the pump)
Monitoring fluid pressure

growth and control
(possible pressure regulation)
Activation of axial punch
NO Has the sensor feed

registered the feed?
(process start)
YES
The control pressure growth
(possible pressure regulation)
Has the sensor feed

NO
registered the end of
forming process?
(max. feed)
YES
Unloading the system
(return oil to tank)
Turning off the pump
Removing the finished piece
Saving experimental data
END
Fig. 3. Experimental flow during tube hydroforming

3 Automatic Control of the Tube Hydroforming System
The tube hydroforming process in experimental conditions, Fig. 5, is conducted using

an Arduino-based electronic device, which consists of hardware – a physical electronic
programmable circuit (microcontroller), and software (IDE - Integrated Development
Environment) that runs on your computer, used to write and upload computer code to
the physical board, Fig. 4.
Fig. 4. Arduino board [6]
The Arduino IDE uses a simplified version of C++, making it easier to learn to
program.
The control of the tube hydroforming process is performed according to the scheme
given in Fig. 5 [7].
The electronic component management system is designed to control and commu‐
nicate with external devices via a serial (COM) computer port [8]. Serial in (Rx) and
Serial out (Tx) pins are used for communicating Arduino board with the computer COM
port. The computer serial port is used for sending the commands, as well as for collecting
feedback from the electronic device, in order to control the hydroforming process [9].
The managing with components of the hydraulic system, i.e. switching the electric
circuits on or off, is conducted by means of a relay. The relay pins (IN1-IN8) need to be
connected to the Digital I/O pins on the Arduino board. The VCC pin is connected to
+5 V voltage from the external power supply. After receiving the instruction, the micro
controller using Digital I/O pins for management and regulation current circuits who
manage and control hydraulic installations.
Automatic Control of Tube Hydroforming Process 105
HYDRAULIC SYSTEM FOR HYDROFORMING TUBES

CONTROL OF HYDROFORMING SYSTEM
computer
THE PROCESS OF HYDROFORMING TUBES
sensor radial feed
radial holder force sensor
sensor force measuring- electronic device
radial holder (tool close) amplifying (the control
axial punch die device hydroforming system)
sensor axial feed

sensor tube
axial
force pressure sensor
ELECTRICAL AND HYDRAULIC COMONENTS WITH INSTALLATIONS

cylinders cylinder of cylinder for
multiplier for axial punches radial holder tool close
proportional
valve
safety
M M M valve
pump
.... .... .... suction
filter
oil tank
Fig. 5. The Scheme of hydraulic system for tube hydroforming
4 Conclusion
The development of production processes requires numerous researches that are based
on the application of knowledge from different fields of education. The hydroforming
process has a wide application in the automotive and aerospace industry, where product
development changes are steadily increasing. The integration of IT and electronics in
automation has a direct impact on the modernization and advancement of systems in
production processes. The introduction of automatic control in the process of tube
hydroformingin experimental conditions gives the possibility of better measurement and
faster implementation of new results. The aforementioned example of introduction of
automatic control of the hydroforming process enables the following:
• shorter preparation/finishing time of the experiment,

• stability of the process with easy execution,
• repeatability of the experiment,
• better control of the process and
• accurate measurement results.
Automatic control of the tube hydroforming process in experimental conditions will
certainly contribute to the development of automotive and other industries, which are
already directing their development towards intelligent automation of production
processes.
References
1. Singh, Harjinder: Fundamentals of Hydroforming. Society of Manufacturing Engineers,

Dearborn (2003)
2. Metal Forming Handbook. Schuler (c) Springer, Heidelberg (1998)
3. http://ccd.uns.ac.rs/aus/autIND/sau_doc/Za%20sajt/07_Meraci_pritiska.pdf
4. https://www.fer.unizg.hr/_download/repository/Mjerenje-Pomaka.pdf
5. Karabegović, E., Šemić, E., Poljak, J.: Control of hydroforming system of welded metal sheets.
J. Technol. Plast. 42(2), 23–31 (2017). https://doi.org/10.24867/jtp.2017.42-2.23-22. ISSN
0354-3870- UDK 621.7
6. https://store.arduino.cc/arduino-uno-rev3
7. Dragiša, V.: Mašine za obradu deformisanjem II deo Hidraulične prese i čekići, Univerzitet u
Novom Sadu, Fakultet tehničkih nauka, Departman za proizvodno mašinstvo, Novi Sad,
decembar 2013, str. 55–59 (2013)
8. Šemić, E., Isić, S., Karabegović, E.: Application of mechatronic system in the automation
hydroforming process. In: Hadžikadić, M., Avdaković, S. (eds.) Advanced Technologies,
Systems, andApplications II. IAT 2017. Lecture Notes in Networks and Systems, vol. 28, pp.
1037–1043. Springer, Cham (2018). https://doi.org/10.1007/978-3-319-71321-2_82, e-ISBN
978-3-319-71321-2
9. Edin, Š., Edina, K.: Razvoj mehatroničkog sistema za hidrooblikovanje cijevi. In: 2nd
International Conference “NEW TECHNOLOGIES NT-2015” Development and Application,
Mostar, April 2015
Analysis and Determination of Friction
in Hydroforming Process of Cross Tube
Mehmed Mahmić ✉ , Edina Karabegović, and Ermin Husak

( )

mmahmic@gmail.com, edina-karabeg@hotmail.com
Abstract. In hydroforming process of cross tube significant influence has a fric‐

tion between tool and workpiece related to a tool load, stability of forming process
and success of obtaining final piece.
Although friction in hydroforming process is a lower intensity it affects the
increase of process parameters and contact stresses, which results in greater tool
load and decreasing its use span.
In this paper analysis of friction in hydroforming process of cross tube and
determination of friction coefficient using experimental results and simulation
with software simufact.forming is given.
Keywords: Hydroforming · Friction · Experiment · Analysis · Simulation
1 Introduction
In the plastic forming processes, which includes hydroforming, the friction between the die
and the workpiece significantly affects the stability of the forming process, the wear and the
life of the die, the energy consumption required for forming, the quality of the finished
piece. The conditions of contact friction are changeable during the forming process and
represent a complex analytical problem. Also, the numerical simulation of the forming
process depends largely on the definition of the initial conditions, which also includes the
coefficient of friction. The determination of friction can be carried out by setting the numer‐
ical simulation on the basis of experimental results of the forming process. Experimental
results can be related to models that are designed to determine the tribological conditions of
the forming process or the real part where the experimental conditions will be much closer
to the real design process. Though in the process of hydroforming, friction of low intensity
is still the area of research by a large number of researchers [1–5].
2 Analysis of Friction in Hydroforming Process
The tribological conditions between the workpiece-tube and die in the hydroforming
process, for the sake of easier understanding, need to be analyzed by zones [2]. Figure 1
shows three friction zones in the hydroforming process of cross tube: linear zone or
compression zone, transition zone and free expansion zone.

https://doi.org/10.1007/978-3-319-90893-9_12
108 M. Mahmić et al.
Expension zone
Fluid
Transition zone
Die Compression zone
Axial Axial
force force
Workpiece
Fig. 1. Friction zone in hydroforming of cross tube
The friction coefficient analysis in the hydroforming process that the authors in the
reference [2] deals with is the determination of the coefficient of friction in the compression
zone. The analytical model for determining the friction coefficient for the symmetrical
workpieces-tubes of the diameter (d0) and the initial height (h0), while the hydroforming
process for determining the coefficient of friction is given in Fig. 2, where: 1 - workpiece-
tube, 2 - die, 3 - upper movable puncher (force F1) and 4 - lower fixed puncher (force F2).
d0
s0 di
F1
3 F1
di1 s1 A1
1
puc h0
h
2
h0
4 puc
F·
F2
di2 s2
A2
F2
Fig. 2. Tube hydroforming process and forces in hydroforming process
( )
If it is assumed that stress in contact between tube and die is 𝜎k 𝜎k = puc [3] the analyt‐
ical model for friction coefficients is obtained:
Analysis and Determination of Friction in Hydroforming Process 109
{( ) [ ( ( )/ ( ))]n }
d02 − di12 ⋅ ln s1 d0 − s1 s0 d0 − s0 − ( )
1, 15k ( 2 ) [ ( ( )/ ( ))]n + puc ⋅ di22 − di12
d0 − di2 ⋅ ln s2 d0 − s2 s0 d0 − s0
2
(1)
𝜇=
4 ⋅ puc ⋅ d0 ⋅ h
where: s0 – initial tube wall thickness; s1 – tube wall thickness in contact with upper
puncher; s2 – tube wall thickness in contact with lower puncher; puc – fluid pressure in tube;
h – current high of tube; di1, di2 – tube inner diameters at upper and lower part of tube; di –
initial tube inner diameter [4].
Based on the Eq. (1) it is seen that the coefficient of friction at the hydroforming between
the workpiece-tube and the die of the functions of the following parameters:
( )
𝜇 = f s0 , s1 , s2 , h, d0 , puc , k, n (2)
where: (s0, s1, s2, h, d0) – geometrical parameters showed at Fig. 2; n – is defined with
experiment.
Since the analytical expression (1) contains the parameters to be determined experimen‐
tally, and only applies to the compression zone in hydroforming, a similar experiment anal‐
ysis should be carried out for both the transition zone and the expansion zone. The paper
will show the determination of the total coefficient of friction for all three zones.
3 Determination of Friction in Hydroforming Process
3.1 Experimental Investigation
Experimental investigations in this paper included the hydroforming of the cross tube
(Fig. 3) with the use of laboratory measuring equipment consisting of: hydroforming
machine (Fig. 4) and measuring equipment [5]. The measuring equipment consists of:
sensor for measuring the pressure of the fluid in the tube, the force measuring dynamom‐
eter on the axial punch (2 pcs.), sensor for stroke/displacement measuring, measuring –
amplifying equipment Spider and the computer.
Fig. 3. Cross tube Fig. 4. Hydroforming machine

Specimen dimensions: outer diameter of the tube 20 mm, length 80 mm and wall thick‐
ness 2 mm. The specimen material is steel Č.1212 and the lubricating oil is: INOL
HIDROL - X 46. The experimental results of the change of feed of axial puncher and
the pressure of the fluid in the tube depending on the time are given in Figs. 5 and 6.
Fig. 5. Fluid pressure in tube Fig. 6. Puncher stroke
3.2 Simulation of Hydroforming Process of Cross Tube

The numerical simulation of the hydroforming process of cross tube has been performed
using MSC.SuperForm, software simufact.forming. At the beginning of the simulation,
the required simulation data is entered: tool geometry and specimen geometry, work‐
piece material data, forming parameters according to experimental data, presumed fric‐
tion coefficient, Fig. 7.
INPUT DATA
- geometry
- material
- parameters, ...
Presumed friction coefficient
FEM simulation
Simulation results
Comparison with
experimantal data NO
YES
Friction coefficient μ
Fig. 7. Determination of friction coefficient using simulation of forming process

Analysis and Determination of Friction in Hydroforming Process 111
Figures 8 and 9 show the data entered in the simulation software for the hydroforming
process according to the experimental results, Figs. 4 and 5.
Fig. 8. Fluid pressure from simulation Fig. 9. Puncher stroke from simulation
The results of the simulation are shown, according to the drawing, in Fig. 7, for
reducing the tube to 52 mm, or the stroke of a 14 mm of axial puncher. In the hydro‐
forming process there is a change in the thickness of the workpiece wall, Fig. 10. At the
part of the workpiece in the compression zone, just outside the forehead of the axial
puncher, the wall thickening occurs, and in the expansion zone a wall thinning occurs.
Figure 11 shows the contact stresses in the process area of the die, between workpiece
and die.
Fig. 10. Wall thickness of workpiece Fig. 11. Contact stresses
The numerical simulation of the hydroforming process of cross tube process has been
performed using the MSC.SuperForm software based on the input parameters of the
process obtained by experimental measurement and the assumption of the friction coef‐
ficient according to the flow diagram (Fig. 7).
By analysing the workpiece obtained by an experimental experiment in which the
thickness of the pipe 2.2 mm, Fig. 3, in front of the forehead of the axial puncher and
the thickening of the workpiece in the same place obtained by numerical simulation of
the hydroforming process, Fig. 10, it is evident that the thickening coincide.
The assumed friction coefficient of simulation whose results were displayed had a
value of 0.05. From this it can be concluded that the hydroforming process is carried
out under favourable tribological conditions.
4 Conclusion
The tribological conditions between the workpiece-tube and the contact zone of die in
the hydroforming process have a significant impact on the forming process itself, on the
wear and life of the die, significantly on the energy consumption required for forming
and ultimately on the quality of the finished piece. In the analytical approach to hydro‐
forming of cross tube, friction must be observed in three zones. The analytical expression
for the friction coefficient, in the compression zone, contains the component that is
experimentally determined on the model. This means that for expansion zone and tran‐
sient zone is also necessary to determine analytical expressions and new experiments
on the model for the purpose of defining analytical expressions. The approach by which
the initial conditions and the results of numerical simulations are matched to the exper‐
imental results of the real model yields more accurate results. In this way, the die load
simulation is obtained in the contact zone between the workpiece-tube and the die. By
this approach the intensity of the total friction coefficient has been determined in the
hydroforming process of cross tube which corresponding to the experiment performed
in this paper.
References
1. Plančak, M., Vollertsen, F.: A contribution to determining the friction coefficient in

hydroforming of tubes. Tribotest 9(3), 219–230 (2003)
2. Ngaile, G., Yang, C.: Analytical model for the characterization of the guiding zone tribotest
for tube hydroforming. J. Manuf. Sci. Eng. 131, 021008-1 (2009)
3. Vollertsen, F., Prange, T., Sander, M.: Hydroforming needs. In: Geiger, M. (ed.) Developments
and Perspectives Advanced Technology of Plasticity, pp. 1197–1210. Springer, Berlin (1999)
4. Mamić, M., Karabegović, E., Jurković, M., Karabegović, I.: Analitički model određivanja
koeficijenta trenja kod hidrooblikovanja. In: 16th International Research/Expert Conference
on Trends in the Development of Machinery and Associated Technology, TMT, 10–12
September, Dubai, United Arab Emirates, pp. 55–58 (2012). ISSN 1840-4944
5. Mahmić, M.: Modeliranje i optimizacija alata za plastično oblikovanje primjenom
eksperimenta i baza znanja, Univerzitet u Bihaću, Tehnički fakultet Bihać, doktorska
disertacija (2012)
Science of Metals Through Lens of Microscope
Belma Fakić(&), Adisa Burić, and Edib Horoz
University of Zenica, Metallurgical Institute “Kemal Kapetanović”,

72000 Zenica, Bosnia and Herzegovina
{belma.fakic,adisa.buric,edib.horoz}@ikk.unze.ba
Abstract. Man atempts to know and see what is not visible with the naked eye
goes back to the far past. By the combination of concave and convex lenses, in
the 16th century, Dutchman Janssen made the first ever-increasing instrument.
Robert Hooke was first used to illuminate objects using artificial light and gave a
first description of plant cells still afar 1665. Henry Clifton Sorby, in the 1960s,
was the first to apply a microscope in metal testing. Today, microscopes are used
daily for the testing purity of metals and alloys, the presence of nonmetallic
inclusions, the microstructural phase present in the alloys as a result of various
forms of heat treatment. This paper gives an overview of the possibilities of light
microscopes in the testing of metal materials.
Keywords: Light microscope Metalic materials Steel Microstrukture
1 Introduction
The material science explores metallic and non-metallic materials. Metallurgy is part of
material science, separated as a special science, because of the extraordinary impor-
tance of metal in everyday life. Metal materials have a significant role in the economic
development of the state and the world in general. The most widely used in the modern
world are iron, carbon, chromium, nickel, molybdenum, steel and iron alloys. By
studying the basic properties of metal and alloys, connection between chemical com-
position, microstructure and the principle of the change in their structure and properties
is concerned with the physical metallurgy.
2 Metallography
As an independent science Physical metallurgy was started in the XIX century, initially
called “Metalography”. The term “Physical Metallurgy” (Germ. Physikalische Metal-
lurgie, Russ. Металловедение) was introduced in Germany in the 20s of the 20th
century, where it was proposed to preserve the term “Metallography” (Germ. Metal-
lografie, Russ. Металлография) only for science about the macro and microstructure
of metals and alloys. In many countries Physical Metallurgy is still referred to as
Metallography [1].
Metallography means metal and graphs - the description of metals is basically a
science of the structure of metals and alloys, and studies the principles of structure

https://doi.org/10.1007/978-3-319-90893-9_13
114 B. Fakić et al.
formation and examines the macrostructure and microstructure of metal, either by

observing the naked eye, at small increases either by light (optical) or electronic
microscope. In addition, X-ray diffraction microscopy (X-ray structure analysis) is also
used for the study of microstructure. Metallography also studies the changes of
mechanical, electrical, magnetic, thermal and other physical properties of metals
depending on the change in its structure. Structural examination is necessary to find the
“structure - property” connection and to establish the principle of structure formation to
predict the properties of new alloys.
2.1 First Testing of Metal Structure

The first study of the metal surface macrostructure, by etching, without additional
optical aids, historically, was carried out by Reomir (1683-1757). He used different
types of reagent to distinguish different type of steels.
By etching in reagents, in 1774, Rinman recognized the damask steel - a steel for
making a sword (a type of sword was forged in Damascus; and in the Balkans in
Prizren and Sarajevo until 1878).
The Russian scientist P.P. Anosov (1799–1851) first applied a special metallographic
microscope, with work in reflected light on specially prepared sample surfaces, for
studying the structure of steel and its changes after forging and heat treatment, and
established the existence of a connection between the structure and properties of steel [1].
2.2 First Examination of Metal Structure Under Microscope

Henry Clifton Sorby, in 1863 in England. Was the first person who tested polished and
crumpled metal samples under the microscope. H.C. Sorby was a geologist, a petro-
graph, and a minerologist. His samples were Swedish wrought iron and wrought iron
armour plate, and blister steel [2, 3]. Representation of the microstructure of steel
observed by Sorby at the 9x increase is given in Fig. 1. This macrograph shows distinct
grain boundaries.
Fig. 1. The first macrograph of blister steel, etched in very dilute nitric acid, 9x [2, 3]
Science of Metals Through Lens of Microscope 115
A few years later, A. Martens (1878) made the first metallographic samples and
recorded the structures of steel and iron under a larger magnification in Germany.
Then, F. Osmond in France set up significant basics of metal science, especially on
steel quenching. W.C. Roberts Austen made the first diagram of iron-carbon system in
England at the end of the 19th century. The American H.N. Howe has introduced the
names of microconstituents: ferrite, cementite, pearlite, hardenite, and Osmond names
martensite, trustite and austenite [2].
3 Microscope
The meaning of the word microscope is obtained from the Greek words - mikros =
small and skopein = see. The principle of complex microscope as an instrument was
developed by the German manufacturer of glasses Zacharias Jansen since 1590. The
inventor of the microscope is considered Anthony Van Leewenhoek, who first
observed microscopic life in a little of water. Although the microscope used more than
200 years for various tests, the first time for metal testing was used by H.C. Sorby in
the second half of the 19 th century [3]. The application of a microscope in the
investigation of the metal with later application compared to the beginning of the
application of the optical microscope in biology is obvious, since the metal microscopy
requires a careful preparation of the samples surface, and in addition, the metals are
untransparent and require observation in rejected light.
3.1 Human Eye

The human eye observes the wavelength of 360 to 780 nm, which corresponds to the
interval of color from violet to dark red. It is maximally susceptible to the green light
wavelength of 560 nm. Therefore, green filters are often used in optical microscopes to
obtain monochromatic images.
The ability to distinguish two separate points is called resolution (power of dis-
cernment). Two points lying next to each other can still be seen as separate when their
spacing, according to Abe, is:
d ¼ k=2nsina ð1Þ
where k represents the wavelength of light with which the pattern is lightened; n is the
refractive index between the eye and the object (for air n = 1); angle a represents a
angular aperture [2]. For the human eye it is sina 0.02. When using green light
(k = 560 nm) in air (n = 1) it is obtained that the resolving power of human eye is
d 0.02 mm. Accordingly, the naked eye can differentiate the dimensions of a few
tenth parts of millimeter [1]. It should be kept in mind that the resolving power of the
human eye varies considerably [4].
3.2 Light (Optical) Metallographic Microscope

Light (optical) microscope, in which the light reflected from the surface of the sample is
used, is called a metallographic microscope. There are two types of metallographic
microscopes: upright and inverted. On the upright microscope, the sample is placed
below lens, while the inverted microscope sample is set upside down the lens.
Several important features of light microscope are: lenses (objectives), eyepieces
(oculars), light source, and camera.
4 Possibilities of Light (Optical) Microscope
A major factor in metal testing with an light (optical) micro-

scope is the preparation of the sample surface. There are few
steps in preparation of samples [5]:
– cutting,
– mounting (if it is necessary),
– grinding – SiC grinding paper, grid: 120, 240, 480, 600, 1000
– polishing – using different clothes and diamond suspension, gradation 9, 6, 3 and
1 lm.
The aim is get mirror surface of sample.
Light (optical) microscope has few capabilities:
– observation in bright field,
– observation in dark field,
– polarization light,
– Differential interference contrast (DIC).
4.1 Different Views Through Ocular Microscope

Light (optical) microscope is used for:
– determination of contamination of non-metallic inclusions;
– determination type, size and distribution of graphite;
– microstructure of alloys after different heat treatment;
– determination size and form of grain;
– measure of microcracks, pores, cavities;
– measurement of thickness layer.
Different micrographs getting on light (optical) microscope OLYMPUS PMG3
which is located in Metallographic laboratory of Institute “Kemal Kapetanović” of
Zenica.
The same micrograph of low carbon steel, using different filters is shown at Fig. 2.
filter 0 filter IF 550 filter 0560
Fig. 2. Different filter
Based on amount of ferrite and pearlite in steel, the content of carbon can be
determined, what is shown at Fig. 3.
steel (0,04%C) steel (0,3%C) detail of pearlite
Fig. 3. Microstructure with different amount of carbon (NITAL)
Non-metallic inclusion contamination is shown at Fig. 4.
aluminates oxide sulfide
Fig. 4. Non-metallic inclusions, unetched
Microstructure of graphite, their different type, distribution and size are given at
Fig. 5.
dendritic type nodular type lamelar type
pearlite with cementite pearlite and phosphide eutectic ferrite around nodulas
and very fine pearlite
Fig. 5. Graphite (first line – unetched, second line – etched in NITAL)
Microstructure of carbon steel C60, after different heat treatment (normalization,

quenching and tempering) are given at Fig. 6.
normalized quenched in water quenched and tempered

pearlite and ferrite martensite tempered martensite
Fig. 6. Microstructure of steel C60 after different heat treatment (NITAL)
Different imperfections in weld joints of steel and pores in Zn alloy are shown at
Fig. 7.
structural steel Zn alloy wrought steel

microcrack (ADLER) micropores (unetched) porosity at grain boundaries
Fig. 7. Microcracks, pores, porosity

Metallic material with different chemical composition has different form of grain.
Grain size of stainless steel, steel 100Cr6 and brass are given at Fig. 8.
stainless steel bearing steel 100Cr6 steel 17CrNiMo6

austenite (Kalling) oxidation (Villela) McQuaid—Ehn (NITAL)
low carbon steel low carbon steel alloy of CuZn

grain size in bright field grain size in dark field grain size of brass
(NITAL) (NITAL) (FeCl3)
Fig. 8. Size and form of grain
Different types and thickness of layers can be seen at surface of metallic materials.
Figure 9 represents Zn and paint layer, decarburization and oxide layer.
layer of Zn and paint decarburization layer oxide layer, penetration

(unetched) (NITAL) (unetched)
Fig. 9. Thickness layer
5 Conclusion
Metallography is part of physical metallurgy and it is irreplaceable instrument for research

of metal and their alloys. Through the lens of light microscope can be seen all what is good
in metal and that what is bad. Can be seen what we did in previous steps of production –
metallography shows us did we use correct temperature and cooling rate during heat
treatment. Metallography is eyes and soul of science of metal and their alloys.
References
Books:
1. Marković, D.: Fizička metalurgija 1. Univerzitet u Beogradu, Tehnički fakultet u Boru (2013)
2. Šuman, H.: Metalografija. TMF, Beograd (1989)
3. Bramfitt, B.L.: Metallographer’s Guide, Practices and Procedures for Irons and Steels. ASM
International (2002)
4. Vander Voort, G.F.: Metallography Principles and Practice, Fourth printing. ASM Inerna-
tional, June 2007
5. ASTM E 3 -11 - Standard Guide for Preparation of Metallographic Specimens
Increase of Performance of Grinding
by Plate Circles
Tonkonogyi Vladimir(&), Yakimov Alexey, and Bovnegra Liubov
Institute of Industrial Technologies, Design and Management, Odessa National

Polytechnic University, Odessa 65044, Ukraine
{vmt47,dlv5}@ukr.net, nikolonpu@gmail.com
Abstract. A method is proposed for increasing the grinding efficiency of gear

wheels by two disk wheels. The possibility of increasing grinding capacity on
machines working in two circles by the method of bending without reducing the
accuracy of processing has been revealed. The conditions for increasing the
processing capacity are determined when placing the disc wheels in one and two
adjacent cavities of the treated wheel.
Keywords: Disc wheels Cross travel Angle of adjustment

Machining accuracy
1 Introduction
Grinding with disk circles is used as a finishing operation in the production of gears of
heavy loaded high-speed gears, which are usually made of surface hardened alloy
steels.
The machining by two disk circles can be carried out according to the schemes with
zero and with 15° profile grinding angles. The process of grinding is accompanied by
the release of a large amount of heat in the cutting zone, resulting in phase and
structural changes in the surface layer of the teeth of the wheels being treated, leading
to a decrease in operational reliability. The zero-degree method of gear grinding is
characterized by a greater thermal stress in conjunction with the 15° method. This is
explained by the fact that when grinding on the zero scheme, each point of the lateral
surface of the tooth is subjected to multiple thermal effects [1–6]. Despite the fact that
the 15° grinding method is less heat-stressed, it is rarely used because of low
productivity.
It is proposed in [7–10] to increase the productivity of gear grinding by using
circles of cubic boron nitride. The disadvantage of using such circles is the need for
expensive modernization of the machines to enable the adjustment of the wheels.
This predetermines the need to search for alternative technical solutions aimed at
increasing the productivity of the grinding process with disk circles.

https://doi.org/10.1007/978-3-319-90893-9_14
122 T. Vladimir et al.
2 Materials and Research Results
Angle of installation of grinding wheels on the gear grinding machine ai is equal to the
sum of the angles (Figs. 1 and 2):
ai ¼ a1 þ b ð1Þ
Fig. 1. Scheme of grinding with the arrangement of circles in adjacent cavities of the gear wheel
Fig. 2. Scheme of grinding with the arrangement of circles in a single groove in the gear wheel
Increase of Performance of Grinding by Plate Circles 123
Angle a1 is determined by the formula:
amax þ invamax arctgamax

a1 ¼ inv ; ð2Þ
2 2
R@:0 cos a
amax ¼ arccos : ð3Þ
R@:0 þ 1; 2 m þ 2
Angle b depends on the grinding method and is determined:

– when the grinding wheels are located in the adjacent cavities of the gear wheel
(Fig. 1) according to the formula:
p
b¼ ðinv20 inva1 Þ; ð4Þ
2z
– when grinding wheels are located in the same hollow of the gear wheel (Fig. 2),
using formula:
3p
b¼ ðinv20 inva1 Þ ð5Þ
2z
The length of the transverse stroke of the table depends on the angle of installation
of the circles, and the length of the transverse stroke of the table depends on the
machine time.
Analytical dependences are proposed for calculating the length of the transverse
course of the table and the angle of installation of the grinding wheels when they are placed
in one depression (6)–(8) and in adjacent valleys of the treated gear wheel (9)–(11).
ð6Þ
2qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
mz 2 ffi 3

þ 1; 2 m þ 2 r02 15
p cos 20
¼24 þ inv20 5 R@:0

ð7Þ
2
lx:eð15Þ
r0 57; 32 2 z cos 15
ð8Þ
ð9Þ
2qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 2ffi 3

mz
þ 1; 2 m þ 2 r0 15 3p cos 20
lx:eð15Þ ¼2 4 2
þ inv20 5
R@:0 ð10Þ
r0 57; 32 2 z cos 15
ð11Þ
Analysis of these dependencies shows that the length of the transverse stroke of the
table depends on the characteristics of the treated gear: the module, the number of teeth,
the radius of the dividing circle. With a decrease in the length of the transverse stroke
of the table, the inertia of the swinging parts of the machine decreases, the deformation
of the lines of the rolling mechanism decreases and, as a consequence, the accuracy of
machining of the gear wheel increases.
An analytical formula is proposed for calculating the number of double transverse
table moves per minute
ð12Þ
where lxc15 – length of idling; n15 – number of table swings, J15 – moment of inertia of
the swinging parts of the machine; Rr15 – radius of the rolling roll, corresponding to a
15° setting of the machine.
With a decrease in the length of the transverse stroke of the table, the number of
strokes per minute increases, and this leads to a decrease in the machine time
Lx
Tm ¼ zKM ð13Þ
nS
where Lx – length of longitudinal travel of table, m; S – longitudinal feed, mm/min; n –
Number of passes; K = 1 with bilateral separation; K = 2 at unilateral separation.
A formula has been derived for calculating the inertia force of the oscillating parts
of the machine, the analysis of which shows that by reducing the length of the
transverse stroke of the table it is possible to increase the number of its oscillations,
while maintaining the inertia of the swinging parts of the machine and the accuracy of
processing:
a)
b)
Fig. 3. Dependence of the length of the transverse stroke of the machine table (a) and machine
time (b) on the number of teeth and the machined wheel module

lx:c: J 2pn 2
Pj ¼ ð14Þ
R2p 60
where – J – moment of inertia of the swinging parts of the machine; n – the number of

double transverse table moves per minute; Rp ¼ mZcos20
cosau – radius of the breakout
roller.
In Fig. 3 shows the dependence of the length of the transverse stroke of the table
(a) and the machine time (b) on the number of teeth and modules of the wheels being
machined.
Analysis of these dependencies shows that the length of the transverse table travel
and machine time on machines with a 15° circle setting is greater than on machines
configured for the design angle of the producing strip. On average, the optimization of
the setting angle of the wheels reduces the machine time by 20%. If we take into
account that the total processing time is 6–7 h, then the reduction of machine time by
20% reaches 40 min.
It can be seen from the nomogram (Fig. 4) that if the gears differ in a different
combination of numbers of teeth and modules, then for each of them there is an
optimum angle for setting the circles corresponding to the minimum length of the
transverse stroke of the table. With the help of the developed nomogram, it is possible
to select the optimal angle of adjustment of the disc wheels for any wheel, in which the
machine operating time of the gear grinding will be minimal.
Fig. 4. Nomogram for selecting the angle of installation of disk circles ai.r on the machine
(circles are located in one cavity), at which the machine operation time will be minimal
3 Conclusion
1. The possibility of increasing the productivity of gear grinding on the machines
working with two disk abrasive wheels by the method of bending around the 15–20°
scheme without reducing the accuracy of processing by increasing the number of
swings of the table and turning of the grinding wheels to the calculated angle at
which the inertia of the swinging parts of the machine decreases, as a result, the
deformation of the rolls of the rolling mechanism decreases.
2. A nomogram was developed with the help of which it is possible to select for each
wheel the optimal adjustment angle of the disk wheels, in which the machine
operating time of the gear grinding will be minimal.
References
1. Ryabchenko, S.V.: The processes of grinding cogwheels with disk circles from SHM,
optimization of modes. In: Superhard Materials, Production and Application: Monograph in
6 b. General editing. of N.V. Novikov. -T.6 Diamond-Abrasive Tools in Fur Technologies.
Processing. Editing of A.A. Shepelev. ISHM n. V.M. BakulNAS of Ukraine, Кiev, pp. 287–
299 (2007)
2. Ryabchenko, S.V.: Ensuring the quality of machining of gears during tooth grinding with
circles of KNB. In: Modern Processes of Mechanical Processing with Tools with NTM and
Quality of Surface of Details of Masks: Collection of Sciences Works. ISHM name of V.M.
Bakul NAS of Ukraine, К., pp. 80–87 (2009)
3. Ryabchenko, S.V.: The state of the surface layer of steel with precision gear grinding with
CBN circles. In: Surface Engineering and Renovation of Products: Mat. 7th International.
Scientific-Techn. Conf., 29–31 May 2007, Yalta. ATM of Ukraine, К., pp. 162–164 (2007)
4. Ryabchenko, S.V.: Quality of processing of gear wheels after grinding from KNB. In:
Quality, Standardization, Control: Theory and Practice: Mat. 9th International.
Scientific-Techn. Conf., 2009, Yalta. ATM of Ukraine, К., p. 228 (2009)
5. Yakimov, A.V.: Investigation of thermal processes during grinding of gear wheels with bowl
circles. In: Yakimov, A.V., Ryabchenko, A.V., Shmat, I.A. (eds.) Equipment and Tools to
Professionals: A Series of “Metalworking”, № 3(148), pp. 56–58 (2012)
6. Yakimov, A.A.: Improvement of the quality of the surface when grinding gears on machine
tools MAAG. In: Yakimov, A.A., Perevezentseva, A.Y., Angel, A.F., et al. (eds.) Equipment
and Tools for Professionals: The Series “Metalworking”, № 4(168), pp. 88–90 (2014)
7. Ryabchenko, S.V.: Grinding of cogwheels with disk circles. In: Equipment and Tools for
Professionals: The Series “Metalworking”, № 4(168), pp. 44–49 (2014)
8. Ryabchenko, S.V.: Increase the efficiency of grinding high-precision gears in circles from
the KNB. In: Instrument Reliability and Optimization of Technological Systems: Collection
of Scientific Papers. DGMA, Kramatorsk, issue 23, pp. 95–102 (2008)
9. Ryabchenko, S.V.: Development of technology for grinding cogwheels with disk circles
from SHM. In: Modern Processes of Machining with Tools from SHM and the Quality of
Surfaces of Machine Parts: A Collection of Scientific Papers, a Series of “Processes of
Machining of Machine Tools”. IMS them. Bakul NASU, Kiev, pp. 161–168 (2006)
10. Ryabchenko, S.V.: Grinding of cogwheels with disk circles from STM. In: Superhard
Materials: A Collection of Scientific Papers, №. 6, pp. 81–98 (2014)
Analysis of Torsional Vibration of the Engine
Connected with Propeller Through Pair of Gears
Ermin Husak ✉ and Erzad Haskić

( )

erminhusak@yahoo.com
Abstract. The determination of natural frequencies and the mode shapes of

vibration of discrete and continuous systems is one of the basic tasks that are
solved in the vibrations of these systems. By knowing these values, it is possible
to avoid unwanted occurrences in the system. In this paper the torsional vibrations
of the shafts of a marine engine connected with a propeller through pair of gears
are analysed. A well-known process of transforming gear system with a multiple
shafts to a single shaft or torsion chain with definite number of degrees of freedom
has been carried out. From this analysis natural frequencies and mode shapes of
vibration have been obtained. Then, a numerical analysis of the torsional vibration
of the shafts of the engine connected with propeller through a pair of gears has
been carried out using the ANSYS software.
Keywords: Torsion · Vibration · Natural frequencies · Mode shapes
1 Introduction
The main task of the designer is to define the final shape of the product or device so that
the system is as far away from the critical conditions during exploitation. These critical
conditions in vibration can be described by the appearance of imbalance, large ampli‐
tudes, the appearance of a gyroscopic effect, etc. Shafts as one of the main structural
elements are exposed to different forms of dynamic loads, which with the characteristic
of the shaft can cause critical conditions. That is why shaft vibration analysis is part of
the analysis of great importance. Shafts with multiple rotors can be modelled as discrete
systems in the form of free or fixed torsion chains from which simple analysis can yield
the value of natural frequencies and mode shapes of vibrations. The specific form of this
analysis is the torsional vibration of gearbox shafts where the multi-shaft gearbox is
transformed into one free or fixed torsional system with multiple rotors. In this case
more reduction of the moments of inertia and stiffness of the shafts in the system is
required. As another possible way of analysing such systems is the use of commercial
software for numerical analysis. An example for the analysis in this paper is a marine
engine connected with propeller through pair of gears. The analysis has been carried out
by transforming this system to the torsional chain as well as the numerical analysis by
the commercial software ANSYS [1–4].

https://doi.org/10.1007/978-3-319-90893-9_15
Analysis of Torsional Vibration of the Engine Connected with Propeller 129
2 Analysis of the System Transformed to Fixed Torsional Chain
Figure 1 shows schematic diagram of the marine engine connected with propeller
through pair of gears. Moments of inertia in kg ⋅ m2 are: flywheel 9000, engine 1000,
gear (1) 250, gear (2) 15 and propeller 2000. It is necessary to find natural frequencies
and mode shapes of the torsional system.
Flywheel Engine Gear 1
Shaft I Shaft II Propeller
Gear 2
Fig. 1. Schematic diagram of the marine engine
Since moment of inertia of flywheel is considerably larger compared to other compo‐

nent moments of inertia, the flywheel can be considered to be stationary. The whole
system is transformed to the two degree of freedom model which is shown on Fig. 2.
Moment of inertia of rotor 1 is composed of moment of inertias of engine, gear 1 and
reduced moment of inertia of gear 2. Rotor 2 represents reduced moment of inertia of
propeller [5, 6].
Fig. 2. Two degree of freedom model

130 E. Husak and E. Haskić
Pair of gears with number of teeth z1 = 40 and z2 = 20 ratio
z2 𝜑z 1 𝜑∗p 1 20 1
i= =− =− = = = , (1)
z1 𝜑z 2 𝜑p k 40 2
where reduction coefficient k value is
k = 2. (2)
Taking into account that the kinetic and potential energy of the reduced part of the
system remain unchanged due to reduction, it can be written that
( )2
1 1
Ek∗ = Jp 𝜑̇ 2p = Jp∗ 𝜑̇ ∗p ,
2 2 ( )2 (3)
1 1
Ep∗ = c2 𝜑2p = c∗2 𝜑∗p .
2 2
Using Eq. (1) can be written
1
𝜑∗p = − 𝜑p , (4)
k
In order to satisfy the condition (3) the following equality must be satisfied
Jp∗ = k2 Jp = 22 ⋅ 2000 = 8000 kg ⋅ m2 ,

(5)
c∗2 = k2 c2 .
Substitute symbol is included for reduced moments of inertia for both gears and
engine on the shaft 1, and denote it as
J1 = Jz1 + k2 Jz2 + JE = 250 + 22 ⋅ 15 + 1000 = 1310 kg ⋅ m2 . (6)

( )
If shear modulus for steel is G = 8 ⋅ 1010 N∕m2 then stiffness of the shaft 1 and shaft
2 can be calculated according following equations:
( ) ( )
G ⋅ J01 G 𝜋 ⋅ d14 8 ⋅ 1010 𝜋 ⋅ 0, 14 Nm
c1 = = = = 981747, 704 ,
l1 l1 32 0, 8 32 rad
( ) (7)
∗
G ⋅ J02 2G
𝜋 ⋅ d24 Nm
c2 = k c2 = k
2 2
=k = 15904312, 81 .
l2 l2 32 rad
The kinetic and potential energy values for two degree of freedom model can be
determined from following Eq. (8):
[ ( )2 ]
1
Ek = J1 𝜑̇ 1 + Jp 𝜑̇ ∗p
2 ∗
,
2
[ ( )2 ] (8)
1 ∗ ∗
Ep = c1 𝜑1 + c2 𝜑p − 𝜑1
2
.
2
Differential equation can be derived from the Lagrange’s Equation (9)
d dEk dEp
+ = 0,
dt d𝜑̇ 1 d𝜑1
(9)
d dEk dEp
+ = 0.
dt d𝜑̇ ∗p d𝜑∗p
After kinetic and potential energy from Eq. (8) is included in the Lagrange’s Equation
(9) differential equations of two degree system motion are obtained
( )
J1 𝜑̈ 1 + c1 𝜑1 + c∗2 𝜑1 − 𝜑∗p = 0,
( ) (10)
Jp∗ 𝜑̈ ∗p + c∗2 𝜑∗p − 𝜑1 = 0.
Respecting procedure of solving Eq. (10) following equations are obtained

( )
c1 + c∗2 −(J1 𝜔2 A1 −)c∗2 A2 = 0,
(11)
−c∗2 A1 + c∗2 − Jp∗ 𝜔2 A2 = 0.
This equation system will have trivial solutions for A1 = A1 = 0, but these solutions
represent a system without motion, so such a solution is not necessary to consider. There
are nontrivial solutions if the determinant of the system is zero, i.e. if it is
|( ) |
( 2 ) | c1 + c∗2 − J1 𝜔2 ( −c∗2 ) |
Δ 𝜔 = || | = 0,
| −c∗2 c∗2 − Jp∗ 𝜔2 || (12)
| |
or in developed form
( ) ( )( )
Δ 𝜔2 = c1 + c∗2 − J1 𝜔2 c∗2 − Jp∗ 𝜔2 − c∗2 2 = 0, (13)
respectively
( ) ( )
Δ 𝜔2 = J1 Jp∗ 𝜔4 − Jp∗ c∗2 + Jp∗ c1 + J1 c∗2 𝜔2 + c1 c∗2 = 0, (14)
from which follows

( ) [( )2 ]1
∗ ∗ ∗ ∗ ∗ 2
Jp∗ c2 + Jp∗ c1 + J1 c2 ± Jp∗ c2 + Jp∗ c1 + J1 c2 − 4J1 Jp∗ c1 c2
𝜔21 , 𝜔22 = ,
2J1 Jp∗
[( )2 ]1
1, 559231339 ⋅ 1011 ± 1, 559231339 ⋅ 1011 − 6, 545398268 ⋅ 1020 2
𝜔21 , 𝜔22 = ,
209, 6 ⋅ 105
𝜔1 , 𝜔2 = 7439, 080816 ± 7338, 258358
2 2
𝜔21 = 100, 822 respectively 𝜔1 = 10, 041 rad∕s

𝜔22 = 14777, 339 respectively 𝜔2 = 121, 562 rad∕s.
In order to determine the mode shapes of vibration, it is first necessary to determine

the amplitude ratio 𝜂1 i 𝜂2 from Eq. (11)
( )
A2 c1 + c∗2 − J1 𝜔21
𝜂1 = = = 1, 053
A1 1 c∗2
( ) (15)
A2 c1 + c∗2 − J1 𝜔22
𝜂2 = = = −0, 155
A1 2 c∗2
Mode shapes can be determined from following

{ }
{ } 1 1
𝛷1 = = ,
𝜂1 1, 053
{ } (16)
{ } 1 1
𝛷2 = = .
𝜂2 −0, 155
3 Analysis of the System in ANSYS
After the analytical results were obtained, a numerical analysis of the torsional vibration
of the shaft of a marine engine connected to the propeller through a pair of gears was
performed using the ANSYS software. Unlike the analytical procedure where the mass
is concentrated in the rotors and the shafts represent an elastic system which is then
viewed as a system with two degrees of freedom, the numerical analysis process in
ANSYS is somewhat different.
The whole system is modelled with its real geometric character and appropriate
material, which is then discredited to the finite elements. In addition to setting appro‐
priate boundary conditions and solving, the results of natural frequencies and mode
shapes have been obtained. The first value of the natural frequency is 1.8106 Hz which
corresponds to 11.376 rad/s while the first mode shape can be seen in Fig. 3.
Fig. 3. First mode shape
The second natural frequency has a value of 18.65 Hz which corresponds to

117.181 rad/s while the second mode shape is shown in Fig. 4.
Fig. 4. Second mode shape
4 Conclusion
The development of contemporary products follows 3D graphics preserved in some of

the appropriate formats. The advantage of software for numerical analysis is the ability
to import these formats, making it easier to further analysis. This particularly applies to
systems that have multiple structural elements that are interconnected by disjoint and
unbreakable connections. In this case, particular attention should be paid to the contacts
between these elements and their status. The results obtained in ANSYS are almost
identical to the results obtained analytically. Such results were obtained only after
successfully defined contacts between the elements of the system.
References
1. Karabegović, I., Novkinić, B., Husak, E.: Experimental identification of tool holder
acceleration in the process of longitudinal turning. J. FME Trans. 43(2), 131–137 (2015)
2. Radosavljević, Lj.B.: Theory of Oscillation. Mašinski fakultet Beograd (1972)
3. Özkal, F.M., Cakir, F., Arkun, A.: Finite element method for optimum design selection of
carport structures under multiple load cases. Adv. Prod. Eng. Manage. 11(4), 287–298 (2016)
4. Husak, E., Kovačević, A., Rane, S.: Numerical Analysis of Screw Compressor Rotor and
Casing Deformations, Advanced Technologies, Systems, and Application II, Lecture Notes in
Networks and Systems, pp. 933–940. Springer (2018)
5. Benson Tongue, H.: Principles of Vibration, 2nd edn. Oxford University Press, Oxford (2002)
6. Karabegović, I., Husak, E., Pašić, S.: Vibration analysis of tool holder during turning process.
In: 14th International Conference Mechanika 2009, Kaunas, pp. 200–204, 2–3 April 2009
FEM Model of Misaligned Rotational
System with Rotating Looseness
Emir Nezirić1 ✉ , Safet Isić1, Isak Karabegović2, and Avdo Voloder3

( )
1
‘‘Džemal Bijedić’’ University of Mostar, 88000 Mostar, Bosnia and Herzegovina
emir.neziric@unmo.ba
2
Technical Faculty, University of Bihać, Bihać, Bosnia and Herzegovina
3
Faculty of Mechanical Engineering, University of Sarajevo, Sarajevo, Bosnia and Herzegovina
Abstract. Importance of knowing characteristics of rotational systems vibra‐

tions is significant, since it is main cause of machine failure. Vibration analysis
could facilitate determination of the machine fault. As a part of the research on
machine faults and its reflection on the machine behavior, numerical modeling
could give excellent starting information for future research. Finite element model
of rotational system motor – flexible coupling – rotor is presented in this paper.
As a fault analyzed in this paper, misalignment and rotating looseness are used
and modelled as external loads. For this modelled rotational system is shown that
it is suitable for analysis of rotational machinery vibrations.
Keywords: Shaft misalignment · FEM · Vibrations · Frequency spectrum

Waveform
1 Introduction
Misalignment is the most often fault of the rotating machinery [1]. That is the reason to
get familiar with its behaviour and impact on the rotational machinery. It is possible to
detect misalignment by vibration measurement, since vibration characteristics of
misaligned rotational systems are well known in practical maintenance and are often
topic in the scientific researches. Main vibration characteristic is 1X and 2X order of
rotational frequency in the frequency spectrum, and 3X up to 10X orders for severe
misalignment [1–5].
Another fault of the rotating machinery which have similar frequency spectrum
characteristics is rotational looseness. To determine how each of the faults is affecting
on the frequency spectrum and other characteristics of vibrations, it is necessary to
investigate vibrations on the rotational machine with both of those faults present.
First step in researching combination of those mentioned faults in the rotational
machinery should be to set up appropriate theoretical model and carry out research on
the model. One of the best ways to define model is the finite element method (FEM). In
this paper an FEM model of rotational system motor – flexible coupling – rotor with
misalignment and rotational looseness as faults is presented. It is possible to solve this
presented model with physical characteristics of the real rotational system to obtain
vibrational data, which could be analysed.

https://doi.org/10.1007/978-3-319-90893-9_16
136 E. Nezirić et al.
2 Basic Assumptions of the Model
System that is going to be modeled consists motor and rotor with flexible shaft connected
by flexible coupling. To simplify the model, motor shaft and motor bearings are consid‐
ered as rigid. Shaft with rotor is flexible and is represented with 2D beam elements
connected in series. First node is on the beginning of the rotor shaft, and it is connected
by flexible coupling to motor shaft. One node which is on 1/3 of the shaft length is
flexible bearing where rotating looseness would be introduced. In last node on shaft,
second flexible bearing would be defined. Flexible coupling and flexible bearings are
going to be introduced to model as external loads, which are dependent on node
displacement.
Parallel and angular misalignment would be used as faults which would cause lateral
vibrations of shaft. In the bearing closer to the coupling would be defined rotating
looseness. Shaft node is free to move inside bearing, until its displacement exceed value
of clearance between bearing outer ring and bearing housing (Fig. 1).
Fig. 1. FEM model of misaligned rotational system with rotating looseness
3 Governing Equations
Kinetic and potential energy for 2D Euler-Bernoulli element could be derived, so it is

possible to determine mass and stiffness matrices using shape functions for element
shown on Fig. 2.
Fig. 2. Euler-Bernoulli 2D beam element with parallel and angular misalignment

FEM Model of Misaligned Rotational System with Rotating Looseness 137
3.1 Centrifugal Force

Since this model is rotating, centrifugal force is going to be derived as external load.
Centrifugal force on element could be written as follows:
qx = 𝜌A𝜔2 ((𝛿 + 𝛼L1 + p) cos(𝜔t) + ux (z)) (1)
where ω is angular velocity of rotating element, δ is parallel misalignment, α is angular

misalignment, L1 is distance from coupling to node 1, p is product of α and z.
Centrifugal force vector could be obtained from total functional of potential energy.
Inserting Eq. (1) in force vector, following equation is obtained:
L
{𝐑 } =
e
[𝐍e ]T 𝜌A𝜔2 ((𝛿 + 𝛼L1 + 𝛼z) cos(𝜔t) + ux (z))dz (2)
∫
0
e
where [N ] is shape matrix for Euler-Bernoulli beam element. Those parts of centrifugal
force are obtained as:
{ e } me 𝜔2 (𝛿 + 𝛼L1 ) cos(𝜔t) [ ]T { e } me ω2 𝛼 cos(𝜔t) [ ]T
𝐑Ω𝛿 = 6 L 6 −L , 𝐑Ω𝛼 = −12 −L 12 −L
12 60
⎡ 156 22L 54 −13L ⎤

[ ] me 𝜔2 ⎢ 4L2 13L −3L2 ⎥
𝐊eΩ = ⎢
420 ⎢ 156 −22L ⎥ (3)
⎥
⎣ sim. 4L2 ⎦
where me is element mass equal to me = ρAL. Since centrifugal force depends upon
variable distance of element from rotation axis, stiffness matrix of centrifugal force is
separated from centrifugal force vectors. Separating stiffness matrix is done because it
would simplify solving the equation of motion.
3.2 Generalized Forces and Moments
Generalized forces and moments are loads which occurs in nodes where bearings and
coupling are introduced. That loads are acting only in one node and could be obtained
by differentiating of potential energy on matching generalized node coordinate.
Potential energy for node where first bearing is introduced could be written as
1 ( )2
V1 = k (𝛿 + Lb.1 sin 𝛼)(1 − cos(𝜔t)) + u1 (4)
2
where k is bearing stiffness, Lb.1 is bearing 1 to coupling distance and u1 is node displace‐
ment of bearing 1. Potential energy for node where second bearing is introduced (end
of shaft) is same as potential energy for the first bearing. Only thing that is changed is
coupling – bearing distance and node displacement. Bearing forces could be obtained
by differentiating the potential energy on corresponding bearing node displacement.
Potential energy of the flexible coupling could be written as shown on Eq. (5).
1 2
Vc = k d (sin(𝛼 + 𝜃) sin(𝜔t) + sin(𝛼 + 𝜃))2 (5)
2 c c
where kc is coupling stiffness, dc is coupling diameter and θ is angular displacement of
node where coupling is introduced. Coupling moment could be obtained by differenti‐
ating the potential energy on angular displacement.
Now, generalized forces and moments vector and matrix could be formed. Since
only elastic force and elastic moment in bearings and coupling are members of gener‐
alized vector, it is indexed as “elastic”.
⎧ 0 ⎫ ⎡0 ⎤
⎪ kc dc2 𝛼(sin(𝜔t) + 1)2 ⎪ ⎢ kc h ⎥
⎪ ⎪ ⎢ ⎥
⎪ ⋮ ⎪ ⋱
{ } ⎪ k((𝛿 + Lb.1. sin 𝛼)(1 − cos(𝜔t))) ⎪ [ ] ⎢ ⎥
k
𝐑el. = ⎨ ⎬, 𝐊el. = ⎢ ⎥ (6)
⎪
0
⎪ ⎢ 0 ⎥
⎪ ⋮ ⎪ ⎢ ⋱ ⎥
⎪ k((𝛿 + Lb.2. sin 𝛼)(1 − cos(𝜔t))) ⎪ ⎢ ⎥
⎢ k ⎥
⎪ 0 ⎪ ⎣ 0⎦
⎩ ⎭
2
where h = dc (sin(ωt) + 1)2. Elastic force matrix is diagonal matrix. Both elastic force
vector and matrix have only values at positions which corresponds to displacements of
nodes where bearings and coupling are introduced.
3.3 Rotating Looseness Modeling
Space between the bearing outer ring and the baring housing represents rotating loose‐
ness. Unloaded rotor shaft is taken as initial condition. While node which represents
bearing has smaller displacement than looseness gap, force in that bearing would be
zero. In the moment when bearing outer ring get in contact with the housing, the bearing
force starts to increase. During the contact between the bearing and housing, an impact
occurs. Potential energy for the bearing 1 could be written as
⎧ 1 ( )2
k (𝛿 + Lb.1 sin 𝛼)(1 − cos(𝜔t)) + u1 − x′ x < x′
⎪ 2 ( )2
V1 = ⎨ 1
k (𝛿 + Lb.1 sin 𝛼)(1 − cos(𝜔t)) + u1 − x′′ x > x′′ (7)
⎪ 2 x < x < x′′
′
⎩ 0
where xʹ = δ + Lb.1 sin α – γ, xʹʹ = δ + Lb.1 sin α + γ and γ is gap between the bearing outer
ring and the bearing housing. As it could be seen, there is no potential energy while the
bearing is moving inside the gap. Differentiating this potential energy would give us
force in node where bearing 1 is introduced.
3.4 Equation of Motion

From the derived expressions above, the equation of motion could be written as
{ } ( [ ] [ ]) { } { } { }
̈ + [𝐊] − 𝐊Ω + 𝐊el. {𝐃} = 𝐑Ω𝛿 + 𝐑Ω𝛼 − 𝐑el.
[𝐌] 𝐃 (8)
{ } { }
where 𝐃 ̈ and 𝐃̇ are acceleration and displacement vectors, [M] and [K] are mass
and stiffness matrices for lateral displacement, [KΩ] is centrifugal force stiffness matrix,
[Kel.] is diagonal matrix of bearings and coupling stiffness, {RΩδ} and {RΩα} are centri‐
fugal force vectors and {Rel.} is bearings and coupling elastic forces vector.
3.5 Solution for the Equation of Motion
Method of direct integration – implicit method would be used for solving the equation
of motion. This method is used since it is unconditionally stable. Its only limitation is
on time step size which impacts on solution accuracy. For every step of the solution
equation of motion is written. Since stiffness of the bearing with rotating looseness
depends on position of the bearing node, it is necessary to check does stiffness changes
in transition between two consecutive steps. Part of the algorithm which is used to
determine stiffness of the bearing with rotating looseness in each step is shown of Fig. 3.
Fig. 3. Algorithm for stiffness determining

4 Numerical Examples
Previously presented FEM model would be analysed with physical characteristics which
are taken from the real motor – coupling – rotor system. Physical characteristics are
shown in Table 1.
Table 1. Physical characteristic of model

Description Symbol Unit Value
Modulus of elasticity E Pa 200*109
Shaft diameter d m 0,02
Rotor diameter drot m 0,20
Shaft length L m 0,35
Rotor length Lrot m 0,05
L1 to coupling distance L1 m 0,10
Rotor to coupling dist. LR m 0,20
L2 to coupling distance L2 m 0,35
Material density ρ kg/m3 7800,00
Bearing stiffness k1, k2 N/m 20*106
Coupling diameter dc m 0,05
As faults in this rotational system would be used parallel misalignment (δ = 0,1–

5,0 mm), angular misalignment (α = 0,001–0,005 rad) and bearing gap (γ = 0–0,4 mm),
for which values would be analysed vibration characteristics.
4.1 Parallel Misalignment
Model equation of motion with the shown characteristics in Table 1 is solved for
displacement. On Fig. 4. is shown displacement of node 5 during 0,5 s period.
δ = 0,1 mm
0,006 0,00195
δ = 0,5 mm
δ = 1,0 mm 6,10352E-5
0,004 δ = 2,5 mm
δ = 5,0 mm
Amplitude [m]
1,90735E-6
Displacement [m]
0,002
5,96046E-8
0,000 1,86265E-9
-0,002 5,0
]
mm
2,5
t[
en
-0,004
nm
lig
1,0
sa
mi
-0,006
l el
0,5
ra
Pa
0,00 0,05 0,10 0,15 0,20 0,1

0 100 200 300 400 500
Time [s] Frekvency [Hz]
Fig. 4. Displacement (left) and Frequency spectrums waterfall plot (right) for node 5 (δ = 0,1–
5,0 mm)
As it could be seen, there is characteristic shape of the vibration diagram for all
parallel misalignment values. There are M and W shapes visible, where only difference
is amplitude of vibrations. Also on other nodes those shapes are noticeable. On Fig. 4
are shown frequency spectrums waterfall plot for the node 5 displacements for different
values of parallel misalignment.
As it could be seen in Fig. 4, frequency spectrum contains 1X (rotational frequency)
and its higher orders (2X, 3X, …, 9X). Presence of the orders of the rotational frequency
confirms source of the vibrations as misalignment. Shape of frequency spectrums are
not changing, only amplitude of each frequency is increased with parallel misalignment
increment.
4.2 Angular Misalignment

Model equation of motion with the shown characteristics in Table 1 is solved for
displacement values. On Fig. 5. is shown displacement of node 5 during 0,5 s period
and frequency spectrums waterfall plot for the node 5 displacements.
0,0010 1E-4
α = 0,001
α = 0,002 1E-5
Amplitude [m]
α = 0,003
α = 0,004 1E-6
0,0005 α = 0,005 1E-7
Displacement [m]
1E-8
1E-9
0,0000
]
rad
n t[
me
-0,0005
n
lig
sa
mi
lar
gu
An
-0,0010
0,00 0,05 0,10 0,15 0,20
0 100 200 300 400 500
Fig. 5. Displacement (left) and Frequency spectrums waterfall plot (right) for node 5 (α = 0,001–
0,005 rad)
As it could be seen on Fig. 5, presence of M and W shapes is noticeable. Also increase

of amplitude is present with angular misalignment increment. Other nodes also have
those characteristics in their displacement-time graphs.
Frequency spectrums for angular misalignment looks pretty much same as frequency
spectrum for parallel misalignment. Presence of rotating frequency with higher orders
is frequency spectrum main characteristics. Only difference is that higher orders have
higher amplitude for angular misalignment. Shape of frequency spectrums are not
changed with angular misalignment increment. Only amplitude is increased on
frequency spectrums.
4.3 Rotating Looseness

Model equation of motion with the shown characteristics in Table 1 and with 0–0,4 mm
of bearing gap is solved for displacement values during 10 s period. On Fig. 6. is shown
displacement of node 5 during 0,2 s period for different bearing gap values.
1E-3
0,0000
1E-4
Amplitude [m]
Displacement [m]
1E-5
-0,0025
1E-6
1E-7
-0,0050 1E-8
0,4
γ=0 0,3
]
mm
-0,0075
γ = 0,1
p[
ga
0,2
γ = 0,2
ng
ari
γ = 0,3
Be
0,1
-0,0100 γ = 0,4
0,0
0,00 0,05 0,10 0,15 0,20
0 100 200 300 400 500
Fig. 6. Displacement (left) and Frequency spectrums waterfall plot (right) for node 5 (γ = 0–0,4
mm)
Displacement of the node 5 is heavily impacted by bearing gap. Amplitude of the

displacement is increasing disproportionally to bearing gap increment. Also the shapes
of W and M are not present, even parallel misalignment is used as initial disturbance in
the rotational system.
With bearing gap increment, floor noise is also increased. Increment of floor noise
is hiding presence of the rotational frequency orders. The eigenfrequency amplitude to
floor noise ratio is decreasing with bearing gap increment, which means that unpredict‐
able behaviour of node inside the bearing gap is trying to hide all usual characteristics
of the misaligned rotational system.
5 Conclusions
Presented misaligned rotational system with rotating looseness model is capable to show
basic characteristics of analysed faults. According to the presented paper, it could be
concluded:
– Presented FEM model of rotational system have vibrational characteristics of the
analysed faults in rotational system. Consequently, those faults are properly
modelled.
– This model confirms presence of the M and W shapes in time waveform for parallel
and angular misaligned rotational system. It also confirms presence of rotational
frequency and it higher orders in the frequency spectrums for both misalignment
types.
– Presence of the rotating looseness in rotational system brings system in unpredictable

movement inside the bearing. That is confirmed by floor noise increment caused by
bearing gap increment.
– Presence of the rotating looseness in rotational system could hide some of the
misalignment characteristics (waveform shape M and W) or make them less notice‐
able.
– Presented model is possible to expand with other faults in rotational systems, so it
could be also used for analysis other faults and its combinations.
References
1. Piotrovski, J.: Shaft Alignment Handbook. CRC Press, New York (2007)
2. Xu, M., Marangoni, R.D.: Vibration analysis of a motor-flexible coupling-rotor system subject
to misalignment and unbalance, Part I: Theoretical model and analysis. J. Sound Vibr. 176(5),
663–679 (1994)
3. Nezirić, E., Isić, S., Doleček, V., Karabegović, I.: Vibration analysis of theoretical SDOF
model of shaft parallel misalignment. J. Technol. 5, 131–134 (2014)
4. Nezirić, E., Isić, S., Doleček, V., Karabegović, I.: Impact of bearing and coupling stiffness on
characteristics of rotary system vibrations. In: Proceedings of 11th International Symposium
on Stability, Vibration, and Control of Machines and Structures, Belgrade, 3–5 July (2014)
5. Redmond, I.: Shaft misalignment and vibration – a model. In: Proceedings of IMAC-XXV:
Conference & Exposition on Structural Dynamics, Orlando, Florida, 19–22 February (2007)
Application of Explosives in Metal Forming
Darko Šunjić1 ✉ and Stipo Buljan2

( )
1
University of Mostar, 88000 Mostar, Bosnia and Herzegovina
darko.sunjic@sve-mo.ba
2
Federal Ministry of Energy, 88000 Mostar, Bosnia and Herzegovina
stipobuljan1@gmail.com
Abstract. The use of explosives in deformation process is one of the most inter‐
esting and still insufficiently explored areas. This unconventional technology has
an increasing use because of being much cheaper and not requiring expensive
tools, machines. Explosive as a source of energy has a low price as well. This
paper involves deep-drawing explosion technology as well as one of the tools
which is used for this technology. High-alloy steels are used to make dies strength,
concrete, plastics and others. Also types of media that can be used in processes
of explosion are mentioned. Because the calculated mass of required explosives
can differ up to 200%, three distinct expressions are suggested, as well as expres‐
sion for calculating of detonation rate by Dautriche method.
Keywords: Metal forming · Explosive · Unconventional technology
1 Introduction
The need for workpieces of larger and more complex dimensions led to research into new
ways of obtaining the same or to unconventional forming methods. With the conventional
approach it was almost impossible to get such workpieces, and the technologies did not
have much “space” for further development. Charles E. Monroe’s research in 1888 marks
the beginning of the use of explosives for the processing of materials and only in the
mid-20th century there is a great deal of research and progress in this technology [1]. These
researches appear as the need for the space industry for new ways of processing large-
dimensioned materials and complex geometries. By the end of the 1960s this unconven‐
tional technology was used to produce submarine, floor, tank and other parts [2].
Today, explosion forming is used for deep drawing, welding, hardening, cutting and
expansion, and has been increasingly applied in all branches of industry in recent times.
Except in technical branches, explosive forming is also used in art [3, 4]. In this tech‐
nology explosive is used as an energy source. Its advantages are that, unlike other energy
sources, it is relatively cheap and can easily be obtained [5]. Explosion is an extremely fast
chemical reaction followed by a rapid release of energy capable of performing a certain
work. Explosions can be caused by physical, chemical or nuclear processes. Because of the
rapid release of energy, the process itself is very short. The advantage of this process is the
ability to produce every amount of energy that we need for the forming process.

https://doi.org/10.1007/978-3-319-90893-9_17
Application of Explosives in Metal Forming 145
2 Tools and Media in Explosive Forming
Materials used to make dies in explosion processing can be highest quality steels, cast
iron, concrete or plastics. Several factors influence the choice of materials, and some of
them are the type of processing material as well as the number of work items [1]. When
designing a tool, make sure that the tool is on an elasticated base, which is usually a
rubber hose filled with water used for shock absorption. Figure 1 shows a deep drawing
tool with an explosion. When designing this tool, it has been taken into account that the
tool itself must achieve the underpressure but also that it must fulfill all the elements of
extreme loads. Unless a certain underpressure is obtained, air compression can occur in
one spot and this workpiece would not have been shaped the way it was intended [6].
Fig. 1. Tool for deep drawing with explosion [6]
Figure 2 shows an example of deep drawing with explosion. The workpiece on which
the workpiece clamps are mounted is placed on the die. Above the workpiece, an explo‐
sive is placed at a certain height. This workpiece on a die with a workpiece clamps is
immersed in a water tank where an explosion is being initiated. In this kind of sheet
forming it is necessary to ensure that the impact on the die is as small as possible, and
Fig. 2. Example of deep drawing with explosion

146 D. Šunjić and S. Buljan
this can be ensured if the rubber is placed underneath the die. In this example, water was
used as a medium. After bringing into water as a media after the air, the mass of the
explosive decreased to 80%. Table 1 shows the types of media used in this technology.
Table 1. Types of media used in this technology

Type of media used in explosive forming
Liquid Solid Gaseous
Water Sand Air
Oil Rubber
Metal balls
Salt
3 Calculation of Explosive Mass
The calculation of the required mass of explosives is very important for explosive
processing because if more explosive is used, workpiece and tools can be destroyed and if
a small amount of explosive is used, the work will not be completely formed. Determining
the mass of an explosive can be calculated in several ways.
(𝛼+1)
⎧ ⎡ ⎤⎫
⎪ ⎢ ⎥⎪
𝛱D2 𝛿B ⎪ ⎢ 1 ⎪
G = (1−cos 𝜑)(1+𝛼) ⎨2ln 1, 5 ( ) ⎥⎬ (1)
e ⎪ ⎢ ⎢ 4f 2 ⎥
2 1 + 2 ⎥⎪
⎪ ⎣ D ⎦⎪
⎩ ⎭
where D – diameter of die aperture (mm), δ – sheet thickness (mm), β and α – coefficient
of approximation of real stress curves: α = 0,24 for Al alloys and β = 327, f – die depth
(mm), φ – the angle between the die and the explosive charge. The distance R between the
explosive charge and workpiece of diameter D is taken in the limits [7]: R = (0,2–0,4)D for
steels and alloys with high mechanical properties; R = (0,3–0,5)D for steels with low
mechanical properties; r0–√the explosive charge range depends on the size of filling G and
3
is determined by r0 = 0,53 G. The height of the water column H over the explosive charge
can be taken within the limits H = (20–25)r0. The mass of explosive for copper and steel
can be determined by the expression:
G = 0, 0113s0,972 𝛽 4,134 R0,582 (2)
where s blank sheet thickness, β the coefficient of drawing for the blank material and the
distance R between the explosive charge and the workpiece. The mass of the explosive can
be determined by an expression that depends on the tensile strength of the sheet σm [8]:
( )4,134
G = 2, 69 ⋅ 10−5 𝜎m s0,233 𝛽R0,14 (3)
The total mass of explosives can also be obtained through the following equation:
Application of Explosives in Metal Forming 147
[ ]0,8
𝜔𝛿R1,8
G= N (4)
1 − v2
where ω represents the specific deformation energy (Ncm/cm2), δ – sheet thickness

(mm), R – explosive distance form workpiece (cm), N – coefficient depending on the
material density, (1−v) – coefficient depending on the type of material.
In addition to the required mass of explosives, velocity of detonation can be calcu‐
lated. It is calculated according to the following formula (Dautriche method):
v⋅L
VE = (6)
2a
where VE - is the velocity detonation of the explosive being tested in m/s; L - the length
of the explosive charge in mm, a - the distance of the detonation wave collision space
in mm; v–velocity of detonation of blasting caps in m/s. The test is performed three times
and as a result the average of all three tests is taken.
The picture shows an explosive pipe that is placed in the ground or sand in an already
prepared groove. The plates should be as far away from the explosive, and the cord must
be laid in the form of mild arches to avoid sharp corners. The initiation of the primer is
performed electrically or by a slow burning cap. When the detonation wave reaches a
longer cord it will pass on it. On its further path, passing through the pipe in which the
explosive detonating wave is located, it will come to a shorter cord that will also detonate.
There is already a detonation wave in the longer cord that goes to the other end of the
cord, as well as a shorter cord, and their crash on the lead plate will give a notch [9].
Fig. 3. Determination of velocity of detonation
4 Conclusion
Is not always possible to create workpieces of the required surface or dimension quality
with traditional or conventional processing technology. Explosive forming as an uncon‐
ventional metal forming technology was created precisely from this need for products
of larger dimensions which, due to the limited tools, could not be obtained by
148 D. Šunjić and S. Buljan
conventional methods. This technology has been present in the market for a long time
and has a wider and wider application in the metal forming industry, metal welding using
explosives, metal cutting with energy generated by detonation etc. Its application is
“justified” due to the very cost of a process that is much cheaper because this technology
does not need expensive tools and machines, and an explosive as a source of energy is
relatively inexpensive. In addition, this technology is applied daily to a growing number
of materials, so it is used for shaping: aluminum, steel, titanium, titanium alloys, nickel
and others. Explosive forming is used in aircraft, automobile, missile industry and art.
References
1. Bohanek, V., Dobrilović, M., Škrlec, V.: Primjena energije eksploziva pri obradi metala.
Rudarsko-geološko-naftni zbornik 26, 29–37 (2013)
2. Brdarević, S., Jeleč, A.: Nekonvencionalni postupci obrade, Zenica (2016)
3. Tong, Z., Li, Z., Cheng, B., Zhang, R.: Precision control of explosive forming for metallic
decorating sphere. J. Mater. Process. Technol. 203(1–3), 449–453 (2008)
4. Fengman, H., Zheng, T., Ning, W., Zhiyong, H.: Explosive forming of thin-wall semi-spherical
parts. Mater. Lett. 45(2), 133–137 (2000)
5. Karabegović, E., Brezočnik, M., Mahmić, M.: Nove tehnologije u proizvodnim procesima,
Mostar (2014)
6. Buljan, S.: Primjena genetskih i stohastičkih metoda u istraživanju procesa dubokog vučenja
eksplozijom. Fakultet strojarstva i računarstva, Mostar (2007)
7. Pavelić, V.: Specifične tehnologije u proizvodnji oružja. Ministarstvo obrane Republike
Hrvatske, Novi Travnik Zagreb (1995)
8. Đukić, H., Nožić, M.: Obrada deformisanjem. Mašinski Fakultet Mostar, Mostar (2013)
9. Tete, A.D., Deshmukh, A., Yerpude, R.R.: Design and implementation of sensor for velocity
of detonation measurement. IOSR-JEEE 9(3), 77–83 (2014). http://www.iosrjournals.org/iosr-
jeee/Papers/Vol9-issue3/Version-1/K09317783.pdf
Application of Iterative Methods to Solve
Inverse Kinematics Problem of Robot
Avdo Voloder(&)
University of Sarajevo, 71000 Sarajevo, Bosnia and Herzegovina

voloder@mef.unsa.ba
Abstract. One of the most important segments of robotics is solving inverse

kinematics of the robot. The paper presents the application of numerical
methods iteration by individual coordinates, to solve the inverse problem of
kinematics of the robot. This problem involves the determination of unknown
internal coordinates of the robot, if they know the external coordinates. Using
this method, the solution converges quickly, as can be seen from the illustrative
examples in the paper.
Keywords: Inverse kinematics Iteration Convergence
1 Introduction
The kinematic model of robot means the knowledge of the connection between the
vectors of the internal coordinates fug and the external coordinate vectors fsg. A direct
problem of robot kinematics is the finding of a vector fsg based on a known vector fug.
The inverse problem of the kinematics of a robot implies the finding of a vector fug
based on a known vector fsg.
Inverse kinematic problem is in principle more complicated to solve than a direct
problem.
2 Solution of Inverse Problem of Kinematics of Robot
There are two ways to solve the inverse kinematic problem of a robot: analytical and
numerical. Analytical way produces exact expressions from which the internal coor-
dinates can be obtained directly from the known external coordinates.
The advantage of the analytical method is that the exact terms are obtained, which
can be predicted and the singular solutions are taken into account, there is no diver-
gence of the procedure. This is a very complex way and there is no analytical solution
for the arbitrary configuration of robots. Using this method, transcendental expressions
are often obtained which are difficult to solve if the robot has more than three degrees
of freedom of movement. Special solutions are obtained for each type of robot [1, 2].
In a numerical way, the solving system of equations, which are generally tran-
scendent, yields approximate solutions for the internal coordinates (qi), where in
general the accuracy of the solution depends on the type of numerical method, the

https://doi.org/10.1007/978-3-319-90893-9_18
150 A. Voloder
iteration step, and the assumed initial solution. The advantage of a numerical method is
that in this way one can make a universal algorithm for the robot arbitrary configu-
rations, with many degrees of freedom of movement.
3 Procedure for Numerical Solution of the Inverse

Kinematics Problem
Let are the functions
y1 ¼ y1 ðx1 ; x2 ; . . .; xn Þ;
y2 ¼ y2 ðx1 ; x2 ; . . .; xn Þ;
ð1Þ
...
yn ¼ yn ðx1 ; x2 ; . . .; xn Þ:
or
fy g ¼ ff ðxÞg: ð2Þ
Then the differentials of the individual functions will be
@y1 @y1 @y
dy1 ¼ dx1 þ dx2 þ . . . 1 dxn
@x1 @x2 @xn
@y @y2 @y
dy2 ¼ 2 dx1 þ dx2 þ . . . 2 dxn
@x1 @x2 @xn ; ð3Þ
...
@y @yn @y
dyn ¼ n dx1 þ dx2 þ . . . n dxn
@x1 @x2 @xn
The last system can be expressed in matrix form

8 9 2 38 9
> dy1 > @ y1 @ y1 @y
>
> >
> @ x @ x . . . @ x1n >>
>
dx1 >
>
>
>
> >
> 6 7> >
= 6 @ y2 @ y2 . . . @ y2 7 > >
1 2
< < =
dy2 ¼ 6 @ x1
6 .
@ x2 @ xn 7
7 dx2 ; ð4Þ
> .. > 6 . .. . 7> . >
>
>
> .
>
>
> 4 . . . . . .. 5> > .. >
> >
>
>
: >
; @ yn @ yn @ yn > : >
;
dyn @ x1 @ x2 . . . @ xn dxn
apropos
dfyg ¼ ½J dfxg: ð5Þ
where the matrix ½J represents the Jakobi’es matrix (Jakobian).

Application of Iterative Methods 151
For the vectors of internal and external coordinates, we can write
dfsg ¼ ½J dfug: ð6Þ
Different methods are used for the numerical solution of the inverse kinematical
problem of a robot: the iteration method by individual coordinates, the Chebyshev
method, the gradient method, etc. [3]. For this problem, it is a very convenient method
of iteration by individual coordinates, which is based on the determination of Jakobian,
which connects the external and internal coordinates of the robot, as well as its inverse
matrix.
The process converts quickly.
Let’s look at the function
fwðuÞg ¼ ff ðuÞg fsg; ð7Þ
where:
fug - the assumed, i.e. the iteratively obtained vector of the inner coordinates,
fsg ¼ ff ðuÞg - the exact external coordinate vector,
ff ðuÞg - an approximation of the external coordinate vector obtained on the basis of
iteratively obtained vectors of internal coordinates.
We can write
dfwðuÞg ¼ dff ðuÞg ¼ ½J df

ug: ð8Þ
If instead of differentiating the individual sizes, we apply their final differences,

then the difference between the first iterative and start values of individual sizes can be
written
fwðu1 Þg fwðu0 Þg ¼ ½J0 ðf

u1 g f
u0 gÞ: ð9Þ
In doing so, we adopt the start vector fu0 g, based on this ½J0 , and also the vector
ff ðu0 Þg.
Provided that: fwðu1 Þg¼ 0, next
1
fu1 g ¼ fu0 g ½J0 ðff ð
u0 Þg fsgÞ: ð10Þ
Analogously, the general iterative expression for a vector f

ui þ 1 g is obtained
1
fui þ 1 g ¼ fui g ½Ji ðff ð
ui Þg fsgÞ: ð11Þ
It is shown that by using this method, a solution is quickly reached.

152 A. Voloder
4 Example
For a robot in a Fig. 1 using the iteration method by individual coordinates, determine
the vector of the inner coordinates, where the external coordinate vector is given:
T
fsg ¼ px py ¼ ð 21 1; 9 ÞT . It’s default: L = 10. Values: px , py , h0 , L are in cm,
u0 is in radians. Adopt the starting vector of internal coordinates: f
ug 0 ¼ ð h0 u 0 Þ T ¼
T
ð 5 0; 2 Þ , where is h in (cm). Iteration is performed until the difference between
consecutive solutions for both internal coordinates is less than 0.001.
y1
y0
x 0 x1 A y2
A
1
h 2 L
P y3
x2
x3
Fig. 1. Robot by example
Solution
The transformation matrix between the individual robotic coordinate systems shown
are [3–5]:
2 3 2 3
cos u sin u 0 0 0 1 0 h
6 sin u 07 6 1 0 07
0 6 cos u 0 7 1 6 0 7
½D ¼ 6 7; ½D ¼ 6 7;
1 4 0 0 1 05 2 4 0 0 1 05
0 0 0 1 0 0 0 1
2 3
1 0 0 L
6 07
2 60 1 0 7
½D ¼ 6 7;
3 40 0 1 05
0 0 0 1
2 3
sin u cos u 0 h cos u þ L sin u
0 0 1 2 6 cos u sin u 0 h sin u L cos u 7
½D ¼ ½D ½D ½D ¼ 6
4
7
5
3 1 2 3 0 0 1 0
0 0 0 1
The external robot coordinate vector is

px h cos u þ L sin u h cos u þ 10 sin u
fs g ¼ ff ð uÞ g ¼ ¼ ¼ :
py h sin u L cos u h sin u 10 cos u
Jakobian is
2 @p @ px
3 " #
h sin u þ 10 cos u
x
@h @u cos u
4
½J ¼ @ p 5¼
@ py
y sin u h cos u þ 10 sin u
@h @u
The initial value of Jakobian is

" #
cos u0 h0 sin u0 þ 10 cos u0
½J0 ¼
sin u0 h0 cos u0 þ 10 sin u0

cosð0; 2Þ 5 sinð0; 2Þ þ 10 cosð0; 2Þ
¼ ;
sinð0; 2Þ 5 cosð0; 2Þ þ 10 sinð0; 2Þ

0; 980066 8; 807319
½J0 ¼ :
0; 198669 6; 887026
It’s inverse Jakobian

1 1 6; 887026 8; 807319 1; 377405 1; 761463
½J0 ¼ ¼
det½J0 0; 198669 0; 980066 0; 039734 0; 196013
It’s next

h0 cu0 þ 10 su0 5 cosð0; 2Þ þ 10 sinð0; 2Þ
ff ðu0 Þg ¼ ¼
h0 su0 10 cu0 5 sinð0; 2Þ 10 cosð0; 2Þ

6; 887026
¼ ;
8; 807319
so that the first solution is obtained for the internal coordinates

1
fug1 ¼ fug0 ½J0 ½ff ðu0 Þg fsg;

5 1; 377405 1; 761463 6; 887026 21; 0000
fug1 ¼
0; 2 0;
039734 0; 196013 8; 807319 1; 9000
5; 578733
¼
1; 738014
Analogously, the following values of the internal coordinates are calculated.

154 A. Voloder
Table 1. Values of the internal coordinates obtained in the illustrative example

Variant hcm urad
A 5.000000 2.000000E-01
5.578733 1.738014
18.137810 2.379403E-01
10.700180 6.880919E-01
19.174670 5.134518E-01
18.526510 5.833398E-01
18.563640 5.843485E-01
18.563670 5.843462E-01
B 5.000000 3.000000E-01
9.405127 1.421843
16.144140 3.069723E-01
17.206980 6.455181E-01
18.566280 5.806513E-01
18.563510 5.843494E-01
18.563670 5.843462E-01
C 6.000000 2.000000E-01
8.142085 1.481678
16.961470 1.616770E-01
16.023700 6.625049E-01
18.611720 5.739544E-01
18.562650 5.843482E-01
18.563670 5.843462E-01
18.563670 5.843462E-01
D 7.000000 4.000000E-01
14.978990 9.103186E-01
18.000900 5.486090E-01
18.537100 5.858104E-01
18.563670 5.843447E-01
18.563670 5.843462E-01
E 7.000000 5.000000E-01
17.055180 7.284958E-01
18.437380 5.782205E-01
18.562800 5.843983E-01
18.563670 5.843462E-01
On the basis of the program, the values of the internal coordinates for the individual
variants of the given starting values and for certain steps of iteration are obtained. The
values obtained are shown in Table 1.
Checking the accuracy of the obtained results can be done by inserting them into
the external coordinate vector:
( )
px h cos u + L sin u
fs g ¼ ff ð uÞ g ¼ ¼ ;
py h sin u L cos u

18; 563670 cosð0; 5843462Þ þ 10 sinð0; 5843462Þ 21
fs g ¼ ¼
18; 563670 cosð0; 5843462Þ 10 sinð0; 5843462Þ 1; 9
We get those values of the outer coordinates, which are set at the beginning.
5 Conclusion
The shown numerical method of iteration by individual internal coordinates is very

suitable for solving the inverse kinematic problem of a robot. The method gives good
results after a small number of iteration steps, as illustrated and obtained by the results
of the robot in the example.
References
1. Sciavicco, L., Siciliano, B.: Modeling and Control of Robot Manipulators, International
Editions edn. McGraw-Hill, New York (1996)
2. Lenarcic, J., Thomas, F. (eds.): Robot Kinematics, Theory and Applications. Kluwer
Academic Publisher, Dordrecht (2002)
3. Dolecek, V., Karabegovic, I.: Robotika, Univerzitetski udžbenik, Bihać (2002)
4. Christensen, H.I.: Robot Kinematics, Centre for Autonomous Systems Kungl Tekniska
Hogskolan (2006)
5. Lenarcic, J., Stanisic, M.: Advances in Robot Kinematics (2010). 12th ARK International
Symposium, pp. 475–482
Parameter Fitting for Soft Dielectric
Elastomer Actuator
Timi Karner ✉ , Janez Gotlih, Boštjan Razboršek, and Karl Gotlih

( )

{timi.karner,janez.gotlih}@um.si
Abstract. Big progress has been made recently in humanoid robotics with the
development of human like robots. Humanoid robots are becoming more and
more similar to humans, but what they still lack are proper mechanical actuators
which are still based on rotary motors or pneumatic cylinders. With the develop‐
ment of soft actuators one is getting closer to mimic the natural behavior of
muscles. Dielectric elastomer actuators can be used as soft actuators. They can
be described with fractional derivative Kelvin-Voigt model. With the help of
fractional derivatives and least-square method one can use this combination in
algorithm for parameter fitting. Fractional Kelvin-Voigt model can describe wider
working frequency range of dielectric elastomer actuator as can basic Kelvin-
Voigt model do.
Keywords: Soft actuators · Dielectric elastomers · Viscoelasticity · Kelvin-Voigt

Fractional derivatives
1 Introduction
With the developing area of soft actuators, new and innovative actuators are being made.
Soft actuators mimic the movement of biological muscles. Today’s humanoid robots
still uses rotational or translational motors which were invented by humans and are not
recognizable in nature. Up to date they exhibited good performance, good power weight
ratio and good efficiency. However, in order to mimic the movement of biological
muscle a soft actuator is needed. They do not exhibit nor rotational nor translational
movement, but they change their structure influenced by the stimulation signal which
can be electricity, chemical reaction or something else. First soft actuator also known
as artificial muscle was the so called McKibben artificial muscle [1]. It was made from
rubber and inflated with compressed air. When air was inflated it expanded in transversal
direction and contracted in longitudinal direction. On the other hand, when air pressure
was removed rubber retracted it back to original shape. These actuators are called pneu‐
matic artificial muscles or PAMs and are one of the first made soft actuators. Shape
memory alloys or SMAs can also be used as soft actuators [2]. These are materials that
are made by alloying copper-aluminum-nickel as main components. Material ‘remem‐
bers’ its original shape and it can be deformed up to certain point. In order to retract its
original shape, material needs to be heated up to a certain temperature. It can produce
relatively large stresses up to 200 MPa, small strains less than 8% and their retraction

https://doi.org/10.1007/978-3-319-90893-9_19
Parameter Fitting for Soft Dielectric Elastomer Actuator 157
speed is slow. They can also be arranged as resistors in form of SMAs. Electrical current
heats up the resistor SMA and retracts it to its original shape. Another field of soft
actuators are called ionic electro active polymers or IEAPs [2]. Those are polymers that
reacts on change of cations and anions surrounding it. With the applied voltage chemical
reaction begins and it change the state of cations and anions in the IEAPs. In order to
maintain the balance of ions the material begins to change its shape. Once the voltage
is removed the chemical reaction reverse and material regain its original shape. The
advantage if IEAPs is small voltage required to start the chemical reaction. Their disad‐
vantages are slow speed, small strain and slow retraction to original shape. Another
interesting soft actuators are called dielectric elastomer actuators or DEAs [1–3]. Their
main advantage is that they can produce up to 100% of strain, up to 3.2 MPa stresses,
high speed of response and relatively high speed of retraction to its original shape. Their
main disadvantages are large actuation voltage which can go up to 5000 VDC with
currents not exceeding 1 mA, they need to be pre-stretched and they exhibit viscoelastic
characteristics. In order to mimic the work-cycle of biological muscle, soft actuators
need to exhibit more than 20% of contractions, stresses need to be more than 0.35 MPa
and movement needs to be repeatable. Since DEAs are the most promising soft actuators
so far, they have been chosen to make an artificial muscle that will simulate the work
of a biological muscle. Their specific structure is actually similar to planar capacitor,
they have dielectric as isolator and two conductive electrodes on upper and lower part
of dielectric [1]. Since dielectric is made from elastomer material it changes shape when
voltage is applied. With the activated DC voltage there is a negative and positive voltage
on electrodes. This cause the electrostatic forces between the electrodes. Since dielectric
is made of elastomer it contracts in transversal direction and expand in longitudinal
direction. Elastomers are also known to be incompressible, their Poisson ratio equals
0.5. Generated stress is also known as Maxwell stress [2] and can be calculated as shown
in Eq. 1, where 𝜀0 is permittivity of vacuum, 𝜀r is relative permittivity, V is applied
voltage and d is thickness of elastomer.
𝜀 0 𝜀r V 2
𝜎Maxwell = (1)
d2
With maximizing the stress activated with the applied voltage the thickness of elas‐
tomer should be minimized. In order to minimize the thickness of the elastomer it should
be pre-stretched to certain point. Since it is incompressible it contracts in thickness when
it is stretched in longitudinal direction [4]. Figure 1 shows the principal operation of
DEA when there is no voltage and when voltage is applied.
158 T. Karner et al.
Fig. 1. Structure and actuation principal of DEA. Original state of DEA above, actuated state of
DEA below
In order to mimic the biological muscle soft actuator needs to perform not only single
actuation but also a periodic actuation. Whenever a periodic or a single cycle actuation
is performed with biological muscle, its final extension is usually never the same, except
if it is the same movement over longer period of time, like walking for example. In order
to be able to describe the behavior of dielectric elastomer the mathematical model needs
to contain the elastic part and viscous part since elastomer is a viscoelastic material.
Usually, DEAs have been described with the help of continuum mechanics using Helm‐
holtz free energy which was combined with the electrostatic force also known as
Maxwell stress [4]. Experimental results from tensile tests were approximated with the
Ogden, Mooney or Yeoh energy functions [5–8]. In order to be able to control the
movement of soft actuator made of DEA a different approach has been taken. For
displacement or force control of DEA an equation of motion is needed.
2 Methods
Since DEA are viscoelastic materials they exhibit elastic and viscous behavior. An
appropriate candidate for getting equation of motion for a viscoelastic material is
Kelvin-Voigt model [9]. It is constructed of two elements, a spring and a dashpot.
Spring represents the elasticity in material and dashpot represents a viscosity of the
material. Since DEA is going to be used in different regimes of working cycles, its
frequency and elongation will change. Basic Kelvin-Voigt model only have two
parameters for material fitting. Even more, it has been proven that different frequen‐
cies with the same amplitude generates different parameters of basic Kelvin-Voigt
model [10]. These can be arranged with the help of using Fractional Derivatives [11].
Fractional derivatives are derivatives where order of differentiation is real number and
is not limited to integer numbers. The dashpot in basic Kelvin-Voigt model can be
replaced by the fractional representation of dashpot, named springpot, since its char‐
acteristics change from being strictly an energy dissipater to a combination of an
energy dissipater and energy conserver [12, 13]. Its characteristics depends on the
order of fractional derivation. Difference of schematic representation of basic Kelvin-
Voigt model and fractional Kelvin-Voigt model shows Fig. 2.
Fig. 2. Difference between basic Kelvin-Voigt model left and fractional Kelvin-Voigt model
right
Basic advantage of using fractional Kelvin-Voigt model is that one is not limited to
only one spring and one dashpot. Later can be interpreted as a spring or as a dashpot or
as a bit of each one respectively. All depending on the order of differentiation. A math‐
ematical formulation of basic and fractional Kelvin-Voigt model is presented in Eqs. 2
and 3, respectively. Parameter k represents the stiffness of the spring, parameter c repre‐
sents the damping of the dashpot, x(t) represents the displacement of DEA, F(t) repre‐
sents the generated force of DEA and 𝛼 represents the order of differentiation.
F(t) = kx(t) + cx(t)

̇ (2)
d𝛼 x(t)
F(t) = kx(t) + c (3)
dt𝛼
Properties of fractional derivatives suggests that when 𝛼 = 1 than fractional Kelvin-
Voigt becomes basic Kelvin-Voigt model which means that order of differentiation is
equal to 1. If 𝛼 = 0 this means that one have two springs in parallel structure. What
should be mentioned here is that value of 𝛼 must be limited between 0 and 1 since values
bigger than 1 have no physical interpretations and values smaller than 0 represents inte‐
gration not differentiation anymore [13]. All values between 0 and 1 are possible and
represents a springpot, a model that contains elastic and viscous part.
2.1 Experiments
Elastomer 4910 VHB™ from 3M™ was chosen to represent dielectric part [14] of DEA.
It is an acrylic elastic polymer with adhesive layer on both sides with 1 mm of thickness.
Its initial dimensions were 49 × 33 × 1 mm. Elastomer was first tested on internally build
machine primarily used for tensile tests. Only low frequencies were taken into
consideration. As previously mentioned biological muscles contract and retract with

different frequencies and different amplitudes. In order to know material parameters for
different regimes, elastomer was tested with periodic oscillations where frequency and
amplitude was changing. The frequency and amplitude range of experiments represents
Table 1.
Table 1. Frequency and amplitude range of periodic oscillations of elastomer

Frequency/ A1 = 0.5 mm A3 = 1.5 mm A6 = 3 mm A9 = 4.5 mm A12 = 6 mm
Amplitude
F1 = 0.01288 Hz A1/F1 A3/F1 A6/F1 A9/F1 A12/F1
F2 = 0.02576 Hz A1/F2 A3/F2 A6/F2 A9/F2 A12/F2
F3 = 0.03864 Hz A1/F3 A3/F3 A6/F3 A9/F3 A12/F3
F4 = 0.05152 Hz A1/F4 A3/F4 A6/F4 A9/F4 A12/F4
F5 = 0.06440 Hz A1/F5 A3/F5 A6/F5 A9/F5 A12/F5
Elastomer was mounted into testing machine and it was excited with periodic oscil‐
lations. With measuring extension, a force response was measured too. With the help of
least-square method and experimental results values of k, c and 𝛼 can be calculated.
Periodic oscillations of displacement and force generated in elastomer can be described
as shown in Eqs. 4 and 5.
x(t) = xc cos(𝜔t) + xs sin(𝜔t) (4)
F(t) = Fc cos(𝜔t) + Fs sin(𝜔t) (5)
Parameters Fc , Fs , xc , xs can be calculated using least-square method from experi‐

mental results as shown in Eq. 6.
1
t2
[ { }]
J1 (xc , xs ) = xe (t) − xc cos(𝜔t) + xs sin(𝜔t) dt (6)
t2 − t1 ∫t1
Using fractional derivative on Eq. 4 and inserting results into the Eq. 3 one can arrive
to the following equations describing residuals for forces generated in elastomer, where
parameters r and s represents residuals for parameters Fc and Fs.
[ ]
r = Fc − kxc + c𝜔𝛼 cos(𝛼𝜋∕2)xc + c𝜔𝛼 sin(𝛼𝜋∕2)xs ≈ 0 (7)
[ ]
s = Fs − kxs + c𝜔𝛼 cos(𝛼𝜋∕2)xs − c𝜔𝛼 sin(𝛼𝜋∕2)xs ≈ 0 (8)
Using least-square method on Eqs. 7 and 8 leads to Eq. 9.
∑
n
( 2 )
J1 (k, c, 𝛼) = ri + s2i (9)
i=1
This system of equation is solvable if parameter 𝛼 is systematically increased for

every iteration. Method calculates residual for every values of 𝛼, k and c. At the end it
plots the value of 𝛼 versus the value of residual. The minimal value of residual represents
the parameters of elastomer material. Figure 3 shows the calculated residual versus 𝛼
parameter.
Fig. 3. Value of parameter 𝛼 vs. calculated residual
It can be seen that the smallest value of residual is calculated when 𝛼 parameter equals
to 0.47. Consequently at that value of 𝛼, value of k equals to 0.138645 N/mm and value
of c equals to 0.086904 Ns/mm. This means that springpot element is behaving nearly
as half spring and half dashpot. It has been mentioned before that equation of motion is
being investigated. Since all parameters were identified for fractional Kelvin-Voigt
model equation of motion can be stated and it is shown in Eq. 10 where m is mass of
elastomer.
d𝛼
F(t) = m̈x(t) + kx(t) + c x(t) (10)
dt𝛼
3 Conclusion
Soft actuators are becoming more and more sophisticated and will slowly mimic the
behavior of biological muscles. With the vast possibilities of soft actuators dielectric
elastomer actuator was chosen for experimental test. DEA offers the most advantages
among others as soft actuators. Elastomer VHB 4910 from 3M was chosen and it exhibit
viscoelastic behavior. Combination of fractional derivatives and Kelvin-Voigt model
shows that calculated parameters can be applied into fractional Kelvin-Voigt model.
This model contains third parameter which determines if the springpot element is more
similar to spring or dashpot. Value of a determines its specification. They evaluate all
possible combinations of frequency and amplitude of periodic oscillations presented in
Table 1. In future work this will be used for determination of control algorithm for DEA
in sense of controlling or displacement of DEA or force generated within.
References
1. Bar-Cohen, Y.: Electroactive Polymer (EAP) Actuators as Artificial Muscles, 2nd edn. SPIE
PRESS, Bellingham (2004)
2. O’Halloran, A., O’Malley, F., McHugh, P.: A review on dielectric elastomer actuators,
technology, applications, and challenges. J. Appl. Phys. 104(7), 071101 (2008)
3. Moscardo, M., et al.: On designing dielectric elastomer actuators. J. Appl. Phys. 104(9),
093503 (2008)
4. Suo, Z., Zhao, X., Greene, W.H.: A nonlinear field theory of deformable dielectrics. J. Mech.
Phys. Solids 56, 467–486 (2008)
5. Lu, T., et al.: Computational model of deformable lenses actuated by dielectric elastomers.
J. Appl. Phys. 114(10), 104104 (2013)
6. Mooney, M.: A theory of large elastic deformation. J. Appl. Phys. 11(9), 582–592 (1940)
7. Yeoh, O.H.: Characterization of elastic properties of carbon-black-filled rubber vulcanizates.
Rubber Chem. Technol. 63(5), 792–805 (1990)
8. Wissler, M., Mazza, E.: Electromechanical coupling in dielectric elastomer actuators. Sens.
Actuators A Phys. 138, 384–393 (2007)
9. Malkin, A.Y., Isayev, A.I.: Rheology. ChemTec Publishing, Toronto (2006)
10. Lewandowski, R., Chorążyczewski, B.: Identification of the parameters of the Kelvin-Voigt
and the Maxwell fractional models, used to modeling of viscoelastic dampers. Comput. Struct.
88, 1–17 (2010)
11. Podlubny, I.: Fractional Differential Equations. Mathematics in Science and Engineering, vol.
198, p. 311. Academic PRESS, San Diego (1999)
12. Mainardi, F.: Fractional Calculus and Waves in Linear Viscoelasticity: An Introduction to
Mathematical Models. World Scientific, Singapore (2010)
13. Mainardi, F., Spada, G.: Creep, relaxation and viscosity properties for basic fractional models
in rheology, United States, North America (2011)
14. 3M™: 3M™ VHB™ Tape 4910 (2017). Technical information about VHB 4910 tape. http://
www.3m.com/3M/en_US/company-us/all-3m-products/~/3M-VHB-Tape-4910?
N=5002385+3293242444&rt=rud. Accessed 24 July 2017
Timber Construction and Robots
Salah-Eldien Omer ✉
( )
SAG CONSULTING d.o.o., Vramceva 17, 10000 Zagreb, Croatia

salah-el-dien.omer@zg.t-com.hr
Abstract. The timber construction is using Robots from the phase of production
to the assembling of the prefabricated houses. The usage of Robots is based on
the need of the producer of prefabricated houses in the sense of software’s
programmed specially for certain production. The light timber construction are
produced and assembled with programs, mainly organized from the side of the
CNC centers and the Robots from the same machine producers. Unlike light frame
constructions the heavy wait timber constructions needs different programs for
production and additional Robots for assembling. Justification of usage of Roberts
should be based on a long term strategic plans of production and assembling when
they use digital designs.
Keywords: Timber construction · Software programming · Special CNC centers

Roberts for prefabricated houses
1 Introduction
Despite strong advancements in timber prefabrication using CNC systems, the timber
construction sector is still characterised by a relatively high proportion of (manual)
assembly tasks. Together with the inherently limited flexibility and working areas of
conventional CNC machinery, this handicaps the sector when trying to take advantage
of the rapidly spreading trend to use complex digital designs directly as input for
comprehensively automated construction processes. Here, robotic systems are
extremely useful—not only can their use lead to significant time savings, but their
ability to transfer digital design data directly to 1:1 assembly operations enables the
fully automated construction of non-standard timber structures. As a result, their use
opens up entirely new possibilities for future timber construction that is not limited
by the same constraints – such as, for example, work intensive joinery and/or addi‐
tional scaffolding – as manual assembly processes of pre-machined components; its
most evident and radical consequences are therefore the ability to digitally oversee
and control a large number of aspects of the design and construction (for instance the
sequencing of the single elements and their assembly) and the ability to freely posi‐
tion building components in space.
Robotic Timber Construction (RTC) in the beginning of year 2002 researches are
still in its infancy, and presents many theoretical, practical and methodological chal‐
lenges to architecture. Obvious examples are wide-ranging and include, for example,
the need for advanced computational design tools and novel constructive systems for

https://doi.org/10.1007/978-3-319-90893-9_20
164 S.-E. Omer
automated construction processes, and the integration of robust and adaptive robotic
fabrication technology.
While some heavy timbers are still hand crafted today, power tools and computer
numerical control (CNC) machines have assisted in the resurgence of heavy timber
construction. The use of CAD/CAM and CNC machines allows for the production of
heavy timbers to exact specifications, eliminating repetitive hand labour, increasing
production, decreasing labour costs, and making heavy timber more affordable. The
machine production of heavy timbers is initiated with a CAD/CAM design that directs
a CNC machine to mill, saw, mortise or drill heavy timber, per specification. Using this
method, one heavy timber member can be exactly reproduced an infinite number of
times. In some cases, a heavy timber member cannot be CNC machined, whether because
the joinery is too complex or the size of the tree is too large or oddly shaped, in which
case the heavy timber can be produced by hand or with power tools. Hand hewn timber
today is more commonly milled, sawn and planed with power tools than with axes and
saws, although some hand tool finishing is required.
Heavy timber structures have similar member types as light-frame structures, but
they use larger wood members – and fewer of them – to form the buildings structure.
Members that are used in vertical applications, providing support to members above,
are called posts or columns. Members that are used in horizontal applications, which
carry the live and dead loads from above to the posts below, include beams, girders and
joists. Wood floor planks are commonly installed over floor beams, similar to wood roof
planks which are installed over purlins. Purlins are connected to the top of the roof
structure, which is usually constructed with a number of heavy timber trusses or rafters.
The purlins carry the live and dead loads from the roof above to the trusses or rafters
below, where the loads are then transmitted to the outer bearing walls.
Unlike light-frame structures, heavy timber structures are not fastened together
with nails or screws. Two types of connections are used in heavy timber construc‐
tion: mortise and tenon joinery or modern metal or wood connections. Traditional
mortise and tenon joinery involves two pieces of timber that are cut and shaped in
such a way as to fit together. One piece is mortised with a slot or hole to accept the
other piece, which has a tenon or projection. Depending on the load requirements,
tenons can be 1 1/2 in. to 3 in. thick, and 4 in. to 10 in. long, After the two pieces are
joined they are pinned together with hardwood pegs. Several types of mortise and
tenon joinery are used, depending on the timber member application. They include:
double through, brace, dovetail, shouldered, or tongue and fork mortise and tenon.
Double through mortise and tenon is commonly used in cases with significant
tension; it has two slots on one member and two tenons on the other member. Brace
mortise and tenon is used where a diagonal brace is joined with a post or girder.
Dovetail mortise and tenon is used where a joist or purlin is joined with a truss, beam
or girder. Shouldered mortise and tenon is used to join a girder and a post, and tongue
and fork mortise and tenon joins the ends of rafters at the ridge. Modern metal joinery
includes heavy steel gusset plates with bolts, which are typically specified for struc‐
tural or aesthetic concerns. Additional modern connections include post anchors, post
and beam connections, hangers and anchors. Post anchors include metal shoes bolted
with bearing plates, metal straps that are bolted and/or embedded in concrete, and
Timber Construction and Robots 165
metal angles with bolts. Post and beam connections include metal caps with lag
screws, metal brackets with bolts and shear plates, metal anchor straps with bolts,
wood bearing blocks with split ring connectors, and precast reinforced concrete caps
with lag screws.
Heavy timber construction requires the use of a crane to lift timber members or pre-
assembled portions of the frame into place. The framer will make the joinery connec‐
tions, and when all of the pieces are in place, the connections can be tightened and the
frame is completed. The construction of an average size heavy timber structure frame
can typically be erected in two to three days. The exposed structural timber frame
requires immediate construction of exterior walls, along with the roof, to dry in the
building. When light wood framing material is used in the exterior wall assembly, heavy
timber construction is classified as Type V construction type. This building code clas‐
sification identifies construction types in relation to combustibility. Type V, Wood
Frame is the highest combustible construction type and includes single family residential
dwellings. When masonry or concrete material is used in the exterior wall assembly,
heavy timber construction is classified as Type IV - Heavy Timber. Type IV classifica‐
tion is also a combustible building type, but it requires the use of minimum sized timbers
which, due to their size, inherently provide protection from structural failure as the result
of a fire.
2 Digital Situation in Timber Processing
Compared to other sections in the construction industry, the wood industry is already
highly advanced in the field of digital fabrication. The unique feature of the Wood robot
and its driving software, is the ability to perform unique large-scale assembly tasks.
Furthermore, the robot is directly driven from BTLx input data, the industry standard
CAD interchange format for timber construction, framing and carpentry. The informa‐
tion of the CAD data is automatically translated into handling and processing steps for
the robot, rendering any additional work steps obsolete. In this respect, software plays
a crucial role in bridging the fabrication gap – an imperative to efficiently produce non-
standard elements that otherwise would need to be individually programmed.
Researchers at the University of German develop a lightweight timber construction
system combining robotic prefabrication with computational design and simulation
processes, as well as three-dimensional surveying technologies used in engineering
geodesy. In collaboration with Kuka, a manufacturer of industrial robot arms, and the
timber construction and engineering company Müller Blaustein, the partners investigate
the potentials of robotic prefabrication in timber construction. Their goal is to develop
innovative, performative and sustainable construction systems made from wood, which
also expand the repertoire of architectural expression in timber construction.
The investigation of the potentials of robotic fabrication requires the development
and application of novel architectural design, planning and simulation processes. A main
focus lies on the coherent “digital chain” from the geometry modelling, to the structural
analysis and digital fabrication, as well as the subsequent monitoring of tolerances and
geometrical deviations. The University of Stuttgart therefore follows an
166 S.-E. Omer
interdisciplinary and integrative approach, which also incorporates the material’s char‐
acteristics and processing possibilities. Utilizing the University’s own robotic fabrica‐
tion equipment, innovative and practical construction principles are continuously being
investigated and further developed into architectural systems.
Robotic Timber Construction (RTC) – where robotic fabrication is used to expand
additive digital fabrication techniques towards industrial full-scale dimensions. Featuring
robotic systems that grasp, manipulate, and finally position building components
according to a precise digital blueprint, RTC combines robotic assembly procedures and
advanced digital design of non-standard timber structures. The resulting architectural
morphologies allow for a convergence of aesthetic and functional concerns, enabling
structural optimisation through the locally differentiated aggregation of material.
Initiated by the group of Gramazio Kohler Research at ETH Zurich, this approach
offers a new perspective on automated timber construction, where the focus is shifted
from the processing of single parts towards the assembly of generic members in space.
As such, RTC promotes unique advantages over conventional approaches to timber
construction, such as, for example, CNC joinery and cutting: through the automated
placement of material exactly where it is needed, RTC combines additive and largely
waste-free construction with economic assembly procedures, it does not require addi‐
tional external building reference, and it offers digital control across the entire building
process, even when the design and assembly information are highly complex. This paper
considers (1) research parameters for the individual components of RTC (such as
computational design processes, construction methods and fabrication strategies), and
(2) the architectural implications of integrating these components into a systemic,
unifying process at the earliest stages of design. Overall, RTC leads to profound changes
in the design, performance and expressive language of architecture and thus fosters the
creation of architecture that profoundly reinvents its constructive repertoire.
In recent years, timber has increasingly become an important building material in
the construction industry. This is not only due to the fact that timber construction systems
are most suitable for reaching future goals for resource and energy efficiency. In this
context, timber construction is already characterized by a high degree of prefabrication.
It allows for higher precision, as well as fast construction and low noise emission on
site. However, many processes in prefabrication are still manual. While the production
of single elements is already digitalized and automated (subtractive fabrication
processes), the subsequent additive processes, when single parts are joined into building
groups, is mostly done by hand.
3 Robot in Production of Prefabricated Houses
Companies across the world are making substantial investments in robotics and auto‐
mation technology in order to increase efficiencies and make up for shortages in the
current workforce.
According to statistics presented during the last years, 85% of current robotics use
in the wood products industry is on the handling side, while 10% of mechanisms are
used for assembly and 5% for machinery. However, the use of robotics can provide
companies with solutions to issues that are currently roadblocks to growth, or are causing
headaches within their own labour force.
Robotics can be used to perform low-level tasks needed in the production process,
allowing companies to stop competing on wages with competitors and other industries.
Robotics is also allowing companies to provide easier systems for customization of
product orders by using new product lines that are completely automated. The produc‐
tion line in cabinet production, can produce eight cabinets per hour, 24 h a day, seven
days a week, with staff needed only for supervision and stocking supplies. Starting at
$1.2 million for the 5,100-square-foot production line, the producers’ representatives
believed that the numbers added up, making such a production facility an affordable
alternative to standard staffing for a similar workload and efficiency.
The Structural Timber design software’s are programs for design and dimensioning
of timber components, and timber roofs according to Eurocode 5 (EC5). In a graphic
environment you design roof trusses and elements of timber structures. The software
simplifies all the repetitive and time-consuming every day calculations for timber
elements and timber roofs. Detailed report with calculations, automatic generation of
truss drawings and details of connections.
To design a timber component or truss, choose from ready roof models, specify the
main dimensions, loads, design code parameters, and the design is immediately
performed and drawings produced automatically. Default values and checks for erro‐
neous input values, facilitate the input data process. The material properties, the loads
and the design code parameters, timber sections, can be adjusted by the user according
to the requirements of the National application document. A context-sensitive Help
system, guides you through the use of the program and the Eurocode provisions. On-
line user’s manual and frequently asked questions (F.A.Q.) are included in the program.
All the above-mentioned information is included in the software integrated in the Robot
Centre which was tested in several exact cases of timber construction of fabricated
houses (Figs. 1, 2 and 3).
Fig. 1. Robots in preparing elements for Prefabricated houses

168 S.-E. Omer
Fig. 2. Special designed robot to help in timber construction
Fig. 3. Special timber construction produced and assembled by Robots

4 Conclusions
After the digital programs with CNC machines followed with Robots in timber
processing industry, they find a wide usage in the production of wood prefabricated
houses and their assembling. Many software’s developed special programs for the
production of wood prefabricated houses as well as the assembling based on the
produced parts of the buildings. So, based on the represented data in this paper we can
conclude the following:
• Justification should be a STRATEGIC decision made by informed management that
balances the short-term goal of SURVIVAL with the longer-term goal of GROWTH
to posture a company for competitive advantage. It should not be relegated to an
accounting function without foresight of long-term strategy.”
• Robot usage improved manufacturing flexibility by reprogramming robots for
different projects, or having multiple robot programs for short production runs,
• They also reduced work-in-process inventory by combining operations in production
and assembly,
• They mostly offer more efficient production planning and scheduling due to predict‐
ability of automation in both cases production and assembling,
• Reduce operating costs generally.
References
1. Andres, J., Bock, T., Gebhart, F., et al.: First results of the development of the masonry robot
system ROCCO. In: Proceedings of the 11th ISARC (International Symposium on Automation
and Robotics in Construction), pp. 87–93. Elsevier, Oxford (1994)
2. Dalacker, M.: Schriftenreihe Planung, Technologie, Management und Automatisierung im
Bauwesen. In: Bock, T. (ed.) Entwurf und Erprobung eines mobilen Roboters zur
automatisierten Erstellung von Mauerwerk auf der Baustelle, vol. 1. Fraunhofer IRB Verlag,
Stuttgart (1997)
3. Nof, S.Y., Rajan, C.N.: Robotics. In: Handbook of Design, Manufacturing and Automation.
Wiley, London (2007)
4. Gramazio, F., Kohler, M.: Digital Materiality in Architecture, pp. 7–11. Lars Mueller
Publishers, Baden (2008)
5. Ficca, J.: Inclusion of performative surfaces material and fabrication research. In: Iwamoto, L.
(ed.) Digital Fabrications: Architectural and Material Techniques, Association for Computer
Aided Design in Architecture (ACADIA), New York (2009)
Conceptual Solution of the Robotic Arm/Plotter
Milena Djukanovic ✉ , Rade Grujicic, Luka Radunovic, and Vuk Boskovic

( )
University of Montenegro, 81000 Podgorica, Montenegro

{milenadj,radeg}@ac.me, luka.radunovic@mensa.me,
vukboskovic@outlook.com
Abstract. In everyday life - a person performs actions that require a lot of repe‐
tition, and therefore a lot of time. By creating a robot that would process and store
data in real-time, repeating the given actions a certain number of times, it would
eliminate the human presence in the process itself. In this regard, the aim of this
paper is the development of a conceptual robot solution that would replace a
human in performing some routine hand-executed functions, such as writing,
drawing, and engraving. Realization of the resulting conceptual solution would
have low cost of production and small dimensions, while easy portability would
allow a greater accessibility and application, both in industrial plants and in
personal use.
Keywords: Robot arm · Plotter · Conceptual solution · Easy portability

Routine hand-executed functions
1 Introduction
It is obvious that worldwide usage of robotics and its components is vigorously growing,
both in the industry and in the private use. People have always striven to do every day
routine tasks as quick as possible, so they could have time left to focus on more important
work. A solution they came up to is assistive technology and robotics.
Robot, as an electro-mechanical system which can autonomously perform a given
function or a task using the program, or controlled by human command, has a big variety
of usage, starting from small housework and finishing with complex tasks, such as
precise drawing, cutting, welding and assembling parts. The intelligence that one robot
possesses directly depends on the program itself and its adaptability to unpredictable
situations [1].
Major advantages of robotization over humans are greater economy, precision, speed
and infinite number of repetitions of a given task [1–4]. They differ, from autonomous
robots hands, used in industry, which have many different movements and functions, to
simpler mechatronic systems or devices that work in automation.
The main focus of this paper is on robotic components that perform repetitive tasks.
Thus, this paper is focused on the development of the design solution of robot arm/plotter
which would perform some routine hand-executed functions, such as writing, drawing,
engraving, but its primary purpose is signing of documents when the user is not directly
next to them.

https://doi.org/10.1007/978-3-319-90893-9_21
Conceptual Solution of the Robotic Arm/Plotter 171
2 Plotter Components
Plotter is a computer controlled machine designed to plot a given vector. Depending on

the mechanical extension of the plotter, it can perform various actions, such as writing,
drawing and engraving the given form.
Pen plotters use a pen, marker or similar type of drawing instrument as their mechan‐
ical extension. They perform their function by moving a pen on a flat surface of paper
or a desired substrate. The performance of pen plotters is based on vector fields, as
opposed to the usual dotted printers, which allows the plotting of complex lines with
high precision.
Plotter, as a mechatronic machine, is a set of mechanical elements and electronic
components. The most common mechanical components are: construction frame, slider,
lead screw, and lead screw nut. These components are controlled and aligned with
programmed electronic components, usually: programmable logic device(microcon‐
troller), sensors, stepper and servo motors.
The slider allows translatory movement of the plotter head positioned by the given
coordinates in relation to XY Cartesian coordinate system. The usual plotter contains
two sliders for each axis separately. Stepper and servo motors are sources of rotational
movement of the lead screw that, with proper nut, enable the translatory movement of
the plotter head on the slider. The sensors allow measurement and calibration of the
necessary data for performing the function. The microcontroller, with the role of the
control unit, coordinates the entire assembly. All these components are located and
adapted to the appropriate design and construction, which integrates the entire device.
The development of various plotters and realization of its management was a topical
topic of numerous researches, such as [5–7]. The breakthrough made in this paper,
compared to the previous solutions, is reflected in the specificity of the purpose of the
device itself and the changes it brings with itself, such as the formation of the appropriate
management base, the realization of management, communication between the machine
and the user, and so on.
3 Characteristics and Operating Principle of the Robotic Arm
Conceptual design solution for the engraving robotic arm, defined in this paper, is shown
in Fig. 1. The solution is developed in SolidWorks software [8].
This machine, which presents an upgraded real-time pen plotter, finds a wide purpose
in both business and private life. As an executive (output) unit, the plotter would allow
a real-time plotting of the vector inserted using a smartphone as an input device. As an
adequate example of its application, verification of documents by signature can be taken,
which would be easy accessible to any smartphone user, regardless of his/her distance
from the plotter. Namely, the signature entered on the input device would automatically
be scaled and positioned on the given document according to the user’s wishes.
The main advantage of this signing method is maintaining the authenticity and
validity of the signed document. The main reason for this is that, in this case, the motion
of the plotter’s head simulates the human hand motion in signing, allowing maximum
172 M. Djukanovic et al.
Fig. 1. Conceptual design of the plotter: 1-frame, 2-slider, 3-lead screw, 4-lead screw nut, 5-
stepper motor, 6-plotter head
precision and originality of the signature. Predicted dimensions of this machine are not
large (Table 1), which enables easy portability and compactness, allowing the plotter to
be placed on the office desk.
Table 1. Plotter dimensions

Length [mm] Width [mm] Height [mm]
Dimensions of plotter 600 600 500
Working area-platform 420 300 –
The complete executed process of the plotter operation would follow the flow
diagram defined in Fig. 2 through the following steps:
1. Setting a document or other substrate on the working platform where the vector
would be drawn;
2. Camera takes a photograph of the working surface;
3. The photo is forwarded to the phone via the application;
4. The user inputs his signature with a pen/finger by touching the smartphone screen;
5. Signature is processed by the program and ready for the user’s position selection on
the document itself;
6. Signature vector and all of the necessary coordinates are stored in the database on
the server and are forwarded to plotter;
7. Microcontroller gets, processes the data and coordinates the plotter head movements;
8. The document is signed and ready for taking.
Fig. 2. The process flow diagram
The focus in designing the machine is placed on lightness, compactness and porta‐
bility. Therefore, the materials planned for usage in the design are defined in Table 2.
Table 2. Materials and its use

Machine part Material
Construction frame Constructional aluminium
Plotter platform Aluminium
Sliders Stainless steel
Lead screw Steel
Lead screw nuts Bronze
Machine cover Plexiglas
Two stepper motors with the lead screws placed on XY directions are allowing
coordinated movement of the plotter head by the given coordinates. Driving of the
stepper motor is controlled by the output unit – microcontroller. According to a number

of practically developed robot designs [9, 10], the most commonly used stepper motor
for this purpose is NEMA 17 [11], controlled by Arduino Uno via motor drivers, so this
motor and microcontroller types are going to be used in the proposed design as well.
Servo motor, also controlled by Arduino, would allow a vertical movement of the pen.
3.1 Compatibility of Input and Output Devices

Smartphone, as an input unit, can vary with a variety of display resolutions. This feature
must be in accordance with format of the plotter platform. This problem is solved by
proportionally pixel scaling (Fig. 3), i.e., by the format ratio of the input device and the
plotter platform, according to the following expressions:
currx
X= x ,
scx max (1)
curry
Y= ymax , (2)
scy
Fig. 3. Scaling the pixels
Where X and Y are point coordinates on the plotter platform, currx and curry – pixel
coordinates on the smartphone, scx and scy – smartphone resolution, xmax and ymax –
dimensions of the platform.
3.2 Smartphone Application and Its Security
Access to signing and executing of the plotter’s task will be provided by a special appli‐
cation on the input device (smartphone). The task of this application is scaling and
forwarding all data to the database server, which are necessary for plotter to execute its
function. By designing the simple user interface, user would use the application with
ease, allowing him/her fast and efficient work.
The data that this application is forwarding to the database server will been crypted
with special algorithm in order to prevent signature falsification and to ensure data
security.
Relational database management system (RDBMS) would be used, as it allows easy
access to the users, and is performing fast enough for this type of application. This
database system uses a structured query language (SQL), which makes it simple for
programming and using. Administrator of the database may grant or restrict privileges
to the users that are using the base.
3.3 Calculation of Plotter Head Movement

A drawing of the lead screw with nut that is going to be used in this design is presented
in Fig. 4. Appropriate dimensions are given in Table 3, while Table 4 presents the char‐
acteristics of chosen stepper motor.
Fig. 4. Lead screw with nut: L – nut length, D – nut diameter, B – nut head thickness, d – lead
screw diameter, P – pitch of the lead screw
Table 3. Lead screw dimensions

Lead screw L [mm] B [mm] P [mm] D [mm] d [mm]
dimensions 16 4 4 16 8
Table 4. Stepper motor specifications [11]

Stepper Angle per step [°] Voltage [V] Current [mA] Max no. of Power [W]
motor rounds per
specifications minute. [rpm]
(NEMA 17) 1.8 12 400 200 5
In order to ensure precise movement of the plotter head, certain calculation of plotter
head movement with stepper motor rotation must be made. The lead screw with the lead
screw nuts transform rotational to the translatory movement, and they position plotter
head on given coordinates. Considering the threaded rod dimensions and stepper motor
specifications, data on the speed of the sliders and head movement itself can be gotten.
Connection between traveled distance of lead screw nut and angle that motor makes is
given with the following expression:
𝜑
X =P⋅ , (3)
2𝜋
Where: φ – turning angle of stepper motor, X – distance that slider or plotter head
covers in x-axis direction, P – pitch of the lead screw.
Differentiating the previous equation by time, we get an expression that connects the
translatory speed of the lead nut and angular speed of the stepper motor:
𝜔
v=P⋅ , (4)
2𝜋
Where: v – translational speed of the lead nut, ω – angular speed of stepper motor.
To provide the right stroke of the pencil, it is necessary to perform calibration of two
stepper motors. Correct use of these formulas, and their integration in the program which
coordinates and moves stepper motors, would result in high precision of vector drawing
on paper.
3.4 An Example of the Functioning of a Process
In this example, background process from the vector inputting to its plotting is shown.
Vector is inserted by a real-time data processing application, and it is automatically
forwarded to output unit that is plotting it on a given format.
Graphical user interface of the application displays an image of the working platform
in the moment when User2 puts his document on platform and permits the drawing. User
1 now can sign or draw vector which will be drawn on plotter.
While the user inputs his signature, vector is processed by application (Fig. 5), stored
in database and forwarded to the plotter.
Fig. 5. Vector drawn in application

Data is stored in database as the numerical coordinates of all dots in vector (Fig. 6).
In this example, given shape is made out of two vectors, red one and blue one. Their
coordinates are scaled based on expressions (1) and (2) and plotted on plotter. The graph
on which the points of the inserted vectors are dotted is shown in the Fig. 7. The figure
also shows the way in which plotter draws the vectors. By positioning the pen on first
coordinate of vector 1, plotter begins its writing. Pen will remain down on paper, i.e.
the vector will be plotted until it reaches the last coordinate, moving straight from point
to point. When the plotter finishes drawing of the first vector, pen is lifted from the paper
and will be positioned at the beginning of the next vector. This process is continuous all
until all vectors are drawn on paper/document.
Fig. 6. Saved data Fig. 7. Plotted dots
Servo motor attached to the head of the plotter, which moves pen and slider with its
rotation, is shown in Fig. 8.
Fig. 8. Plotter head with motor and pen

This assembly has two states, active and inactive one. The active state is when the
pen is drawing, and inactive when the pen is lifted from the paper. Also, user may
regulate the drawing pressure, allowing him to make the thinner or thicker lines,
depending of type of pen that is attached to the plotter. An example of finished work
and signed document ready for future use is shown in Fig. 9.
Fig. 9. Plotter platform after plotting
4 Conclusion
A conceptual solution for robotic arm/plotter that perform routine hand-executed func‐
tions is presented in this paper. This type of plotter would help speed up actions when
a person is away from the document to sign, whether it is on a business trip, stuck in
traffic, on vacation, etc.
Future elaboration would consider improving the application and its security. By
improving the application, data flow speed would increase from the input device to the
plotter, resulting in much more precise and fluent plotter pen movement. The final step
would be physical realization of the design, according to developed model, and selected
components, materials and adopted dimensions.
Acknowledgment. This paper is prepared within two projects - “Montenegrin Wearable Robots
(MWR)” supported by Ministry of Science of Montenegro and COST project - Action CA16116
“Wearable Robots for Augmentation, Assistance or Substitution of Human Motor Functions”.
References
1. Wisskirchen, G., Biacabe, B.T., Bormann, U., Muntz, A., Niehaus, G., Soler, G.J., von
Brauchitsch, B.: Artificial Intelligence and Robotics and Their Impact on the Workplace. IBA
Global Employment Institute (2017)
2. UK Essays.: Machines Vs Human Workers Business Essay (2013). https://

www.ukessays.com/essays/business/machines-vs-human-workers-business-essay.php
%3fcref%3d1
3. Soffar, H.: Advantages and disadvantages of using robots in our life (2016). https://
www.online-sciences.com/robotics/advantages-and-disadvantages-of-using-robots-in-our-
life/
4. Arkin, R.C., Moshkina, L.: Affect in Human-Robot Interaction. https://
www.researchgate.net/profile/Mohamed_Mourad_Lafifi/post/Should_we_research_the_
behaviour_of_robot_when_it_has_a_negative_emotion/attachment/59d6407479197b8077
99cab3/AS:431172468645892%401479810935588/download/Affect%2bin%2bHuman-
Robot%2bInteraction.pdf
5. Shivakumar, M., Stafford, M., Ankitha, T.H., Bhawana, C.K., Kavana, H., Kavya, R.: Robotic
2D plotter. Int. J. Eng. Innov. Technol. (IJEIT) 3(10), 300–303 (2014)
6. Karthik, S., Reddy, P.T., Marimuthu, K.P.: Development of low-cost plotter for educational
purposes using Arduino. In: IOP Conference Series: Materials Science and Engineering, vol.
225, no. 1 (2017)
7. Bhuiyan, M., BinHanafi, M.M.: Two degree of freedom plotter using robotic arm. Adv. Mater.
Res. 1115, 499–502 (2015)
8. Solidworks modeling software. http://www.solidworks.com/
9. Instructables. https://www.instructables.com/
10. Community of robots. http://communityofrobots.com/
11. Nema stepper motor specifications. http://www.mosaic-industries.com/embedded-systems/
microcontroller-projects/stepper-motors/specifications
Robot for Cleaning Ventilation Ducts
Milos Bubanja ✉ , Marina Mijanovic Markus, Milena Djukanovic,

( )
and Mihailo Vujovic

milosbubanja@hotmail.com, {marinami,milenadj}@ac.me,
vujovic_mihailo@yahoo.com
Abstract. Conceptual solution of robot for cleaning ventilation channel will be

shown and elaborated upon in this work. Our research focuses on the non-
destructive techniques to clean the dust, fungi, foreign particles and it also detects
flaws, cracks, and corrosion inside the ducts. This research involves a semi-
autonomous robot powered by rechargeable batteries, with four driven wheels for
movement inside different cross-sectional ducts. It comprises of several sensors
such as ultrasonic distance sensors for autonomous movement, quality of air
sensor and HD camera to provide the great visuals for inspection of various
parameters inside the ducts. Additional equipment like LED light and DC motor
for brushes are also envisaged. The main objectives are to clean the ducts by a
cleaning brushes attached to the robot, and ducts inspection by a HD camera,
helping to provide the healthy conditioned air supply to the chambers. The simple
design of robot makes it more convenient and useful for varying cross section
ducts and can helps to possible vendors build its bond at consumer ends.
Keywords: Robot · Cleaning · Inspecting · Duct · Ventilation system

Air conditioning system
1 Introduction
Air that flows through air conditioners – ventilation systems are, because of humidity
and temperature, is suitable substrate for developing fungus, mold, and bacteria. Expo‐
sure to fungi and mold, for example aspergillus, can lead to allergic reactions. These are
most commonly reported as inflammation of sinus, asthma or aspergillosis. The signs
of illnesses caused by hypersensitivity to fungal antigens cannot be told apart from those
caused by other allergens such as pollen, house dust or mold since these also cause
itching of the mucous membranes of the eyes and nose, sneezing, coughing, leaky nose
and difficult breathing. In many countries of the world there is a strict regulation which
makes it necessary to clean ventilation openings in catering facilities. Successful
cleaning of these ventilation openings is only possible with the help of a robot.
Cleaning of the ventilation channels and their inspection presents popular applica‐
tions of modern robotics. Offering HVAC (Heating, ventilation, and air conditioning)
system inspection and documentation services are a great business opportunity for air

https://doi.org/10.1007/978-3-319-90893-9_22
Robot for Cleaning Ventilation Ducts 181
duct cleaning contractors everywhere, especially as ASHRAE1 Standard 180 and latest
ANSI/ASHRAE/ACCA Standard 180-2012 [1] is accepted and adopted inspections has
to be a part of maintaining HVAC systems [2]. There are several commercial applica‐
tions of mobile robots for these services. Small robots are used only for duct inspection
and others for inspection and cleaning (like in [2]).
Commercial robots for duct inspection and cleaning are oft truck robots (like in [3]),
Fig. 1. There are also wheeled robots, Fig. 2(a), (like some DUNDUCT products [4])
and some special constructions as for example JettyRobot, a belt driven robot, Fig. 2(b),
[5], or whiled robot with adjustable main body [6].
(a) (b)
Fig. 1. (a) Vac-Systems Int.: The Super Trac Robotic Air Duct Cleaning System; (b) BioVac
System Inc.: Wolverine Robot used for inspecting, cleaning and sanitizing ventilation ducts.
(a) (b)
Fig. 2. (a) Danduct Clean small inspection robot and cleaning robot: (b) JettyRobot
In recent years there are plenty researches with robot duct cleaning. Some of them
are focused on compliance device [7]. But authors of this reading aimed to propose low
cost effective cleaning robot acceptable to potential producers and customers in the
region.
1
ASHRAE, founded in 1894, is a global society advancing human well-being through sustain‐
able technology for the built environment.
182 M. Bubanja et al.
Robot systems for duct inspection are as a rule connected to laptop computer and
supplied with camera. Operator can see what camera sees on the LCD monitor, record
the inspection for example with voice over and then store and edit data.
Robot systems for duct cleaning have additional equipment for different ways of
cleaning, and are usually supplied as customer wishes. They can be also autonomous [8]
if they are used only for cleaning, but semi-autonomy it is more desirable because of
possible operator’s action. Combination of autonomy with wireless operator control
allows to operator to act according to conditions on the ground, which is important by
using robot for duct inspection and cleaning at the same time.
Because of many different shapes of ventilation ducts, it is very hard to develop
universal robot that can clean all of them. There are many designed robots for cleaning
and inspection of ventilation system, but most of them have problem with overcoming
different this variety of geometrical shapes ventilation channels come in.
In this paper is presented design of robot that can easily be modified and adapted to
different geometrical shapes of ventilation channels. Main advantages of robot that is
presented in this paper is its capability to clean both rectangular and round ventilation
channels of different dimensions that range from 400 × 300 mm for rectangular ones
and with diameter more than 400 mm for round ones.
2 Design of the Robot
Prototype of the robot (Fig. 1) is developed by 3D modeling software Solid Works.

Maximum number of rotations of engine is 152 RPM. The robot is driven by all four
wheels which give him to have good maneuverability through ventilation ducts. The
wheels are TETRIX MAX all terrain tires with diameter 127 mm. Based on this infor‐
mation, maximum speed of robot can be calculated as follows:
2𝜋
𝜔= ⋅n (1)
60
v=𝜔 ⋅ r (2)
Where: 𝜔-angular velocity, n – number of rotations, v – speed and r – radius.
rad m
It is calculated from (1) and (2): 𝜔 = 15.91v = 1.01 .
s s
DC engine is controlled via HiTechnic DC motor controller for Tetrix, which allows
us to easily change movement speed of robot, which goes up to 1,01 m/s.
Brain of entire system is NI-MyRIO 1900 microcontroller, which has 10 AI pins, 6
AO pins, 40 DIO pins. Communication with DC driver for motors is done through I2C
network and for this microcontroller must have special adapter for I2C network. In our
case Tetrix + myRIO adapter board (rev2) was used (Fig. 3).
Fig. 3. 3D CAD model of the duct inspecting and cleaning robot without additional equipment
Due to need to monitor robot in real-time, as well as monitoring the state of venti‐
lation channels, robot uses Microsoft HD camera which is connected to microcontroller
through USB port on the controller.
In case of using robot only in the purpose of inspecting ventilation channel, his
control can be achieved through wireless network with microcontroller. In case of losing
connection with robot, it can autonomously return to starting position. Autonomous
movement through ventilation channels is accomplished with the use of ultrasonic
distance sensors.
Because of low visibility through ventilation channels, robot is equipped with LED
light whose power can be adjusted. Robot can be equipped with different extensions for
cleaning; analysis of this will be done later in the paper. In inspection purposes, its power
supply comes from 12 V battery with capacity of 3000 mAh.
In case of using robot just for inspection and checking the state of ventilation ducts,
following autonomous capabilities of robot were calculated (Table 1):
Table 1. Current, voltage and power values in the case of autonomous work
Motor Microcontroller Lighting
Imax (A) 1,37 1,16 1
Imin (A) 0,19 0,21 0
U (V) 12 12 0–12
Pmax(W) 16,44 14 1
Pmin (W) 2,28 2,6 12
I exp 0,8 0,9 1
Pexp 9,6 10,8 12
Iexp and Pexp are values that were calculated through experimental means. Also
minimal and maximal values of current and power are also presented.
Based on battery capacity and known power, based on following calculations values
of how long robot autonomy can last was calculated:
Q
Q=I∗t→t= (3)
I
Q 3
tmin = = = 23.52 [min] (4)
Imax 7.64
Q 3
tmax = = = 185.4 [min] (5)
Imin 0.97
Where: tmin – time of robot autonomy with maximal load, tmax – time of robot
autonomy with minimal load, texp – time of robot autonomy calculated experimentally.
From (3), (4) and (5) are calculated:
3
tmin = = 23.52 [min]
7.64
Q 3
tmax = = = 185.4 [min]
Imin 0.97
If we take minimal time autonomy of robot can last, which is 23.52 min, maximal
length of ventilation channel it can travel would be 1407.738 m in one direction.
Robot is also equipped with quality of air sensor MQ – 135. Data about air quality
is constantly being recorded and delivered to user for monitoring and storing.
3 Cleaning of Ventilation Channels
Robot is capable to clean rectangular, as well as round ventilation openings. This is done
via filled tires with large friction coefficient. Brushes can be mounted on robot, and are
able to move vertically using linear actuator with stroke length of 50 mm and power of
36 W. For additional equipment mounting such as brushes mechanical parts are added
to mobile robot platform (Fig. 4).
3.1 Additional Cleaning Equipment
In additional cleaning equipment there are three types of brushes. Different types of
brushes enable robot to clean different types of ventilation ducts and perform different
kinds of cleaning. Depending on state and shape of ventilation duct, it is required to
choose the correct brush for the job. Mounting of brushes on the robot is very simple.
Fig. 4. 3D CAD model of the duct cleaning robot with ability to mount additional equipment
3.1.1 Brush with Pneumatic Whips

This brush is mounted at the front part of the robot, on the shaft, which allows better
height adjustment and positioning of it.
This type of brush (Fig. 5) works on the principle of air flowing through tubes. This
causes tubes to chaotically move in all directions. By doing this, tubes hit inner walls
of ventilation duct removing dust and other polluting particles. Air flowing out of the
tubes pushes those particles forward.
Fig. 5. Example of brush with pneumatic whips

This process is continued throughout robots journey through ventilation system, at

the end all the pent up dirt is being expelled through opening.
This type of brush can have as few as 2–3 tubes or more depending on amount of
dirt in ventilation system. The more they are used, deeper the cleaning will be, but also
air losses exponentially increase, and with it the price.
3.1.2 Rotational Brush for Ventilation Channels with Round Cross Section
Rotational brush (Fig. 6) is also mounted on shaft on the front part of the robot. Shape
like that makes it more suitable to clean ventilation channels with round cross section.
Brush can have different diameters and be made from different materials depending on
the needs of the user.
Fig. 6. Example of rotational brush
If there is need for detailed cleaning and removing stuck particles, it is advised to
use the brush made out of harder plastic polymer which, even though it has greater
hardness, will not damage inner walls of ventilation channel. Brush is rotated using DC
motor which allows it to rotate at speeds ranging from 0 to 200 RPM.
This brush can also be used to clean ventilation channels with rectangular cross
section as well, as its threads are made out of plastic with good elastic properties. This
allows it to adapt to shape and size of channel without lowering its cleaning capabilities.
3.1.3 Ventilation Brush for Ventilation Channels with Rectangular Cross Section
During movement of robot through ventilation ducts, this type of brush (Fig. 7) rotates
at high speeds using DC motor. Direction of rotation of this brush is away from robot,
which causes it to remove dust and other polluting particles from walls of ventilation
channel and push it forward. This process is the slowest from all the above, but it is also
the most detailed.
Fig. 7. Example of brush for cleaning ventilation channels with rectangular cross section
Brush can also be positioned in horizontal position to be able to not only clean the
floor and roof of channel, but also the side walls.
4 Robot Control
Microcontroller that is used to control this entire system has been programmed in
LabView. This allows great flexibility and ease of coding.
This is done on this program also because microcontroller of robot is also made by
National Instruments, and thus it was made for it.
In the coming part of the paper, codes which control movement (Fig. 8), camera that
sends information to user (Fig. 9) as well as code that allows robot to be controlled by
joystick (Fig. 10) are shown.
Figure 8 shows how robots movement is coded in LabView. This part of the code
controls all four motors which drive the robot.
The code shown in the Fig. 9 allows camera to send captured information in real-
time. The code allows precise monitoring of robots position as well as inspecting the
state of ventilation system. Example of what the user sees is given on the left side of the
picture, while the code itself is presented on the right side.
In Fig. 10 is shown part of the code which allows robot to receive commands directly
from joystick that the user uses to control robot, and acts upon those commands.
Fig. 8. Code for movement
Fig. 9. Code for camera

Fig. 10. Code for joystick
5 Conclusion
In this paper robot for cleaning and inspection of ventilation systems is presented. The
robot is of great help in the air conditioning and ventilation systems as it provides
cleaning of the duct at the easiest and cheapest way. The wheeled movement of the robot,
combined with the autonomous and wireless control, helps the robot provide active
motion inside the ducts’ different profiles and variable cross section. The duct cleaning
and inspection robot provides an efficient non-destructive method for cleaning the ducts
using brushes. The HD camera provides the ability to inspect ducts and to see undesired
states of ducts such as flaws, cracks and leakages.
Regular cleaning can increase the efficiency and life of ducts and the cooling units.
Dust and dirt in ventilations system is a great health hazard and keeping them clean is
a requirement controlled by law in most countries. Since Montenegro is not part of these
countries, this is a major issue. But robot that was developed in this paper isn’t too much
expensive and is easy to made, so we hope in the near future it will be part of ducts
cleaning in our country.
Acknowledgment. This paper is prepared within two projects - “Montenegrin Wearable Robots
(MWR)” supported by Ministry of Science of Montenegro and COST project - Action CA16116
“Wearable Robots for Augmentation, Assistance or Substitution of Human Motor Functions”.
References
1. http://www.philcoaircontrol.com/wp-content/uploads/2016/07/ACCA-180.pdf
2. http://www.vacsysint.com/products/inspection-and-robotic
3. http://biovacsystem.com/en/manufacturing-our-products/robotic-air-duct-cleaning-
equipments
4. http://www.danduct.com/multi-purpose-robot
5. http://www.jettyrobot.com/
6. Singh, A.P., Malhotra, J., Gupta, N., Jaiswal, P., Kausar, A., Bhandwal, M.: Duct inspection
and cleaning robot. Int. J. Mech. Prod. Eng. 4(2), 45–48 (2016). ISSN 2320-2092
7. Jeong, W., Jeon, S., Park, D., Kwon, S.: Force control of a duct cleaning robot brush using a
compliance device. In: Proceedings of the 9th International Conference on Informatics in
Control, Automation and Robotics (ICINCO-2012), pp. 372–376 (2012). https://doi.org/
10.5220/0004119703720376, ISBN 978-989-8565-22-8
8. Irvandy, D., bin Mohd Zaid, A., bin Zakaria, M.F.: Autonomous Duct Cleaning Robot. http://
www.elektro.undip.ac.id/el_kpta/wp-content/uploads/2012/05/L2F004469_MTA.pdf
Cloud Robotics
Samir Vojić(&)
Technical Faculty Bihać, University of Bihać,

77000 Bihać, Bosnia and Herzegovina
samir.vojic@unbi.ba
Abstract. Cloud Robotics is a developing field of robotics, which currently

includes various applications and network paradigmas. The robots are limited in
terms of computing ability, memory and data storing. Cloud offers unlimited
computing ability, memory, data storing and especially a possibility for coop-
eration. In this paper, an overview of definitions, concepts and technologies
related to cloud robotics is given. Cloud robotics may have an important role in
the future asa a technology which is directed at man, capable of solving
society’s problems.
Keywords: Robotics Cloud Cloud robotics
1 Introduction
Cloud computing and cloud based services are on the rise because they offer a high
scalability and are very efficient because of capability to adjust to the current needs.
Due to the distribution around the world it is able to provide quick results (the task is
computed physically close to the data it is working on, to reduce network delay), has
low down-time, and most of all is financially reasonable (pay per use) [1].
Robotic systems have brought significant socioeconomic impacts to human lives
over the past few decades. For example, industrial robots (especially robot manipu-
lators) have been widely deployed in factories to do tedious, repetitive, or dangerous
tasks, such as assembly, painting, packaging, and welding.
These preprogrammed robots have been very successful in industrial applications
due to their high endurance, speed, and precision in structured factory environments.
To extend the functional range of these robots or to deploy them in unstructured
environments, robotic technologies are integrated with network technologies to foster
the emergence of networked robotics [2].
Networked robotics, similar to standalone robots, faces inherent physical con-
straints as all computations are conducted onboard the robots, which have limited
computing capabilities. Information access is also restricted to the collective storage of
the network. With the rapid advancement of wireless communications and recent
innovations in cloud computing technologies, some of these constraints can be over-
come through the concept of cloud robotics, leading to more intelligent, efficient and
yet cheaper robotic networks [2].

https://doi.org/10.1007/978-3-319-90893-9_23
192 S. Vojić
Cloud robotics is an emerging field of robotics rooted in cloud computing, cloud

storage, and other Internet technologies centered around the benefits of converged
infrastructure and shared services. It allows robots to benefit from the powerful com-
putational, storage, and communications resources of modern data centers. In addition,
it removes overheads for maintenance and updates, and reduces dependence on custom
middleware. Cloud robotics allows robots to take advantage of the rapid increase in
data transfer rates to offload tasks without hard real time requirements. This is of
particular interest for mobile robots, where on-board computation entails additional
power requirements which may reduce operating duration and constrain robot mobility
as well as increase costs.
2 Defining Cloud Robotics
The term “cloud-enabled robotics” was presented by James Kuffner for the first time at
the IEEE RAS Intl. Conference on Humanoid Robotics in 2010 [3]. He was first to point
out the potential of distributed networks combined with robotics, primarily to enhance
the robot agents limited capabilities. Cloud Robotics presented a need of fast, reliable
and ubiquitous network connections. The internet infrastructure has gone through a
radical change in the past 10 years in terms of bandwidth and Quality of Service, and is
now able to enable to support the cloud reliably. In 2011, at the Google I/O developer
Conference, Google and Willow Garage introduced their theory and foreseen appli-
cation of Cloud robotics.
There are several definitions of CR floating around, and various labs use them in
quite different senses. Ken Goldberg (UC Berkeley) emphasizes the benefits of the
unlimited computation capacity, memory and programming capability, leading to a
new form of collective robot intelligence through learning and sharing [4]: “Humans as
a species are getting smarter because we are able to share information much quicker
and build on innovations faster, robots have that potential as well” [4].
The same research group has described more specifically that “Cloud Robotics
proposes a thin-client model where robots are connected to modern cloud computing
infrastructure for access to distributed computing resources and datasets, and the
ability to share training and labeling data for robot learning” [5].
According to Steve Cousins (CEO of Willow Garage Inc.) the role of cloud robotics
is to provide an opportunity for open communication among robots, since “No robot is
an island” [6].
One of the most profound definitions was given by Mario Tremblay: “Cloud
robotics happens when we connect robots to the internet and then, by doing so, robots
become augmented with more capacity and intelligence. The cloud lets them com-
municate with other machines and serve their human operators better. Connected
robots equal augmented robots. By collaborating with other machines and humans,
robot robots transcends transcend their physical limitations and become more useful
and capable, since they can delegate parts of their tasks to more suitable parties” [7].
Cloud Robotics 193
3 Architecture of Cloud Robotics
According to Guoqiang Hu, Wee Peng Tay, and Yonggang Wen the architecture of
cloud robotics is organized into two complementary tiers: a machine-to machine
(M2M) level and a machine-to-cloud (M2C) level. On the M2M level, a group of
robots communicate via wireless links to form a collaborative computing fabric.
On the M2C level, the infrastructure cloud provides a pool of shared computation
and storage resources that can be allocated elastically for real-time demand [8].
In Fig. 1, we illustrate the system architecture for cloud robotics.
Fig. 1. Cloud robotics – two level architecture
Robots in a network can communicate if they are within communication range of

each other, and with the cloud servers if the robots are close to access points of the
cloud infrastructure. A wireless M2M communication network can be formed by robots
working cooperatively with each other to route and relay information. We call either a
robot or an access point a node in the M2M/M2C network. Several standards like
Zigbee, Bluetooth, and WiFi Direct have been developed for short range wireless
communications between robots. For long range communications, radio frequency and
microwave communication technologies may be used. A network of robots is often
formed dynamically and in an ad-hoc manner. There is no central controller to coor-
dinate the communication flow in the network. Robots may leave and join the network,
or may become unavailable because of unpredictable failures or obstructions in the
194 S. Vojić
environment. Furthermore, the network is highly dynamic if robots are mobile. All
these considerations make the design of effective routing protocols difficult and
impractical in some scenarios [2].
4 Benefits of Applications
According to Ben Kehoe, Sachin Patil, Pieter Abbeel, and Ken Goldberg the potential
benefits of Cloud include: Big Data (access to remote libraries of images, maps, tra-
jectories, and object data), Cloud Computing (access to parallel grid computing on
demand for statistical analysis, learning, and motion planning), Collective Robot
Learning (robots sharing trajectories, control policies, and outcomes) and Human
computation (using crowd sourcing access to remote human expertise for analyzing
images, classification, learning, and error recovery) [8].
In fact, cloud technology is proving to be the tipping point from the basic,
single-purpose robotics of years past—think assembly line robots, or the machines that
vacuum our floors and wash our dishes—to devices that can think, act, and work
alongside humans seamlessly. Big Data, artificial intelligence, machine learning, and
more are now being used to develop robots own neural networks, making use of
today’s large data sets to train machines on behavior. In this way, programmers are
now able to improve robots object recognition skills, their natural language processing,
their image classification and more, resulting in smarter machines, all by leveraging the
cloud.
In the future User Interfaces Will Disappear. The experience that most people have
with robotic technology today involves sitting at a desk or picking up a laptop and
using that interface to tell the machine what to do. That is vastly different from how we
all conduct ourselves in our day-to-day lives. We live on our smartphones, on our
portable and wearable devices, and we’re using them to talk to each other, text, and
more. We expect to interact with them in an almost entirely seamless way, and we
expect that from all technologies as well. That kind of human interaction, rather than
“program and walk away,” is the next big target for robotics engineers. Also, prices
will drop. Distributed software, including the Robot Operating System (ROS) powered
by Ubuntu Core, is having two very direct impacts on machine development: prices are
coming down and hardware is getting smaller. Developers are starting to take the
popular machine learning frameworks and put them into robotic contexts, leveraging
distributed software to handle vision, movement, object recognition, and more. But this
is just the learning and training stage. Eventually, having all of this information on the
“robot cloud” will mean that new machines will be able to tap into all the learning that
has been gathered by a larger and larger group of devices. As machines talk to
machines faster and faster, sending more and more data into the cloud, it’s becoming
difficult for humans to make sense of that information. Our brains just don’t work fast
enough. Clearly, the amount of data coming at us every day is not slowing down, but
we need a way to intelligently sort and classify it all. Robotic technology is offering a
solution to this problem, leveraging AI and analytics to break down the reams of data
that exist on the cloud and categorizing them based on the applications that they’re
managing. Over time, as a given company’s robots are trained and train others, they
Cloud Robotics 195
will get better and better as a group at identifying the most important information
available on the company’s cloud servers. As a result, the Big Data that is now
becoming so overwhelming will become more manageable.
5 Conclusion
Cloud robotics is an evolutionary jump for robots, a transformational change of

paradigm, where the communication with peers (robot to humans) has become aa
essential characteristic of a system. Cloud robotics allows robots to share computation
resources, information and data with each other, and to access new knowledge and
skills not learned by themselves. This opens a new paradigm in robotics that we believe
leads to exciting future developments. It allows the deployment of inexpensive robots
with low computation power and memory requirements by leveraging on the com-
munications network and the elastic computing resources offered by the cloud
infrastructure.
With the cloud robotics, it is possible to achieve the “remote-brain”, older concept
in which the hardware (sensors, actuators and low-level actions) and the software
(high-level reasoning, path planning, grasping, etc.) are separated [9].
References
1. Lorencik, D., Sincak, P.: Cloud robotics: current trends and possible use as a service. In:
SAMI 2013, 31 January–2 February 2013, pp. 85–89. IEEE (2013)
2. Hu, G., Tay, W.P., Wen, Y.: Cloud robotics: architecture, challenges and applications. IEEE
Netw. 21–28 (2012)
3. Kuffner, J.: Cloud-enabled robots. In: Proceedings of the IEEE International Conference on
Humanoid Robots, Nashville (2010)
4. Goldberg, K., Kehoe, B.: Cloud Robotics and Automation: A Survey of Related Work. UC
Berkeley Technical Report No. UCB/EECS-2013-5 (2013)
5. Kehoe, B., Matsukawa, A., Candido, S., Kuffner, J., Goldberg, K.: Cloud-based robot
grasping with the Google object recognitionengine. In: IEEE International Conference on
Robotics and Automation (ICRA), Karlsruhe, pp. 4248–4256 (2013)
6. Goldberg, K.: Understanding no robot is an island. In: Humans Invent
7. www.robotshop.com/blog/en/myrobots-com-pioneers-cloud-robotics-interview-with-mario-
tremblay-1348
8. Gowrimol, D.: A study on cloud robotics and automation. Int. J. Adv. Res. Comput.
Commun. Eng. 5(4), 25–26 (2016)
9. https://www.asme.org/engineering-topics/articles/robotics/6-ways-cloud-robotics-change-
world. Accessed 10 Jan 2018
Vibration Analysis of Motorcycle Handles
Zlata Jelačić ✉ and Boran Pikula

( )

{jelacic,pikula}@mef.unsa.ba
Abstract. The exposure of men, or parts of their bodies, to vibration has been
the subject of numerous studies for decades and has resulted in better under‐
standing of many parameters governing the human response to vibrations. In
particular, regarding the hand-arm system, vibration can cause changes in
tendons, muscles, bones and joints and can affect the nervous system to eventually
produce the so-called Hand-Arm Vibration Syndrome (HAVS). Quantification
of the parameters affecting the subjective response of different individuals has led
to define numerous standards as, for example, the EN ISO 5349-1 and EN ISO
5349-2 regarding the measurement and evaluation of human exposure to hand-
transmitted vibrations. Moreover, the European Directive 2002/44/EC on the
minimum health and safety requirements, regarding worker exposure to risks
from physical agents (e.g. vibration), limit the exposure to vibrations. The aim of
this study is to analyse the exposure level of motorcycle drivers to hand-arm
vibration (HAV). For this research, vibration levels of a common 650 cm3 maxi
scooter were experimentally measured and the maximum driving time that could
be safely used was established.
Keywords: Motorcycle vibration · Slam Stick X · Measurement
1 Introduction
The human exposure to vibrations is classified according to his impact into whole-body
vibration (WBV) and hand-arm vibration (HAV). These two types of vibration have
different sources, affecting different areas of the body and produce different symptoms.
HAVs are the most studied and they lie in the range of 6.3–1250 Hz.
The symptoms of exposure to HAV can be classified as vascular, neurological, or
musculoskeletal [1]. Vascular symptoms can be classified under the heading of “Vibra‐
tion White Finger”, a phenomenon characterized by the whitening of fingers when
exposed to low temperatures, while the most common neurological disorders are numb‐
ness and tingling in the fingers. The musculoskeletal damage associated with hand-arm
vibration manifests itself in the form of pain at the upper extremities, with loss of manual
sensitivity and dexterity [2]. More recent studies also point to the appearance of Carpian
tunnel syndrome [3]. According to Friden [4], the exposure to HAV can cause a variety
of vascular and neuromuscular symptoms, like tingling in digits, discomfort or inflam‐
mation in the wrist and hand, digital lightening, cold intolerance, feebleness of the finger
flexors or basic muscles and discoloration and trophic skin lesions of the fingers.

https://doi.org/10.1007/978-3-319-90893-9_24
Vibration Analysis of Motorcycle Handles 197
However, Gerhardsson and Hagberg [5] say that these symptoms depend on several
factors, namely on the intensity and duration of exposure to the vibration, on the type
of processes involved and on the tools used. Regarding the measurement of vibration
transmissibility to the human body, international standards have been developed over
the years, in order to guide and enable uniform experimental methodologies that can
obtain plausible results and encourage follow-up research in this area.
2 Materials and Methods
This research study applied the calculation method outlined in the International Standard
(ISO) 5349, for the case of HAV. The vibration exposure levels of a common 650 cm3
maxi scooter were evaluated. The data were collected from two tests, with an average
velocity of 25 km/h during drive in urban area.
2.1 Calculation of Vibration Exposure
European Directive 2002/44/EC defines the following terms related with HAV received
by workers. In this study, both concepts were adapted to drivers (moto):
• Hand-arm vibration: the mechanical vibration that, when transmitted to the hand-
arm, entails risks to the health and safety of workers, in particular vascular disorders,
osteoarticular injuries and neurological or muscular disorders.
• Whole-body vibration: the mechanical vibration that, when transmitted to the whole
body, entails risks to the health and safety of workers, in particular lower-back
morbidity and trauma of the spine.
• Daily exposure limit value: the vibration exposure value that cannot be exceeded
under any circumstances. For HAV this value is 5 m/s2 to an eight-hour reference
period.
European directive 2002/44/EC bases the evaluation of exposure levels to HAV on
the ISO Standard 5349. In order to assess the level of exposure, these standards
normalize the daily exposure value to an 8-h reference period A(8), defined as:
√
Td
Ai (8) = awi ; (i = 1, 2, 3) (1)
T0
for the HAV. In Eq. (1) the index i is used to indicate either the x, y or z axes, which are
defined according to the ISO Standard and awi is the weighted root means square accel‐
eration, rms, specified in meters per second squared in the direction of the three axes.
Also in Eq. (1), Td is the daily duration of exposure to the vibration awi and T0 is the
8-h reference period (28,800 s). The rms value of the acceleration over a direction i is
defined as:
198 Z. Jelačić and B. Pikula
[ ]1∕2
1 T 2
awi = ∫ a (t)dt ; i = 1, 2, 3 (2)
T 0 wi
Where T is the duration of the measurement and awi (t) is the weighted acceleration
specified in meters per second squared in the direction of the corresponding three axes.
After measuring the vibratory surface in contact with the body, the accelerations are
weighted in frequency domain and, for that, the frequency weighting curves recom‐
mended by the ISO Standard are used on the various directions. For the case of HAV,
the ISO 5349 standards specify that the total magnitude of the vibration should be deter‐
mined from vibration in the (x, y, z) orthogonal coordinates, as:
( )1∕2
av = a2w1 + a2w2 + a2w3 (3)
where av is the Total Magnitude of the Vibration (TMV) from the source producing it,
which is defined in m/s2, and the other quantities have already been defined previously.
Nevertheless, if a dominant axis of vibration exists, Eq. (3) can only be evaluated for
that dominant axis of vibration.
2.2 Measurement Equipment
The coordinate system used to acquire HAV data is defined in the Fig. 1. Our acceler‐
ometer channels were aligned with the motorcycle, in such a way that x is the media-
lateral direction, y is the anterior-posterior direction and z is the superior-inferior direc‐
tion (Fig. 2).
Fig. 1. Coordinate system used in the HAV (ISO 5349-1, 2001)
Fig. 2. Slam Stick X coordinate system
The Slam Stick X data loggers are capable of measuring acceleration in all three axes
while also measuring temperature and pressure. The recorders are available with
different measurement ranges (±16 g to ±2,000 g) and an industry leading high sample
rate (up to 3.2 kHz on the DC response MEMS accelerometer). A micro-USB receptacle
allows for quick and easy connection to a computer where data can be analysed with
Midé’s provided software package - Slam Stick Lab. The software also enables config‐
uration of the device to meet a variety of customer needs. Triggers include time delays,
calendar date/time wake up and acceleration, temperature and/or pressure triggers [6].
2.3 Experimental Setup

Vibration and mechanical shock data were collected during a total of two and half hours
of measurements and were obtained in accordance with the refereed International
Organization for Standardization (ISO 5349; ISO 2631) standards. For all situation two
tri-axial accelerometers were used. On the motorcycle, the accelerometers were placed
on the motorcycle handles as shown in the Fig. 3.
Fig. 3. Experimental setup
The coordinate system used to acquire HAV data is defined in the Fig. 2, the accel‐
erometer channels were aligned with the motorcycle in such a way that y is the anterior-
posterior direction, x is the media-lateral direction and z is the superior-inferior direction.
Vibration data were digitally recorded and stored.
3 Results and Discussion
The effect of vibration on health is said to be duration dependent and its assessment
should be made independently along each axis [7]. Figure 4 shows the motorcycle
trajectory and average speed. Figure 5 show the HAV acceleration for the case of
motorcycle in the three major directions. The analysis was made on the test track in the
urban area.
200 Z. Jelačić and B. Pikula
Fig. 4. Motorcycle trajectory and average speed
Fig. 5. Measured acceleration in x, y and z direction
From the acceleration data, the rms value of the acceleration over a direction i,
awi (t)(i = 1, 2, 3) and the A(8) parameters were calculated (Table 1). The daily duration
considered for the motorcycle was 2.5 h (9000 s).A large dispersion of vibration values
can be seen among axes. Nevertheless, as can be seen in the Fig. 5, the highest vibration
is produced in the longitudinal direction, i.e. along y axis (red), and the lowest is in the
vertical direction, z axis (blue). Moreover, the vibration in the fore/aft direction is almost
twice than is in the vertical direction.
Table 1. Calculated DEA values (daily exposure action)

Mean (StDev) awi(m/s2) Ai(8) (m/s2)
y axis 2,84 1,59
x axis 2,04 1,14
z axis 1,52 0,85
4 Conclusion
In this paper, an exposure analysis of hand-arm vibrations at the motorcycle handle is

presented for the case of paved road. With regard to the exposure of human hand-arm
system, in the outdoor situations, it is possible to conclude that drivers are exposed to
levels of effective acceleration that is higher than the action value recommended by the
European Directive 2002/44/EC. In the HAV the most affected joints are those in
shoulder, for earm and arm, since these joint resonance occurs at frequencies between
4 and 30 Hz, i.e. lower frequencies. The severity of HAV exposure depends essentially
on the type of terrain and of some adjustments of tyre pressures, suspension type and
frame material, among most important.
Directive 2002/44/EC quantifies the HAV exposure and sets limit values for param‐
eter A(8). The daily exposure action(DEA) is of 2.5 m/s2. If the measured values are
above the exposure action value, actions should be taken to reduce the vibration values
to below this value. Moreover, the measured values should not exceed the exposure limit
value and, if they are, immediate actions should be taken to prevent exposure above this
limit. In the case of average speed of 25 km/h on the paved road the set limit values
aren’t exceeded.
References
1. Lopez-Alonso, M., Pacheco-Torres, R., Martínez-Aires, M.D., Ordonez-García, J.:

Comparative analysis of exposure limit values of vibrating hand-held tools. Int. J. Ind. Ergon.
43, 218–224 (2013)
2. Rui, F., D’Agostin, F., Negro, C., Bovenzi, M.: A prospective cohort study of manipulative
dexterity in vibration-exposed workers. Int. Arch. Occup. Environ. Health 81, 545–551 (2008)
3. House, R., Krajnak, K., Manno, M., Lander, L.: Current perception threshold and the HAVS
Stockholm sensorineural scale. Occup. Med. Oxf. Engl. 59, 476–482 (2009)
4. Friden, J.: Vibration damage to the hand: clinical presentation, prognosis and length and
severity of vibration required. J. Hand Surg. Edinb. Scotl. 26, 471–474 (2001)
5. Gerhardsson, L., Hagberg, M.: Work ability in vibration-exposed workers. Occup. Med. Oxf.
Engl. 64, 629–634 (2014)
6. Mide Slam Stick X product documentation. https://www.mide.com/collections/shock-
vibration-data-loggers. Accessed 22 Jan 2018
7. Thrailkill, E.A., Lowndes, B.R., Hallbeck, M.S.: Vibration analysis of the sulky accessory for
a commercial walk-behind lawn mower to determine operator comfort and health. Ergonomics
56, 115–125 (2013)
Acoustic Diagnostics of Lever Mechanisms
with Subsequent Processing of Data
on Neural Networks
Sergiy Kovalevskyy(&), Olena Kovalevska, and Raul Turmanidze
Department of Machine Building Technology, Donbass State Engineering

Academy, Kramatorsk 84313, Ukraine
kovalevskii@i.ua, olenakovalevskaya@gmail.com,
inform@gtu.ge
Abstract. The technique of acoustic diagnostics for machine tools - robots is

developed. A neural network reference model has been constructed that allows
to diagnose the current characteristics of the state of objects under different
conditions, namely, the configuration of the mechanism, the geometric param-
eters of the mechanism with the motor-spindle running, the dynamics of the
movement of the nodes of the experimental stand mechanism with variable
speed and load on the drive, and the temperature of the object. Experiments have
been carried out to investigate the relationship between the parameters of the
spectrum of an acoustic signal with a given discreteness, excited by a perturbing
effect in the form of “white noise.” The possibility of using the proposed
approach to the management of complex technological machines, such as
machines with mechanisms based on parallel kinematics, is shown to improve
the accuracy of the positioning of the actuators, to ensure their dynamic tuning
and to optimize the trajectories of the movements of the working organs of the
equipment.
Keywords: Acoustic diagnostics Machine - robot Neural networks

Reference model
1 Introduction
It is known that increasing the accuracy and efficiency of technological processes of

mechanical processing depends on the quality of management of the executive
mechanisms of equipment. Along with increasing technical requirements for them,
requirements for equipment and process control systems, in particular - for the diag-
nostic channels’ informativeness, are increasing. In works [1–3 and others.] The
analysis of methods and means of diagnostics of technological equipment and pro-
cesses of mechanical processing is presented. However, despite significant progress in
creating new tools, methods and methods for diagnosing objects, the relevance of
search work in this direction is not reduced [4–6].
Among the actively developing methods of diagnosing objects and processes
should be distinguished methods of acoustic diagnostics and the creation of control
systems on this basis [7, 8].

https://doi.org/10.1007/978-3-319-90893-9_25
Acoustic Diagnostics of Lever Mechanisms with Subsequent Processing of Data 203
Previously, in the works [9], the assumption that as an informative source of a

diagnostic signal should be used its amplitude-frequency characteristic of the object’s
own oscillations in the acoustic range was expressed and partially confirmed.
The purpose of the work is to confirm the possibility of multi component analysis
of the parameters of objects (processes and equipment).
In order to create an informative spectrum of an acoustic signal, a hypothesis about
its informational content is adopted in the case of violation of an object by the influence
of “white noise”, which, along with the detection of information active frequencies of
the spectrum, allows forming in the process of diagnostics, single initial conditions.
The purpose of the research is to build a neural network reference model for
diagnosing the current characteristics of the object. The choice of diagnosed charac-
teristics of an object is determined by the tasks of object management:
– increase the accuracy of positioning of the executive mechanism of the object;
– providing dynamic adjustment of the executive mechanism;
– optimization of trajectories of displacements of the working body.
The solution of these tasks is aimed at improving the quality and productivity of
technological machines and processes.
In the works [8] the authors consider issues related to the diagnosis of technological
equipment. The main task of diagnostics is to recognize the state of the object in
conditions of limited information. The algorithm of recognition is partly based on
diagnostic models that establish the relationship between the states of objects and their
reflections of the space of diagnostic features.
The authors [8, 9] proposed to use acoustic control for such algorithms, based on
the amplitude - frequency characteristics of natural oscillations of the objects.
Among methods of acoustic control distinguish: active - use radiation and reception
of acoustic signals, and passive - based only on the reception of acoustic signals.
Acoustic control methods are based on interaction with a controlled product of
elastic oscillations and waves of a wide frequency range. For non-destructive testing of
multilayer structures, these methods are most widely used. The main ones are
low-frequency methods, ultrasonic method of passage and, to a lesser extent, reverb
and acoustic-topographical, although the possibilities are not fully disclosed. Obvi-
ously, the main problem lies in the methodology and is applied to the mathematical
apparatus for the processing of acoustic signals.
2 Testing Methodology
For research, a fast calculation algorithm for the discrete Fourier transform was used
with the help of an FFT analyzer, the digital audio signal input of which is input. The
analyzer selects from the signal successive intervals (“windows”) on which the spec-
trum is calculated, which is displayed in the form of the frequency dependence of the
amplitude of the frequency (Fig. 1).
Similarly to the band analyzer, the logarithmic scale is usually used in the axes of
frequency and amplitude. But due to the linear positioning of the FFT bands in
204 S. Kovalevskyy et al.
Fig. 1. FFT analyzer display
frequency, the spectrum may not be sufficiently detailed at low frequencies or exces-
sively oscillate at higher frequencies.
The horizontal line on the FFT analyzer displays a white noise that has equal
energy in equal linear frequency intervals.
Parameter N - number of analyzed signal responses - is crucial for the type of
spectrum. The larger N, the denser the frequency network, by which the FFT
decomposes the signal, and the more details on the frequency are visible on the
spectrum.
Longer signal areas are analyzed for higher frequency resolution. If the signal
within the FFT window changes its properties, the spectrum displays averaged signal
information from the entire window interval.
To analyze the rapid changes in the signal, the length of the window N is assigned a
small. Then the resolution of the analysis in time increases, and in frequency -
decreases. Thus, the resolution of the frequency analysis is inversely proportional to the
resolution of time, which can be explained by the uncertainty relation [7–9].
For the formalization of the mathematical description one can use the recommen-
dations of work [4], in which the spectrum of the signal is represented by the spectral
characteristic
Zþ 1
Xp ðjwÞ ¼ cp ðtÞejwt dt ð1Þ
1
The spectral density of a signal is determined by the formula:
1 h 2 i
Sv ðxÞ ¼ lim M Xp ðjxÞ ð2Þ
Tp !1 2 Tp
The estimation of the spectral density is made according to the known imple-
mentation of the xp(t) of the signal by generating from it a discrete sequence x(n),
n = 0, 1, …, N −1(n) and processing this sequence in accordance with the given
quantization. The Fourier transform of a valid sequence of finite length x(n), n = 0, 1,

…, N −1, the same
XN1
X ejxT ¼ vðnÞejxTn ð3Þ
n¼0
As a spectral density estimate, the function is used
1 jxT 2
Pv ðxÞ ¼ X e ð4Þ
N
In general, these estimates are not prosperous and there is the possibility of their
fluctuations around the true value of the spectrum. Filters and averaging methods are
used to obtain a well-appreciated spectrum.
Using frequency filters with the frequency characteristic S(k), we can obtain the
modified spectrum of the object X as a function of its properties Y
Ak ffk ½SðkÞg ! XðY Þ ð5Þ
Equipment for experimental research:

– an experimental stand in the form of a delta mechanism, which is a group of
mechanisms based on parallel kinematics, which is equipped with a system of
numerical software control based on the MEGA 2650 board, which provides pro-
gramming of stand control using G-codes;
– Two sound cards are installed, and two computers running the Windows XP
operating system;
– Piezoelectric Emitter and Sensor.
– an amplifier of the excitation signal “white noise”, power 2 Watts.
Software:
– Spectrolab computer package;
– computer package NeuroPro-0,25;
– Statistica 10 computer package.
The plan of experiments was supposed to study the relationship between the
parameters of the spectrum of the acoustic signal with a given discretion, disturbed by
the excited influence in the form of “white noise”. The authors put forward the
assumption about the possibility of diagnosing the characteristics of investigated
objects invariant with respect to the errors made by the hardware part of the diagnostic
system. This assumption is based on the fact that the diagnostic signal has the initial
values of the acoustic spectrum formed by the “white noise” by the same information
channels and the same hardware as the main excited diagnostic signal.
Each dimension has the maximum possible protection against accidental signal
fluctuations. This was achieved by the fact that, at each implementation of the elements
of the experiment plan, the measurements are averaged over a cycle of 100 successive
scans of the spectrum with a spectral discreteness of 178.3 from 0 to 20,000 Hz.
The full-factor experiment with randomization of combinations of factors allowed to

form a database, where the factors were the coordinates of positioning the executive
mechanism of the experimental stand ðXi ; Yi ; Zi Þ. An example of a fragment of such a
base for experimental positions is presented in Table 1.
Table 1. Amplitude of the acoustic spectrum by its discrete

№, Inputs: positioning Outputs: amplitude of acoustic spectrum,
coordinates, mm decibels, by its discrete, Hz
X, mm Y, mm Z, mm 172,3 344,6 516,9 … 21882,1 22054,4
1 −52 −30 5 −29,9 −29,76 −46,87 … −73,68 −81,32
2 52 −30 5 −32,3 −31,94 −46,98 … −75,34 −83,49
3 0 −60 5 −34,49 −33,72 −45,56 … −75,61 −83,99
4 0 0 5 −36,68 −36,11 −48,39 … −76,12 −84,93
… … … … … … … … … …
n−3 −52 −30 250 −54,24 −48,78 −49,53 … −75,75 −83,77
n−2 52 −30 250 −53,57 −48,33 −48,57 … −76,55 −84,44
n−1 0 −60 250 −55,13 −48,41 −48,71 … −76,27 −84,16
n 0 0 250 −54,31 −49,35 −50,21 … −75,86 −84,02
An example of the acoustic spectrum of the diagnostic signal for identifying the
positioning of the executive mechanism of the experimental stand is presented in
Fig. 2.
Fig. 2. Spectrum of acoustic signals for positioning coordinates of the executive mechanism of
the experimental stand
2.1 Results and Discussions

On the basis of the experimental basis, neural network diagnostic models were
constructed:
– configuration of the mechanism;
– geometric parameters of the mechanism when working motor-spindle;
– dynamics of movement of knots of the mechanism of the experimental stand with
variable speed and load on the drive;
– change in the temperature of the object.
The device of neural networks allows handling large volumes of input data rather
quickly and also to represent them in a form convenient for further researches.
Below is a fragment of the mathematical model created by the principles of
connection:
Database Fields (Output Symptoms):
S7
S8
S60
S68
S86
S89
S90
S93
S95
S106
S108
S111
S113
S115
S116
S120
S123
S126
S127
Database fields (end syndromes):
Z
End syndromes:
Z = 0,9342992 * Syndrome1_1 − 0,8471334 * Syndrome1_2 + 0,7881347 *
Syndrome1_3 − 0,9805002 * Синдром1_4 + 0,8393618 * Syndrome1_5 −
0,796478 * Syndrome1_6 + 0,9282416 * Syndrome1_7 + 0,9115893 * Syndrome
1_8 − 0,7316513 * Syndrome1_9 + 0,8834189 * Syndrome1_10 + 0,8092021 *
Syndrome 1_11 + 0,742706 * Syndrome1_12 + 0,7383921 * Syndrome1_13 +
Syndrome1_14 − 0,8908381 * Syndrome1_15 + 0,8708463 * Syndrome1_16 −

0,6654512 * Syndrome1_17 + 0,8319975 * Syndrome1_18 − 0,9760703
* Syndrome1_19 − Syndrome1_20 − 0,308390
…………………………………………..
…………………………………………..
Post treating end syndromes:
Z = ((Z * 245) + 255)/2)
The generated neural network models were tested for Fisher’s criterion adequacy.
The results of assessments of these models are presented in Table 2.
Table 2. Verification of mathematical models constructed according to experimental data on

adequacy
The object of Input coordinate m n Error mm Fтable Fcalculation
the experiment
Configuration X(−75,−50, 16 72 0,01 2,67 8,6 107
of the −25,0,25,50,75)
mechanism Y(−75,−50, 11 72 0,01 2,66 4,5 107
−25,0,25,50,75)
Z(−75,−50, 11 72 0,01 2,66 1,6 108
−25,0,25,50,75)
X(−75,−50, 16 222 0,01 2,60 9,6 103
−25,0,25,50,75)
Y(−75,−50, 17 222 0,01 2,60 3,1 104
−25,0,25,50,75)
Z(−75,−50, 19 222 0,01 2,60 2,1 104
−25,0,25,50,75)
Configuration X(−75,−50, 129 222 0,01 2,63 1,4 103
with working −25,0,25,50,75)
motor-spindle Y(−75,−50, 129 222 0,01 2,60 2,9 108
−25,0,25,50,75)
Z(−75,−50, 129 222 0,01 2,59 1,8 107
−25,0,25,50,75)
Trajectory of X0 ! X1 X0 14 216 0,01 2,60 5,9 102
moving with X1 14 216 0,01 2,60 6,5 102
variable speed Y0 ! Y1 Y0 14 216 0,01 2,60 6,7 102
and load on Y1 14 216 0,01 2,60 6,7 102
the drive
Z0 ! Z1 Z0 14 216 0,1 2,60 8,8 102
Z1 14 216 0,1 2,60 8,8 102
V 14 216 0,01 2,60 4,7 102
m 14 216 0,01 2,60 3,5 102
Changes in Sample1 1 11 0,1 C0 3,17 3,7 102
sample Sample 2 1 13 0,1 C0 3,05 6,9 102
temperature
The adequacy of the models is confirmed by the repeated excess of the calculated
value of Fisher’s criterion over its table value for the given error probability for the
input models - the frequency ranges of the acoustic spectrum, constructed on the line of
the topple of the initial data.
3 Conclusion
The obtained results of the study confirm the provisions set forth in [21], where the
authors state that the management of the executive motions of technological machines
with mechanisms of parallel structure is a complex task, the solution of which can be
obtained on the basis of solving kinematics problems and dynamics of executive
mechanisms. As a result, conditions can be created for the precise positioning of
working bodies of executive mechanisms at optimal speeds and accelerations of their
movements in a given trajectory. For these purposes, management systems should be
equipped with informative object identification systems. To confirm this, we propose
the results of experimental studies that provide the basis for the establishment of
systems for the diagnosis of kinematics and dynamics of the mechanism, on the basis
of which the authors create mobile machines for use in reconfigurable production
systems of mechanical processing.
The method of acoustic diagnostics for mechanisms of technological machines
(machines - robots) is developed in the work, which allows diagnosing their various
states in different conditions. The possibility of using the proposed approach to the
management of complex technological machines, such as machines with mechanisms
based on parallel kinematics, is shown to increase the accuracy of the positioning of
actuators, to ensure their dynamic adjustment and optimization of the trajectories of
displacements of work equipment bodies (cutting tools). The developed method of
acoustic diagnostics for mechanisms of machine tools - robots allows to expand the
range of possibilities to increase the accuracy and productivity of their work.
References
1. Kovalevskyy, S., Starodubcev, I.: Neural network analysis of natural frequencies spectrum of
products for the diagnosis of their parameters. In: Research and Development in Mechanical
Industry, RaDMI, pp. 230–234 (2012)
2. Kovalevskyy, S., Kovalevska, O., Kostina, A.: Perfection of non-destructive methods of
control of surfaces of details of machines. In: Research and Development in Mechanical
Industry, RaDMI 2011, pp. 230–235 (2011)
3. Kovalevskyy, S.V., Kovalevska, O.S.: Acoustic monitoring with neural network diagnostics.
Am. J. Neural Netw. Appl. 2, 39–42 (2015)
4. Kovalevskyy, S.V., Dašić, P., Starodubcev, I.N., Kovalevska, O.S.: Size control products
using spectrum of natural vibrations and neural networks machining. Bull. Donbas State
Mach. Build. Acad. 2(27), 38–43 (2012)
5. Kovalevsky, S.V., Kovalevsky, O.S., Korzhov, E.O.: Diagnostics of Technological Systems
and Products of Mechanical Engineering (Using the Neural Network Approach): Monograph.
DSEA, Kramatorsk (2016)
6. Kovalevskyy, S.V., Kovalevska, O.S., Nagiyeva, A.I.: Acoustic control of dimensions and
indicators of traceability of parts surfaces using neural network models. In: Neuronite
Technologies and Their Applications: Collection of Works of International Scientific
Conference, pp. 24–35. DSEA, Kramatorsk (2014)
7. Kovalevskyy, S.V., Kovalevska, O.S., Tulupov, V.I.: Development of Methods of Acoustic
Diagnostics in Mechanical Engineering: Monograph. DSEA, Kramatorsk (2014)
8. Kovalevsky, S.V., Andronov, A.Y., Kovalevska, O.S.: Development and Research of Special
Methods of Machining of Machine Parts: Monograph. DSEA, Kramatorsk (2015)
9. Kovalevskyy, S.V., Kovalevska, O.S.: Acoustic diagnostics of static and dynamic properties
of mechanisms with parallel kinematics. Complex Qual. Assur. Technol. Process. Syst.
(KJYPSPS) 2(2), 59–60 (2017)
Power and Control System of Knee
and Ankle Powered Above-Knee Prosthesis
Miljan Rupar1 ✉ , Zlata Jelačić2, Remzo Dedić1, and Adisa Vučina1

( )
1
University of Mostar, 36000 Mostar, Bosnia and Herzegovina
miljan_deus@yahoo.com, {remzo.dedic,adisa}@sve-mo.ba
2
jelacic@mef.unsa.ba
Abstract. After above-knee amputation, amputees usually go under intense

rehabilitation process in order to overcome their disadvantage and successfully
reintegrate in the society. Passive prostheses, which are mostly used by amputees,
enable performing of various activities such as walking on levelled and inclined
ground, and even running, riding a bicycle and as of lately swimming. However,
performing high power demanding tasks, such as stair ascent, presents a problem
because the lack of muscles makes it impossible to produce required forces. This
means, that in order for the prosthesis to be able to perform high demanding power
activities it must be powered, primarily in its main joints – knee and ankle. In this
paper, we are presenting hydraulic power system and control system for knee and
ankle powered prosthesis which we are developing in order to achieve required
kinematics and dynamics of the prosthesis which would enable it to perform high
power demanding activities in more natural manner, especially stair ascent.
Keywords: Powered above-knee prosthesis · Hydraulic power system

Control system
1 Introduction
Despite technology advancements of modern age, ascending stairs of above knee ampu‐
tees in natural manner is still challenging to accomplish. Reason for that are large forces
and moments that are required to accomplish stair ascent. After above-knee amputation,
amputees usually undergo intense rehabilitation process in order to overcome their disad‐
vantage and successfully reintegrate in the society. Proper reintegration in the society is
very important because amputation has significant influence on life of amputees, and
their interaction with environment. Today, most amputees use passive prostheses. Passive
means that these devices are not powered in any way, and cannot give any additional
output of energy to its user. They usually work on using elastic or some other potential
energy which is stored during one sequence of gait, and then released in another. This is
far more efficient than using peg leg prosthesis, but still not suitable enough to naturally
perform common gait activities. Various types of passive prostheses enable performing
of different activities such as walking on levelled and inclined ground, and even running,

https://doi.org/10.1007/978-3-319-90893-9_26
212 M. Rupar et al.
riding a bicycle and as of lately swimming. However, performing high power demanding
tasks, such is stair ascent, presents a problem for passive prostheses.
In the process of ascending stairs, during loading response and midstance phase,
body is being lifted from one stair to another and only one leg touches the stair and lifts
the body, while another is in swing phase (in the air). In this phase the entire load is on
one leg, and body is in such position that moments in the knee and ankle joints are very
high, and exceed body weight by several times. In case of above knee amputation, the
lack of muscles makes it impossible to produce these forces, and passive prostheses
which only use stored energy are unable to perform such activities. This means, that in
order for the prosthesis to be able to perform high demanding power activities, it must
be powered, primarily in its main joints – knee and ankle. In this paper, we are presenting
hydraulic power system and control system for knee and ankle powered prosthesis which
we are developing in order to achieve required kinematics and dynamics of the prosthesis
which would enable it to perform high power demanding activities in more natural
manner, especially stair ascent.
2 Prosthesis Power System
Our prosthesis is designed in a way to mimic main leg movements in sagittal plane. Its
key features are knee and ankle joints which are movable in sagittal plane and are exter‐
nally powered. Movements of knee and ankle enable obtaining required kinematics of
the prosthesis, and powering system enables its dynamic features [1]. Powered knee
enables overcoming large forces that occur during loading response. Having only
powered knee in the prosthesis design is not enough, since joint in the ankle is also very
important in natural stair ascent [2]. Powered ankle joint enables dorsal and plantar
flexion movement of the entire foot, which provides better stabilization of the knee and
the entire prosthesis and also provides power needed in push-off phase. Another key
feature is novel prosthetic foot with passive forefoot joint flexion ability which addi‐
tionally helps in making stair ascent more natural [3] (Fig. 1, left).
2.1 Hydraulic Power System of the Prosthesis
By enabling characteristic joint movements of the prosthesis, kinematic features in

sagittal plane are achieved in order to mimic movements of sound human leg. In order to
enable its dynamic features, external power system is needed. For prosthesis power system
we chose hydraulic actuators. Pneumatic and electric actuators were considered, but were
rejected since at the time of our initial tests they did not meet required size/power ratio.
Using pneumatic or electric actuators with sizes appropriate to leg prosthesis could not
give enough power which is required in order to enable stair ascent. Our hydraulic power
system concept is designed to enable characteristic movements of the prosthesis in sagittal
plane while walking and ascending stairs. Since it is far more difficult to achieve, we are
concentrated on testing the prosthesis on stair ascend. The base components of our
hydraulic system are two hydraulic actuators, one for powering the knee and other for
powering the ankle joint, hydraulic power pack unit and accompanying hydraulic
Power and Control System of Knee and Ankle Powered Above-Knee Prosthesis 213
installation. We chose off-the-shelf hydraulic power pack unit because it consists of all
needed hydraulic installation components (electrical motor, hydraulic pump, reservoir,
appropriate valves, connections etc.) integrated in one whole unit [3].
For our initial tests, we used unidirectional power pack unit. Unidirectional power
system was used to power prosthesis during stance phase, and during swing phase pros‐
thesis was free to move/swing (Fig. 1, right).
Fig. 1. Prosthesis design: (a) Powered knee (b) Powered ankle (c) Passive forefoot flexion [3]
(left); prosthesis with externally integrated hydraulic power system and actuators [3] (right)
2.2 Current Development and Future Work on Prosthesis Hydraulic Power

System
Our tests showed a step forward toward more natural stair ascent with powered ankle,
powered knee and novel prosthetic foot. Hydraulic system was able to overcome loads
of stair ascent, and features of the prosthesis (powered knee/ankle and forefoot flexion)
enabled more natural kinematics. However, using unidirectional hydraulic system
lacked control of the prosthesis during swing phase. This showed as downside since
prosthesis could not provide proper kinematics in swing phase. In order to improve this,
switch had to be made from unidirectional to reversible hydraulic power system (Fig. 2).
Reversible power system enables fine tuning of the position of the linear actuators in
Fig. 2. Reversible hydraulic power system of the prosthesis [3]

214 M. Rupar et al.
both directions and hence mimics the movements of the sound leg more precisely. Tests
with reversible hydraulic power system are currently being carried out, and implemen‐
tation of two independent power systems is being tested.
3 Prosthesis Control System
After completing the work on the prosthetic mechanism and power system, which
represented the first major phase of the design of the above-knee prosthesis, and after
demonstrating the initial statement that the problem of stair ascent could be practically
solved, we started with another phase of prosthetic design which focuses on the imple‐
mentation of sensors and electronic control.
Operating the prosthesis in our initial tests was performed completely manually by
switch with two operating and one neutral position. The first position activates the
system and prosthesis is in working mode performing stair ascent. In neutral position
hydraulic valve is open and system can move freely, which is used for prosthesis swing
phase. The second position closes the valve which disables system movements, and is
used for standing mode [3].
For operating the prosthesis, we also used remote Wi-Fi control by Smartphone,
based on Arduino Uno microcontroller. However, this showed to be inconvenient for
amputee compared to manual switch, since for using Smartphone both hands are needed,
and amputee has to have at least one hand free in order to maintain balance by holding
onto a stair railing.
3.1 Current Development and Future Work on Prosthesis Control System
In order to create an ideal prosthesis, control system must be able to recognize and
assume user’s intended motion and to automatically adapt its movements to different
walking speeds and different terrain. In our case that comes down to the fact that the
prosthesis has to follow the same angle trajectories of the knee and ankle joints as those
of a sound leg of an amputee. In order to achieve this, it is necessary to perform a full
range of experiments and optimization until final results are achieved.
The idea is to have the prosthetic controller determine the joint inputs required at
each instant in time to make the prosthesis track a commanded motion [4]. This could
be a sequence of end-effectors’ positions and orientations or a continuous path. Various
control methodologies are available for this kind of control, including proportional-
integral-derivative (PID) control, passivity-based control, robust control, adaptive
control, etc. The particular control method used may cause significantly different
performance. Therefore, each method has its own range of applications. For example,
PID control, the most common type of control algorithm in industry, works well for
independent joint setpoint regulation problems.
In this version, for the first time, six sensors, which are a combination of a gyroscope
and an accelerometer, are integrated along with the MPU 60/50 development board.
Three sensors are placed on the prosthesis and three on the sound subject or the sound
leg of an amputee. The sensors serve to map the angles of the leg and more precisely
Power and Control System of Knee and Ankle Powered Above-Knee Prosthesis 215
the angles of the knee, ankle and the foot. To facilitate the placement of sensors on the
prosthesis and on the leg of a healthy subject, plastic cages were made for each sensor
individually through 3D prototyping. The same is done for the electronics (Fig. 3).
Fig. 3. Electronic and sensors for the real-time prosthetic control
4 Conclusion
Our current developments offer an improvement in above-knee prosthetic design but

further optimisation is still required.
One way to optimise our current prosthetic design is through the integration of power
system into the prosthetic’s housing. This means that smaller components need to be
implemented into a design, but with special attention to their positioning in order not to
affect the moments of inertia of the prosthesis.
Other improvements could be made on control strategy applied. Future solutions
could make use of learning algorithms and adaptive control in order to reduce the adap‐
tation period necessary to adjust to the use of the prosthetic device.
With rapid development of different novel prosthetic solutions, laboratory testing of
the prosthesis becomes a necessary requirement to verify new concepts prior to their
application on human subjects. Our new prosthesis could make the gait of above knee
amputees generally more natural, and in the near future with the latest and more precise
technology, it might be able to fully mimic the natural human gait.
References
1. Rupar, M., Vučina, A., Dedić, R., Đindo, H.: Overview of the development of hydraulic above
knee prosthesis. In: Badnjevic, A. (ed.) CMBEBIH 2017: Proceedings of the International
Conference on Medical and Biological Engineering 2017, pp. 218–222. Springer, Singapore
(2017)
2. Vučina, A.: Prilog istraživanju kinematike sila u natkoljeničnoj protezi pri gibanju na
stepenicama. Doktorska disertacija. Fakultet strojarstva i računarstva, Sveučilište u Mostaru
(2002)
216 M. Rupar et al.
3. Rupar, M.: Contribution to research of above knee prosthesis with active hydraulically powered
knee and ankle joints and prosthetic foot with passive forefoot joint flexion. Doctoral Thesis.
Faculty of Mechanical Engineering and Computing, University of Mostar (2017)
4. Jelačić, Z.: Contact force problem in the rehabilitation robot control design. In: Badnjevic, A.
(ed.) CMBEBIH 2017: Proceedings of the International Conference on Medical and Biological
Engineering 2017, pp. 193–204. Springer, Singapore (2017)
Computer Science, Information and
Communication Technologies,
e-Business
Cyber Security Capacity Building
Planning Within Organisations
Ramo Šendelj ✉ and Ivana Ognjanović

( )
University of Donja Gorica, 81 000 Podgorica, Montenegro

{ramo.sendelj,ivana.ognjanovic}@udg.edu.me
Abstract. Organisations are facing challenges of protecting their data in cyber‐

space since damages of cyberattacks might potentially run into billions, besides
killing investor confidence and denting brand image. Recent data shows that one
of key threats for organisations are insider attacks caused by person who has/had
authorized access to an organization’s information system, and thus having the
knowledge about the system and being a potential point of intrusion, and jeop‐
ardizing the availability, integrity and confidentiality of the organization’s infor‐
mation systems. That is a reason why we analyse key factors influencing planning
of capacity building activities for employees, and thus making optimal planning
of raising awareness, training and education activities.
Keywords: Cyber security · User behaviour · Capacity building · Training

Awareness
1 Introduction
Many governments and multinational companies have recognized insider cyber security
threats as one of the most important challenges of modern life and business. Insider
threats are usually defined as a person who has/had authorized access to an organiza‐
tion’s information system, and thus having the knowledge about the system and being
a potential point of intrusion, and jeopardizing the availability, integrity and confiden‐
tiality of the organization’s information systems [1, 2]. The most popular forms of the
insider threats are: theft, espionage, fraud, and sabotage.
On the other side, several studies were aimed on understanding employees’ behav‐
iour and actions that gratefully affected the security of the system, as well as appropriate
counter measures for addressing this kind of attacks. Reported findings are various,
including psychological characteristics and aspects, such as “majority of people want to
get their jobs done more than they are interested in protecting themselves, a behavioural
tendency that gives surface for attacks” [2], different cyber security awareness level of
end users [3], compliance with security policies and intention and use of antimalware
software [4], etc. However, the organisations need to measure and report on the state of
information security culture in the organisation [5], which shall be used as essential input
for new measures how to increase the level of cyber awareness and knowledge within
organisation and thus how to overcome insider threats.

https://doi.org/10.1007/978-3-319-90893-9_27
220 R. Šendelj and I. Ognjanović
Education planning and maximizing learning progress and achievements attract

large audience during last decades especially encouraged by development of eLearning
approaches and supporting tools. By having rapid changes in dynamic cyber world in
central focus, issue of capacity building activities can be considered as permanent insider
threat treatments approach. That is a reason why we put focus on identification of essen‐
tial indicators needed for effective planning of cyber security training approach at the
level of organisation, i.e. by considering different employees, working at different posi‐
tions and having different background knowledge and different profiles. Research study
reported in this paper presents one step on the way moving towards creation of auto‐
matically and dynamically adaptive training assistance tool, and it is implemented by
conducting experimental study among both, cyber experts and cyber trainers, with key
aim on selecting appropriate set of courses on the bases of different employees’ profiles.
The paper is organised as follows, Sect. 2 provides details about different insider
cyber threats for the organisations, while Sect. 3 presents different educational
approaches in the field of cyber security. Section 4 explains experimental setup, the
method and presents research analyses. Section 5 concludes the paper with critical anal‐
yses and key suggestions for future work.
2 Insider Cyber Threats for Organisations
Rapid use of Internet make businesses placed in cyber space, and thus organisations are
facing challenges of protecting their data since damages of cyberattacks might poten‐
tially run into billions, besides killing investor confidence and denting brand image [6].
2.1 Insider Threats for the Organisations
According to the IBM 2016 Cybersecurity Intelligence Index, 60% of security incidents
were carried out by insiders, either with malicious intent or carelessness, where an
insider is considered as any employee with physical or remote access to the company
assets [1, 7].
Recent reports [7] identify the following biggest risks coming inside the organisa‐
tions and having employees as key actors using Internet and different devices, such as:
(i) Bring-your-own-device (BYOD) practices - one of the greatest security risks when
accessing to business information by using own devices (e.g. mobile devices); (ii)
Threats of Shadow IT - growing risks of using non-approved applications deployed on
clouds thus being outside of network controls scopes; (iii) Unapproved/unsecured
downloads - even being usual activities, it is one of key practices which make system
vulnerable to malwares and hacks; (iv) Unsecure IoT - by using different devices
connected to the Internet, more entry points to organisational system are opened thus
becoming potential attack points; (v) Unprotected unpatched devices - firewalls or/and
anti-virus programs may not be installed on devices that are used inside the organisa‐
tions, and thus making new security gaps; (vi) Risks of unauthorised access and authen‐
tication - omissions in the process of authentication and authorisation might directly
cause huge loses of data and damages; (vii) Illegal/unethical activities - scenarios of
Cyber Security Capacity Building Planning Within Organisations 221
using illegal data usage can be one of drastic ways for disturbance of both, individuals
and organisations; and (viii) The menace of social engineering - one of the most
dangerous attacks distributed via Trojans and phishing by email and other applications
attachments.
However, employees’ behaviours can directly or indirectly impose any of afore‐
mentioned risks, and almost 30% of them are accidental (as reported by the Verizon’s
2016 Data Breach Investigate Report [8]). On the other side, employees’ activities may
be aimed on using some applications (from the web) to improve usual business (known
as “Negligent security risk” [8]), while there are employees whose activities are not for
underestimation and they are related to intentional bad impacts for organizations,
including espionage, financial loss, activity, compromising business security (known as
“Malicious”) and still are the most potent.
Table 1. Security Behaviour Intentions Scale (SeBIS) questions and associated sub-scales [9]
Group. Device securement Group. Password generation
1. I set my computer screen to automatically 1. I do not change my passwords, unless I have
lock if I don’t use it for a prolonged period of to
time 2. I use differen tpasswords for different
2. I use a password/passcode to unlock my accounts that I have
laptop or tablet 3. When I create a new online account, I try to
3. I manually lock my computer screen when I use a password that goes beyond the site’s
step away from it minimum requirements
4. I use a PIN or passcode to unlock my 4. I do not include special characters in my
mobilephone password if it’s not required
Group. Updating Group. Proactive awareness
1. When I’m prompted about a software update, 1. When someone sends me a link, I open it
I install it right away without first verifying where it goes
2. I try to make sure that the programs I use are 2. I know what website I’m visiting based on
up-to-date its look and feel, rather than by looking at the
3. I verify that my anti-virus software has been URL bar
regularly updating itself 3. I submit information to websites without first
verifying that it will be sent securely (e.g., SSL,
“https://”, a lock icon)
4. When browsing websites, I mouse overlinks
to see where they go, before clicking them
5. If I discover a security problem, I continue
what I was doing because I assume someone
else will fixit
2.2 End User Behaviours
There are numerous examples in the research literature examining the cyber security
awareness level of end users, their behaviour and its association with potential cyber
threats. For instance, [3] have over a number of years, analyzed security-related end
users behaviours with recognizable dimensions to assess and/or influence their behav‐
iour within organisation. There is no agreement among researchers how to model user
behaviours since majority of approaches are focused on questionnaires conducted by

users, thus putting in focus the objectivity and the willingness to report the truth (espe‐
cially when we are talking about potentially malicious employees). The recently devel‐
oped model is made by experts who offered end-users to construct a set of Likert scale
questions resulting with “Security Behaviour Intentions Scale (SeBIS)”-an 16-item scale
of four sub-scales that measures attitudes towards choosing passwords, device secure‐
ment, staying up-to-date, and proactive awareness [9] (see Table 1). In this paper SeBIS
scale is used for describing different kinds of user behaviours and actions, as SeBIS scale
has proven correlation with the psychological constructs Consideration for Future
Consequences (CFC) [10], which is a unidimensional scale measuring how much atten‐
tion people pay to long-term consequences.
3 Cyber Education Approaches
Development of eLearning approaches and supporting tools significantly contributed to

modernization and effectiveness of different learning and training approaches. More‐
over, in order to create educational program in IT security, the comprehensiveness of
both human and technological issues are essential, ranging from, security policy, proce‐
dures, and techniques to various management, operational and technical controls [11].
On the other side, IT security programs must be in line with directives and established
security policies [11], and in the same time being adapted to the needs, background
knowledge capacities of each individual trainee.
However, there is no general approach for capacity building in cyber security fields,
and NIST [11] suggest the following three categories of courses and educational
approaches:
(i) Security awareness - aimed on changing behaviour, strengthening good security
practices, and recognition of IT security concerns and relevant counter actions;
(ii) Training - continuous process of getting knowledge aimed on gaining practical
security skills and competencies;
(iii) Education - as defined in NIST [11] “integrates all of the security skills and
competencies of the various functional specialties into a common body of knowl‐
edge, adds a multidisciplinary study of concepts, issues, and principles, and strives
to produce IT security specialists and professionals capable of vision and pro-
active response.”
Furthermore, there is no unique distribution of courses per categories, variety of
courses and curricula are available for all three categories. For the purpose of research
presented in this paper, common characterisation of courses available at [12] is used.
4 Experimental Analyses
4.1 Context
Since the key aim of this study is to analyse the nature of effective planning of cyber
security training, from the perspective of organisation and employees’ characteristics,
the following research questions are defined.
RQ1. What is the impact of employees’ characteristics on cyber security training

planning?
The question addresses the need to consider the employees’ existing level of aware‐
ness and knowledge about cyber security threats for organisations. However, while there
has been much research focused on analysing different security behaviour scales, the
research question addresses different employees’ characteristics, including gender,
working position, and 16-item scale consisting of four sub-scales that measures attitudes
towards choosing passwords, device securement, staying up-to-date, and proactive
awareness [9].
RQ2. What is the impact of organisation’s characteristics on cyber security training

planning?
The question analysis if specific characteristics of organisations might impact the
suggested educational approach, including type of organisation and sector, existing level
of IT security policies, etc.
Implemented research study included active participation of 20 cyber security
experts (including both, practitioners and certified trainers) who were invited to make
own assessment about capacity building training sets for 10 different categories of
employees. The categories of employees were characterized in accordance with 16-item
Security Behavior Intentions Scale [9], and basic demographic data: age, gender, years
of working experience, as well as working positions, and if employee attended at least
3 trainings about security issues. Additionally, basic characteristics about organizations
were also available, including: sector, field, years of existence, the level of soundness
of IT security policies (1-does not exist, 2-exist but no implementation, 3-exist with
implementation mechanisms, 4-exist with strong implementation mechanisms).
Furthermore, cyber security experts were invited to make suggestion of capacity building
plan for each individual employee, by selecting appropriate courses from the list of avail‐
able courses from 3 categories: Raising Awareness, Training and Education [1, 12].
4.2 Data Collection and Analysis

Invited cyber security experts worked on the same assignments (64 different categories
of employees) independently and the only limitation was in total number of suggested
courses which was set to 5. Collected data were encoded (selected courses are coded
with two arrays: one presenting position within belonging group, and the other
presenting groups containing selected courses) in order to allow different statistical
analyses.
Due to the nature and type of the data collected, the analyses employed were standard
descriptive statistics [13] including reported mean (M) and standard deviation (SD)
values. A ANOVA test was used to assess if there are significant differences between
courses and groups of courses suggested by different experts; as well as to assess if
different values of specific characteristics significantly impact on proposed sets of
courses.
4.3 Results
The descriptive statistics for selected courses (groups 1–3) are reported in Table 1.
Average number of selected courses per employee profile was 3.75 (SD = 2.41). The
results demonstrate significant difference between courses groups (F(2, 63) = 5.11; p =
0.003).
Table 2. The descriptive statistics of the selected courses from the conducted study
Measure Group 1: awareness Group 2: training Group 3: education
M, SD 2.21, 1.56 1.53, 1.27 1.15, 0.93
The RQ1 compares the differences between selected courses and identifies charac‐
teristics from different employees’ profiles with significant importance on courses
recommendation, as follows:
– Employees profiles are firstly divided into two groups based on gender value, and
there is no statistical difference revelled: F(1, 63) = 2.53; p > 0.05; while by dividing
them into 4 groups based on generic working positions (e.g. administrative worker
with technical tasks, administrative worker with no technical tasks, manager with
technical tasks, manager with no with technical tasks), the difference is significant:
F(3, 63) = 8.72; p = 0.01.
– When employees are divided based on four SeBISsub-scales, various differences are
identified between groups, as presented in Table 3.
Table 3. Differences in suggested courses between groups of SeBIS sub-scales

SeBIS Choosing Device Staying Proactive
sub-scales passwords securement up-to-date awareness
Choosing - F(15, 63) = 2.02, F(11, 63) = 2.83, F(19, 63) = 1.82,
passwords p = 0.01 p = 0.05 p > 0.05
Device - - F(11, 63) = 0.94, F(19, 63) = 0.15,
securement p > 0.05 p > 0.05
Staying - - - F(14, 63) = 4.17,
up-to-date p = 0.001
Proactive - - - -
awareness
– Finally, the fact if employee attended at least 3 training events did not make any
significance between courses F(1, 63) = 1.78; p > 0.05, while it showed significant
correlation with 16-item SeBISscale.
RQ2 is focused on analysing the impacts of key organisational characteristics on
courses’ suggestions. The results show that in case of different values for sector and
years of experience for the organisation, there are no significant differences in courses:
F(3, 63) = 2.02; p > 0.05 and F(4, 63) = 1.63; p > 0.05; while there is statistically
significant difference in case of different fields: F(4, 63) = 5.14; p = 0.05.
On the other side, the impact of the level of soundness of IT security policies within
organisation is significant F(3, 63) = 7.05, p = 0.000. The Tukey post hoc test revealed
that there was no significant differences in the following two case: between organisations
which does not have any IT security policies and those having it but with no imple‐
mentation framework; as well as between those having IT security policies with strong
implementation mechanisms and those with not such strong mechanisms for imple‐
mentation and monitoring.
5 Discussion with Conclusions
Presented study examined different factors influencing selection of appropriate capacity

building activities aimed on decreasing the level of insider threats within organisation.
Even the study has clear limitations, such as: artificial generation of employees’ char‐
acteristics, lack of realistic data about employees’ personality, organisational environ‐
ment and contexts, reported findings are made on the bases of completely independent
evaluation and participation of 20 cyber security experts (including both, practitioners
and certified trainers), and gave the following findings:
– Even several researchers [14] highlighted the differences in security behaviour
between males and females, our study reported that similar capacity building
approaches can be applied for both, if their other characteristics are the same;
– Organisational perspective on IT security directly impacts the employees’ behaviour
and their attitudes [15], while it is at individual level, due to its psychological profile
and still personal decision how the employee will behave within organisation.
However, our study showed that organisational perspective is important to be consid‐
ered when planning capacity building activities for its employees;
– SeBIS sub-scales is used in this paper as measurement of attitudes towards choosing
passwords, device securement, staying up-to-date, and proactive awareness [9]
exemplified the importance of having high accuracy in assessment of employees’
attitudes, usual actions and habits in cyber space. However, as reported in the liter‐
ature [2] it is still under concerns of using direct questionnaires and survey analyses
(due to identified psychological impacts on the answers [16]) and it is more reason‐
able to expect integration of other techniques (such as, intelligent reasoning, data
mining etc.) to revile characteristics of employees’ attitudes and behaviour [15].
However, this paper provided analyses on identification of key factors influencing
capacity building program for employees at organisations. Furthermore, it is shown that
different sets of information are needed in order to create plan which can address both,
the needs of organisation from the aspect of building security posture and enhancing the
employees’ behaviour from the aspect of potential insider threats. Reported results make
a significant foundation for further analyses of optimal capacity building training plan‐
ning when other external factors shall be examined, like limited time and financial
resources, changes inside organisation, rapid developments in cyber space, etc. Future
research might also go one step further on implementing monitoring applications for
employees’ actions and alerting systems for potential suspicious behaviours and sugges‐
tion of permanent measures.
References
1. NIST: Guide for Conducting Risk Assessments: Information Security. NIST Special
Publication 800-30 (2012)
2. Metalidou, E., Marinagi, C., Trivellas, P., Eberhagen, N., Giannakopoulos, G., Skourlas, C.:
Human factor and information security in higher education. J. Syst. Inf. Technol. 16(3), 210–
221 (2014)
3. Stanton, J.M., Stam, K.R., Mastrangelo, P., Jolton, J.: Analysis of end user security behaviors.
Comput. Secur. 24, 124–133 (2005)
4. Crossler, R.E., Bélanger, F., Ormond, D.: Inf. Syst. Front. (2017). https://doi.org/10.1007/
s10796-017-9755-1
5. Da Veiga, A., Eloff, J.H.: A framework and assessment instrument for information security
culture. Comput. Secur. 29, 196–207 (2010)
6. U. Edition: Clearswift Insider Threat Index (CITI). Clearswift (2015)
7. 2016 Cost of Data Breach Study: Global Analysis. Benchmark Research Sponsored by IBM.
Ponemon Institute LLC (2016)
8. The Verizon’s 2016 Data Breach Investigate Report. Verizon (2016)
9. Egelman, S., Peer, E.: Scaling the security wall: developing a Security Behavior Intentions
Scale (SeBIS). In: CHI 2015 (2015)
10. Joireman, J., Shaffer, M., Balliet, D., Strathman, A.: Promotion orientation explains why
future-oriented people exercise and eat healthy evidence from the two-factor consideration
of future consequences-14 scale. Pers. Soc. Psychol. Bull. 38(10), 1272–1287 (2012)
11. Arden, L., Bement, J.: Building an Information Technology Security Awareness and Training
Program. National Institute of Standards and Technology (2003)
12. Training Modules by SANS Security Awareness. https://www.sans.org/security-awareness-
training/products/end-user
13. Blaikie, N.: Analyzing Quantitative Data. Sage, London (2003)
14. Anwar, M., He, W., Ash, I., Yuan, X., Li, L., Xu, L.: Gender difference and employees’
cybersecurity behaviors. Comput. Hum. Behav. 69, 437–443 (2017)
15. Rees, L.P., Deane, J.K., Rakes, T.R., Baker, W.H.: Decision support for Cybersecurity risk
planning. Decis. Support Syst. 51, 493–505 (2011)
16. Shropshire, J., Warkentin, M., Sharma, S.: Presonality, attitudes, and intensions: predicitng
initial adoption of information security behavior. Comput. Secur. 49, 177–191 (2015)
New Method of Sequences Spiral
Hybrid Using Machine Learning Systems
and Its Application to Engineering
Matej Babič1 ✉ , Isak Karabegović2, Sanda Ipšič Martinčič3, and Gyula Varga4
( )
1
Jožef Stefan Institute, 1000 Ljubljana, Slovenia
babicster@gmail.com
2
3
University of Rijeka, 51000 Rijeka, Croatia
smarti@inf.uniri.hr
4
University of Miskolc, 3515 Miskolc, Hungary
gyulavarga@uni-miskolc.hu
Abstract. In an era of increased emphasis on sustainability and quality assur‐

ance, knowledge about metals and other materials used in products, manufac‐
turing processes, and construction is invaluable. Metallurgy is the study of the
physical and chemical behaviour of metallic elements. CNC operators typically
test many materials with different CNC machine parameters to optimize the topo‐
logical properties of materials. In this article we present a solution to this prob‐
lems. We analyse SEM pictures of the microstructure of robot laser hardened
specimens using graph theory and fractal geometry. Intelligent systems methods
enable predictions for mechanical engineering based on a hybrid of genetic
programming and multiple regression, with applications to metallurgy and
mechanical engineering. Hybrid evolutionary computation is a generic, flexible,
robust, and versatile method for solving complex global optimisation problems
that can also be used in practical applications. Hybrid intelligent systems enhance
laser hardening by decreasing the process time and increasing the topographical
properties of materials.
Keywords: Intelligent system · Hybrid machine learning · Laser · Hardening

Visibility graphs · Fractal dimension · Metallurgy
1 Introduction
Metallurgy [1] is the domain of materials science and materials engineering that studies
the physical and chemical behaviour of metallic elements and their mixtures, which are
called alloys. Metallurgical engineers transform metals into products useful for society,
while minimizing waste, maximizing energy efficiency, increasing performance, and
facilitating recycling. Metallurgy requires a thorough grasp of the sciences, including
chemical and physical properties of metallic elements, compounds, and alloys. Metal‐
lurgy also includes metal-related technologies, such as extraction methods, use in

https://doi.org/10.1007/978-3-319-90893-9_28
228 M. Babič et al.
industry and engineering, and metalworking processes such as casting, forging, and
sintering. A current problem in metallurgy is describing the complexity of hardened
specimens by analysis of SEM pictures. In this article, we combine fractal geometry [2]
and graph theory [3] using the intelligent systems [4], machine learning, and hybrid
systems [5] of intelligent systems to address this problem. Machine learning has several
approaches, so selecting the appropriate approach is highly dependent on the problem,
its complexity, and its simplification options. Approaches may also be combined in a
hybrid system if problems are first segmented. The field of machine learning offers many
possible implementations of solutions to problems from various fields. Machine learning
is a field of artificial intelligence that deals with development of techniques that allow
a computer or a machine to learn. Machine learning is basically a method for creating
computer programs on the basis of data (samples). It relies heavily on statistics, which
also relies on data, but in contrast, machine learning focuses on the algorithms and
computational operations. Machine learning has a variety of applications, such as search
engines (web), medical diagnosis, detection of forged documents, analysis of trends in
prices on the stock markets, identifying DNA sequences, speech and font recognition,
identification of objects in machine vision, computer games, robotics, etc. Some
machine learning systems attempt to eliminate the need for human intuition when analy‐
sing data, while other systems are based on collaboration between humans and machines.
Machine learning algorithms can be divided into different types depending on their goal
or the result of learning. Hardening heat treatments invariably involve heating to a
sufficiently high temperature to dissolve solute-rich precipitate. The metal is then rapidly
cooled to avoid reprecipitation, often by quenching in water or oil. The concentration
of solute dissolved in the metal is then much greater than the equilibrium concentration.
This produces what is known as solid-solution hardening, but the alloy can usually be
hardened further by ageing to allow formation of a very fine precipitate. Ageing is
performed at an elevated temperature but still well below the temperature at which the
precipitate will dissolve. This treatment may improve mechanical properties and
increase hardness, producing a tougher, more durable item. Alloys are heated above the
critical transformation temperature for the material, then cooled rapidly enough to cause
the soft initial material to transform into a much harder, stronger structure. Alloys may
be air cooled or cooled by quenching in oil, water, or another liquid, depending on the
amount of alloying element in the material. Hardened materials are usually tempered or
stress-relieved to improve their dimensional stability and hardness. Steel parts often
require heat treatment to obtain improved mechanical properties, such as increased
hardness or strength. Hardening consists of heating the components above the critical
(normalising) temperature, holding at this temperature for one hour per inch of thickness,
cooling at a rate fast enough to allow the material to transform to a harder, stronger
structure, and then tempering. Steel is essentially an alloy of iron and carbon; other steel
alloys have other metal elements in solution. Heating the material above the critical
temperature causes carbon and the other elements to enter the solid solution. Quenching
“freezes” the microstructure, inducing stresses. Parts are subsequently tempered to
transform the microstructure, achieve the appropriate hardness, and eliminate stresses.
We used three methods for analysing SEM images of robot laser hardened specimens
hardened at different temperatures and speeds [6]. We calculated the fractal dimension
New Method of Sequences Spiral Hybrid Using Machine Learning Systems 229
in 2D space, 3D space, and calculated the density of visibility graphs in 3D space.

Fig. 1 shows metallography tool steel used to study the microstructure of the robot laser
hardened specimen. This report focuses on hardened alloys tested at room temperature
and prediction of the topographical property hardness of hardened specimens with intel‐
ligent systems using a hybrid method, with the goal of improving existing hybrids. We
describe a new hybrid method based on the sequences spiral method.
Fig. 1. Metallography tool steel used to study the microstructure of metals.
2 Experimental Method
First, we hardened tool steel with a robot laser cell. We changed two parameters: speed
v ∈ [2, 5] mm/s and temperature T ∈ [1000, 1400] °C. After hardening, we polished and
etched all specimens. Detailed characterization of their microstructure before and after
surface modifications was conducted using a field emission scanning electron micro‐
scope (SEM), JEOL JSM-7600F. We use three methods for analysing SEM images,
including two methods for calculating the fractal dimension in 2D [7] and 3D [8] space,
and an algorithm for constructing 3D visibility graphs [9], as summarized in Fig. 2. An
intelligent systems method including neural networks [10], genetic programming [11],
and multiple regression [12] was used for analysis. The network consists of layers of
nodes (input, hidden, and output layers) that are interconnected by artificial neurons
which develop weights that modulate signals pass through them. Each node fires an
impulse when the sum of its inputs exceeds a threshold. Supervised classification
involves using a training set (i.e. learning set) consisting of examples of inputs corre‐
sponding to desired outputs to ‘‘train’’ the network. Training involves adjusting the
weights of the artificial neurons so that the desired outputs are achieved. Unsupervised
classification seeks to find the natural classes into which the data may be subdivided.
Evolutionary computation (EC) is the study of computational systems that borrow ideas
from, and are inspired by, natural evolution and adaptation. EC covers a number of
techniques based on evolutionary processes and natural selection: evolutionary strat‐
egies, genetic algorithms, and genetic programming. Evolutionary strategies are an
approach for efficiently solving certain continuous problems, yielding good results for
some parametric problems in real domains. Compared to genetic algorithms, evolu‐
tionary strategies run more exploratory searches and are a good option when applied to
relatively unknown parametric problems. Genetic algorithms emulate the evolutionary
process that takes place in nature. Individuals compete for survival by adapting to envi‐
ronmental conditions. Crossover between individuals, mutation, and death are all part
of adaptation. Genetic programming is an extension of genetic algorithms. Its aim is to
build computer programs that are not expressly designed and programmed by a human.
It can be considered as an optimisation technique whose search space is composed of
all possible computer programs for solving a particular problem. Genetic programming’s
key advantage over genetic algorithms is that it can handle individuals (i.e. computer
programs) of different lengths. In multiple regression analysis, the relationship between
one dependent variable and several independent variables (called predictors) is analysed.
The regression equation takes the form of (1).
Y = b𝟎 + b1 x1 + b2 x2 … + bp + e (1)
where Y is the dependent variable, bs are the regression coefficients for the corresponding
x (independent) terms, b0 is a constant or intercept, and e is the error term reflected in
the residuals. The parameters of the regression equation are estimated using ordinary
least squares (OLS).
Fig. 2. Three methods for image analysis: two methods for calculating the fractal dimension in
2D and 3D space, and the algorithm for constructing 3D visibility graphs
Hybrid evolutionary computation [13] is a generic, flexible, robust, and versatile

method for solving complex global optimisation problems that can also be used in prac‐
tical applications. We present a new hybrid: the sequences spiral hybrid. The intelligent
hybrid systems model is presented in Fig. 3.
Fig. 3. Intelligent hybrid systems model
3.1 Results
In Table 1, the parameters of hardened specimens that have an impact on hardness are
presented. We mark specimens from P1 to P21. Parameter X1 represents temperature in
degrees Celsius [°C], X2 represents the speed of hardening [mm/s], X3 represents fractal
dimension in 2D space, X4 represents fractal dimension in 3D space, X5 represents the
density of visibility graphs in 3D space, and parameter X6 represents the basic hardness
of specimens. The final parameter Y is measured surface hardness of laser-hardened
robot specimens. Table 2 presents experimental and prediction data for the surface
hardness of laser hardened robot specimens. In Table 2, the symbol S stands for the
name of specimens, E stands for the experimental data, NM1 stands for the prediction
with the neural network with a 30% learn set, NM2 stands for the prediction with the
neural network with a 50% learn set, NM3 stands for the prediction with the neural
network with method one live out, R stands for the prediction with regression, GP stands
for the prediction with genetic programming, HM1 stands for the prediction with the
insert hybrid method of intelligent systems, and HM2 stands for the prediction with the
sequences spiral hybrid method of intelligent systems. In Table 1, we can see that
specimen P17 has the largest density of visibility graphs in 3D: 0.2832. Thus, specimen
P17 has the most complex graph. Specimen P1 has maximum hardness after hardening,
60 HRC. The measured and predicted surface hardness of robot laser hardened speci‐
mens is shown in Fig. 3. The genetic programming model is presented in Fig. 2, with
the regression model below it. The genetic programming model has a 1.97% deviation
from the measured data, which is less than the regression model, which has a 3.27%
deviation. The best neural network NN3 has a 0.81% deviation from the measured data.
Neural network NN1 has a 4.47% deviation from the measured data and NN2 has a
2.56% deviation from the measured data. The hybrid mathematical model of intelligent
system GP-MR is presented in Fig. 4 and has an 848.02% deviation from the measured
data. The new sequence spiral hybrid method of intelligent system has a 3.39% deviation
from the measured data. Equation (4) presents a model of the hybrid machine learning
system GP-MR.
Table 1. Parameters of hardened specimens

S X1 X2 X3 X4 X5 X6 Y
P1 1000.0 2.0 1.9135 2.304 0.1936 34.0 60.0
P2 1000.0 3.0 1.9595 2.264 0.2208 34.0 58.7
P3 1000.0 4.0 1.9474 2.258 0.2144 34.0 56.0
P4 1000.0 5.0 1.9384 2.341 0.2256 34.0 56.5
P5 1400.0 2.0 1.9225 2.222 0.2445 34.0 58.0
P6 1400.0 3.0 1.9784 2.388 0.2221 34.0 57.8
P7 1400.0 4.0 1.9540 2.25 0.2036 34.0 58.1
P8 1400.0 5.0 1.9776 2.286 0.2096 34.0 58.2
P9 1000.0 2.0 1.9720 2.178 0.2352 60.0 57.4
P10 1000.0 3.0 1.8580 2.183 0.2288 58.7 56.1
P11 1000.0 4.0 1.9784 2.408 0.2144 56.0 53.8
P12 1000.0 5.0 1.9410 2.210 0.2352 56.5 56.0
P13 1400.0 2.0 1.9784 2.257 0.2208 58.0 55.3
P14 1400.0 3.0 1.5810 2.265 0.232 57.8 57.2
P15 1400.0 4.0 1.9650 2.433 0.1984 58.1 57.8
P16 1400.0 5.0 1.8113 2.289 0.1904 58.2 58.0
P17 800.0 0.0 1.9669 2.232 0.2832 34.0 52.0
P18 1400.0 0.0 1.9753 2.235 0.2688 34.0 57.0
P19 2000.0 0.0 1.9706 2.261 0.2416 34.0 56.0
P20 950.0 0.0 1.9631 2.282 0.2128 34.0 58.0
Table 2. Experimental and prediction data

S E NN1 NN2 NN3 R GP HM1
P1 60.0 60.01087 60.06062 59.9681 59.03441 59.8 35.03963
P2 58.7 57.91091 57.84108 57.84108 56.40501 58.2 44.08994
P3 56.0 56.83736 56.63749 56.73611 57.26038 56.0 26.01645
P4 56.5 56.32328 56.47253 56.45682 54.17947 58.3 44.45627
P5 58.0 57.95455 58.04968 58.1518 56.92762 57.8 86.96554
P6 57.8 58.4084 58.06627 57.89607 56.30819 59.4 84.63897
P7 58.1 58.23035 58.09943 58.14831 61.28071 56.7 89.32436
P8 58.2 57.59059 57.96153 57.86465 59.76056 57.5 86.97091
P9 57.4 53.55737 57.39159 57.32264 55.17812 57.5 83.76528
P10 56.1 48.10235 56.12515 56.15657 55.91692 56.8 46.04427
P11 53.8 53.21436 53.76771 53.78168 53.24036 58.7 27.02772
P12 56.0 54.49331 51.10043 52.55539 54.54833 56.0 43.97276
P13 55.3 57.96328 57.47276 55.28552 58.00671 58.5 89.15543
P14 57.2 44.18085 56.2936 57.2083 56.58243 57.2 86.43758
P15 57.8 56.58687 55.59012 57.89345 57.15531 60.0 85.89897
P16 58.0 49.59263 55.3824 57.92138 61.002 57.5 89.49785
P17 52.0 56.86639 55.04666 53.04916 48.2581 55.1 23.52358
P18 57.0 58.88932 59.97596 56.79023 53.76958 55.7 68.99137
P19 56.0 59.46188 60.33119 57.01954 60.40532 56.3 99.59686
P20 58.0 59.8025 60.96921 58.49307 56.88412 58.0 25.61841
Fig. 4. The measured and predicted hardness of robot laser hardened specimens
Model of genetic programming Eq. (2)
Y =R+E+F (2)
Where:
15.5046
7.03289 +
x22 x5
R = 7.03289 + + 0.280007 ⋅ x2 + 19.6057 ⋅ x4 +
x22 2.3256
G
E=
H
K
F = 0.142189 ⋅
L
( )
x
G = 0.142189 ⋅ −140.037 + x1 + 1
P⋅Q
( )
x2 x − x2 8.03287
H = − + 0.103739 ⋅ x4 ⋅ −7.14267 + 0.142189 ⋅ x1 + 4 +
x4 x2 x6
( )
7.03289 ⋅ x2
P = 0.142189 ⋅ x1 − 18.6057 ⋅ 7.14267 +
x1
( )
7.03287 + 1.93014 ⋅ x2
Q = 0.142189 ⋅ x1 − 18.6057 ⋅ 7.14267 +
x1
( )
0.142139 ⋅ x1 ⋅ −0.0518079 ⋅ x2 + x3 + M
K=
N
x1 ⋅ x2 + 7.03287
M= ( )
x2 ⋅ 7.03287 + 1.93014 ⋅ x2
( )
N = 0.142189 ⋅ x1 − 18.605 ⋅ 7.14267 + 0.142189 ⋅ x2
( )
x x − x2 8.03287
L = − 2 + 0.103739 ⋅ x4 ⋅ −7.14267 + 0.142189 ⋅ x1 + 4 +
x4 x2 x6
Model of regression Eq. (3)
Y = 124.7836 + 0.006336 ⋅ x1 − 0.05668 ⋅ x2 − 0.4049 ⋅ x3

(3)
− 19.8671 ⋅ x4 − 123.117 ⋅ x5 − 0.0467 ⋅ x6
Equation (4) gives the hybrid mathematical model of the intelligent system GP-MR.
Variable x5 of the multiple regression model (4) is then used in the genetic programming
model.
B⋅B−A+B⋅C
Y= (4)
D−1
where:
A = 124.7836 + 0.006336 ⋅ x1 − 0.05668 ⋅ x2 − 0.4049 ⋅ x3 − 19.8671 ⋅ x4 − 0.0467 ⋅ x6

B = 7.03289 + R
C =E+F
D = 123.117 ⋅ 2.3256
3.2 Discussion
Hardening is a metallurgical and metalworking process used to increase the hardness of

a metal. The hardness of a metal is directly proportional to the uniaxial yield stress at
the location of imposed strain. A harder metal will have a higher resistance to plastic
deformation than a less hard metal. We use the method of intelligent systems, genetic
programming, neural networks, and multiple regression to predict the topographical
property hardness of robot laser hardened specimens. The best parameters of the robot
laser cell that give us maximum hardness after hardening are a speed of 2 mm/s and a
temperature of 1000 °C. Hardness is a good predictor of the performance of a mechanical
component, since irregularities in the surface may form nucleation sites for cracks or
corrosion. A statistically significant relationship was found between hardness, the
parameters of the robot laser cell, the topological property of visibility graphs in 3D,
fractal dimension in 2D space, and fractal dimension in 3D space. The fractal analysis
in 2D and 3D space and topological property density of 3D visibility graphs of a series
of digitized surface microstructures from the robot laser surface modified specimens
indicated that useful correlations can be derived between the fractal dimensions and
surface microstructural features, such as hardness. We show that the neural network
gives us the best prediction results. The genetic programming model is better than the
regression model. The new hybrid system method of intelligent systems has a 848.02%
deviation from the measured data. This hybrid system method of intelligent system is
presented as a quadratic equation. In this quadratic equation, we must eliminate negative
solutions, because the hardness of robot laser hardened specimens must be positive. The
new sequence spiral hybrid method of intelligent systems has a 3.39% deviation from
the measured data. In addition, image analysis of SEM images of robot laser hardened
specimens is an interesting approach. X5 has the largest impact on the regression model,
while X4 has the largest impact on the genetic model.
4 Conclusions
This paper presents a new method using sequences spiral hybrid systems of machine
learning for use in engineering. We use the mathematical methods of graph theory,
fractal geometry, and its application in the robot laser hardening process. We present a
method of visibility graph in 3D space and fractal geometry to analyse the complexity
of robot laser hardened specimens, which is a very difficult problem. Our new method
has many applications in pattern recognition, computer graphics, computational geom‐
etry, among other uses. The main findings can be summarized as follows:
1. The method of graph theory and fractal geometry is used to analyse robot laser
hardened specimens.
2. The porosity of the hardened specimens using the topological properties hardness
of the visibility graphs in 3D space and the fractal dimension in 2D and 3D space
are described.
3. For prediction of the hardness of hardened specimens, the methods of intelligent
systems, e.g. neural networks, genetic algorithms, and multiple regression are used.
4. With the new hybrid system of intelligent systems, sequences spiral hybrid method,
we increase production of the robot laser hardening process, because the reducing
the process time and increasing the topographical property porosity of materials.
5. A new machine learning method is applied to metallurgy.
6. Two methods of hybrid intelligent systems are compared.
Acknowledgement. Project no. NKFI-125117 has been implemented with the support provided
from the National Research, Development and Innovation Fund of Hungary, financed under the
K_17 funding scheme.
References
1. Radivojević, M., Rehren, T., Pernicka, E., Šljivar, D., Brauns, M., Borić, D.: On the origins
of extractive metallurgy: new evidence from Europe. J. Archaeol. Sci. 37, 2775 (2010). https://
doi.org/10.1016/j.jas.2010.06.012
2. Chang, Q., Chen, D.L., Ru, H.Q., Yue, X.Y., Yu, L., Zhang, C.P.: Three-dimensional fractal
analysis of fracture surfaces in titanium–iron particulate reinforced hydroxyapatite
composites: relationship between fracture toughness and fractal dimension. J. Mater. Sci.
46, 6118–6123 (2011)
3. Ben-Moshe, B., Hall-Holt, O., Katz, M.J., Mitchell, J.S.B.: Computing the visibility graph of
points within a polygon. In: Proceedings of the 20th ACM Symposium on Computational
Geometry, pp. 27–35 (2004)
4. Woolf, B.: Intelligent tutoring systems: a survey. In: Shrobe, H.E. (ed.) Exploring Artificial
Intelligence: Survey Talks from the National Conferences on Artificial Intelligence, pp. 1–
43. Morgan Kaufmann, San Mateo (1988). American Association for Artificial Intelligence
5. Arosha, S.M.N., Senanayke, O., Malik, A., Mohammad, P.I., Zaheer, D.: Anterior cruciate
ligament recovery monitoring system using hybrid computational intelligent techniques. Int.
J. Hybrid Intell. Syst. 10, 215–235 (2013)
6. Grum, J., Žerovnik, P., Šturm, R.: Measurement and analysis of residual stresses after laser
hardening and laser surface melt hardening on flat specimens. In: Proceedings of the
Conference “Quenching 1996”, Ohio, Cleveland (1996). J. Surf. Engin. Mat. Adv. Technol.
3, 146. https://doi.org/10.4236/jsemat.2013.32019
7. Li, J., Du, Q., Sun, C.: An improved box-counting method for image fractal dimension
estimation. Pattern Recogn. 42, 2460 (2009). https://doi.org/10.1016/j.patcog.2009.03.001
8. Babič, M., Kokol, P., Guid, N., Panjan, P.: A new method for estimating the Hurst exponent
H for 3D objects. Materiali in Tehnologije 48, 203 (2014)
9. Babič, M.: Analizakaljenih materialov s pomočjo fraktalne geometrije. Doctoral dissertation
(2014)
10. Graves, A., Schmidhuber, J.: Offline handwriting recognition with multidimensional
recurrent neural networks. In: Bengio, Y., Schuurmans, D., Lafferty, J., Williams, C.K.I.,
Culotta, A. (eds.) Advances in Neural Information Processing Systems 22 (NIPS 22), 7th–
10th December 2009, Vancouver, BC, Neural Information Processing Systems (NIPS)
Foundation, pp. 545–552 (2009)
11. Koza, J.: Genetic Programming: On the Programming of Computers by Means of Natural
Selection. MIT Press, Cambridge (1992). ISBN 978-0262111706
12. Box, G.E.P.: Some theorems on quadratic forms applied in the study of analysis of variance
problems, I. effect of inequality of variance in the one-way classification. Ann. Math. Stat.
25, 290 (1954). https://doi.org/10.1214/aoms/1177728786
13. Ravi, V., Naveen, N., Pandey, M.: Hybrid classification and regression models via particle
swarm optimization auto associative neural network based nonlinear PCA. Int. J. Hybrid
Intell. Syst. 10, 137–149 (2013)
A Multifunctional Platform for Elders’
Assisting to Live Alone
Blerina Zanaj1 ✉ , Fatjon Shaba2, Majlinnda Belegu1, and Gerti Boshnjaku2

( )
1
Agricultural University of Tirana, 1029 Tirana, Albania
{bzanaj,mbelegu}@ubt.edu.al
2
New Media Communications, 1029 Tirana, Albania
{design,marketing}@newmedia.al
Abstract. Living well demands for many facilities in order to assist the people
suffering from different disease, especially for those who are unattended by a
nurse. The houses of elder people that prefer to live alone require the equipment
with such facilities that we think will help them to carry on with their life on their
own. According to the recent statistics the prevalence of dementia is increasing
as a trend with a sharp curve. Where the observations has shown that belonging
to this group are mostly the people of middle or greater age, so 25%–30% of people
of age 85 years old or older suffer from this disease. In order to make more enjoy‐
able their life we have thought to build a multifunctional platform that will assist
them during the day. Another purpose of this platform will be the notification of
their relatives mostly their children, so the platform will show them that they
parents or relatives are in good health. So they can be notified if something unusual
and bad happened to them and this will be achieved through a non invasive system.
The platform LAB (Living Alone Better) that we are proposing is a multifunc‐
tional application. The main features of it is to remember them to take their medi‐
cines, when it is time to have their meals and as well to notify them to attend the
visits to their doctors when it is scheduled beforehand, in the day and the hour
when the appointment is fixed. Beside it they can do some physical training by
following the instructions said by voice of a trainer. Through the interaction with
the platform they can interact as well with their friends, they can read or hear the
paper headlines in their own language being the news local or from a larger region
of interest from the continent of world wide. Also in case of any home incidence
they can dial the emergency number by just pressing a button and a registered
voice call with their data and geo-location will part toward the hospital as well a
sms which will be sent in automatic to their family members. In this sms is written
the information regarding the incidence when it happens in real time. By reading
their location the platform will be able to connect them to their nearest friends.
So they can hang out together to a near place of interest that can be: a green space,
a museum, a theatre, some bars or restaurant and the system could tell them about
the weather conditions and the forecast for their zone. Also it will show the
shortest path to follow or the nearest bus station or train that they can catch to
reach a destination. For their entertainment they can listen to the radio, or hear
some music from a predefined list, also some brain training games are imple‐
mented in the platform. This platform is designed to be very simple in use without
any complexity added, so it can help the elder people to use it and live better even

https://doi.org/10.1007/978-3-319-90893-9_29
A Multifunctional Platform for Elders’ Assisting to Live Alone 239
that those have preferred to live alone. Using such platform is a necessity also in
our country as the middle age is increasing.
Keywords: Application · Medicine · Taking care · Dementia
1 Introduction
We live in the era of information and the technology development has given us the
opportunity to think about new ways to entertain people. When we speak about enter‐
tainment is not suppose to mention only the games and the social networks and the virtual
world that those create. Entertainment should include all ages and its role is to keep the
mind busy, to help you communicate with people you know, to be informed about the
news all around the world. Different applications perform such tasks and those are
mainly used on single purpose. If we want to find the location of a place we go and
search on google maps. Then we speak with our friends through the chat of facebook,
in hangout, or we use whats up or viber. Then if we want to find out regarding the historic
of our training, we can keep track of the steps we walked, and the calories burned we
can check them by activating another soft on our mobile. There is also another kind of
applications those are embedded software on different gadgets that might measure you
blood pressure, the temperature of your body, the different parameter for those that suffer
from diabetes, etc. The number of applications is enormous and it is not our intention
to mention all of them in this work, but we wanted to show only some of them and to
emphasis their usage. These applications mentioned before are used mostly by young
to middle age. The interest for using them is less probable with the middle and older
people. They might find difficulties on downloading and registering on different social
network to mention just one problem. So we have thought why not to create a multi
functional platform. A platform where we combine the chat, the google map location,
some kind of agenda, and an sos application, all in one. This multifunctional platform
will help elders to organize their day a bit different. Due to the many functionalities it
will offer them different activities that might be suggested or planned beforehand in
collaboration with the physician and the physical trainer whose opinions regarding the
person that might use it and it will be helpful and very important. The agenda will hold
the schedule about the days where they have to meet their physician. And if their children
or other parents are required to accompany them some notification might as well start
and reach their mobile like a sms to remember them about this appointment. The agenda
will be as well part of this application.
2 System Description and Functioning
The idea of a multi functional platform were dictated by the necessity that some elder
people might need such a system in order to remember them what tasks they need to
perform during the day. There are many elder persons who prefer to live alone. But the
problem is that living alone for some person is a bit dangerous if they suffer from some
illness. When the illness make them forget the actions and activities they need to perform
240 B. Zanaj et al.
during day a system that might tell and help them to remember will be very helpful. This
platform will have different applications all included in a single stand lone computer
program. This program will have different interfaces or better to say voices where the
person that might use it needs to check and view. This application will hold a schedule
of the visit to the doctor this person should attend during week, or month. When it will
be the due time some kind of alarm will switch on and showing to him/her the time when
he/she needs to see her doctor. This schedule as well is organized about taking the
medicines and the meals during day. Besides the scheduling kept the application will
also read to the old person the news or it might switch on the radio to the channel where
the news are transmitting. The finding of the friends not only starting for the last job
position the person has held but also persons who attended courses and classes on the
same year in university or high school. But most importantly the system starting from
their location will show you who from those you know well and you are good friend too
are near to your location. So you can let and meet with them to hang out together. The
system might as well tell you about different amusements around and how you can reach
them. By telling you the nearest bus stop or train/metro station that brings you there.
Also some suggestions might come to help by showing the restaurants in that zone and
the kind of food they prepare there. Within the application we have thought to introduce
some physical training exercises that might be in the form of a video with the trainer
that shows you how you could do them. These kinds of exercises should be something
simple and that can be afforded by elders without any harm to their health. Another
important feature will be the starting of a call from the application by just touching a
button on the screen. This will be the case when the person is not feeling well and
he/she require the intervention of a doctor or a nurse in the hospital. So with the press
of the button a sms will start to reach the doctor but at the same time a sms will be sent
to the person that is in care of the elder person some parents of his/her. In this case the
alarm will reach to the care centre that has him/her as a patient and some doctor will
come in his/her help by an ambulance. The application will send with a sms the location
of the elder person. The same sms will go to the parent that is related to the old person.
So he/she who is in care of him/her will run to their house and help them together with
the doctor. Also within the application some brain training games might be included as
well. These games like chess or domino should be played with others that might be your
friends or someone else from another country that you might play with him/her from
internet with online games but without bets.
3 System Interaction with Other Web Applications
LAB platform is thought to be a web application. The interaction with different sites and
as well to Google maps and geo location system will be required in order to send with
a sms the location of the person in case if some incidence inside the house might happen.
Also one of the features of the system will be the creation of a social media where they
can find their friends and as well to play some games online with them. The location of
these other person will start the search not only on behalf of who they already know,
but mostly from those that live near to him/her. For the usage of such system we have
thought to build some menu and windows for the user interface. This interface should
be not too complicated and it might be very accurate in each detail it holds. Otherwise
it might create confusion in usage and the person will become very soon bored and will
abandon the system. The first view that they will find when they will switch on the
software will be as in Fig. 1. The first moment he/she is logged in it can start the inter‐
action with the platform by following the guide.
Fig. 1. System first window when the menu will be in the kind of a command button
So first you choose what you want to do and everything will be shown immediately
in a very simple way with just a click to choose between the options. So the different
functionalities of such application will be shown as in Fig. 2.
Fig. 2. All system features where are the options that bring you to another window or wizard
The different reminders that the application can held are for the meal consumption
or for the visit to the doctor. This reminder will switch on an alarm to remember the old
person to go to the doctor or take the medicines, meals. If the person does not switch it
242 B. Zanaj et al.
off it will continue to snooze till you switch it off, which means that the person have
performed already the action it tries to tell you to do. These remainders will be listed as
in the Fig. 3.
Fig. 3. Different reminders on the pills the person has to take during the day
Beside it the activation of the alarm when it is the date to go and make a visit to the
physician like in the Fig. 4.
Fig. 4. Doctor visit reminder as scheduled beforehand
The different alarm states are shown as in the Fig. 5.

Fig. 5. The different state of the same reminder depending on the action of the elder person
The leisure activities that the application has will be regarding the physical training
voice and video that will show them to perform some simple exercises to keep fit like
in the Fig. 6.
Fig. 6. Different exercises that the person might do to keep fit their body
244 B. Zanaj et al.
Also there might be included some games that might help to train their brain in order to
try to improve somehow their memory and to keep it under exercise helps like it is shown
in the Fig. 7.
Fig. 7. The different games that might be included within the application
The location of the friends will be another option that is offered by the geo location with
GPS, where the identification of their present location is shown in the display like in the Fig. 8.
Fig. 8. Finding friend location and starting of a call through one social media like what’s up or
Skype call
In the case of some incident that might happen to the old person, like a fall he/she
will be able to start a SOS voice message to the hospital and a sms will be sent to the
parents that take care of them in distance. This message will notify them with the inci‐
dence that might have happened. This other feature of the system is brought as in the
Fig. 9.
Fig. 9. A voice call will be sent to the hospital with the patient location and personal data already
saved and an sms will be sent to the person that is responsible for him/her
4 Conclusions
The multifunctional platform will be very handy to take care in distance for old person
suffering from Dementia. This application will make their everyday life more enjoyable
and they can be more active by meeting their friends, exercising by following the
instructions of a trainer, visiting some place of interest near to their apartment, etc. Also
it will be like an interactive agenda for their visits to the hospital or doctor, where the
agenda will start an alarm when it will be the date. This application is thought to be
coded for both as an application running on a mobile phone and in a PC as well. It will
require a Wi-Fi connection in order to find and update the street direction when the
person is going to visit some new place he/she has not been before. And the google map
introduced in the application will show the street view and direction updated if they
perform some path change during their walk. Similar applications to this one have been
already built and are in usage but mostly they perform single function and they are too
complicated for the old person. This one is far too simple to use with not to many menus
and options where you can get lost with no return back to the main menu. Other feature
246 B. Zanaj et al.
might as well be included but as this application is going to run on mobile phone and
on PC its code might be lightweight and simple in use.
References
1. Chan, M., Campo, E., Esteve, D., Fourniols, J.-Y.: Smart homes at current features and future
perspectives. Maturitas 64, 90–97 (2009)
2. Down, K.: Review of the Assisted Living Economic and Business Modelling projects (2014).
https://connect.innovateuk.org/documents/3301954/10546325/Review+of+EBM
+projectsfull.pdf/fe005c83-95db-4e43-810c-12508d27279f
3. Lin, H.-T.: Implementing smart homes with open source solutions. Int. J. Smart Home 7, 289–
296 (2013)
4. Greenhalgh, T., Wherton, J., Sugarhood, P., Hinder, S., Procter, R., Stones, R.: What matters
to older people with assisted living needs? A phenomenological analysis of the use and non-
use of telehealth and telecare. Soc. Sci. Med. 93, 86–94 (2013)
5. Gibson, G., Newton, L., Pritchard, G., Finch, T., Brittain, K., Robinson, L.: The provision of
assistive technology products and services for people with dementia in the United Kingdom,
Dementia (2014). https://doi.org/10.1177/1471301214532643
6. Tolkiehn, M., Atallah, L., Lo, B., Yang, G.-Z.: Direction sensitive fall detection using a triaxial
accelerometer and a barometric pressure sensor. In: Proceedings 33rd Annual International
Conference of the IEEE Engineering in Medicine and Biology Society (EMBC 2011), pp. 369–
372 (2011)
Economic Aspects of the Application of Cloud
Computing
Mirha Bičo Ćar ✉ , Savo Stupar, and Elvir Šahić

( )

{mirha.car,savo.stupar,elvir.sahic}@efsa.unsa.ba
Abstract. The Paradigm, on which cloud computing was built, as a new model
for using IT services, has been known since the beginning of computing, and was
created as a result of analogy with the way public utility services are used and
paid for: water, gas, electricity, telecommunication services, etc. However, the
practical realization of this idea was only possible after the emergence of the
Internet and the rapid development of a range of other information and commu‐
nication technologies. One of the root causes of both the development and the
progress of cloud computing is significant time savings, lower risks and fewer
barriers to the introduction of new applications, as well as significant cost savings
in the implementation of IT projects. The aim of this paper is to briefly introduce
some areas of cloud computing, then explain the key benefits and attractiveness
of cloud computing in relation to other types of computing, and ultimately
consider the economic aspects and issues that are related to it.
Keywords: Economics of cloud computing · On-demand computing

IT resource scalability · Comparative analysis of cost models
Application of cloud computing
1 Introduction
If we begin from the generally accepted definition of cloud computing [1] as a model
that provides a comprehensive, appropriate network approach to demand a common
pool of flexible computing resources (for example, networks, servers, warehouses,
applications and services) that can be quickly and reliably used with minimal manage‐
ment efforts or interaction of service providers, we see that it has 5 essential character‐
istics. One of the most important features of “On-demand self-service” on the principle
of “pay as you go” has provided cloud computing, as a new paradigm of computer
resources, the ability to first of all change the economy of using information technolo‐
gies. Due to the already mentioned principle, cloud computing changes the relationship
between capital costs (CAPEX) and operating costs (OPEX) [2]. If we want to analyze
the economic cost-effectiveness and sustainability of various models of information
technology implementation, it is necessary to get acquainted with their characteristics
as well as their requirements and costs that they cause when using.

https://doi.org/10.1007/978-3-319-90893-9_30
248 M. Bičo Ćar et al.
2 IT Management Models
There are 4 different models of IT implementation and management in companies to

which we assign an appropriate cost model. These models are [2]: Model of conven‐
tional/traditional internal IT management; Common Location Model (Colocation);
Managed Services model and Public Cloud Computing Model.
In a conventional internal IT model or “zero-outsource model”, all aspects of IT
applications or purchase service, implementation, and administration are performed
using internal resources. It is such a model in which the IT department of the company
independently buys, implements and manages equipment and software applications. The
most common form is the office IT infrastructure. In many offices, the internet connec‐
tion is provided by the ISP (Internet Service Provider) and is connected to the internal
network via a router. This internal network consists of firewall software, switching
devices, central servers, print servers, desktop computers and possibly wireless networks
and laptops. The internal IT department buys, installs and manages all the equipment,
as well as automation software for office work. Specialized business applications can
be managed the same way, using existing ones or customized software applications,
which are installed for this purpose on the company’s hardware. In this model, appli‐
cations for external users can be started, using e.g. corporate website.
In the common location model, companies implement and manage applications in
an independent data center of the bidders, otherwise known as colocation. The company
is responsible to purchase hardware or server, and to develop or purchase the software
to run the application. The colocation facility supplies data center with electricity,
cooling, space (metal shelves) for servers and provides network connection of hardware.
A co-location facility usually provides redundant (backup) network connectivity, an
alternative source of electricity and physical security. Co-location services are usually
purchased in the form of annual contracts with initial fee for services and monthly fees,
calculated on the basis of the size of the used space and the performed data traffic.
In the Managed services model, users entrust the execution and management of an
IT outsourcing to a service provider, but do not have the obligation to acquire and main‐
tain servers, network, or any other hardware. The service provider rents a complete IT
infrastructure and takes responsibility for hardware management and basic operating
system. In some cases, the service provider also rents standard software such as data‐
bases or the basic services of the database management system as part of its total rental
services. Similar to the collocation model, contracting with Managed Services model
providers usually includes minimum annual commitments, with an initial fee for imple‐
mentation, followed by a repeating monthly cost, the amount of which depends on the
hardware and software configuration that is being leased. In this model bandwidth is
usually not charged separately. Instead, the user is provided with a standard share based
on the number of server numbers that are contracted. In this model, some additional
services can be arranged, such as backups. These services are usually charged on a
monthly basis and are calculated based on the size of the memory space for which you
want to make a copy.
In the Cloud model, as in the Managed services model, the company dislocates or
relocates IT infrastructure, but in a completely different way. Instead of fixed hardware
Economic Aspects of the Application of Cloud Computing 249
resources, the company uses virtualized resources, which are dynamically allocated only
at the time when it is needed. An analogy can be made with the principle Just In Time
concept. Similarly, a dynamic allocation of resources to a cloud service allows the client
to use computer resources only when they need it. Servers do not need to have an empty
walk during non-working periods. The cloud payment service model is in line with this
type of usage profile, with the provision of services that often do not require upfront
charges, and the monthly charge is made, not in the form of a subscription, but based
on the actual amount of funds spent in that month. This can be considered a significant
cost advantage compared to traditional implementation models.
2.1 Areas and Reasons to Apply Cloud Computing

At the beginning of the development and implementation of cloud computing, the
emphasis was on the design of scalable web applications developed by companies with
limited IT infrastructure [3]. Well-known examples are the TimesMachine project,
which launched the New York Times [4] and Animoto video service [5], launched by
the same company. In both cases, public cloud services were used for the reduction of
barriers to entry in the work, which the two companies dealt with. It was precisely
scalable cloud computing infrastructure that enabled the rapid expansion of the business
of these two companies and was available at a very low price. Cloud services are often
not limited to specific applications. In principle, each application can work in the cloud
or combine with cloud services. The reasons for using cloud services can vary accord‐
ingly, so that they can range from one-off, temporary and sometimes periodic demand
and handling difficult predictions, sometimes to extremely high loads, overcoming the
seasonal effects of demand or outsourcing functionality and services to third parties in
whole. The use of cloud services can make sense both for testing purposes and for
production environments. The service provider may disguise the use of one or more
cloud services for end-users, such as Animoto companies [5], or these services may be
directly visible, as is the case with Office Web applications.
With dynamic use (scaling or reducing) of IT resources in cloud computing and
introducing a billing system that depends on the degree of use, companies are able to
transform their fixed to operational costs. “Pay as you go” billing models are based on
actual use of resources over a specific period and payment of compensation for use in
the same period. This is in contrast to the purchase or leasing model where the resources
are purchased or rented and paid in a specified period, which is defined by the contract,
regardless of their actual use. The advantage of using a billing based on real use is in
elasticity and in the fact that there is a small risk of overvaluation and insufficient
utilization (too much resources available, which are not sufficiently used) or underesti‐
mation and saturation (too little or insufficient resources available that jumps or falls in
the cue can not be overcome). When it comes to elasticity, then this term implies the
ability to add or remove a resource with a high degree of precision within a few minutes
(and not for a few weeks or even months), thus enabling better application of the actual
application requirements.
In the practice of conventional data centers, overestimation and insufficient utiliza‐
tion are the rule, and the optimal use of resources represents an almost unachievable
goal. If there is an unexpectedly high demand for resources, it can not be met immedi‐
ately, and vice versa, in case of an unexpected fall in demand, the degree of under-
utilization is increased. In doing so, it is often not possible to predict the resource demand
reliably: Many e-commerce applications face a very changing demand, which is not
conditioned solely by seasonal changes, but can be under the strong influence of certain,
temporary trends and effects. The financial consequences of such situations are impos‐
sible to ignore and need to be carefully considered.
In comparative cost calculations, the costs of real use of cloud supply are compared
with the cost of a data center, or the costs of an IT infrastructure need to be procured
and maintained. One very interesting and simple model of cloud service cost estimation
is suggested in the work of a group of authors from Berkeley University [6]. There are
costs of a cloud service, whose services theoretically do not have a capacity limit above
the cost of data centers, whose services always have a fixed capacity. The cost of the
data center besides the cost of hardware acquisition also includes operating costs, such
as computer power, computer cooling systems, real estate rental costs, as well as the
costs of administrative staff. In this context, electricity costs constitute the largest part
of Total Cost of Ownership (TCO) [7]. In principle, it can be concluded that more
efficient use of energy can be easily achieved in large data centers, and more difficult in
smaller ones. Owners of large data centers are in a more favorable position to make
significant savings by contracting lower energy prices or establishing data centers in
regions with lower energy costs. In addition, they benefit from a significantly lower
purchase price of hardware, so that the money saved here can be invested in failsafe
systems and data security systems. Microsoft’s study [8] accordingly identifies short‐
comings for small corporate data centers and predicts a long-term trend of multi-user
utilization of very large data centers. Today, building data centers of incredible dimen‐
sions is becoming a trend around the world. They physically occupy huge land surfaces,
and investments in establishing data centers are in the range of several hundred million
dollars. Global data centers comprise more than 100,000 servers, and thus can dramat‐
ically reduce TCO per server. According to this study, costs will not be paid for organ‐
izations with less than 100 servers. For organizations that work with about 1,000 servers,
the cost of using a private cloud will be ten times higher than those for the public
cloud [8].
2.2 Comparative Analysis of Different Models

From previous considerations, it can be clearly concluded that with different imple‐
mentation models, different types of costs and fees appear in pairs. For a clearer and
better comparison of all models, on a hypothetical example [2] we will show the amount
of costs according to different models. For example, if we take a small business appli‐
cation, which is composed as a three-layer system. The first layer is the web server, the
middle layer is the application server and the lowest layer is the database. Since users
in the conventional model need to implement support themselves in the form of backup,
we assume that for each hardware component one needs to acquire and implement
another one of such a component. Which means you should procure and install: 2 firewall
devices, 2 load-balancer devices and 6 servers. The cost of the software is not included
in the price because of additional complications in making a comparison, but in such

cases, usually open-source solution is installed. The price would be $ 31,000.
Internal IT Model: In the internal IT management model, capital costs prevail over
operating costs [2]. The capital cost in our case is the cost of hardware that is necessary
for the application to work, and totaling about $ 31,000. For the sake of more precise
comparisons with other models, we assume that the planned lifetime of the equipment
is 3 years and 36 months. When the total capital costs are transferred to the months of
the estimated usage period, we reach a sum of $ 861 per month. Of course, there should
also be added operating costs, such as electricity costs, administration costs, internet
usage charges, etc., but these costs are difficult to determine because they vary depending
on the usage intensity, the size of the company, etc. (Table 1, according to [2]).
Table 1. Comparative analysis of cost models of internal IT and model of collocation - according
to [2: pp. 53–55]
Elements Model of internal IT Model of collocation
6x Server $ 18,000 $ 18,000
2x Load-balancer-a $ 10,000 $ 10,000
2x Firewall $ 3,000 $ 3,000
Total capital costs (for 36 months) $ 31,000 $ 31,000
Monthly $ 861 $ 861
Bandwidth 10 M bit/s (per month) – $ 1,000
Total Monthly $ 861 $ 1,861
Common Location Model (Collocation) [2]: This model is much like a conventional
internal IT model, because it is necessary to acquire and install computers and equip‐
ment, but the essential difference is that they are located at the third parties. This, in
addition to the aforementioned capital expenditures, requires additional operating costs
due to the establishment of a network connection with servers located in data centers,
which can be significant expenditure on a monthly basis, e.g. for 10 Mbits per second,
server bandwidth should cost about $ 1,000. The monthly cost in the collocation model
is much higher than in the traditional model, and when added to the cost of renting space
in the data center, the figure increases further. The advantage of this model is the rela‐
tively easy establishment of additional infrastructure.
Managed Services [2]: The characteristic of the managed-services model is that it only
generates operating costs, excluding capital, compared to the previous two models. The
equipment is rented and the property does not switch to users at any time. If we return
to the previous example of a business application, this model requires the same infra‐
structure as in the previous two models. So, two firewall devices, two load-balancers
and six servers. When we talk about permeability, usually each server receives 500 GB
per month, which is quite enough for an average business application. An additional
cost in this model is the initialization, the installation of equipment, which is paid once
and usually amounts as much as one monthly rental of hardware. The cost difference
between the model of entrusted services and the previous two models is high, even more
than nine times in comparison to conventional v.s. internal model. The main advantage
is that it is not necessary to hire a workforce to administer infrastructure and resources,
which can be a significant saving item at the monthly level (Table 2, according to [2]).
Table 2. Analysis of cost model of managed IT service and cloud model, according to [2: pp.
55–58]
Elements of managed IT Managed IT Cloud Elements of cloud model
service model
6x Server $ 6,000 $ 759 6x Large Instances
2x Load-Balancer $ 1,500 $ 25 1x Load-Balancer
2x Firewall $ 300 $ 216 1x Firewall (or VPN)
Hardware costs (per m.) $ 7,800 $ 300 Memory ($ 300/month)
Initialisation ($ 7,800/36) $ 217 $ 1,300 Hardware & memory costs
– – $ 135 Bandwith (780 GB x $ 0,17/GB)
Total Monthly $ 8,017 $ 1,435 Total Monthly
Cloud Model [2]: Cloud computing, as the most current model, offers high flexibility
to users. The use of resources is based on rentals, as well as in the model of trusted
services. The basic difference between these two models is that resources are not dedi‐
cated to specific users, but are virtualized and shared. The utilization of resources in this
model is much higher, and accordingly, it has the space to create the appropriate price
classes that match the resources actually used. The way the resources are charged in the
cloud model is slightly different from that in the model of trusted services: the total cost
is formed according to the time of use of resources. Installation and usage of resources
is called an instance. Every time a user establishes a new instance, he gets a portion of
virtualized resources, and pays them as much as he uses them [2]. The cost of an instance
of computer resources per hour varies, depending on the allocated capacity and location
of the resources, e.g. the cheapest resources are in the east of the United States, while
they are the most expensive in the North-East Pacific. The average hourly price is some‐
what higher than $ 0.17, which when multiplied by an average number of hours per
month (730.48) is $ 126.53, and the total monthly cost for six server instance is $ 759.
Firewalls also provide the ability to establish Virtual Private Network access (exactly
specified way of moving data to a computer network), which is an additional advantage,
and is usually $ 0.30 per hour, which is $ 216 per month. Additional cost in the cloud
model is the storage capacity, which is paid extra, and for one business application of
the example mentioned it would be enough to have 2 terabytes of memory capacity,
which when added to the input-output traffic costs about $ 300.
Bandwidth is paid to the amount of transferred bytes at a given time interval, and
Amazon for a different quantity of bytes has different price ranges (more transferred
bytes imply a lower price per byte). So in a category of up to 10 Terabytes per month,
the price is $ 0.17 per Gigabyte, if we assume that for a business application from the
example we need 780 GB per month, we come at a cost of $ 135 per month for bandwidth.
The bandwidth is the cost of a load-balancing service, which is approximately $ 25
monthly. The total monthly cost, when it comes to the Cloud model, is $ 1.435, and this
is the second most affordable solution, after conventional internal IT. So, the Internal
IT model, assuming these assumptions, is the most favorable solution. It is very impor‐
tant to note that resource utilization is a key factor for the economic aspect. If the planned
period was reduced to 18 months then the monthly spending in the internal IT model
would be $ 1722, which is more than the cost in the cloud model. It can not be said that
there is one solution for all users, and therefore a detailed analysis of all factors and
needs should be done when individual users and companies are thinking about the intro‐
duction and construction of IT infrastructure, switching from conventional to another
model, or vice versa.
3 Conclusion
Cloud Computing is a new technology that realizes the idea that computer data
processing is executed and charged as communal services. One of the main reasons for
accepting and popularizing this concept is finding an acceptable solution to overcome
the problem of changing needs for computing resources. Whether the application of
cloud computing or some software applications pays out for a company is not the ques‐
tion that can be answered unequivocally, but it is necessary to perform the analysis for
every case. The answer will depend on the characteristics and goals of the specific
application.
References
1. Mell, P., Grance, T.: The NIST Definition of Cloud Computing, Recommendations of the
National Institute of Standards and Technology, National Institute of Standards and
Technology, Gaithersburg, MD 20899-8930 (2011). http://dx.doi.org/10.6028/NIST.SP.
800-145. Accessed 27 Jan 2018
2. Rosenberg, J., Mateos, A.: The Cloud at Your Service: The When, How, and Why of Enterprise
Cloud Computing. Manning Publications Co., Greenwich (2011)
3. Baun, C., Kunze, M., Nimis, J., Tai, S.: Cloud Computing - Web-Based Dynamic IT Services.
Springer, Heidelberg (2011)
4. Gottfrid, D.: The New York Times Archives + Amazon Web Services = TimesMachine, The
New York Times (2008). https://open.blogs.nytimes.com/2008/05/21/the-new-york-times-
archives-amazon-web-services-timesmachine/. Accessed 26 Jan 2018
5. Von Eicken, T.: Animoto’s Facebook Scale-Up, Right Scale (2008). https://
www.rightscale.com/blog/enterprise-cloud-strategies/animotos-facebook-scale. Accessed 27
Jan 2018
6. Armbrust, M., Fox, A., Griffith, R., Joseph, A.D., Katz, R.H., Konwinski, A., Lee, G., Patterson,
D.A., Rabkin, A., Stoica, I., Zaharia, M.: Above the clouds: a Berkeley view of cloud computing,
Technical report No. UCB/EECS-2009-28; EECS Department, University of California (2009).
https://www2.eecs.berkeley.edu/Pubs/TechRpts/2009/EECS-2009-28.html. Accessed 07 Feb 2018
7. Menadžment Centar Beograd: Total Cost of Ownership http://mcb.rs/recnik/tco-total-cost-
ownership-ukupni-troskovi-posedovanja/. Accessed 09 Feb 2018
8. Harms, R., Yamartino, M.: The Economics of the Cloud. Microsoft http://www.microsoft.com/
presspass/presskits/cloud/docs/The-Economics-of-the-Cloud.pdf. Accessed 09 Feb 2018
The Role of Hadoop Technology
in the Implementation of Big Data Concept
Savo Stupar ✉ , Mirha Bičo Ćar, and Elvir Šahić

( )

{savo.stupar,mirha.car,elvir.sahic}@efsa.unsa.ba
Abstract. Thanks to the incredible speed of the development of Information and

Communication Technologies, a huge amount of information is generated in the
world today that is increasing with every day. Since such a large amount of data
could not be stored and processed (in real or reasonable time) using conventional
methods of storing and processing data, a new paradigm of storage, processing
and managing of large quantities, variety and data processing speed called the
Big Data concept appeared. The aim of this paper is to try to explain the need and
significance of the Big Data concept, the changes in data manifestations from the
Big Data concept and the role of Hadoop technology as a whole and its compo‐
nents individually in the practical realization of this concept. Since the practical
realization of the Big Data concept would not be possible without the corre‐
sponding components of Hadoop technology, the paper will deal with the aspects
of economics and productivity of storage, processing and data analysis.
Keywords: Big Data · Hadoop · HDFS · MapReduce

Data processing productivity
1 Introduction
Every minute, Internet users send 204 million e-mails, generate 1.8 million likes on
Facebook, send 278,000 tweets, post up 416,000 images on Facebook and Instagram,
search 2.4 million terms on Google, etc. [1, 2]. All these huge amounts of data need to
be stored, processed, analyzed, presented and then suggesting concrete business solu‐
tions based on these results. Realization of these activities in a reasonable time and
obtaining very often unexpected and surprising conclusions enables the Big Data
concept. It is a concept that belongs to the field of artificial intelligence, although it does
not attempt to “teach” a computer to think as a human being [1: p. 11]. Technically, the
Big Data concept is based on the application of sophisticated statistical methods on huge
amounts of data in order to increase the accuracy or probability of the results obtained.
Big Data is a concept based on predictions, and the fact is that the system works well
because it uses large amounts of data, and with the passage of time and the use of new
data produces more and better results. In addition to the basic condition for the successful
operation of the system that relates to a very large amount of data, an additional condition
is required, which is a large processor power and memory space in which all these data
will be stored, as well as the special technologies for more efficient storage, access and
processing of these data, and later analysis and presentation of the obtained results

https://doi.org/10.1007/978-3-319-90893-9_31
The Role of Hadoop Technology 255
(Hadoop). Since these factors needed for its implementation were not sufficiently devel‐
oped or were too expensive a few years ago, Big Data technology appears only in these
times when these conditions can be provided without significant investments.
2 Big Data Concept
Big Data represents collecting, storage and analysis (processing) of such a large amount
of data, which exceeds the possibilities of the previously used software for storage,
processing and data management. Big Data is a term created in 2008. The practical
realization of this new concept was not possible until innovative storage technologies
were developed and a large number of data was processed in a fast and simple manner
(Hadoop and Spark technology). The main reasons for collecting and processing large
data sets according to Berman are [3]:
• Collecting data on individuals for investigative purposes (detective hypothesis)
• Collecting data on the population to control each member (Big Brother hypothesis)
• Gather information about the individuals to learn everything about the population
• Gathering information to make it a place to put all things together
• Big Data is a collection of data with a purpose of searching for specific items and
facts
• Collecting data for carrying out general scientific conclusions
• Big Data is a social archive that generates money
• Big Data does not exist as a separate venture; This is what we have always done, but
with more data and details “[3: p. 201]”.
If we were looking for a concept, which is similar to the Big Data concept, then it
would be the concept of Business Intelligence. There is almost no difference between
those two concepts, except in the amount and structure of the data that need to be
collected, stored, analyzed and processed into real time results. Hence, there are differ‐
ences in storage techniques and methods, processing and data management.
After such conclusions we can define Big Data as a technology that enables collecting
and processing of large amounts of structured and unstructured data in real time.
According to another definition, Big Data is made of all the data that could not be
managed or analyzed using standard tools and data analysis techniques [4]. And finally,
the definition which is most often used and which emphasize characteristics of large
datasets, which are the subject of collection, storage and analysis, is a definition, often
referred to as “3V” according to the initial letters: Volume, Velocity and Variety [5].
The “4V” concept is also often mentioned, where the 4th characteristic starts from the
assumption that the data themselves have a certain value [6]. Some authors, as the 4th
characteristic instead of Value, use Veracity-truthfulness. Other also add 5th and 6th
data Dimensions (6V) where the 6th dimension is the Viscosity, as the effectiveness.
Other dimensions of the Big Data concept are mentioned, such as the complexity of data
management since they come from a variety of sources, obscurity or uncertainty [7].
Finally, a definition that is an attempt to compile multiple definitions, which was
suggested by Ward and Barker [8] – “Big Data is a term that denotes the placement and
256 S. Stupar et al.
analysis of a large, complex data structure using a variety of techniques, including but
not limited to NoSQL, MapReduce and Machine Learning technologies” [8].
2.1 Hadoop: Definition and Its Tools

Hadoop is the open-source software framework of the Apache Foundation, the creation
and application of which enabled the first practical realization of the Big Data concept.
The theoretical concept of Hadoop is based on the possibility that large problems can
be broken into smaller parts so that analysis of smaller parts of the problem can be done
faster and cheaper [9]. Breaking these big problems into smaller parts (which can be
processed later in parallel) a huge savings in the processing speed are being achieved.
Hadoop is written in Java and is considered a computer environment based on HDFS
(Hadoop Distributed File System) as the basis for storing data and MapReduce, which
is an interface for processing large amounts of data stored in HDFS. These two concepts
represent the basic elements of Hadoop, but besides them, the environment is made up
of many other components. According to SAS Institute [10] the four basic modules in
the Hadoop platform are: Hadoop Common - libraries and tools used by other modules;
Hadoop Distributed File System (HDFS) - an expanding Java-based system that stores
data; MapReduce - a software model for parallel processing of large data sets; and
YARN (Yet Another Resource Negotiator) - a resource management platform. Other
components of the Hadoop ecosystem are shown in the Table 1 on the next page [11].
2.2 Hadoop’s Distributed File System - HDFS
HDFS is a Google file system that maximizes scalability. It is an integral part of Hadoop
and is suitable for storing large amounts of data. This file system is highly bug resistant
and is not hardware-demanding, which is a huge advantage over other file systems (it
can work even on very cheap home computers). In principle, as many new computers
are added to a cluster as much capacity increases and almost as much performance
increases which shortens the time needed for some query to be made.
Since it is intended for installation on servers (not on a desktop computers or laptops)
or more specifically for hundreds or even thousands of servers that have different
components and for which there is a probability of failure, the main goal and one of the
essential characteristics of the HDFS architecture is to quickly detect errors and to auto‐
matically remove them. Another important feature is that it is specifically designed to
work with large amounts of data of the order of 1 GB or more. HDFS is not designed to
interact with the user but is configured so that other tools can quickly and easily process
the data stored in it.
HDFS consists of NameNode and DataNode, which function according to the master
slave principle. Master nodes take care of where replicated data blocks are and maintains
the “health” of the system. Slave nodes manage blocks of data. It would mean that one
Hadoop cluster has one and only one NameNode and more DataNodes. Usually there
is one master server on which NameNode is installed, while DataNodes are installed on
other servers. NameNode is the central part of the HDFS that contains meta data about
cluster (where the file is), the data block or part of the file is located (location address).
Table 1. Components of the Hadoop ecosystem (according to [11]).
Function and description

Ambari Hadoop Starting, stopping, reconfiguring Hadoop services is very easy
Web-based Cluster using Ambari. It allows the system administrator to select any
support Management cluster server and to perform one of the actions listed on it
tool
Hadoop Ambari provides interface for monitoring key cluster indicators.

Cluster The system administrator can easily see how much unused
Tracking space has left in the cluster, what is the stat us of NameNode,
etc
Installation, Hadoop cluster can be installed using Ambari too. At the

adjustment beginning it is enough to install the Ambari server on just one
of Hadoop server, which does not even have to be in the cluster
Avro A Data Serialization System
Cassandra NoSQL distributed database system which provides the ability to select
scalability and availability or scalability and consistency of data
Chukwa A system for collecting data for managing large distributed systems
HBase A scalable database that suppo rts structured data storage for large tables
Hive Query Language, similar to SQL (Data Warehouse Infrastructure) that presents
data in the form of a summary table for summarizing and ad -hoc queries
Mahout Scalable machine learning and data interpretation libraries, searching

resemblance among data and data mining library
Pig High-level language for data flow control and execution environment for
parallel calculation. It can be said that it is a platform for manipulating data
stored in the HDFS system, containing a compiler for MapReduce programs and
a high programming language called Pig Latin. An integral part of this platform
are methods for extraction, tra nsformation and filling of data
Spark Fast and universal gene rator for calculating Hadoop data: provides a simple and
high-quality programming model that supports a wide range of applications,
including ETL, machine learning, word processing, and chart calculations
Tez Multiple Data Flow Programming Framework, built on Hadoop's YARN, which
provides a powerful and flexible generator for performing an arbitrary DAG
(direct acyclical graph) tasks for data processing for serial and interactive use
cases. Tez adopts Hive, Pig and other frames in the Hadoop's ecosystem, but
also with other commercial software (ETL tools) in order to replace Hadoop's
MapReduce as a basic execution mechanism
ZooKeeper High Performance Service for coordination of distributed applications:

maintains configuration information, naming, distributed synchronization, etc
Sqoop Data transfer interface between Hadoop environments and relational databases
The master server controls file access and manages the File System Namespace, which
supports the traditional hierarchy. The user can create a directory and file, store the file
in it, change their names, delete and move them. Files, which are divided into blocks,
are stored on other servers where DataNode is installed. DataNode serves for storing
and creating of blocks, deleting, copying and restoring deleted data. Therefore, Name‐
Node stores meta data about blocks and ensures that end users can see the file rather
than blocks that are not readable to them, while DataNode stores the data. This could
be simplified as if the NameNode contains block addresses, the name and number of
copies.
The HDFS envisages the possibility that blocks are entered multiple times, which
reduces the risk of losing data. The default number of copies is three, which means that
each block will be copied three times. It does not have to be on a single server, but on
all servers in the cluster where DataNode is installed. It is true that 1 TB of original data
will require 3 TB of space, which increases the price of the whole process, but if security
is a priority, this price is not big to pay. It can be set to increase the number of copies
so that additional security is provided. If something unintended happens with NameNode
(since it is a “single point of failure”) there is no way to return data because it is not
known which file belongs to which data block. To avoid such a case, Secondary Name‐
Node is introduced, which is located on another server and represents cluster metadata
backup.
2.3 MapReduce
One possible approach to solving the problem of processing large amounts of data, which
results in a significantly higher speed, and productivity in terms of a smaller engagement
of human labor and lower costs (spending less electrical energy, use of cheaper
computers, less engagement of human labor, etc.) of computer data processing can be
solved in the following way: Break large problem or job into smaller subproblems (by
the principle of “divide and conquer” or “mischief and rule”); Independent sub-tasks
are assigned to independent “workers” and resolved in parallel; and Combine inter‐
mediate results from each independent “worker” to obtain a unique output or end result.
“Workers” in the context of computer processing may be: Multi-core processors (cores);
Multiple processor computers (multiple processors) or more cluster machines.
This is a paradigm used in creating tools that will enable fast and inexpensive
processing of large amounts of data stored with the help of the HDFS system. If Hadoop
consisted only of HDFS and Common packets, that would be enough as a storage plat‐
form for large amounts of data. As it is not rational to just store the data (they should be
processed, analyzed and presented), Hadoop has a special component that just makes it
one of the most important, which is MapReduce. An important feature of the MapReduce
algorithm is that it is run in parallel and distribute. MapReduce is an API (Application
Programming Interface), a computer model that is very suitable for processing large
amounts of data. It is implemented in Java, but it is possible to write MapReduce
programs in almost any programming language using the Streaming API.
MapReduce includes an important job planning component - TaskTracker, respon‐
sible for selecting the server that starts the job, as well for scheduling multiple jobs on
a shared cluster. TaskTracker consults the NameNode to find the location of all the
servers needed to perform the work. Each server is “ordered” to analyze the code in
parallel. Since TaskTracker is located on the same computer as the DataNode and the
data being processed, this results in an increase in the processing speed (even 1tb in
1 min). MapReduce processing begins with a step - Input Split, which allows each block,
containing the files to be processed, to be divided into individual records. A code that
defines the job of a mapper, allows developers to write and run a program that works
directly on each DataNode server in a cluster. The code recognizes the data format that
is stored in all blocks and can perform algorithms for processing these data, e.g. number
of occurrences of one word in a text, pattern detection, facial recognition, etc. MapRe‐
duce deals with scalability [12] and it’s made out of two elements (and two words) that
are: (1) Map: apply operations to selected data items and provides direct outputs; and
(2) Reduce: unites the direct outputs from the Map-step and provides the final output.
Let’s consider how MapReduce works on the example text file [13].
The system operates by input viewed as key/value pair and the result of MapReduce
operation is a list of key/value pairs or list. There are also two optional phases in the
MapReduce program that can, and do not need to be implemented, which are Sort and
Shuffle. These two phases serve as data preparation after the Map phase for the Reduce
phase. Output Map phases are key/value pairs, and these are the input data for Sort and
Shuffle, whose output is also key/value pairs that are input to the Reduce phase whose
output is one or more key/value pairs. The Table 2 illustrates how MapReduce works
on the example of WordCount program. The program aims to count how many times a
word appears in the document. The key advantages of the MapReduce are: Parallel
processing and Place (of processing) of data. Therefore, “Share and Dominate” para‐
digm, helps us to process data more easily and quickly with multiple computers, so the
time required for data processing is significantly reduced. Then, instead of moving data
from the input to the processing unit, we move the data processing unit to the MapReduce
environment. At that point data is distributed between nodes, where each node processes
part of the data assigned to that node. This give us the following advantages: nodes do
not get overloaded, process is shortened and more economical.
Table 2. Example of counting the words using the MapReduce function [13].
If we need to count the words in the file (“Dean, Bean, Rean, Clean, Clean, Rean, Dean, Clean
and Bean”), we will find unique words and the number of occurrences of these words in a file
The whole job (counting words in a document) is divided into three simpler tasks, which will be
distributed for work on three nodes (Map Node)
We mark the words in each of the three folders and assign each word a fixed value “1”. This
value “1” is actually a single occurrence of a word in a folder so regardless of the fact that same
word can appear two or more times in a folder, each occurrence is assigned a fixed value of 1
A list of key/value pairs will now be created where the key is nothing but a single word, and the
value is the number of impressions. So for the first folder (Dean, Bean, Rean) we have 3
key/value pairs - Dean,1; Bean,1; Rean,1. The mapping process remains the same on all nodes,
so there are two other nodes: (Clean,1; Clean,1; Rean,1) and (Dean,1; Clean,1; Bean;1)
Next phase is splitting the list of pairs –the pairs are sorted and mixed so that all pairs with the
same key (the word) are sent to the corresponding reducers
Therefore, after the sorting and mixing phase, each reducer will have a unique key and a list of
values that correspond to that key. For example: Bean,(1.1); Clean,(1,1,1), etc
Each Reducer sums the values that are present in that list of values. As shown in the figure, the
reducer receives a list of values, e.g. (1.1) for the Bean key. After that, it collects the number of
those on the list itself and gives the final result as - Bean, 2
All key/value output pairs are then grouped (collected) and recorded in the output file
3 Conclusion
Big Data, whose basic task is to process enormous amounts of data that have a high
degree of heterogeneity in terms of sources, formats and structure, in the real or reason‐
able time, as a theoretical concept, has arisen much earlier than the conditions for its
practical application have been established. Exponential development of information
and communication technologies from the half of the twentieth century created the need
and simultaneously enabled the conditions for realization of the Big Data concept.
Regardless of the fact that there are other technologies for storing and manipulating large
amounts of data, Big Data is identified with the emergence of Hadoop technology. The
two basic elements of Hadoop technology are HDFS and MapReduce. HDFS is a special
type of operating system that allows fast and simple storage of huge amounts of data on
cheap hardware, while the MapReduce software enables fast processing of this infor‐
mation on the basis of subdivision and parallel execution of these sub-tasks. Thanks to
its performance, Big Data finds extensive use in all phases of complex production,
scientific research in all fields of science, business and even politics.
References
1. Mayer-Schönberger, V., Cukier, K.: Big Data: A Revolution that will Transform How we
Live, Work, and Think. Houghton Mifflin Harcourt, Boston (2013)
2. Manyika, J., Knobaluch, M.: Internet users send 204 million emails per minute. https://
mashable.com/2014/04/23/data-online-every-minute/#MBzXmZxc6gqB
3. Berman, J.J.: Principles of Great Data: Preparing Sharing and Analyzing Complex
Information, p. 201. Elsevier, Morgan Kaufman, Amsterdam (2013)
4. Manyika, J., Chui, M., Brown, B., Bughin, J., Dobbs, R., Roxburgh, C., Byers, A.H.: Big
Data: the next frontier for innovation, competition, and productivity. Report by McKinsey
Global Institute (2011). https://cdda.cs.stonybrook.edu/sites/default/files/cddafiles/
CDDA_University%20Consortium.pdf. Accessed 3 Dec 2017
5. Gantz, J., Reinsel, D.: Extracting Value from Chaos. IDC iView. Sponsored by EMC (2011).
http://www.emc.com/collateral/analyst-reports/idc-extracting-value-from-chaos-ar.pdf.
Accessed 3 Dec 2017
6. Zhang, L.: How structured data helps in big data analysis -expand patent data with linked
data cloud. Electrical Engineering and Computer Sciences, University of California, Berkeley
(2013). http://www.eecs.berkeley.edu/Pubs/TechRpts/2013/EECS-2013-96.pdf. Accessed
17 Nov 2015
7. Kocijan, K.: Big Data: how we got to the big data and where are they leading us.
Komunikacijski obrasci i informacijska znanost, Zagreb, IIS, pp. 37–62 (2014)
8. Ward, J.S., Barker, A.: Undefined by data: a survey of big data definitions. School of
Computer Science, University of St Andrews (2013)
9. Dean, J., Ghemawat, S.: MapReduce: simplified data processing on large clusters, In: Sixth
Symposium on Operating System Design and Implementation, OSDI 2004, San Francisco,
CA, December 2004
10. Cutting, D., Cafarella, M.: What is it and why does it matter? SAS Institute Inc, Hadoop
(2014). http://www.sas.com/en_us/insights/big-data/hadoop.html. Accessed 27 Jan 2018
11. ApacheTM Hadoop®: What is Appache Hadoop. Apache Hadoop Org. http://
hadoop.apache.org/. Accessed 27 Jan 2018
12. Perera, S., Gunarathne, T.: Hadoop MapReduce Cookbook. Packt Publishing, UK (2013)
13. Bakshi, A.: MapReduce tutorial – fundamentals of MapReduce with MapReduce example.
Edureca (2016). https://www.edureka.co/blog/mapreduce-tutorial/#mapreduce_word_
count_example. Accessed 27 Jan 2018
Cybernetization of Industrial Product-Service
Systems in Network Environment
Elvis Hozdić1 ✉ and Zoran Jurković2

( )
1
University of Ljubljana, 1000 Ljubljana, Slovenia
ehozdic@yahoo.com
2
University of Rijeka, 58000 Rijeka, Croatia
zoran.jurkovic@riteh.hr
Abstract. The beginning of the new millennium, from the point of view of
industrial production, is characterized by the rapid development of information
and communication technologies and the Internet on one side, as well as global‐
ization, uncertainty and ever-increasing demands of the contemporary market on
the other. The basic question that arises from this relationship is: How to balance
these two diametrically opposite “half”, which are interconnected at the same
time? The answer can be sought undoubtedly in the innovations of products,
production technologies and production management methods on the underlying
principles advocated by the new Industrial Revolution named in Industry 4.0. The
proposed paper is part of the concept of cybernetization of advanced production
systems developed in the framework of research work in the field of structuring
reference models of cyber-physical production systems. The focus is on showing
the cybernetic participants of modern industrial manufacturing and service
systems and their connection to the Internet of Service Network.
Keywords: Autonomous work system · Cybernetics · Cyber-physical systems

Product-Service System
1 Introduction
The term cybernetization refers to the application of scientific-theoretical knowledge of

cybernetics and methodological concepts of system engineering. Cybernetics as a
science directs production systems towards more efficient functioning and management
of the development of the production system as a complex, dynamic, network and
distributed system.
The concept of cybernetics derives from the Greek word kybernáô, which at the same
time means managing, controlling, steering. In contemporary scientific terminology, the
term cybernetics is introduced by Norbert Wiener in 1948 [1] through his book Cyber‐
netics or Control and Communication in the Animal and the Machine. Wiener explains
that under the term “organism” is also meant a combination of man-machine, and it
follows that cybernetics is a science of management in living beings, machines and their
combinations.
However, in today’s literature there are different definitions that point to the diversity
of the concept of cybernetics.

https://doi.org/10.1007/978-3-319-90893-9_32
Cybernetization of Industrial Product-Service Systems 263
According to Bertalanffy’s opinion, cybernetics is a management theory as part of

a general system theory. Basically, such a concept of cybernetics is the use of feedback
and information principles. In 1954, in his book The Human Use of Human Beings,
Cybernetics and Society, Norbert Wiener approached the cybernetics of a wider circle
of readers. Founded on these Wiener works, the cyberspace has become a major science
of cybernetic management, regardless of their physical form. Cybernetics has been
raised to a very high degree of abstraction where general principles and governance are
defined and set. Peklenik defines cybernetics from the point of view of system control
in real-time [2]. Under the term “control”, Peklenik means the process we are organizing
to operate on a “management object” with the effort to reach the object or the goals.
Product Cybernetics is oriented towards studying the principles and the legality of
the management of technical-technological production systems. Today, its focus on the
field of development and management in complex, distributed, complex and advanced
production systems is very strong, based on real physical and cybernetic elements.
With globalization, the competitiveness of production systems undoubtedly
increases, and the rapid development of science and information-communication tech‐
nologies (ICT) open up new opportunities and challenges to successfully manage
complex, distributed, network and advanced production systems. The global market
demands greater efficiency of the above-mentioned production systems and better
service support for such production systems. Global advances in the ICT and the Internet,
as well as global market demands, at the same time create new momentum for the
development of cybernetics itself, and its applications in the domain of modern produc‐
tion systems, their concepts and models. In the following section, the development of
production and service systems and their cybernetization will be presented. The cyber‐
netic view is given through the representation of reference models of industrial produc‐
tion and service systems involved in the Internet of service network.
2 Product-Service Systems
By reviewing relevant literature, it is concluded that today’s production systems are

complex. Complexity [3, 4], as a feature of advanced production systems, is largely
conditioned by the technological, economic and sociological nature of the system, and
the dynamic environment in which the system operates.
Starting from Peklenik’s cybernetic approach to production [2], the production
system can be defined as a complex technological-economic-social system involving
Subject – people, knowledge, production processes and associated machines and devices
for performing processes with the intent of transforming materials, energy, information
and knowledge in products and services that have a use and market value, with constant
dynamic disturbances, constraints and goals, and which operate from the environment
in which such a production system exists. The production process is performed through
the interactive actions of material, physical elements, such as machining, workpiece,
tools, transportation means, etc. as well as non-material, cybernetic elements such as
knowledge, information flows, cybernetised management and organizational structures.
264 E. Hozdić and Z. Jurković
Production processes in the cybernetic spirit model the Peklenik [2] and introduces
three levels in production systems: the manufacturing level, the managerial level and
the corporate level. The manufacturing level consists of a number of subsystems neces‐
sary for transforming an innovation or a prototype into a marketable product. At the
manufacturing level, processes, related to products and production systems, such as
construction, planning, control, machining, assembly, refurbishing and the like are
taking place. The managerial level is responsible for implementing the policy set up by
the board on the corporate level. At the managerial level, processes that support produc‐
tion at the production level, such as management, research and development, sales,
procurement and trading, take place. The corporate level set the business policy, related
to the type of products, the volume and product mix, management, financing, location,
profit targets, etc. At the corporate level, business strategies are formed and strategic
decisions are made in manufacturing companies.
The traditional view of the production system, as a sum of functions and activities,
because of the increasing complexity, needs to be replaced by a dynamic and integrated
concept with intertwined and interdependent functions. From this perspective, the High
Level Group defines advanced production systems as [5]: “Advanced production
systems incorporate production systems and related services, processes, devices and
equipment, including automation, robotics, metering systems, advanced information
processing, signal processing, and production management with high-performance
information and communication systems”.
The concept of service is found today in various fields of everyday life (healthcare,
financial counseling, car servicing, etc.) in the field of production that has emerged in
recent years in many papers, including papers [6, 7].
In paper [8], MacDonald and Payne defines service as an activity and highlights its main
features: ineffectiveness, timely execution and use, and inclusion of external elements in
service delivery. Tomiyama, within service engineering, defines the service as an activity
that changes the user service state [6]. In doing so, the activity is understood as the sequence
of certain actions or procedures in which the bidding and service user are involved.
At the end of the 20th and beginning of the 21st century, a concept that integrates
production and service systems into a unified whole – Product-Service Systems (PSS)
concept [9] is emerging. The PSS concept proposes the production of products supported
by the service to reduce the impact on the environment. By applying the concept of PSS
in the Meier industry domain and other authors, in the paper [10], they propose the
Industrial Product-Service System (IPS2) concept. IPS2 is a PSS type that is intended
for business entities. IPS2 is an industry product that is supported through service life‐
cycle throughout its life span. The industrial product is such a product that the industry
buys in order to create new value. In IPS2, the key element is providing business rela‐
tionships between participants, manufacturers of production equipment, service
providers and the social environment with the aim of creating an environmentally
friendly and socially acceptable solution [10].
The user of IPS2 and IPS2 network elements are most often found in different physical
relationships. The physical part of IPS2 is most commonly available to the user, which
means that for the proper functioning of the system, the ICT required by the main
provider of the service and the individual module of IPS2 provides remote support for
the product. This kind of support for industrial products, through the implementation of
the IPS2 concept, is presented in the paper [11]. Service support for processing processes
using ICT is presented to paper [12].
Linking products and service to generate new value is an advanced way of production
realized through the Service-oriented Manufacturing concept proposed by Zhao and
other authors in the work [13]. Unlike the production of standalone physical products
in service-oriented production, integration of products and service services is achieved
with the aim of increasing their value during different stages of life. If a particular service
is used as production support, such service is considered a production-oriented service.
The concept of production-oriented service was introduced by Schmenner in the work
[14]. As an example of production-oriented services Schmenner cites bank services,
education services, air transportation, hotel nights, recreation, lawyer services, health
services, counseling services and the like. Unlike Schmenner, product-oriented services
are in Huang and other authors [15] defined as service-supporting services.
For service support through the implementation of functionality in the Autonomous
Work System (AWS) [16], Zupančič and other authors in the work [17] propose a new
participant in network manufacturing systems that can offer their service services in the
form of different functionalities for each individual AWS. Such network participants
are [17] appointed Service Unit (SU). AWS and SU represent Manufacturing-oriented
Service Networks (MOSeN) The following are defined network manufacturing systems,
defined basic elementary work system in the spirit of cybernetization and its networking
in the Internet of Things (IoT) network, as well as the basic participants of Manufac‐
turing-oriented Service Networks integrated into the Internet of Service (IoS) network.
3 Cyber Participants of Product-Oriented Service Networks
Network Manufacturing Systems (NMS) are a network of related manufacturing systems

that strive to utilize the connectivity advantages and present themselves in the form of a
larger production system. For NMS it is important that the participants in such production
networks retain their autonomous decision-making, while at the same time co-ordination in
the production network contribute to the achievement of competitive advantage [18].
In the paper [16], Butala and Sluga defines the concept of production networks as: “The
production network is the basis for competitiveness, innovation, agility and adaptability and
enables affiliated partners to (1) shape a long-term business coalition, (2) develop mutual
understanding, (3) jointly respond to business opportunities, (4) participate in synergistic
effects, and (5) share information, knowledge, resources, competencies and risks.
3.1 Autonomous Work System as a Network Participant
The Autonomous Work System (AWS) was introduced in the concept of Adaptive
Distributed Manufacturing System (ADMS) in work [19], see Fig. 1. Autonomy of AWS
is reflected through its specific competence in terms of knowledge, experience and
resources. In work [20], authors define autonomy as the ability of the system to inde‐
pendently make decisions about itself, without external assertions, and perform their
own activities without external coercion.
Fig. 1. Autonomous work system based on the concept of cyber-physical production systems,
rearranged [16]
The key elements of AWS are: production operations, management and feeding, and
a monitoring system based on measurement of input and output values. Surveillance
enables real-time processes and operations to be realized, as well as finding significant
deviations and uncertainties, enabling high-quality decision-making in crunching and
managing systems, designing corrective measures to improve performance, and identi‐
fying the system’s performance at the network coordination level. The autonomous work
system is defined as a system with rounded technological functionality and an appro‐
priate management function. Technological functionality is basically built as the
Elementary Work System (EWS), which by definition [2] is the process, process
Fig. 2. Elementary work system based on the concept of cyber-physical production systems [21]
implementation device, and Subject (e.g. operator). The elementary work system in
cyberspace, as a participant of the Internet of Things, is shown in Fig. 2.
The physical working system is a functional unit consisting of a process, a process
implementation device (PID), a program logic controller (PLC) and a communication
interface through which the other two ESCPWS components are connected. The
communication interface, which links the physical and cyber component of ESCPWS,
has the role of mapping the state of the elementary work system, while the intelligent
communication interface has the role of linking the human component of ESCPWS to
the other two components of the new system.
The cybersystem is made up of agents through which control, supervision and control
of processes taking place in the physical system are performed. In the cyber world,
platforms for multi-agent systems are created that have the task of: the physical system
configurator, witch enables the reconfigurability of the physical system; planner and
coordinator of the physical system work activities; intelligent interfaces that make the
connection between cyber and social space on one side, and cyber and physical space
on the other.
According to Butali and Slugi [16], the basic functionality of AWS is the transfor‐
mation of inputs into output values. Performing such functions enables the AWS
elements, such as process knowledge, machining machines, tools, measuring devices,
subject, etc. The aforesaid elements AWS round out the functional competence of AWS,
which is required for the execution of the process. In doing so, AWS functionality is
limited to key functionalities, which AWS allows for a certain advantage over the
competition. In addition to key functionalities, AWS also needs other functionalities
that are not part of the system’s functionality, and are implemented in AWS (mainte‐
nance of machines, logistics, quality control, process parameter optimization, tool
management, and the like). To achieve such functionality, Zupančič and other authors,
in the paper [17], propose a concept of production-oriented service networks in which
the service unit has the role of network participants.
3.2 Service Unit as a Network Participant

The concept of a production-oriented service network suggests sharing certain func‐
tionalities within the AWS and SU networks. Functionality is realized through the real‐
ization of the service. The function diagram of the service unit is shown in the following
Fig. 3. Service unit creates new value with service delivery. The service is performed
through a service process that changes the particular property of the service facility. The
service process is performed by means of resources and Subject, which has the task of
controlling and evaluating the service process. The concept of support manufacturing
operations based on ADMS and MOSeN, is the conceptual framework for the network
of related AWS and SU for which the authors in the work [17] propose the concept of
Manufacturing-oriented Service Architecture (MOSA).
Fig. 3. Functional diagram of Service Unit, replaced by [17]
MOSA consists of four key network participants, which are divided into three levels:
(1) the level of support available to support the execution of operations in AWS, (2) the
level of operations in which AWS is executed in which individual production operations
are performed, and (3) cybernetic level, in which the Virtual Coordination Unit (VCU)
and the Virtual Competence Center (VCC) are developed in the work [19], and the
structures represented by AWS and SU in the cyber space. AWS in cyber space is
represented by its cyber work system (CWS), while SU represents its cyber service
system (CSS). Connecting AWS and SU to the cyber level is a continuation of the
Fig. 4. Cybernetised manufacturing-oriented service architecture

cybernetic structure of manufacturing systems in the cybernetic spirit. At the cyber level,
a cyber product-oriented service system is formed.
CWS, CSS, VCC and VCU in the cyber system are connected to the cyber manu‐
facturing-oriented service network (CMOSeN), which is integrated into the wider
Internet of Service, Fig. 4.
4 Conclusion
The paper presents an overview of the part of the research work on the theme of cyber‐
netization of production structures and the structuring of the model of cyber-physical
production systems. Through the content of the work, the expansion of the principles of
production and technical cybernetics has been expanded in the domain of the develop‐
ment of advanced production systems with the focus on autonomous work systems and
service units to support production operations. By means of cybernetization and network
connectivity, AWS and SU have been enabled to create network manufacturing service
systems connected to the Internet service network. By structuring the Internet of Service
Network, the key principles of vertical and horizontal integration in the domain of
production are outlined, highlighted by the European Production Platform - “Future
Fabrics” in response to the challenges of the new Industrial Revolution – Industry 4.0.
References
1. Wiener, N.: Cybernetics or Control and Communication in the Animal and the Machine. The
Massachusetts Institute of Technology, Cambridge (1948)
2. Peklenik, J.: Fertigungskybernetik, eine neue wissenschaftliche Dusziplin fur die
Produktionstechnik. Festvortrag anlasslich der Verleihung des Georg - Schlesinger Preises
1988 des Landes Berlin, Berlin (1988)
3. Peklenik, J.: Complexity in manufacturing systems. CIRP J. Manuf. Syst. 24, 17–25 (1995)
4. Suh, N.P.: Complexity in engineering. CIRP Ann. Manuf. Technol. 54(2), 46–63 (2005)
5. High Level Group: High Level Group on Key Enabling Technologies (2010)
6. Tomiyama, T.: Service engineering to intensify service contents in product life cycles. In:
Proceedings of the Second International Symposium on Environmentally Conscious Design
and Inverse Manufacturing, pp. 613–618 (2001)
7. Aurich, J.C., Mannweiler, C., Schweitzer, E.: How to design and offer services successfully.
CIRP J. Manuf. Sci. Technol. 2(3), 136–143 (2010)
8. MacDonald, M., Payne, A.: Marketing Plans for Service Businesses: A Complete.
Butterworth-Heinemann, Oxford (2006)
9. Goedkoop, M.J., van Halen, C.J.G., te Riele, H.R.M., Rommens, P.J.M.: Product Service
Systems, Ecological and Economic Basics, The Netherlands (1999)
10. Meier, H., Roy, R., Seliger, G.: Industrial Product-Service Systems—IPS2. CIRP Ann.
Manuf. Technol. 59(2), 607–627 (2010)
11. Mori, M., Fujishima, M., Komatsu, M., Zhao, B., Liu, Y.: Development of remote monitoring
and maintenance system for machine tools. CIRP Ann. Manuf. Technol. 57, 433–436 (2008)
12. Judež, Ž., Sluga, A., Butala, P., Kopa, J., Peklenik, J.: Toward a web-based machinability
data service, no. 3
13. Zhao, L., Wan, J., Jiang, P., Qin, Y.: Service design for product lifecycle in service-oriented
manufacturing. In: Intelligent Robotics and Applications. Lecture Notes in Computer
Science, pp. 733–742. Springer, Heidelberg (2008)
14. Schmenner, R.W.: How can service business survive and prosper? Sloan Manag. Rev. 27(3),
21–32 (1986)
15. Huang, G., et al.: Establishing production service system and information collaboration
platform for mold and die products. Int. J. Adv. Manuf. Technol. 52(9), 1149–1160 (2011)
16. Butala, P., Sluga, A.: Autonomous work systems in manufacturing networks. CIRP Ann.
Manuf. Technol. 55(1), 521–524 (2006)
17. Zupančič, R., Sluga, A., Butala, P.: A service network for the support of manufacturing
operations. Int. J. Comput. Integr. Manuf. 25(9), 790–803 (2012)
18. Westkämper, E., Hummel, V.: The Stuttgart Enterprise - integrated engineering of strategic
& operational functions. In: 38th CIRP ISMS (2005)
19. Sluga, A., Butala, P., Peklenik, J.: A conceptual framework for collaborative design and
operations of manufacturing work systems. CIRP Ann. Manuf. Technol. 54(1), 437–440
(2005)
20. Scholz-Reiter, B., Höhns, H.: Integrated software agents: enabling technology for
collaborative E-logistics and E-business. Int. J. Comput. Integr. Manuf. 16, 517–525 (2003)
21. Hozdić, E., Vrabič, R., Butala, P.: Conceptualization of socio-cyber-physical production
systems (unpublished manuscript). In: Research Work: Model of Cyber-Physical Production
System (2016)
Technology-Enhanced Systems in Idiopathic
Scoliosis 3D Diagnosis and Screening
Saša M. Ćuković1 ✉ , William R. Taylor1, and Ionuţ G. Ghionea2

( )
1
Department of Health Sciences and Technology, Institute for Biomechanics,
Swiss Federal Institute of Technology – ETH Zurich, Zurich, Switzerland
sasa.cukovic@hest.ethz.ch, bt@ethz.ch
2
Faculty of Engineering and Management of Technological Systems,
University Politehnica of Bucharest, Bucharest, Romania
ionut.ghionea@imst.pub.ro
Abstract. This paper presents some of the recently developed solutions for non-
invasive scoliosis assessment and diagnosis. Scoliosis is a complex 3D deformity
that affects younger population especially in early adolescence, usually without
a known reason. The main standard in traditional scoliosis evaluation is single or
biplanar radiography in frontal and sagittal plane of the patient and so-called Cobb
angle measured on the dominant deformity curve. Nowadays, it is recommended,
by many professional societies, to avoid intensive and often X-ray (ionizing)
rehearsals and monitoring due to cumulative harmful effect on patients. In that
course many research and development institutes and companies offer wide
variety solutions for non-invasive diagnosis. Many of these are based on an optical
3D digitizer and software necessary to extract or estimate the degree of deformity
and to generate its visualization. We also gave a contribution to this topic with
innovative optical method - 3D Scoliosis simulator.
Keywords: Scoliosis · Optical diagnosis · Non-invasive solutions

Radiation-free techniques
1 Introduction
Scoliosis is a very complex and multifactorial 3D deformity of the spine which affects
4-6% of the global population. In this area, particularly interesting cases are adolescents
with idiopathic scoliosis (AIS) developed without reasonable factors. Currently, the
most relevant insight in spinal deformity is one upright PA and/or sagittal X-Ray film.
As a gold standard for deformity evaluation clinician relay on a Cobb angle which is 2D
parameter of deformity, measured over planar patient’s X-ray film (Fig. 1a). Unfortu‐
nately, Cobb angle does not consider axial vertebral rotations and 3D dimensional nature
of the spinal structure. This parameter requests highly specialized observers who are
able to identify end vertebrae of the most significant segment of the deformity curve in
planar films. Other techniques include Adam’s bending test, scoliometer, visual assess‐
ment and qualitative and quantitative measurements, indexing of scoliosis, etc.

https://doi.org/10.1007/978-3-319-90893-9_33
272 S. M. Ćuković et al.
a) b) c)
Fig. 1. Conventional diagnosis: (a) Radiography, (b) Low-dose radiography system SterEOS®,
(c) CT
New computer-aided methods based on a 3D surface topography and analysis provide

noninvasive acquisition of the dorsal or trunk surface with preliminary estimation and
quantifying the degree of the deformity or postural insufficiency [1–6]. Due to high prices,
these techniques are not always available in clinical praxis and on the other side, require
meticulous verification with some “internal” modality to be finally employed in daily
routine. By comparing back 3D surfaces and belonging X-ray film, many authors claim
that there is a high correlation between curvature of the spine and external surface defor‐
mation [7–9]. This statement causes the controversy to those who stress that relation
between internal and external parameters of the deformity is not so clear or it is very low
[10]. Third group of authors forces combined approach. They believe that for full and
adequate description and estimation of spinal deformity it is necessary to use X-Ray films
for the preliminary diagnosis and then to do a follow-ups with a less harmful acquisition
methods, thus to reduce ionizing probes on patients [11–13].
2 Ionizing Techniques
Some deformities request double X-rays films, so called biplanar imaging in two normal
planes, to increase precision of the diagnosis [14]. Also, some classification schemes
are based on a banding posture and its X-ray radiographs to estimate the structural nature
of the deformity increasing the need for exposing the patient to the negative radiation.
Each exposure to ionizing radiation increases the risk for cancer development, and in
that course abovementioned procedures are restricted to the special cases and should be
avoided in follow-up deformity monitoring. In [15], it is showed that from all breast
cancer cases developed in 5573 women, 69% of them had some exposure to ionizing
techniques for spinal deformity diagnosis during adolescence.
Technology-Enhanced Systems in Idiopathic Scoliosis3D Diagnosis and Screening 273
Recently, there are some advances in ionizing imaging and low-dosing methods,
such as stereo-radiography device EOS® 2D/3D (BioSpace, Paris, France) which offers
fast biplanar scanning (180 × 45 cm, 10–25 s) and 3D vertebral and hip reconstruction
(Fig. 1b). This innovation allows clinicians to have better insight in spinal condition and
significant internal indicators of the deformity, but it is also appropriate for preoperative
planning and postoperative monitoring based on high-quality images [16, 17].
Despite many advantages compared to traditional radiography, this device is still
based on radiation principle and thus not acceptable for often application to same
patients. In some cases, scoliosis occurs due to the structural and morphological changes
on the isolated vertebral body, like congenital scoliosis, spinal tumors, discus hernia‐
tions, etc. In that cases MRI and CT techniques are recommended [18, 19], (Fig. 1c).
As a particularly important tendency in scoliosis screening is to reduce a number of
unnecessary radiographic investigations or total elimination of such methods from clin‐
ical praxis. To do so, new methods are developed to estimate deformity and to perform
complete screening based on the external shape, only.
3 Non-ionizing Techniques
Development of a new techniques that will decrease harmful effect on the exposed
patients in adolescence is the key direction in field of spinal deformity diagnosis and
have many potentials and clinical benefits [20]. There are many diagnostic systems
available on the market such as: ISIS optical system, AUSCAN - Automatic Scoliosis
Analyzer, ELITE optoelectronic system, FORMETRIC Diers optic system, SIPS - The
Spinal Image Processing System, QSIS - Quantec Spinal Image System, COMOT/
METOS topography system, etc., [1].
Rasterstereography is a photogrammetry method based on a spatial triangulation
principle and surface topography [21–24]. 3D acquisition and reconstruction begin by
projecting a structured-light or laser pattern on the back (dorsal) surface of the patient
in standing, neutral position. First solutions were optimized for point clouds containing
up to 25000 points, but with further software development, back shape of the patient is
reduced down to 5000 representative points. The main advantage of this method is
absolute non-invasiveness and applicability in many other fields in medicine. Thus, these
methods are useful for esthetic surgery, breast augmentation planning, maxillo-facial
reconstruction, exoskeleton and orthopedic devices design, etc.
In order to estimate deformity based on a digitalized 3D trunk or dorsal surface of
the patient it is necessary to recognize and identify some important anatomical land‐
marks on the skin, to estimate back symmetry line (BSL) and to precisely calculate
middle spinal line (MSL) and its 3D coordinates [25] (Fig. 2).
The basic, usually visible or palpable landmarks on the patient’s back surface are:
prominent point of the 7th cervical vertebra C7 (Vertebral Prominence – VP), location
of SIPS (DL/DR) points (lat. “Spina Iliaca Posterior Superior”), and sacrum point (SP)
[26]. Additional markers could be: Shoulder Left – SL, Shoulder Right – SR, Scapula
Left 1 – C1, Scapula Left 2 – C2, Scapula Left 3 – C3, Scapula Right 1 – R1, Scapula
Right 2 – R2, Scapula Right 3 – R3. BSL line is usually estimated after symmetry
a) b)
Fig. 2. Reconstructed 3D dorsal surface, (a) transversal profiles and back symmetry line (BSL),
blue (b) middle spinal line (MSL), black dotted
analysis of the back-transversal profiles [21] (Fig. 2a). The key line in non-invasive
approach is MSL, often described as a polynomial and smooth 3D curve which connects
centroids of vertebral bodies and intervertebral discs (Fig. 2b). In normal postures
without spinal deformity, the MSL line is vertical straight line that connects C7 and DM.
3D optical systems are usually verified with radiography scans and their application
is limited due to the missing correlation with radiographic parameters [10]. Many efforts
were involved in these systems improvements aimed to be applicable to the wide range
of patients (old/young, obese/skinny, male/female, whole trunk/dorsal, etc.), to be
resistant to soft tissue artifacts, posturing habits, muscle tonus, BMI and different spinal
deformities.
System ISIS. Integrated Shape Imaging System (ISIS) is a light-based scanner devel‐
oped by Oxford Orthopedic Engineering Centre (UK) [23, 27] and it requires labeling
prominent landmarks on the dorsal surface with black stickers before the scanning.
Thanks to optical profilometry and image processing, scanning time is significantly
reduced to 100 ms and measuring precision is increased to ± 1 mm [7] (Fig. 3).
Besides the curvature and height mapping, ISIS generates transversal profiles on the
dorsal surface from C7 to sacral point and they are basis to create polynomial curve that
represents back symmetry line (blue) (Fig. 3a and b). Besides frontal profile, clinicians
can see sagittal profile with similar parameters of deformity curvature, and so called
Quasi Cobb angle of the BSL. The main disadvantage of this system is the lack of 3D
model of the spine and estimation of internal deformity parameters.
Raster stereography system DIERS Formetric 3D/4D. System FORMETRIC III

3D/4D (Formetric 2, Diers International GmbH, Schlangenbad, Germany) provides
a) b) c)
Fig. 3. ISIS method (a) Curvature map – 3D reconstructed surface, (b) back symmetry line –
points extraction, (c) Quasi Cobb angle estimation
contactless photogrammetric 3D scanning of the dorsal surface using Video Raster

Stereography – VRS technique [26] (Fig. 4a). The most important advantage of this
system is automatic analysis and recognition of anatomical landmarks, estimation of the
MSL, BSL and vertebral rotations (Fig. 4b). By extensive comparison of 500 optical 3D
scans from this device and its corresponding X-ray films, it is proved that this device
offers reliable results and precise 3D reconstruction. This device also comes with addi‐
tional modules for functional, postural and gait analysis.
a) b) c)
Fig. 4. System FORMETRIC III 3D/4D (a) device, (b) BSL, MSL and anatomical landmarks,
(c) 3D virtual diagnosis
FORMETRIC III 3D/4D also estimates the shape of the middle spinal line and
generates a 3D model of the spine and that is another advantage of this system compared
to existing non-ionizing methods (Fig. 4a). Deviation of this line from the standardized
X-Ray films is up to 2 mm and thus, diagnosis of scoliosis is limited to 50º of primary
Cobb angles. This system allows 3D curvature analysis of BSL and MSL and generates
many external deformity indicators in the sagittal (kyphosis and lordosis) and in frontal
plane (scoliosis). Besides many benefits and noninvasive-based diagnosis, this system
does not provide internal indicators of deformity.
3.1 3D Scoliosis Simulator

3D Scoliosis simulator is our contribution to non-invasive, markerless techniques in
scoliosis monitoring and it is the main module of the ScolioMedIS information system
(Fig. 5). This system is optical solution for a 3D image acquisition and point-clouds
processing of AIS patients’ back surface. We applied Knowledge-Aided technologies
to simulate the 3D spinal shape based on the reconstructed dorsal surface and generate
important deformity indicators.
Fig. 5. 3D Scoliosis simulator – General scheme
3D visualization of the patient-specific deformity model is a result of a 3D registra‐

tion between generic spinal model developed in Mimics (Materialise, Belgium) and
CATIA v5 (Dassault Systèmes, France) and reconstructed dorsal surface. This method
analyses 3D model of the dorsal surface topography and extracts over 100 internal and
external deformity indicators, along with the 3D visualization of the spine. Report on a
deformity assessment contains 3D visualization of the patient-specific 3D model along
with visualization of the apex vertebra and ends curvature vertebrae (Fig. 6).
The main advantages of our solution compared to analyzed solutions are: over 100
deformity indicators (especially internal), visualization of end vertebrae which define
primary and secondary curve of the scoliosis, apex vertebra localization, Cobb and
SOSORT-recommended angles and vertebral rotations. We are performing its final
validation towards clinical implementation.
Fig. 6. Report and visualization - Patient-specific deformity model
4 Conclusion
Contactless methods based on the optical medium detects thousands of points in second,
they are not harm for the human health and thus suitable for medical applications.
Besides basic clinical benefit, such methods are suitable for fast deformity screening
and monitoring after applied therapy, without the need to set markers on the patient’s
skin, but also to analyze the data and correlation of external and internal deformity
indicators. Despite many reservations surrounding non-ionizing scoliosis screening, we
believe that described 3D optical methods and particularly our solution has a great
potential in accurate and reliable diagnosis of AIS deformities in daily clinical routine.
Acknowledgement. This research is supported by the Swiss Government Excellence grant (Ref.
2017.0024) approved by the Federal Commission for Scholarships for Foreign Students (FCS).
References
1. Devedžić, G., Ćuković, S., Luković, V., Milošević, D., Subburaj, K., Luković, T.:
ScolioMedIS: web-oriented information system for idiopathic scoliosis visualization and
monitoring. Comput. Methods Progr. Biomed. 108(2), 736–749 (2012)
2. Don, R., Capodaglio, P., Cimolin, V., Benedetti, M.G., D’Osualdo, F., Frigo, C., Vismara,
L., Negrini, S.: Instrumental measures of spinal function: is it worth? A state-of-the art from
a clinical perspective. Eur. J. Phys. Rehabil. Med. 48(2), 255–273 (2012)
3. Legaye, J.: Recent Advances in Scoliosis – 8. Three-Dimensional Assessment of the Scoliosis.
InTech, London (2012)
4. Gorton, G., Young, M., Masso, P.: Accuracy, reliability, and validity of a 3-dimensional
scanner for assessing torso shape in idiopathic scoliosis. Spine 37(11), 957–965 (2012)
5. Cheung, C.W., Law, S.Y., Zheng, Y.P.: Development of 3-D ultrasound system for
assessment of adolescent idiopathic scoliosis (AIS): and system validation. In: Conference
of the IEEE Engineering in Medicine and Biology Society (2013)
6. Treuillet, S., Lucas, Y., Crepin, G., Peuchot, B., Pichaud, J.C.: SYDESCO: a laser-video
scanner for 3D scoliosis evaluations. Stud. Health Technol. Inform. 88, 70–73 (2002)
7. Berryman, F., Pynsent, P., Fairbank, J., Disney, S.: A new system for measuring three-
dimensional back shape in scoliosis. Eur. Spine J. 17(5), 663–672 (2008)
8. Zubović, A., Davies, N., Berryman, F., Pynsent, P., Quraishi, N., Lavy, C., Bowden, G.,
Wilson-Macdonald, J., Fairbank, J.: New method of scoliosis deformity assessment: ISIS2
system. Stud. Health Technol. Inform. 140, 157–160 (2008)
9. Jaremko, J., Poncet, P., Ronsky, J., Harder, J., Dansereau, J., Labelle, H., Zernicke, R.:
Estimation of spinal deformity in scoliosis from torso surface cross sections. Spine 26(14),
1583–1591 (2001)
10. Petros, P., Theodoros, G., Angelos, K., Costas, A., Evangelos, D.: A review of the trunk
surface metrics used as scoliosis and other deformities evaluation indices. Scoliosis 5(12), 1–
20 (2010)
11. Sakka, S.A., Macindoe, S., Mehta, M.H.: Correlation of the quantec scanner measurements
with x-ray measurements in scoliosis. Res. Spinal Deform. 1(37), 313–317 (1996)
12. Xue-Cheng, L., Tassone, J., Thometz, J., Paulsen, L., Lyon, R., Marquez-Barrientos, C.,
Tarima, S., Paul, R.: Development of a 3-dimensional back contour imaging system for
monitoring scoliosis progression in children. Spine Deform. 1(2), 102–107 (2013)
13. Liu, X.C., Thometz, J.G., Tassone, J.C., Paulsen, L.C., Lyon, R.M.: Historical review and
experience with the use of surface topographic systems in children with idiopathic scoliosis.
OA Musculoskelet. Med. 1(1), 1–8 (2013)
14. Labelle, H., Aubin, C., Jackson, R., Lenke, L., Newton, P., Parent, S.: Seeing the spine in 3D:
how will it change what we do? J. Pediatr. Orthop. 31(1), 37–45 (2011)
15. Ronckers, C.M., Land, C.E., Miller, J.S., Stovall, M., Lonstein, J.E., Doody, M.M.: Cancer
mortality among women frequently exposed to radiographic exams for spinal disorders.
Radiat. Res. 174(1), 83–90 (2010)
16. 09 February 2018. http://www.eos-imaging.com/
17. Wade, R., Yang, H., McKenna, C., Faria, R., Gummerson, N., Woolacott, N.: A systematic
review of the clinical effectiveness of EOS 2D/3D X-ray imaging system. Eur. Spine J. 22,
296–304 (2013)
18. Birkfellner, W.: Applied Medical Image Processing: A Basic Course. CRC Press, Vienna
(2010)
19. Preim, B., Bartz, D.: Visualization in Medicine: Theory, Algorithms, and Application.
Elsevier, Morgan Kaufmann Publishers, Burlington (2007)
20. Huysmans, T., Moens, P., Van Audekercke, R.: An active shape model for the reconstruction
of scoliotic deformities from back shape data. Clin. Biomech. 20, 813–821 (2005)
21. Drerup, B., Hierholzer, E.: Back shape measurement using video rasterstereography and
three-dimensional reconstruction of spinal shape. Clin. Biomech. 9, 28–36 (1994)
22. Burkhard, D.: Rasterstereographic measurement of scoliotic deformity. Scoliosis 9(22), 1–
18 (2014)
23. Dangerfield, P.: Research into Spinal Deformities 6. IOS Press, Amsterdam (2008)
24. Aubin, C.E., Stokes, I.A.F., Labelle, H., Moreau, A.: Research into Spinal Deformities 7.
Studies in Health Technology and Informatics, vol. 158. IOS Press, Amsterdam (2010)
25. Turner-Smith, A., Harris, D., Houghton, G., Jefferson, R.: A method for analysis of back
shape in scoliosis. J. Biomech. 21(6), 497–509 (1988)
26. 04 March 2018. http://www.diers.de
27. Tanguy, A., Peuchot, B.: Research into Spinal Deformities 3, vol. 88. IOS Press, Amsterdam
(2002)
Contributions to Improve the Sustainability
in Services Based Organizations
Mihail Aurel Titu1 ✉ , Bianca Alina Pop2, and Stefan Titu3

( )
1
Lucian Blaga University of Sibiu, 10, Victoriei Street, Sibiu, Romania
mihail.titu@ulbsibiu.ro
2
SC TEHNOCAD SA, 72, Vasile Alecsandri Street, Baia Mare, Romania
bianca.bontiu@gmail.com
3
The Oncology Institute “Prof. dr. Ion Chiricuță” Cluj Napoca,
34-36 Republicii Street, Cluj-Napoca, Romania
stefan.titu@ymail.com
Abstract. The sustainability of the services based organization consist on its

ability to monitor the external environment about opportunities, changes, trends
and risks, as well as the ability to learn, change and innovate. To reach the
sustainability threshold, the organization must focus on both results and
processes. A viable solution is to adopt a quality strategy that allows a perceptible
perception of its products to its customers. This research purpose is to identify
the extent of which organizations are sustainable and seek to meet the interests
of all the involved stakeholders, as well as to help the organizations to identify
areas where they can improve their performance.
Keywords: Quality · Quality management · Efficiency · Sustainability

Organization
1 Introduction
The services based organizations need to set some performance indicators which support
a long-term development strategies and organizational changes to ensure a resource-
efficient management, to streamline the processes, and show a continuous improvement.
Consistently data collection of the organization processes provides information to
evaluate, control and improve the processes, products and services to meet both - busi‐
ness objectives and customer preferences, in a dynamic change.
The economic agents’ activity on the service industry differs in many ways to the
carried out activities in the industrial enterprises that are material goods producers.
Therefore, it has been found that the profit, the performance, the organizational
structure and the quality perceiving of the service providers will differ in many respects
from those of producing enterprises. A closer analysis of the contemporary activity on
the economic sectors, leads inherently to the conclusion that services are an essential
component of production and distribution with significant cost considerations. The
services economy is not opposed to industry, but has become a more advanced phase of

https://doi.org/10.1007/978-3-319-90893-9_34
280 M. A. Titu et al.
national economy development where the goods and services demand is mutually inte‐
grated into value chains.
The services purpose is to stimulate the gained benefit resulted by the consumers
satisfaction. In this way the services can no longer be separated from the constitutive
set of the goods supply and is no longer a simple “substitute” for consumer satisfaction.
Broadly speaking, services are distinguished by the major characteristics, detailed in
Fig. 1.
Fig. 1. The services characteristics
Products or services have a higher or a lower life expectancy, they are physically
and morally exploited. Therefore, they need to constantly adapt to the customers’
requirements, even if this process requires new financial resources, new business strat‐
egies, other structures or other people. None of products is safe from the aging process
and if the organization does not protect itself against this threat, competition will accel‐
erate its disappearance.
Customers who are satisfied with a service are loyal, but if they are not satisfied, they
focus their attention on competition.
Creating and maintaining quality in the organization is dependent on the systematic
approach of the quality management, in order to ensure that the customer needs are
understood and satisfied. Achieving the quality requires commitment to the quality
principles to all the organization levels, as well as a continuous improvement analysis
of the quality management system [1].
The company managers are aware that the company’s only sustainable competitive
advantage is the ability to learn faster than other competitor and to turn companies into
organizations focused on learning and people investing.
2 Modern Guidelines for Quality Definition
Quality is a concept that is used in all economic and social areas but which has a subjec‐
tive character and has a particular significance for specific sectors, functions or
objects [2].
Contributions to Improve the Sustainability in Services Based Organizations 281
There are many definitions of the “quality” term, which express a different meaning
for different areas, formulated by different quality experts [3].
Dr. Joseph M. Juran [4–6] - a reference name for the quality world history and one
of the “quality management” parents, has indicated the transition from passive quality
control to quality management; he highlighted the importance of quality planning and
grounded the idea that quality implementation should be the result of a collective effort.
He has demonstrated how quality planning affects the different levels of an enterprise
activity and highlight the importance of the “quality trilogy”, presented in Fig. 2, which
makes possible to implement the quality strategic planning across the enterprise.
Fig. 2. The quality trilogy
The Total Quality Control (TQC) was transferred from TQC to Total Quality
Management TQM. Kaori Ishikawa discussed the content of total quality control in
Japan [7], which is different from the total quality idea in Western countries.
Following K. Ishikawa’s explanation, the concept means “Company Wide Quality
Control” (CWQC), which involves all employees, from top management to quality
control workers. Ishikawa pleaded for using statistical quality control methods to
promote quality control by learning the statistical tools. Ishikawa has created one of
these tools - the cause-effect diagram, which is also known as the Ishikawa
diagram [8].
According to Crosby [9] “quality is free”, the non-quality have cost. He advocated
a quality-based approach to quality, considering quality as “compliance with require‐
ments.”
This expert argued that “zero-defect” production is a feasible objective and consid‐
ered that in order to achieve “zero defects” the path is the prevention techniques
improvement.
His opinion on quality was expressed through four basic principles of quality
management, detailed in Fig. 3.
W. Edwards Deming states that the “good quality means a predictable uniformity
degree and dependability with a market-appropriate quality” [10].
Webster’s Online Dictionary presents the following definitions for the quality term:
– An essential and distinctive attribute of something or someone;
– Excellence or value degree;
– A characteristic property that defines the seemingly individual nature of something;
– A superior social state, “a good man”.
Fig. 3. The quality management principles
However, the more commonly accepted definitions are:

– “appropriate for use”;
– “compliance with the requirements”;
“an entity characteristics set that gives the ability to meet expressed or implied needs”
[11]; “the extent to which an assembly of intrinsic characteristics meets the require‐
ments”; according to this definition, the quality is not expressed by a single characteristic
but by a set of characteristics.
3 The Services Quality Perceived by the Client in the Current

Economic Context
The service enterprises activities differ in many ways to the activities that are carried
out in enterprises that are producing material products, hence the quality of the perceived
services is more difficult.
The service is “the activities generated result to the vendor and customer interface
and through internal activities of the vendor to meet the customer requirements”.
Service provision is a person’s or process activity that is geared to meet the customer
requirements (external or internal). Internal customers are those from their own service
organization.
Providing a service may involve the following aspects:
– a material product activity provided by the customer;
– a non-material product activity provided by the client;
– an immaterial product delivering;
– the customer environment creation.
The quality strategy is a winning strategy. By realizing it, the whole society wins:
consumers, producers, managers [12].
A well-served service represents a profitable strategy for the company and also offers
greater consumers satisfaction. A quality service results in attracting new customers,
increasing business volume with current customers, and losing fewer clients.
Quality is the main element that creates loyal consumers, satisfied consumers that
have chosen the company after having experienced their services, consumers who will
call in the future the company services and that they will recommend it to others.
According to the ISO 9001:2015, 7 principles have to be used by the organization
top management who wants to achieve a superior performance - Fig. 4.
Fig. 4. Quality management principles according to ISO 9001:2015
According to Jacques Horovitz (by his book “La qualite de service a la conquete du
client”), the essential principles on which the service quality is based, can be synthesized
as follows [13]:
– only the client is the judge of the service quality;
– only the client determines the service “excellent’’ level, level he always wants;
– a “promise” must be clearly grounded by the organization to allow to satisfy the
customer’s wishes by earning money and differentiating itself from competition;
– nothing prevents the promise from turning into clear quality standards;
– the organization must “manage” the client’s expectations;
– in the service matter there is no middle way, it must be perfect, zero defect must be
achieved; is required to be applied a constant effort to defects or mistakes elimination;
the clients must remain the evaluation barometer of the company’s products and of
those belonging to the competition.
4 The Management System Components to Improve

the Sustainability in the Service Based Organizations
Starting with the above-mentioned principles of ISO 9001:2015 quality management,

as well as the PDCA cycle principles that can be applied to all the processes and to the
quality management system as a whole detailed in Fig. 5, the key components of a
management system of the sustainability improvement in services based organizations,
were identified.
Fig. 5. The PDCA cycle applied in QMS [14]
The conceptual model contains four components that interrelate each other - Fig. 6.
These components are: strategic planning and deployment; process management; results
evaluation and analysis, i.e. learning, improvement and innovation.
They also interact with the organizational context and improve sustainability.
The model is based on the fact that the quality management system requires a
systemic approach. The first construct correlates the organizational context, including
opportunities, changes, trends and risks with strategic planning and strategic deploy‐
ment, process definition, and resource provision. In the second and third constructs,
process management, evaluation and results analysis are specified as key factors of
learning, improvement and innovation. In the fourth construct, innovation and change,
improvement and corrective actions are considered decisive factors in assuring the
sustainability improvement.
Fig. 6. Conceptual model: management system for sustainability improving
5 Increasing Performance Through the Integrated Quality

Assurance Concepts Application
The concerns about ensuring the goods quality have deep roots in history and have
evolved over the centuries through our attempt to “master the quality.” Insurance modes
have experienced several evolution stages:
– quality assurance through the final product control, the simplest form of the 20th
century beginning;
– quality assurance by statistical methods – in the 1950s, when the technical issues is
the priority;
– quality assurance by staff motivation - corresponding to the ‘60s and’ 70s, when the
focus was on the teamwork, implementation of zero defect programs and quality
circles;
– integrative quality assurance concepts that have been applied since the 1980s: Total
Quality Control - Feigenbaum - USA, Company Wide Quality Control - Ishikawa -
Japan, Total Quality Management – Juran;
– the integrating concepts aim to achieve the enterprise success by meeting its
customers’ requirements;
Their integrity is given by the following features:
– quality ensuring and improving, involves the participation of all enterprise depart‐
ments;
– all staff involving in the quality assurance and improvement process;
– quality must be tracked and ensured at all product life cycle stages;
– company’s agents’ involvement in the quality assurance and improvement process
(suppliers, intermediaries, customers, service units, etc.).
The Japanese quality system is a humanistic one and requires the efforts focus on all
employees awareness through education and training in quality system terms.
The American theory sets a quality limited level, which can be improved with
substantial additional costs that the customer will not be able to bear.
Japanese theory does not accept defects and accountability to the operator with the
performed activities inspection, the operator being able to stop the manufacturing line
when he has found a defect or a deviation from the working instructions. So it was
demonstrated that the American system is inferior to the Japanese system.
The application of the quality-oriented strategy in the company, has the conse‐
quences presented in Fig. 7.
Fig. 7. The quality oriented strategy advantages
Improving the company’s reputation - the company’s reputation is created by the

quality of the offered products and services. It is gained over time by a concern to offer
customers the products they want at predetermined standards.
The product market developing – the quality increasing efforts and product perform‐
ance improvement lead to increased market sales, even if the prices are increased.
Cost reduction - by the additional costs eliminating that are by the non-quality
generated. This has been demonstrated by Ph. Crosby in “Quality is Free” [9], in which
he analyzed the quality costs in the slightest detail and clearly showed that non-quality
generates additional costs that the enterprise will incur from its own revenue.
Compliance with the firm’s commitments - in 1940 G. Taguchi presented a produc‐
tion quality system that demonstrates that the process robustness lack leads to insufficient
performance to satisfy the customer [15].
The standards deviations requires corrections, tests, certifications, due to the product
robustness lack. On a system level the robustness lack leads to the obtained performance
degradation and implicitly to the products quality decreasing, with registered financial
losses, most of the time, after the product reaches to the customer.
Taguchi notes that the degradation of the industrial system’s performance is caused
by three types of disturbances that affect the product quality: on-line control, production
process disturbances, and process tolerances [15].
The quality considered singular is not enough to ensure the competitiveness, it must
be integrated into the enterprise strategy, taking into account its environment and the
competitive situation in which it operates. In other words, the quality strategic impor‐
tance for the company success is canceled if it is not integrated into the company’s
overall strategy.
The quality assurance policy integration into the enterprise’s overall strategy is a
Total Quality Management (TQM) characteristic.
A definition of TQM is: “The set of properties and management features of an
organization that gives it the ability to meet the demands of external and internal quality,
seeking all the competitive advantages” [16].
6 Conclusion
The expressed quality interest by the top management of many organizations is due to
an external threat generated by actual or potential market losses due to the non-quality.
That is why quality is regarded as a strategic contribution that cannot be neglected to
achieve competitiveness.
In this research paper, the following aspects were highlighted:
– Profiling and analyzing the quality in a services based organization as a fundamental
area of economic and social activities. In this respect, it was considered important
the quality scientific treatment of the services by addressing the evolution,
contouring, defining and characterizing as an economic and social activity, with the
presentation of the main implications.
– The systematically clarifying of the service quality importance in the economy. In
this respect, the elements related to the evolution, the quality definition and classifi‐
cation, were clarified.
– The relation between the customer orientation and economic and financial results
that contribute to improve the management service, were identified.
– Through the new concepts and knowledge, increasing the processes and services
quality in industrial organization, contributes to a much better company development,
providing a clear and predictable development framework backed by a proper
employee training.
– Within this strategy an important role is played by the quality that represents a
strategy of the company products/services differentiating, and on the other hand,
represent a way by which it creates its competitive advantages. Ensuring the long-
term companies competitiveness in national and international markets, can be
achieved through the application of quality-oriented strategies at their level, with the
client’s central focus - with his expectations, wishes and preferences.
The advantages of a quality strategy implementation to improve the competitiveness,
are:
– The company’s products quality contributes to improve its image and notoriety.
– High-quality products have the ability to expand the company’s market.
– Product quality entails better prices and, implicitly, higher profits, which in turn can
be reinvested in R&D activities (improving and permanently adapting product/
service characteristics to the customer requirements, new products designing, manu‐
facturing methods and technologies improvement, etc.).
– A quality-oriented strategy implementation does not necessarily entail additional
costs for the enterprise – the solutions can be identified on quality improving while
maintaining or even reducing the costs by attracting the employees support in this
direction.
It is considered that the work quality represent the main difficulty that arises is
attracting and motivating all the company employees in the quality assuring process of
the finished products.
Managers and employees of the services companies - should have an interest in
offering quality services at the lowest prices on the market. But in order to the market
quality services, the managers and employees should: want the quality, to be able and
to know how to achieve the quality, to evaluate and improve it permanently.
References
1. Olaru, M.: Managementul calității. Editura Economica București (2013)

2. Enătescu, A.M., Enătescu, M.: Calitate. Terminologie comentată. Ed. Tehnică, București
(2000)
3. Encyclopedia Britannica: Encyclopǣdia Britannica Online, “Quality” (2009)
4. Juran, J.M., Gryna Jr., F.M.: Calitatea produselor. Tratat de planificare, proiectare, realizare
și control. Tradus din limba engleză. Editura Tehnică, București (2012)
5. Juran, J.: La qualité dans les services. AFNOR Gestion (1994)

6. Juran, J.M.: Quality Control Handbook. McGraw-Hill, New York (1988)
7. Wilkinson, A., Marchington, M., Dale, B.: Manufacturing more effective TQM: implications
for the management of human resources. Res. Pract. Hum. Resour. Manage. 2(1), 69–88
(1994)
8. Ishikawa, K.: What Is Total Quality Control? The Japanese Way. Prentice-Hall International,
London (1985). Transl. David Lu
9. Crosby, P.B.: Quality Is Free. McGraw-Hill, New York (1979)
10. Deming, W.: The New Economics for Industry, Government, and Education, p. 132. MIT
Press, Boston (1993). ISBN0262541165
11. SR ISO 8402:1995: Managementul și asigurarea calității. Vocabular (1995)
12. Cronin, J.J., Taylor, S.A.: SERVPERF versus SERVQUAL: reconciling performance-based
and perceptions-minus-expectations measurement of service quality. J. Mark. 58(1), 125–131
(1994)
13. Horovitz, J.: La qualite de service a la conquete du client, Edité par intereditions (1987). ISBN
10: 2729601961/ISBN 13: 9782729601966
14. BS EN ISO 9001:2015: Quality management systems. Requirements (2015)
15. Ealey, L.A.: Quality by design: Taguchi methods and U.S. Industry. Dearborn, Mich. ASI
Press (1988)
16. Madu, C.N.: Handbook of Total Quality Management. Springer Science & Business Media,
6 December 2012 (2012). ISBN 978-1-4613-7409-1. Technology &Engineering. https://
www.google.ro/search?tbo=p&tbm=bks&q=subject:%22Technology+%26+Engineering
%22&source=gbs_ge_summary_r&cad=0
Applying Weighted Particle Swarm
Optimization to Imbalanced Data
in Software Defect Prediction
Lucija Brezočnik ✉ and Vili Podgorelec

( )
Faculty of Electrical Engineering and Computer Science,

{lucija.brezocnik,vili.podgorelec}@um.si
Abstract. Imbalanced data typically refers to class distribution skews and

underrepresented data, which affect the performance of learning algorithms. Such
data are well-known in real-life situations, such as behavior analysis, cancer
malignancy grading, industrial systems’ monitoring and software defect predic‐
tion. In this paper, we present a W-PSO method, which comprises weighting of
instances in a dataset and the Particle Swarm Optimization algorithm. The
presented method was combined with classification methods C4.5 and Naive
Bayes, respectively, and tested experimentally on ten freely accessible software
defect prediction datasets. Based on the results achieved, the presented W-PSO
method creates better classification models than classification methods C4.5 and
Naive Bayes in the majority of the cases.
Keywords: Machine learning · Swarm intelligence · Particle swarm optimization

Software defect prediction
1 Introduction
Predicting defect-prone software components is an economically significant activity and

thus has received a good deal of attention. Current defect prediction work focuses on
(1) Estimating the number of defects remaining in software systems, (2) Discovering
defect associations, and (3) Classifying the defect-proneness of software components,
typically into two classes, defect-prone and not defect-prone [15]. In this paper, we
concentrated on the third approach.
The prediction result can be used as an essential measure for the software developer
[16], and can be used to control the software process. Being able to predict which
components are more likely to be defect-prone supports better-targeted testing resources
and can, therefore, improve efficiency.
Recent research results from the fields of Machine Learning and Data Mining have
provided practical tools for using data miners to learn predictors for software quality [9]
automatically. In a never-ending quest of optimizing the software projects’ budgets, such
predictors can be used to focus the testing on parts of the system that seem defect-prone.
For this purpose, the predictors are built upon static code attributes from the software code

https://doi.org/10.1007/978-3-319-90893-9_35
290 L. Brezočnik and V. Podgorelec
repositories. In general, many different static code attributes can be used for learning
predicting models. Although some researchers oppose the value of static code attributes as
defect predictors [3], most of them nevertheless endorse them [7, 8, 14].
On the other hand, numerous learning methods can be used to construct the predic‐
tors, like Neural Networks, Support Vector Machines or Decision Trees, to name a few.
To choose the right one is not the most straightforward task to do, and the choice of
learning method is even far more important than which subset of the available data is
used for learning [9]. Different Decision Tree learning methods were often used in prior
work [7, 14], due to their efficiency and straight-forwardness. However, the big problem
of most of the existing learning methods is that they tend to maximize the overall
predictive accuracy. Overall accuracy is a good measure of a learner’s performance
when the possible outcomes occur with similar frequencies. In the software engineering
data, however, this is not the case. Therefore, a good learner to be used for predicting
software errors should provide high predictive accuracy of positive cases (defect-prone
software modules), while preserving low predictive error rate of negative cases (not
defect-prone software modules).
In our previous work, we have already designed different Decision Tree-based
methods to predict defect-prone software modules [10, 13], improved mining of soft‐
ware complexity data using evolutionary filtered training sets [11] and tackled the
problem of balanced prediction [12]. One of the significant issues, we had encountered,
and which is still being researched heavily, was the problem of dealing with imbalanced
data. For this purpose, we developed an evolutionary method that effectively weights or
samples the tuples from the training dataset and tries to minimize the adverse effects of
imbalanced data [5].
Encouraged by some very promising results, most recently, we developed a swarm
intelligence-based method W-PSO for improving the classification of imbalanced data,
which comprises weighting of instances in a dataset and the Particle Swarm Optimiza‐
tion algorithm [2]. As software engineering datasets are very imbalanced in nature, this
paper aims to adapt the W-PSO to software engineering data, and to evaluate it on a
number of software engineering datasets.
The structure of this paper is as follows. Section 2 outlines the fundamentals of the
W-PSO algorithm. The latter was tested with the experiment presented in Sect. 3. In
Sect. 4, the results are depicted, while, in Sect. 5, results of the statistical analysis are
given. Section 6 concludes the paper and outlines the directions for future work.
2 Weighted Particle Swarm Optimization Method
The presented method, proposed by Brezočnik et al. [2], is an enhancement of the PSO
algorithm [6] with weighting mechanism on the instances of the dataset. Algorithm 1
depicts the pseudocode of the presented method.
In the initialization phase of the algorithm, the population is defined of the possible
solutions, i.e., particles, X = [X1, X2,…,Xn]. Thus, i-th particle is represented as real-
valued vector Xi = [x1, x2,…,xd], where d denotes the dimension of the particle.
Figure 1 shows an example of a particle, where its size is the same as is the number of
Applying Weighted Particle Swarm Optimization 291
instances in the chosen dataset. Each square represents the value of the weight for a
single instance, i.e., the first instance has a weight 0.01, the second 0.89, the third 0.54,
etc. The larger the weight is, the greater is the importance of the instance.
Fig. 1. Particle representation of the W-PSO method
The algorithm explores new solutions by moving the particles throughout the search
space in the direction of the current best solution. Movement is possible because of the
following equations:
( ) ( )
vnew
id
= 𝜔 vold
id
+ c1 r pBestold
id
− xold
id
+ c2 r gBest − xold
id
,
(1)
xnew
id
= x old
id
+ v new
id
,
where
ω denotes inertia weight,

r denotes a random value in the interval [0, 1],
c1 denotes a cognitive component and
c2 denotes a social component.
As stated before, particles with movement track two extremes. The first one is known
as a global extreme, and comprises the best solution found so far in the entire swarm
(gBest), and the second one is an individual extreme. The latter represents the best posi‐
tion of each particle found so far (pBesti).
Since the W-PSO method is designed modularly, many components can be changed
or replaced easily, e.g., fitness function and classification method. In this paper, particles
were evaluated with the fitness function presented by Eq. 2.
number of correctly classified instances

fitness = (2)
number of instances
3 Experiment Setup
The experimental approach was used to show the power of the W-PSO algorithm on
imbalanced datasets. The proposed method was tested on ten different datasets. A
detailed description of them is presented in Table 1. The first column comprises names
of the datasets, the number of attributes and instances are defined in the following two
columns, and the last two columns present the distribution of the instances per class,
i.e., either a module has/has not one or more reported defects.
Table 1. Description of the used datasets

Dataset Number of Number of Modules with Modules
attributes instances no defect with defect
AR3 30 63 55 8
AR6 30 101 86 15
CM1 22 498 449 49
KC1-class- 95 145 60 85
level-binary
KC2 22 522 415 107
MC2 40 161 109 52
Mozilla4 6 15545 5108 10437
PC3 38 1563 1403 160
jEdit_4.0_4.2 9 274 140 134
jEdit_4.2_4.3 9 369 204 165
In the first part of the experiment, the W-PSO method was combined with the clas‐
sification method C4.5 (W-PSO_C4.5), and in the second part, with the classification
method Naive Bayes (W-PSO_NB).
In both cases, the same parameter settings were used: 100 particles were created
(Np), which were moving through 100 generations. The value of ω was set to 0.8, and
values of c1 and c2 were both set to 2. The dimension of the problem depended on the
tested dataset.
To prevent overfitting, we used 5-fold cross-validation jointly with a separate vali‐
dation set. The process is as follows. The model learns on the weighted training set, from
which 1/4 of the instances were removed for validation (validation set). When a classi‐
fication model is built on a training set, it is evaluated on a validation set. The best particle
of each fold, i.e., a particle that has the best result on the validation set, is then evaluated
on the test set. The final result is the value obtained as an average of 5 folds (see line 18
of Algorithm 1).
4 Results
Tables 2 and 3 give results of the presented method on ten datasets (see Table 1). As
one can see, method W-PSO outperformed classification method C4.5 in classification
accuracy (Acc) and F-measure in 80% of the cases. In one case, the methods gave the
same result (dataset Mozilla4), and, in one case, (dataset CM1), classification method
C4.5 performed better.
Table 2. Results of the classification accuracy and F-measure of C4.5 classification method and
W-PSO_C4.5
Dataset C4.5 W-PSO_C4.5
Acc F-measure Acc F-measure
AR3 0.692 0.567 0.906 0.704
AR6 0.800 0.444 0.801 0.664
CM1 0.880 0.634 0.858 0.540
KC1-class-level-binary 0.586 0.561 0.697 0.682
KC2 0.752 0.638 0.818 0.704
MC2 0.656 0.639 0.708 0.662
Mozilla4 0.942 0.933 0.942 0.933
PC3 0.853 0.610 0.878 0.640
jEdit_4.0_4.2 0.673 0.672 0.715 0.713
jEdit_4.2_4.3 0.568 0.563 0.602 0.592
Table 3. Results of the classification accuracy and F-measure of NB classification method and
W-PSO_NB
Dataset NB W-PSO_NB
Acc F-measure Acc F-measure
AR3 0.769 0.629 0.937 0.750
AR6 0.850 0.460 0.841 0.592
CM1 0.850 0.600 0.870 0.534
KC1-class-level-binary 0.621 0.542 0.724 0.708
KC2 0.886 0.799 0.831 0.712
MC2 0.844 0.800 0.696 0.613
Mozilla4 0.694 0.691 0.804 0.789
PC3 0.457 0.411 0.857 0.610
jEdit_4.0_4.2 0.618 0.583 0.690 0.686
jEdit_4.2_4.3 0.581 0.577 0.572 0.556
When W-PSO was combined with the Naive Bayes classification method, results
show, that Acc and F-measure were improved in 60% of the cases, while in the rest of
them, Naive Bayes performed better.
In Fig. 2, two box-plots of the obtained predictive performance results are presented–
one for accuracy and the other for F-measure score. Although the results were quite in
favor of the W-PSO method, this is not very obvious from the accuracy measure, because
accuracy is not an optimal metric for such skewed data. In this manner, we also present
the results of F-measure score, which can be interpreted as the harmonic average of the
precision and recall, where precision is the ratio between correct positive predictions
and all positive predictions, while recall is the ratio between correct positive predictions
and all positive instances. The box-plot of F-measure scores shows that W-PSO outper‐
formed the baseline methods considerably.
Fig. 2. The comparison of W-PSO with baseline methods on ten software defect prediction
datasets
5 Statistical Evaluation
Statistical analysis was conducted with the IBM SPSS Software [4]. The goal of statis‐
tical analysis is to check if the results of Acc and F-measure obtained from W-PSO
algorithm are statistically significantly better than results of classification methods C4.5
and Naive Bayes. The used alpha value α was 0.05.
In order to select the right statistic tests, the test of normality was performed first.
Since all results are not distributed normally, parametric tests were chosen. Using the
Paired Samples Test, we have proven a significant difference between W-PSO_C4.5 and
C4.5 (t9 = 2.419, p < 0.05). When the same test was applied to results of the W-PSO_NB
and NB (t9 = 1.388, p > 0.05), a significant difference was not proven. However, W-
PSO_C4.5 results are, on average, 5.2% better than C4.5 results (95% CI [0.003, 0.101]),
and W-PSO_NB results are 6.5% better than NB results (95% CI [−0.041, 0.171]).
For F-measure analysis, non-parametric tests were used as some results follow a
normal distribution. First, we compared W-PSO_C4.5 and C4.5 using the Friedman test.
Results show that statistical difference exists (p = 0.011), thus, the post hoc analysis
with the Wilcoxon Signed Ranks Test was conducted. Based on the fact, that p =
0.037, the statistical significance of W-PSO_C4.5 over C4.5 is confirmed. As we
compared W-PSO_NB and NB using the Friedman test, statistical significance between
them was not proven.
6 Conclusion
The presented method W-PSO was implemented and tested successfully on ten imbal‐
anced datasets. As results and statistical analysis show, the presented method outper‐
formed classification methods C4.5 and Naive Bayes in the majority of the cases. Thus,
there are many options for continuing research in this rather new area.
As the software engineering datasets are strongly imbalanced, different authors
suggest new metrics for evaluating the predictive performance of models for software
defects’ prediction. In this manner, Menzies et al. proposed balance as a metric to oper‐
ationalize the notion of balance between tpr (true positive rate, correctly predicting
software modules containing errors) and fpr (false positive rate, falsely predicting soft‐
ware without defects as erroneous) [9]. Following this approach, we plan to extend the
W-PSO method to optimize predictions with regard to this and other similar metrics.
It would also be interesting to apply the feature selection method as presented in [1]
to such skewed datasets, and observe the effect of selected informative features on tested
metrics.
References
1. Brezočnik, L.: Feature selection for classification using particle swarm optimization. In: IEEE
EUROCON 2017 – 17th International Conference on Smart Technologies, pp. 966–971.
IEEE, Ohrid, Macedonia (2017)
2. Brezočnik, L., Karakatič, S., Podgorelec, V.: Weighted particle swarm optimization for
imbalanced data. In: Twenty-sixth International Electrotechnical and Computer Science
Conference ERK 2017, pp. 387–390. IEEE Slovenia Section, Portorož, Slovenia (2017)
3. Fenton, N.E., Ohlsson, N.: Quantitative analysis of faults and failures in a complex software
system. IEEE Trans. Softw. Eng. 26(8), 797–814 (2000)
4. IBM: IBM SPSS Software. https://goo.gl/djbcCa. Accessed 10 Jan 2018
5. Karakatič, S., Heričko, M., Podgorelec, V.: Weighting and sampling data for individual
classifiers and bagging with genetic algorithms. In: 7th International Joint Conference on
Computational Intelligence IJCCI, pp. 180–187. IEEE, Lisbon, Portugal (2015)
6. Kennedy, J., Eberhart, R.: Particle swarm optimization. In: IEEE International Conference
on Neural Networks, vol. 4, pp. 1942–1948. IEEE, Perth, Australia (1995)
7. Khoshgoftaar, T.M., Allen, E.B., Deng, J.: Using regression trees to classify fault-prone
software modules. IEEE Trans. Reliab. 51(4), 455–462 (2002)
8. Menzies, T., Di Stefano, J.S., Chapman, M., McGrill, K.: Metrics that matter. In: 27th Annual
NASA Goddard/IEEE Software Engineering Workshop, pp. 51–57. IEEE (2002)
9. Menzies, T., Greenwald, J., Frank, A.: Data mining static code attributes to learn defect
predictors. IEEE Trans. Softw. Eng. 33(1), 2–13 (2007)
10. Pighin, M., Podgorelec, V., Kokol, P.: Fault-threshold prediction with linear programming
methodologies. Empirical Softw. Eng. 8(2), 117–138 (2003)
11. Podgorelec, V.: Improved mining of software complexity data on evolutionary filtered
training sets. WSEAS Trans. Inf. Sci. Appl. 6(11), 1751–1760 (2009)
12. Podgorelec, V., Karakatič, S.: Predicting software defect-proneness from software repository
data – a case of eclipse bug data. In: 18th International Multiconference Information Society
– IS 2015, Collaboration, Software and Services in Information Society, pp. 5–8. InstitutJožef
Stefan, Ljubljana, Slovenia (2015)
13. Podgorelec, V., Kokol, P.: Evolutionary induced decision trees for dangerous software
modules prediction. Inf. Process. Lett. 82(1), 31–38 (2002)
14. Porter, A.A., Selby, R.W.: Empirically guided software development using metric-based
classification trees. IEEE Softw. 7(2), 46–54 (1990)
15. Song, Q., Jia, Z., Shepperd, M., Ying, S., Liu, J.: A general software defect-proneness
prediction framework. IEEE Trans. Softw. Eng. 37(3), 356–370 (2011)
16. Wahono, R.S.: A systematic literature review of software defect prediction: Research trends,
datasets, methods and frameworks. J. Softw. Eng. 1(1), 1–16 (2015)
Intelligent Transport Systems, Logistics,
Traffic Control
Cooperative Systems in Traffic Technology
and Transport
Sadko Mandžuka ✉
( )
Faculty of Traffic and Transport Sciences,

Vukelićeva 4, 10000 Zagreb, Croatia
sadko.mandzuka@fpz.hr
Abstract. Traffic and transport engineering of the 21st century requires new
approaches to make transport safer, more efficient and reliable. The requirements
for urban traffic are especially high with demands in integration of different trans‐
port modes and significant reduction of noise and air pollution. Technological
approach known as intelligent transport systems is based on the use of modern
information-communication technologies and artificial intelligence algorithms.
A particularly important area are cooperative systems that provide wirelessly
communication between vehicles, infrastructure (roads and related equipment)
and other users (pedestrians, vulnerable road user, etc.). The main characteristics
of Traffic Cooperative Systems Architecture and some typical applications are
described in the paper.
Keywords: Intelligent transport systems · Cooperative systems

Information-communication technologies
1 Introduction
Traffic and transport engineering of the 21st century requires new approach in order to
make transport safer, more efficient and reliable, with minimal impact on the environ‐
ment and society. Basic characteristics of this new approach are increasing demands
followed by the imperative of lower costs. Traffic safety is especially very important. The
challenge is to increase traffic safety by 50% until 2020 in order to reach the main goal
of zero fatalities in 2050. The goal for more efficient traffic system is set for reducing
congestion – 2% Gross Domestic Product (GDP) and – 60% energy efficiency & emis‐
sions until 2050. The expectations for urban traffic are especially high with demands in
integration of different transport modes and significant reduction of noise and air pollution.
In this regard, in the 21st century, traffic engineering explores some new approaches
to achieving goals for safe, efficient and reliable transport with minimal impact on the
environment and society. Also, the basic today feature is increasing requirements with
the imperative of lower costs.
While great progress has been achieved in the past decades making the EU the
world’s safest road transport region, the high number of fatalities and serious injuries
still cause great human suffering and unacceptable economic costs, estimated at EUR
100 billion annually. In 2016, 25,500 people lost their lives on EU roads and a further

https://doi.org/10.1007/978-3-319-90893-9_36
300 S. Mandžuka
135,000 people were seriously injured, [1]. In key EU Commission document: White
Paper 2011: “Towards a “zero-vision” on road safety” is particularly emphasized
“Harmonize and deploy road safety technology - such as driver assistance systems,
(smart) speed limiters, seat belt reminders, eCall, cooperative systems and vehicle-
infrastructure interfaces - as well as improved roadworthiness tests for alternative
propulsion systems”.
Traffic Engineering professionals have recognized intelligent transport systems
(ITS) as the technological basis for today’s traffic challenges, [2, 3]. ITS is an holistic,
control and information & communication technologies (ICT) upgrade of classic traffic
and transport systems which significantly improves system performance, traffic safety,
efficiency in transportation of goods and passengers, increases passenger protection and
comfort, reduces pollution, etc. [12].
The possibilities of ITS traffic safety improvements can be divided into following:
(a) Systems related to infrastructure (roads, bridges, tunnels…)
(b) Systems related to vehicles
(c) Systems based on cooperation
A particularly potent approach was recognized in the possibility of application of
cooperative systems in traffic and transport.
2 Basic Characteristics of a Cooperative Approach
Cooperative system can be defined as a combination of technologies, people and organ‐

isations which enables communication and coordination required for achieving a
common goal of a certain group which performs various activities, for the benefit of all
participants. Traffic systems Cooperation also can be viewed in narrow sense as a
problem of communication of a moving entity (vehicle) with road infrastructure and/or
other moving entities. In this narrow sense the following communications were recog‐
nised: V2V – vehicle to vehicle, V2I – vehicle to infrastructure, V2P – vehicle to pedes‐
trian, I2P – infrastructure to pedestrian, [14], Fig. 1.
ITS as superstructure
Classical Traffic Infrastructure
Fig. 1. Definition of ITS, [12]

Cooperative Systems in Traffic Technology and Transport 301
In this manner cooperative concept can expand the “sensory” capabilities of the
driver (visual field - dead angle) or detection of vehicles not seen (from other vehicles,
roads infrastructure, buildings, etc.). This system reduce reaction time (early warning
systems, automatic drive mode) and warn of reducing driver’s focus on driving (other
vehicles, surrounding infrastructure etc.). Also, cooperative system reduces the effect
of monotonous car driving: (intelligent vehicle speed control, virtual road train) (Fig. 2).
Fig. 2. Cooperative systems topology, [14]
The main characteristics of a cooperative approach are:

1. Consider the driver, vehicle, infrastructure and other road users as a unique system
2. Consider operational and management needs of the entire system
3. Integrated approach to safety of traffic and all participants
4. Applies technology in a coherent manner in order to support overall integration of
system entities.
Currently the following systems recognized where the cooperative approach can be
successfully applied: navigation systems and travel information systems, warning systems,
emergency services’ vehicle management, priority management in urban public transport,
intelligent systems for speed control, support systems for vulnerable transport users and
others. The main components of traffic cooperative system is presented in Fig. 3. The main
components are On-board unit in car and Roadside unit as infrastructure component. Some
communication technologies can be applied: Short-range and ad hoc systems (CEN DSRC,
European 5.9-GHz ITS, wireless LAN, Bluetooth, Infrared), Cellular systems (WiMAX,
302 S. Mandžuka
global public mobile network (GSM/GPRS, UMTS, 5G) and Digital broadcast systems
(Digital audio broadcasting – DAB, digital multimedia broadcasting - DMB, digital video
broadcasting-terrestrial - DVB-T) etc.
Fig. 3. The main components of traffic cooperative system, [15]
3 Cooperative ITS Architecture
3.1 Basic Unit - Cooperative Station
The main entity of cooperative communications architecture is cooperative ITS station,

[4, 5]. It follows the principles of the standard OSI model for layered communication
protocols which is extended for inclusion of ITS applications, [13]. Figure 4 presents
the ITS station reference architecture. As represented in figure, the upper layers (above
networking and transport) are composed of Application and Facilities layer. The appli‐
cations layer provides ITS services. Three classes of applications have been distin‐
guished at this level: Road Safety, Traffic Efficiency and Other Applications. Each
application can be assigned to one of the three identified application classes. In order to
define the Basic Set of Applications (BSA), use cases are firstly examined and selected.
The Facilities layer provides generic support facilities to applications. This layer is
further composed of three main components: Application, Information and Communi‐
cation support.
The application support is the kernel of common functions supporting the applica‐
tions. This consists of station lifecycle management, automatic services discovery,
download and initialization of new services, HMI generic capabilities, and many others.
A key concept in ITS is the ability of transport entities (vehicles, roadside infrastructure,
pedestrians, etc.) to collect knowledge of their local environment, from a range of sensor
equipment, and to share that knowledge in order to make more intelligent use of the
transport infrastructure. This is described in the term “co-operative awareness”.
Fig. 4. ITS station reference architecture, [13]
The information support covers the presentation layer of the OSI reference model
and holds the role of data management. In any ITS system, there will be an abundance
of data sources, both mobile and static ones, [13]. These data will mostly be location
referenced, time specific and attached with life time value and with accuracy and reli‐
ability parameters. Therefore, fusing data and keeping the information up to date is one
of the challenges of information support. Main entity that supports this is Local Dynamic
Map (LDM) that is able to take data both from different sources and from received ITS
messages to build a data model of the local environment, [13].
The communication support, includes the session layer of the OSI Reference model.
It will cooperate with the transport and network layer to achieve the various communi‐
cation modes required by the applications.
3.2 V2X Facilities Layer Model and Facilities

The facilities layer is defined to provide generic functionalities currently identified that
can be shared by several applications and their assigned use cases. Figure 4 gives the
basic architecture of the facilities layer and the enclosed functional blocks. Application
support consists of next features, [13]: (a) Station positioning (3D position information
of the station); (b) Mobile station dynamic monitoring (vehicle real time kinematics
parameters); (c) Station state monitoring (current static state of ITS station); (d) Services
management (life cycle services); (e) Local Dynamic Map (LDM) management (update
the LDM data base); (f) Messages management (generating, structuring and formatting
application messages by available European standard; (g) Security access management
(security requirements of the ITS applications); (h) Time management (maintaining a
global time reference which is used for the time stamping data elements).
304 S. Mandžuka
Information support consists of (a) LDM data base (it contains all the data necessary
to build LDM); (b) Location referencing (positioning static and mobile objects on a
LDM); (c) Data presentation (coding and decoding messages according to formal
language); (d) Station type/capabilities (static and dynamic information related to ITS
station) (Fig. 5).
Fig. 5. Facilities layer and corresponding functional blocks, [13]
Communication support function provides capabilities to manage the various

communication modes e.g. broadcasting, geocasting, unicasting and communication
sessions.
3.3 Basic Set of Applications
There are 4 main applications classes:

1. Cooperative road safety: The primary objective of applications in the active road
safety class is the improvement of road safety and some function in Incident
Management, [6].
2. Cooperative traffic efficiency: The primary objective of applications in the traffic
management class is the improvement of traffic fluidity. However it is recognized
that in improving traffic management they may offer secondary benefits not directly
associated with traffic management, [7].
3. Co-operative local services and global internet services: Applications in the co-
operative local services and global internet services classes advertise and provide
on-demand information to passing vehicles on either a commercial or non-commer‐
cial basis. These services may include Infotainment, comfort and vehicle or service
life cycle management. Co-operative local services are provided from within the ITS
network infrastructure. Global internet services are acquired from providers in the
wider internet.
Table 1 identifies the application class, the application, and the use case that they are
considered in the development of ITS standardization, [13].
Table 1. Basic Set of Applications (BSA) - Taxonomy

Applications class Application Use case
Active road safety Driving assistance - Emergency vehicle warning
Co-operative Slow vehicle indication
awareness Intersection collision warning
Motorcycle approaching indication
Driving assistance - Emergency electronic brake lights
Road Hazard Warning Wrong way driving warning
Stationary vehicle - accident
Stationary vehicle - vehicle problem
Traffic condition warning
Signal violation warning
Roadwork warning
Collision risk warning
Decentralized floating car data - Hazardous
location
Decentralized floating car data - Precipitations
Decentralized floating car data - Road adhesion
Decentralized floating car data - Visibility
Decentralized floating car data - Wind
Cooperative traffic Speed management Regulatory/contextual speed limits notification
efficiency Traffic light optimal speed advisory
Co-operative Traffic information and recommended itinerary
navigation Enhanced route guidance and navigation
Limited access warning and detour notification
In-vehicle signage
Co-operative local Location based Point of Interest notification
services services Automatic access control and parking
management
ITS local electronic commerce
Media downloading
Global internet Communities services Insurance and financial services
services Fleet management
Loading zone management
ITS station life cycle Vehicle software/data provisioning and update
management Vehicle and RSU data calibration
306 S. Mandžuka
4 Use Cases
4.1 Emergency Vehicle Warning
This use case allows an active emergency vehicle to indicate its presence. In many
countries the presence of an emergency vehicle imposes an obligation for vehicles in
the path of the emergency vehicle to give way and to free an emergency corridor (Fig. 6).
Fig. 6. Emergency vehicle warning, [13]
The purpose is to reduce emergency vehicles intervention time to rescue and/or

protect people. This is also an active safety use case as reducing the risk of collision
between an emergency vehicle and another vehicle.
4.2 Wrong Way Driving Warning
This use case indicates to vehicles in the affected area that a vehicle is driving against
the planned direction of traffic. The affected area is primarily the road in which the
vehicle is driving in the wrong direction and the affected vehicles are those vehicles
approaching the violating vehicle (Fig. 7).
Fig. 7. Wrong way driving warning, [13]

4.3 Traffic Light Optimal Speed Advisory

This use case allows a traffic light to broadcast timing data associated to its current state
(e.g. time remaining before switching between green, amber, red). The purpose is Traffic
regulation at an intersection. Also, this application allows economical driving with
minimal emissions of exhaust gases. Communication mode is Periodic, as permanent
messages broadcasting (Fig. 8).
Fig. 8. Traffic light optimal speed advisory, [13]
5 Conclusion
Cooperative systems in traffic technologies and transport provide great possibilities. A

large number of projects in the implementation phase are being implemented in the EU
today. In addition, new use cases are also under developing, [8–10]. In this regard, great
importance is attached to the development of 5G infrastructure for the needs of coop‐
erative transport systems. Particular importance is given to security, especially in
communication technologies, [11].
Acknowledgment. The research presented in this paper is supported with University of Zagreb
Program funds Support for scientific and artistic research (2017) through the project: “Cooperative
Systems for Traffic Control”.
References
1. EU Commission, White Paper 2011: Towards a “zero-vision” on road safety (2011)

2. EU Parliament, Directive 2010/40/EU on the framework for the deployment of Intelligent
Transport Systems in the field of road transport and for interfaces with other modes of
transport (2010)
3. Mandžuka, S., Marijan, Ž., Horvat, B., Bicanic, D., Mitsakis, E.: Directives of the European
Union on intelligent transport systems and their impact on the Republic of Croatia. Promet
Traffic Transp. 25(3), 273–283 (2013)
4. Mandžuka, S., Škorput, P., Vujić, M.: Architecture of cooperative systems in traffic and
transportation. In: 23rd Telecommunications Forum TELFOR, Belgrade, pp. 25–28 (2015)
308 S. Mandžuka
5. Moreno, A., Onieva, E., Perallos, A., Iovino, G., Fernández, P.: Cooperative decision-making
ITS architecture based on distributed RSUs. In: Proceedings of 9th International Conference,
Puerto Varas, Chile (2015)
6. Škorput, P., Mandžuka, S., Jelušić, N.: Real-time detection of road traffic incidents. Promet
Traffic Transp. 22(4), 273–283 (2010)
7. Huzjan, B., Mandžuka, S., Kos, G.: Real-time traffic safety management model on
motorways. Tehničkivjesnik 24(5), 1457–1469 (2017)
8. Vojvodić, H., Mandžuka, S., Vujić, M.: The use of cooperative intelligent transport systems
for vehicle emission charging. In: 24th Telecommunications Forum (TELFOR), Belgrade,
pp. 1–4 (2016)
9. Vujić, M., Mandžuka, S., Gregurić, M.: Pilot implementation of public transport priority in
the city of Zagreb. Promet Traffic Traffico 27(3), 257–265 (2015)
10. Mandzuka, S., Gregurić, M., Kljaić, Z.: Cooperative environment for E-mobility
infrastructure. In: 24th Telecommunications Forum TELFOR, Belgrade, pp. 1–4 (2016)
11. Škorput, P., Vojvodić, H., Mandžuka, S.: Cyber security in cooperative intelligent
transportation systems. In: International Symposium ELMAR, Zadar, pp. 35–38 (2017)
12. Bošnjak, I.: Intelligent transport systems 1, Faculty of Traffic and Transport Sciences, Zagreb
(2006). (In Croatian)
13. ETSI, Intelligent Transport Systems (ITS); Vehicular Communications; Basic Set of
Applications; Definitions 102 638 V1.1.1, June 2009
14. Mandžuka, S., Ivanjko, E., Vujić, M., Škorput, P., Gregurić, M.: The Use of Cooperative ITS
in Urban Traffic Management, Intelligent Transport Systems: Technologies and Applications.
Wiley, NewYork (2015)
15. Bishop, R.: Intelligent Vehicle Technology and Trends. Artech House Publishers (2005)
Wireless Sensor Network Based Alarm Detection
and Monitoring of Cyber-Physical System
with Mobile Robot Inspection
Lejla Banjanovic-Mehmedovic ✉ and Mirzet Zukic

( )
University of Tuzla, 75000 Tuzla, Bosnia and Herzegovina

lejla.mehmedovic@untz.ba, mirzet.zukic@fet.ba
Abstract. The cyber-physical systems has been an evergrowing terminology in

today’s evolving Industry 4.0, combining improved information technology and
automation solutions, data and physical elements and the ability to connect
devices to one another using wireless sensor networks. In this paper we present
alarm detection and monitoring system using the wireless sensor network (WSN)
with the mobile robot inspection. An abnormal event could happen in any uncer‐
tain time, so we need more sensor nodes in industrial environment to set alarm
precisely if abnormal event happen. Our framework includes the autonomous
robot (agent) as an integral part of wireless sensor network and the mobile robot
travels to the positions, where the alarm was detected to investigate. The aim of
our work was to develop alarm detection system that could help in different
factory to install reliable alarm detection systems with localizing and monitoring
capability within a relatively lower cost. As such, it could be replicated anywhere
including complex event processing.
Keywords: Alarm detection · Cyber-physical systems · Industry 4.0

Mobile robot inspection · Monitoring · Wireless sensor networks
1 Introduction
Cyber-physical systems (CPS) are enabling technologies which bring the computational
and physical worlds together to create a networked world in which intelligent objects
communicate and interact with each other. Cyber-physical systems represent the next
evolutionary step from existing embedded systems, which together with the internet and
data and available services form cyber-physical systems [1].
Cyber-physical production systems, relying on the latest development of computer
science, information and communication technologies, manufacturing science and tech‐
nology are leading way to a new industrial age, defined by fourth industrial revolution
[2]. Industry 4.0 represents a leap forward from more traditional automation to a fully
connected and flexible systems, based on the application of artificial intelligence (AI)
and increasing sophistication of cyber-physical systems [3].
There are that underpin Industry 4.0 consists of several technological areas, which
are horizontal and vertical system integration, the internet of things, cybersecurity, the

https://doi.org/10.1007/978-3-319-90893-9_37
310 L. Banjanovic-Mehmedovic and M. Zukic
cloud, big data analytics, simulation, additive manufacturing (3d printing), augmented
reality and robot [4].
The concept Industry 4.0 is best defined by the project’s “smart factory” through the
merging of the virtual and physical worlds through cyber-physical systems and the
resulting fusion of technical and business processes [5]. The smart factories require the
underlying processes and materials to be connected through transparent network with
real-time tracking and monitoring in order to generate the data necessary to make real-
time decisions. An optimized smart factory allows operations to be executed with
minimal manual intervention and high reliability. In order to minimize failures of manu‐
facturing processes and avoid the huge economic losses, it is important to act before
issues or challenges arise, rather than simply reacting to them after they occur. This
feature can include regularly inspected activity, identifying anomalies and monitoring
safety concerns [5].
Sensor networks have recently emerged as an effective tool for monitoring and
detecting alarms states in industrial environments. Each individual node in the network
can monitor its local region and communicate through a wireless channel with other
nodes to collaboratively produce a high-level representation of the environment’s states.
In order to deploy sensor networks in such large environments, often with hundreds of
nodes, wireless sensors must be low-cost and affordable. Hence, wireless sensors are
typically highly limited in terms of sensing, computation, communication, battery life
and the actions they can perform. These limitations can be addressed through the addi‐
tion of more capable mobile robot sensors to form a hybrid wireless network. Higher-
capability mobile robots may be dispatched to gather more accurate temperature or
humidity readings, or recharging a static sensor’s battery which the static sensors are
not equipped for. Although the static sensors are less capable than mobile robots, they
are also much less expensive and can be used to cover a vast area at a low cost. The
static sensors must assign these events to mobile robots while simultaneously mini‐
mizing the distance travelled by the robots and the communication among the static
sensors to prolong their battery life [6].
There are few examples of hybrid sensors networks in different task allocation. The
event assignment problem for assigning mobile robots to events in hybrid wireless sensor
networks is presented in [6]. An achievements in the field of electric power substation
equipment’s inspection by mobile robot are presented in [7]. An intruder detection by
using both a WSN and mobile robots is presented in [8].
In this paper, we focus on the alarm detection and monitoring of cyber-physical
system using a wireless sensor network and automatic inspection approach using a
mobile robot. We first present alarm detection techniques in Sect. 2. Then, in Sect. 3,
we present the architecture of cyber-physical system. In Sect. 4 inspection mobile robot
localization based on topological approach is presented. Section 5 presents our experi‐
mental results of alarm detection and monitoring capabilities. Finally, we summarize
our findings in Conclusion.
Wireless Sensor Network Based Alarm Detection and Monitoring 311
2 Alarm Detection Techniques
In factories, manufacturing processes depends of corrected manufacturing or operational

environment values at all stages of the work. To prevent the accidents in factory, a key
method to be taken is to effectively monitor the certain physical values (changing
temperature of the environment or equipment such as transformers, cable connectors,
etc.). Additional recent data analytics capabilities include event stream processing and
complex event processing in critical operating plant situations [9]. An abnormal situation
or anomaly is defined as a disturbance or series of disturbances in a process that cause
plant operations to deviate from their normal operating or overrun limits. An alarm
system comprises hardware and software components, which can signal an alarm state,
transmit the signal to the process automation system, record the signal, and display a
message about the signal to the operator (monitoring).
There are many event detection techniques [10]. The initial idea of event detection
is to use a threshold value, if the value of sensor node exceeds the pre-defined alarm
threshold, an event is considered to happen. Pre-alarm function is a programmable option
which determines the system’s response to real-time detector sensing values above the
programmed setting as early warning of incipient or potential abnormal conditions.
Model-based event detection technique model event phenomena in a specific form such
as mathematical formulas or maps. Pattern matching-based event detection methods
define a data pattern for event and then an event is detected if data pattern matches with
event pattern. Fuzzy logic is a kind of machine learning which can tolerate unreliable
and imprecise sensor reading.
In our research, we used pre-alarm detection and alarm detection techniques in hybrid
wireless sensor network for cyber-physical system monitoring. This situation allows the
possible cause of the pre–alarm to be investigated prior to a full alarm condition. The
pre-alarm signal also provides an opportunity to filter out false alarms.
3 Architecture of Cyber-Physical System
Wireless Sensor Network (WSN) includes a number of battery powered sensor nodes,
a processing subsystem for local data processing and memory and a wireless commu‐
nication module [11]. The sensing nodes have a sensing device that acquires data from
the physical environment.
The designed system in our research was principally based on wireless sensor
network (WSN) and monitoring center. Most of the sensor nodes act as terminal devices,
can only collect data and send it to the coordinator node. Each static node contains a
Arduino processing unit, BMP180 sensor for temperature and pressure measuring and
a communication module. The communication between the nodes was established
through a wireless communication using XBee module and ZigBee/IEEE 802.15.4
standard. XBee radio supports AT and API mode for sending and receiving data [12].
Xbee modules used AT mode for this application.
In our approach, we enhance the alarms detection process by adding autonomous
robot as mobile sensor node and the coordinator of sensor network. The coordinator
node is responsible for collecting data from the end devices. After that, it sends this
information joined with the address of the appropriate sensor node to the computer. The
system framework is shown on Fig. 1.
Fig. 1. System framework
Arduino based mobile robot is designed with mechatronic features. The following
several basic technical indicators need to be met: (1) Move autonomously along the
guideline, which can be set up in advance and identified by visual identification unit on
inspection robot; (2) Communication capabilities of mobile sensor node in frame of
wireless sensor network; (3) In case of anomaly appearance, detected with static sensor
nodes, mobile robot node can immediately start inspection task.
The inspection robot generally can move along the guideline in the plant by using
two wheel differential steering mobile platform. In order to accurately locate the part of
environment which could be inspected in the plant, the static sensor nodes are used as
navigation target point in environment.
The operator can monitor status of static nodes and mobile robot via the monitoring
system instead of onsite manual inspecting.
4 Inspection Mobile Robot Localization
Topological maps ar typically represented as graphs in which navigation-relevant places

are modeled as graph nodes and connections between places are indicated by graph
edges [13]. Using the some form of visual based navigation, inspection mobile robot
can detect intersection points in path and change states. The topological form of navi‐
gation for our experimental setup is presented on Fig. 2.
Fig. 2. General form of topological navigation in our experimental testbed
In case, when inspection mobile robot gets information from perception unit
Pi = {1, 2, 3, 4, 5}, it makes possible actions Di = {Forward, Left, Right, Rotation 180,
[ ]
Back}, marked as Pi , Di . It moves forward or back through hallway Hi. (transition mode)
between intersection points (topological nodes i and i + 1) or right or left toward sections
points (sensor node position) SRi or SLi.
5 Experimental Results
The idea of experimental presentation of alarm detection and monitoring of cyber-phys‐

ical system is to organize four sections inside an industrial plant, two on the right and
two on the left side, with two corridors among them. The inspection mobile robot as part
of network communicate with static sensor nodes and with GUI application at the moni‐
toring device. The one example of experimental scenario is presented on Fig. 3.
Fig. 3. Experimental setup for alarm detection with sensor node in plant 4
Figures 4 and 5 show the results of measured level of temperature and pressure for
selected wireless sensor node, where pre-alarm and alarm were detected.
Fig. 4. Temperature and pressure measuring for selected sensor node; temperature pre-alarm
detection in plant 2
Fig. 5. Temperature and pressure measuring for selected sensor node; temperature alarm
detection in plant 4
The inspection robot states in environment during navigation process are presented
on Fig. 6. The navigation target point is defined with sensor node, where we have infor‐
mation that alarm was detected.
Fig. 6. Inspection mobile robot states in case of alarm detection in plant 4
The graphical based monitoring system based on the wireless sensor network for one
example of alarm detection scenario is presented on Fig. 7. Microsoft Visual Studio is
used for realization of the graphical monitoring, which can display the measuring data
collected by wireless sensor nodes in real-time and the inspection robot state in topo‐
logical map. The control panel can give three different visual indication: for stable state,
green indication (if the temperature and pressure are lower than the pre-alarm limit);
pre-alarm state (orange indication) and alarm state (red indication) for each sensor node.
This situation allows the pre–alarm to be investigated prior to a full alarm condition.
The pre-alarm signal also provides an opportunity to filter out false alarms.
Fig. 7. Monitoring system
6 Conclusion
Real-time environmental conditions and equipment monitoring is very important in the

smart factories in order to monitor potential threats or predict potential quality issues.
This research focuses on alarm detection and monitoring of cyber-physical system using
the advantages of a hybrid wireless sensor network. The developed system is very
flexible for different smart factory application using specific perception unit for inspec‐
tion mobile robot. Future work is based on improving perception part of inspection robot
and reducing false alarm occurrence by multi-sensor decision algorithms using different
type of sensors on the same location.
References
1. Willliam, M.D.: Industry 4.0 - Smart Manufacturing for the Future. Germany Trade & Invest,
Berlin (2014)
2. Working Group, acatech: Securing the future of German manufacturing industry:
recommendations for implementing the strategic initiative INDUSTRIE 4.0. Final Report of
the Industrie 4.0 (2013)
3. Bahrin, M.A.K., Othman, M.F., Azli, N.H.N., Talib, M.F.: Industry 4.0: a review on industrial
automation and robotic. Jurnal Teknologi (Sci. Eng.) 78(6–13), 137–143 (2016)
4. Michael, R., Markus, L., et al.: Industry 4.0: The Future of Productivity and Growth in
Manufacturing Industries. Boston Consulting Group (2015)
5. Hartigan, M., Laape, S., Mussomeli, A., Gish, D.: The Smart factory-responsive, adaptive,
connected manufacturing (2017)
6. Coltin, B., Veloso, M.: Mobile robot task allocation in hybrid wireless sensor networks. In:
IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pp. 2932–
2937 (2010)
7. Lu, S., Zhang, Y., Su, J.: Mobile robot for power substation inspection: a survey. IEEE/CAA
J. Automatica Sinica PP(99), 1–18 (2017)
8. Li, Y.Y., Parker, L.E.: Intruder detection using a wireless sensor network with an intelligent
mobile robot response. In: IEEE Southeast Conference 2008, Huntsville, Alabama (2008)
9. Us, T., Jensen, N., Lind, M., Jørgensen, S.B.: Fundamental principles of alarm design. Int. J.
Nucl. Saf. Simul. 2(1), 44–51 (2011)
10. Sun, X., Zhuang, A., Wang, B., Li, T., Liu, Q.: INFLE: an improved neighhbor-based fuzzy
logic event detecting algorithm for wireless sensor networks. Int. J. Future Gener. Commun.
Netw. 7(3), 13–24 (2014)
11. Ammar, M.B., Souissi, R.: A new approach based on wireless sensor network and fuzzy logic
for forest fire detection. Int. J. Comput. Appl. 89(2), 48–55 (2014). ISSN 0975-8887
12. Banjanovic-Mehmedovic, L., Husejnovic, A., Bosankic, I., Kasapovic, S.: Monitoring of
cooperation between autonomous vehicles in roundabout environment. In: 3rd International
Conference “New Technologies 2016 NT-2016 Development and Applications”, Bosnia and
Herzegovina, pp. 377–383 (2016)
13. Berns, K., von Puttkamer, E.: Autonomous Land Vehicles, Steps toward Sevice Robots.
Vieweg+Teubner, Verlag (2009)
Review of Simulation Based Comparison
of VANET Protocols
Zlatan Jukic1 ✉ and Muhammad Arshad2

( )
1
HTL Rankweil & Vienna University of Technology, Vienna, Austria
zlatan.jukic@htlr.snv.at
2
University of Engineering and Technology Lahore, Lahore, Pakistan
engineerarshad@hotmail.com
Abstract. Vehicle-to-Vehicle communications comprising of VANET is one of

the best way to transfer information between vehicles [1]. VANET protocols are
the rules that perform the role of governing body for all this communications
process. There are many parameters and factors that are taken into account while
evaluating the performance of any protocols including VANET protocols as well.
Protocol that performs best could be used for V2V communication in VANET.
Another part of the VANET system is so-called Vehicle-to-Infrastructure
communication that handles the communication of the vehicle with road side units
or RSUs. RSUs play a vital role in communication between vehicles that are far
away. There are certain factors again that effect the V2I communication. Vehic‐
ular Adhoc Network can be considered as the backbone of Intelligent Transpor‐
tation System (ITS) [2].
Keywords: VANET · Intelligent transportation system · Simulation topology

NCTUNs
1 Introduction
Nowadays, we witness emergence of technological paradigms, known collectively as

Ambient Intelligence, which influence on all the important segments of our lives. The
aim of related research is to create an intelligent daily space and enable, among others,
a more efficient use of various transport and transfer infrastructures. In particular, the
research focuses on architectures, applications and corresponding communication
models making our cars and roads safer. Central to this concept is possibility of efficient
exchange of various types of information, such as: traffic intensity, possible dangers,
weather conditions, need for emergency help and alternative routes. Such applications
are also examples of the so-called Intelligent Transportation System (ITS), whose
general goal is to provide the highest level of safety, efficient communication and well-
being on the road.
Vehicular Adhoc Networks fall under the hierarchy of wireless network that makes
the communication between the vehicles possible following some protocols. With the
passage of time various methods have been used to indicate the drivers about the
upcoming situations. There are many major challenges to be addressed in the research

https://doi.org/10.1007/978-3-319-90893-9_38
318 Z. Jukic and M. Arshad
of VANET mainly VANET protocols their stability and better performance and relia‐
bility. In this paper we are going to perform a brief comparison via simulation of routing
protocols used in VANET [3].
There are many different routing protocols for routing information from source to
destination. Some of them are unicast, multicast, geocast and broadcast [4]. VANET is
a sub category of MANET. The fundamental difference between VANET & MANET
is that MANET is a network which is infrastructure-less while VANET have particular
infrastructure to work with [5].
There are two basic types of protocols named astopology based protocols and posi‐
tion based protocols. VANET protocols are categorized as position based protocols [6].
In Sect. 2 we elaborate basic introduction and comparison of VANET and MANET.
Section 3 briefly explains the protocol types that can be used in VANET routing.
Section 4 is the simulation using different scenarios and results are obtained. Finally the
conclusion discusses the results of simulation.
2 Related Work
In [2], a taxonomy of the VANET simulation software is discussed comparing the

performance of the software in the form of results that they provide in similar cases. In
[7] another network simulation software TraNS is discussed that works in two different
modes application eccentric and network eccentric. Both [2] and [7] perform realistic
model simulation of VANET.
[8] Is about tutorial survey of the VANET providing different models also providing
the application of VANET in detail along with the hazards that can occur during the
V2V and V2I communication. The [9] discusses how to model the mobility of the cars
using OMNET++. OMNET++ is another software used for efficient VANET simula‐
tion and modeling.
[10] exposes the idea of beaconing in VANET. It discusses the vehicle-to-vehicle
communication in detail for increasing safety and comfort. It is also explained how the
802.11p standard provides a novel technique to decrease packet loss and ensure that data
packets are delivered successfully.
3 VANET/MANET
Mobile Adhoc network is a network type that is completely dynamic or mobile in nature
and it helps the nodes of ITS or intelligent transportation system to connect at runtime
and establish communication between each other to transfer messages through the
network [11]. Vehicular Adhoc Network as the name suggest the nature of the network
being Adhoc meaning that it is always an on demand type of network that is established
when required for communication between cars of ITS or intelligent transportation
system [12]. The term Adhoc refers to the idea that it is dynamic in nature and no specific
topology is used in this protocol as vehicle establish network on demand, which is clearly
Adhoc in nature. Different routing protocols have been proposed for use in VANET to
Review of Simulation Based Comparison of VANET Protocols 319
achieve better performance and reliability. Many types of problems occur in VANET
routing such as:
• Path Discovery for Packing Routing
• Maintenance of route for reliable communication
• Maintaining Adhoc nature of the network
Figure 1 below depicts a group of heterogeneous wireless networks, such as VANETs
with mobile access networks to support the mobile users, respectively cellular BS, WiMaX,
and WiFi as heterogeneous access networks. Each network consists of a number of vehi‐
cles, and a set of gateways, such as the Roadside Units (RSUs) inside VANETs. An impor‐
tant component of the VANET network is RSU. The task of RSU is to provide accurate and
detailed position information of vehicles. Furthermore the RSUs route the packets reliably
and efficiently and is responsible for various message transmission between vehicles and
other entities in VANETs networks. VANETs are used to protect against collisions between
vehicles in that drivers are informed whether a vehicle is being halt when the speed is too
high or the distance between vehicle and other objects on the road. There are two impor‐
tant communication protocols in these networks, Vehicle-to-Vehicle (V–V) and Vehicle-
to-Infrastructure (V–I) communications.
Fig. 1. VANET network communication architecture
The information between vehicles is exchanged through the (V2V) protocol the
communication between infrastructure and the vehicle via the (I2V) protocol. The
former one is characterized by a better performance. The protocol (I2V) has greater
quantity and higher quality. Vehicle-to-Infrastructure (V2I) communication is a part of
the (I2V) protocol. In turn, communications between elements of infrastructure, for
example between RSU or the communication between RSU and WiMAX Base Station
is established by the (I2I) protocol.
In order to enable the (V2V) connection, the vehicles must built a kind of network.
Because the vehicles are in movement all the time, there is no fixed infrastructure
supporting the communications – an immanent problem of ad-hoc networks.
When it comes to fully automatic vehicles (driven without human control), equipped
with many sensors, it is most crucial to ensure their capability of mutual communication
about dangers and exchange of warnings. In VANET, it is the GeoNetworking that
provides wireless communications between vehicles and between vehicles and base
station located along the road. GeoNetworking works in completely distributed manner
and is based fully on the ad-hoc network concept. It offers the functions of “geographical
addressing”, and “geographical forwarding”. In contrast to conventional networks,
where the parties are addressed by their identities (e.g. IP-address), data packets are sent
to a node, determined by its position, or to a number of nodes located in some geograph‐
ical area (Fig. 2).
Fig. 2. Inter-vehicle information model
4 Routing Protocols
Routing protocols are protocols of the network layer whose task is to guide the packets
from source to destination nodes. The exchange of messages between the two instances
is controlled by one or more efficient route selection protocols; the number of the unnec‐
essary overhead information must be minimal so that network resources are not wasted.
In order to make an optimal route selection the following parameters must be taken into
account: number of hops, real-time behavior and transmission speed. Furthermore, the
selection depends on: delay, availability and congestion of particular links and the volume
of interference of adjacent links. It is important to stress that the shortest route must not
necessarily be the best route. Many routing protocols allow the use of routing metrics to
select and to evaluate the quality of a route. Hop Count is possibly the most commonly
used metric, but there are many other metrics concerning conditions of wireless connec‐
tions such as expected number of transmissions (ETX), expected transmission time (ETT),
and weighted cumulative time ETT (WCETT).
The routing protocols are the backbone of any network communication for estab‐
lishing a communication path and deciding the recovery procedure in case of failure or
malfunction [13]. The routing protocols for VANET ad-hoc wireless networks can be
classified based on routing information update mechanism in following categories, Fig. 3:
1. Proactive (sensitive or table-driven) routing protocols
2. Reactive (practical or on-demand) routing protocols
3. Hybrid Routing Protocols
Fig. 3. MANET/VANET Routing protocols
4.1 Proactive Routing Protocols

In this section, we briefly introduce the fundamental concepts of proactive and reactive
routing and path selection protocols in wireless multi-hop networks. We also describe
how to combine proactive and reactive approaches to optimize routing and path selec‐
tion. In proactive routing protocols, which are table-driven protocols, it is necessary to
determine the paths to be used between two nodes before the data transfer begins.
The source nodes in ad-hoc mobile networks maintain routes of all accessible desti‐
nation nodes. This happens regardless of whether these routes are needed or not needed.
To select the correct route a node must regularly send control messages. Therefore,
proactive routing protocols can waste bandwidth (because the control messages must
be sent also when there is no traffic at all). However, the advantage of this category of
protocols is that the Hosts can quickly get the route information and quickly set up a
session. They try to find and to store the reliable routing information in their routing
tables. The routing tables are updated regularly to react on changes in network topology.
For these reasons they are not suitable for large networks because the node entries
must be maintained for each node in each node’s routing table. Some examples of proto‐
cols from this category are DSDV, STAR, and OLSR. They are also the table-driven
routing protocols. Proactive or link-state protocols ensure that each source node in the
network has information about the current routes to all other destinations. Such protocols
are very attractive since they find optimal routes without delays. The OLSR (Open Link
State Routing) protocol is widely used in wireless networks (originally developed for ad-
hoc mobile networks). It is described in RFC 3626. The protocol finds the optimal route
based on the smallest number of hops between nodes. The mechanism by which OLSR
distributes the link status updates throughout the network is particularly interesting.
The backbone algorithm behind this type of protocols is the shortest path algorithm
as it keeps track of all the nodes in the system in a table format which is visible to other
neighbors as well [14]. The major drawback behind these type of protocols is that they
keep track of all the nodes of a network always including nodes that are not being used.
So is a path is not in use every now and then and are still broadcasted which consumes
a lot of bandwidths of the network. This is the main deficiency of using this type of
protocols in VANET.
4.2 Reactive (on-Demand) Routing Protocols

Reactive or on-demand protocols work best in dynamic networks, where there are many
mobile nodes and where the network topology changes rapidly. This type of routing
differs significantly from the traditional proactive routing. Its basic idea is to find and
maintain only the required routes only when are needed and, thus, to avoid costs of
maintenance of the unused routes. Two phases of operation must be distinguished:
(a) Route determination: The source node initiates route detection based on a request.
The source node consults its route cache for the possibly available route from source
to the destination or, if the route does not exist, the route identification is initiated.
The data packet sent by the source node contains the destination addresses of the
node as well as addresses of the intermediate nodes which are on the way to the
destination node.
(b) Route maintenance: Due to the network’s dynamic topology, it comes by reason
of the connections breaks to a route failure between the nodes so that the route
maintenance must be carried out. The connections break imply for example the
frequently disconnected connection: it means the highly dynamic topology results
in frequent changes in its connectivity, thus the link between two vehicles can
quickly disappear while they are transmitting information.
Reactive routing protocols are efficient from proactive routing protocols as they
overcome the drawback of high bandwidth utilization in proactive routing by main‐
taining record of all paths [14]. Reactive routing protocols only keep track of the paths
or nodes that are active and in use currently. AODV is one of the reactive routing proto‐
cols that keep tracks of the nodes dynamically and efficiently updates the list or table
formed for keeping track of nodes in the network. AODV protocol mainly relies on
maintaining routing tables at intermediary nodes but DSR routing protocol utilized
source routing. Which states that source indicates the order of nodes in which the packet
will travel from source to destination [15].
4.3 Hybrid Routing
Hybrid routing is a combination of the proactive and reactive procedures and achieves
better performance optimization. Hybrid protocols exploit a multilevel hierarchical
network by clustering of nodes in close proximity (referred as a layer 1 network). This
node grouping is organized so that each cluster has a cluster head that acts as a gateway
(default router) for all other nodes. The cluster heads of different clusters communicate
with each other, forming the level 2 network.
The hybrid approach uses a reactive routing in the level 1 (within the cluster) and a
proactive routing in the level 2 network (between clusters). Such collaboration optimizes
the performance because a minor node mobility changes within a cluster often change
routes, but the inter-cluster routes are affected only when the nodes travel a long distance.
Hybrid protocols, Fig. 4 are therefore extremely effective in scenarios where nodes
move frequently but within a small geographic area. Proactive protocols exhibit high
overhead and small delays, while reactive protocols have less overhead but more latency.
The disadvantage of hybrid routing, as with all topology-based methods, is its limited
scalability and extensive overhead in managing routes in large ad-hoc networks. In
addition, hybrid routing uses the reactive-protocol route detection mechanism and the
proactive-protocol table maintenance mechanism to avoid network latency and overhead
problems. For these reasons, the topology-based routing methods are not suitable for
highly dynamic VANETs.
Fig. 4. VANET hybrid architecture
The proactive and reactive routing protocols had their own pros and cons so a new
type of protocols was proposed known as hybrid protocols utilizing the best features of
reactive and proactive protocols that resulted in a new protocol also known as ADV
protocols. These protocols work as DSDV at start but converts to on-demand type to
save bandwidth of the network. Hence, the drawbacks of proactive and reactive routing
are resolved [14].
4.4 Description of AODV and ADV Protocols
A routing protocol has major influence on the throughout and packet loss rates of the
wireless networks, given their physical layer efficiency and topology. There exists a
great number of ad-hoc routing protocols, each having its advantages and drawbacks.
In what follows, we give a short description of the Adaptive Distance Vector (ADV)
algorithm and the Ad Hoc On-Demand Distance Vector (AODV) algorithm.
4.5 ADV-Routing Protocol

The Adaptive Distance Vector (ADV) is a Distance Vector Routing algorithm or ADV is
a Distance-Vector routing algorithm that uses sequence numbers to determine loops in the
networks. This protocol reduces the routing overhead thanks to the fluctuations of traffic
intensity and node mobility. The fluctuations depend on the frequency and amount of
routes’ actualizations. Only the routes connecting active users are maintained, so that the
number of the announced entries is reduced and full regular actualizations are avoided.
4.6 AODV Routing Protocol
Ad Hoc On-Demand Distance Vector (AODV) is a routing protocol designed in 1991

for wireless mobile ad-hoc networks. This protocol establishes source-destination routes
on demand and supports unicast as well as multi-cast routing. The routes are established
only when required by the end-nodes and maintained just so long as required by them.
Therefore, minimum additional overhead traffic is produced. In case of multicast group
communication a tree-like routes are built. The update of routes is performed by help
of sequence numbers.
A node that receives such a message and maintains established routes, sends
acknowledgements about status of these routes back to the source. The node that initial‐
ized the connection, uses the route with the smallest number of hops (i.e. the shortest
path approach). The entries in the routing tables that are not used, will be re-used after
some time. Upon breakdown of a connection, the error will be acknowledged at the
sending node and the process of establishing the route will be repeated.
5 Simulation
5.1 Performance Metrics - Packet Drop

The analysed performance measure its packet drop parameter. The parameter shows
total number of packets dropped during the communication between nodes. This mainly
happens due to congestion or obstacles like buildings and mountains that hinder the
communication.
5.2 Throughput
Throughput is the ratio of number of packets delivered accurately to the destination to

the total number of packets transferred. Measured as follows:
Total Packets Delivered ∗ Packet Size

T.P =
Total Packets transfered
Throughput is measures in bytes/sec.

5.3 Total Simulation Time

This is specified in NCTUNs 6.0 tool as to show how much time the simulation took for
each scenario with every different protocol in consideration.
5.4 Introduction of the Tool
The tool used for simulation of VANET protocols is known as NCTUns. NCTUns is perhaps
the finest tool to use for simulation of network for ITS. It has the ability to merge network with
traffic simulation to and give real time network and vehicular simulation of VANETs. Speed
and direction upon receiving messages from nearby vehicles are taken into account [16].
5.5 Simulation Scenarios
We have considered six cases for the highway and city scenario whereby in every scenario
different number of nodes (vehicles) and other parameters as different speeds are considered.
The topology design as discussed previously is done in NCTUNs 6.0. Before going in to the
scenarios simulation attributes need to be specified as given in the Tables 1, 2 and 3.
Table 1. Highway scenario parameter

Attributes Values
Number of nodes 7,14
Max. Speed of node 30 m/s, 18 m/s
Packet type UDP
Simulation time 70 s
Table 2. City scenario parameter

Attributes Values
Radio obstacles no 4
Attenuation provided 20 dBm
Max. Speed of node 15 m/s, 7 m/s
Packet type UDP
Simulation Time 70 s
Table 3. Highway & city scenario parameter

Attributes Values
Packet type UDP
Simulation time 200 s
Traffic signal alteration Standard time
5.6 Highway Scenario

Zoomed image of the highway topology is shown below which is a two row road.
5.7 City or Urban Scenario
Zoomed image of topology shown below displays a mix of two row and four row roads
normally in use in city area (Figs. 5, 6, 7, 8, 9 and 10).
Case 1:
Case 2:
Fig. 5. Highway scenario depiction with 7 Fig. 6. Highway scenario depiction with 14
nodes and 30 m/s nodes and 18 m/s
Case 4:
Case 3:
Fig. 7. City Scenario depiction with 17 nodes Fig. 8. City Scenario depiction with 42 nodes
and 15 m/s and 7 m/s
Case 5: Case 6:
Fig. 9. City Scenario depiction with 22 nodes Fig. 10. City scenario depiction with 45 nodes
and 15 m/s and 7 m/s
6 Simulation Result
Similar to city scenario, in highway scenario ADV protocol is dominating in its perform‐
ance as it has lesser number of packets dropped during the communication. When the
speed increases from 20 to 30 m/s, the amount of packets dropped with AODV increases
up to 440 packets.
Similar to highway scenario Fig. 11, in highway scenario ADV protocol is domi‐
nating in its performance as it has lesser number of packets dropped during the commu‐
nication.
Fig. 11. Packet Drop comparison in highway scenario case 1 with 7 nodes and 30 m/s
Similar to highway scenario Fig. 12, in highway scenario ADV protocol is domi‐
nating in its performance as it has lesser number of packets dropped during the commu‐
nication (Fig. 13).
Fig. 12. Packet drop comparison in highway scenario case 2 with 14 nodes and 18 m/s
Fig. 13. Packet drop comparison in highway scenario case 2 with 14 nodes and 18 m/s
Fig. 14. Packet drop comparison in city scenario case 3 with 17 nodes and 15 m/s
The above graph of Fig. 14 shows that ADV protocol is performing better than
AODV as we can see value of packet drop is small for ADV as compared to AODV
shown by blue and red lines respectively.
Fig. 15. Packet drop comparison in city scenario case 3 with 17 nodes and 15 m/s
The above graph of Fig. 15 shows that ADV protocol is performing better than
AODV as we can see value of packet drop is small for ADV as compared to AODV
shown by blue and red lines respectively.
Here also, the above graph of Fig. 4 shows that ADV protocol is performing better
than AODV as we can see that the value of packet drop is smaller for ADV compared
to AODV, which is illustrated by the blue and red lines. In the urban scenario, the ADV
completely outperforms AODV. The increase of the speed leads to the drop of 479
packets, while ADV improved its performance (Figs. 16, 17, 18 and 19).
Fig. 16. Packet drop comparison in city Fig. 17. Packet drop comparison in city
scenario case 5 with 22 nodes and 15 m/s scenario case 6 with 45 nodes and 7 m/s
Fig. 18. Packet drop comparison in city Fig. 19. Packet drop comparison in city
scenario case 4 with 22 nodes and 15 m/s scenario case 4 with 42 nodes and 7 m/s
7 Conclusions
Simulation results reveal a very important fact which is one of the drawbacks of VANET
protocols like ADV and AODV: the packet drop increases due to obstacles such as
mountains or buildings on the way of communication. Another fact is that it is difficult
to maintain continuous connection because the distance between the vehicles increases,
and, in the absence of RSUs, glitches may occur in the communication. The results are
in favor of having a hybrid protocol like ADV which can perform better than proactive
and reactive routing protocols separately, but still there is a lot of space for improvements
and further research.
References
1. Kumar, R., Dave, M.: A Comparative Study of Various Routing Protocols in VANET. arXiv
preprint arXiv:1108.2094 (2011)
2. Martinez, F.J., Toh, C.K., Cano, J.C., Calafate, C.T., Manzoni, P.: A survey and comparative
study of simulators for vehicular ad hoc networks (VANETs). Wirel. Commun. Mobile
Comput. 11, 813–828 (2011)
3. Hartenstein, H., Laberteaux, K.: VANET: Vehicular Applications and Inter- Networking
Technologies, vol. 1. Wiley Online Library (2010)
4. Lin, Y.-W., Chen, Y.-S., Lee, S.-L.: Routing protocols in vehicular ad hoc networks: a survey
and future perspectives. J. Inf. Sci. Eng. 26, 913–932 (2010)
5. Ranjan, P., Ahirwar, K.K.: Comparative study of VANET and manet routing protocols. In:
Proceedings of the International Conference on Advanced Computing and Communication
Technologies (ACCT 2011) (2011)
6. Umashankar, I.R., Purnima, K.: A Comparative Study of Topology and Position Based
Routing Protocols in Mobile Ad Hoc Networks (2014)
7. Piorkowski, M., Raya, M., Lugo, A.L., Papadimitratos, P., Grossglauser, M., Hubaux, J.-P.:
TraNS: realistic joint traffic and network simulator for VANETs. ACM SIGMOBILE Mobile
Comput. Commun. Rev. 12, 31–33 (2008)
8. Hartenstein, H., Laberteaux, L.: A tutorial survey on vehicular ad hoc networks. IEEE
Commun. Mag. 46, 164–171 (2008)
9. Báguena, M., Tornell, S.M., Torres, Á., Calafate, C.T., Cano, J.-C., Manzoni, P.: VACaMobil:
VANET car mobility manager for omnet ++. In: 2013 IEEE International Conference on
Communications Workshops (ICC), pp. 1057–1061 (2013)
10. Noori, H., Olyaei, B.B.: A novel study on beaconing for VANET-based vehicle to vehicle
communication: probability of beacon delivery in realistic large-scale urban area using 802.11
p. In: 2013International Conference on Smart Communications in Network Technologies
(SaCoNeT), pp. 1–6 (2013)
11. Larsson, T., Hedman, N.: Routing Protocols in Wireless Ad-Hoc Networks-A Simulation
Study. University of Technology, Luleå, Stockholm. Master Thesis (1998)
12. Rehman, S.-u., Khan, M.A., Zia, T.A., Zheng, L.: Vehicular ad-hoc networks (VANETs)-an
overview and challenges. J. Wirel. Netw. Commun. 3, 29–38 (2013)
13. Jazayerifar, E., Sabbaghi-Nadooshan, R.: Performance Evaluation of AODV and ADV
Routing Protocols in Safety Situations in Highway Environment (2015)
14. Abolhasan, M., Wysocki, T., Dutkiewicz, E.: A review of routing protocols for mobile ad
hoc networks. Ad Hoc Netw. 2, 1–22 (2004)
15. Awerbuch, B., Mishra, A., Hopkins, J.: Dynamic Source Routing (DSR) Protocol
16. Wang, S., Chou, C.: NCTUns tool for wireless vehicular communication network researches.
Simul. Modell. Pract. Theory, 17, 1211–1226 (2009)
Ramp Metering on Urban Motorways
Martin Gregurić ✉ , Sadko Mandžuka, and Edouard Ivanjko

( )
University of Zagreb, 10000 Zagreb, Croatia

{martin.greguric,sadko.mandzuka,edouard.ivanjko}@fpz.hr
Abstract. Reduced Level of Service on urban motorways, which actually repre‐

sents evolved urban bypasses, is the product of two overlapping problems. First
of them is related to the heavy congestions, and the second one is related to the
unavailable constructional build-up of their capacities since they are surrounded
by the urban and traffic infrastructure. In order to cope with those problems, it is
necessary to introduce urban motorway control methods such as for example ramp
metering (RM). The main goal of RM is to increase the throughput of urban
motorways by restricting access of on-ramp traffic to mainstream traffic by using
special traffic lights. In this paper, an overview of currently most used RM algo‐
rithms (ALINEA, SWARM, and HELPER) and their fundamental deficiency in
partial problem solving for different traffic scenarios is given. A new RM algo‐
rithm based on the Adaptive Neuro-Fuzzy System neural network called
INTEGRA is also described.
Keywords: Urban motorways · Ramp metering · Traffic control

Machine learning
1 Introduction
Urban bypasses were constructed in order to serve transit traffic mainly composed of
heavy vehicles. Such transit traffic presents a major burden for the urban traffic network
because of the mentioned heavy vehicles. With the expansion of the urban area, the
urban bypasses have become surrounded by urban infrastructure and the adjacent urban
traffic network. The local population started to use urban bypasses as the fastest connec‐
tion between suburbs and urban center, as well as a convenient way for avoiding traffic
lights and congestion in the urban center. This is possible since the urban bypasses are
connected with the urban traffic network by numerous on- and off-ramps. At this time
it is possible to talk about the transformation of urban bypasses into the urban motorways
which serve both, local and transit traffic. Over the years increase of the transit and local
traffic has led to the decreased Level of Service (LoS) on urban motorways. LoS can be
furthermore reduced during severe periodic and non-periodic fluctuations in traffic
demand that are enabled by a combination of increased traffic demand, the proximity of
the urban area, and numerous on- and off-ramps fairly close to each other. Since the
constructional build-up of existing urban motorways is infeasible in most cases due to
its urban surrounding, it is necessary to introduce new motorway traffic control
approaches from the domain of intelligent transportation systems (ITS). ITS can be

https://doi.org/10.1007/978-3-319-90893-9_39
Ramp Metering on Urban Motorways 333
defined as holistic, control, information and communication upgrade to classical trans‐

port and traffic systems enabling significant improvement in performance, traffic flow,
the efficiency of passenger and goods transportation, safety and security of transport,
reduction of pollution, etc., [1, 2]. ITS based traffic control can improve the LoS of urban
motorways without building new infrastructure.
On- and off-ramps are places where it is possible to make a significant impact on
urban motorway mainstream traffic flow and on-ramp queues by using Ramp Metering
(RM) [1, 3]. The main goal of RM is to increase the throughput of urban motorways by
restricting access of on-ramp flow to mainstream flow by using special traffic lights [2].
The core element of RM is an algorithm that computes on-ramp flow restrictions rates
according to the measured traffic parameters. In order to create a conventional RM
algorithm, an exact model of traffic flows on the urban motorway is needed. The most
challenging problem in the design of the exact traffic models is the stochastic nature of
the traffic flows. Therefore, only the main characteristic of traffic flows can be captured
in current models [3, 10]. The consequence of the absence of the exact traffic model is
the lack of robustness of conventional RM algorithms on sudden fluctuations in traffic
demand [4, 5]. Latest approaches in RM algorithm design are based on model-inde‐
pendent techniques from the domain of Artificial Intelligence (AI). Those techniques
are based on machine learning and generalization of control knowledge with the main
goal to provide amore comprehensive nonlinear feedback control for the RM control
problem. In line with the latest approaches, this paper will provide a description of an
approach, which includes integration of selected existing RM algorithms into a new RM
algorithm called INTEGRA by using an advanced learning framework based on the
Adaptive Neuro-Fuzzy System (ANFIS) neural network.
2 Ramp Metering
Large traffic flow from an on-ramp entering the urban motorway can significantly reduce
speed and increase the density of the mainstream flow, what consequently can lead to
downstream congestions. Downstream congestions can create shock waves spreading
upstream on the urban motorway. Additionally, if an oversized vehicle platoon from an
on-ramp is stopped long enough it will produce a long on-ramp queue that potentially
can block arterial roads of the adjacent urban network [3]. The core idea of RM is to
transform on-ramps, as the entrance point of all traffic to the urban motorway, into a
controlled buffer zone. This buffer zone stores vehicles that are not allowed to merge
with the mainstream traffic due to the high traffic density in the motorway mainstream.
Visualization of a complete installation of a local RM system on an urban motorway is
shown in Fig. 1. Most important part of the RM system along with the detectors, appro‐
priate traffic lights, and other signalization is the RM algorithm which determines the
“access rate reduction” for the respective on-ramp flow [6].
334 M. Gregurić et al.
On-ramp data/so ware

Underground fiber LAN
server Warning sign to warn drivers
that ramp metering is on
Ramp metering Urban arterial road

signal device Queue
override
Central detector
traffic Check-out sensors
management Check-in sensors
server
Mainstream
Highway Upstream
Downstream
sec on data
road sensors server
road sensors
Underground fiber LAN
Fig. 1. Basic ramp metering installation on a motorway section [6]
2.1 Conventional Ramp Metering Algorithms
RM algorithms can be based on local and system-wide control strategies. Local strat‐
egies take into account only the local traffic situation, and the most representative algo‐
rithms based on those strategies are Demand-Capacity, Percent-Occupancy, and
ALINEA. ALINEA is the most widespread local RM algorithm with the main goal to
keep the downstream occupancy of the main traffic at a specified level by adjusting the
metering rate of the controlled on-ramp [5].
Other mentioned local RM algorithms are furthermore simplified by measuring
downstream occupancy and comparing it with predefined values. The system-wide
(coordinated) strategies consider the overall traffic situation on the entire controlled
motorway section [5]. For the sake of the research that is presented in this paper, only
two RM algorithms from two sub-categories of system-wide RM strategies will be
presented. The HELPER algorithm for ramp metering firstly detects the place of a major
bottleneck and enrolls several upstream on-ramps to create virtual on-ramp queues.
Virtual queues have the primary goal to stop entering additional traffic flow from one
or more upstream on-ramps into the mainstream flow in order to mitigate any down‐
stream congestion. If a bottleneck is not currently present, RM is conducted using the
local RM approach [7]. SWARM as a competitive RM algorithm contains a local and
global RM algorithm. The local algorithm in SWARM defines the metering rate
according to the difference between the current and critical traffic density for a particular
on-ramp. The global algorithm conducts coordination and predicts bottlenecks on the
motorway by using short-term historical data. The more restrictive metering rate value
between these two control logic is chosen as the final one [3].
2.2 Ramp Metering Algorithms Based on Artificial Intelligence Techniques

Artificial neural networks (ANN) and fuzzy logic based control system are the most
recent methods for application of AI techniques in RM algorithm design. The fuzzy
logic based control approach tackles the sensitivity problem of crisp discretization in
rule-based algorithms, which cannot adequately present traffic situations due to the non-
linear and non-stationary nature of traffic flows [8]. Fuzzy Inference System (FIS) eval‐
uates all defined fuzzy rules in parallel based on combination and aggregation of their
inference results and consequently produces crisp control outputs – metering rates. In
the other hand, ANN is widely used in RM design and prediction of traffic demand due
to its spatiotemporal approach in sorting time-space sequences that are adequately
mapped with metering rates. Estimation of ANN control outputs is derived according
to the dataset of rules that is created by adjusting the ANN weight structure during the
learning process. In RM algorithm design regarding the application of AI, it is also
possible to use the following techniques: Iterative learning control (ILC), evolutionary
and reinforcement learning algorithms, and Deep Learning (DL).
2.3 INTEGRA
The usability and effectiveness of the used RM algorithms significantly depend on their
ability to react to various different traffic scenarios including the unforeseen rapid
changes in traffic demand. In other to cope with those challenges, this paper describes
an integration of existing standard RM algorithms into a new algorithm named
INTEGRA. INTEGRA is based on the ANFIS framework, which uses an adaptive ANN
to modify parameters of a Takagi–Sugeno FIS during the learning process. The learning
dataset is derived from teaching RM algorithms (ALINEA, SWARM, and HELPER),
which are simulated on the same traffic scenario. This dataset is processed with an
adequate criteria function, which essentially evaluates achieved results of teaching
algorithms according to the parameters that represent the overall state on urban motor‐
ways such as Travel Time (TT) and Delay [7]. Additionally, metering rates of the orig‐
inal INTEGRA can be adjusted with predictions of on-ramp demand like it is described
in [9]. Functionality scheme for learning the INTEGRA algorithm can be seen in Fig. 2.
Fig. 2. Functionality scheme of the INTEGRA approach [7]

336 M. Gregurić et al.
3 Simulation Results
Prior to implementation, RM control systems have to be tested in simulations using real-

world traffic data. Zagreb bypass between the nodes Lučko and Jankomir is simulated
as the use case scenario in the CTMSIM macro-simulation tool [6]. According to
Table 1, the proposed original INTEGRA has produced the second lowest average TT
value compared to the other RM algorithms.
Table 1. Results of comparative analysis between different RM algorithms

No ALINEA SWARM HELPER INTEGRA Predictive
Control INTEGRA
Average TT 14.46 7.39 5.58 6.82 6.43 6.69
Average Delay 6.06 8.8 8.03 7.29 10.01 7.03
Average queue length 0 16 18 17 19 16
Maximal queue length 0 40 49 40 42 42
The proposed INTEGRA with prediction has produced a slightly higher value of TT,
but it has achieved a lower average value of Delay and lower average number of vehicles
in on-ramps queues compared to the original INTEGRA. This is an important result
since the original INTEGRA has produced highest Delay values compared to the other
RM algorithms. No control scenario produces zero on-ramp queues due to simulator
settings.
4 Conclusion
Most of the today’s urban motorways are characterized by the high fluctuation in traffic
demand, what in combination with numerous on-ramps, which are close to each other,
induce the need for RM. Main drawbacks of the conventional RM algorithms is their
dependency upon exact traffic models, and the fact that they provide best results under
specific traffic scenarios. Modern approaches in the design of algorithms for RM involve
the application of AI techniques that are based on machine learning. Therefore, no exact
models of the traffic system are required. This paper describes the INTEGRA algorithm
which conducts integration of control knowledge from several different teaching RM
into the one control strategy by using the ANFIS framework. Obtained results have
shown that INTEGRA algorithm can produce better results in comparison with its
teaching RM algorithms. Those results can be improved if the metering rates computed
by INTEGRA are adjusted according to the prediction of traffic demand.
Acknowledgment. The research presented in this paper is supported with University of Zagreb
Programme funds Support for scientific and artistic research (2017) through the project:
“Cooperative Systems for Traffic Control”.
References
1. Mandžuka, S., Žura, M., Horvat, B., Bićanić, D., Mitsakis, E.: Directives of the European
Union on Intelligent Transport Systems and their impact on the Republic of Croatia. Promet-
Traffic Transp. 25(3), 273–283 (2013)
2. Mandžuka, S., Horvat, B., Škorput, P.: Development of ITS in Republic of Croatia. In:
Proceedings of 19th ITS World Congress, ERTICO, Wien (2012)
3. Gregurić, M., Ivanjko, E., Mandžuka, S.: The use of cooperative approach in ramp metering.
Promet-Traffic Transp. 28(1), 11–22 (2016)
4. Papageorgiou, M., Diakaki, C., Dinopoulou, V., Kotsialos, A., Wang, Y.: Review of road
traffic control strategies. Proc. IEEE 91(12), 2043–2067 (2003)
5. Gregurić, M., Ivanjko, E., Mandžuka, E.: Cooperative ramp metering simulation. In: 2014
37th International Convention on Information and Communication Technology, Electronics
and Microelectronics (MIPRO), Opatija, pp. 970–975 (2014)
6. Gregurić, M., Ivanjko, E., Mandžuka, S.: A neuro-fuzzy based approach to cooperative ramp
metering. In: Proceedings of the 18th IEEE ITSC, Las Palmas, Spain, pp. 54–59 (2015)
7. Zhang, M., Kim, T., Nie, X., Jin, W.: Evaluation of On-ramp Control Algorithms, California
PATH Program - Research Report UCB-ITS-PRR-2001-36, University of California,
Berkeley (2001)
8. Wei, C.H.: Analysis of artificial neural network models for freeway ramp metering control.
Artif. Intell. Eng. 15(3), 241–252 (2001)
9. Treiber, M., Kesting, A.: Traffic Flow Dynamics. Springer Science Business Media,
Heidelberg (2013)
10. Feng, C., Yuanhua, J., Jian, L., Huixin, Y., Zhonghai, N.: Design of fuzzy neural network
control method for ramp metering. In: Proceedings of the 3rd ICMTMA IEEE, 6–7 January,
Shanghai, China (2011)
Terminology Extraction to Build an Ontology
of Intelligent Transport Systems
Pero Škorput ✉ , Sadko Mandžuka, and Markus Schatten

( )

{pero.skorput,sadko.mandzuka}@fpz.hr, markus.schatten@foi.hr
Abstract. Ontologies in the field of information sciences represent the basic

building blocks in formal semantic infrastructure construction, which are needed
for the development of various semantics-aware applications. At the beginning
of any ontology development, it is important to create a dictionary of concepts
which will make-up the ontology. The methodological basis for the construction
of an Intelligent transport systems ontology and iterative procedures for its expan‐
sion is described in this paper. Furthermore, knowledge base analysis methods of
terminology extraction for building ontologies within the domain of intelligent
transport systems is presented.
Keywords: Intelligent transport system · Ontology · Dictionary
1 Introduction
The use of innovative technologies in traffic and transport are necessary for economic
growth and development. Intelligent Transport Systems (ITS) become associated with
innovative technologies such as linked data, semantic web, online enterprise mobility
solutions, etc., and their use is widespread and there are a wide range of research works
and applications. ITS can be defined as a technology platform or application that
improves the efficiency of the mobility, transport productivity, energy consumption and
capacity of the transport systems [1, 9]. Unfortunately, existing ITS applications have
not reached a sufficient level of semantic interoperability. The ITS applications should
be able to exchange linked traffic data or structured traffic information and thus be more
useful for users. Formal ontologies could be used as a mechanism to specify and reuse
knowledge between different ITS applications. In computing, the ontology term refers
to the formulation of a comprehensive and rigorous conceptual schema given within one
or more domains [2]. Being independent of language and understandable by computers,
ontologies are useful because they help to achieve a common and integrated under‐
standing of descriptive information.
Through this research paper a terminology extraction method to build an ontology
of ITS is presented. In the second chapter the methodological conceptualization of
semantic modelling the ITS domain and constellations of different ontologies in traffic
domain are described. The third chapter describes the terminology extraction stage of

https://doi.org/10.1007/978-3-319-90893-9_40
Terminology Extraction to Build an Ontology of Intelligent Transport Systems 339
the ontology development. This step enables a first point of traffic data representation
in machine-recognizable form. Those methods used are described in more detail below.
2 Methodology
2.1 Conceptualization for Semantic Modeling In ITS Domain
The development of ontology models for ITS starts with an analysis of their features in
a theoretical and applicative sense [3–5]. The demand for the ontology model stems
from the consideration of problem domain, as well as its limitations. They stem from
the fact that the ontological model cooperative processes still cannot be fully modelled
in the field of ITS, because there are no coordinated or cooperative models in traffic
sciences.
The main part of ontology modeling consists of the following considerations:
• Observing the ITS problem domain as an integral part of the overall transport system;
• Abstraction in terms of theoretical extraction of all phenomena, things, events, infor‐
mation, etc., as well as entities involved in the processes like traffic management;
• Mapping entities as objects, defining attributes, operations and relations of objects
in the semantic model of the domain;
• Mapping of the domain model into the ontology [4]
While defining concepts, we should start from the simplest and general concepts
which will serve as building blocks for complex or composite classes. These classes will
be used to model complex concepts and relationships. The primary method for building
ontologies should follow the top-down systematic approach, and, if necessary, comple‐
mented with a bottom-up approach. The principles of building an open ontology should
be based on an open or framework approach.
Figure 1 shows the mapping process of a real traffic environment in the ontological
model of ITS. The mapping from the left to the right shows the research process in which
the phenomenological model gradually formalizes or conceptualizes specialized soft‐
ware tools. The path from the right to the left confirms the certification process in which
the model compares with reality. The reality is modeled with traffic science theories,
which is reflected in the phenomenological models. UML models are based on infor‐
mation theories, modeling diagram classes, and object-oriented modeling [5].
Fig. 1. Cycle mappings of real traffic environment in the ontological model of ITS [3]
340 P. Škorput et al.
In accordance with currently used ontologies, we suggest a new ontology model

based on the expressiveness of description logic and more specifically OWL [2] (web
ontology language), which was standardized by the W3C and the selected ontological
editor, accompanied by a tool for the acquisition of knowledge (ontology editor and
knowledge-based framework).
2.2 Phenomenological Model

To describe knowledge in the field of cooperative ITS different ontologies are used,
which define concepts such as events, routes, locations, etc. [6]. Together with other
ontologies that describe the time and geographical relationships, they form one unit
using one of the basic ontology properties which is knowledge sharing. This feature is
extremely important for the establishment and/or expansion of the ontology because it
is not necessary to reintroduce knowledge that has already been described.
There are a number of relations between different ontologies that describe the knowl‐
edge of a problem domain that only speaks in favor of the fact that knowledge is global,
shared and reused between different problem domains, while the corresponding constel‐
lations of defined concepts, relations and properties interpret certain problem domains.
Figure 2 shows the constellations of different ontologies in traffic domain and how
they share concepts, relations or properties. An ontology of ITS is an ontology which
aggregates the main concepts detailed in other ontologies.
Fig. 2. Constellations of different ontologies in the traffic domain [3]

3 Terminology Extraction
At an early stage of the ontology establishment, it is important to clearly define a glossary

of terms that are to be included in the ontology. Previous chapters generally describe
guidelines for semantic analysis from which it is possible to extract terminology used
for building and expanding the ontology in an iterative process.
One of the most important sources is the Datex II standard [7], which is the backbone
of information exchange between European transport centers. Its data model is proven
to be a good foundation and source of traffic knowledge in building a domain ontology.
For the establishment of the open ontology this protocol is important in the following
areas:
• Traffic events (congestion, accidents, barriers, road conditions, etc.)
• Service activities (traffic management, works, roadside assistance and the like)
• Consequences of adverse events (delays, the length of the column and the like)
• Information on the environment (transit information, parking, etc.).
• Derived data (data obtained by calculation)
• Measurement data (data obtained by measuring the road)
This protocol is adequate for marking traffic events, but it does not specify how to
use this data. This fact actually represents a kind of technological niche and an area of
innovative improvements. Application of semantic technologies offers such improve‐
ments, i.e. the open ontology model exactly describes the meaning and more importantly
types of use of such information. The reports that are sent via this protocol, from the
viewpoint of ontology, represent factual knowledge. There are four types of such traffic
reports:
1. Reports about the events on the road - describe one or more traffic scenarios that
were causally related, i.e. in addition to a detailed description, it provides location
information;
2. Traffic reports - elaborated traffic data that are sent periodically, and actually consists
of interpretations of the traffic conditions in a given area, which are generated and
sent by centers for traffic management based on the collected data;
3. Measurement reports - represent a form in which measured data that are obtained at
a certain monitoring station (e.g., weather stations, traffic counters, etc.) are sent
periodically, or a group of monitoring stations in a given area;
4. Traffic conditions - makes cross-section of events in a given transport area for a
period of time.
The model of this Protocol was created in UML (unified modelling language) which
will be discussed in the next section.
The protocol dictionary is a significant benefit because it is comprehensive, i.e., a
list of concepts used for the exchange of traffic information between traffic centers as
shown in Table 1.
Table 1. List of concepts used for the exchange of traffic information between traffic centers
Package name Definition Package name Definition
… … … …
Accident Situations in which TrafficData A collection of
one or more vehicles information
lose control and do associated with
not recover. These measured or derived
include collisions values relating to
between vehicle(s) or vehicle movements at
other road user(s), a specified location
between vehicle(s)
and obstacle(s), or the
result of a vehicle
running off the road
TrafficConcentration Averaged TrafficFlow Averaged
measurements of measurements of
traffic concentration traffic flow rates
TrafficSpeed Averaged TravelTimeData Measured or derived
measurements of travel time
traffic speed information relating
to a defined linear
stretch of the road
network
WeatherRelated Weather related TrafficConcentration Averaged
information measurements of
traffic concentration
… … … …
Source: DATEX II protocol
The traffic environment is a collection of events that may be related to the interaction
of humans, vehicles and traffic infrastructure [6]. It can include one or more situations
that are caused by the influence of external or internal factors. The ontological model
can be presented as an abstract class. The traffic environment as a class within the
ontology model, in addition to properties (or knowledge) mapped from the real world,
contains logical features that are copied from the data model or databases that contain
traffic data.
Figure 3 shows a UML class diagram that describes traffic events that will be
conceptualized in the ontology. It is viewed as an abstract class that aggregates four
lower-level classes. These are Accident, People involved, Vehicles involve and Vehicle.
The depicted section presents a clear example of the conceptualization of traffic
environment. From an ontological perspective, this model does not contain any formal
semantics, but it contains detailed descriptions of entities and hierarchies that signifi‐
cantly facilitate the creation of ontologies.
Fig. 3. UML class diagram of traffic events
4 Conclusion
This research is in the area of building an ontology of ITS. The significance of this work
is in the theoretical-cognitive sense based on the investigations that led to the develop‐
ment of resulting in a higher degree of semantic interoperability between various ITS
applications. This enabled the first point of traffic data presentation in the context of
machine-recognizable form. The use of ontology is located in establishment of a
common and open conceptualized domain model, necessary for use of multi-agent tech‐
nologies in area of ITS [8].
Multi-agent technologies with the use of ontologies can carry out specification tasks
for the traffic environment, can perform semantic search, multiple use of knowledge,
integrate traffic information from heterogeneous information sources and establish
semantic interoperability between various ITS applications.
The research and development of a new methodology for ITS is necessary to expand
research activities for further development of ontologies. Comparative experience in the
parallel development of ITS application and domain ontology show that stages of
ontology construction through various informal models of design and analysis and
formal in the form of expanding ontologies new concepts, using the evaluation should
follow the life cycle of the system.
References
Union on intelligent transport systems and their impact on the Republic of Croatia. Promet
Traffic Transp. 25(3), 273–283 (2013)
2. Gruber, T.R.: A translation approach to portable ontology specifications. Knowl. Acquisition
5, 199–220 (1993)
3. Škorput, P.: Open Ontology Model of Cooperative Intelligent Transport Systems, Ph.D thesis
(in Croatian), Faculty of Traffic and Transport Sciences University of Zagreb (2014)
4. Škorput, P., Mandžuka, S., Kljaić, Z.: Applying of semantic web in the critical infrastructure
protection management. In: 19th Telecommunications Forum–TELFOR, Beograd (2011)
5. Škorput, P., Mandžuka, S., Bošnjak, I.: Ontology development in ITS domain. In: 19th
Intelligent Transport Systems World Congress, Wien (2012)
6. Škorput, P., Mandžuka, S., Schatten, M.: Ontologies in the area of cooperative intelligent
transport system. In: 21st Telecommunications Forum-TELFOR (2013)
7. Wei-feng, L., Wei, C., Jian, H.: Research on a DATEX II based dynamic traffic information
publish platform. In: Second International Symposium on Intelligent Information Technology
Application, IITA 2008, pp. 412–416 (2008)
8. Schatten, M., Tomičić, I., OkrešaÐurić, B.: A review on application domains of large-scale
multiagent systems. In: Central European Conference on Information and Intelligent Systems,
pp. 201–206 (2017)
9. Mandžuka, S., Ivanjko, E., Vujić, M., Škorput, P., Gregurić, M.: The use of cooperative ITS
in urban traffic management. In: Intelligent Transport Systems: Technologies and Applications.
Wiley (2015)
Cooperative Vehicle Actuated Traffic Control
in Urban Areas
Miroslav Vujić ✉ , Sadko Mandžuka, and Luka Dedić

( )

{miroslav.vujic,sadko.mandzuka}@fpz.hr,
dedic.luka7@gmail.com
Abstract. Rapid increasing in the number of vehicles in urban areas directly

affects traffic quality because of the physical limitations of road infrastructure.
Frequent congestions in the cities cause the increase of delays, stop-and-go
actions and especially travel times. Up-to-date methods of traffic control are
focused on advanced intelligent transport systems solutions in order to increase
the quality of urban traffic network. This is directly affects improvement of traffic
parameters quality such as travel time, delay, etc. Also, defined parameters affect
the quality of environmental factors in urban traffic network (emission of pollu‐
tion from motor vehicles, fuel consumption, noise level, etc.). Higher level of
traffic control is enabled by direct real-time communication between vehicles
(personal vehicles, public transport vehicles, etc.). This method outperforms
traditional fixed cycle management of signalized intersections, which use prede‐
termined signal plans. Benefits of using cooperative actuated control where model
with implemented control algorithms is compared with existing traffic model with
fixed time signal plans are described. The main goal is to harmonize traffic flow
with optimal vehicle speed and to reduce the number of stop-and-go actions which
directly affects emission of pollution and fuel consumption.
Keywords: Intelligent transport systems · Energy efficiency · Level of Service

Urban traffic control
1 Introduction
Rational usage of time and space means that the traffic management must be oriented
to modern traffic control strategies and technologies. In order to improve modern trans‐
port systems it is necessary to insist on increase of traffic system quality, especially in
urban areas where the population is bigger and the consequences of traffic are increasing
exponentially. Traffic Engineering professionals have recognized intelligent transport
systems (ITS) as the technological basis for new approaches in solving the problems of
today’s (and especially the future) transport system. ITS is an holistic, control and
information & communication technologies (ICT) upgrade of classic traffic and trans‐
port systems which significantly improves system performance, traffic safety, efficiency
in transportation of goods and passengers, increases passenger protection and comfort,
reduces emission of pollution, etc. [1, 2, 14].

https://doi.org/10.1007/978-3-319-90893-9_41
346 M. Vujić et al.
The research within the eCoMove project shows a great development and progress
in the mentioned area – “beyond-state-of-the-art” strategies are defined with the imple‐
mentation of the cooperative systems (vehicle-to-vehicle (V2V) and vehicle-to-infra‐
structure V2I communication). Research within the eCoMove project is presented
through the “green priority strategies”, where a part of the urban network was controlled
by UTOPIA (Urban Traffic Optimization by Integrated Automation) software. Research
resulted in overall reduction of stop-and-go actions with the emphases on the stop
reduction of heavy duty vehicle [3]. Furthermore, research conducted by authors [4]
compared the “state-of-the-practice” transit signal priority (TSP), a coordinated/actu‐
ated signal control, with multi-modal V2I communication and results showed reduction
in bus delay and pedestrian delay while still providing similar car delay. The CVIS
(Cooperative Vehicle-Infrastructure Systems) project defines that the efficiency of the
system is drastically increased with the number of vehicles that are equipped with the
new ITS applications [5]. According to the project results [6], by the implementation of
the cooperative approach and the generic algorithms, the improvements were achieved
at the stop-and-go actions and the average delay decrease which is directly related to
Level of Service (LoS). The research [7] presented an impact of the implementation of
the adaptive priority strategies on the travel time’s reduction in urban areas regarding
public transport. Simulation based optimization was used on the model of the city of
Lausanne, Switzerland, and a microscopic evaluation model of fuel consumption was
implemented in the Aimsun simulation software. The presented research upgrade was
made by the integration of the q-learning algorithms (model-free reinforcement learning
technique) in the SUMO (Simulation of Urban MObility) simulation software, where
the improvements in the number of vehicles in queue length and CO2 reduction were
presented [8].
The progress is made worldwide in upgrading the traffic management system.
Furthermore it shows the usage of a micro and macro simulation models that can repre‐
sent a real time traffic situation in simulated environment making on that way research
efficient. The main purpose of this research was implementation of the vehicle actuated
traffic control algorithm and related benefits in reductions in the number of stop-and-go
actions, vehicle delay and emission of pollution.
Vehicle actuated traffic control algorithms were implemented in a simulation model
with PTV VISSIM simulation tool [9]. The demonstration corridor is selected in the city
of Zagreb and described in the second chapter. The description of the developed algo‐
rithms is presented in the third chapter, while the results are presented in the fourth
chapter. Improvements were made regarding queue lengths, number of stops, vehicle
delay, and vehicle speed harmonization. The emission of pollution CO2, NOX and
PM10 gases evaluated in EnViVer [10] at high traffic congestion. Conclusions and
further research plans are given in the last chapter.
2 Use Case Description
The main problem in the city of Zagreb is traffic fluctuation (westbound and eastbound)
in the morning and afternoon peak hours. The defined demonstration corridor is located
Cooperative Vehicle Actuated Traffic Control in Urban Areas 347
in the west part of the arterial route which is important for connecting the west part of
the city with its center. The defined demonstration corridor with its signalized intersec‐
tions is presented in the Fig. 1.
Fig. 1. Demonstration corridor - Zagrebačka Avenue in the city of Zagreb with its signalized
intersections
The demonstration corridor is 3500 [m] long and is consisted of seven signalized
intersections where six of them have four approaches and one of them has three
approaches. The side approaches on every intersection are problematic regarding traffic
demand, so the algorithm had to be personalized for each intersection. The existing
traffic control situation has fixed signal plans. The main direction has a green-wave
simulation from both sides (offset). Every intersection signal cycle is 150 [s] long, with
the implemented intergreen matrix regarding pedestrians. In this research, the pedestrian
flow was not generated, but it was included in the calculations.
2.1 Traffic Data Collection
The afternoon peak hour (4 PM–5 PM) was selected as relevant for simulation modeling
because of major traffic demand. The traffic data was collected on every signalized
intersection simultaneously and two major road vehicles categories were defined -
personal vehicles (PV) and heavy duty vehicles (HDV). Regarding vehicle composition,
personal vehicles make 95,95%, while heavy duty vehicles make 4,05%. The speed limits
on the selected corridor are set to 60 [km/h] on main approach and 40 [km/h] on side
approach. Traffic flow saturation was calculated, so the capacity of the defined demon‐
stration corridor could be compared with the gathered traffic flow data. Traffic flow
saturation was calculated according to Highway Capacity Manual [11]:
S = s0 × N × fW × fHV × fg × fp × fbb × fa × fLT × fRT (1)
The defined correction factors are:
fW -The lane width factor,

fHV -The group of vehicles homogeneity factor,
fg -The longitudinal tilt factor,

fp -The parking factor,
fbb -The public transit vehicles factor,
fa -The area type factor (urban area, motorway, etc.),
fLT -The left-turn vehicles factor,
fRT -The right-turn vehicles factor.
According to the previously defined equation, the gathered traffic on the defined
demonstration corridor is oversaturated. According to the fixed traffic control of the
defined corridor’s signalized intersections, the main path flow is oversaturated by 32,5%
westbound and by 26,3% eastbound. Oversaturation is detected on the north approaches
(21%), while on the south approaches the saturation is within limitations.
2.2 Simulation Model Calibration and Evaluation Parameters
After calibration of the simulation VISSIM traffic model and EnViVer emissions of
pollution model the next evaluation parameters were selected:
(a) Evaluation parameters regarding quality of the traffic network
• The average queue length [veh],
• The traffic flow [veh/h],
• The number of stops,
• The delay on the intersection/Level of Service [s].
(b) Evaluation parameters regarding emission of pollution
• CO2,
• NOX,
• PM10.
Apart from the parameters regarding quality of the traffic network, parameters for
the exhaust emission were gathered and evaluated. For exhaust emission the following
parameters were considered; CO2, NOX, PM10.
For the evaluation, two scenarios were taken into account; comparison of the simu‐
lated network with fixed and adaptive management with oversaturated flow, and scenario
with fixed and adaptive management with reduced demand at a saturation flow. The
simulations were conducted ten times for each defined scenario with the random seed
number increment of 1. The simulation duration was set on 4500 [s] where the data from
the last 3600 [s] was collected considering 900 [s] for network charging.
3 Algorithm Development
The algorithms for adaptive traffic control were developed for each signalized intersec‐
tion separately because of different physical situations on the intersections (the number
of signal phases, the duration of phases, etc.). There are two levels of traffic control for
the whole demonstration corridor – the global and the local level.
3.1 The Global Level

The global level of traffic control is essential for the maintenance of the “green wave”
on the main path throughout the corridor. This kind of control is significant for traffic
flow harmonization and the elimination of unnecessary stop-and-go-actions. Figure 2
presents the global level and the requirements for maintaining a “green wave” within
the algorithm.
Fig. 2. The global level
The global level of the algorithm is consisted of three parts in which the first part
sets a “green wave”. The values measured in seconds are depending on the distance
between signal controllers. After the first requirement is met, the second part initializes
the cycle within the individual signal control. The last part of the global level is testing
the conditions of the given parameters on the local level where the time factor is the
first condition. If the maximal cycle second is not fulfilled, the algorithm continues to
check the local level. Otherwise the algorithm returns to the beginning.
3.2 The Local Level
The local level of traffic control is programmed according to the maximum utilization
of effective green times [12]. A real-time traffic situation is monitored through the
detectors that are located on the approaches on every intersection. The detectors monitor
traffic demand on each segment of every approach so the optimal distribution of green
times can be achieved [13].
Local level of the algorithm is divided to particular phases. Every phase has a certain
sets of conditions, where the main phase is the “master” phase which decides the order
of green lights for other phases. In demonstration corridor the exclusive approach is
given to main direction, where the secondary phases (phases in which north and south
bounds have green lights), are tested after meeting the conditions in primary “master”
phase. The priorities of activation of the next phase are defined with the order of detection
in algorithms (firstly, main approaches are examined, than the left-turn approaches on
main path, etc.).
The sequence of phase activation is not predetermined; algorithm is adapting phase
activation according to traffic demand on each approach, but with accent on main path
approaches. In other phases on each signalized intersection after the examination of
minimum green times, the conditions which are defined on global level are analyzed.
Regarding existing traffic situation and calculated saturation, data priorities were defined
in algorithms so the first priority is allocated to main path lanes (westbound and east‐
bound directions), after the completion of adaptive traffic control on main path, north
approaches are the second level of calculation, and finally, south approaches have the
minimum advantages regarding phase activation correction. This allows the reduction
of stop-and-go actions on main path approaches and the group of vehicles is released
through the corridor. Figure 3 presents the local level and conditions within algorithm.
Fig. 3. The local level of the algorithm
When there are no special traffic demands on the detectors, the algorithm works as
defined on every intersection in the fixed time signal plan. In every phase, signal groups
that are not in conflict are “opened”, but minimum green times are calculated regarding
minimum pedestrian crossing times on every signalized intersection. When vehicles are
detected on each approach detector, green times extend upon repeated detection or the
expiry of maximum green time for the selected approach/phase. The algorithm ensures
that all of the phases get activated in every cycle without exceeding the psychological
barrier of 120 [s]. The traffic situation on every approach of each signalized intersection
is monitored in every second of the cycle.
4 Cooperative Approach
The presented algorithms can be upgraded with implementation of communication

between vehicles and infrastructure. Cooperation between vehicles and infrastructure is
essential for development of new traffic control systems to improve the quality of the
urban traffic system in general.
As shown in Fig. 4, the vehicle enters the cooperative control zone and cooperative
control mode is enabled. The vehicle sends expanded set of information (vehicle location
and speed, current number of passengers, possible delays according to predefined time‐
tables, etc.) to Traffic Management Centre – TMC. V2I communication is achieved
between vehicle and TMC. According to received information and current traffic situa‐
tion TMC makes a decision about the need for optimization activation, and selects
favorable technique. In cases of incident in the cooperative control zone TMC sends
vehicle Infrastructure driver

(TMC)
Vehicle enters cooperative

control zone
Traffic Management
System
Sending Receiving
information information
(V2I) (V2I)
Activation of optimization
algorithms
Receiving
Sending
information
information (I2V)
(I2V)
Manual reaction of driver

Information displayed
(speed adjustment, route
on-screen (to driver)
alteration, etc.)
Fig. 4. Basic diagram of cooperative approach scenario [15]

real-time information to the driver (detailed information about the incident, expected
delay, alternative routing, etc.) who then manually makes the decision about next
actions. It can be seen that cooperation is achieved within traffic and transport system,
and the improved system enables the observation of the driver, vehicle and infrastructure
as one unified (upgraded) system. With cooperative (advanced) approach the public
transport system can be observed as one unique system, with real-time communication
between vehicles and infrastructure.
5 Conclusion and Further Research
The results of this research show that the impact of the adaptive traffic control affects
the quality of urban traffic network, which is demonstrated through a simulation model
of the selected corridor in the city of Zagreb. The corridor, consisting of several signal‐
ized intersections, has been simulated with the adaptive traffic control algorithms, which
ensure green lights to dedicated groups of vehicles. The algorithm enables giving partial
green times to defined groups of vehicles and with this concept, the “lost” green times
are minimized. The algorithm shows great improvements in traffic flow performance
and affects flow rates at high congestion, which has been shown by decreasing number
of stops, queue lengths and delays. Although the algorithm works well under saturated
flow, room for improvement exists. Further research will be focused on improvement
the existing algorithm in the mean of self-learning abilities. Expected benefits are higher
robustness regarding lower demands of traffic flow and overall improvement of traffic
network management.
References
Union on Intelligent Transport Systems and their impact on the Republic of Croatia. Promet
Traffic Transp. 25(3), 273–283 (2013)
2. Vujić, M., Mandžuka, S., Gregurić, M.: Pilot implementationofpublic transport priorityinthecityof
Zagreb. Promet Traffic Traffico 27(3), 257–265 (2015)
3. Robertson, D.I.: Research on the TRANSYT and SCOOT methods of signal coordination.
ITE J. 56(1), 36–40 (1986)
4. He, Q., Head, K.L., Ding, J.: Multi-modal traffic signal control with priority, signal actuation
and coordination. Transp. Res. Part C 46, 65–82 (2014)
5. Leistner, D.: Impact of cooperative systems, University of Dresden, Germany (2009)
6. Uffmann, A., Friedrich, B.: Online optimization within cooperative systems in urban road
networks. In: 13th World Conference on Transport Research, Rio de Janario, Brazil (2013)
7. Vujić, M., Šemanjski, I., Vidan, P.: Improving energy efficiency by advanced traffic control
system. Trans. Marit. Sci. 02, 119–126 (2015)
8. Stevens, M., Yeh, C.: ReinforcementLearning for Traffic Optimization. Stanford.edu
9. PTV AG: PTV 6 User Manual. PTV, Karlsruhe (2014)
10. Eijk, A., Ligterink, N., Inanc, S.: EnViVer 4.0 Pro and Enterprise-Manual (2014)
11. Group of Authors: Highway Capacity Manual. Transportation Research Board (2000)
12. Hunt, P.B., Roberetson, D.I., Bretherton, R.D., Winton, R.I.: A traffic responsive method of
coordinating signals. Transport and Road Research Laboratory, Crowthorne (1981)
13. Group of Authors: How to Solve the 15 Toughest Traffic Management Challenges with in
Sync Adaptive Traffic Signal Control, Rhythm Engineering (2013)
14. Bošnjak, I.: Intelligent transport systems 1, Faculty of Traffic and Transport Sciences, Zagreb
(2006). (in Croatian)
15. Mandžuka, S., Ivanjko, E., Vujić, M., Škorput, P., Gregurić, M.: The use of cooperative ITS
in urban traffic management. In: Perallos, A., Hernandez-Jayo, U., Onieva, E., García-
Zuazola, I.J. (eds.) Intelligent Transport Systems: Technologies and Applications. Wiley,
New York (2015)
Application of Mini Computers and RFID
Technology in Automation
Malik Čabaravdić ✉ , Sanela Čančar, and Anel Husaković

( )

mcabaravdic@mf.unze.ba, sanela.cancar@hotmail.com,
anelhusakovic88@gmail.com
Abstract. The advancement of technology is increasing the need for its appli‐
cation in everyday life. RF technologies one of the interesting area and is widley
used in Europe but also in Bosnia and Herzegowina. Point of this paper is to
investigate usage, abilities and limitations of this contactless technology by
simple access control of parking lot via RFIDtechnology and Arduino mini
computer.
Keywords: RF · RFID · Single-board computers · Mini-computers · Arduino

Parking lot
1 Introduction
Under the term “access control” is understood any system that controls access to a
particular object or space and, depending on the purpose of the system, issues permis‐
sions to enter the facility or monitors the passage of physical persons or vehicles through
the access point of the space. Evaluation of this control is defined from “what I have” -
a card, over “what I know” - PIN to “what I am” - biometric data. This paper discusses
the use of RFID (radio frequency identification) cards.
2 Access Control System
Within this paper, an example of access control system in the parking lots is shown,
using the RFID card, RFID reader and Arduino mini computer.
2.1 The Principle of Work of the Realized Parking Access Control System
Each parking user must have a contactless card. The card used in this example is passive,
operates at a frequency range of 125 kHz and is being read by 125 kHz RF readers.
Using serial communication the data is forwarded to Arduino UNO controller. Arduino
processes the information received and controls the other periphery, depending on the
user’s authorization. In the case that user ID is authorized to enter the parking lot, the
ramp will open and the text on the LCD screen will be displayed. Near the ramp, an
ultrasonic distance sensor is positioned, which detects the object at a certain distance.

https://doi.org/10.1007/978-3-319-90893-9_42
Application of Mini Computers and RFID Technology in Automation 355
After passing through the ramp, for safety reasons, it is detected whether the user is near
the ramp via the ultrasonic sensor and if it is not the case, the ramp is being closed. In
the case that the user is close to the ramp, the ramp remains open.
2.2 About RFID Technology

The most important RFID components are tag (transponder) and reader (intergator). Tag
is an identification device (we used a card) which contains an integrated antenna and an
integrated circuit with a unique identification number, which allows contactless
exchange data with the RFID reader, at a certain frequency. The reader is a device that
communicates with tags, at a certain frequency, and can write or read information,
depending on the type of RF technology and the frequency of operation. Tags are divided
into passive, semi-passive and active. Passive tags are RFID tags that do not have their
own power source, but the energy is received by the RFID reader through the magnet‐
ization of tag’s antenna. These tags are the smallest, the cheapest, but also have weakest
range. The storage capacity of the data is the smallest and they are very sensitive to the
electromagneticnoise of the environment. Semi-passive tags use batteries for the oper‐
ation of the integrated circuit, but the communication is realized using the energy of the
RFID reader. Active tags have batteries, which are used to operate the integral circuit
and transmit signals to the reader. Active readers are used for bigger distances, but are
also the most expensive [3] (Fig. 1).
Fig. 1. RFID transponder (antenna and integrated circuit) [3]
Also, there are several ways for communication of tags with a reader, which are
called air interface. Depending on this interface, the range of tags and the identity of the
reader that is compatible with the tag can be determined. The basic parameters of the
interface are energy source, operating frequency, communication mode, data entry,
coding information and matching.
Operating frequency is the electromagnetic frequency that the tag uses to commu‐
nicate or to receive energy. The electromagnetic spectrum in the RFID area is divided
into low frequency (LF), high frequency (HF), ultra high frequency (UHF) and micro
waves. RFID systems must not interfere with other, protected applications and users,
such as emergency radio or television signals. The frequencies in which they operate
are frequencies known as Industrial Scientific Medical - ISM (Tables 1 and 2).
356 M. Čabaravdić et al.
Table 1. RFID frequency range [3]

Name Frequency range ISM frequencies
LF 30300 kHz to 135 kHz
HF 330 MHz 6.78 MHz, 13.56 MHz,
27.125 MHz, 40.680 MHz
UHF 300 MHz – 3 GHz 433.920 MHz, 869 MHz,
915 MHz
Micro waves >3 GHz 2.45 GHz, 5.8 GHz, 24.125 GHz
Table 2. Range of the tags and applications [3]

Name Range Typical situations
LF 50 cm Identification of animals, products
HF 3m Building access control
UHF 9m Boxes and pallets
Micro waves >10 m Identification of vehicles
Readers communicate with tags. The readers contain an antenna for reading (has not
the same form as tag-antenna) and have connection to the data processing system. They
can be also used for writing of data on tags. The simplest readers allow the reading of
only one type of tag, using only one frequency and one protocol, while the more complex
ones use different protocols, allowing selection of data, checking and correcting errors.
The principle of the work is that the reader sends a radio wave request for reading or
writing to a tag, which then gives feedback - a unique identification code and/or a set of
data stored in an integral circuit. The received information is converted into a digital
data set. This information may contain information about the location of the product,
price, expiration date, etc. After collecting data from one or more tags, the reader sends
this data to the computer, and in that sense it represents an agent between the tag and
the information system. There are also problems related to interference. An example is
when a number of tags must be read in a short time. This problem is solved by the
Command Response Protocol, by which the reader sends the command to the tag to stop
transmission after repeated sending of signals. [4] (Fig. 2)
MAIN COMPUTER
RFID
RFIDTAG
RFID READER
RFID RFID INLAY

CHIP
Fig. 2. RFID working principle [4]

Application of Mini Computers and RFID Technology in Automation 357
2.3 Simulation of a Parking Lot Using RFID Technology

In order to test RFID technology, a simulation model for parking lots was made.
Furthermore, an Arduino connection scheme is given, with a servomotor running a ramp,
an ultrasonic sensor, which enables the vehicle’s safety after passing through the ramp,
an RFID reader, which registers ID tag, an LCD display for displaying input/output
messages and the message when the user is not authorized to pass (Fig. 3).
Fig. 3. Connection scheme
The model of the system is shown in Fig. 4.
Fig. 4. Model of the system

358 M. Čabaravdić et al.
3 Conclusion
Nowadays, it is almost impossible to imagine life without the technology that has been
used in all aspects of modern society. With the vast development of the IT industry,
there is an expansion of the offer so that products become accessible to a wider range
of people. Over the last few years, technology seeks to make computers as small and
cheaper as possible, while trying to provide as many opportunities as possible. One of
these is the Arduino computer, which is characterized by its small dimensions and great
capabilities at a very affordable price [1, 2].
With its hardware and free integrated development environment, Arduino provides
users with a simple interaction with the environment. Arduino is also a good choice for
connecting sensors and other devices because of having already finished libraries that
make it easy to connect and use them. The advantage of using the Arduino interface is
the possibility of further refinement and improvement. The advantage of working with
the Arduino interface is the availability and low price of the components. This paper
presents RFID technology whose future potential is great and opens doors to numerous
applications in real life.
The future of the RFID system will enable organizations that use RFID to improve
their business as well as exchange information between business partners according to
established standards through a large and well-secured network. All these components
are expected to be a part of the Internet of Things (IoT) in the future.
References
1. Muračević, A., Husaković, A., Čabaravdić, M.: Application of the mini computers in the
automation. In: Proceedings of “TMT 2016” Conference, Mediterranean Sea Cruising, 24
September–1 October 2016
2. Zenzerović, P.: Arduino kroz jednostavne primjere. Hrvatska zajednica tehničke kulture,
Zagreb (2013)
3. Shepard, S., Puglia, A., Puglia, M.: RFID: Radio Frequency Identification, 1st edn. Wiley
(2007)
4. Miles, S.B., Sarma, S.E., Williams, J.R.: RFID Technology and Applications, Reissue edn.
Cambridge University Press (2011)
Application of Unmanned Aerial Vehicles
in Logistic Processes
Jasmina Pašagić Škrinjar ✉ , Pero Škorput, and Martina Furdić

( )

{jpasagic,pero.skrput,mfurdic}@fpz.hr
Abstract. In order to respond to the challenges of modern society, the organi‐

zation of logistic processes requires synergistic optimization effects of physical
processes and application of innovative technologies. The emphasis is based on
continuous optimization of the processes with the aim of faster, more quality and
more cost-effective services to the end user. In this paper possibility and potential
of UAV (Unmanned aerial vehicle) application in logistics processes will be
explained. Processes as stock inventory in warehouses and others can be improved
with application of UAV.
Keywords: Unmanned aerial vehicle · Logistic processes

Warehouse operation
1 Introduction
The UAV is an acronym for Unmanned Aerial Vehicle which is defined as an aerial
vehicle that does not carry a human operator, uses aerodynamic forces to provide vehicle
lift, can fly autonomously or be piloted remotely, can be expandable or recoverable, and
can carry a lethal or nonlethal payload [6]. It is controlled either autonomously on
computer based on pre-programmed flight plans or more complex dynamic automation
systems [17] or by remote control of a pilot on the ground. Its usage is currently limited
by difficulties such as satellite communication and cost. A drone that can be operated
by radio frequency controller and send live audio-visual feedback has been built [6].
The first appearances of UAV were in the military, but modern trends and globalization
scattered UAV applications in the spectrum of commercial applications. UAVs have a
wide range of applications and models. They are categorized into three categories, which
are safety control, scientific research, and commercial applications [4]. In business,
companies such as Amazon and Facebook have been grabbing the headlines and the
public’s imagination with multimillion-dollar research into UAVs for same-day deliv‐
eries and to expand the internet to remote parts of the world. While these are a long way
from commercial reality, other industries are finding numerous uses for UAVs tech‐
nology – uses that could bring wider benefits to the environment and society, as well as
greater efficiencies [16]. Commercial UAVs are used in field of agriculture, construction,
transportation, traffic management, inspection, public safety and lots of other civil-
government uses. Except transportation, use of UAV can be applicated in different

https://doi.org/10.1007/978-3-319-90893-9_43
360 J. P. Škrinjar et al.
phases of supply chain, and this paper will refer to the use of UAV in various logistic
processes. In this review article we will explain wide range of UAV application. In
Sect. 2 we will review different fields of UAV use, and in Sect. 3 its use in the narrower
area of the logistics. In Sect. 4 we will give a conclusion to the explored topic.
2 Literature Review
In this section, we will present several articles related to UAV application. This section
will explain the application of UAV in various areas and how it can be transposed into
the logistic sector.
Mohammed et al. discuss the applications of unmanned aerial vehicles (UAVs) in
smart cities, their opportunities and their challenges in their paper [4]. They wrote about
various UAV applications in cities which include monitoring traffic flow to measuring
and detecting floods and natural disasters by using wireless sensors. They also explain
challenges and issues of UAV usage such as safety, privacy and ethical uses. Figure 1
shows a Law enforcement UAV used in police activities. They conclude that integrating
UAVs with smart cities will create a sustainable business environment and a peaceful
place of living. And that UAV systems and smart cities can significantly impact and
benefit any country when used effectively and efficiently.
Fig. 1. A Law enforcement UAV [4]
Patra and Sengupta [1] in their paper present mechanism to deploy and rearrange the
positions of the UAVs considering the irregular user densities and requirements. Their
mechanism consists of an initial deployment scheme for the UAVs and the UDPR algo‐
rithm to enable the UAVs to move according to the heavy demands of the users. UDPR
algorithm calls either of the two approaches, the primitive Juxtapositioning or the
proposed Hot Center Positioning, as a method for calculating the distance that the UAVs
must travel to serve the users in high density subregions. The former approach, as is
rudimentary, expects the UAVs to attain a position closest possible to the UAV in need
of assistance. Whereas, the latter approach meticulously calculates the effective position
which can be attained by traveling the minimum distance so that the requesting UAV
can be effectively assisted and at the same time least amount of energy will be consumed
in traveling the distance. They simulate their approaches using C++ language. The
Application of Unmanned Aerial Vehicles in Logistic Processes 361
simulation results shown that the proposed Hot Center Positioning approach outper‐
forms the primitive Juxtapositioning approach by covering more area (serving more
number of users) and at the same time keeping more number of UAVs active.
Science, technology and the future of small UAVs is reviewed by Floreano and Wood
[2]. They explained the important socio-economic impacts of small UAVs such as fixed-
wing UAVs with a long flight time which could provide bird’s-eye-view images and a
communication network for rescuers on the ground, rotorcrafts with hovering capabil‐
ities which could inspect structures for cracks and leaks; UAVs for transport od medical
supplies from nearby hospitals, swarms of dispensable UAVs with flapping wings which
could enter buildings to search for chemical hazards and multi-modal caged robots with
purpose to fly and roll into complex structures to safely search for signs of life. They
shown the advantages of small drones and the future possibilities for their implemen‐
tation in various fields.
In their paper, Maza et al. [5] described a multi-UAV distributed decisional archi‐
tecture developed in the framework of the AWARE Project together with a set of tests
with real UAVs and Wireless Sensor Networks (WSNs) to validate this approach in
disaster management and civil security applications. Their paper present the different
components of the AWARE platform and the scenario in which the multi-UAV missions
were carried out. The missions described in their paper included surveillance with
multiple UAVs, sensor deployment and fire threat confirmation. In order to avoid redun‐
dancies, instead of describing the operation of the full architecture for every mission,
only non-overlapping aspects are highlighted in each one. Key issues in multi-UAV
systems such as distributed task allocation, conflict resolution and plan refining were
solved in the execution of the missions.
With the use of Unmanned Aerial Vehicles negative impacts of forest fire can be
significantly reduced. This field of application was explored in paper “The use of
Unmanned Aerial Vehicles for forest fire monitoring” [6]. Through research it was
shown how UAVs can contribute to reducing probability of errors made by tactics on
the ground and in air, reaction time, accuracy in decision making and load of people and
equipment in peak days. Review of the challenges for using small UAVs was made for
the regulations in the Republic of Croatia, also with the detail described proposed system
architecture including module for communication, data receiving module, video play
module, fire detection module and GIS display module. When you take into account the
state of development and equipment of other countries in relation to the Republic of
Croatia, it can be concluded that the current system control forest fires should be supple‐
mented with additional systems for decision support to enhance the existing system. In
support of these efforts, application of UAVs significantly contributes. Such systems
can contribute to reducing the probability of errors, shortening reaction time, increasing
accuracy in decision making, and shortening load of people and techniques in peak days.
3 Application of Unmanned Aerial Vehicles in Logistic Processes
Unmanned Aerial Vehicles (UAVs) have received attention in the last decade because
of their low cost, small size, and programmable features. Drone delivery is one of the
most promising applications to deliver packages efficiently [10]. At the end of 2016
Amazon developed Prime Air - a delivery system designed to get packages to customers
in 30 min or less using unmanned aerial vehicles.
To qualify for 30-min UAV delivery, the order must be less than five pounds
(2.26 kg) and small enough to fit in the cargo box that the UAV will carry. The recipient
must also be within a 10-mile radius of a participating Amazon fulfillment center. In the
UK, Amazon is only permitted to operate its UAVs, which fly below 400 feet, during
daylight when there’s low wind and good visibility.
When Amazon first unveiled its UAV, dubbed the Prime Air Drone, it showed off
an octocopter with eight rotors. It weighed 25 kg and could carry up to 2.26 kg at speeds
of 80 kph. However, the latest UAVs designed by Amazon are smaller with no fixed
wings. They actually appeared very different from the original drones, which were more
massive with fixed wings [12]. One of the Amazon UAVs is showed on Fig. 2. In 2017
Amazon has received a patent for a shipping label that includes a built-in parachute, to
enable the delivery of packages by UAVs or other aerial vehicles [5]. They still didn’t
put it in use but this move announces the expansion of UAV use for commercial
purposes.
Fig. 2. Amazon Prime Air – UAV for package delivery [18]
France’s General Directorate for Civil Aviation (DGAC) has authorised logistic
company DPDgroup to operate a regular commercial UAV delivery service. According
to a statement posted on 15 December 2016 on the DPD website: “For the first time
worldwide, a regular commercial route makes it possible to deliver parcels by drone.
“Once a week, the DPDgroup drone travels between Saint-Maximin-La-Sainte-Baume
and Pourrières, in south of France (Provence region). The route makes it possible to
delivery items to a cut-off incubator of companies including a dozen start-ups specialized
in technology. Their parcels now delivered by Chronopost and DPD France will be sent
by drone from the Pick-up point in Saint-Maximin.”
DPD has been testing drone deliveries for some time – and it uses a model where
the drone deposits and collects parcels in a “delivery terminal” which “protects access
to the drone during landing and take-off phases” [11].
In partnership with the French start-up company Atechsys, DPD has worked on this
project for two years (Parcel Delivery 2.0), within which 500 trial flights were
performed. The UAV delivery service has developed to meet specific needs such as
supply of isolated areas, emergency delivery, and so on.
The DPD drone radius is 25 km, its capacity is max. 1.5 kg of cargo and its maximum
speed is 30 km/h. It is consisted of six electric rotors and is equipped with a flight recorder
and live streaming camera. It also has a built-in GPS system so it can track its route in
time. There is also an interest in safety; an autonomous parachute involving in cases of
deviations from normal movements [14].
Kim and Matson [10] in their paper shown a proposal to solve problems with drone’s
limited capacity and battery life by collaborating a drone delivery system with existing
public transportation. Their study allocated the tasks to the UAVs and buses in the
context of the multi-agent delivery system. Also, this authors found a path for each
package by solving the vehicle routing problem (VRP) to find the cost-optimized path
given the heterogeneous multi-agent system and minimize the number of UAVs needed
for deliver. The experimental results in paper showed that the routing algorithm will
reduce the total mileage and the number of UAVs given the same set of orders.
In their paper, Beul et al. [9] describe micro aerial vehicles (MAVs), such as multi‐
rotors, which are envisioned for autonomous inventory-taking in large warehouses.
Fully autonomous operation of MAVs in such complex 3D environments requires real-
time state estimation, obstacle detection, mapping, and navigation planning. To this end,
we employ a cognitive MAV equipped with multiple sensors including a dual 3D laser
scanner, three stereo camera pairs, an IMU, an RFID reader, and a powerful onboard
computer running the ROS middleware. Tasks with hard real-time requirements such
as attitude control and state estimation are processed on a Pixhawk Autopilot, which
communicates with the main computer via the MAVLink protocol. The authors
described integrated system for autonomous MAV-based inventory in warehouses and
detailed the involved components and evaluated their system with the real autonomous
MAV in a realistic scenario.
Harik et al. [3] in their paper present a novel warehouse inventory scheme towards
the automation of the warehouse inventory task. The main purpose of work is to make
the inventory process completely autonomous. They used UAV to replace the manual
inventory of the available stock. They put an Unmanned Ground Vehicle (UGV) and an
Unmanned Aerial Vehicle (UAV) to work cooperatively. The UGV is used as the
carrying platform, and considered as a ground reference for the indoor flight of the UAV.
While the UAV is used as the mobile scanner which flies vertically to scan Barcodes
using the onboard frontal scanner. The UGV navigates among rows of racks carrying
the UAV. At each rack to be scanned, the UGV stops, and the UAV takes off to fly
vertically scanning goods in that rack. Once the UAV at the top, the UGV moves to the
next rack, and since the UAV takes the UGV as the ground reference, it will follow it
autonomously, which results in placing the UAV at the top of the second rack, and
scanning goods from top to bottom starts. The process is repeated until the row of racks
is fully scanned. The UAV then lands on the UGV, and recharge its batteries while the
UGV moves to the next row of racks. They present in their paper the proposed archi‐
tecture and the first experimental results of the proposed scheme. The autonomous
warehouse inventory global architecture defined in this paper is shown on Fig. 3.
Fig. 3. Autonomous warehouse inventory global architecture [3]
They developed autonomous flight controller and shown the efficiency of the results
compared to a previous work. Indeed, the warehouse inventory requires more work. For
now the UGV navigation is based on the reflection of a flat wall for the lidar sensors,
which is not the case with racks, where holes and empty spaces are present in racks, thus
further development needs to be done in this direction. The UAV may also need to be
equipped with sensors to avoid collision with the racks during the scanning process [3].
In their paper, Bae et al. [7] proposed new approach to investigate open storage yard
using unmanned aerial vehicle and RFID. It reduces cost of inventory checking and
mismatch of real world and cyber world. They also implemented a prototype system to
show the feasibility. RFID enables identification from a distance, and unlike earlier
barcode technology [8]. Many companies adopt RFID technology to monitor flow of
factory and manage the inventory [7].
The proposed inventory checking system based on drone and RFID is shown in
Fig. 4. It is composed of three major components. The first is a RFID reader that detects
RFID tags of products distributed outdoor inventory yard. It stores the detected tag id,
Fig. 4. Proposed approach for inventory checking [7]

location, time, and product count. The second is a drone that navigates whole outdoor
yard manually or automatically. The last is server program for inventory checking that
compares database values and real product data.
They used C# as programming language and the .net framework was also used for
the program. They developed a data collection program to detect and save tag informa‐
tion. After the flight, the gathered tag data is transferred to the inventory checking server
and is compared with the inventory data stored in database. The server program is
developed using java and MySQL server. It lists all product of inventory database and
compares each states defined in above. Abnormal states like location error, missing
product or unregistered were highlighted in inventory checking system. The authors [7]
proposed new approach for inventory checking in open stock yard using UAV and RFID
technology. They concluded it will decrease the mismatch between information stored
in inventory system and real world data, also reduce labor cost of investigating stock
and equipment cost for the system.
UAVs are also useful farther up the construction supply chain, in mining and aggre‐
gates. Working out how much material is sitting in a stockpile in a mine or quarry usually
involves taking a few dozen measurements with manual surveying equipment and then
calculating the volume. UAV can measure the volume of dozens of stockpiles in a single
flight, taking thousands of measurements that are converted into an accurate point cloud
within an hour. As well as being far quicker and more accurate, it is also much safer.
Falling off stockpiles is one of the industry’s biggest occupational hazards. Using UAVs
to survey quarries and building sites also means that human surveyors do not need to
venture close to dangerous sheer drops [13].
4 Conclusion
Throughout this paper a review of various UAV applications was made. Over the last
few years, apart from the use of UAVs that are piloted remotely, UAVs that fly inde‐
pendently based on pre-programmed flight plans or more complex dynamic automation
systems have been developed. Given the ever increasing demands of the users, but also
the increasing competition in the field of logistic processes and general supply chains,
it is necessary to improve the current capabilities by applying contemporary technolo‐
gies such as expert systems, which include UAV.
Companies Amazon and DPD are already using UAV for packet delivery, but UAVs
can also be useful in warehouse operations, which is shown through analysis of several
papers.
The use of UAV primarily results in optimization of logistic processes, with the aim
of reducing inventory costs, significantly shortening the process, reducing use of human
resources, and so on. It can be concluded that the use of UAV, apart from optimization
effects, also has an effect on increasing worker safety, increase the quality of work done,
reducing errors and else. Today, the UAV application in logistic processes such as
delivery and warehouse operations has just begun, which means that in the future we
can expect a broad application and implementation of expert systems in supply chain
operations.
References
1. Patra, A., Sengupta, S.: Dynamic deployment of UAV-enabled floating access points for
serving hot zones. In: 2017 International Symposium on Performance Evaluation of Computer
and Telecommunication Systems (SPECTS) (2017)
2. Floreano, D., Wood, R.J.: Science, technology and the future of small autonomous drones.
Nature 521(7553), 460–466 (2015)
3. Harik, E.H.C., et al.: Towards an autonomous warehouse inventory scheme. In: 2016 IEEE
Symposium Series on Computational Intelligence (2016)
4. Mohammed, F., Idries, A., Mohamed, N., Al-Jaroodi, J., Jawhar, I.: UAVs for smart cities:
opportunities and challenges. In: International Conference on Unmanned Aircraft Systems
(ICUAS), pp. 267–273 (2014)
5. Maza, I., Caballero, F., Capitán, J., Martínez-de-Dios, J.R., Ollero, A.: Experimental results
in multi-UAV coordination for disaster management and civil security applications. J. Intell.
Rob. Syst. 61(1), 563–585 (2011)
6. Škorput, P., Mandžuka, S., Vojvodić, H.: The use of unmanned aerial vehicles for forest fire
monitoring. In: Proceedings of 58th International Symposium ELMAR - 2016, pp. 93–96
(2016)
7. Bae, S.M., Han, K.H., Cha, C.N., Lee, H.Y.: Development of inventory checking system
based on UAV and RFID in open storage yard. In: 2016 International Conference on
Information Science and Security (ICISS), pp. 1–2 (2016)
8. Want, R.: An introduction to RFID technology. IEEE Pervasive Comput. 5(1), 25–33 (2006)
9. Beul, M., Krombach, N., Nieuwenhuisen, M., Droeschel, D., Behnke, S.: Autonomous
navigation in a warehouse with a cognitive micro aerial vehicle. In: Koubaa, A. (ed.) Robot
Operating System (ROS). Studies in Computational Intelligence, vol. 707. Springer, Cham
(2017)
10. Kim, M., Matson, E.T.: A cost-optimization model in multi-agent system routing for drone
delivery. In: Bajo, J., et al. (eds.) Highlights of Practical Applications of Cyber-Physical
Multi-Agent Systems, PAAMS 2017. Communications in Computer and Information
Science, vol. 722. Springer, Cham (2017)
Others
11. postandparcel.info
12. www.amazon.com
13. www.economist.com
14. www.fpz.unizg.hr/prom
15. www.pocket-lint.com
16. www.theguardian.com
17. www.theuav.com
18. www.theverge.com
New Technologies in the Field Energy:
Renewable Energy, Power Quality,
Advanced Electrical Power Systems
The Turbulence Intensity of the Wind Bora
Blago Pehar ✉ , Elvir Zlomušica, and Suad Zalihić

( )
‘‘Džemal Bijedić’’ University of Mostar, 88000 Mostar, Bosnia and Herzegovina

blago1982pehar@gmail.com,
{elvir.zlomusica,suad.zalihic}@unmo.ba
Abstract. The Bora is a cold and gusty wind which occurs along the eastern coast
of the Adriatic Sea and its hinterland. It is most commonly from the north or north-
east direction and often achieves stormy strength. Due to the direct connection of
the Bora and the complex terrain, the speed of the wind Bora is rapidly changing
while crossing obstacles and the turbulent flow occurs. For the classification of
wind turbines the wind turbulence intensity with the wind speed are the basic
parameters. The measurements of the Bora characteristics were made at Hrgud
and Morine in the area of Eastern Herzegovina on an extremely complex terrain.
The standardized measurement equipment was used which is mounted on a meas‐
uring mast at the height of 78 m. The measured values of the turbulence intensity
of the Bora, have been analyzed and compared with the corresponding standards
and recommendations.
Keywords: Turbulence intensity · Complex terrain · Wind Bora
1 Introduction
According to climatological research, the stormy Bora is formed by the flowing of cold air
formed in the deep elevated valley characterized by a large amplitude and a small wave‐
length [1–4]. The wind Bora is a strong cold and katabatic wind which mostly blows from
north to north–east, starts suddenly and decelerates slowly. It occurs most frequently during
the winter season with a duration of several hours to several days [5]. There are anti-
cyclonal (dry) and cyclonal (with clouds) types of Bora. Furthermore, there are several
conditions needed for its formation (mountain massifs, different values of temperature and
pressure in the highland and lowlands, etc.) [6, 7]. The Dinarids are characterized by
numerous transitions and peaks. Their three-dimensionality contributes to the observed
spatial variability of the wind speed and direction along the coast. The higher part of the
terrain provokes a stronger wave fracture than the surrounding terrain, so the hydraulic jump
is stronger downstream from the peaks. This results in the formation of areas of leeward
turbulence (wakes) due to the strong Bernoulli dissipation within the jump [7–10].
One of the most important parameters for wind turbine design is the turbulence inten‐
sity, represented by the ratio of the standard deviation and the mean wind speed averaged in
some time interval. Turbulence intensity is a measure of the gustiness of the wind. Turbu‐
lence is a flow regime characterized by chaotic, stochastic property changes. This includes

https://doi.org/10.1007/978-3-319-90893-9_44
370 B. Pehar et al.
a low momentum diffusion, high momentum convection, and a rapid variation of pressure
and velocity in space and time [9–11].
The aim of the research is to determine the intensity of the turbulence on the analyzed
locations for the classification of the wind turbines.
2 Method and Materials
The comparative measurements are performed over a period of more than four months. The
analyzed period for this study started July, 27th, 2012 and ended December, 10th, 2012.
2.1 Site Description
The measurement sites Hrgud and Morine in Bosnia and Herzegovina can be categorized
as complexed. The terrain is characterized by karsts with small meadows, bushes and a low
forest vegetation, typical for a submediterranean surronding. The measurement sites are
located approximately 35 km southeast (Hrgud) and 40 km east (Morine) of the city of
Mostar, Fig. 1.
Fig. 1. Locations of Hrgud and Morine in Bosnia and Herzegovina
Fig. 2. Position of met mast at Hrgud (left) and Morine (right) sites
The Turbulence Intensity of the Wind Bora 371
Surrounding terrain that features significant variations in topography and terrain obsta‐
cles which may cause flow distortion, as shown in Fig. 2.
2.2 Equipment and Measurement Configuration

The measuring equipment is mounted on a classical triangular lattice mast, 78 m high.
In accordance with the IEA standard [12], and IEC 61400-12-1 [13], compliant recom‐
mendations, the equipment was calibrated according to MASNET procedures [14, 15].
The wind characteristics (speed and direction) were measured with a sampling rate of
1 Hz, and the measured values were calculated at 10-min averaged intervals. Recalcu‐
lated values are associated with the corresponding statistical magnitude. Air tempera‐
ture, relative humidity and atmospheric pressure were additionally measured. The wind
speed was measured with the met mast, at the height of 30, 55, 75/77.5 m (cup anemom‐
eters Thies Classic) and the wind direction was measured at the height of 53 and 75 m
(wind vanes Thies Compact) Fig. 3.
Fig. 3. Met mast used in measurement campaign
The mast positions were selected to evaluate the wind potential of the locations and
there coordinates are shown in Table 1.
Table 1. Overview of coordinates specified for the two met masts

Location UTM WGS 84 zone 34 Gauss Krüger zone 6 (BiH)
X Y X Y
Hrgud 258 390 4 776 041 6 502 939 4 772 697
Morine 277 376 4 797 784 6 521 126 4 795 099
Data availability in the analyzed period was 99.32% for Hrgud and 99.79% for Morine,
which meets the standards and recommendations [12–15] on the 10-min wind parameters.
At both locations, data were obtained by an anemometer at 77.5 m and vane at 53 m.
According to the frequency roses, Bura was isolated between the sectors 0°–90°. Based on
372 B. Pehar et al.
these data, ambient characteristic turbulence intensity parameters were computed and then
compared with the turbulence classes according to IEC 61400-1 [16].
2.3 Assessment of the Ambient Turbulence Intensity

If Vave is the mean of wind speed variable and σv is the standard deviation of longitudinal
wind speed of measured data, then the measured 10-min turbulence intensity is the ratio
of wind speed standard deviation to the mean of wind speed [16, 17].
According to IEC 61400-1 [16] there are four different classes of wind turbines to
suit differing site wind conditions. The wind speed and turbulence parameters for each
class are given in Table 2.
Table 2. Basic parameters for wind turbine classes [16]

Wind turbine class I II III S
Vave (m/s) 10 8.5 7.5 Values
Vref (m/s) 50 42.5 37.5 specified
by the
V50,gust (m/s) 70 59.5 52.5
designer
A Iref 0.16
B Iref 0.14
C Iref 0.12
In Table 2, Rayleigh distribution is assumed, i.e. k = 2. Vave is the annual mean wind
speed at hub height, Vref is the 50-year extreme wind speed over 10 min, V50,gust is the
50-year extreme gust over 3 s, Iref is the mean turbulence intensity at 15 m/s. A, B and
C are the categories of higher, medium and lower turbulence intensity characteristics
respectively [16].
Expected standard deviation σ1 is described based on an approximation of the 90th
percentile of the standard deviation of the longitudinal wind speed Vhub by Eq. 1:
( )
𝜎1 = Iref 0.75Vhub + b ; b = 5.6 m∕s (1)
Representative ambient turbulence Iref, for a set of 10-min time interval, is equal to
the 90th percentile of the TI values. Assuming a normal distribution of these values, it
represents the mean value plus 1.28 times standard deviations, Eq. 2:
TI90 = TIave + 1.28 ⋅ 𝜎TI (2)
Analyzing the wind directions, Fig. 4, it is proved that Bora blows from north-north‐
east (NNE) and east-northeast (ENE) with dominant directions 30° and 40° at Hrgud
site. At Morine site Bora blows from north-northeast (NNE) with the dominant direc‐
tions 10° and 20°, which corresponds to the researches [7–10, 18–20]. The percentage
of the north direction is 62.66% at Hrgud site and 67.34% at Morine site.
N N
20,0 20,0
15,0 15,0
10,0 10,0
5,0 5,0
W 0,0 E W 0,0 E
S S
Fig. 4. Frequency roses at 53 m at Hrgud site (left) and at Morine site (right)
The highest mean wind speed of 9.50 m/s at Hrgud site and 9.60 m/s at Morine site
corresponds to the dominant directions, Fig. 5.
N N
12,0 12,0
8,0 8,0
4,0 4,0
W 0,0 E W 0,0 E
S S
Fig. 5. The mean wind speeds roses at 77.5 m high, Hrgud site (left) and Morine site (right)
The representative turbulence intensity I15 according to IEC 61400-1 [16] is defined
as hub height turbulence intensity at a 10-min averaged wind speed of 15 m/s. It is
determined as the sum of mean value of the turbulence intensity and the standard devi‐
ation, taking into account only the values with wind speed exceeding 10 m/s. The turbu‐
lence intensity has been calculated for the hub height at the reference points. For the
normal turbulence model, the representative value of the turbulence standard deviation,
σ1, shall be given by the 90% quantile for the given hub height wind speed, Table 3 and
Fig. 6.
374 B. Pehar et al.
Table 3. Ambient representative turbulence intensity at reference points at Hrgud and Morine
sites
Position Hrgud Morine
Height above ground 77.5 m 77.5 m
Average turbulence for all wind speeds 18.71% 16.96%
Average turbulence for wind speeds > 4 m/s 15.31% 12.46%
Average turbulence for wind speeds > 10 m/s 14.00% 8.91%
Binned standard deviation of turbulence for wind speeds > 10 m/s 2.95% 2.22%
Representative ambient turbulence (I15) for wind speed > 10 m/s acc. 17.78% 11.75%
IEC 61400-1 [16]
30%
25%
Ambient representative TI
20%
15% Class A
Class B
10% Class C
Morine
5% Hrgud
0%
0 2 4 6 8 10 12 14 16 18 20 22 24
Wind speed m/s
Fig. 6. Ambient representative turbulence intensity in comparison to the turbulence classes of

IEC 61400-1 [16], at the reference points
The ambient characteristic turbulence intensity at the site does not meet the require‐
ments of IEC 61400-1 [16] turbulence classes C, B as well as A at Hrgud site.
The ambient characteristic turbulence intensity at the site meets the requirements of
IEC 61400-1 [16] turbulence classes C as well as B and A at Morine site.
4 Conclusion
The turbulence at higher wind speed is of the most important interest for the wind turbine
manufacturers. Analyzing the measured data, the intensity of turbulences decreases at
high speed, except for speeds 10–12 m/s at Hrgud site where we can see slight increasing.
According to the IEC 6100-1 [16], the turbulence intensity at Hrgud site does not corre‐
sponds the requirements turbulence classes C, B as well as A. While the turbulence
intensity at Morine site corresponds class as well as B and A.
Due to significant differences in the class at sites, which may be conditioned by the
complexity of the terrain and/or short period of analysis, it is necessary to extend the
periods of analysis.
This study, the same as the other in the region [5–10, 18–20], confirms unpredicta‐
bility of the wind Bora. Thus it is essential to continue the measurement through longer
period of time and on different locations.
Acknowledgement. The authors are grateful for the considerable technical support from the staff
of the Center for Renewable Energy Sources Impro Impex Mostar.
References
1. Poje, D.: Bura (Bora) and burin at Split. Croat. Meteorol. J. 30, 1–19 (1995)
2. Vučetić, V.: Severe Bora on the mid-Adriatic. Croat. Meteorol. J. 28, 19–36 (1993)
3. Horvath, K., Ivatek-Šahdan, S., Ivančan-Picek, B., Grubišić, V.: Evolution and structure of
two severe cyclonic Bora events: Contrast between the northern and southern Adriatic.
Weather Forecast. 24, 946–964 (2009)
4. Jurčec, V.: On mesoscale characteristics of Bora conditions in Yugoslavia. Pure. appl.
Geophys. 119(3), 640–657 (1981)
5. Večenaj, Ž., Belušić, D., Grisogono, B.: Along-coast features of Bora related turbulence.
Bound.-Layer Meteorol. 143(3), 527–545 (2012)
6. Bourgeois, S., Winkelmeier, H., Meissner, C.: Turbulence intensity and high wind speeds
above complex terrain: measurements and CFD-modelling. In: DEWEK, 10th German Wind
Energy Conference (2010)
7. Zlomušica, E., Zalihić, S., Bejdić, J.: Comparison of wind measurements by LIDAR and
measurement mast for Bora wind in Bosnia and Herzegovina. In: The 5th International
Symposium on Sustainable Development, Sarajevo (2014)
8. Zlomušica, E.: Particular review on SODAR and LIDAR measurements of Bora wind
in Mostar, Bosnia and Herzegovina. Int. J. Eng. Technol. IJET-IJENS 13(06), 53–61 (2013)
9. Zlomušica, E., Zalihić, S., Čampara, M., Zaimović, R.: Using of remote sensing methods in
wind analysis in conditions of local wind Bora and complexity of terrain. In: The International
symposium Innovative and Interdisciplinary Applications of Advanced Technologies (IAT)–
ISAME (2017)
10. Zalihić, S.: Using of remote sensing methods in analysis of vertical wind profile on complex
terrains in conditions of local wind Bora, Ph.D thesis, ''Džemal Bijedić'' University of Mostar,
Faculty of Civil engineering (2017)
11. Burton, T., Sharpe, D., Jenkins, N., Bossanyi, E.: Wind Energy Handbook. Wiley, West
Sussex (2001)
12. IEA: IEA Recommended practices for wind turbine testing and evaluation 11. Wind Speed
Measurement and use of cup anemometry, 1st edn. (2003)
13. International Standard, IEC 61400-12-1 Wind turbines - Part 12-1: Power performance
measurements of electricity producing wind turbines, 1st edn. (2005)
14. MEASNET, Cup Anemometer Calibration Procedure - Version 2 (2009)
15. MEASNET, Evaluation of Site-Specific Wind Conditions - Version 2 (2016)
376 B. Pehar et al.
16. International Standard, IEC 61400-1 Wind turbines - Part 1: Design requirements, 3th edn.
(2005)
17. International Standard, IEC 61400-1 Wind turbine generator systems - Part 1: Safety
Requirements, 2nd edn. (1998)
18. DEWI, Assessment of the site wind conditions regarding IEC wind turbine classes at the site
Pag/Croatia, DEWI GmbH, Wilhelmshaven (2007)
19. DEWI, Assessment of the site wind conditions regarding IEC wind turbine classes at the site
Rudine/Croatia, DEWI GmbH, Wilhelmshaven (2010)
20. DEWI, Site-related Wind Potential Analysis and Energy Yield Assessment at the Site
Podvelez, DEWI GmbH, Wilhelmshaven (2009)
Lab-Scale Tests as Support to Selection
of Sustainable Coal Combustion Technology
- Case Study: Support to Design
of TPP Kakanj Unit 8 -
Nihad Hodžić ✉ , Sadjit Metović, and Anes Kazagić

( )

{hodzic,metovic}@mef.unsa.ba, a.kazagic@epbih.ba
Abstract. In the Laboratory for coal and biomass combustion at the Mechanical
Engineering Faculty - University of Sarajevo, extensive research was carried out
on the characteristics of combustion of mixtures of coals from Middle Bosnian
basin. The aim of the research was to obtain reliable data for further analysis and
optimization of the combustion system relied on these coal blends, and the answer
to the question of whether it is possible to efficiently burn these coals in boilers
with dry ash bottom. Combustion of various coal blends from coal mines Kakanj,
Breza and Zenica were investigated at various temperatures (950–1350 °C) and
under various technological conditions in the combustion chamber. The following
parameters were varied: fraction of individual coals in the mixture, combustion
temperature, thermal load of the combustion chamber, air distribution including
the combustion air staging system. During the study the composition of flue gas,
in particular CO2, CO, NOx and SO2 emissions, then combustion efficiency, solid
ash sample characteristics, reactivity of CaO and binding of sulfur to alkaline
compounds of the coal ash were analyzed. Summary of the results of this research
is given in this paper. It has been shown that high efficiency combustion can be
established under certain conditions. It is also estimated that pulverized coal
combustion with dry ash bottom furnace is acceptable and sustainable technology
for treated coals/blends.
Keywords: Coal · Combustion · Emissions · Boiler · Furnace · Solid ash

Optimization
1 Introduction
All boilers in TPP Kakanj are based on pulverized combustion technology with slag
tap furnace. They are designed for firing coal from Middle Bosnian mine basin:
Kakanj (K), Breza (B) and Zenica (Z) in a mass ratio of 70:20:10, respectively -
K70B20Z10 blend. Applied combustion technology is characterized by high temper‐
ature and high specific heat loads in the combustion chamber - particularly in the
melting chamber and the burners’ zone, high emissions of pollutant components (e.g.
NOx and SO2), intense fouling of boiler surfaces with frequent slagging, and

https://doi.org/10.1007/978-3-319-90893-9_45
378 N. Hodžić et al.
consequent cracking of boiler tubes - first of all evaporator pipes and superheaters.
The choice of this combustion technology was made due to the composition and ash
characteristics of the coals, i.e. the ash melting temperatures. Thereby, coals from the
mines Breza and Zenica fall into group of coals with low ash melting temperatures.
The subject of laboratory research was the behavior of various mixtures of coals
from Kakanj, Breza and Zenica under conditions that correspond to pulverized combus‐
tion technology with dry bottom furnace. In this regard, test schemes have been set up
to suit the conditions of the real operating conditions of the Large Boiler with this
combustion technology. This is primarily reflected through the process temperature
values in the combustion chamber at various thermal loads of the boiler and through the
amount and manner of introducing and distribution combustion air into the combustion
chamber. Basically, four different temperature regimes for the Lab-scale furnace (called
Automatic Controlled Reactor - AUCR), are selected: 950, 1150, 1250 and 1350 °C.
The aim of the research was to answer the critical question on the basis of the tests and
results obtained: Is the pulverized coal combustion technology with dry bottom furnace
applicable for treated mixtures of coals - especially from technical, but also economic
and environmental aspects? The obtained set of results has provided, among other things,
the evaluation of the ability of slag and ash to be discharged from the combustion
chamber in dry state, the ash/deposit deposition intensity, the propensity of ash to slag‐
ging etc. To answer the aforementioned questions, laboratory tests basically included
the change of: (a) coal mixture composition: the fraction of Kakanj coal in the blend as
a dominant was varied from 60 to 80% w, (b) the process temperature in the interval
corresponding to the relevant combustion technology, (c) the excess air ratio with basic
total air supply to the combustion chamber (λ = 1.15), including test regimes with air
staging (OFA), e.g. λ1/λ = 0.95/1.15 [1–3].
2 Lab-Scale Furnace
AUCR experimental line is shown in Fig. 1. In the AUCR it is possible to burn pulverized
coal and/or waste woody biomass under various temperatures and techno-logical condi‐
tions. The facility is designed to operate at wide temperature range (up to 1560 °C) and
to supply different quantities and distributions of base fuel and combustion air, including
the possibility of testing of reburning combustion technologies using additional fuel,
e.g. natural gas. Output data obtained from the tests are related to combustion efficiency,
deposition intensity and slag/ash deposition characteristics, as well as emissions of flue
gas components: CO, O2, CO2, NO, NO2, NOx and SO2. For the purpose of analysis, the
ash samples deposited during the tests are excluded from the reactor by a water cooled
probe which can be brought by a slider to a pre-selected position within the reaction
tube. Ash samples are also taken from ceramic probes attached to the probe through a
special ceramic console. To determine the slag deposition rate as well as the slagging/
fouling characteristics, five indicators are used for evaluation of the ash samples taken
from the reactor: shape, melting condition, structure and the degree of adhesion and
cohesion. To provide additional information on ash deposits, the chemical composition
of ash deposits is determined and Ash Fusion Test (AFT) is performed. The chemical
Lab-Scale Tests as Support to Selection of Sustainable Coal Combustion Technology 379
composition of ash deposits allows the determination of redistribution rate of the ash
components which is important input data for boiler design. Additionally, the fraction
of unburnt in the ash is determined, which determines the combustion efficiency for the
given fuel under various conditions of combustion, considering also fuel particle size
distribution [4, 7, 8].
Fig. 1. (a) AUCR experimental line: front view of four visible zones with temperature
measurement of the reaction tube and process control; (b) pulverized fuel tank, dosing and supply;
(c) worktable with measurement equipment during the test
During the test regimes, the process temperature, fuel consumption, air flow
(primary, secondary/tertiary and OFA) are measured and the flue gas composition (O2,
CO, CO2, NO, NO2, NOx and SO2) is continually analyzed. In addition, sampling of
different deposits, slags and ashes from the pre-selected test modes and sampling posi‐
tions allows the determination of the burnout rate. Based on the measured and analyzed
samples of the flue gas, the properties of the combustion process are determined: slag‐
ging/fouling/abrasion of boiler heating surfaces, including the dependency of the CaO
reactivity of the mineral part of the fuel from ambient conditions; emissions of CO,
CO2, NOx and SO2 under conditions with and without air staging; and combustion
efficiency. Based on the results of the measurements, a database of combustion charac‐
teristics for the given fuel is established, on the basis of which reasonable recommen‐
dation and/or choice of sustainable combustion technology can be made, i.e. the choice
of the suitable combustion conditions for the given fuel consequently contributing to
selection of suitable furnace and boiler type and equipment [8].
3 Results of the Research
Basic properties of the test fuels: In the Table 1 the composition and heating values
of Kakanj, Breza and Zenica coals from which the test mixtures with different fractions
are formed, are given. There are significant differences in heating value, ash content and
sulfur content in component coals.
Table 1. Lower heating value and composition of component coals [7, 8]

Hd, kJ/kg C, % H, % S, % N, % O, % A, % W, %
Kakanj (K100) 11436 29,39 2,25 2,44 0,71 7,94 47,20 10,05
Breza (B100) 15867 43,55 2,38 1,62 0,78 11,01 28,28 12,39
Zenica (Z100) 19925 51,96 2,75 3,78 0,92 7,18 21,44 11,96
A particular difference is expressed in the ash fusion temperatures for these coals,
Fig. 2-left. These temperatures are significantly lower for the Breza and Zenica coals
than for the Kakanj coal and fall into easy-melting coals (t4B = 1220 °C, t4Z = 1200 °C).
The ash softening temperature (t2) from Breza and Zenica coal is 1170 °C and 1160 °C,
so a logical design requirement for the design of the units of TPP Kakanj is that the flue
gas temperature at the outlet of the furnace does not exceed tF2 ≤ 1150 °C. On the other
hand, ash from Kakanj coal has almost double the SiO2 content (53.82%) compared to
the ash from other two component coals and somewhat more pronounced content of
Al2O3 and K2O. In this case, the ash from Breza coal has the highest CaO content
(31.40%) and the ash from Zenica coal has the highest SO3 content (16.92%), Fig. 2-
right. One of the common criteria for assessing the propensity of coal ash to the slagging/
fouling the boiler heating surfaces is the base/acid ratio:
B CaO + Fe2 O3 + MgO + K2 O + Na2 O
= ; for B/K > 0.5 it is assumed that ash is prone
K SiO2 + Al2 O3 + TiO2
to the fouling of the heating surfaces - in this case are: (B/K)K = 0.32, (B/K)B = 1.11 and
(B/K)Z = 0.90. Another important criterion is the silicate value of the ash:
SiO2
SSiO2 = 100 on the basis of which the slagging tendency
SiO2 + Fe2 O3 + CaO + MgO
is estimated: for SSiO2 < 65 the slugging is high and for SSiO2 > 72 there is no slugging.
Based on this, it is concluded that the ash from Kakanj coal (SSiO2K = 72.15) has no
propensity to slugging - the value is at the limit, while the ashes from Breza and Zenica
coals fall into the high slagging category (SSiO2B = 37.9 and SSiO2Z = 43.9) [8].
Fig. 2. Characteristic ash temperatures (left); ash composition (right) for component coals
Keeping in mind the previous difference in the ash characteristics from the compo‐
nent coals, change in the ash composition proportional to the ash fraction in the
component coals in the formed coal mixtures is also noticed. This is particularly notice‐
able for the SiO2 content (43.8% in the mixture K60B25Z15 to 49.5% in the K80B15Z5
mixture) and CaO content (decrease from 14% to 9.7% with increases of Kakanj coal in
the blend). In this way, the ash fusion temperatures, as well as the tendency of ash from
the mixture to fouling of boiler heating surfaces changes. For example, with the increase
of the fraction of Kakanj coal in the blend the ash melting temperature (t4) increases,
i.e., the ratio of base and acid numbers (B/K < 0.5) decreases – the ash from coal
mixtures are not prone to fouling, Fig. 3-left. With the growth of the Kakanj coal fraction
in the blend, the silicate number of the ash increases, i.e. the tendency of the ash to
fouling decreases. For example, for coal ash from mixture containing 80% Kakanj coal
little slagging is expected compared to ash from mixtures with low Kakanj coal content
when strong slagging is expected, Fig. 3-right [4–6].
Fig. 3. Ash flow temperatures and base/acid ratio (left); silicate number of the ash (right), for
different coal mixtures
It is important to note here that during the laboratory research of coal and/or biomass
combustion at the Mechanical Engineering Faculty in Sarajevo, criterions were estab‐
lished for estimating the propensity of ash to the fouling of heating surfaces which, in
addition to the composition of ash, take into account both the process temperature of
combustion and other technical and technological settings of the process, e.g. the amount
and location of combustion air intake into the combustion chamber [3].
Emission of flue gas components: NOx, SO2, CO2 and CO emissions for combustion
of coal mixtures at different temperatures, various amounts and modes of combustion
air intake into the reaction zone. Thus, in Fig. 4 the emissions of these components for
the coal mixture K80B15Z5 are presented - the mixture with the lowest fraction of
Zenica coal. In the combustion regimes with OFA air supply, with increasing the process
3
temperature, the increase of NOx emissions is evident (from 388 mg/mn at 950 °C up to
3
755 mg/mn at 1350 °C). In the case of the primary air supply regime (λ1/λ = 1.15/1.15),
3 3
NOx emission was 601 mg/mn at 1150 °C, which is 72 mg/mn higher than the emission
at the same temperature with OFA air supply switched on, Fig. 4-left [1, 7, 8].
Fig. 4. Coal mixture K80B15Z5: NOx and SO2 emissions (left), and CO2 and CO emissions
(right) for different process temperatures and air supply
Contrary, there is no significant change in SO2 emission related to the distribution

3
of air supply at the same process temperature (difference is only 17 mg/mn ), Fig. 4-left.
CO emission drastically decreases with process temperature increase, where at 1150 °C
it is noticeably lower for the primary air supply (without OFA) compared to the test
regime with OFA switched on, Fig. 4-right.
Figure 5 shows emissions for the coal mixture K70B20Z10. As can be seen, observed
for one of the process temperatures, the lowest NOx emission is generally for the case
when air staging is applied (OFA switched on, λ1/λ = 0.95/1.15). NOx emission is higher
with primary air supply- when OFA is switched off. With increasing process tempera‐
ture, NOx emission is rising. When OFA is switched on, NOx emission is: min(tF1 =
3 3
950 °C) = 502 mg/mn , max(tF1 = 1350 °C) = 710 mg/mn . There is no significant change
in SO2 emission depending on air distribution, but on the other hand there is notable
increase in this emission with process temperature increases. Thereby SO2 emission
values are quite close one another regardless of the total amount of combustion air. In
3
these test regimes SO2 emission is: min(tF1 = 950 °C) = 2329 mg/mn , max(tF1 =
3
1350 °C) = 4063 mg/mn , Fig. 5-left [1, 7].
Fig. 5. Coal mixture K70B20Z10: NOx and SO2 (left) and CO2 and CO emissions (right) for
different process temperatures and air supply systems
On the other hand, the highest CO emissions were recorded at the lowest process
3 3
temperature of 950 °C (339 mg/mn ), and the lowest at maximum of 1350 °C (71 mg/mn ).
The CO emission, at the same process temperature, is lower for primary air supply then for
3
air staging - at λ1/λ = 1.15/1.15 it is 223 mg/mn , Fig. 5-right [1].
Characteristics of ash/slag deposits and the sulfur capture rate: Examples of solid
samples of deposits taken from ceramic probes attached to the water-cooled lence and
from bottom of reactor/combustion chamber during the selected test regimes and refer
to the projected coal mixture K70B20Z10, are presented. It is noteworthy that all the
samples, irrespective of the place taken from, are in loose (powdery) state for all process
temperatures below 1250 °C. At this temperature the initial formation of some larger
pieces of slag is observed, but is still very porous and does not stick on firm surface and
is easily crushed, i.e. easy-to-remove, Fig. 6-left. By analyzing solid samples of deposits
and slags, it was concluded that pulverized coal combustion technology with dry bottom
surface is applicable to coal mixtures from Middle Bosnian basin, both from the aspect
of combustion efficiency and from the aspect of undisturbed slag discharge. The sulphur
capture rate decreases with the increase of temperature, and is lowest for the mixture
with the lowest Kakanj coal fraction (K60). Additionally, there is no significant differ‐
ence between these values in relation to the amount and the air distribution, Fig. 6-right
[2, 3, 7].
Fig. 6. Samples of deposits and slag (left), degree of Sulfur binding in ash (right)
4 Conclusion
Based on the results of laboratory tests, the following conclusions can be drawn:
– The chemical analysis of flue gas was carried out. It was found that the emissions of
NOx and SO2 flue gas components, due to lower process temperatures, are lower in
relation to pulverized combustion systems with slag tap furnace. This applies in
particular to the test regimes in which an air staging is applied. However, these emis‐
sions (in particular SO2 emission) are still higher than the emission limit values
(GVEs) defined in the LCP and IED of EU directives. Regardless of the type of coal
tested, the temperature conditions and the air distribution, NOx emissions are always
3
greater than 400 mg/mn . Thereby the SO2 emission is halved in relation to this emis‐
sion for actual combustion of this coal in TPP Kakanj with slag tap furnace(ca
3 3
8000 mg/mn ) and in the most unfavorable cases its value is about 4000 mg/mn .
– High-efficiency combustion with low CO emissions can be established. In this sense,

regimes with medium process temperature ranging from 1150 to 1350 °C are distin‐
guished. Also, the determined fraction of unburnt is very low or within the limits of
usual values when using pulverized combustion technology with dry bottom furnace.
– From the aspect of the form, structure and characteristics of the ash deposit taken
from the ceramic probes and water-cooled lence and the slag from the bottom of the
combustion chamber, here is also shown that more favorable combustion regimes
are those with lower process temperature, i.e. combustion of the tested coal mixture
with dry ash discharge (temperature in combustion chamber is about 1250 °C), and
no major problems are expected regarding the fouling of the boiler heating surfaces
and slagging of this kind.
– The test regimes at the limit test temperature of 1350 °C characterizes the intense
formation of solid deposits on ceramic probes and the formation of larger pieces of
slag at the bottom of the reactor/combustion chamber. Generally, the test regimes
with process temperature of 1350 °C are those whose temperature is higher than the
ash flow temperature of all component coals, in particular Breza and Zenica coals.
– Keeping in mind the results of the tests in laboratory conditions, including the fore-
going conclusions, it can be estimated that pulverized combustion technology with
dry bottom furnace is acceptable for the tested coal mixtures and present more
sustainable way of combustion of these coals compared with PF technology with slag
tap furnaces.
References
1. Hodzic, N., Kazagic, A., Smajevic, I.: Influence of multiple air staging and reburning on NOx
emissions during co-firing of low rank brown coal with woody biomass and natural gas. Appl.
Energy 168, 38–47 (2016)
2. Smajević, I., Kazagić, A., Musić, M., Becic, K., Hasanbegović, I., Sokolović, S., Delihasano-
vic, N., Skopljak, A., Hodzic, N.: Co-firing Bosnian Coals with woody biomass: experimental
studies on a laboratory-scale furnace and 110 MWe power unit. Therm. Sci. 16(3), 789–804
(2012)
3. Kazagić, A., Smajević, I., Duić, N.: Selection of sustainable technologies for combustion of
Bosnian coals. Int. Sci. J. Therm. Sci. 14(3), 715–727 (2010)
4. Hodžić, N., Smajević, I., Kazagić, A.: Mitigation ash-related problems by cofiring coal with
woody biomass in case of Kakanj thermal power plant - experimental and trial run. In: 23rd
European Biomass Conference and Exhibition (EUBCE 2015), Wienna, 01–04 June 2015
5. Hodžić, N., Kazagić, A., Smajević, I.: Motivzauvođenje u praksukosagorijevanjaugljaidrvn
ebiomase u TE Kakanj - iskustvaizprobnogpogonapodržanalaborato-rijskimistraživanjima, 12.
Savjetovanjebosansko-hercegovačkogkomitetaCigré, Neum (2015)
6. Smajevic, I., Hodzic, N., Kazagic, A.: Lab-scale investigation of middle-bosnia coals to
achieve high-efficient and clean combustion technology. In: 8th International Conference on
Sustainable Development of Energy, Water and Environmental Systems, Dubrovnik, 22–27
September 2013
7. Hodžić, N.: Istraživanjekosagorijevanjaugljaibiomaseusmjerenonasmanjenjeemisijaprimarnimmj

erama u ložištu, Doktorskadisertacija, MFS/UnSa, Sarajevo (2016)
8. Hodžić, N., et al.: Combustion tests of coals Kakanj, Breza and Zenica at app. 70-20-10 mass
ratio for pulverised fuel combustion with dry botom furnace (in Bosnian), MEFS/UnSa,
Department for Energy, Process and Environmental Engineering for JP
ElektroprivredaBiHd.d.-Sarajevo, Contract No. 144-KI/11, Final Report, Sarajevo, July 2012
Influence of New Technologies on Higher
Energy Efficiency of Hydrostatic Devices
and Systems
Milutin Živković, Predrag Dašić(&), and Zvonko Petrović
High Technical Mechanical School of Professional Studies,

37240 Trstenik, Serbia
milutinzivkovicts@gmail.com, dasicp58@gmail.com,
zvonko.petrovic@vtmsts.edu.rs
Abstract. Both the experience and data from literature show that current
hydraulic components and development of electric control systems have signifi-
cantly increased energy efficiency. Therefore, all world manufacturers of hydraulic
devices and systems use advanced technologies and methods with the final aim to
integrate them into mechanical system. Energy losses, made as a result of all forms
of friction, are minimized. It is safe to say that modern growth is based on devel-
opment of new materials and accompanying technologies as well as on control
techniques, i.e. on mechatronic approach and design of such systems. Ecological
requirements and legal regulations should not be neglected, either. Based on the
above issues and available data the authors of the paper try to answer the following
question: What is the position of once famous factory of hydraulic devices from
Trstenik in comparison to other competitive factories worldwide?
Keywords: Development and research Efficiency Energy losses

New technologies
1 Introduction
Fluid power can be transmitted remotely and can be controlled, which allows mecha-
nization and automation of work processes. Thanks to its advantages, in terms of other
energy transfer modes, as well as the way of converting the energy, all the advantages and
disadvantages of hydrostatic power transmission are well known [1, 2]. They are char-
acterized by the possibility of achieving drives both in demanding and precisely controlled
systems as well as in the general purpose system. The financial statements, on the pro-
duction and placement of these devices, show that fluid power shows an increasing growth
index from year to year. It is even bigger than the entire mechanical engineering sector and
is characteristic of industrially developed countries, in particular [3]. Mechanical engi-
neering is the science and skill of the designer/constructor in the selection of mechanisms,
which have moving parts, which are controlled or trigered mechanically, i.e., power of the
fluid. Technical requirements have become more and more demanding due to the
development of design technologies and expressed requests by the contracting authority.
For this reason, the fluid power system designer is faced with the need to know the

https://doi.org/10.1007/978-3-319-90893-9_46
Influence of New Technologies on Higher Energy Efficiency 387
integrated design of the components and all their interactions, the control theory and
measurement of the drive parameters as well as the microprocessor control. This means
that the technology of fluid power control, in recent years, has improved due to the
increasing integration of electronics, i.e., control technologies within the components of
fluid technology, hydraulic circuits and control systems [2]. Current trends and future
requirements for energy efficient systems will not only significantly increase their number
as well as the number of possible constructive solutions, but will significantly change their
performance requirements [2, 4, 5]. Intensive development of other technologies, both
productive and technological, influence the significant improvement both of the
mechanical and functional characteristics of the device [7, 8]. We should not ignore
neither the ecological requirements and greater competitiveness on the market, as well as
the accelerated development of new industrial countries (China, Turkey, India, Brazil.)
[3]. All this affects on a tendency for constant innovation and technical and technological
improvements. In this paper, the authors will analyze what is based on all improvements in
the world by analyzing the influence factors on the production and placement of hydraulic
devices. The question is where is now the place of the factory Prva Petoletka from
Trstenik, once known manufacturer of hydraulic and pneumatic devices and systems. An
example of a wing pump of power steering installations will show the importance of
investment in the development, and new technologies and modern production
lines/machines [6].
2 Assumptions on Modern Development of Hydraulic

Components and Systems (HCS)
Scientific and technological circumstances, in the modern and developed world, relate
to the development of electronic digital computer and communication technologies
leading to the automation of machines, production processes and even entire factories.
This is possible thanks to the accumulated knowledge and development of digital
devices as well as measuring and control devices, which significantly shortened the
development time of new products based on: utilization of their advantages, to com-
pensate for their deficiencies as using the advantages of other technologies and
acquired knowledge in the stage development of products (research, development and
production) as well as: design, construction, production, exploitation, maintenance and
their recycling.
Modern development (HCS) is based on the development: design technique,
development of new materials and control techniques as well as other associated
technologies, which is shown in Fig. 1 [1].
2.1 Development of Design Technique

Design of hydraulic components, the last decade, defined new concepts which includes
control of: pumps, motors and valves. It is today mostly generated through electrical
signals, which is due to the development and improved performance of digital control
systems, Fig. 1. With the reduction of the drive power it is possible to reduce all losses:
flow, pressure and mechanical, with the choice of components for the system’s
388 M. Živković et al.
Fig. 1. Conversion and control of energy in modern hydrostatic systems
Fig. 2. Examples of solutions of hydrostatic systems with a linear actuator
customized purposes. Valve-controlled hydraulic systems as well as proportional-

to-control systems adapted to the drive show the examples given in Fig. 2 [4].
They relate to the regulation of the pump flow, both constant and variable operating
volumes (gear pump) and the possibility of regulating the speed of the drive (electric
motor), within the limits of min. and max. values, Figs. 2 and 3 [4]. Manufacturers in
mobile hydraulics in particular, uses the pressure sensor in order to reduce all operating
losses in progress [2, 4, 5]. The speed and direction of the actuator are controlled by
changing (controlling) the pump/motor [4].
Figure 4 shows all sources of energy losses in any hydraulic system as well as
measures for their energy improvement.
Fig. 3. Basic characteristics of the drive solutions in Fig. 2
Fig. 4. Energy losses in hydraulic system and measures for increasing efficiency
2.2 Development Based on New Materials and Technologies

For all components of hydraulics, the chosen tribopar material is important because it
will significantly affect the reduction of frictional forces and the intensity of wear. By
analyzing all contacts, with components of hydraulic devices, their number and
importance are observed, Fig. 5.
For parts that make sliding joints, most commonly used are tinned bronze and
special types of steel. They must have good wear resistance, high tensile strength and
hardness as well as toughness. However, a new study has shown that the contact
surfaces of steel, thermally treated, achieve a significantly lower friction coefficient
than the most common bronze paired with steel [6]. Increasingly, special types of
Fig. 5. Analysis of possible contacts of hydraulic components and their significance
graphite are used for parts that work in difficult conditions. A combination of hardened
steel parts and graphite parts is used. Graphite can be enriched with metals such as: tin,
copper, lead and others. The material obtained in this way has good mechanical
properties and exceptional sliding properties. Modern methods used in modern pro-
duction of materials and measures for their improvement are shown in Fig. 6. Graphite
is often used for the production of heavily loaded parts of various hydraulic compo-
nents, considering the range of good features such as: low coefficient of friction, good
self-lubricating properties, very good coefficient of heat conductivity, small coefficient
of linear expansion (when changing temperature). Also, these parts slightly change the
dimensions in operation, have increased strength with increasing temperature, high
corrosion resistance, easy mechanical handling, low specific weight (reduced weight of
parts). When choosing a type of graphite, for the production of pump parts, it is
necessary to take into account the product of pressure and speed.
Fig. 6. Overview of the procedure and methods of contacting surface

In order to obtain a quality part, it is necessary to provide a reliable combination of

graphite and metal, and the highest quality compounds are obtained by gluing special
adhesives. In the case where the parts are glued, it is necessary to produce certain
preparations of graphite that have the shape of sleeves or rings, depending on where it
will be installed. For the bonding of graphite parts on the metal surfaces of the
hydraulic components, good results were given by the adhesive “ECCOBOND” [6]. It
is applied to the prepared parts according to the strictly prescribed procedure. After
application, baking is performed at a certain temperature in the time interval defined in
the instruction. Parts are subsequently processed in order to obtain desired dimensions
and quality of processing. Figure 7 shows a cylinder block with pistons and pedals of
an axial-piston fuel pump that uses kerosene as a working fluid. On the cylinder block
there are cylindrical openings in which graphite shells are incorporated and nitride
pistons are moving inside the sleeve. The radial bearing of the cylinder block, in the
form of a sleeve, is also made of graphite. Pedal, pos. 1, has a graphite ring on the
sliding surface, pos. 2, so that graphite contact is obtained on the nitrided surface of the
pivot or tilting plate. The shape of the pedals and the rings enables hydrostatic and
hydrodynamic swimming which contributes to the long service life of the pedal [6].
Potential loss spots (mechanical and volumetric) are shown in the same picture, and at
the same time are the places most affected by the improvement of their energy char-
acteristics. In Prva Petoletka Trstenik there are good experiences in the application of
graphite parts in axial-piston pumps, which have been producing them for 40 years.
Considering the field of application, extensive and long-term testing was carried out on
the basis of which data were obtained which were confirmed in exploitation [6]. When
it comes to reliability, long-term tests gave the following data: for pumps used in
aviation within the 106 h interval there were 16 failures, while for industrial applica-
tions, there were 19 failures in the same interval.
Fig. 7. Cylindrical block with graphite parts (a) and display possible power losses with piston
pumps (b)
All these data show that this is a reliable solution with a long life in exploitation [6].
In addition to graphite, which has been widely used in hydraulic and pneumatic
components, ceramics have been used more often in various fields in the last 30 years
at their place. For more important use of ceramics, it was necessary to make efforts in
order to reduce price and brittleness from one and increase the toughness and reliability
on the other. The problem of brittleness is overcome by incorporating fine fibers or

particles into the structure of the parts, which are phase-transformed. The coefficient of
linear expansion of ceramics is small, so it is suitable for making parts that slightly
change the dimensions with temperature change. It has extremely high hardness and
wear resistance (with temperature change), ceramic parts do not significantly change
mechanical properties and possess relatively good slip properties. Hydraulics and
pneumatics are used and increasingly used in sliding elements, sealing elements and
parts that are extremely thermally charged [6].
When it comes to high-tech ceramics, it should be said that parts are shaped,
normalized by structure and grits under strictly controlled conditions, most often by the
process of hot isostatic pressing (HIP). High-tech ceramics are inorganic non-metallic
materials of fine-grained crystalline micro structure, with precisely controlled content
of additives. In the area of piston pumps, ceramics are made of shells and rings of
various shapes that are attached to certain places. For example, cylinder blocks of
high-pressure piston pumps have cylindrical ceramic shells. Especially the ceramic
material is interesting for the parts of the sliding seals at high pressures and angular
velocities. Parts are made by the surface modification technology, while certain
dimensions are obtained by a special processing method, so it is possible to obtain
quality better than Ra = 0,1 lm. In order to overcome the problems of external leakage
of oil and compatibility with the environment, the use of biodegradable working media,
as well as water, becomes a necessity instead of mineral oils. The recommendations for
their selection are defined by the manufacturers of these devices, prescribing the type of
pump as well as the range of working pressures, which can be seen in Fig. 8 [7].
Fig. 8. Recommendations for selection of working fluid according to the selected pump
2.3 Development Based on Control Techniques

The development of microprocessor technology has increased the field of application of
proportional and servo techniques, and therefore the management techniques. This
significantly contributes to improving the dynamic characteristics of the drive, reducing
the cost of the device [1, 2]. Trends have transformed fluid power technology, from
multiple individual devices for a function, to integrated technology, and more varia-
tions of design. The use of control systems of the fluid power initiative, coupled with
the need to control and manage the system, helped the researchers to find alternatives
with proportional and/or servo valves, Fig. 2. The second way is the concept of digital
valves, which allows simple connection between the valve and the system, based on the
capabilities of the microprocessor [10, 11]. Known characteristics of fluid power ele-
ments, combined with advanced electronics/electronics technologies, with great flexi-
bility in calculations and all control transformations, Fig. 9 [8].
Fig. 9. The appearance of modern piston axial pump/engine control
In control of the drive fluid power, the convection and nonconvection control is
known. Conventional control is: linear and classical null control. The basic noncon-
ventional control is: optimal, adaptive, variable structure control, monitoring control as
well as intelligent control. Conventional controls (P, PD, PID, feedback, on/off con-
dition), regardless on many new control algorithms, are widely used in the pump
control and motor control industry. The optimal time control implies control of the
optimal values of the system’s default values (minimal peak load, minimum position
error, readiness to work with minimal absolute error, …) [9–11].
3 Analysis of the Possibilities of Constructive Improvements

of Wing Pumps of Production “PPT–Trstenik”
Prva Petoletka from Trstenik, once known manufacturer of hydraulic and pneumatic
devices and systems, in cooperation with the largest manufacturer of power steering
components (SU) in Europe (ZF) has successfully produced them for 40 years.
Products and Wing Pumps (Fig. 10), which has in the body a valve for limiting flow
and pressure. In the case when a special pressure limiting valve is installed in the
system, the pump is only run with a flow limiting valve. Operating elements are
dynamically balanced and the pump generates low noise, which is of particular
importance to motor vehicles as well as to other control systems.
Depending on the drive mode, three types of performance are distinguished: pumps for
drive over pulleys, gears (directly to the engine) or built directly on the compressor shaft.
Fig. 10. Wing pump from PPT Trstenik Fig. 11. Wings in contact (force analysis)
The maximum pressures are up to 135 bar, but in most systems it is limited to 100 bar. The
constructive development of the pumps initiated the needs of the users, so today there are
dozens of types of solutions. They relate to: installation measures, the way in which flow
control is regulated as well as the specific flow. Standard pump solutions, for installation on
trucks and buses, have a specific flow of 13.5 and 20 cm3/o and rpm (500 to 3500) min−1,
regulated flow (9 to 20) l/min and max pressure up to 135 bar. New solutions for hydraulic
systems for SU require a higher pressure level, and they have values up to 180 bar. Higher
pressure levels can be achieved only by reducing the thickness of the wings, without
increasing the contact pressure force. Because of the working volume (cm3/o), it is assumed
that the length of the wings (l) must remain the same. If the thickness of the wings would
decrease from b1 = 1,98 mm to b2 = 1,4 mm (Fig. 10), the pressure could be increased
from p1 = 135 bar to over p2 = 190 bar, with the requirement that the working life remains
same (Fig. 11).
Reducing the thickness of the wings reduces its mass, which leads to a reduced
centrifugal force, which also contributes to less wear. Achieving higher pressures, in
addition to reducing the thickness of the wings, was also conditioned by some of its
structural improvements [9]. By the standard solution, which has been produced for
years in the PPT-SU, the oil pressure works on both front and rear surface of the wing,
which at values above 135 bars leads to a reduction of the slip gap.
This may result in possible wedging. With a vigorous solution, the oil pressure only
affects the front of the wing so that at higher pressure values there is no change in the
gap. However, serial production of a new pump solution is not possible on existing
production lines as well as in existing technological processes. This requires new
investments both in production and in serious research and significant investment in
personnel and control devices at all stages of development. As a contribution to this,
reduced competitiveness in the world market, expressed through reduced production
and placement for the same ones, Table 1, and Fig. 12 [9].
Table 1. Review of production and distribution of wing pumps (1999–2017) year

Years Sold (pieces) Produced (pieces)
1994 2482 2670
1995 5566 5722
1996 5275 5850
1997 10935 10687
1998 14667 13946
1999 16000 17922
2000 20500 22148
2001 27200 27200
2002 29846 29906
2003 23328 22650
2004 27254 27271
2005 14279 14344
2006 7445 7445
2007 11038 11184
2008 2430 2430
2009 1403 1403
2010 795 836
2011 918 943
2012 663 670
2013 606 610
2014 501 495
2015 856 865
2016 582 595
2017 1096 1100
Fig. 12. Production - commercial results of wing pumps production (PPT Trstenik)
4 Conclusion
Based on all of the foregoing, it can be concluded that:

• competitiveness implies continuous improvement of product characteristics,
• development of microprocessors influences the development of components and
hydraulic systems, also has a significant impact on the development of modern
components as well as the system,
• application of new materials, management algorithms and development of auton-
omous systems,
• There is no modern production and competitiveness in the market without new
products and constant investment in both production equipment and research,
• what is the problem of surviving on the market and having a safe placement of
products show the achieved production results of the factory’s production of wing
pumps of factory PPT from Trstenik.
References
1. Ivantysyn, J., Ivantysynova, M.: Hydraostatische Pumen und Motoren (Konstruktion und
Berechnung). Vogel Buchverlag (1993). ISBN 3-8023–0497-7
2. Tan, K.K., Putra, A.S.: Servo hydraulic and pneumatic drive. In: Drives and Control for
Industrial Automation. Advances in Industrial Control. Springer, London (2010). https://doi.
org/10.1007/978-1-84882-425-6_2
3. Report by Global Industry Analyst, Inc. (2017). www.strategyr.com
4. Komesta, J.: Variable Speed Pump Drives for Industrial Machinery – System Considera-
tions, p. 40. Bosch Rexroth Group, Chicago (2013)
5. Rydberg, K.-E.: Energy efficient hydraulics. In: National Conference on Fluid Power,
Linköping University, Sweden, 16–17 March, pp. 1–9 (2015)
6. PPT Trstenik: Test reports on pumps and hydraulic units (2000–2014)
7. Asaff, Y.: Systematization of hydraulic systems design for use with biodegradable fluids. In:
LASHIP, São Paulo, Brazil, 2–3 September, p. 40 (2014)
8. Ivantysynova, M.: Technologies and innovations for hydraulic pumps. In: 10th Fluid Power
Conference/Dresden 2016, Group 3 – Pumps/Paper 3-0, pp. 135–138 (2016)
9. Živković, M.: Study on improving the design characteristics of the wing pumps. PPT
Trstenik 7673. xx. 955, p. 35 (2017)
10. Yigen, C.: Control of a digital displacement pump, Supervisors; H. Clemmensen Pedersen,
Per Johansen: Project Group: MCE4-1022, p. 111 (2012)
11. Heikkilä, M., Linjama, M.: Hydraulic energy recovery in displacement controlled digital
hydraulic system. In: 13th Scandinavian International Conference on Fluid Power, SICFP
2013, Linköping, Sweden, 3–5 June, pp. 513–519 (2013)
Design of Pico Hydropower Plants
for Rural Electrification
Krsto Batinić(&), Dušan Golubović, Stojan Simić,

and Goran Orašanin
University of East Sarajevo, 71123 East Sarajevo, Bosnia and Herzegovina

bkrsto94@gmail.com, dusan.golubovic54@gmail.com,
stojans@modricaoil.com, goranorasanin@yahoo.com
Abstract. Pico hydro power stations include all hydropower systems with
output of up to 5 kW. These systems have proven themselves to provide a
reliable source of electricity for rural areas. Apart from rural area electrification,
pico hydro power stations are convenient for utilizing the hydropower potential
of water supply systems. This paper shows the calculation procedure and the
selection criteria for basic components of a pico hydropower station based on a
Pelton turbine. The paper provides an example of calculation procedure as well
as the possibility of using them in Bosnia and Herzegovina.
Keywords: Hydropower Pico hydropower plants Pelton turbine

Rural electrification
Nomenclature
c1 - absolute velocity of water jet [m/s]
D - pitch circle diameter [m]
d - optimal jet diameter [m]
Da - runner outside diameter [m]
dN - nozzle diameter [m]
du - inside diameter of the penstock [m]
s - wall thickness [m]
g - gravitational constant (g = 9,81) [m/s2]
h - bucket height [m]
Hn - net head [m]
i - transmission ratio [-]
kc - nozzle coefficient [-]
ku - coefficient [-]
nG - rotational speed [min−1]
P - power output [kW]
Q - installed flow [m3/s]
t - bucket depth [m]
u1 - optimal peripheral velocity [m/s]
z - approximate number of buckets [-]
η - efficiency [-]
wD - jet contraction coefficient [-]

https://doi.org/10.1007/978-3-319-90893-9_47
398 K. Batinić et al.
1 Introduction
In this paper, there will be presented the possibility of producing electricity in the pico
hydropower plants. Pico hydropower plants include all hydropower systems up to
5 kW. In the last 20 years, pico hydropower plants found their highest use in the
electrification of rural areas, replacing aggregates with fossil fuel, photovoltaic systems
and wind power plants. Pico hydropower plants are most used today in the countries of
South America and Southeast Asia. In addition to the hydrologic wealth of these areas,
the main reason for the dominance of the picohydroelectric systems in the electrifi-
cation of rural areas is low cost and negligible operating costs.
2 Low-Power Hydropower Plants – Pico Hydropower Plants

and Their Use
In most literature, hydropower systems of up to 50 kW are classified as “micro

hydropower systems” [1], but it is not uncommon for systems of power from 5 to
50 kW to be classified as “micro hydropower systems”, and those of up to 5 kW
classified as “pico hydropower systems’’ [2]. In this paper, systems of power up to
5 kW will be classified as “pico hydropower systems” in order to make a difference
between them and micro hydropower plants, which according to the current law in the
Republic of Srpska and Bosnia and Herzegovina are all hydropower plants up to
500 kW. Low-power hydropower systems (up to 5 kW) have proven to be a very stable
source of electricity for the electrification of rural areas [3]. Their highest advantage is a
long lifetime and a small number of components.
In this paper, there will be presented the guidelines for the selection of components
and the design of a turbine wheel of a pico hydropower plant based on the Pelton
turbine are presented. Input parameters for design are: installed flow Q [m3/s], and
gross head H [m]. Also, here will be considered only systems up to 5 kW. The design
principle of the turbine and all components is the same for both 5 and 50 kW, with
more rigorous safety requirements, precise manufacturing of components, material
selection, etc.
2.1 The Use of Pico Hydropower Plants in the Electrification

of Rural Area
Hydropower systems with power up to 5 kW have a very important role in the elec-
trification of rural areas. These systems are suitable for the supply of individual
households with electricity. For example, it is estimated that there are between 60,000
and 100,000 picohydropower plants only in rural parts of the state of Laos (Laos
People’s Democratic Republic) [4].
The first country that began to use the picohydropower plant for rural electrification
is Nepal. In 1990, more than 600 picohydropower plants were installed in this country.
Of all turbines, the highest use had Crossflow and Pelton turbines. The reason for this is
the greatest adaptability to the parameters of the net head and flow, and the lowest cost
of production [8].
Design of Pico Hydropower Plants for Rural Electrification 399
The basic precondition for using picohydropower plants in the electrification of

rural areas is the existence of a widespread hydrological network of water sources. For
this reason, the pico hydropower plants found their greatest use in the countries of
South America and Southeast Asia.
2.2 The Use of Pico Hydropower Plants in the Water Supply Systems
Except the use in the electrification of rural areas, the pico hydropower plants can also
be used for the exploitation of hydropower potential in water supply systems. If there is
a sufficient height difference between the two values on the transport pipeline, it is
justified to invest the installation of the turbine and the generator on the already existing
pipeline. [6] When installing the hydropower plant on the transport pipeline, the costs
of the water intake and the supply pipeline are eliminated. The biggest lack of these
systems is variability of flow due to different consumption during the day. Therefore,
the turbine and the generator must be dimensioned so that they can work in both
increased and reduced loads. This will lead to a reduction in the efficiency coefficient at
nominal parameters, but will allow the system to operate throughout the year.
3 Design of a Pico Hydropower Plants
The first step in designing a pico hydropower plant is to estimate the hydropower
potential of the water source. This is done by measuring the gross head and the water
discharge. After dimensioning of the supply pipeline, the net head is calculated, and it,
together with the flow, represents the hydropower potential of the hydropower plant.
After determining the net head and water discharge, approach to designing a system is
used to make the most of this potential. According to some analyzes, for the optimal
diameter of the pipeline, the diameter for which the hydraulic losses are 5–15% of the
gross head is taken [3]. After previously selection of the diameter of the supply
pipeline, we approach to estimate of the net head. In this paper there will be not
describe of the calculation of the net head, but it will be approached to the estimate of
the Pelton’s turbine with known parameters of the watercourse (flow, gross and net
head).
In the electrification of rural areas, the largest use found action-type turbines. The
reason for this is that action turbines can work efficiently with much less installed flow.
In such small systems, Crossflow turbines were first used. The reason for this is their
simple production. The Pelton turbine can work with a smaller flow from the Crossflow
turbine and to operate at a significantly higher net loss.
3.1 Design of Pelton Turbine

There are three main principles in designing of the turbine:
• calculation of optimal jet diameter,
• calculation of turbine wheel diameter, and
• calculation of turbine buckets dimensions.
3.1.1 Calculation of the Nozzle

Nozzle is an inlet pipe of Pelton turbine. On account of pressure loss, kinetic energy is
increasing, so velocity of water jet is very high:
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
c1 ¼ kc 2 g Hn ð1Þ
For absolute velocity of water jet, calculation needs to be calculated optimal jet
diameter:
rffiffiffiffiffiffiffiffiffiffiffi
4Q
d¼ ð2Þ
p c1
By instructions given in [8], water jet diameter should not be smaller than 4 mm.
Number of buckets depend on water jet diameter. Also, if number of buckets is not at
least possible in calculation ðz ¼ 16Þ, there needs to be increased number of nozzles in
it and calculation have to be done again. There are two main construction imple-
mentations of nozzles: Nozzles with variable cross section and nozzles with constant
cross section. In variable ones, width of water jet is regulated by position of regulation
spear. Adjustment of that can be done manually or automatically.
In calculating of nozzle’s diameter, jet contraction coefficient must be considered.
This coefficient represents relation between nozzle’s diameter and water jet diameter:
dN
wD ¼ ð3Þ
d
Jet contraction coefficient wD is in amplitude of 1,6 for holes with sharp edges till
1,05 for ideally rounded edges.
Beside nozzles with an option of regulation of the cross section, it is possible to use
even nozzles with the constant cross section. The only advantage of it is lower price
and better lasting than first ones.
3.1.2 Calculation of Pelton Wheel Dimensions

After calculation of the absolute velocity and water jet diameter, now is the turn of
calculation of the wheel turbine. Geometric characteristics of turbine buckets are cal-
culating based on water jet diameter, most important of which are (Fig. 1):
• bucket width - b [m]
b ¼ ð2; 5 3; 2Þ d ð4Þ
• bucket height - h [m]
h ¼ ð2; 1 2; 7Þ d ð5Þ
• bucket depth - t [m]
t 0; 9 d ð6Þ
For net head amount, extensive optimal peripheral velocity is calculated:

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u1 ¼ k u 2 g Hn ð7Þ
u1
ku ¼ ¼ 0; 45 0; 49 ð8Þ
c1
Beside coefficient of nozzle usage and hydraulic losses, coefficient ku highly affects
on coefficient of Pelton turbine usage. Relying on extensive velocity of the turbine
wheel and designed number of generator’s rotational speed, pitch circle diameter is
calculated by this form:
60 u1 i
D¼ ð9Þ
p nG
In the case of the pico Pelton turbine, processed in this paper, transmission ratio
will have the value i = 1, because the turbine and the generator are directly connected,
which means there is no change rotational speed. Figure 1 shows the main dimensions
of the Pelton turbine buckets.
Fig. 1. Main dimensions of Pelton buckets [8]

In some cases, it is possible to install a turbine with larger buckets of calculated

ones. Small increase in dimensions slightly affect the reduction of the useful efficiency
coefficient. Increasing the dimensions obtained from the calculation is justified if an
already existing turbine is used, or if the reduction in dimensions would lead to changes
in production, thereby affecting the increase in production costs.
Depending on the pitch circle diameter and the height buckets, the runner outside
diameter is calculated:
Da ¼ D þ 1; 2 h ð10Þ
On the efficiency of the Pelton turbine, a big influence has the number of buckets
adopted. Optimal number of Pelton turbine buckets depend on the pitch circle diameter
and water jet diameter [7]:
Dp
z¼ ð11Þ
2d
According to the same author, the number of buckets not be less than 16. Various
authors give different formulas for calculating the optimal number of buckets. In
principle, any cutting of a water jet on the bucket causes a disruption in the flow.
Therefore, the unnecessary increase in the number of buckets reduces the efficiency of
the turbine. On the other hand, insufficient number of buckets causes volumetric losses
in turbine wheel.
4 Calculation Example
In this chapter will be displayed calculation of pico hydropower plants, using the
following parameters:
• installed flow – Q = 0,005 m3/s;
• gross head – H = 50 m;
• the length of the supply pipeline – l = 200 m;
The first step in the calculation is the previous choice of the diameter of the supply
pipeline. A high density polyethylene pipe (HDPE pipe), a nominal operating pressure
of 10 bar (PN 10) and an absolute roughness coefficient, k = 0.002 mm was selected
for the supply pipeline.
Respecting the recommendation on the choice of the optimal diameter of the supply
pipeline referred to in Sect. 3, the order of standard pipe sizes, the pipe DN 110 NP 10
is selected, with the following parameters:
• inside diameter – du = 93,8 mm;
• wall thickness – s = 8,1 mm;
• pressure nominal – 10 bar.
For selected pipe DN 110, approaches to a detailed calculation of Pelton turbine
parameters.
Table 1. Results of the Pelton turbine calculation for different rotational speed
Quantity Symbol Unit Calculation
Rotational speed nG [min−1] 650 1000 1200
Net head Hn [m] 47,2
Power output for efficiency coefficient 55% P [kW] 1,27
Absolute velocity of water jet c1 [m/s] 29,22
Optimal jet diameter d [mm] 14,76
Nozzle diameter dn [mm] 18,45
Optimal peripheral velocity u1 [m/s] 13,70
Pitch circle diameter D [mm] 402,45 261 218
Optimal number of buckets z [-] 43 28 23
Bucket width b [mm] 47,23
Bucket height h [mm] 39,85
Helpful sizes h1 [mm] 5,16
h2 [mm] 22,14
a [mm] 17,71
Runner outside diameter Da [mm] 450 309 266
Bucket depth t [mm] 13,28
The calculation is done for different rotational speed. The results of the calculations
are shown in Table 1.
As can be seen from Table 1, for different rotational speed of the turbine, the
different pitch circle diameter and number of buckets are obtained. Other values do not
depend on the projected rotational speed, but only on the parameters of the projected
flow. The turbine parameter calculation process is calculated for available rotational
speed of the generator. For this turbine, 1200 min−1 is chosen. For the given number of
turns, the generator 145STK2 M, manufacturer ALXION, is selected, the character-
istics shown in Fig. 2:
Fig. 2. Nominal power output of generator depending of rotational speed [5]

A flange ball valve must be installed before the regulating nozzle, so that the supply
pipeline should not discharge due to possible repairs on the nozzle, turbine or housing.
Figure 3 shows the system of concrete pico hydropower plants.
Fig. 3. Concrete pico hydropower plant
5 Conclusion
Pico hydropower plants proved to be a stable source of energy for the electrification of
rural areas. Apart from lower investments comparing to the solar systems, wind power
plants and aggregates on fossil fuels, the advantage of the pico hydropower plant is the
simplicity of construction, easy assembly and disassembly, low operating costs, as well
as greater resistance to weather conditions.
Pico hydropower plants, so far, are not specifically defined by the laws of the
RepublikaSrpska and BiH. By resolving the legislation for the pico hydroelectric power
plants, the conditions for further progress in this area would be met.
References
1. Manjanutha, N., Kuldeepak, K., Thammaih, G.: Design of a Pelton wheel turbine for micro
hydro power plant. In: 5th International Conference on Science, Technology and
Management, New Delhi, pp. 447–454 (2016)
2. Williamson, S.J., Stark, B.H., Booker, J.D.: Low head pico hydro turbine selection using a
multi-criteria analysis. Renew. Energy 61, 43–50 (2014)
3. Williams, A.A., Simpson, R.: Pico hydro-reducing technical risk for rural electrification.
Renew. Energy 34, 1986–1991 (2009)
4. Smits, M., Bush, S.: A light left in the dark: the practice and politics of pico-hydropower in
the Lao PDR. Energy Policy 38, 116–127 (2010)
5. http://www.alxion.com/e_produit/82.pdf. Accessed 25 Sept 2017
6. Batinić, K.: Proizvodnjaelektričneenergije u vodovodnimsistemima, COMETa, Jahorina
(2016)
7. Zaid, M.: Pelton wheel driven mycro-hydro plant. In: IJREAT (2014)
8. Eisenring, M.: Micro Pelton Turbines, MHPG Series: Harnessing Water Power on a Small
Scale, vol. 9. Swiss Center for Appropriate Technology (1991)
Development of Construction of Mini Hydro
Power Plant Model Based on Pelton Turbine
Radoslav Tomović(&), Aleksandar Tomović, Marko Mumović,

and Vuk Vujošević

radoslav@ac.me, aleksandartomovic23@gmail.com,
markomumovic@gmail.com, vukvujosevic7@gmail.com
Abstract. In this paper is presented constructive development of a mini hydro

power plant model, based on Pelton turbine. The 3D model of the Pelton turbine
was constructed and drawn in working environment SOLIDWORKS. The con-
struction is implemented to the end, where the model of mini hydro power plant
was obtained. The parts of a model where made of various materials. Supporting
parts were made of steel, and on them is attached plexiglass, so the model look
like an aquarium. The main part of the turbine, runner, is compiled using two
disks made of aluminium alloy, and buckets which were printed on 3D printer,
using PLA plastic. For the simulation of water head, water pump is used.
Keywords: Power plant Turbine Bucket Pelton
1 Introduction
The importance of energy gained by the power of water (hydropower) becomes more
important due to the influence of people on the natural environment. The majority
countries want to substitute energy sources, like coal and gas, with renewable sources
like hydropower. At the same, the worldwide electricity consumption is increasing,
putting a further demand on better turbine performances. Hydroelectric power plant is a
facility for the production of electricity obtained by the power of water. Any water
turbine is a mean of extracting energy from falling water. Different types of turbines
works under different conditions of operating. Propeller turbines work best in case of
low head and high flow, crossflow turbines needs medium heads and flows, and the
pelton turbines work best under conditions of high head and low flow.
For our design of mini hydro power plant, we decided to construct and make the
model of Pelton turbine.
The input parameters that were most important for dimensioning and design of the
turbine were given by the technical task. The water pump was used as simulator of
water head, so the data given in technical task were taken from the description of pump
parameters (Table 1).

https://doi.org/10.1007/978-3-319-90893-9_48
406 R. Tomović et al.
Table 1. Setting the task and changing properties
First step: Setting the task and changing properties

Input product Output product
Water from nozzle Work of turbine Output voltage
Properties of input products Properties of output produtcs

Operation from the nozzle, flow 50 l/min, and
Voltage 12V
pressure 4 bar
Additional Requirements
Fixed requirements: Tolerated requirements: Rating parameter:
-max dim. - flow 50 l/min -High current
a x b x c= 1000 mm x - pressure 4 bar
1000 mm x 1000 mm - position of nozzle
2 The Working Principle

(1) Water in housing - First, the water is filled into the turbine housing
(2) Pumping of water – The pump pumps water and simulates the head
(3) Water outlet on the nozzle – Water from the pump comes to the nozzle
(4) The jet impact to turbine – The turbine starts to rotate
(5) Rotation and transfer of energy – The energy of the jet impact is transmitted over
the shaft
(6) Reduction over belt - increasing the rpm speed, transmission of mechanical
energy
(7) Electrical motor – Converting the mechanical energy to the electrical energy
(8) Electric consumer (Fig. 1).
Rotation and Reduction over Electrical motor Electric consumer

transfer of energy belt
Water in housing Pumping of water Water outlet on the nozzle The jet impact to turbine
Fig. 1. The working principle of Pelton turbine

Development of Construction of Mini Hydro Power Plant Model 407
3 General Description of Mini-Hydro Pelton Turbine
Pelton turbines work best under conditions of high head and low flow. For micro-hydro
sites, the best option is Pelton turbine and have many advantages:
– Peltons can exploit high head and produce a lot of energy from a small unit,
– they are very easy to make,
– they can be used for various heads and flows and can be very useful for driving
small electrical generators.
“Water is brought down the penstock pipe to a nozzle, and it comes out into the
turbine casing as fast, circular jet. The jet is directed at a wheel, or runner, which has a
number of buckets around its edge. The force of the jet on this wheel makes it turn, and
gives the output powerˮ [1].
For better turbine exploitation and efficiency, the proper bucket design is one of the
most important tasks in constructing a turbine.
“The buckets have the shape of two cups joined together, with a sharp ridge
between them. There is a notch cut out of the bucket at the outside end of the ridge.
[…] The first part of a bucket that the jet should hit is the splitter ridge. This should be
sharp and smooth, so that it cleanly splits the jet into two halves. […] To achieve this,
the position of the end point of the splitter ridge is important. The optimum position
will have most of the water hitting the bucket when it is nearly at right angles to the jet.
The angle of the splitter ridge is set so that it is approximately at right angles to the jet
when the full cross-section of the jet is hitting it. The rest of the internal shape is
designed to allow the water to flow freely around, and out the of the edgesˮ [1] (Fig. 2).
Fig. 2. A Pelton turbine runner
“The notch allows the jet to pass through the outside edge of the preceding bucket
so that the first place it touches each bucket is at the end of the splitter (see Fig. 3).
Note the jet passing through the notch of the bucket a before hitting the splitter ridge on
bucket b. The pressure of the water in the penstock produces the speed of the jet, but
Fig. 3. Diagram of a Pelton runner showing PCD, and design flow
the water in the jet itself is at atmospheric pressure. […] Pelton turbines are specified by
their Pitch Circle Diameter, PCD. This is the tangential diameter at which the jet
centerline passes the wheel. Jets are normally controlled by spear valves or injectors.
A spear valve consists of a nozzle with a spear-shaped needle in the centre of it. The
spear is moved in and out to vary the flow and some position of spear will be the
‘design’ position, which gives the design flowˮ [1].
3.1 Number and Design of the Bucket

During selection a number of buckets for this turbine the perfect solution is to fit them
as close to the shaft and use as few as possible. The problem is that if there are too few
of the buckets some of the water in the second jet will not be caught, and therefore
efficiency of the turbine will be smaller. The exact number depends on the bucket
design, but is generally between 18 and 22 buckets. The number of buckets can be
found by drawing or by using following equations.
rffiffiffiffiffiffi
D1 D1
zk ¼ þ 17; when is [ 12 ð1Þ
d0 d0
D1 D1
zk ¼ 0:4 þ 16; when is ¼ 12 ð2Þ
d0 d0
D1 D1
zk ¼ 0:5 þ 15; when is 6\ \35 ð3Þ
d0 d0
D1 – pitch circle diameter, a diameter where the jet hits the bucket
d0 – diameter of the nozzle
In this project 18 buckets as an optimal number for this application were choosen.
The drawing of the bucket is presented at the Fig. 5. The buckets are made on the 3D
printer using PLA plastic, and for one bucket it took nearly 4 h of printing (Fig. 4).
Fig. 4. Buckets of PLA plastic
Fig. 5. Drawing of the bucket
After the printing was completed, the buckets should be processed to remove
unevenness, and to get better fit into the runner.
3.2 Calculations
Based on the cross-section of the outlet pipe, measurements for various pipe thick-
nesses in order to determine the best possible variant were performed, by measuring the
time of flowing of 50 l of water. Based on this, the water speed on the pipe outlet and
the dynamic pressure was calculated. In the Table 2, the calculation results are given.
1
dynamic pressure : pd ¼ qv2 ð4Þ
2
D2 p 4Q
flow : Q ¼ A v ¼ v!v¼ 2 ð5Þ
4 D p
Table 2. Results of measurements and calculations

Measurement – 50 l 1 inch – 0.0254 m 1=2 inch – 0.0127 m 5 mm – 0.005 m
1. 42 s 45 s 67 s
2. 42 s 46 s 68 s
3. 42 s
Avg. 41.67 s 45.5 s 67.5 s
Flow ½l=min 72 66 44.4
Flow ½m3 =s 0.0012 0.0011 0.00074
Power ½W 3.36 41.47 525.54
Velocity ½m=s 2.37 8.68 37.69
Dynamic pressure ½bar 0.028 0.37702 7.10
Calculating the forces on the shaft in a Pelton turbine is relatively straight forward.
The main forces come from the water jets, the runner and the shaft weight, the bearings,
and the drive system. Formula for the calculation of the jet force:
Fjet ¼ q v2o Ao ð1 þ sin 75 Þ ð6Þ
2np
vo ¼ vm þ vt ¼ vm þ rt ð7Þ
60
3.3 Generator
It was not easy to choose the most suitable generator that would be used to convert
mechanical energy into electricity. The first attempt was with the car’s alternator, but
this did not work because of an inadequate level of exploitation that the alternator had.
Our turbine had a sufficient number of rotation (rpm) for the alternator, but the alter-
nator’s starting current could not be overcome, which slowed down the operation of the
turbine, so that the alternator could not go into the generator mode and could not begin
to generate electricity. So we replaced it with an electric motor with permanent magnets.
Fig. 6. 3D model and the real- time model
Such a motor did not have the starting current, and it did not need power supply, but only
needed to connect it with the electric consumer and he would start to produce electricity.
The generator feeds a 50 W bulb, which was used as a consumer (Fig. 6).
3.4 Model of Pelton Turbine

It started from the 3D model that was done in Solid Works. Based on 3D models and
technical drawings, the construction itself began. The construction consisted of two
parts, a casing which is covered with plexiglass, in which the turbine was placed, and
the other part on which the bearings were placed, and the generator of electricity.
Transmission from the shaft to the electric motor is carried out over wheel belt, with a
transfer ratio of 3.5. As a fall simulator, a water pump was used, which pumped water
from the casing, into which the water was poured to some level, so that the casing itself
looked like an aquarium.
4 Conclusion
The Pelton turbine has a wide application range. Pelton wheels are most suitable for
making demonstration models, they are used for outputs of only a few hundred watts
[W], on heads down to several tens of meters. They can work at higher heads better
than any other type of turbine, and hold up their efficiency at partial flow better
compared to other types. One general fact is that the one kWh of energy produced in
hydro power plants is significantly cheaper than that in the thermal power plant, and
has a less negative impact on the environment. For this reason hydro power plants are
more popular and more desirable than the source of energy for a single country.
References
Books
1. Thake, J.: The Micro-Hydro Pelton Turbine Manual, Design, Manufacture, and Installation
for Small-Scale Hydropower (2000)
2. Tomović, R.: Osnove konstruisanja. Mašinski fakultet Podgorica (2015)
3. Durali, M.: Design of Small Water Turbines for Farms and Small Communities (1976)
Gas Escape from Combustion Chamber
to Crankcase, Analysis of a Set of Parameters
Affecting the Blow by
Cristiana Delprete1, Arian Bisha2, and Erjon Selmani2(&)

1
Politecnico di Torino, 10100 Turin, Italy
cristiana.delprete@polito.it
2
Universiteti Politeknik i Tiranes, 1000 Tirana, Albania
arian_bisha@fim.edu.al, erjon.selmani@polito.it
Abstract. One of the several losses of a combustion chamber is the gas leakage
toward the crankcase due to imperfect sealing of the rings. Commonly known as
blow by, it affects efficiency and emissions. A bibliographic review concerning
the phenomenon is mentioned and starting from that, the equations that rule the
ring dynamics, inter-ring pressures and mass flows are described and solved for
a diesel engine using ©RicardoRINGPAK Solver. Ring and groove dimension
together with engine speed and load were the investigated parameters. However,
blow by gasses shows to depend upon by many other factors.
Keywords: Blow by Internal combustion engines Piston rings
1 Introduction
Internal combustion engines plays an utmost importance on the transportation on road,

sea and often in air. The power obtained from fuel combustion is used only partly due
to several losses, namely thermodynamic and mechanic. Piston rings are the
mechanical elements interposed between the piston and the liner to guarantee air
tightness, oil distribution and oil scrapping. However, their tightness is not perfect and
part of the intake gas mixture is lost toward the crankcase, known as blow by gases. In
addition, some trapped gasses may go back to the combustion chamber during the
exhaust stroke and ejected as unburned hydrocarbons, also known as blow back. Piston
rings perform also secondary duties such as heat transfer. In 20th and 21st century,
problems of air pollution and fuel shortage have gained large importance due to
stringent regulations, and due to the increasing awareness of the pollution effect on the
climate change.
2 Ring Axial Dynamics
Piston rings are curved beams with an end gap that must seal the contact and guarantee
a radial force toward the cylinder liner. This force is an important element of the sealing
capacity of a ring but the gap is the first way of escape for the high-pressure gas. In [1],

https://doi.org/10.1007/978-3-319-90893-9_49
414 C. Delprete et al.
a historic review of piston ring technology is given. The standard solution for the
majority of the applications is a three-ring pack, made by two-compression rings ad one
oil-ring [2], however in literature different solutions have also been proposed [3]. In
order to compensate the thermal expansion in axial direction, ring axial width is smaller
than the respective groove height. This clearance is responsible for an additional way
for the gas to escape when rings are in motion. Eweiss [4] and Englisch [5] are among
the first scientific publications on this topic. They employed the ideal gas law applied to
an orifice and made the assumption of adiabatic gas flow (no heat exchange with
surrounding walls). Furuhama in [6] proved experimentally that the gas temperature in
a region is almost the same to the wall temperature (Isothermal). Namazian and
Heywood in [7] proposed more realistic models of the system, including ring motion,
the effect of the squeezed oil on the grooves and the friction force due to the oil in the
liner wall, according to Fig. 1.
Equation 1 describes the mechanical equi-
librium of forces, meanwhile Eq. 2 gives the
ideal gas law combined with the continuity law,
the flow was assumed to be fully laminar,
isothermal and compressible.
d2h
mr ¼ FP þ IP þ FF FS ð1Þ
dt2
mr = mass of the ring

h = ring-groove clearance
FP = Force due to gas pressure
IP = force due to inertia
FF = Force due to the oil in the ring-liner
interface Fig. 1. Ring axial movement
FS = Force generated from the squeeze of
the oil in the groove.

1 dV dP dmin dmout
P þV ¼ ð2Þ
RT dt dt dt dt
Similar models were also developed by Kuo et al. [8], Keribar et al. [9], Koszalka
[10], Tian [11, 12]. The most important forces are gas pressure and inertia, friction
remains almost constant and of low magnitude. The most critical situation is at high
speed and low load, where inter ring pressures are lower and inertial force are higher.
Under this situation the second ring flutters, increasing the amount of gas escaped in the
crankcase. Ring lift is not uniform over its circumference and seems to lift initially near
the gap. Tian in [12] explains in detail also the phases of the blow back.
Gas Escape from Combustion Chamber to Crankcase 415
2.1 Ring Radial Dynamics

The radial degree of freedom of a piston ring, is imposed by the necessity to com-
pensate thermal expansion. This motion of the ring gives the so-called “conformability”
to the cylinder bore, i.e. the capacity to adapt to the bore surface. In Fig. 2 are showed
forces acting in the radial direction of a ring cross section. When the inward force
overcomes the outward one, the ring face lifts from the liner and a direct path opens for
the gasses, this is called “radial collapse” and increases sharply the quantity of escaped
gas. Dukes [13], recorded experimentally that collapse occurs when ring lifts from
lower to upper groove flank. Tian in [12] and Rabutè et al. in [14], discovered that
radial collapse is sensitive to operating conditions, to the design of ring cross section
and to the extent of lubricated area on ring face. Lijima et al. [15] and Przesmitzki et al.
[16] made experiments and observed.
Fig. 2. Ring twist angle Fig. 3. Ring radial movement
2.2 Ring Twist

The third degree of freedom is the twist of the ring around its circumference axis.
Figure 3 shows the positive twist angle for a ring section. When twist angle is
accounted, the equation of the moments acting on the ring must include the effect of
asperity of the contact and oil squeeze. References [9, 12, 17, 18] addresses the topic.
Namazian [7] defines the blow by flow as a direct power loss and quantifies to
weigh between 0.5 and 1.2% of the total cylinder charge for a new engine. The power
and efficiency loss from blow by were calculated to lay between 2–7%. According to
Min [19] inter ring crevices are accounted to be responsible for nearly 38% of the total
hydrocarbon emissions meanwhile Rao [20] quantifies the effect of the gas leakage on
engine performances.
2.3 Simulation
In [21], Selmani et al. developed a similar analysis for the blow by, where the
investigated parameters were the ring gap variation, the ring mass variation and the
static twist variation. In the present work the effort was focused on the effect of ring and
groove geometry and volume variation and the effect of engine speed and load on ring
dynamics and gas blow by. Simulations were performed using RINGPAK Solver of
©
Ricardo Software and Matlab®, the test case is a turbodiesel engine of medium size.
Where not stated, engine speed used in simulations is 2000 r.p.m and full load con-
dition. In Tables 1 and 2 are summarized the variations of parameters adopted during
the simulations.
Table 1. Ring and groove geometry variations Table 2. Speed and load variation
Case Ax. Ax. Land 2° land Case B Speed r.p.m Load
A width width diameters height B_1 2000 Full
R1 R2
B_2 2000 Half
A_1 −7% −7% Original Original
B_3 2000 Zero
A_2 Original Original −3% Original
A_3 Original Original Original −20%
B_4 4000 Full
B_5 4000 Half
B_6 4000 Zero
2.4 Results
The results are grouped for each simulation case. In Fig. 4 are depicted the results for
the “base” condition, 2000 r.p.m at full load. Together with the ring positions, angles
and pressures are given also the mass flowrate of gas in the different grooves and the
cumulative blow by, expressed as fraction of the base cylinder gas mass at BDC
(0.2 * 10^−2 [kg] at 1 bar and 298 K).
Fig. 4. Results for the base case.

According to Fig. 4, at 2000 rpm and full load, top ring lifts at the end of the
expansion stroke, in correspondence with the second land pressure surpassing the first
land pressure (crossover point). From mid of exhaust to mid of intake stroke it stays
lifted, in contact with groove upper flank. The second ring stays on the lower groove
flank mean while third ring lifts even during the expansion stroke, driven by the inertia
force. The maximum pressure in land 2 is 5.767 [bar]. Since the minimum oil film
thickness is 5 µm, none of the rings undergoes to radial collapse. The total gas mass
flowrate below the third groove, which ends in the crankcase as blow by is equal to
0.0103117 [kg] for 1 cycle of engine operation (720° of crank angle). In Figs. 5, 6, 7
and 8 are depicted the results for the cases A and B.
Following the top-down direction for each case in Fig. 5, it is possible to see the
effect of geometry variation on the inter-ring volumes. In case A_1 we obtain groove
volume increase and in case A_2 we obtain an inter ring volume increase. In terms of
sealing capabilities, it is desirable to reduce crevice volumes. In fact, increasing
inter-ring volumes will lead to higher gas flow and inter-ring pressures, having second
and third ring lift. In case A_3 the effect is opposite to the two previous, reducing
second land height and consequently the volume. This reflects in a higher resistance to
the gas flow, thus a lower second land pressure and blow by rates. The top ring stays
lifted for a longer time in this case, although it is expected the opposite to occur. The
reason lays in the higher time the second land pressure takes to decrease.
Fig. 5. Results for case A.

Fig. 6. Results for Case B, 2000 r.p.m.
Results for the 2000 r.p.m. case at various loads (cases B_1 to B_3) are provided in
Fig. 6. In Fig. 7 are given the results for the 4000 r.p.m. and various loads (cases B_4
to B_6) meanwhile in Fig. 8 are given the gas flow rates in each inter-ring region.
According to Fig. 6, at the medium speed of 2000 r.p.m. when load decreases,
combustion chamber pressure will decrease. As a consequence, inter ring pressures, gas
flows and forces in axial directions will be lower. For this reason, in rings 2 and 3,
where this reduction is more marked, inertia force will prevail, making rings 2 and 3
become instable and expect axial flutter more frequently.
At the regime of high speeds, 4000 r.p.m. the situation is considerably more
complicated, as Fig. 7 shows. When the engine speed is increased, the cycle time will
decrease, doing so, the time for the gas to flow is reduced accordingly. But, when this
occurs, it means that the same gas cannot flow using the usual paths, i.e. ring gaps and
ring-groove flanks, but will search other paths. Under this situation, high pressure gas
will press the ring face and under certain conditions, will push it inwards the groove,
causing a radial collapse and opening a direct path for the gas to flow. This situation
occurs frequently under the 4000 r.p.m. case, as Fig. 6 shows.
Fig. 7. Results for case B, 4000 r.p.m.
Fig. 8. Gas flow in several regions for cases A and B.

On the other hand, inertia force quadruplicates when engine speed doubles. It
means that at higher speeds and no loads, inertia force will be high enough to compete
with low pressures, making them to be instable and undergo both axial flutter and radial
collapse.
Figure 8 provide the gas flow rates above and below each ring groove. Since the
result is obtained from the solution of Eq. (2), the gas flow will follow the trend of the
inter ring pressures. Consequently, the B_6 case will mirror perfectly the axial dynamic
of the piston rings.
2.5 Discussion
In order to have a direct impact the large amount of results and data obtained from
simulations, Tables 1 and 2 will be updated with the numerical result of the cumulative
amount of gas blow by for each case (Tables 3 and 4).
Referring to the B_1 as the base case is possible to compare the blow by, expressed
in kg per one cycle of engine operation, to the other test cases under examination. Case
A_2, where land diameters were reduced, show the worst result in terms of sealing
capacity, reaching the highest value of calculated flow rates per cycle. The best in class
is the B_5 case, high speed and half load. It seems that inertial forces are more properly
equilibrated by gas pressure force, obtaining a better balance. Between the two case
studies, the most frequent in normal engine operation is the 2000 r.p.m. speed, thus,
reducing the inter-ring volumes has a positive effect in terms sealing efficiency.
Table 3. Cumulative blow by for case A

Case Ax. Ax. 1°,2°, 3° land 2° land Blow by
A width-R1 width-R2 diameter height [kg/cycle]
A_1 −7% −7% Original Original 0.009821
A_2 Original Original −3% Original 0.0481848
A_3 Original Original Original −20% 0.0094545
Table 4. Cumulative blow by for case B

Case B Speed r.p.m Load Blow by [kg/cycle]
B_1* Medium (2000) Full 0.0103116
B_2 Medium (2000) Half 0.0075031
B_3 Medium (2000) Zero 0.00669085
B_4 High (4000) Full 0.0030745
B_5 High (4000) Half 8.5917*10^−6
B_6 High (4000) Zero 0.021165
*
B_1 is the base case
3 Conclusion
In the present paper the behaviour of piston rings and the inter ring gas flows was in the
focus. According to previous studies, piston ring is considered as a curved beam with
three degree of freedom, two translations and one rotation around its central axis.
Furthermore, gas flow between inter-ring crevices is considered fully laminar,
isothermal and compressible and ideal gas equation is employed for the solution. Two
parameters were investigated in relation to their effect on blow by, piston lands,
grooves and ring geometry, engine speed and load. According to results, smaller
inter-ring volumes have positive effects on ring dynamics and gas flow, reducing it
considerably. Smaller volumes are synonymous to higher resistance to the flow. On the
otherhand, higher volumes are responsible also for higher housing capacities and not
only for lower resistance to flow. At medium speed of 2000 r.p.m., blow by increased
with load, at high speed, blow by decreased with load, obtaining the lowest value at
4000 r.p.m and half load. When load increases, inter ring pressure increases due to the
higher pressure developed in the combustion chamber. When instead speed increases,
inertia force quadruplicates, making the rings to become highly instable at higher
speeds and low loads.
References
1. Economou, P., Dowson, D., Baker, A.: Piston ring lubrication-Part 1. The historical
development of piston ring technology. J. Lubr. Technol. 104, 118–126 (1982)
2. Ellermann, J., Röhrle, M.D., Schelling, H.: Oil consumption and blowby of truck diesel
engines-test bench results. SAE Technical Paper, No. 810937 (1981)
3. Yoshida, H., Sugihara, H., Kusama, K.: Practical use of two piston ring set for gasoline
engine. SAE Technical Paper, No. 950817 (1995)
4. Eweiss, M.: Forsch-h. 371 (1935)
5. Englisch, C.: ATZ, Jg. 41, HT, 22, S. 579 (1938)
6. Furuhama, S., Tosio, T.A.D.A.: On the flow of gas through the piston-rings: 2nd Report, The
character of gas leakage. Bull. JSME 4(16), 691–698 (1961)
7. Namazian, M., Heywood, J.B.: Flow in the piston-cylinder-ring crevices of a spark-ignition
engine: effect on hydrocarbon emissions, efficiency and power. SAE Technical Paper,
No. 820088 (1982)
8. Kuo, T., et al.: Calculation of flow in the piston-cylinder-ring crevices of a
homogeneous-charge engine and comparison with experiment. SAE Technical Paper,
No. 890838 (1989)
9. Keribar, R., Dursunkaya, Z., Flemming, M.F.: An integrated model of ring pack
performance. J. Eng. Gas Turbines Power 113, 382–389 (1991)
10. Koszałka, G.: Modelling the blowby in internal combustion engine. Part I. A mathematical
model. Arch. Mech. Eng. 51(2), 245–257 (2004)
11. Tian, T., et al.: Modeling piston-ring dynamics, blowby, and ring-twist effects. J. Eng. Gas
Turbines Power 120, 843–854 (1998)
12. Tian, T.: Dynamic behaviours of piston rings and their practical impact. Part 1: ring flutter
and ring collapse and their effects on gas flow and oil transport. Proc. Inst. Mech. Eng. Part J
J. Eng. Tribol. 216(4), 209–228 (2002)
13. de Dykes, P.K.: Piston ring movement during blow-by in high-speed petrol engines. Proc.
Inst. Mech. Eng. Automob. Div. 1(1), 71–83 (1947)
14. Rabutè, R., Tian, T.I.A.N.: Challenges involved in piston top ring designs for modern SI
engines. J. Eng. Gas Turbines Power 123(2), 448–459 (2001)
15. Iijima, N., et al.: An experimental study on phenomena of piston ring collapse. SAE
Technical Paper, No. 2002-01-0483 (2002)
16. Przesmitzki, S., Tian, T.: An Experimental study of the time scales and controlling factors
affecting drastic blow-by increases during transient load changes in SI engines. SAE Paper
2008-01-0794 (2008)
17. Ruddy, B.L., Parsons, B., Dowson, D., Economou, P.N.: The influence of thermal distortion
and wear of piston ring grooves upon the lubrication of piston rings in diesel engines. In:
Proceedings of the 6th Leeds-Lyon Symposium on Tribology. ASME, NY (1979)
18. Richardson, D.E.: Comparison of measured and theoretical inter-ring gas pressure on a diesel
engine. SAE Technical Paper, No. 961909 (1996)
19. Min, K., Cheng, W.K., Heywood, J.B.: The effects of crevices on the engine-out
hydrocarbon emissions in SI engines. SAE Technical Paper, No. 940306 (1994)
20. Rao, V.K., Gardiner, D.P., Bardon, M.F.: Effects of gas leakage and crevices on cold starting
of engines. SAE Technical Paper, No. 940078 (1994)
21. Selmani, E., Bisha, A., Delprete, C.: Blow by gases from combustion chamber to crankcase.
A dynamic analysis of parameters affecting the phenomenon. In: Proceedings of the 1st
International Conference on Engineering and Enterpreneuship, ICEE 2017, Tirana,
ALBANIA (2017)
Comb-Based Decimator for Multiples-of-Five
Decimation Factors
Gordana Jovanovic Dolecek1(&) and Isak Karabegovic2

1
Institute INAOE, 72740 Puebla, Mexico
gordana@ieee.org
2
Abstract. This paper presents novel comb-based decimator for multiples-of-

five decimation factors. The proposed decimation structure has two stages. In
the first stage is a comb filter decimated by one fifth of the overall decimation
factor, while in the second stage is a comb decimated by five. Additionally, the
simple multiplierless filter is inserted in the second stage. This filter improves
alias rejection in all folding bands except in bands which are multiples of five, in
comparison with the original comb filter. Like original comb filter, the proposed
filter is multiplierless.
Keywords: Decimation Alias Comb filter Folding bands

Multiplierless filters
1 Introduction
Analog-to-digital (A/D) converters are the interface between the analog world and the
digital signal processing. The most efficient A/D converters are Sigma Delta (SD) A/D
converters [1] due to the oversampling of analog signal with a frequency much larger
than the minimum sampling frequency required for preserving the quality of the
information (Nyquist frequency). The rate of the oversampled signal must be decreased
to the Nyquist frequency, to be efficiently processed by the DSP (Digital Signal Pro-
cessing). This process is performed in digital format and is called down-sampling.
Down-sampling may introduce the aliasing effects that deteriorate the signal and have
to be eliminated by the decimation filter. Figure 1 shows the decimation structure,
where the second block presents downsampling, and M is the decimation factor.
The goal is to design decimation filters with reduced complexity while performing
efficient aliasing rejection [1, 2]. The most simple decimation filter is comb filter which
has all coefficients equal to unity, and consequently does not require multipliers. Its
system function in the recursive form is given as,
K
1 1 zM
HðzÞ ¼ : ð1Þ
M 1 z1

https://doi.org/10.1007/978-3-319-90893-9_50
424 G. Jovanovic Dolecek and I. Karabegovic
Decimation
filter ↓M
Fig. 1. Decimation.
Similarly, the nonrecursive form is:

" #K
1MX1
HðzÞ ¼ zk ; ð2Þ
M k¼0
where M is a decimation factor and K is order of the filter and presents the number of
the cascaded comb filters.
The very simple comb structure, called CIC (Cascaded-Integrator-Comb), structure,
[2] is derived from (1). It consists of the cascaded integrators and combs, separated by
downsampler.
In order to reject aliasing, the magnitude characteristic of the comb
K

Hðejx Þ ¼ sinðxM=2Þ ; ð3Þ
M sinðx=2Þ
must have high attenuations in the bands around of comb zeros, called folding bands.
However, comb filter itself does not provide enough alias rejection. Different methods
have been proposed to improve comb aliasing rejection keeping its simplicity, [3–6].
The goal of this paper is to present a simple method to improve alias rejecting of a
two-stage comb filter for decimation factors which are the factors-of-five. This is
achieved by inserting a simple multiplierless filter at the second stage.
The rest of the paper is organized in the following way. Next section presents a
two-stage structure and the mathematical background for improving alias rejection.
Section 3 describes the proposed filter and its structure. The magnitude response is
compared with the original comb filter and magnitude response from method in [5].
2 Two-Stage Structure and Mathematical Background
2.1 Two-Stage Structure

Considering that the decimation factor M is a multiple-of-five we write:
M ¼ 5 ðM=5Þ: ð4Þ
Using (1) and (4) we write:

K K K
1 1 zM 1 1 zM=5 1 1 zM
HðzÞ ¼ ¼ ¼ H1 ðzÞH2 ðzM=5 Þ; ð5Þ
M 1 z1 M=5 1 z1 5 1 zM=5
Comb-Based Decimator for Multiples-of-Five Decimation Factors 425
where:
K K
1 1 zM=5 1 1 zM
H1 ðzÞ ¼ ; H2 ðzM=5 Þ ¼ : ð6Þ
M=5 1 z1 5 1 zM=5
Using the multirate identity [2] we get the structure shown in Fig. 2. Next issue is
to propose simple filter to improve alias rejection in this structure.
Fig. 2. Two-stage comb decimation structure.
2.2 Improving Alias Rejection of Two-Stage Structure by Simple Filter

We take advantage of the special relation between the coefficients of a symmetric
polynomial, and the position of its zeros on the unit circle in complex z-plane [6]. This
special relation is explored in [6] for the efficient comb filter design. Using the idea
from [6] we propose to insert in the second stage simple multiplierless filter denoted as
G(z). Omitting the normalization factor we have:
" #2
X
4
m
GðzÞ ¼ z 0:5z5 : ð7Þ
m¼0
Figure 3 shows pole-zero plots of the overall comb filter and filter G(zM/5) for
M = 15.
Fig. 3. Pole-zero plots of comb and filter G(z3).
Note that the filter G(z3) introduces zeros around the original comb zeros in the
folding bands 1, 2, 3, and 4, and thus increase alias rejections in those folding bands.
This idea is illustrated in the Fig. 4 which contrasts magnitude responses of the original
comb and comb cascaded with filter (7).
Note that the alias rejection is improved in all folding bands except in fifth folding
band of the original comb filter, due to inserting the filter (7) in second stage.
Fig. 4. Magnitude responses.
3 Proposed Filter
3.1 System Function
The proposed decimator has the simple filter (7) at the second stage of the structure
from Fig. 2. From (1), (6) and (7) we get the system function of the proposed filter:
K M=5
1 1 zM
Hm ðzÞ ¼ G zM=5 ¼ H1 ðzÞH2 zM=5 G z : ð8Þ
M 1 z1
Method is illustrated in the following example.

Example 1. Consider decimation factor M = 20 and K = 4. The first and second
stages are decimated by 4 and 5, respectively.
Figure 5 contrasts the magnitude responses of the proposed filter and the original
comb filter. Note that the proposed filter has better alias rejection than comb filter.
Fig. 5. Magnitude responses in Example 1.

Comb-Based Decimator for Multiples-of-Five Decimation Factors 427
3.2 Structure
From (8) we get structure shown in Fig. 6.
Fig. 6. Proposed two-stage structure.
Note that the only difference between the proposed structure from Fig. 6 and
two-stage comb structure from Fig. 2, is in the second stage, in which the filter G(z) is
cascaded with the comb filter H2(z). However, this simple filter in second stage results
in a significant improvement in comb alias rejection, as shown in Example 1.
Combs in both stage can be implemented either in recursive or nonrecursive form.
3.3 Comparison
Comparison with Method in [5]
The novel efficient two-stage structure for even decimation factors was recently pro-
posed in [5]. The decimation factors of the first and second stages are M/2, and 2.
At the second stage is inserted multiplierless filter [1 + (2 − a)z−1+z−2], where a = 2−k,
and k is an integer. For the sake of comparison we consider an even value of M which is
a factor of five, considering M = 10, and K = 3. The magnitude responses are con-
trasted in Fig. 7. The proposed method provides better alias rejections in all folding
bands except in the last one. However, the proposed method is more complex, being
the filter G(z) has lesser order and works at lower rate. Additionally, method in [5] is
applicable for even decimation factors, while the proposed method works only for
factors of five decimations.
Fig. 7. Comparison with method in [5].

4 Conclusion
This paper presents novel two-stage structure for the decimation factors which are
multiples of five. The first stage is a comb filter decimated by M/5, while the second
stage is comb decimated by 5, where M is the overall decimation filter. Simple mul-
tiplierless filter, inserted at the second stage, improves alias rejections in all folding
bands which are not multiple of five.
References
1. Laddomada, M.: Comb-based decimation filters for sigma-delta A/D converters: novel
schemes and comparisons. IEEE Trans. Sig. Process. 55(5), 1769–1779 (2007)
2. Jovanovic Dolecek, G. (ed.): Multirate Signal Processing: Design and Applications. IGP,
Hershey (2003)
3. Jovanovic Dolecek, G. (ed.): Advances in Multirate Systems. Springer, New York (2017)
4. Coleman, J.O.: Chebyshev stopbands for CIC decimation filters and CIC-implemented array
tapers in 1D and 2D. IEEE Trans. Circ. Syst. I: Regul. Pap. 59(12), 2956–2968 (2012)
5. Jovanovic Dolecek, G., Garcia Baez, R., Laddomada, M.: Design of efficient multiplierless
modified cosine-based comb decimation filters: analysis and implementation. IEEE Trans.
Circ. Syst. I Regul. Pap. 64(5), 1051–1063 (2017)
6. Jovanovic Dolecek, G., Dolecek, L.: Exploiting features of symmetric polynomials for
improved comb filter design. In: Proceedings of IEEE SPA 2016, Poznan, Poland, 26–29
September 2016
The Filter-Compensation Device Applications
to the AC 25 kV 50 Hz AC of Serbian Railways
Branislav Gavrilovic ✉ and Zoran Bundalo

( )
Railway College of Vocational Studies, ZdravkaCelara 14, Belgrade, Serbia

gavrilovicbranislav5@gmail.com, cheminot@gmail.com
Abstract. New technologies in rail, including devices for infrastructure and

transport, have meant improved safety, comfort, and speed of services. It has also
brought new problems for the AC 25 kV 50 Hz traction power supply system with
a specific filter-compensation device in traction substations of the Serbian Rail‐
ways. These problems have resulted in some specific behavior issues during
operation modes. For these reasons, research activities during last several years
in engineering departments at the Railway College of Vocational Studies in
Belgrade have been focused on studying the behavior of the traction system during
selected operation states with traction vehicles with diode converters. The main
findings of these problems are detailed in this paper.
Keywords: AC railway · Filter-compensation device · Power quality
1 Introduction
The general configuration of the Serbian AC 25 kV 50 Hz traction power supply system

with rail vehicles for transport in Fig. 1 contains the following:
• contractor feeding line of 110 kV,
• traction substation with filter-compensation device (FCD),
• catenary (the whole structure—contact line) and
• electric vehicles.
The electromagnetic compatibility is discussed more and more. Therefore, usage of
the FCD in traction substations is necessary. This FCD is utilized for power factor
corrections and to reduce current harmonics caused by older electric single-phase loco‐
motives with diode converters, which are still very common on Serbian railways, see
Figs. 2 and 3. This FCD during transient effects also influences the behavior of the
traction system, see Fig. 4.

https://doi.org/10.1007/978-3-319-90893-9_51
430 B. Gavrilovic and Z. Bundalo
Fig. 1. The AC 25 kV 50 Hz traction power supply system in the Serbian Republic
Fig. 2. Locomotive type ZS 441 with a diode rectifier, (inductive power factor- PF ≅ 0.84) and
locomotive type ZS 444 with a thyristor-diode bridge, (PF ≅ 0.77)
Fig. 3. Electric multiple unit FLIRTs (Fast Light Innovative Regional Train) type ZS 413/41
with three-phase induction traction motors and input an converter filter (PF ≅ 0.87)
The Filter-Compensation Device Applications to the AC 25 kV 50 Hz AC 431
Fig. 4. Current waveform at 25 kV side of railway substation transformer of the electric

locomotives with diode rectifiers [1]
Electric locomotives with diode rectifiers generate all odd current harmonics (i.e.
3rd, 5th, 7th, and so on), see Fig. 5. The current harmonics pass through a catenary, traction
substation transformer, a 110 kV contractor feeding line and then to the main power
supply system.
Fig. 5. Current harmonics at 110 kV side of railway substation transformer [1]
The harmonics pass through the catenary section, independently of the impedance
of the external main power supply system, and then they pass through a series of alternate
impedance of the traction substation transformer, which are changed only by transmis‐
sion ratio of this transformer. According to Ohm’s law, voltage harmonics originate
from the input alternate impedance of the external main power supply system. The
harmonic currents of various frequencies cause a voltage drop from the impedance of
the main power supply system and voltage deformation [2].
The direct results from the above lead to a:

• rise in network losses,
• drop of active power supply, thereby resulting in an efficiency drop.
These direct results can create others problems with their system configurations:
• Creation of system resonance, which usually produces increased current or voltage.
• Faulty function of protections, measuring equipment and registering equipment.
• Interference of telecommunication equipment and control circuits.
The FCD helps eliminate 3rd, 5th, 7th, and so on, harmonics and compensates reactive
power. The FCD used in Serbian Railways consists of 28 branches of capacitor banks
for compensation of inductive reactive power consumed by electric locomotives (total
QC = 2716 kVAr). He is connected to 25 kV network via power transformer 2.7 MVA
(27.5/0.69 kV). One has a resonant frequency of 150 Hz and the other one, 1400 Hz.
For first harmonic, both of these circuits have a capacitive character, and can thus be
used to compensate reactive power. Because their capacitive reactive power is too high,
a decompensation reactor is connected to the circuit. Power in the reactor is phase
controlled by a thyristor controller, which makes it possible to set power factor to a
desired value very quickly [7]. A simplified diagram of the filter-compensation device
for compensation of capacitive reactive power of the 25 kV contact network (QL = 30
kVAr) can be seen in Fig. 6.
Fig. 6. The filter-compensation device for compensation of capacitive reactive power of the
25 kV contact network (total QL = 30 kVAr).
This FCD significantly suppresses on 3rd to 29th current harmonics, and compensates
power factor very well. In the Serbian Republic, the required value of power factor is
0.95–1 (inductive). Traction transformer substations equipped with a filter-compensa‐
tion device have a power factor of about 0.98.
2 Control of the Filter-Compensation Device Installed

at Substation
The 25 kV AC electrified railway system can be considered as a lumped circuit. In case

of the overhead transmission line, the conductance in the parallel admittance of π-circuit
can be neglected. Figure 7 shows the single-phase feeder line with the series resistance,
the series inductance, and the parallel capacitance. In addition, the feeder line consists
of three 10 km-long π-sections, and the traction substation transformer is described as
an AC voltage source with internal impedance. The locomotive is represented by a SCR
bridge converter which generates the harmonic currents and the reactive power demand.
Fig. 7. Modeling of the feeder line
Figure 8 shows the AC railway system with the FCD provided at the traction substa‐
tion. The load current flowing into the locomotive, the FCD current, and the DC-link
voltage are measured for the controller input signals. Hysteresis control method is
employed for the current control of the FCD.
Fig. 8. System configuration with FCD installed at the traction substation
In this compensation system, in order to look into the relationship between the loco‐
motive voltage and the reactive power compensation, an equivalent circuit can be
described at the fundamental frequency as shown in Fig. 9.
Fig. 9. Equivalent circuit of AC railway network with compensation system installed at the
traction substation
For the sake of simplicity, the source impedance and the parallel capacitance of line
impedance are not considered. By defining VL = VL + j0 as the reference phasor, the
voltage drop across the feeder line is calculated as [3]:
R ⋅ PL + X ⋅ QL X ⋅ PL − R ⋅ QL
ΔV = (R + jX) ⋅ IL = +j = ΔVR + jΔVX (1)
VL VL
where, the bold denotes phasor, and VL, R, X, PL, and QL mean the load voltage, the
resistance of feeder line, the reactance, the active power of the load, and the reactive
power, respectively. From (1), the phasor diagram of the feeding system without
compensation can be depicted as shown in Fig. 10.
Fig. 10. Phasor diagram with no compensation
Using the phasor diagram, the square of source voltage magnitude is derived as:
( )
VS = VL + ΔV = VL + ΔVR + jΔVX (2)
( )
|V |2 = V 2 + |ΔV|2 + 2 ⋅ R ⋅ P + X ⋅ Q
| S| L L L (3)
If the FCD is an ideal compensator and QFCD = −QL is satisfied, the reactive power
supplied from the source becomes zero and (3) can be rewritten as:
( )
|V |2 = V 2 + |ΔV|2 + 2 ⋅ R ⋅ P − X ⋅ Q
| S| L L FCD (4)
Since the locomotive consumes the( inductive reactive ) power, the value of QFCD is
negative and the value of |ΔV|2 + 2 ⋅ R ⋅ PL − X ⋅ QFCD is always positive, not zero.
The load voltage is not equal to the source voltage even when any compensation is taken.
It is known that even though the unity power factor is achieved with the help of the FCD,
the magnitude of the source and load voltage cannot be the same. When the FCD is
installed at the traction substation, the associated phasor diagram is shown in Fig. 11.
Fig. 11. Phasor diagram with FCD installed at the traction substation
When the FCD is installed at the traction substation, it is possible that the harmonic
currents are effectively compensated by measuring the load current directly. The control
objectives of the FCD installed at the traction substation are to compensate the harmonic
currents and reactive power demand for improving the input power factor.
2.1 Harmonic Compensation
The single-phase voltage and current signals can be handled on the stationary reference
frame by assuming the imaginary values 90° phase-shifted. The imaginary value can be
calculated from the practical value through the various methods such as the Transport
Delay, the Hilbert Transform, the Inverse-park Transform, and the Adaptive Notch Filter
[4]. On the basis that the Inverse-park Transform and the Adaptive Notch Filter are
reported to be relatively superior to the other methods, the Inverse-park Transform is
applied in this paper [5]. The α – β axis values of source voltage and load current in
Fig. 8 can be expressed as:
℘
∑
vS (t) ≅ vS𝛼 = VS cos (𝜔 ⋅ t) + VS (n) ⋅ cos (n ⋅ 𝜔 ⋅ t) (5)
n=3,5,7,…
℘
∑ ( )
iL (t) ≅ iL𝛼 = IL cos (𝜔 ⋅ t − 𝜑) + IL (n) ⋅ cos n ⋅ 𝜔 ⋅ t − 𝜑n (6)
n=3,5,7,…
℘
∑
vSI (t) ≅ vS𝛽 = VS sin (𝜔 ⋅ t) + VS (n) ⋅ sin (n ⋅ 𝜔 ⋅ t) (7)
n=3,5,7,…
℘
∑ ( )
iLI (t) ≅ iL𝛽 = IL sin (𝜔 ⋅ t − 𝜑) + IL (n) ⋅ sin n ⋅ 𝜔 ⋅ t − 𝜑n (8)
n=3,5,7,…
where, VS, IL, and n denote the maximum value of source voltage, the maximum value
of load current, and the harmonic order, respectively. The α-axis value is measured from
the system and the β-axis value is calculated from the measured α-axis value. From (5)–
(8), it is possible that the source voltage and the load current vectors can be assumed
and expressed on the stationary reference frame coordination as shown in Fig. 12.
Fig. 12. Voltage and current vector representation
The voltage and current vectors in Fig. 12 can be projected onto the synchronous
reference frame rotating at the source frequency. The corresponding d-q axis values of
the load current are derived as:
[ ] [ ] [ ] [ ]
iLd i cos (𝜔 ⋅ t) sin (𝜔 ⋅ t) i
= T L𝛼 = ⋅ L𝛼
iLq iL𝛽 − sin (𝜔 ⋅ t) cos (𝜔 ⋅ t) iL𝛽
⎡ I cos (𝜑) + ∑ I (n) ⋅ cos ((n − 1) ⋅ 𝜔 ⋅ t − 𝜑 ) ⎤
℘
[ ] (9)
⎢ L n=3,5,7,…
L n
⎥ ⃗iLd + iLD
= ⎢ ℘ ( )⎥ =
∑
⎢ IL sin (𝜑) + IL (n) ⋅ sin (n − 1) ⋅ 𝜔 ⋅ t − 𝜑n ⎥ iLq + iLq
⎣ n=3,5,7,… ⎦
where, the upper bar and the upper tilde means the DC value and the AC value of d-q
axis. It is known that the load current has both the fundamental and harmonic compo‐
nents. The AC value depends on the harmonic contents of load current. The DC value
represents the fundamental load current and is easily obtained using the low pass filters
(LPF) without any phase-shifting. When the d-q axis DC values are subtracted from the
d-q axis load currents, only the d-q axis AC values to be compensated remain. These
components are provided as the current reference for the harmonic compensation to the
controller.
2.2 Reactive Power Compensation

If the harmonic currents have already been compensated by the FCD installed at the
traction substation and the d-axis is aligned on the source voltage vector as shown in
Fig. 12 the source voltage and current are sinusoidal, and expressed as:
[ ] [ ] [ ]
iSd iLd IS ⋅ cos 𝜑
= = (10)
iSq iLq −IS ⋅ sin 𝜑
[ ] [ ] [ ]
vSd vSd VS
= = (11)
vSq vSq 0
where, IS denotes the magnitude of source current. The source voltage and current vector
are represented by the d-q axis components on the synchronous reference frame as
follows:
vS = vSd + jvSq (12)
iS = iSd + jiSq (13)
From (12) and (13), the following equation can be considered.

∗ ( ) ( )
vS ⋅ iS = vSd ⋅ iSd + vSq ⋅ iSq + j vSq ⋅ iSd − vSd ⋅ iSq (14)
By substituting (10) and (11), (14) is expressed as follows:

∗
vS ⋅ iS = VS ⋅ IS cos (𝜑) + jVS ⋅ IS sin (𝜑) = Pdq + jQdq (15)
Using (15), the single-phase instantaneous power can be represented as follows:
1 1
p(t) = V ⋅ I ⋅ cos (𝜑) ⋅ (1 + cos (2𝜔 ⋅ t)) + VS ⋅ IS ⋅ sin (𝜑) ⋅ sin (2𝜔 ⋅ t)
2 S S 2
1 1
= Pdq ⋅ (1 + cos (2𝜔 ⋅ t)) + Qdq ⋅ sin (2𝜔 ⋅ t) (16)
2 2
1 1
= vSd ⋅ iSd ⋅ (1 + cos (2𝜔 ⋅ t)) + vSd ⋅ iSq sin (2𝜔t)
2 2
The first term of (16) is defined as the single-phase instantaneous active power and
the second term is defined as the single-phase instantaneous reactive power [6]. It is
known from (16) that when the source voltage is constant, the single-phase instantaneous
passive or active power depends on the d-axis current and the single-phase instantaneous
reactive power relies solely on the q-axis current. Therefore, the FCD installed at the
traction substation can compensate the reactive power demand by controlling the q-axis
value of source current to be zero instead of controlling the instantaneous reactive power.
Figure 13 shows the overall control scheme of the FCD installed at the traction
substation. The d-q transform block includes both the single-phase to two-phase trans‐
formation and the transformation into the synchronous reference frame. The reference
of q-axis source current in the reactive power compensation loop is set at zero. The error
between the reference and the actual value of the q-axis source current enters the PI
controller, and then it is added to the q-axis current reference for the harmonic compen‐
sation because the q-axis source current determines the instantaneous reactive power.
No additional current sensor to measure the source current is required. Among the
outputs of inverse d-q transform block, only the a-axis value becomes the current refer‐
ence for the hysteresis controller. The DC-link voltage regulator plays a role in compen‐
sating the power losses of the FCD as well as the voltage regulation.
Fig. 13. Overall control scheme of FCD installed at the traction substation.
3 Modeling of Electric Railway System with Filter-Compensation

Device
A model of electric railway system connected to 110 kV network was developed in order
to determine power quality parameters of voltage and current. A model consists of elec‐
tric railway substation with FED and contact line feeding electric locomotives equipped
with diode rectifiers. Figure 14 shows the model in which is used for analysis of elec‐
tromagnetic transients [1]. An electric railway substation consists of one 110/25 kV
transformer with rated power 7.5 MVA which is connected to the transmission grid. The
impedance of 110/25 kV power transformer referred to the 110 kV side is R = 0.5 Ω
and L = 4 mH (uk% = 10%). 110 kV transmission network is represented by Thevenin
equivalent (impedance in series with voltage source). The positive and zero sequence
impedance was calculated from single-phase and three-phase short-circuit currents.
Fig. 14. Model of electric railway system with FCD
The catenary system was modeled using a frequency dependent J. Marti model which
is based on the approximation of the line characteristic impedance Z(ω) and uppropa‐
gation function A(ω) by rational functions of the higher order. Ground resistivity was
assumed 100 Ωm.
4 Analysis of the Simulation Results
Constant speed of the diode locomotive was analyzed. Electric railway system is
connected between two phase of the 110 kV network. All calculated values relate to the
single diode locomotive 1 km away from the electric railway substation. Voltage and
current waveforms were calculated on 25 kV and 110 kV level at the railway substation.
The diode electric locomotive causes voltage distortion in the 25 kV catenary system.
Figure 15 shows voltage waveform Urms = 27.9 kV, Fig. 16 active power calculated
Prms = 1.3 MW and Fig. 17 reactive power calculated Qrms = 511.8 kVAr on 25 kV side
of railway substation transformer without compensation.
Fig. 15. Voltage waveform on 25 kV side of railway substation transformer without

compensation level Urms = 27.9 kV
Fig. 16. Active power calculated at 25 kV level in electric traction substation without
compensation: Prms = 1.3 MW
Fig. 17. Reactive power calculated at 25 kV level in electric traction substation without
compensation: Qrms = 511.8 kVAr
The harmonic compensation performance is illustrated in Fig. 18. It is seen that the
input current becomes perfectly sinusoidal after the harmonic compensation starts.
Fig. 18. Harmonic compensation (a) Substation voltage and current; (b) Compensation current.
Figure 18 shows the performance of the harmonic and the reactive power compen‐
sation when the FCD is installed at the traction substation with the proposed control
strategy. It is shown that when the harmonic and the reactive power compensation are
provided, there exists no phase deviation between the source voltage and the source
current, whereas the result with only the harmonic compensation shows a certain phase
shift.
5 Conclusion
This paper presents a study on the FCD applications to the AC 25 kV 50 Hz AC of

Serbian Railways and proposes a novel control algorithm based on the synchronous
reference frame. The results show that the proposed FCD control algorithm properly
compensates the power quality problems in the AC railway systems, such as the
harmonic current pollution, the voltage perturbation, and the reactive power demand.
When the FCD is installed at the traction substation, the performance to compensate the
harmonic currents is vastly superior. Also, the reactive power demand can be completely
compensated so that the input power factor is nearly unity. However, the voltage drop
and the voltage sag cannot be fully compensated.
The control strategy proposed by the authors allows the voltage to be maintained
within the prefixed range, increasing the number of locomotives that can be fed by a
single track feeder. Moreover, by tuning the FCDs on 3rd to 29th harmonic, the voltage
form factor can be improved and, consequently, the power that can be drawn by
locomotives increases. Practical tests on scale models of the system and software simu‐
lations have proved the effectiveness of the solution proposed both for steady state and
transients situations.
References
1. Župan, A., Tomasović Teklić, A., Filipović-Grčić, B.: Modeling of 25 kV electric railway
system for power quality studies. In: Proceedings of EuroCon 2013, Zagreb, Croatia, pp. 844–
845, 1–4 July 2013
2. Hlava, K.: The limitation of FCD effect to centralized ripple control signal of electric energy
contractor. Report no. D 237 4026, part no. 1, 2 TÚDC Department EMC Prague (1996). (in
Czech)
3. Acha, E., Agelidis, V., Anaya-Lara, O., Miller, T.: Power Electronic Control in Electrical
Systems, Chap. 2, p. 45. Newnes, Oxford (2002)
4. Silva, S., Lopes, B., Cardoso Filho, B., Campana, R., Boventura, W.: Performance evaluation
of PLL algorithms for single-phase grid-connected systems. In: Proceedings of IEEE Annual
Conference on Industry Applications, pp. 2259–2263 (2004)
5. Santos Filho, R., Sixas, P., Cortizo, P., Torres, L.: Comparison of three single-phase PLL
algorithms for UPS applications. IEEE Trans. Ind. Electron. 55(8), 2923–2932 (2008)
6. Saitou, M., Matsui, N., Shimizu, T.: A control strategy of singlephase active filter using a novel
d-q transformation. In: Proceedings of IEEE International Conference on Industry
Applications, pp. 1222–1227 (2003)
7. Sopov, V., Biryukov, V., Prokushev, Y., Rylov, Y.: Electric transport vehicle power-supply
system analysis with various traction networks topologies. In: Proceedings of Strategic
Technologies (IFOST 2008), Novosibirsk, Tomsk, pp. 462–464 (2008)
Optimization of Water Supply System
Using Software EPANET 2.0
Aleksandar Košarac ✉ , Dejan Romić, Goran Orašanin, and Jovana Blagojević

( )
University of East Sarajevo, 71123 East Sarajevo, Bosnia and Herzegovina

kosarac88@hotmail.com, dejan.romic@gmail.com,
goranorasanin@yahoo.com, jovana.pajkic@gmail.com
Abstract. Concepts of planning water supply networks in countries in transition

were based at the beginning and during 20th Century, when most of the existing
objects were built. In that period main aim of water supply system was to provide
sufficient amount of water for the population and economy, so the water objects
were planned based on the input available at that time. Upgrades of water supply
systems were done in different scale, without systematic approach to an integral
overview to a weakness in whole system. New concept of Water Utilities, which
has appeared in countries in transition, changes approach where in old systems
water utilities have to satisfy demand driven consumption towards the system
where utilities start to managing with demand managed consumption. First task
to use new concept is introduction of efficiency in management and maintenance
of water supply system.
In this paper there is presented the process of water supply system optimiza‐
tion by using software package Epanet 2.0 on real example of DMA Pavlovac
which is part of water supply system under the authority of the UC “Vodovodi‐
kanalizacija” East Sarajevo.
Keywords: District metered area · Optimization · Water supply system
1 Introduction
Water utilities have a difficult task to provide fresh and clean water for normal func‐
tioning of cities with increasing urbanization. As The United Nations Commission on
Sustainable Development (CSD) stated, the amount of water per person decrease as
population grows, increasing possibility for water shortages [1]. Therefore, it is neces‐
sary to invest significant efforts to ensure enough drinking water with the aim of meeting
the needs of future generations and preserving natural resources.
The premise in work of water utilities so far was to provide a continuous supply of
water with maintaining the water quality at an acceptable level for all consumers. This
especially applies to the developing countries, where it is often considered that water is
a social category and that there is no market value. In order to satisfy water consumption,
water utilities must gradually change their priorities, rather than the permanent expan‐
sion and the opening of new sources, they have to turn towards inner reserves, to reduce
losses from the network and to reduce irrational use of water of their customers [2].

https://doi.org/10.1007/978-3-319-90893-9_52
444 A. Košarac et al.
New priorities for water utilities, instead of permanent seek for new sources, should
be rehabilitation of old systems, reducing of water losses, increasing economic efficiency
as well as raising awareness of consumers about the value of water, thereby reducing of
water consumption. With the new priorities system of water supply will be satisfactorily
kept under control by reducing losses in water systems or with optimization of water
supply systems.
Through this paper there will be presented the optimization process on real example
of water supply zone Pavlovac which is part of the water supply system under the
authority of the UC “Vodovodikanalizacija” East Sarajevo.
2 Influence of Pressure on Water Losses
In optimizing water system it is needed to particularly consider at what point the loss
reduction is economically justified, which includes monitoring of all data, from the input
data, costs of optimization, output as well selection of optimization methods. Control
of pressure in the network can directly influence the degree of water losses. If assumption
is that the water is leaking from pipes through the cross section A, then the outbound
flow is equal to [2]:
√ ( )
Q = CQ A 2gH m3 ∕s (1)
where:
A – Area of cross-section of orifice (crack) (m2),

CQ – Flow coefficient (depends on the shape of the orifice) (-),
g – Gravitational acceleration (m/s2),
H – Pressure at the orifice location (m).
√ given form it is obviously that with double pressure increasing, the loss increases
In the
for 2 = 1.41 times. This is true if the area of cross-section of the leak is constant.
However, with increasing pressure the area of crack through which water leaks is
increasing too, so resulting relationship between flow of leaking water and pressure is [3]:
Q = CQ × HN1 (2)
Where:
N1 – leak coefficient, depends on the type of materials, hydraulics and flow consump‐
tion.
3 Optimization Process of the Water Supply
Optimization of the water supply system is implemented in several related steps:

– collecting of input for mathematical model of the system,
Optimization of Water Supply System Using Software EPANET 2.0 445
– development of mathematical models based on the input,

– model calibration,
– simulation of different operating conditions on the mathematical model,
– analysis of simulation results and identifying possible activities in order to improve
the water supply system.
The input for mathematical model of the water supply system can be classified into four
main groups [4]:
– Geographic information consist of topographical maps, cadastral data, aerial photo‐
graphs and other supporting information that are useful for determining the physical
location of the system.
– Objects data should contain all information about pipes, pumps, valves, reservoirs
and other physical elements of water system.
– Operating data are important for the establishment of a boundary condition in math‐
ematical model of the system. They consist of the values of pressures, flow rate, water
level in the reservoirs, adjusted valves.
– Data on needs, or water consumption should be available from a database of
consumers.
The collection of input for model is not a one time job, it’s a permanent process. The
model must be constantly updated in order to attain the correct results of the simulation.
Model calibration is done so that the model be in more realistic condition. The basis
for the calibration are measurement of consumption during some periods of day so that
on the basis of these measurements there can be possible to determine the unevenly of
consumption and performs the calibration.
With simulation of various operating conditions the information is obtained about
the behavior of system which can determine further steps to improve the system.
After the simulation, faults of the system are recorded and solutions are being found
for their mitigation or removal. Also in this process mathematical model plays an essen‐
tial role, because they can model objects or elements in the water supply system, or
setting up the parameters can be made by estimated measures of optimization.
By repeated simulation of models with improved infrastructure and configuration
there can be determined level of enhancements designed to optimize measures and based
on the results the most advantageous solutions are adopted.
In this paper the EPANET software package is applied developed by the U.S. Envi‐
ronmental Protection Agency (EPA-Environmental Protection Agency). The program
does not require special licence (it is free of charge) so it’s affordable for all utility
companies in countries in transition. With application of mathematical model EPANET
it is possible to display real water flow and pressure in the water supply system. More‐
over, it provides the possibility of simulation of different steady states and dynamic
states of water systems. EPANET is easy to use and the results are obtained in the form
of tables and charts.
4 Calibration of the Model of the District Metered Area (DMA)

Pavlovac
Description of the existing situation and view of basic characteristics of the system is
extremely important if we want to define the conditions for good quality implementation
of the optimization process of water supply system.
Selected DMA Pavlovac (Fig. 1) is located in city of East Sarajevo. DMA covers
the settlements Pavlovac and the central part of the municipality East New Sarajevo.
Reservoir with volume of 180 m3 is located at 617 m above sea level. The reservoir
Pavlovac is supplied from the main reservoir of II height zone of East Sarajevo water
supply system.
Fig. 1. DMA Pavlovac
Water network consists of PEHD pipes with a diameter from PE63 to PE110, and
pipes of cast iron with a diameter from DN50 mm to DN150 mm. Total length of the
pipeline is around 11 km. Over 4400 inhabitants are supplied in this DMA with 1450
home water meters and also 165 business entities.
On the basis of collected data the model of hydraulic system is made. It is necessary
to point out that the first model is not calibrated so this model illustrates the theoretical
hydraulic condition of Pavlovac water supply system.
Consumption in the water supply system is not constant but it changes during the
day. The smallest consumption is during night hours, and maximum is in the afternoon.
This uneven consumption in a hydraulic model is displayed using the time pattern. The
basis for the time pattern is data of the company “Vodovodikanalizacija” East Sarajevo
obtained from measurements of flow at the input and output of reservoir Pavlovac i.e.
flow and water consumption monitored during 24 h on June 22nd 2017. The interval
between measurements is one minute.
It is necessary to point out that there is a smaller variation of the intake, and increased
water consumption during the day. Consumption is reduced during the night hours, and
it reaches maximum value in the afternoon, while the input in a reservoir is constant.
Time pattern multipliers are calculated as follows [5]:
Average hour demand

Time pattern multiplier = (3)
Average daily demand
With value of daily and hourly average demand for June 22nd 2017 and with use of time
pattern multipliers we are obtaining time pattern of a 24-h uneven consumption. Figure 2
shows the window for setting the time pattern in EPANET for model of DMA Pavlovac.
Fig. 2. The adjusted time pattern of DMA Pavlovac model
After the calibration of consumption the model of the actual state of the water supply
system is obtained, the next step is the implementation of simulations and analysis of
the existing situation.
5 Simulation and Analysis of Water Supply System
Key parameters for analysis of the existing situation and system optimization, obtained
as a result of the conducted simulation, are the value of pressures in nodes, speed of
flow through pipeline network. Also, an important parameter is the water consumption
in the system and the speed of charging and discharging the reservoir on the basis of
which it is being determined whether the reservoir capacity is optimal.
For better insight into DMA Pavlovac, apart from the calibrated model simulation
(of the existing actual state) there was also conducted model simulation with a fire
protection demand lasting for 2 h.
Pressure analysis – Lower critical points of pressure in water supply systems are
2.5 bars and higher critical point is pressure more than 5 bars. The analysis found out
that there are no nodes with pressure lower than 2.5 bars, while in the network there are
43 nodes with the pressure higher than 5 bars.
Velocity analysis – Recommended flow velocities in water pipes are within the
limits 0.2–2 m/s. Higher and lower velocity of these limits are not recommended, and
optimal velocity is around 0.8 m/s. After the model tests it’s found out that the flow
velocities in the distribution grid pipes are in range of 0.2–0.7 m/s. There are no sections
with flow velocity above 2 m/s in system but there is the total of 112 sections with flow
velocity less than 0.2 m/s. These data suggest that the pipeline network from hydraulic
point of view is oversized and the velocity of water flow can lead to decreasing of water
quality, however there should be kept in mind that these oversized pipes are to a certain
extent in correlation with fire protection regulations that require certain pipe diameter
for hydrant connections.
Analysis of water consumption and capacity of the reservoir – The average daily
consumption is around 10 l/s which corresponds to the data from the company, and inlet
is approximately constant with around 11.5 l/s. Based on the preceding considerations
it is concluded that current inlet flow corresponds to the average consumption, and the
reservoir can meet current needs, given that in the night hours when consumption is the
lowest there is full water charging and spillover from the reservoir.
In order to determine the response of the system to unforeseen demand we include
in the model in two nodes a consumption for the case of fire extinguishing in the amount
of 5 l/s. Consumption is given to the most unfavorable hydraulic nodes in the system in
the time period from 15:00 to 17:00 h when there is greatest consumption in DMA.
After repeated simulations it’s found out that there are no great changes in values of
pressures in pipe network, except in the nodes where there is assigned fire protection
consumption and another node where pressure in the time period is below a critical value.
Also there has been no big change in flow velocity in pipes, and there has been no
occurrence of critically high velocities during testing of fire duration. During the testing
of the fire duration, consumption has risen to around 20 l/s. Due to the increased
consumption the level in the reservoir is decreasing, and a longer period of time is
required to refill the reservoir. Accordingly, it is concluded that the reservoir at higher
demand cannot respond to the needs of consumers.
6 System Optimization
After the execution of simulations and determining the values of key parameters for the
system operation the activities for upgrade of those parameters commence i.e. system
optimization. This paper deals with optimization by decreasing the critical values of
pressure and optimization of reservoir capacity.
6.1 Pressure Optimization in System
The principle of optimization by decreasing the pressure is based on the installation of

pressure reduction valve (PRV) on a specific place in the water supply system. PRV
decreases the variable input pressure to a constant output pressure.
As previously stated the pressure on several sections exceeds the critical value of 5 bars.
For the place of the PRV installation setting at 3 bars one location in the model is
identified. After another simulation the number of nodes with pressure higher than crit‐
ical pressure was reduced from 43 to 6 and average pressure reduced by 2.5 bars.
Figure 3 shows the values of pressures after installation of the PRV.
Fig. 3. Pressure values on places with critical pressure after installation of PRV
6.2 Optimization of Reservoir Volume

As stated above that the present volume of the reservoir cannot adequately respond to
unplanned demand increase, that is why it is necessary to determine the optimal reservoir
volume. Calculation of the optimal reservoir volume is defined as [6]:
V = V1 + V2 + V3 (4)
where,
V – total reservoir volume (m3),

V1 – storage for several hours equalizing (m3),
V2 – storage needed for fire suppression (m3),
V3 – reserve volume (m3).
The volume for several hours equalizing V1 is obtained as the product of the coefficient
for several hours equalizing “a” and the maximum daily consumption Qmax,dn (m3/day).
In this case it’s taken eight hour equalizing so the value of coefficient a = 0.3, maximum
daily consumption Qmax,dn (m3/day) is obtained as the product of the average daily
consumption Qsr,dn (m3/day) and the highest coefficient of uneven daily consumption (Kd):
Average daily consumption for the system that is considered is Qsr,dn = 864 m3/day,
and the highest coefficient of uneven daily consumption is in the period between 19:00
and 20:00 h and it amounts to Kd = 1.44, so maximum daily consumption is
Qmax,dn = 1244 m3/day, therefore V1 amounts to 375 m3.
Storage for fire suppression is calculated as the amount of water of 5 l/s needed for
the fire duration of 2 h and it amounts to V2 = 40 m3, and reserve volume is V3 = 65 m3.
Based on the previously determined volumes the optimal reservoir volume is
V = 480 m3.
After modelling the reservoir with the calculated volume and after conducting new
simulation lasting for 72 h, with fire protection demand of 5 l/s for 2 h per day, the
response of the reservoir level has been obtained on 2 unfavorable hydraulic nodes in
the system as shown on Fig. 4.
Fig. 4. Water level in reservoir during 72 h simulation period

Based on the reservoir response it can be seen that longer time is needed to fully
charge the reservoir, and after its full charging the level fluctuations are relatively small
even during the fire demand time, so it is concluded that the capacity of the reservoir is
optimal and consumers’ needs can be satisfied as well as an unplanned demand which
may occur in DMA.
7 Conclusion
With application of the software package EPANET 2.0 for hydraulic modelling and
simulation of the system operation one obtains data which help to determine critical
points regarding water losses. The critical points in system are determined based on
critical pressure values which can lead to higher leakages or even to pipe break and
damage. Also the sections of pipes with critical velocity can be determined. In the DMA
Pavlovac the zones with high pressures are located and it is determined through the
analysis that reservoir capacity is not optimal. Moreover, with application of EPANET
possible activities for removing those shortages in the system are defined, and that is
installation of pressure reduction valve (PRV) and calculation of optimal reservoir
volume.
The main advantage of EPANET is relatively fast and easy model building and
simulation of the system operation, as well as fast simulation of potential problem solu‐
tions and analysis of their results without financial investments. Precondition for
achieving good results with EPANET optimizations are field measurements which are
input for system modelling and quality and accuracy of model depend on them. All of
the above stated lead to a conclusion that software EPANET 2.0 and similar software
packages for modelling and simulation of hydraulic systems will have significant role
in planning, managing and optimizing of water supply systems.
References
1. Vučijak, B., Ćerić, A., Silajdžić, I., MidžićKurtagić, S.: Water for life, basic of integrated water
management (Vodazaživot, osnoveintegralnogupravljanjavodnimresursima), Sarajevo (2011)
2. Prodanović, D.: Rational use of water as an alternative to new sources
(Racionalnokorišćenjevodekaoalternativanovimizvorištima), predavanje, Univerzitet u
Beogradu (2002). http://hikom.grf.bg.ac.rs
3. Lambert, A.O., Hirner, W.: Losses from Water Supply Systems: Standard Terminology and
Recommended Performance Measures. International Water Association (2000)
4. AWWA: Computer Modeling of Water Distribution Systems Manual of Water Supply
Practices
5. Rossman, L.A.: EPANET 2-Users Manual, EPA/600/R-00/057, Water Supply and Water
Resources Division, National Risk Management Research Laboratory, Office of research and
development, U.S. Environmental Protection Agency, Cincinnati
6. Milojević, M.: Water supply and sewer (Snabdevanjevodomikanalisanjenaselja),
Naučnaknjiga Beograd (1995)
The Possibilities of the Cadastral Land Use
Assessment by the Methods of Remote Sensing
Admir Mulahusić ✉ , Jusuf Topoljak, Nedim Tuno, and Karlo Ajvazović

( )

{admir_mulahusic,jusuf.topoljak,nedim_tuno}@gf.unsa.ba,
ajvazovic.karlo@gmail.com
Abstract. This study shows integration of remote sensing methods and conven‐
tional geodetic methods as well as processing and analysis of the results of cadas‐
tral land use assessment for the area of cadastral municipality Butmir in Sarajevo.
For that purpose high-resolution multi-spectral WorldView-2 satellite images
were used, with the object-oriented approach to image classification and image
classification algorithm creation. The main objective is to compare the results
obtained with the conventional methods of cadastral land evaluation and the
results obtained by the methods of remote sensing. The possibilities of method‐
ology development for the assessment of cadastral land use for each individual
cadastral parcel have been assessed. Such a methodology would provide updated
information on the real estate cadastre, reduce the costs and time of data collection
and land use map creation procedure, with the aim to solve problems in economic
development of the country, urbanization of the area, development of science and
exploration of the land potentials.
Keywords: Land cadastre · Cadastral land evaluation · Remote sensing

WorldView2 · Image classification in remote sensing · GIS
1 Introduction
Since the time humankind learned to settle on land, cadastres have developed and
evolved to suit society’s needs. Changes in the relationship of humankind to land have
invoked matching evolutionary changes in the function of cadastres [1]. Required and
primarily important data are always about the part of the land’s surface (parcel of land
or land lot), about the person who owns the land, a land utilization types, the location
of a land parcel, etc. A dual system of real estate records is in effect in Bosnia and
Herzegovina and it fully meets needs of the society [2]. Today’s era of information
technologies has largely updated real estate records, cadastre and land registry (in legal
frameworks). A large number of real estate data in Bosnia and Herzegovina is available
through the Internet [5]. Good management of land resources implies a quality and rapid
management of land information [7]. By establishing a new survey in Bosnia and
Herzegovina (i.e. detailed survey in the period between 1966 and 1984), the cadastral
land evaluation (assessment of the land use of each single parcel) and land classification
for valuation purposes were made. The constant changes, which were the result of social

https://doi.org/10.1007/978-3-319-90893-9_53
The Possibilities of the Cadastral Land Use Assessment 453
development, have caused disagreement between actual situation on the ground and the
situation in the cadastral documents. Fast and efficient cadastral work is subject to
updated and accurate data, which depend on speed, accuracy of data collection and
updating the changes. Since the creation of the cadastre to this day, methods of data
collection have evolved and improved. Depending on needs, the most efficient and
fastest methods are remote sensing methods [4]. Remote sensing is the science of
obtaining information about the Earth’s surface without having immediate contact.
Thereby, we observe and record reflected or radiated electromagnetic waves, process
them, analyze them, and use them in different applications [6].
2 Test Area
The paper analyzes the area of the cadastral municipality (CM) Butmir (Fig. 1), located
approximately 8 km southwest of downtown Sarajevo in the Ilidža municipality. Cadas‐
tral municipality Butmir has an average altitude of 521 m and occupies an area of
approximately 7 km2.
Fig. 1. Part of the observed area of CM Butmir (digital orthophoto from 2013)
Sarajevo International Airport, also known as Butmir Airport, the main airport in
Bosnia and Herzegovina, is located in the CM Butmir.
3 Used Data
The following data were used during the analysis:

(a) Database of cadastral land use,
(b) Digital orthophotos of the analyzed area and
(c) High resolution satellite images of the analyzed area.
454 A. Mulahusić et al.
ESRI ArcGIS is software used to create the database of cadastral land use. Database
was used as a basis for the analysis of cadastral land use, that were obtained by classical
cadastral land evaluation methods, for the purpose of defining cadastral land use by the
method of remote sensing. It also served as a basis for analyzing and comparing data
obtained by classical methods of cadastral evaluation and remote sensing methods.
Database of cadastral land use contains basic data about parcels, in terms of their shape,
surface and the way of use according to land cadastre register.
For the purposes of compiling the cadastral land use assessment database, the
following were used:
• Digital cadastral map of the analyzed area,
• Katastar.ba Geoportal of the Federal Administration for Geodetic and Real Property
Affairs (Fig. 2) and
• Digital orthophotos of the analyzed area from 2013.
Digital Cadastral Map represents an ‘‘updated’’ state of cadastral parcels in terms
of their surface and shape. Using the GIS applications, the digital cadastral map is asso‐
ciated with the attribute data of the cadastral register. The digital cadastral map (obtained
from the Federal Administration for Geodetic and Real Property Affairs) is in vector
format, enabling easier manipulation, processing and analysis of spatial data. The Real
Estate Cadastre Database was created with the ArcGIS software by allocating cadastral
land use from the cadastral register CM Butmir to the values of polygons within the
digital cadastral map.
The database data refers to the month of May 2017. Figure 2 shows a part of the land
cadastre database with the corresponding legend.
Fig. 2. Part of database of cadastral land use within CM Butmir
Digital orthophoto maps represent basic source of spatial data. They could be used
for the spatial planning and realization of many different activities. Digital orthophoto
maps used in this study are in official use by the Federal Administration for Geodetic
and Real Property Affairs. These were created by the aerial photogrammetric survey in
the period between 2011 and 2013. Images spatial resolution is 0.25 m in RGB (red,
green, blue) composites, and they represent quality basis for analysis and accuracy
assessment of the classification. For CM Butmir 8 images at a scale of 1:2500 were used.
Three multispectral WorldView-2 satellite images were used in this study. Satellite
images were from a different period of time (all from 2013). The year of collection
corresponds to the reference data, or to orthophoto images that will be used in data
analysis. Satellite image from May 2013 was chosen for better differentiation of the
arable land in the classification process. All images were provided by the commercial
company Digital Globe Foundation. Digital Globe Foundation is a leading company in
the field of satellite imagery and geospatial content. Digital Globe Foundation program
encourages research projects and the development of geospatial applications.
4 Data Processing and Analysis
Nowadays, a large number of GIS software for manipulating, analyzing and interpreting
spatial data have been created [3]. Most software applications support the processing of
remote sensing data, whether they are commercial or freely available programs. In the
practical part of the study, ArcGIS 10.3 software package was used in combination with
the eCognition program for the object-oriented classification of satellite images. A
diagram of the classification is given in Fig. 3. Prior to the classification of the satellite
images, a detailed analysis of the classification area and cadastral land use assessment
was done in order to extract the information classes. An interactive classification and
analysis of results yielded to the largest possible number of classes that could be iden‐
tified for the analyzed area.
Fig. 3. Classification diagram of the WorldView-2 satellite images

Classification accuracy assessment provides the user with an insight into classifica‐
tion errors, on the basis of which the user could evaluate whether the results of the
classification are useful, or the classification procedure needs to be repeated. Sometimes
it is necessary to change the methodology. Assessment of the accuracy for the individual
land use classes is done by overlapping and comparison of test points, for which the land
use class is known (i.e. the way of land use), with the corresponding value of the clas‐
sified image area. The known values of the test points could be obtained from multiple
sources. Usually they are obtained by field measurements or, for example, observing
aerial photogrammetric images of a larger scale (or accurate thematic maps). When using
other satellite or aerial images as reference data, it is necessary to be particularly careful
because there might be a significant disagreement between the actual quality of derived
classification and the worse accuracy indicators obtained by making the error matrix,
due to the fault interpretation of land cover [8].
Table 1. Comparison of the assessments of the classification accuracy and the real estate cadastre
database.
Class Accuracy assessment of the object Accuracy assessment of the real
oriented classification estate cadastre database
User’s accuracy Producer’s User’s accuracy Producer’s
accuracy accuracy
Building 80.65% 80.65% 27.42% 58.62%
Road 91.67% 91.67% 45.83% 97.06%
Surface water 96.30% 96.30% 64.29% 69.23%
Stream 77.78% 82.35% 97.22% 87.50%
Pool 90.00% 100.00% 100.00% 100.00%
Forest 64.71% 61.11% 4.17% 50.00%
Field 62.50% 88.24% 96.23% 32.48%
Meadow 90.57% 67.61% 34.69% 34.69%
Orchard 81.63% 72.73% 24.24% 66.67%
Access road 84.85% 100.00% 6.67% 100.00%
Other land use 78.95% 93.75% 31.58% 8.70%
Overall accuracy 85.34% 46.17%
Kappa coefficient 0.82 0.40
The orthophoto image of the analyzed area was used to determine the actual values
of land use and for the purpose of comparison in this study. The accuracy assessment
was performed with 431 well-distributed check points. To assess the accuracy of land
use in the cadastral register CM Butmir, the same procedure as in the classification was
used. In this way, it is possible to directly compare the obtained results with the results
obtained by the object-oriented classification (Table 1). Identical check points were used
to assess the accuracy of the classification and database of cadastral land use. This
procedure eliminates errors of accuracy assessment and provides an objective compar‐
ison of the obtained results. The cadastral land use data were adapted to the information
classes of the land classification prior to the accuracy assessment. All land use data are
thus classified into an object-oriented classification classes (Fig. 4).
Fig. 4. The land use classification results of the WorldView-2 multispectral image
5 Conclusion
The results obtained in this study reveal that the overall accuracy of land use data in the
cadastral register is around 50%. This indicates a large disagreement between the situa‐
tion in the field and the situation in the cadastral records. The interpretation of satellite
imagery brings many significant improvements compared to classical methods of data
collection and processing. A very good accuracy of spatial data is often obtained through
satellite imagery interpretation, which enables the process of updating the land use
information to be performed in a timely and cost-effective manner. Remotely gathered
data still don’t have the level of detail comparable to the data directly collected on the
ground. However, due to obsolete data in cadastral records, image processing and anal‐
ysis techniques can have a great importance for land registers.
Land use classification accuracy was over 85% for high-resolution satellite imagery.
It clearly indicates the necessity for remote sensing methods application in a land use
determination. The classification usually produces land-use and land-cover maps, which
are suitable for urban design, physical and spatial planning, landscape design, efficient
forest and agricultural asset management, real estate mass appraisal, expropriation and
land consolidation, topographic map updating etc. The accuracy of the classification
depends on the quality of the input satellite image. High-resolution multispectral satellite
imagery is a good choice for object-oriented classification algorithms. Correct
processing of these images requires that the temporal, spectral, spatial and radiometric
resolution must be considered, as well as the technical specifications of a particular
image, such as cloud amounts, sun elevation angle, etc. A qualitative and well-defined
geospatial data, derived from satellite imagery, provide a strong foundation for various
operations and data integration in GIS environment.
References
1. Ting, L., Williamson, I.P.: Cadastral trends: a synthesis. Aust. Surv. 4(1), 46–54 (1999)
2. Begić, M.: Geodetska služba Bosne i Hercegovine 1880–2012. Geodetski glasnik 42, 53–105
(2012)
3. Steiniger, S., Bocher, E.: An overview on current free and open source desktop GIS
developments. Int. J. Geogr. Inf. Sci. 23(10), 1345–1370 (2009)
4. Ali, Z., Tuladhar, A., Zevenbergen, J.A.: An integrated approach for updating cadastral maps
in Pakistan using satellite remote sensing data. Int. J. Appl. Earth Obs. Geoinf. 18, 386–398
(2012)
5. Mulahusić, A., Topoljak, J., Tuno, N.: Geodezija za građevinske inžinjere. Univerzitet u Zenici
(2017)
6. Oštir, K., Mulahusić, A.: Daljinska istraživanja. Građevinski fakultet Univerziteta u Sarajevu
(2014)
7. Roić, M.: Upravljanje zemljišnim informacijama – katastar. Geodetski fakultet Sveučilišta u
Zagrebu (2012)
8. Congalton, R.G., Green, K.: Assessing the Accuracy of Remotely Sensed Data: Principles and
Practices. CRC Press, Boca Raton (2009)
Analysis of the Energy Potential of Organic
Bioradable Part of Municipal Waste
Mahmut Jukić ✉ and Ifet Šišić

( )
University of Bihac, 77000 Bihać, Bosnia and Herzegovina

mahmut.jukic@gmail.com, sisic_btf@yahoo.com
Abstract. Sustainable waste management, and especially its biodegradable part

at the European Union level, is becoming one of the priorities, not only because
of the necessity of conserving soil and reducing waste, but also the possibility of
using significant energy potentials of this type of waste. This is particularly true,
bearing in mind that about 90 million tons of biowaste is produced annually in
the EU, 40% of which is still disposed of in landfills.
Considering the composition of biodegradable municipal waste (food waste,
garden waste, agricultural waste, etc.), as well as a high percentage of moisture
in it, the process of anaerobic digestion is the most appropriate way of converting
biodegradable municipal waste into renewable energy. This process is carried out
using biogas plants that use biodegradable municipal waste as a raw material for
producing biogas as an energy product for energy production.
Therefore, the use of biodegradable municipal waste for the purpose of biogas
production has two important beneficial aspects: environmental protection
through the avoidance of waste disposal on non-sanitary landfills, and the produc‐
tion of energy from renewable sources.
Keywords: Biodegradable waste · Biogas · Anaerobic digestion · Digestate

Compost · Cogeneration
1 Introduction
Sustainable waste management and especially its biodegradable part is becoming one
of the priorities at the European Union level, not only because of the necessity of
conserving soil and reducing waste, but also the possibility of using significant energy
potentials of this type of waste. This is especially true, bearing in mind that about 90
million tons of bio-waste are produced annually in the EU, 40% of which is still disposed
of in landfills (2006/12/EC; EC 2010).
According to the available data of the Federal Ministry of Environment and Tourism
from 2011, the total amount of collected waste in the Federation of B&H annually
amounted to 735 051 t/year, while in Una-Canton Canton the amount of collected waste
according to these data amounted to 100 840 t/year. Given the unreliability of the data
in this paper, the data derived from the research carried out by Mahmut Jukic in 2014
will be used in the preparation of a master’s thesis on the “Contribution to Determining
the Energy Potential of Municipal Waste”. The research was carried out in the Una-Sana

https://doi.org/10.1007/978-3-319-90893-9_54
460 M. Jukić and I. Šišić
Canton area, specifically in the “Mezdre” landfill, which is being disposed of by the
waste of the municipalities of Cazin, Bosanska Krupa and Buzim, using the official
method: SWA-Tool (Development of a Methodological Tool to Improve Precision and
Comparability of Solid Waste Analysis Data), which aims to develop an increase in the
accuracy and comparability of municipal waste data at the European level. According
to the data from the mentioned master thesis, the total generated amount of municipal
waste in the Una-Sana Canton area amounted to 34,000 t/year. Considering the high
proportion of biodegradable waste that can be treated with anaerobic digestion of 58%,
the estimated quantity is 19 720 t/year.
In addition to biodegradable household waste, which is 17.5%, organic waste is to
be treated as garden waste in the amount of 22.7% and fine elements (residues of wood
waste, sawdust, branches, other waste from the yard and the carriageway, etc.) in the
amount of 17.8%. Thus, the share of the organic biodegradable fraction of municipal
waste to be treated by anaerobic digestion is 58%.
2 Problems of Research
Sustainable management of biodegradable organic waste includes basic hierarchical

settings as well as in the management of all types of waste: avoiding the occurrence of
waste - prevention, that is, reduction at the source), to separate different forms of bio-
waste, material processing, thermal treatment and eventually disposal.
Regardless of the fact that the reduction in waste generation - prevention in the first
place, it is not possible to completely avoid what is necessary to implement measures
that will reduce the disposal of waste to landfills (Kirac 2009).
Anaerobic digestion is an appropriate way of converting biodegradable municipal
waste into renewable energy since this process is exclusively bacterial, and anaerobic
bacteria act best in aqueous medium or very humid conditions. It is this fact that makes
the process of anaerobic digestion very important. When talking about energy potential,
from every ton of such waste under biological treatment from 100 to 200 m3 of biogas
can be obtained (EC 2008).
The waste treatment methods differ in the EU member countries. In developed coun‐
tries, the 2014 data suggests that less communal waste was deposited in landfills, while
methods of composting and recycling waste predominated in relation to other methods
of treatment. Thus, recycling and composting in 2014 almost two one third (64%) of
waste treatment in Germany, followed by Slovenia (61%), Austria (58%), while the
largest share of municipal waste deposited in landfills was recorded in Latvia (92%),
Malta (88%), Croatia (83%) and Romania (82%) (EUROSTAT 2016).
In addition to disposal to landfills, thermal processing and composting, the most
optimal way of treating organic biodegradable waste is the treatment of anaerobic diges‐
tion through a biogas plant (Kricka et al. 2009).
All countries of the European Union have set clear targets before them, which can
be achieved by applying procedures for the rational utilisation of the organic part of
municipal waste. One of these methods is certainly the production and use of biogas
from organic waste, the method of anaerobic digestion (Amon et al. 2006). Biogas is a
Analysis of the Energy Potential of Organic Bioradable Part 461
mixture of several gases, where methane and carbon dioxide account for 90% of the total
mixture. The complete composition of biogas is shown in Table 1. (Kricka et al. 2009).
Table 1. Chemical composition of biogas

Flammable part of biogas Non-flammable part of biogas
Gas Volumepart % Gas Volumepart
Methane CH4 55–75 Carbon dioxide CO2 25–45
Hydrogen H2 0–1 Nitrogen N2 0–2
Sulfurichydrogen H2S 0–1 Oxygen O2 0–0,5
Aeratedwater H2O 0–2
Ammonia NH 0–2
3
In addition to the many advantages inherent in anaerobic digestion in the treatment

of the organic part of municipal waste, there are certain disadvantages that are reflected
in the inability to provide constant quality or quantity as well as the space constraints
for plant accommodation. Despite the difficulties of the European country, they increas‐
ingly use anaerobic digestion (Ostrem 2004). The largest number of installed waste
treatment plants with anaerobic digestion are in Germany, Switzerland, the Netherlands,
that is, the countries that first implemented the anaerobic digestion method as treatment
of municipal waste, while France and Spain have started to use the anaerobic digestion
method much later (De Baere, Mattheeuws).
2.1 Objective and Methods of Research
Considering the significant share of organic waste in municipal waste, the aim of this
paper is to analyse the energy potential of this waste. The aim of this analysis is to
determine the energy potential of the organic biodegradable fraction of municipal waste
in the field of research (Una-Sana Canton), or certain energy factors that will serve to
dimension the possible biogas plant, as well as to analyse the financial effects of a
possible cogeneration plant for production electrical or thermal energy by using biogas.
Based on the amount of biodegradable organic waste in the field of research, this
analysis will determine the thermal value of organic biodegradable waste (Hd), the total
amount of energy contained in the waste (E), the amount of biogas produced in the biogas
plant (Qpl), the chemical energy of the produced biogas (Eh), the power of biogas at the
entrance to the cogeneration plant (Ppl), the biogas mass (mpl) and the mass of wet and
dry digestate (md, msd). This paper will use mathematical, statistical and methods of
technical-economic analysis.
2.2 Results of Research and Discussion
The total energy of the biodegradable organic part of municipal waste is calculated
according to formula 1 (Table 2):
E = M ⋅ Hd (MJ) (1)
Table 2. Data for the analysis of the energy potential of organic municipal waste (Jukić 2015).
Area USK-Region Bihac
Population 306.849 people
Surface 4 125 km2
Population density 74,3 citizens/km2
Produced amount of waste 108 840 t/yeara
Amount of collected waste 34 000 t/year
Organic waste 19 720 t/year
Amount of waste per capita 157 kg/capita/year
a
Federal Ministry of Environment and Tourism, 2011
With the following:

M – mass of generated biodegradable waste
Hd – the lower thermal value of the organic waste, for the calculation of the lower
thermal value of this waste structure, reference values are used for the following:
Hd = 6 MJ/kg for garden waste
Hd = 4,2 MJ/kg for food residues,
Hd = 12 MJ/kg for other fine waste with the smallest structure (Kiely 1998).
The average lower thermal value of the biodegradable organic part of the municipal
waste is Hd = 7.2 MJ/kg, so that the total energy contained in this waste is:
E = 19.720 × 1000 × 7,2 = 141,984 × 106 (MJ)
E = 0,142 × 106 (GJ)
If 1 ton of fuel oil has approximately 40 GJ of energy, the energy equivalent of the
organic biodegradable fraction of municipal waste in the Una-Sana Canton amounts to
3550 tons of that oil.
The amount of biogas that can be obtained in a biogas plant is calculated based on
formula 2:
[ ] [ 3]
t m
Qbp = Q ⋅ qs . (2)
year t
The specific amount of biogas obtained from the biodegradable organic part of the
municipal waste is qs = 150 m3⁄t (Rohlik 2016).
Qpl – the amount of biogas per year
Q – annual production of biodegradable organic part of municipal waste
[ ] [ 3] [ 3 ]
t m m
Qbp = 19.720 ⋅ 150 = 2.958.000 .
year t year
The chemical energy contained in the biogas is calculated on the basis of the biogas
heat value and the total amount of biogas per formula 3. For this calculation the heat
value of the biogas from 6.5 kWh/m3 (Rohlik 2016) is taken.
[ ] [ ]
m3 kWh
Ebp = Qbp ⋅ Hdbp (3)
year m3
[ ] [ ] [ ]
m3 kWh kWh
Ebp = 2.958.000 ⋅ 6, 5 = 19.227.000
year m3 year
If the AD plant works 8000 h per year with full capacity, the average biogas power
at the entrance to the cogeneration unit is calculated according to the formula 4:
[ ] [ ]
kWh kWh
Ebp 19.227.000
year year
Pbp = [ ] = [ ] = 2403.37[kW] (4)
h h
𝜏 8000
year year
In addition to biogas as the basic product that is obtained from an organic biode‐
gradable part of municipal waste by anaerobic digestion, in order to do further researches
of the plant for AD sustainability, it is also necessary to calculate the mass of the digestate
as a beneficial product. In order to calculate the mass of the digestate it is necessary to
calculate the biogas mass according to the Eq. 5:
[ ] [ ] [ 3 ]
kg kg m
mbp = 𝜌bp 3 ⋅ Qbp (5)
year m year
[ ] [ 3 ] [ ]
kg m t
mbp = 0, 717 ⋅ 2.958.000 = 2.120, 8
m3 year year
]
[
kg
Biogas density 𝜌bp = 0, 717 (Nakomcic-Smaragdakis et al. 2013).
m3
Mass of the digestate is obtained if the mass of the biogas is taken from the total
mass of organic waste (Formula 6):
[ ] [ ]
t t
md = Q − mbp (6)
year year
[ ] [ ] [ ]
t t t
md = 19.720 − 2.120, 8 = 17.599, 2
year year year
The amount of this digestate md is digestate in which the water content is about 40%
so it is necessary to calculate the mass of dry digestate:
[ ] [ ]
t t
msd = 60% ⋅ md = 10.535, 2 where 50% (5267,7 t/year) there is fine
year year
compost that will be sold, and 50% is a rough compost which is a cost because it is
necessary to dispose it in a proper manner.
Based on the above factors, it is also possible to calculate the final products that can
be obtained from biogas by means of cogeneration plants, i.e. electrical and thermal
energy. First, it is necessary to calculate the electric power of a biogas plant using
formula 7:
Pe = Pbp[kW] ⋅ 𝜂 = 2403, 37 ⋅ 0, 3 = 721.011 kWel ≈ 0, 72 MWel (7)
𝜂 – the electrical efficiency of the biogas plant is 30% for gas engines (Husika 2016)
Annual electricity production is:
[ ] [ ]
h MWh
Eel = 8000 ⋅ 0, 72 [MWel] = 5.760
year year
Pth
The ratio between the thermal and electric power of the big-scale plant is = 1, 7
(Husika 2016), Pe
Bearing in mind the above ratio, the thermal power of a biogas plant is:
Pth = 1, 7 ⋅ 0, 72 [MW] = 1, 22 [MW]
The total amount of heat is obtained as follows (Table 3):

[ ] [ ]
h MWh
Qth = 8000 ⋅ 1, 22 [MW] = 9.760
year year
Table 3. Products obtained by analysing the energy potential of biodegradable organic waste
Product Annual production
Electricity (MWh/year) 5760
Heat (MWh/year) 9760
Dry digestate (t/year) 10535,2
It is important to note here that 20% of the heat remains for the operation of the
biogas plant, and the other part, i.e. power of 1 MW can be used to warm the surrounding
houses, or to heat the sanitary water. If we consider that the average boiler utilisation
rate in B&H is 42% (Husika 2016) and that the number of days in the heating season is
180, the heat can be calculated according to formula 8:
Qisk. = P ⋅ 𝜏 ⋅ k (8)
where it is:
P[MW] – the heat power of buildings available for heating buildings

𝜏[h] – duration of the heating season
k – coefficient of heat load capacity of the heating plant
Qisk. = 0, 976 MW ⋅ 180 ⋅ 24 h ⋅ 0, 42 = 1.770, 85 MWh
The average consumption of thermal energy in households is about 7900 kWh per
year (Statistics Agency of B&H 2015), which means that 224 households can be supplied
by the heat from a biogas plant.
3 Conclusion
The aim of this paper was to present the most important benefits (products) from treating
this type of waste and using it for energy purposes due to a significant share and assuming
that organic biodegradable waste was initially separated from the rest of the municipal
waste. The use of municipal waste for energy purposes is inevitable in modern societies.
Technologies for energy valorisation of communal waste are continuously developed
and improved with the aim of achieving the most favorable effects on the community
and the environment.
In this paper, for the area of research based on the annual amount of biodegradable
municipal waste (19 720 t/year), the energy contained in (0.142 × 106 GJ) was calcu‐
lated, which equals to 3550 tons of fuel oil.
Based on the amount of biogas available in the biogas plant with anaerobic digestion
(2.958.000 m3/year), the energy contained in the biogas (19 227 kWh/year) was calcu‐
lated. Based on the calculated power of 0.72 MW and 1.22 MW of heat, the produced
electricity is calculated in the amount of (5760 MWh/year), the total amount of thermal
energy (9760 MWh/year). In addition to these two beneficial products based on the
calculation of waste mass and biogas, the amount of dry digestate for sale in the amount
of (10 535,2 t/year) was calculated. Based on the calculated utilised amount of heat (1
770.85 MWh), it was concluded that 224 households could be supplied with it.
References
De Baere, L., Mattheeuws, B.: Report of the Intergovernmental Panel on Climate Change
Anaerobic Digestion of the Organic Fraction of Municipal Solid Waste in Europe – Status,
Experience and Prospects (n.d)
Hublin, A., Kralik, D., Čurlin, M.: Energija iz biorazgradivog organskog otpada, Internet,
dostupno: http://gospodarenjeotpadom.yolasite.com/resources/ANDREA%20HUBLIN.pdf,
pristupljeno, 04 January 2018 (n.d)
Husika, A.: Tehnologije za proizvodnju energije iz biomase, seminar Proengineer, Razvojni
program ujedinjenih nacija, Amabasada Kraljevine Švedske, Fond za zaštitu okoliša
Federacije Bosne i Hercegovine i Fond za zaštitu životne sredine i energetsku efikasnost
Republike Srpske, Sarajevo i Banja Luka (2016)
Jukić, M.: Doprinos određivanju energijskog potencijala komunalnog otpada, Magistarski rad,
Mašinski fakultet, Univerzitet u Sarajevu (2015)
Kirac, M.: Najbolje raspoložive tehnike, Diplomski rad, Fakultet strojarstva i brodogradnje,
Sveučilište u Zagrebu (2009)
Mihajlović, V.: Model upravljanja otpadom zasnovan na principima smanjenja negativnog uticaja
na životnu sredinu i ekonomske održivosti, Doktorska Disertacija, Fakultet tehničkih nauka,
Univerzitet u Novom Sadu (2015)
Rohlik, P.: Koncept iskorištavanja biootpada za proizvodnju biometana, Diplomski rad, Fakultet
Strojarstva i brodogradnje, Sveučilište u Zagrebu (2014)
Voća, N., Krička, T., Ćosić, T., Rupić, V., Jukić, Ž., Kalambura, S.: Digested residue as a fertilizer
after the mesophilic process of anaerobic digestion. Plant Soil Environ. 51(6), 262–266 (2005)
New Technologies in Agriculture and
Ecology, Chemical Processes
Modeling the Extraction Process of Sage (Salvia
Officinalis L.) with Supercritical CO2 at Different
Temperatures
Sabina Begić1 ✉ , Vladan Mićić2, and Darko Manjenčić3

( )
1
sabina.begic@untz.ba
2
University of East Sarajevo, 75400 Zvornik, Bosnia and Herzegovina
micicvladan@yahoo.com
3
University of Novi Sad, 21000 Novi Sad, Serbia
manjend@gmail.com
Abstract. In this paper an investigation was done on the extraction of sage

(Salvia officinalis L.) using supercritical carbon dioxide at different extraction
temperatures (313, 323 and 333 K). For the qualitative and quantitative analysis
of the obtained extracts methods GC-FID and GC-MS were applied. Qualitative
and quantitative composition of the obtained CO2 extracts and essential oils
isolated from CO2 extracts at different temperatures of extraction was also deter‐
mined. The extraction system was modeled by using the modified equation
Reverchon - Sesti Osseo, whereby the effect of temperature on the yield of the
extract was studied. Based on the results of modeling the overall yield of the
extract depending on the temperature, it is concluded that a given model approx‐
imates the experimental results very well in all cases.
Keywords: Supercritical extraction · Temperature · Sage · Modeling
1 Introduction
Genus Salvia (Lamiaceae) comprises about 900 species, spread throughout the world,
some of which are economically important since they are used as spices and flavouring
agents in cosmetics, perfumery and pharmaceutical industries. Salvia officinalis L.
extracts have been reported in a use for a wide range of applications due to its significant
antioxidant [1–6], antibacterial [4, 7–9], hypoglycemic, anti-inflammatory, fungistatic
[7, 8], virustatic, astringent, eupeptic and anti-hydrotic and antidiabetic properties [10].
The processes for isolation of sage essential oil (EO) and extracts such as hydrodis‐
tillation, the conventional organic solvent extraction and subcritical or supercritical
extraction using CO2 (SFE or SCCO2 extraction) have advantages and disadvantages
with respect to operating cost, capital cost, yield, and quality of obtained extracts [11].
Having in mind the disadvantages of the process of distillation by steam and extrac‐
tion with organic solvents, the extraction by pressurized gasses, which is developing
rapidly in recent years, presents a good alternative because it has a number of advantages

https://doi.org/10.1007/978-3-319-90893-9_55
470 S. Begić et al.
[12]. Depending on the state of solvent, it can be distinguished extraction by gasses in

liquid and in supercritical state. Low viscosity, high value of the coefficient of diffusion,
negligible surface tension, despite the density which is similar to the density of liquids,
make that the mass transfer during extraction with supercritical fluids is better in
comparison with organic solvents [12–15].
The flexibility of manipulating with selectivity and soluble power of fluids and better
transport characteristics that allow faster penetration into the porous matrix and diffusion
of substance from it, are some of the main advantages of using liquid and supercritical
fluids in extraction methods [12, 14, 15]. Another important property of these gasses as
solvents is easy separation of the extracted substances from the solvent by reducing its
power of dissolution, i.e. reducing the density below the level needed for the extract to
dissolve.
The main advantage of SFE over traditional extraction techniques is a possibility of
continuous modulation of the solvent power/selectivity of the dense CO2 as well as
elimination of polluting organic solvents and expensive post-processing of the extracts
for elimination of solvent’s trace [16]. Therefore, the influence of the origin of the plant,
growing conditions, harvest time and the part of the analyzed plant on the yield and
chemical composition of extract must be considered as well [17–19].
During the past years there are a several papers regarding SFE extraction of sage
from different origin [2, 3, 5, 6, 9, 11, 20–25]. The yield of essential oil significantly
varying and highly depended on extraction conditions and plant origin. Most of these
investigations were concentrated on isolation of light fractions or sage’s essential oil but
most valuable compounds present in sage extract are diterpenes which poses antioxi‐
dative characteristic [25]. Data published in the literature related to SFE of sage did not
covered the wide range of pressure and temperature (or density) [2, 6, 9, 11, 20–25] and/
or did not analyzed chemical composition of obtained fractions and total extract [2, 3,
5, 11, 22, 23]. Further more, all of above-mentioned study did not give an explicit
conclusion related to the best operating conditions of SFE for different compound
group’s extraction. The sage’s essential oil could be extracted according to the literature
data at mild SFE conditions (40–60°C) [3, 5, 6, 9, 20–24].
From a large number of different gasses under pressure which are applied as solvents
for the extraction of natural substances (fragrant, spicy and pharmaceutically active
substances), carbon dioxide has a distinct advantage [12].
There are some reports about the choice of nitrous oxide (N2O) as an extraction fluid.
The chemical properties of this fluid make N2O more suited for the extraction of polar
compounds. However, in the presence of a high organic content, the gas can cause violent
explosions. This drawback strongly limits its use [23, 26–28].
Even water was investigated as a possible supercritical fluid but, unlike CO2, the
high critical temperature and pressure, together with the corrosive nature of H2O under
these conditions, has limited its practical applications [28]. Even so, water was used in
some cases as a co-solvent for the extraction of more polar compounds from aromatic
plants. The presence of water as pretreatment of plant material or added to CO2 at
supercritical and subcritical state as a co-solvent has shown to influence the qualitative
and quantitative composition of the extract [29–32].
Modeling the Extraction Process of Sage (Salvia Officinalis L.) 471
Ethane, propane and dimethyl ether have been used as supercritical solvents for the
extraction of bioactive compounds from plants. Beside critical points that are compa‐
rable with CO2, these solvents have higher polarizability than CO2, resulting in a stronger
interaction with the more polar compounds and co-solvents [32–35]. However, experi‐
mental results show that SC CO2 offers a wider versatility for the fractionation of
extracted compounds using different operative pressures in the extractors or separators
[1, 7, 11, 18–20].
With supercritical extraction with carbon dioxide, extraction is carried out in a
supercritical area, and separation under conditions of pressure and temperature that are
appropriate for the gaseous state of carbon dioxide [6].
For this experiment Sage (Salvia officinalis L.) native to the area Berkovic, Eastern
Herzegovina, was used. For extraction, laboratory device High Pressure Extraction Plant
(HPEP, Nova Swiss, Switzerland) was used. The extraction process is carried out by the
plant material being milled and ground to a certain degree of fragmentation [36–39].
Preparation of drug for extraction was done by fragmentation using the comercial mill
(Multi Moulinex, 260 W), taking care that the drug is not overheated during milling.
For determining the granulometric composition of drug, a set of sievs from the
manufacturer Erweka Apparatebau GmbH (Germany) was used. As extragent, commer‐
cial carbon dioxide of 99% purity was used, produced by Tehnogas (Novi Sad).
A sample of sage (Salvia officinalis L.), with mass of 60 g and particle size d = 0.22 mm
is extracted in the extractor of internal volume V = 200 ml, at a pressure p = 100 bar,
flow rate of extractant w = 3.23 ∙ 10−3 kg/min and temperatures 313, 323 i 333 K (8, 9).
Test results of kinetic extraction, i.e. total yield of extract depending on the temperature
are shown by diagrams in Fig. 1. The highest yield of extract was achieved at a temper‐
ature of extraction 313 K (4.45 g/100 g of drug). This is consistent with the theory that
with increasing temperature, the volumetric mass of the solvent decreases, and thus the
solubility of extractives. Higher temperatures than 313 K are not recommended to be
applied for the extraction of aromatic plants since they contain components which are
transformed at high temperatures (thermodegradable components). At the end of the
extraction process, after the expiry of time τ = 10 h, CO2 – extracts were separated and
essential oil was allocated from them. The quantitative and qualitative analysis of
obtained extracts and essential oils was conducted with GC/MS method. Results of these
researches are given in Table 1. The results show that extracts obtained at temperatures
313 K and 323 K have the same qualitative composition, while the extract obtained at
a temperature of 333 K does not contain β – thujone and terpineol L – 4. The dominant
component in obtained extracts is filokladen whose content range from 23.57 to 34.12%.
Regarding the content of α – thujone and camphor, it is the lowest in the extract obtained
at a temperature of extraction 333 K and amounts to 0.32 i.e. 0.88%, and significantly
higher contents are at a temperature of extraction 313 K and 323 K. Essential oils
extracted from CO2 extracts contain the highest percentage of α – thujone whose content
ranges from 17.44 to 28.79%. The separated oils also contain a high percentage of
camphor in the amount of 19.31 to 25.62%.
5
4,5
4 313K
323K
y (g/100 g drug)
3,5 333K
3
2,5
2
1,5
1
0,5
0
0 100 200 300 400 500 600 700
time (min)
Fig. 1. Diagrams of yield of extract in function of time
Table 1. Qualitative and quantitative composition of CO2 – extracts and essential oils extracted
from them
Component Component content (%, w/w)
CO2 extract/essential oil
Temperature (K)
313 323 333
1,8 – cineole -/1.24 -/3,47 -/-
α – thujone 3.42/17.44 5.11/28.79 0.32/22.60
β – thujone 0.29/3.11 0.40/5.21 -/4.39
camphor 9.89/19.31 12.03/25.62 0.88/21.63
isoborneol 8.14/8.20 9.95/8.67 8.62/8.83
terpineol L – 4 0.20/0.53 0.23/0.61 -/0.53
bornyl – acetate 0.19/3.56 3.29/3.76 1.44/3.94
sabinyl – acetate 2.62/0.41 0.18/- 0.09/0.47
isocaryophyllene 0.54/1.04 0.62/0.95 0.27/1.07
α – gurjunene 0.80/0.72 0.84/0.60 1.01/0.38
γ – elemene 7.08/8.28 7.97/7.70 5.48/8.84
selina–3,7(11) diene 9.92/7.92 11.04/5.61 19.32/7.66
1,11-epoxyhumulene 3.57/2.17 4.07/1.49 5.43/2.34
caryophyllene oxide 2.92/2.46 3.07/1.74 4.88/2.35
phyllocladene 34.12/16.76 23.57/3.16 31.19/7.23
Total 83.09/93.15 82.38/97.38 82.60/92.26
Using a modified Reverchon – Sesti Osseo equation the given extraction system was
modeled, whereby the effect of temperature on the yield of extract was researched, [40].
The final form of Reverchon – Sesti Osseo equation is:
[ ]
t
Y = 100 1 − exp(− ) (1)
ti
where:
t – time of extraction (s),

ti – time of internal diffusion (s),
Y – normalised yield of extract (%).
Time of internal diffusion is defined as:
1 Ap K
= (2)
ti (1 − 𝜀)V
where:
Ap – area of mass transfer between the particles of plant material and fluid (m2),
K – total coefficient of mass transfer (m/s),
ε – porosity of plant material,
V – volume of the extractor (m3).
If it is assumed that the internal diffusion time (ti) is a constant value for a defined
system and extraction conditions, from the Eq. (1) follows:
Y ′ t
Z = ln (1 − =) = a t = − (3)
100 ti
where:
′
a – constant, which represents the negative reciprocal value of the internal diffusion
time.
Normalised extract yield Y (%) is defined by the expression:

y
Y= ⋅ 100 (4)
ymax
where:
y – yield of the extract after the extraction time t (g/100 g of drugs),

ymax – yield of the extract at the end of the extraction process (g/100 g of drugs).
In order to simplify the Eq. (3), the same is extended by adding a constant b, to get
the expression:
Z = at + b (5)
By introducing the Eq. (5) in (1) a modified Reverchon – Sesti Osseo equation is
obtained (6):
[ ]
Y = 100 1 − exp(at + b) (6)
where a and b are constants determined from experimental data.

The modified Reverchon - Sesti Osseo equation (Eq. 6) is used in this paper for
modeling of the extraction system. Results are shown in Tables 2, 3 and 4 and in
Figs. 2, 3, 4 and 5.
Table 2. experimental and calculated values of the extract yield (Y) and values Z depending on
temperature
Time T (K)
(min) 313 K 323 K 333 K
Y Yrač Z Y Yrač Z Y Yrač Z
30 38.43 39.36 −0.485 47.84 46.56 −0.651 33.54 28.29 −0.409
90 59.10 53.71 −0.894 68.77 61.12 −1.164 52.92 46.23 −0.753
150 67.62 64.66 −1.128 72.76 71.71 −1.300 60.92 59.69 −0.940
210 73.70 73.02 −1.336 73.75 79.41 −1.337 68.31 69.77 −1.149
270 75.73 79.41 −1.416 82.06 85.02 −1.718 71.69 77.34 −1.262
390 86.18 88.00 −1.979 92.36 92.07 −2.572 83.69 87.26 −1.813
510 94.16 93.00 −2.840 96.34 95.80 −3.308 94.77 92.84 −2.951
600 100 95.34 – 100 97.39 – 100 95.35 –
ymax = 4.45 ymax = 3.01 ymax = 3.25
Sy,x = 3.143 Sy,x = 3.695 Sy,x = 4.272
Table 3. The values of the parameters a and b in Eq. (6) for given temperatures
T (K) a b |r|
313 −0.0045 −0.3652 0.987
323 −0.0053 −0.4676 0.984
333 −0.0048 −0.1885 0.993
Table 4. The modified Eq. (6) for calculating the normalised yield of total extract at different
temperatures of extraction
T (K) Modified equation
313 Y = 100[1−exp(−0.0045t – 0.3652)]
323 Y = 100[1−exp(−0.0053t – 0.4676)]
333 Y = 100[1−exp(−0.0048t – 0.1885)]
0
100 200 300 400 500 600 700
-0,5
-1
-1,5
-2
313K
Z -2,5 323K
333K
-3
-3,5
-4
-4,5
-5
time (min.)
Fig. 2. Dependance of Z on extraction time
120
100
80
Y (%)
Yexp
60
Yrač
40
20
0
0 100 200 300 400 500 600 700
time (min.)
Fig. 3. Results of fitting the normalised yield of total extract at a temperature of 313 K
120
100
80
Y (%)
Yexp
60
Yrač
40
20
0
0 100 200 300 400 500 600 700
time
120
100
80
Y (%)
60 Yexp
Yrač
40
20
0
0 100 200 300 400 500 600 700
time (min)
Based on the results of modeling the extraction system: sage leaf – supercritical
CO2, using the modified equation of Reverchon Sesti Osseo (6) where the yield of
extraction was examined in dependance of mean particle diameter, flow, pressure and
temperature of carbon dioxide, it can be concluded that the applied model approximates
experimental results very well in all cases where the correlation coefficient ranges from
0.939 to 0.996, and the standard error of regression from 1.597 to 10.011.
4 Conclusion
As in the cases of examining the influence of pressure and in the cases of testing the
effect of temperature on the yield of CO2 – extract (isothermal process) two periods of
extraction are noticed. At the beginning fast, and later after 4.5 h is slow extraction. The
highest yield of extract was achieved from the extraction of sage at the extraction
temperature of 313 K, while at temperatures of 323 K and 333 K similar results of
extraction were achieved, but with much lower yield. With increasing the temperature
the volumetric mass of the solvent decreases, and thus the solubility of the extractives.
On the other hand, higher temperatures than 313 K are not recommended to be applied
for the extraction of aromatic plants since they contain components which are trans‐
formed at high temperatures. The highest yield of extract was achieved at a temperature
of 313 K (4.45 g/100 g of drug) with the lowest essential oil content in the extract
(33.58%). Based on the results of modeling the extraction system: sage leaf – supercrit‐
ical CO2, using modified equation of Reverchon Sesti Osseo during wherein the yield
of extract was examined depending on the median particle diameter, flow, pressure and
temperature of carbon dioxide, it was concluded that the applied model approximates
experimental results very well with a correlation coefficient, which ranges from 0.939
to 0.996, and standard error of regression from 1.597 to 10.011.
References
1. Miura, K., Kikuzaki, H., Nakatani, N.: Antioxidant activity of chemical components from
sage (Salvia officinalis L.) and thyme (Thymus vulgaris L.) measured by the oil stability index
method. J. Agric. Food Chem. 50, 1845–1851 (2002)
2. Dapkevicius, A., Venskutonis, R., van Beek, T.A., Linssen, J.P.H.: Antioxidant activity of
extracts obtained by different isolation procedures from some aromatic herbs grown in
Lithuania. J. Sci. Food Agric. 77, 140–146 (1998)
3. Ðarmati, Z., Jankov, R.M., Schwirtlic, E., Ðulinac, B., Ðorđević, A.: High antioxidant activity
of extracts obtained from sage by supercritical CO2 extraction. J. Am. Oil Chem. Soc. 68,
731–734 (1991)
4. Bozin, B., Mimica-Dukic, N., Samojlik, I., Jovin, A.: Antimicrobial and antioxidant
properties of rosemary and sage (Rosmarinus officinalis L. and Salvia officinalis L.
Lamiaceae) essential oils. J. Agric. Food Chem. 55, 7879–7885 (2007)
5. Ivanovic, J., Djilas, S., Jadranin, M., Vajs, V., Aničić, N., Petrović, S., Žižović, I.:
Supercritical carbon dioxide extraction of antioxidants from rosemary (Rosmarinus officinalis
L.) and Sage (Salvia officinalis L.). J. Serb. Chem. Soc. 74, 717–732 (2009)
6. Hayouni, E.A., Chraief, I., Abedrabba, M., Bouix, M., Leveau, J.Y., Mohammed, H., Hamdi,
M.: Tunisian Salvia officinalis L. and Schinus molle L. essential oils: their chemical
compositions and their preservative effects against Salmonella inoculated in minced beef
meat. Int. J. Food Microbiol. 125, 242–251 (2008)
7. Weckesser, S., Engel, K., Simon-Haarhaus, B., Wittmer, A., Pelz, K., Schemppa, C.M.:
Screening of plant extracts for antimicrobial activity against bacteria and yeasts with
dermatological relevance. Phytomedicine 14, 508–516 (2007)
8. Menaker, A., Kravets, M., Koel, M., Orav, A.: Identification and characterization of
supercritical fluid extracts from herbs. C. R. Chim. 7, 629–633 (2004)
9. Eidi, A., Eidi, M.: Antidiabetic effects of sage (Salvia officinalis L.) leaves in normal and
streptozotocin-induced diabetic rats, Diabetes Metab. Syndr., Clin. Res. Rev. 3, 40–44 (2009)
10. Catchpole, O.J., Grey, J.B.: Near-critical extraction of sage, celery and coriander seed. J.
Supercrit. Fluids 9, 273–279 (1996)
11. Hortacsu, O.: Modeling of natural materials extraction. In: Kiran, E., et al. (eds.) Supercritical
Fluids, Funtamentals and Aplication, vol. 366, pp. 499–516. Springer, Dordrecht (2000)
12. Perrut, M.: Supercritical fluid application: industrial developments and economic issues. Ind.
Eng. Chem. Res. 39, 4531–4535 (2000)
13. Lang, Q., Wai, C.M.: Supercritical fluid extraction in herbal and natural product studies – a
practical review. Talanta 53, 771–778 (2001)
14. Aleksovski, S.A., Sovova, H.: Supercritical CO2 extraction of Salvia officinalis L. J. Supercrit.
Fluids 40(2), 239–245 (2007)
15. Reverchon, E., De Marco, I.: Supercritical fluid extraction and fractionation of natural matter.
J. Supercrit. Fluids 38, 146–166 (2006)
16. Perry, N.B., Anderson, R.E., Brennan, N.J., Douglas, M.H., Heaney, A.J., Mc Gimpsey, J.A.,
Smallfield, B.M.: Essential oils from Dalmatian sage (Salvia officinalis L.): variations among
individuals, plant parts, seasons and sites. J. Agric. Food Chem. 47, 2048–2054 (1999)
17. Santos-Gomes, P.C., Fernandes, M.: Ferreira, Organ and season dependent variation in the
essential oil composition of Salvia officinalis L. cultivated at two different sites. J. Agric.
Food Chem. 49, 2908–2916 (2001)
18. Bettaieb, I., Zakhama, N., Aidi Wannes, W., Kchouk, M.E., Marzouk, B.: Water deficit effects
on Salvia officinalis fatty acids and essential oils composition. Sci. Hortic. 120, 271–275
(2009)
19. Reverchon, E., Taddeo, R.: Extraction of sage oil by supercritical CO2: influence of some
process parameters. J. Supercrit. Fluids 8, 302–309 (1995)
20. Kutta, G., Pluhár, Z., Sárosi, S.: Yield and composition of supercritical fluid extracts of
different Lamiaceae herbs. Int. J. Hortic. Sci. 13, 79–82 (2007)
21. Nguyen, U., Frankman, G., Evans, D.A.: Process for extracting antioxidants from Labiatae
herbs, US patents US 5,017,397 (1991)
22. Langa, E., Della Porta, G., Palavra, A.M.F., Mainar, A.M.: Supercritical fluid extraction of
Spanish sage essential oil: optimization of the process parameters and modelling. J. Supercrit.
Fluids 49, 174–181 (2009)
23. Raynie, D.E.: Warning concerning the use of nitrous oxide in supercritical fluid extractions.
Anal. Chem. 65, 3127–3128 (1993)
24. Pourmortazavi, S.M., Hajimirsadeghi, S.S.: Supercritical fluid extraction in plant essential
and volatile oil analysis. J. Chromatogr. A 24(2), 1163 (2007)
25. Pereira, C.G., Meireles, M.: Supercritical fluid extraction of bioactive compounds:
Fundamentals, applications and economic perspectives. Food Bioproc. Technol. 3, 340–372
(2010)
26. Ong, E.S., Cheong, J.S.H., Goh, D.: Pressurized hot water extraction of bioactive or marker
compounds in botanicals and medicinal plant materials. J. Chromatogr. A, 92–102 (2006)
27. Leal, P.F., Maia, N.B., Carmello, Q.A., Catharino, R.R., Eberlin, M.N., Meireles, M.: Sweet
basil (Ocimum basilicum) extracts obtained by supercritical fluid extraction (SFE): Global
yields, chemical composition, antioxidant activity, and estimation of the cost of
manufacturing. Food Bioprocess Technol. 1, 326–338 (2008)
28. Gamiz Gracia, L., de Castro, M.D.L.: Continuous subcritical water extraction of medicinal
plant essential oil: comparison with conventional techniques. Talanta 51, 1179–1185 (2000)
29. Bicchi, C., Rubiolo, P., Frattini, C., Sandra, P., David, F.: Offline supercritical fluid extraction
and capillary gas chromatography of pyrrolizidine alkaloids in senecio species. J. Nat. Prod.
54, 941–945 (1991)
30. Mohamed, R.S., Saldana, M.D.A., Mazzafera, P., Zetzl, C., Brunner, G.: Extraction of
caffeine, the obromine, and cocoa butter from Brazilian cocoa beans using supercritical CO2
and ethane. Ind. Eng. Chem. Res. 41, 6751–6758 (2002)
31. Catchpole, O.J., Grey, J.B., Perry, N.B., Burgess, E.J., Redmond, W.A., Porter, N.G.:
Extraction of chill, black pepper, and ginger with near critical CO2, propane, and dimethyl
ether: analysis of the extracts by quantitative nuclear magnetic resonance. J. Agric. Food
Chem. 51, 4853–4860 (2003)
32. Illes, V., Daood, H.G., Perneczki, S., Szokonya, L., Then, M.: Extraction of coriander seed
oil by CO2 and propane at super and subcritical conditions. J. Supercrit. Fluids 17, 177–186
(2000)
33. Reverchon, E.: Supercritical fluid extraction and fractionation of essential oils and related
products. J. Supercrit. Fluids 10, 1–37 (1997)
34. Hamdan, S., Daood, H.G., Toth-Markus, M., Illes, V.: Extraction of cardamom oil by
supercritical carbon dioxide and subcritical propane. J. Supercrit. Fluids 44, 25–30 (2008)
35. Simandi, B., Deak, A., Ronyai, E., Gao, Y.X., Veress, T., Lemberkovics, E., Then, M.:
Supercritical carbon dioxide extraction and fractionation of fennel oil. J. Agric. Food Chem.
47, 1635–1640 (1999)
36. Lawless, J.: The Encyclopedia of Essential Oils. Harper-Collins Publisher Ltd., London
(1992)
37. Muller, J., Muhlbauer, W., Koll-Weber, M., Kraus, W.: Effects of drying on essential oil of
Chamomilla recutita and Salvia officinalis. Medicinal Plant Report 3, 65–67 (1996)
38. Piccaglia, R., Marotti, M., Dellacecca, V.: Effect of planting density and harvest date on yield
and chemical composition of sage oil. J. Essent. Oil Res. 9, 187–191 (1997)
39. Putievsky, E., Ravid, U., Sanderovich, D.: Morphological observations and essential oils of
sage (Salvia officinalis L.) under cultivation. J. Essent. Oil Res. 291–293 (1992)
40. Zeković, Z., Lepojević, Ž., Tolić, A.: Modeling of the thyme supercritical carbon dioxide
extraction system. The influence of extraction time and carbon dioxide pressure. Sep. Sci.
Technol. 38, 541–552 (2003)
The Effect of Concentration of Methanol as a Solvent
on the Antioxidative Activity of Sage Extract
Selma Osmić1 ✉ , Sabina Begić1,2, and Vladan Mićić1,2

( )
1
{selma.tuzlak,sabina.begic}@untz.ba
2
University of East Sarajevo, 75400 Zvornik, Bosnia and Herzegovina
micicvladan@yahoo.com
Abstract. The production of high quality plant extracts is the basis for the
production of herbal preparations and the isolation of bioactive ingredients.
Recently, there is a tendency of replacing synthetic antioxidants, due to their toxic
and potentially cancerous effects, with natural antioxidants. For this reason, it is
necessary to intensify research of antioxidative and biological effects of plant
extracts, as well as the influence of the process conditions of extraction on the
mentioned effects and the yield of extraction. In this paper, the antioxidative
activity of sage extract was tested, depending on the concentration of methanol
as a solvent in the process of maceration with occasional mixing. Total phenolic
and flavonoid content in obtained sage extracts was determined spectrophoto‐
metrically. The radical scavenging capacity was determined by the DPPH
method, where the extract concentration required for the neutralization of 50% of
the initial concentration of the DPPH radical was also determined.
Keywords: Methanol · Antioxidativity · Sage · Extract
1 Introduction
Numerous studies in chemistry, biochemistry and medicine fields confirm that raw
materials of plant origin, as well as their extracts, contain natural antioxidants. Antiox‐
idants are substances that, present in small amounts, postpone or slow down oxidative
changes of the substrate. Literature sources provide insight into a broad antioxidant
classification. The division by their origin, which differentiates between natural and
synthetic antioxidants, is widely accepted. Natural originate from living systems, while
synthetic antioxidants are chemically synthesized. According to the place of emergence,
antioxidants that are important to human organism are divided into endogenous, that are
formed within the organism and antioxidants of exogenous origin, i.e. those that are
introduced into the organism by diet, dietary supplements or medicines. Based on the
chemical nature, enzymatic and non-enzymatic antioxidants differ, and according to
solubility water-soluble antioxidants and fat-soluble antioxidants [1].
Based on the way they show their activity, the following types of antioxidants are
mentioned: inhibitors of free radicals generated by oxidative reactions (preventive
oxidants), antioxidants interrupting the autoxidation chain reaction propagation, singlet

https://doi.org/10.1007/978-3-319-90893-9_56
The Effect of Concentration of Methanol as a Solvent 481
oxygen eliminators, reducing agents that convert hydroperoxides into stable compounds,
metal chelators which convert metal prooxidants (iron and copper derivatives) into
stable products and inhibitors of proxidative enzymes. Antioxidants, according to the
level and mode of action in the human body, are divided into: preventive antioxidants,
antioxidants “scavengers” and reparative antioxidants. Preventive antioxidants prevent
the formation of free radicals. “Scavengers” have the ability to react with free radicals,
while reparative antioxidants act by special mechanisms, restoring or removing of vital
biomolecules damaged by oxidation. Another proposed classification of antioxidants,
according to a different mechanism of action, is primary or true antioxidants, that delay
initiation or interrupt the chain reaction of propagation, and secondary antioxidants that
reduce the rate of oxidation [2–4].
Many medical scientific studies have confirmed the importance of plant antioxidants,
as a complement to human endogenous antioxidants. The interest of the food industry
for plant antioxidants has increased especially after a series of published scientific
studies on the harmful effects of synthetic antioxidants, added to food, on human health.
The excessive presence of free radicals in living organisms and the oxidative processes
induced by them, that can lead to damage of biomolecules, are inhibited by substances
labeled with the term antioxidants. The need for plants and their extracts in the form of
pharmaceutical and dietary products is increasing year by year [5]. The use of plant
antioxidants in the food, cosmetic and pharmaceutical industries is a promising alter‐
native to synthetic antioxidants in terms of price, compatibility with dietary intake and
no harmful effects on the health of the body [6]. For the aforementioned reasons, in the
last few decades, a large number of research is aimed at identifying new antioxidants
from natural raw materials. Most of natural antioxidants are phenolic compounds, e.g.
polyphenols, of which the most important are phenolic acids and flavonoids [7]. Phenolic
compounds are secondary plant metabolites that are present in all plant tissues. Unlike
primary plant metabolites, necessary for life and cell development, secondary metabo‐
lites are compounds that do not directly participate in photosynthesis and breathing
metabolism but are essential for survival of the plant in the environment and are the
most common and most active plant antioxidants. Their presence is a consequence of
evolution, that is, the changes that were necessary for the survival of the plant in the
environment. Their quantity and composition in plants depend on various internal and
external factors, in the sense that their share increases due to stress conditions [8]. They
make one of the most numerous groups of compounds in nature, and to date more than
8,000 [9] have been discovered.
The structure of phenolics differs, depending on the number of aromatic nuclei
present in the molecule, so there are different monophenols that contain one benzene
ring to which one or more -OH groups are attached, and polyphenols containing a higher
number of benzene rings within one molecule [10].
Phenolic compounds, depending on their structure, act as antioxidants in many ways,
one of which is due to the presence of hydroxyl groups that are good proton donors and
which can react with reactive oxygen and thus prevent the formation of new radicals
[11]. The interaction of hydroxyl groups with the benzene ring electron gives the mole‐
cules special properties of free radicals that are capable of altering the oxidation process
[12]. Because of their redox potential, phenolics can act as free radical “scavengers”.
482 S. Osmić et al.
Antioxidativity of phenolic acids increases with increasing the number of hydroxyl

groups, so for example dihydroxybenzoic acid derivatives have higher antioxidative
activity than hydroxybenzoic acid derivatives. According to Pokorný [7], the presence
of more than three hydroxyl groups on the aromatic nucleus does not influence the
increase in the antioxidant effectiveness of molecules.
Phenolic acids and their activity have been subject to numerous biological, chemical
and medical studies for years. In food, they appear as glycosides or esters with other
compounds such as sterols, alcohols, glucosides and fatty acids [13]. They have strong
antioxidant, anti-inflammatory, anticancer and antimicrobial activity, and have also
shown a positive effect in the treatment of many cardiovascular diseases and the preven‐
tion of atherosclerosis [14]. Dai and Mumper [8] have shown that phenolic acids exhibit
a stronger antioxidant effect on free radicals than vitamins C and E and carotenoids.
Flavonoids, as extremely important natural antioxidants, are one of the most commonly
present phenolics that are widespread in plant tissues. About half of the phenolic
compounds that occur in plants are flavonoids [15]. In addition to being one of the most
important plant pigments, they also contribute to taste, mostly bitterness and pungency
of plants in which they are found. This class of compounds undoubtedly also participates
in the defense system, as the pungency and bitterness they give to plants are unpleasant
to herbivorous, including man, and also protect the plant from the harmful effects of UV
radiation. Plant extracts rich in flavonoids have been used for centuries as an effective
antibacterial agents against many bacterial strains such as Escherichia coli, Staphylo‐
coccus nervous, Staphylococcus epidermis, Saccharomyces cerevisiae [9]. They exhibit
antioxidant properties by inhibiting the enzymes responsible for the production of
superoxyl radicals, by chelating metals involved in the reaction of the formation of free
radicals and by preventing peroxidation [16]. Phenolics can be extracted from fresh,
frozen, or dried plant samples. Usually, prior to extraction, the herbal sample is treated
by grinding or crushing and homogenization and may be previously air-dried or freeze-
dried. The classical extraction of dried plant material with solvents is the most commonly
used extraction method because of simplicity of application and efficiency [17]. In addi‐
tion to the above mentioned method, a number of new methods, such as supercritical
extraction and ultrasonic assisted or microwaves assisted extraction, have been devel‐
oped to reduce the time of extraction and solvent consumption and increase the yield
and quality of the extract.
The extraction yield depends on the type of solvent, i.e. its polarity, time and
temperature of extraction, the ratio of plant material and solvent, as well as the chem‐
ical composition and physical properties of the material being subjected to extrac‐
tion. The choice of the appropriate solvent affects the yield and the extraction rate
[18], therefore the solvent should satisfy high polarity, low viscosity, low boiling
point, chemical inertness, thermal stability, non-degradability, inflammability and
non-toxicity, environmental acceptability, availability, low cost and the ability to
reuse [19]. For the extraction of phenolic compounds from plant material, aqueous
solutions of methanol [20], ethanol [21] and acetone [22] are most commonly used.
Shan et al. [23] found high effectiveness of the methanol in the extraction of phenolic
compounds from spices. It was also found that methanol was more efficient in extrac‐
tion of lower molecular weight polyphenols, whereas aqueous acetone is good for
extraction of higher molecular weight flavanols [24]. In preparing extracts rich in

flavonoid anthocyanidin, most commonly used are acidified organic solvents (meth‐
anol or ethanol), which enable denaturation of the cell membrane while simultane‐
ously dissolving and stabilizing anthocyanidins. Caution is required when adding
acid, which in excess can hydrolyze unstable acyl groups and sugar residues. For
obtaining the best yield of extraction of anthocyanin, the use of weak organic acids
such as formic, acetic, citric, tartaric and phosphoric acid or low concentrations of
strong acids, such as 0.5–3.0% trifluoroacetyl and <1, 0% hydrochloric acid, is recom‐
mended [25, 26]. Altiok et al. [21] confirmed the importance of the presence of water
in an organic solvent, which promotes the diffusion process, thus facilitating the
extraction of phenolic compounds from plant tissue. Using different polarity solvents,
it has been found that flavonoid compounds are easier to extract from plant tissue
using a more polar solvent using a more polar solvent. Process conditions also have
a significant effect on the extraction process. By temperature increase, the solubility
and the diffusion coefficient of the extractable matter increase, and thus the mass
transfer rate in the solution. By increasing the time of extraction, the amount of the
transferred matter increases, whose diffusion rate gradually decreases due to the
decrease of concentration gradient and increase of the viscosity of the solution.
However, when it comes to extraction of plant antioxidants, i.e. phenolic compounds,
the increase in temperature and extraction time can cause hydrolysis and oxidation
of phenolic compounds, leading to a decrease in the yield of phenolic extract [27].
Excessive extraction time and high temperature increase the possibility of oxidation
of phenolic compounds. Apart from the choice of solvent, temperature and extraction
time, extraction efficiency is also affected by other parameters, such as solvent and
sample ratio and particle size [28]. By reducing the particle size of the plant material
that is extracted, the contact surface between the two phases increases, increasing the
amount of the transferred substance in the unit of time.
Depending on the solvent used, during the extraction process from the plant material,
along with the phenolic compounds, other compounds, such as sugars, organic acids,
fats, proteins, pigments etc. which need to be removed from the extract are also
extracted [28].
In this paper an experimental study was carried out to investigate the influence of
the concentration of aqueous methanol solvents on the antioxidant properties of meth‐
anol extracts of the sage leaf.
2 Experimental Part
Plant material - dried leaves of the sage plant (Salvia officinalis L.), which has been
grown at the plantations of MP Ljekobilje Trebinje in 2015, was provided for the exper‐
imental part of the work. Herbal material was grinded in an electric coffee mill. Methanol
(95–96%), chlorogenic acid (99%), distilled water, catechin hydrate, Folin Ciocalteu
reagent, AlCl3, 0.135 mM solution of 2,2-diphenyl-1-picrylhydrazyl (DPPH), aqueous
solution of Na2CO3, sodium hydroxide and sodium nitrite (5%) were used in the work.
The required aqueous solutions of methanol, 40%, 50% and 60%, were prepared by
dilution with distilled water.
Determination of degree of fragmentation of the plant material was performed using
vibratory sieves. Average diameter of particles of plant material is obtained by applying
the expression:
( )
100 ∑ mi
=
d di
where:
mi – mass percentage of the fraction I (%),

di – average diameter of the fraction i (mm),
d – degree of fragmentation (mm).
Extraction of plant material by maceration method was performed using 40%, 50%
and 60% aqueous solutions of methanol. At each procedure, plant material of the defined
degree of fragmentation was introduced into the Erlenmeyer of 250 ml and then poured
with a selected solvent, whereby the ratio of plant material to solvent was 1:10. The
procedure was carried out at different time intervals (30, 60, 90, 120, 150 and 180 min)
at room temperature, with occasional stirring. In the obtained extracts, the mass of total
phenolic (mgCAE/g) and their yields (%) were determined depending on the time of
extraction. In the extracts obtained by extraction with various concentrations of aqueous
methanol solutions during the extraction time of 60 min, total phenolics, flavonoids and
free radical scavenging activity were determined.
Determination of total phenolics content was performed by Folin-Ciocalteu’s spec‐
trophotometric method and using chlorogenic acid as standard [29–33]. Determination
of total flavonoids was conducted by method according to Markham and expressed
as milligrams of catechin equivalents per gram of dry sage (mg CE/g) [29–33]. Deter‐
mination of free radical scavenging capacity (RSC) was performed by the DPPH method
[29–33]. Based on the known RSC value, the IC50 value was determined, representing
the concentration of extract required to inhibit 50% of the initial DPPH radical concen‐
tration.
3 Research Results and Discussion
For the purpose of carrying out the research, the dried plant material was grinded and
then, by using vibratory sieves, its separation on the fractions was performed according
to the size of the particles. Based on the calculation of the obtained mass fractions, the
degree of fragmentation, which was 0.206 mm, was calculated.
For each individual concentration of the aqueous methanol solution (40%, 50% and
60%), six samples of the sage of a given degree of fragmentation were prepared, which
were subjected to maceration at different time duration of the procedure (30, 60, 90, 120,
150 and 180 min). Extraction was carried out at room temperature, which was 25.6 °C.
The obtained yields of the total phenolic extract of the sage leaf, depending on the time
of extraction, for different concentrations of aqueous methanol solutions as solvent are

shown in Fig. 1.
8
The yield of phenolic extract
7
6
5
4 40% methanol
(%)
3 50% methanol
2
60% methanol
1
0
0 50 100 150 200
Extraction time (min)
Fig. 1. The yield of the phenolic extract depending on the extraction time
The highest yield of phenolic extract was obtained using a 50% aqueous methanol
solution after a 90 min extraction time. Changes in the concentration of aqueous meth‐
anol solutions did not greatly affect the change of yield of phenolic extract of sage.
Taking into account practical and economic aspects, as well as the possibility of plant
phenolics degradation, or their undergoing undesirable reactions by extended extraction
times [8, 34–36], in further research the content of total phenolics and flavonoids was
determined in methanol sage extract obtained at room temperature of 25.6 °C during the
period of extraction time of 60 min (Fig. 2). The highest content of total phenolics was
obtained by extraction with 60% aqueous methanol (67.340 mg CAE/g), and the smallest
with 50% aqueous methanol (60.664 mg CAE/g). Extraction with 40% methanol solvent
gave the phenolic content of 63.196 mg CAE/g. Bouterfas et al. [37] also obtained the
highest value of total phenolics extracted from Marrubium vulgare L. dried leaves when
60% aqueous methanol was used, compared to the other studied concentrations (20%,
40% and 80%; v/v). Do et al. [38] obtained the maximum total phenolics from Limno‐
phila aromatica with 75% aqueous methanol, while 100% methanol and 50% aqueous
methanol gave lower total phenolics values.
The highest flavonoid content was obtained by extraction with 50% aqueous meth‐
anol (37.944 mg CE/g) and the lowest with 40% solvent (24.092 mg CE/g). Extraction
with a 60% aqueous methanol solution yielded a flavonoid content of 32.304 mg CE/g.
The obtained maximum values of total phenols and flavonoids with solvents of different
polarity can be explained by the chemical composition of the plant sample. Plant mate‐
rials may contain phenolics varying from simple (e.g., phenolic acids, anthocyanins) to
highly polymerized substances (e.g., tannins) in different quantities [8] and differences
in the structure of phenolic compounds also determine their solubility in solvents of
different polarity [39].
80
Flavonoid content (mg CE/g)

Phenolic content (mg CAE/g)
70
60
50
40 phenolic content
30 flavonoid content
20
10
0
40% methanol 50% methanol 60% methanol
Fig. 2. Content of total phenolics and flavonoids in the extract
In this paper, the research of antioxidant properties of extracts, obtained by extraction

with methanol of different concentrations (40%, 50% and 60%) during a 60 min extrac‐
tion time at room temperature, was also carried out. The IC50 value was determined for
the obtained methanolic extracts and the highest IC50 value (0.101 mg/ml), i.e.
the lowest antioxidant activity had the extract obtained by extraction with 40% aqueous
methanol solution, while the extracts obtained with 50% and 60% aqueous methanol
solutions had the same value for IC50 (0.096 mg/ml). Antioxidant potential of the plant
material usually positively correlate with the phenolic content [30, 39–42]. However,
in the present work, the correlation between the total phenolic content and antioxidant
activity was not observed. Mohammedi and Atik [43] reported that aqueous methanol
crude extracts of the leaves from Tamarix aphylla had higher total phenolic content than
aqueous ethanol, water and aqueous acetone extracts, but they did not showed the highest
antioxidant activity. Various antioxidant compounds with different chemical character‐
istics and polarities may or may not be soluble in a particular solvent [44]. The presence
of electron-donating and electron-withdrawing substituents in the ring structure of
phenolics as well as the number and arrangement of the hydroxyl groups determines
their antioxidant potential [45], which is why it is possible that phenolic compounds
existing in 40% aqueous methanol extract do not possess an ideal structure for the scav‐
enging of free radicals, compared to 50% and 60% aqueous methanol extracts.
4 Conclusion
By performing the procedure of extraction of sage leaf, at extraction temperature of

25.6 °C (room temperature of extraction), the phenolic extract yield increased over the
first 90 min, while by the further prolongation of time, it decreased. A different behavior
was demonstrated by 40% methanol, whose yield grew in the first 60 min of extraction
and then decreased. This behavior can be explained by the fact that during longer time
of extraction, hydrolysis and oxidation of the phenolic compounds occur, which leads
to a reduction in the phenolic extract yield.
The highest phenolic extract yield was achieved using 50% methanol (7.2456%),
while at the same extraction conditions the lowest yield (5.9872%) was obtained with
40% methanol as a solvent. The yield of phenolic extracts obtained by extraction with
50% methanol as a solvent compared to 40% methanol was higher by 21.02% and
compared to 60% methanol by 12.02%.
As for the total phenolic content, it was the highest in the extract obtained with 60%
aqueous methanol (67.340 mg CAE/g), and the lowest in the extract obtained with 50%
aqueous methanol (60.664 mg CAE/g). The yield of the phenolics in the extract obtained
by extraction with 60% aqueous methanol was higher by 11.05% compared to the extract
obtained by 50% aqueous methanol and by 6.56% compared to the extract obtained with
40% aqueous methanol.
Total flavonoid content was highest in the extract obtained by extraction with 50%
aqueous methanol (37.944 mg CE/g), and the lowest by using 40% aqueous methanol
(24.092 mg CE/g). The yield of total flavonoids in the extract obtained by extraction
with 50% aqueous methanol as solvent was by 57.50% higher than yield obtained with
40% aqueous methanol, and by 17.46% compared to yield obtained using 60% methanol.
The effect of the concentration of aqueous methanol solutions (40%, 50, 60%) on the
antioxidant activity of methanol extracts was also tested. On the basis of the obtained
results, it can be concluded that the sage leaf extracts obtained with 50% and 60%
aqueous methanol solutions have a good antioxidant ability to inhibit free DPPH radi‐
cals. The lowest antioxidant activity showed the extract obtained by extraction with 40%
aqueous methanol solution.
References
1. Beattie, S.D.: Bioenergetics and oxidative metabolism. In: Devlin, T.M. (ed.) Textbook of
Biochemistry with Clinical Correlations, 6th edn, pp. 978–989. Wiley-Liss, New York (2006)
2. Kancheva, V.D.: Phenolic antioxidants – radical-scavenging and chainbreaking activity: a
comparative study. Eur. J. Lipid Sci. Technol. 111, 1072–1089 (2009)
3. Shi, H., Noguchi, N., Niki, E.: Introducing natural antioxidants. In: Pokorny, J., Yanishlieva,
N., Gordon, M. (eds.) Antioxidants in Food, Practical Applications, 1st edn, pp. 22–70.
Woodhead Publishing Limited, Cambridge (2001)
4. Reische, D.W., Lillard, D.A., Eitenmiller, R.R.: Antioxidants. In: Akoh, C.C., Min, D.B.
(eds.) Food Lipids: Chemistry, Nutrition and Biotechnology, 3rd edn, pp. 410–413. CRC
Press, Boca Raton (2008)
5. Halliwell, B., Murcia, M.A., Chrico, S., Aruoma, O.I.: Free radicals and antioxidants in food
and in vivo: what they do and how they work. Food Sci. Nutr. 35, 7–20 (1995)
6. Lobo, V., Patil, A., Phatak, A., Chandra, N.: Free radicals, antioxidants and functional foods:
impact on human health. Pharm. Rev. 4(8), 118–126 (2010)
7. Pokorný, J., Yanishlieva, N.V., Gordon, H.: Antioxidants in Food – Practical Applications,
pp. 22–70. Woodhead Publishing, Cambridge (2001)
8. Dai, J., Mumper, R.: Plant phenolics: extraction, analysis and their antioxidant and anticancer
properties. Molecules 15, 7313–7352 (2010)
9. Cushnie, T.P., Lamb, A.J.: Antimicrobial activity of flavonoids. Int. J. Antimicrob. Agents
26, 343–356 (2005)
10. Katalinić, V., Smole, M.S., Skroza, D., Generalić, I., Abramović, H., Miloš, M., Ljubenkov,
I., Piskernik, S., Pezo, I., Terpinc, P., Boban, M.: Polyphenolic profile, antioxidant properties
and antimicrobal activity of grape skin extracts of 14 Vitis vinifera varieties grown in
Dalmatia (Croatia). Food Chem. 119, 715–723 (2010)
11. Castellano, G., Tena, J., Torrens, F.: Classification of phenolic compounds by chemical
structural indicators and its relation to antioxidant properties of Posidonia Oceanica (L.)
Delile. MATCH Commun. Math. Comput. Chem. 67, 231–250 (2012)
12. Heim, K.E., Tagliafero, A.R., Bobilya, D.J.: Flavonoid antioxidants: chemistry, metabolism
and structure-activity relationships. J. Nutr. Biochem. 13, 572–584 (2002)
13. Mandal, S.M., Chakraborty, D., Dey, S.: Phenolic acids act as signaling molecules in plant-
microbe symbioses. Plant Signal. Behav. 5(4), 359–368 (2010)
14. Zhang, Y., Seeram, N.P., Lee, R., Feng, L., Hebe, D.: Isolation and identification of strawberry
phenolics with antioxidant and human cancer cell antiproliferative properties. J. Agric. Food
Chem. 56, 670–675 (2008)
15. Harborne, J.B., Baxter, H. (eds.): The Handbook of Natural Flavonoids, vol. 2. John Wiley,
Chichester (1999)
16. Buhler, D., Miranda, C.: Antioxidant action of flavonoids. Micronutrients Information Center,
Linus Pauling Institute, London (2000)
17. Xu, B.J., Chang, S.K.: A comparative study on phenolic profiles and antioxidant activities of
legumes as affected by extraction solvents. J. Food Sci. 77(2), 159–166 (2007)
18. Albu, S., Joyce, E., Paniwnyk, L., Lorimer, J.P., Mason, T.J.: Potential for the use of
ultrasound in the extraction of antioxidants from Rosmarinus officinalis for the food and
pharmaceutical industry. Ultrason. Sonochem. 11, 261–265 (2004)
19. Bimakr, M., Rahman, R.A., Taip, F.S., Ganjloo, A., Md Salleh, L., Selamat, J., Hamid, A.,
Zaidul, I.S.M.: Comparision of different extraction methods for the extraction of major
flavonoid compounds from spearmint (Mentha spicata L.) leaves. Food Bioprod. Process.
89, 67–72 (2011)
20. Ross, K.A., Beta, T., Arntfield, S.D.: A comparative study on the phenolic acids identified
and quantified in dry beans using HPLC as affected by different extraction and hydrolysis
methods. Food Chem. 113, 336–344 (2009)
21. Altiok, E., Baycin, D., Bayraktar, O., Ülkü, S.: Isolation of polyphenols from the extracts of
olive leaves (Olea europaea L.) by adsorption on silk fibroin. Sep. Purif. Technol. 62, 342–
348 (2008)
22. Shi, J., Nawaz, H., Pohorly, J., Mittal, G., Kakuda, Y., Jiang, Y.: Extraction of polyphenolics
from plant material for functional foods-engineering and technology. Food Rev. Int. 21, 139–
166 (2005)
23. Shan, B., Cai, Y.Z., Sun, M., Corke, H.: Antioxidant capacity of 26 spices extracts and
characterization of their phenolic constituents. J. Agric. Food Chem. 53, 7749–7759 (2005)
24. Guyot, S., Marnet, N., Drilleau, J.: Thiolysis-HPLC characterization of apple procyanidins
covering a large range of polymerization states. J. Agric. Food Chem. 49, 14–20 (2001)
25. Mattila, P., Kumpulainen, J.J.: Determination of free and total phenolic acids in plant derived
foods by HPLC and diode array detection. J. Agric. Food Chem. 50(13), 3660–3667 (2002)
26. Robards, K.: Strategies for the determination of bioactive phenols in plants, fruit and
vegetables. J. Chromatogr. A 2003, 657–691 (1000)
27. Areias, F.M., Valentäo, P., Andrade, P.B., Ferreres, F., Seabra, R.M.: Flavonoids and phenolic
acids of sage: influence of some agricultural factors. J. Agric. Food Chem. 48, 6081–6084
(2000)
28. Durling, N.E., Grey, O.J., Webby, R.F., Mitchell, K.A., Foo, L.Y., Perry, N.B.: Extraction of
phenolics and essential oil from dried sage (Salvia officinalis) using ethanol-water mixtures.
Food Chem. 101, 1417–1424 (2007)
29. McDonald, S., Prenzler, P.D., Antolovich, M., Robards, K.: Phenolic content and antioxidant
activity of olive extract. Food Chem. 73, 73–84 (2001)
30. Stanković, M.: Total phenolic content, flavonoid concentration and antioxidant activity of
Marrubium peregrinum L. extracts. J. Sci. 33, 63–72 (2011)
31. Pourmorad, F., Hosseinimehr, S.J., Shahabimajd, N.: Antioxidant activity, phenol and
flavonoid contents of some selected Iranian medicinal plants. Afr. J. Biotechnol. 5(11), 1142–
1145 (2006)
32. Pontis, J.A., da Costa, L.A.M.A., da Silva, S.J.R., Flach, A.: Color, phenolic, and flavonoid
content and antioxidant activity of honey from Roraima, Brazil. Food Sci. Technol. 34(1),
69–73 (2014)
33. Brighente, I.M.C., Dias, M., Verdi, L.G., Pizzolatti, M.G.: Antioxidant activity and total
phenolic content of some Brazilian species. Pharm. Biol. 45(2), 156–161 (2007)
34. Naczk, M., Shahidi, F.: Extraction and analysis of phenolics in food. J. Chromatogr. A
1054(1–2), 95–111 (2004)
35. Chan, S.W., Lee, C.Y., Yap, C.F., Wan Aida, W.M., Ho, C.W.: Optimisation of extraction
conditions for phenolic compounds from limau purut (Citrus hystrix) peels. Int. Food Res. J.
16, 203–219 (2009)
36. Chirinos, R., Rogez, H., Campos, D., Pedreschi, R., Larondelle, Y.: Optimization of
extraction conditions of antioxidant phenolic compounds from mashua (Tropaeolum
tuberosum Ruíz and Pavón) tubers. Sep. Purif. Technol. 55(2), 217–225 (2007)
37. Bouterfas, K., Mehdadi, Z., Benmansour, D., Khaled, M.B., Bouterfas, M., Latreche, A.:
Optimization of extraction conditions of some phenolic compounds from white horehound
(Marrubium vulgare L.) Leaves. Int. J. Org. Chem. 4, 292–308 (2014)
38. Do, Q.D., Angkawijaya, A.E., Tran-Nguyen, P.L., Huynh, L.H., Soetaredjo, F.E., Ismadji,
S., Ju, Y.H.: Effect of extraction solvent on total phenol content, total flavonoid content, and
antioxidant activity of Limnophila aromatica. J. Food Drug Anal. 22, 296–302 (2014)
39. Złotek, U., Mikulska, S., Nagajek, M., Świeca, M.: The effect of different solvents and number
of extraction steps on the polyphenol content and antioxidant capacity of basil leaves (Ocimum
basilicum L.) extracts. Saudi J. Biol. Sci. 23(5), 628–633 (2016)
40. Kim, I.S., Yang, M.R., Lee, O.H., Kang, S.H.: Antioxidant activities of hot water extracts
from various spices. Int. J. Mol. Sci. 12(6), 4120–4131 (2011)
41. Butsat, S., Siriamornpun, S.: Effect of solvent types and extraction times on phenolic and
flavonoid contents and antioxidant activity in leaf extracts of Amomum chinense C. Int. Food
Res. J. 23(1), 180–187 (2016)
42. Teixeira, T.S., Vale, R.C., Almeida, R.R., Ferreira, T.P.S., Guimarães, L.G.L.: Antioxidant
potential and its correlation with the contents of phenolic compounds and flavonoids of
methanolic extracts from different medicinal plants. Revista Virtual de Química 9(4), 1546–
1559 (2017)
43. Mohammedi, Z., Atik, F.: Impact of solvent extraction type on total polyphenols content and
biological activity from Tamarix Aphylla (L.) Karst. Int. J. Pharma Bio Sci. 2(1), 609–615
(2011)
44. Turkmen, N., Sari, F., Velioglu, Y.S.: Effects of extraction solvents on concentration and
antioxidant activity of black and black mate tea polyphenols determined by ferrous tartrate
and Folin-Ciocalteu methods. Food Chem. 99, 835–841 (2006)
45. Zhang, H.Y., Sun, Y.M., Wang, X.L.: Substituent effects on O-H bond dissociation enthalpies
and ionization potentials of catechols: a DFT study and its implications in the rational design
of phenolic antioxidants and elucidation of structure–activity relationships for flavonoid
antioxidants. Chem. Eur. J. 9(2), 502–508 (2003)
Occurrence of Apple Powdery Mildew,
Podosphaera Leucotricha (Ellis. & Everh.)
E. S. Salmon in North-Western Region
of Bosnia and Herzegovina
Zemira Delalić(&)

zemirabtf@gmail.com
Abstract. Apple powdery mildew (Podosphaera leucotricha (Ellis. & Everh.)

E. S. Salmon) along with apple scab pathogen (Venturia inaequalis (Cooke)
Winter) represents economically the most important apple pathogen. Apple
powdery mildew was monitored during 2016 and 2017 in north-western area of
Bosnia and Herzegovina at 10 orchards of the following varieties: Idared,
Braeburn, Gala, Fuji, Elstar, Granny Smith, Cripps Pink, Jonathan, Golden
Delicious and Red Delicious. A scale from 0–5 was used for determination of
infection intensity on leaves of the examined varieties. The results were anal-
ysed using Townsend Heuberger method providing the infection index. High
infection indexes were determined for Idared, Granny Smith, Jonathan
(58–80%) in 2016 and 2017, which were therefore categorized as varieties
highly sensitive to P. leucotricha. Cripps Pink variety was moderately sensitive.
Moderate resistance was registered for varieties: Elstar, Golden Delicious. The
lowest infection indexes were recorded for Fuji and Gala varieties (3–8%), while
the Braeburn variety did not indicate any infection symptoms. More intensive
infections are recorded as a consequence of mild winters, dry and warm weather
during spring, heavy fertilization, sensitive varieties and modern breeding
methods. Sensitive varieties such as Idared, Jonathan and Granny Smith are
usually used for development of orchards, while significance of agrotechnical
and chemical measures are often underestimated in apple production. Important
protection measures include removal of white infected branches and its disposal.
For new orchards farmers should use resistant varieties. The basic protection
measure for P. leucotricha is use of fungicides, which may prevent secondary
infection of apple fruit, leaves and flowers.
Keywords: Apple Apple powdery mildew (Podosphaera leucotricha)

Infection intensity
1 Introduction
Apple is the second most widespread fruit culture in its presence, immediately behind
the plum and it has an important role in the overall fruit production in Bosnia and
Herzegovina. In 2015, there were 6 704 447 apple trees capable of yielding; the total
yield was 91 471 t, with an average of 13.6 kg per tree [1]. The total number of fruit

https://doi.org/10.1007/978-3-319-90893-9_57
492 Z. Delalić
apple trees in the Una-Sana Canton (North-Western part of B&H) was 155 231, the
total yield was 2 015 tons, an average of 13.0 kg per tree (Federation Institute of
Statistics - Una-Sana Canton in Figures, 2016). Total yield and yield per tree is still low
compared to large European apple and pear producers (Italy, Austria, Poland, Spain,
etc.). Apple is the most important continental type of fruit trees in the world in terms of
production. In recent years, world production has been growing slightly [2]. However,
one of the limiting factors in the production of this fruiting species in almost all
breeding areas is the appearance of apple powdery mildew (P. leucotricha), which,
along with the spotted leaves and the scabs on the fruit (V. inaequalis is the eco-
nomically most significant apple disease in our country, especially in sensitive varieties
such as Jonathan and “Honeycrisp”. The attack of pathogens, asbestos-causing agents
in the honeycrisp variety is more or less present every year, which certainly depends on
the climatic conditions and the presence of inoculums in the previous year [3]. In
Vojvodina, the powdery mildew of apple regularly appears along with the sediment of
scabies and causes a great damage [4]. The more massive appearance of apple powdery
mildew in the northwestern part was registered in 2009 [5]. Patogen P. leucotricha is
native to North America, where it was first observed in 1871. In Europe, it was first
registered in Tirol in 1873. Since then, it has been occurring regularly every year at a
stronger or lower intensity [6]. It’s spread all over where the apple is grown. The
appearance of P. leucotricha was also noticed in the Una-Sana Canton (North-Western
part of Bosnia and Herzegovina). Mortality of young fruits were determined from
29.33% to 58.0%, and on fruits from 22.0% to 42.0%. In the event of a strong attack,
leaves and flowers are drying, decreasing the increase in fruit trees, fruit yields and the
market price of fruits [7] (Stojanović 2004). Until the discovery and use of systemic
fungicides, disease was a limiting factor for the production of apples [8]. If protection
from this pathogen is omitted, the genus can be completely absent. The aim of this
paper is to determine the pathogenic fungus, P. leucotricha, and monitor the intensity
of infection on 10 apple saplings with the most commonly grown varieties in the
north-western part of Bosnia and Herzegovina. On the basis of the intensity of infec-
tion, one of the goals is to determine the resistance, i.e. the sensitivity of the varieties,
and recommend quality protective measures.
2 Materials and Methods of Work
In the two-year period (2016–2017), in June and July in the North-Western part of
B&H (Una-Sana Canton), the appearance of fungus was regularly monitored, the
causative agent of apple powdery mildew in 10 fruit plants with varieties: Idared,
Bramley, Gala, Fuji, Sugar Bee, Granny Smith, Cripps Pink, Jonathan and Red
Delicious. Diagnosing the cause of the disease was determined based on typical
symptoms on plants. Then samples of diseased leaves and apples of apple plants were
taken and in the microbiological laboratory the fungus was determined by microscopic
method. The semi-permanent microscopic preparations of fungus were made [9], and
microscopic observations and painted conidia and conidiofora were observed. The
assessment of the infection was carried out on the scales of the scale 0–5. Intensity of
infection is calculated according to Townsend-Heuberger. For each variety, 200 sheets
Occurrence of Apple Powdery Mildew 493
were selected by random selection. The infected leaves are ranked according to the
scale: 0 = no symptoms; 1 = 1–10%; 2 = 10–25%; 3 = 25–50%; 4 = 50–75%; and
5 = 75–100% of infected leaves. The index of the disease was calculated according to
the Townsend Heuberger formula:
I%= ∑(n×v) ⁄ NV×100
where I% = index of the disease; n = degree of infection by scale; v = number of sheets

per category; N = Highest degree of infection; V = total number of sheets examined.
Based on the disease index, the resistance of the varieties was determined in the fol-
lowing way: 0-very resistant variety (all plants are completely healthy, index of disease
0); I-resistant (index of disease 10); II = medium-resistant (disease index 11–25);
III = medium sensitive (index of disease 26–50) and IV = very sensitive (index of
disease above 50).
3.1 Symptoms and Fungus Properties

During the year 2016 P. leucotricha was detected on 6 apple plants (60% of the
examined plants) and in 2017 at 8 plants (80% of the examined plants). Fungus attack
was registered on flower buds, leaves and shoots. The aforementioned pathogen
develops on all green apple parts, creating a characteristic flour and white leaf coat
[10]. In spring, younger buds and flowers emerge from infected buds, and are com-
pletely covered with mycelium and conidia of pathogens. The less affected flower buds
move for 5–8 days later than healthy ones. The flowers are covered with mycelium and
spores of fungus, while the leaflets of infected flowers are light-yellow or light-green
coloured, deformed and obtuse. Spotted flower buds are more sensitive to freezing and
give less fruit. Infected terminal buds are completely infected with spores. Partially
affected buds give “white shoots” (Fig. 1), which rise for several weeks, then discards
leaves [11]. The fungus develops epiphyseal mycelium on leaves, shoots, flowers, and
young fruits, so they look like they were covered with powdery mildew, which is why
the disease is called “powdery mildew”. Because of the development of the coating, the
fungus cover the green parts of the plant that are less assimilated [12]. Individually
infected leaves occurring during vegetation were secondary infected. On the reverse
side of these leaves spruce with a whitish shroud were developed, chlorosis, and later
necrosis and insertion of the leaf [13]. The infected leaves lose their green color, they
dry and drop (Fig. 2). Due to secondary infections the apple is lagging behind, does not
thicken and it is easier to froze. The weakest symptoms are on the fruit. Their epidermis
becomes like net, incused in epidermis, small fruits, deformed, with small cracks on the
epidermis. The multi-annual presence of “ powdery mildew” in the orchard is seen on
the trees of the so-called “rare crown”. In seedlings production, when the top of the
apple seedlings was infected, developmental disruption occurs and such seedlings may
dry out.
It was found that P. leucotricha was developed as an epiphytic. On the surface of the
plant organs it forms a mycelium with conidiophores and conidia, as confirmed by
494 Z. Delalić
Fig. 1. (a) White shoots (Photo by: Delalić Z.) (b) White shoots (Photo by: Delalić Z.)
Fig. 2. Necrosis and leaves twisting caused by fungus P. leucotricha (Photo by: Delalić Z.)
microscopy in the laboratory (Figs. 3 and 4). The mycelium is attached by the
appressorium for plant tissue, and it is fed by haustoria. The conidias are in series (6–9),
single-celled, elliptical size 12–20 23–30 lm. Cleistothecium may occur in groups of
luteur or leaf handles, diameter 75–96 lm, round or pear-shaped. Two types of
appendices are formed on them, the ones that are dichotomous at the top, upright, while
others are basically rudimentary light-headed, simple and rarely split. In cleistothecium
there is an elliptical askus (45–55 50–70 lm) with eight ascospores. Ascospores are
single-celled, colorless, ellipsoidal (size 12–14 22–36 lm). Pathogen development
in the north-western part of Bosnia and Herzegovina was favorably influenced by mild
and wet winters. The greatest influence is on the appearance of the infection [14].
Temperatures of −20 °C to −25 °C can completely destroy fungus. The conidia is
germinated at a temperature of 4–30 °C, and the optimum is 20–22 °C. They can
germinate without a drop of water, but a higher relative humidity is required (75–100%).
Optimum air humidity for the development of P. leucotricha is 90% [15]. The pathogen
develops poorly if the spring is cold and rainy, and at summer temperatures above 33 °C
the mycelium stops developing. Inflammation of the conidia and infection process
negatively influences light [16]. The results of many studies show that increasing the
Fig. 3. Conidia of fungus P. leucotricha Fig. 4. Conidiophore of fungus P. leucotricha

(Photo by: Delalić Z.) (Photo by: Delalić Z.)
intensity of light was caused by higher germination of conidia [17–19]. A large number
of conidia formed in sunny days can be destroyed by UV radiation and reduce the
infectious potential of the pathogen [20]. On the contrary, in the conditions of reduced
light, a smaller number of conditions were formed, and even inhibits them in the dark.
P. leucotricha is a parasite of many species of the genus Malus (apple, pear, quince,
medlar), and the species of the genus Photinia. Data [21] show that mushrooms para-
sitise in almond and peach.
3.2 Indexes of Illness and Resistance Varieties

During 2016, the causative agent of apple powdery mildew, P. leucotricha was
detected in 6 apple trees (60% of examined dairy plants), and in 2017 on 8 plantations
(80% examined). Table 1 shows the indexes for the studied apple varieties, as well as
their sensitivity. The disease indexes ranged from 0–80% and were higher in 2017 due
to more favorable climatic factors for the development of apple powdery mildew. In
2016 and 2017, as the most vulnerable, the variety of Idared was distinguished with the
index of disease of 75–80%, then the varieties Granny Smith and Jonathan. In the
group of medium-sensitive varieties, Cripps Pink with a disease index of 26–38% was
classified. In the Elstar, Golden Delicious and Red Delicious varieties, disease indexes
were 15–23% indicating that they are medium resistant varieties. Fuji and Gala
varieties were resistant to low intensity disease, and there was no infection with
P. leucotricha on Braeburn. The attack of pathogens, asbestos-causing agents on Idared
and Jonathan varieties has been present every year, which certainly depends on the
climatic conditions and on the presence of inoculum in the previous year [22]. Due to
the actuality of this pathogen, protection measures must be implemented each year.
496 Z. Delalić
Table 1. Index of illness and resistance varieties

Sorta Index of Resistance/sensitivity of varieties
illness (%)
2016 2017
Idared 75 80 Very sensitive
Granny Smith 70 74 Very sensitive
Jonathan 58 61 Very sensitive
Cripps Pink 26 38 Medium sensitive
Elstar 20 23 Medium resistant
Golden Delicious 15 22 Medium resistant
Red Delicious 12 18 Medium resistant
Fuji 5 8 Resistant
Gala 3 6 Resistant
Braeburn 0 0 Very resistant
4 Conclusion
The pathogenic fungus P. leucotricha regularly occurs on fruit apple plants in the North-
Western part of Bosnia and Herzegovina and causes significant damage to leaves, young
shoots, flowers and fruits. The appearance is being promoted by mild winters, dry and
warm weather in the spring, stronger intake, sensitive variety, modern way of breeding.
The optimal temperature of 10–20 °C contributes to strong infection. The fungus does not
require the presence of free water for infection. Conidia contain sufficient water resources,
so often neither high relative humidity is necessary. The appearance of the pathogen in
2016 was weaker than in 2017 because the spring was colder and rainier, so the mycelium
was less developed from the buds and the younger. When cultivating plants, most com-
monly used cultivars are susceptible to this pathogen such as Idared, Jonathan and Granny
Smith. According to the farmers’ statements, in practice, the quality of agro-chemical and
chemical protective measures are not implemented, which indicates their lack of infor-
mation on modern apple cultivation. Of the protection measures, it is important to remove
“white shoots” during cutting, both during the winter and spring (green cuttings) and their
destruction (burning or burial). The rows of the cultivations should be placed in the
direction of blowing the dominant winds, in order to ensure good air circulation and low
relative humidity. When raising new plantations, producers need to choose resistant
varieties, since later protection is difficult and expensive. The basic measure of protection
against P. leucotricha is the use of fungicide. Fungicides are most commonly applied at
intervals of 7 to 10 days, starting from the phenophase when the flower buds are visible,
and until the phenophase when the shoots reach their full upgrowth. This treatment
provides a permanent presence of deposits of fungicides on newly formed herba-
ceous parts. In apple seedlings with domination of the apple varieties very sensitive to
P. leucotricha, it is sometimes necessary to carry out several treatments during a vege-
tation season. In practice, sulfur-based preparations have been used for long time, then
protective fungicides. Today, fungicides with systemic action are used.
References
1. Agencija za statistiku Bosne i Hercegovine (2016)
2. Nikolić, D., Keserović, Z., Magazin, N., Paunović, S., Miletić, R., Nikolić, M., Milivojević,
J.: Stanje i perspektive razvoja voćarstva u Srbiji. IV Kongres voćara i vinogradara sa
međunarodnim učešćem, Zbornik radova i apstrakata, Vrnjačka Banja, pp. 3–22 (2012)
3. Trajčevski, T.: Pepelnica jabuke, uzročnik Podosphaera leucotricha (Ell. i Ev.) Salm. i
djelotvornost zaštite novijim fungicidima, Glasnik zaštite bilja 5, pp. 6–42 (2007)
4. Balaž, J., Knežević, T.: Efikasnost novijih fungicida za suzbijanje čađave krastavosti i
pepelnice jabuke. Pesticidi 18, 175–185 (2003)
5. Dervišević, S., Delalić, Z., Jogić, V.: Analiza zdravstvenog zasada jabučastog voća na
području općine Sanski Most, Zbornik radova, Smotra naučnih radova studenata agronomije
sa međunarodnim učešćem, Čačak, pp. 41–48 (2009)
6. Miletić, N.: Podosphaera leucotricha (Ell. i Ev.) Salm. - prouzrokovač pepelnice jabuke,
Biljni lekar, vol. 6, pp. 450–456 (2000)
7. Stojanović, S.: Poljoprivredna fitopatologija, Srpsko biološko društvo “Stevan Jakovljević”,
Kragujevac (2004)
8. Delalić, Z.: Fitopatologija-specijalni dio. Grafičar, Bihać (2007)
9. Milošević, D.: Fitopatologija-praktikum, Agronomski fakultet, Čačak (2008)
10. Cvjetković, B.: Pseudomikoze voćaka i vinove loze. Zrinski d.d., Čakovec, p. 534 (2010)
11. Ivanović, M., Ivanović, D.: Mikoze i pseudomikoze biljaka. Poljoprivredni fakultet, Beograd
(2001)
12. Ciglar, I.: Integrirana zaštita voćnjaka i vinograda. Zrinski d.d, Čakovec (1998)
13. Miletić, N., Tamaš, N.: Integralna zaštita jabuke, Zbornik radova II Savetovanja ‘‘Inovacije u
voćarstvu’’, Beograd, pp. 95–106 (2009)
14. Xu, X.-M.: The effects of constant and fluctuating temperatures on the length of the
incubation period of apple powdery mildew (Podosphaera leucotricha). Plant Pathol. 45,
924–932 (1996)
15. Butt, D.: Apple powdery mildew: an overview of the biology. In: Apple and Pear Mildew
and Canker. East Mailing Research Association, 1 (1995)
16. Stephan, S.: Research into the sporulation and dispersal of spores of apple mildew
(Podosphaera leucotricha (Ell. et Ev.) Salm.). Archiv fur Phytopathologie und Pflanzen-
schutz 24, 491–501 (1988)
17. Xu, X.-M., Butt, D., Ridout, M.: Temporal patterns of airborne conidia of Podosphaera
leucotricha, causal agent of apple powdery mildew. Plant. Pathol. 44, 944–955 (1995)
18. Jacob, D., Rav David, D., Sztjenberg, A., Elad, Y.: Conditions for development of powdery
mildew of tomato caused by Oidium neolycopersici. Phytopathology 98, 270–281 (2008)
19. Mieslerová, B., Lebeda, A.: Influence of temperature and light conditions on germination,
growth and conidiation of Oidium neolycopersici. J. Phytopatology 158, 616–627 (2010)
20. Willocqet, L.D., Colombet, M., Rougier, J., Fargues, M., Clerjeau, M.: Effects of radiation,
especially ultraviolet B, on conidial germination and mycelial growth of grape powdery
mildew. Eur. J. Plant Pathol. 102, 441–449 (1996)
21. Butt, D.: Podosphaera leucotricha (Ell. & Ev.) Salmon. In: Smith, I.M., et al. (eds.) European
Handbook of Plant Diseases, pp. 263–265. Blackwell Scientific Publication, London (1986)
22. Miletich, N., Tamas, N., Vuksa, P., Pfaf-Dolovac, E., Dolovac, N.: The influence of shading
on the development of Podosphaera leucotricha under field conditions. Bul. J. Agric. Sci.
18(2), 178–184 (2012). Agricultural Academy
Application of Multivariate Statistic
to Classify Blueberry Fruits
Vildana Alibabić1 ✉ , Azra Skender1, Melisa Orašćanin1, and Ibrahim Mujić1,2

( )
1
vildana.btf@gmail.com, skender.azra71@gmail.com,
melissa.bajramovic@gmail.com, ibrahim.mujic@ri.t-com.hr
2
Colegium Fluminense Polytechnic of Rijeka, 51000 Rijeka, Croatia
Abstract. Blueberry (Vaccinium myrtillus) is kind of the berry fruit which has
recently been grown in the area of Bosnia and Herzegovina. A few farmers
decided to plant this fruit, but there is no scientific data on the properties of the
fruit so far, hence the need for this research. The research examines the morpho‐
logical, chemical and sensory characteristics of the three cultivars of blueberry
(Goldtraube, Bluecrop and Blueray), grown in the Una - Sana Canton, as well as
a share of total phenols (TP), anthocyanins (TA) and ascorbic acid (TC). There‐
fore, the multivariate statistics were used for highlighting the similarities or
differences in specific parameters.
The Blueray cultivarhas the best morphological properties, as well as the
highest dry matter share (13.72%), total acid (2.33%), sugar (11.76%) and ash
(0.27%). TP content (29.75–59.34 mgGAE/100 g) was lower than in blueberries
grown in the region, while TA (213.91–684.84 mg/100 g) and TC (7.52–9.16 mg/
100 g) were a bit higher. Significant differences (p ≤ 0.05) among cultivars were
observed for all measured parameters. By using the PCA Bluecrop cultivar was
characterized with the share of water and Blueray with the share of ash, dry mater,
sugar and total acid.
Keywords: Antioxidants · Blueberry · Fruit characteristics

Multivariate statistic
1 Introduction
Increase of fruit consumption and fruit demand is a trend present in recent years in the
world (and in Bosnia and Herzegovina as well). This trend occurred for several reasons.
One of the reasons is increased purchasing power among residents, especially developed
countries, and also increased knowledge of the positive effects on health that come from
micronutrients and phytochemicals from fruit has contributed to this trend [1]. It has been
established that health benefits come mainly from natural antioxidants [2]. Each fruit group
has a different phytochemical profile, and a large number of studies show that berries are
one of best sources of phytochemicals. Inspired by these studies, the group of authors from
Norway researched (with FRAP method) the share of antioxidants in over 3000 foodstuffs
and concluded that the highest proportion of antioxidant components can be found in

https://doi.org/10.1007/978-3-319-90893-9_58
Application of Multivariate Statistic to Classify Blueberry Fruits 499
medicinal, spice and aromatic herbs and berries [3]. Among the berries, blueberries are
distinguished by the share of total phenols [4] with a significant proportion of antho‐
cyanin, which was found to be excellent in the fight against heart and cardiovascular system
diseases, as well as neurodegenerative diseases [5]. Anthocyanins are particularly inter‐
esting because they are stable compounds. Besides their antioxidant effect, it has been
discovered that they also have anticancerogenic and antimutagenic [6], as well as anti-
diabetes effect [7]. They are excellent for eye health and good eyesight [8], and they are
good in fight against a number of other chronic illnesses [9]. Blueberries contain a signifi‐
cant amount of vitamin C. The results of some studies suggest that blueberry and vitamin
C may have different mechanisms for the antioxidative effects or different sensitivities to
different oxidative stress-related biomarkers [10]. The nutritional value of blueberry -
which is reflected in the share of sugar, organic acids, vitamins and minerals - is not incon‐
sequential. Blueberry contains a significant share of dietary fibers that positively affect
heart function and increase insulin sensitivity [11]. One of the most important factors for
the consumers is their recognizable sweet taste, fruity aroma and other sensory properties
that this fruit has [12]. Nevertheless, there are many factors influencing the chemical
profile, antioxidant share and the sensory properties of blueberries. Among the most
important factors are climate, the growing season, soil quality, degree of ripeness [13], as
well as the time and the method of harvesting, conditions and the duration of storage after
harvest [14], but one of the most significant factors hides in the cultivar of blueberry. The
aim of this study was to define morphological, chemical and sensory characteristics of three
sorts of blueberries (Goldtraube, Bluecrop and Blueray) which are produced in Una - Sana
Canton. Using the multivariate statistical methods in this study, it has been examined which
of the researched sorts has advantages and/or disadvantages in relation to the investigated
parameters and what are their interrelationships and differences. Based on that, one
cultivar, with the best characteristics for cultivation in the investigated area, was distin‐
guished.
2 Eksperimental Part
Three cultivars of blueberries (Goldtraube, Bluecrop i Blueray) planted in 2014 were

analyzed in this research. The producers are located in the municipality of Cazin, which
geographically belongs to the North - Western part of Bosnia and Herzegovina and admin‐
istratively to the Una - Sana Canton (USC). The season was extremely rainy (the amount
of rainfall was 1934.3 mm) and the average air temperature was 11 °C. The fruit samples
were harvested when they reached their harvesting maturity and transported to the labora‐
tory of the Biotechnical Faculty, University of Bihać. For each sort, the 30 repetitions were
performed, except for chemical content for which three repetitions were performed. The
morphological parameters (fruit width, length and weight) were measured, using the
method described by Milivojević et al. [15]. For the measurement of dry matter (refracto‐
metrically), ash (525 ± 25 °C), acidity (titration with 0.1 N NaOH solution) and reducing
sugar (volumetrically according to Luff - Schoorl method), the method described by Vračar
[16] was used and cellulose (Weender analysis) described by Prosky et al. [17]. For the
determination of the total phenols content (TP) the Folin-Ciocalteu method adapted
500 V. Alibabić et al.
according to Waterhouse [18] was used, for the anthocyanins (TA), the pH differential
method has been used [19] and for vitamin C, iodine titration described by Ciancaglinia
et al. [20]. Sensory evaluation method was obtained following previously published meth‐
odologies described by Vračar [16], with the help of the Panel of 5 members who evalu‐
ated the appearance, flavor, taste and color of blueberry using a scores system for appear‐
ance (5–15), flavor (10–20), taste (1–3) and color (1–10). The members of Panel tested
previously according to the Guide published by ISO Sensory analysis [21] and they were
chosen on the basis of their senses for the color and shades recognition and sensitivity
levels for taste and flavor differentiation. For statistical analysis in addition to basic statis‐
tics, analysis of variance (ANOVA) and Duncan’s test revealed significant differences
between varieties at p ≤ 0.05 significance level. For the correlation between all parame‐
ters, the Pearson correlation coefficient (r) was used and the classification of samples was
performed by using multivariate analysis by the method of Principal Component Analysis
(PCA) and grouping data using Cluster Analysis (CA). For statistical analysis and data
processing, we used the computer program XLSTAT 2011 and STATISTICA 10.0.
2.1 Results and Discussion

The results of morphological parameters were shown in Table 1, as well as the chemical
profile and content of antioxidative components. The best morphological traits of blueber‐
ries were recorded in the Blueray cultivar (the largest average fruit width was 1.41 cm,
height 0.91 cm and weight with 2.04 g). The dimensions of Blueray fruit are significantly
different (p ≤ 0.05) than the dimensions of two other cultivars. Compared with the data from
the literature, the dimensions are smaller in relation to the dimensions of fruits in Poland
[22] where the measured width of the Blueray sort was 1.7 cm and Bluecrop 1.7 cm. Blue‐
crop sort was identified as the best overall. In Serbia, Leposavić [23], measured the weight
of the Bluecrop sort and it was in the range of 1.56 g to 160 g, which is slightly higher on
the average compared to the same sort in Una-Sana Canton. Savić [24] measured the
weight of the four sorts from 1.14 g to 2.21 g and concluded that the average weight of the
fruit was greatly influenced by the sort and vegetation year.
In terms of the chemical composition of the Blueray cultivar, which contains the highest
average dry matter share (13.72%), total acids (2.33), reducing sugar (11.76%) and ash
(0.27%), the sort still significantly different (p ≤ 0.05) from other sorts. Skupień [25] meas‐
ured a somewhat higher share of dry matter in the Blueray cultivar in Poland (14.46%), a
lower value for acidity (0.87%) and ash (0.19%), and a similar proportion of reducing sugars
(11.33%), while Leposavić [23] measured a similar dry matter content (13.55%), a slightly
lower acidity value (0.78%) and reducing sugars (8.95%), but a higher proportion of cellu‐
lose (1.02% in Bluecrop cultivar) compared to the Una-Sana Canton measurements
(0.46%). Comparing the chemical composition with the data from the literature, minor devi‐
ations can be found regarding water share, dry matter, sugar, ash and cellulose.
When it comes to the total acid value measured in this study (0.91–2.33%), overall they
differ from the values in literature (0.16–1.3%) and are higher. Among all three cultivars
statistical significance (p ≥ 0.05) of dissimilarities was determined in the content of antiox‐
idant components, with the highest content of TP found in the Blueray cultivar, TA in the
Goldtraube cultivar and TC in the Bluecrop cultivar. Dragović-Uzelac et al. [26] determined
Table 1. The results of fruit dimensions, chemical composition and content of measured
antioxidative components in different varieties of blueberries
Parameters (n = 30) Blueberries
Goldtraube Bluecrop Blueray
c b
Fruit width (cm) 1.00 ± 0.07 1.17 ± 0.11 1.41 ± 0.12a
c b
Fruit length (cm) 0.77 ± 0.07 0.85 ± 0.07 0.91 ± 0.08a
Fruit weight (g) 1.10 ± 0.16c 1.46 ± 0.24b 2.04 ± 0.35a
a a
Dry matter(%) 13.10 ± 0.21 12.01 ± 1.78 13.72 ± 0.77a
Acidity(%) 1.90 ± 0.08b 1.91 ± 0.01b 2.23 ± 0.09a
a a
Ash (%) 0.18 ± 0.01 0.20 ± 0.01 0.27 ± 0.04a
Reducing sugar(%) 8.56 ± 1.27ab 4.23 ± 1.63b 11.76 ± 0.01a
Cellulose(%) 0.46 ± 0.01a 0.44 ± 0.021a 0.37 ± 0.04a
c b
TP (mg GAE/100 g) 29.75 ± 0.02 31.50 ± 0.26 59.34 ± 0.02a
TA (mg/100 g) 684.84 ± 0.02a 213.91 ± 2.33c 263.16 ± 2.21b
b a
TC(mg/100 g) 9.03 ± 0.07 9.16 ± 0.11 7.52 ± 0.05c
a, b, c
The same letter in the same row indicates no significant differences (p ≤ 0.05) between the varieties.
that the TP content of the Bluecrop cultivar in Croatia (291.56–368.33 mg GAE/100 g)

depends on the air temperature, sun exposure and the growing season, and is much higher,
as well as study conducted by Silva et al. [27] on the Goldtraube and Bluecrop cultivars
(115–630 mgGAE/100 g) of the share in this study, but at the same time the anthocyanins
share is significantly lower (56–263 mg /100 g) compared to their share in the same sorts in
Una - Sana Canton (213.91–684.84 mg/100 g). A similar content of TA is measured in
blueberries from other areas of B&H (north-eastern Bosnia, Kladanj, Sreberenica, Fojnica)
in a range of 380–449 mg/100 g, but it’s a lower rank [28]. Share of vitamin C is slightly
lower in blueberries found in Una-Sana Canton than the percentage found in other studies
[29] for Bluray cultivar (11.8 mg/100 g). Comparing the results of content of TP, TA and
TC with their content in other berries from the same areas of research we observed a much
higher content of TC (35.77–54.92 mg/100 g) in the raspberry cultivars [30] and slightly
higher values of TC (13.4 - 16.2 mg/100 g) in the blackberry cultivars [31], but much lower
content of TP (21.1–25.6 mg GAE/100 g) and TA (153.0–238.7 mg/100 g) in the black‐
berry cultivars [31]. In the same time, the content of TP is much higher (164.54–416.24 mg
GAE/100 g) in the raspberry cultivars and content of TA lower (53.38–58.83 mg/100 g)
[30]. Commonly it can be concluded that compared to the data from the literature the share
of vitamin C in the corresponding ratio, as well as the share of anthocyanins, while in the
case of total phenol the proportion measured in the sorts in Una-Sana Canton is consider‐
ably lower compared to other studies (103.18–689.9 mgGAE/100 g).
Significant (p ≤ 0.05) Pearson correlation between morphological and chemical param‐
eters and share of antioxidative components for all blueberry cultivars was shown in the
Table 2. A significant positive correlation (p ≤ 0.05) was found between the morphological
parameters and share of TP, while a significant negative correlation (p ≤ 0.05) was found
between the morphological parameters and share of TC and TA.
Table 2. The significant (P ≤ 0.05) Pearson correlation between morphological and chemical
parameters and content of antioxidative components
Parameter (+) Correlation coefficient (r) (−) Correlation coefficient (r)
Fruit width (cm) 0.749 (length); 0.718 (weight); 0.783 (TP) −0.738 (TC); −0.639 (TA)
Fruit length (cm) 0.749 (width); 0.508 (TP) −0.522 (TA)
Fruit weight (g) 0.718 (width); 0.783 (TP) −0.742 (TP); −0.601 (TA)
Dry matter (%) 0.745 (sugar)
Acidity (%) 0.617 (DM); 0.703 (ash); 0.833 (sugar)
Ash (%) 0.703 (acidity); 0.653 (sugar)
Reducing sugar 0.745 (DM); 0.833 (acidity); 0.653 (ash)
(%)
Cellulose (%) . .
The sensory evaluation of the best characteristics showed that Blueray cultivar gath‐
ered highest points in terms of appearance (15 points), scent (18 points), taste (3 points)
and color (10 points). The difference in sensory properties was very statistically signif‐
icant (p ≥ 0.05) between all three cultivars. Blueray sort had equal yields in size, color,
shape and degree of maturity, without soil particles, disease or pest damage, with good
taste and smell. According to the Official Gazette of the Republic of Croatia, the National
Code on the Quality of Fruits, Vegetables, Mushrooms and Pectin [32]. Blueray is
classified as the extra class according to established characteristics, while the Goldtraube
and Bluecrop cultivars are classified as the class one with good sensory properties.
Finally, in order to determine the specific differences between the cultivars of blue‐
berries, an independent method of multivariate statistical data was carried out. Principal
components analysis (PCA) was used to determine the interrelationship between the
different parameters, while the Cluster analysis (CA) was used for grouping the data.
The variables in this study case are measured parameters for morphological, chemical
and sensory characteristics for all blueberry sorts, as well as data for antioxidant compo‐
nents in produce (TP, TA, TC). The results are also shown in Fig. 1 for the first two
factors (PC1 vs. PC2), for distinction between measured parameters and for distinction
between cultivars.
The first factor of PC1 has a specific value of 4.71 and explains 33.62% of the total
variability and is in the greatest positive correlation with the width and mass of fruit,
the share of TP, acid, ash and reducing sugar. The second factor (PC2) has a specific
value of 3.68 and explains an additional 26.27% of the total variability with the best
positive correlation with the width and height of the fruits and the water share. PC1
shows a negative correlation with TA, TC and water share, while PC2 shows a negative
correlation with the share of dry matter, acid, sugar and cellulose. There are two groups
in the picture: Bluecrop and Blueray. Bluecrop sort could be characterized by water
share, and Blueray with shares of ash, acid, reducing sugar and dry matter. Goldtraube
has no specific parameter to characterize it by. Using the agglomerative hierarchical data
grouping (AHC) method (Fig. 2), a dendogram was obtained for the researched samples
on the basis of the blueberry cultivar and there were three distinguished homogeneous
Fig. 1. Two-dimensional PCA scatter plot for blueberry cultivars
groups. From left to right, the first homogeneous group being the samples of the Gold‐
traube cultivar, the second group being Blueray cultivar samples and the third is Blue‐
crop which further confirmed the differences in the investigated parameters.
Dendrogram
500
450
400
350
Dissimilarity
300
250
200
150
100 Goldtraube Blueray Bluecrop

50
0
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BC
BR
BR
BR
BR
BR
BR
BR
BR
BR
BR
BR
BR
BR
BR
BR
BR
BR
BR
BR
BR
BR
BR
BR
BR
BR
BR
BR
BR
BR
BR
GT
GT
GT
GT
GT
GT
GT
GT
GT
GT
GT
GT
GT
GT
GT
GT
GT
GT
GT
GT
GT
GT
GT
GT
GT
GT
GT
GT
GT
GT
Fig. 2. Dendogram for blueberry cultivars according differences between measured parameters
3 Conclusion
Determining by the morphological characteristics, the chemical composition and the

sensory properties of the three researched cultivars of blueberries, have shown that
Blueray is the best-characterized cultivar. Blueray showed a significant share of antho‐
cyanins, but a lower share of total phenolic components. The highest share of total
phenols was recorded in the Goldtraube cultivar, which had the highest proportion of
cellulose, while the Bluecrop cultivar had a slightly higher vitamin C share than the
other two cultivars. Positive correlation was determined between the dimension of the
fruit and the share of the total phenols, while the PCA analysis characterized Blueray
by its chemical composition. Variance analysis found statistically significant differences
in all researched parameters between cultivars, and AHC grouping segregated them into
completely homogeneous groups. It can be concluded that the Blueray cultivar generally
showed the best overall characteristics in this study. Considering the range of generally
modest assortment, producers were encouraged to use this cultivar for cultivation.
References
1. Zmarlicki, K.: Global changes on blueberries’ markets. In: International Conference

Vaccinium spp. and Less Known Small Fruits: Cultivation and health benefit, 30 September–
5 October 2008, IPGB SAS, Nitra, Slovak Republic, p. 50 (2008)
2. Wang, H., Cao, G., Prior, R.L.: Total antioxidant capacity of fruits. J. Agric. Food Chem.
44, 701–705 (1996)
3. Carlsen, M.H., Bente, C., Halvorsen, L., Holte, K., Siv, K., et al.: The total antioxidant content
of more than 3 100 foods, beverages, spices, herbs and supplements used worldwide. Nutr.
J. 9, 3 (2010)
4. Castrejón, A.D.R., Eichholz, I., Rohn, S., Kroh, L.W., Huyskens-Keil, S.: Phenolic profile
and antioxidant activity of highbush blueberry (Vaccinium corymbosum L.) during fruit
maturation and ripening. Food Chem. 109, 564–572 (2008)
5. Huxley, R.R., Neil, H.A.W.: The relation between dietary flavonol intake and coronary heart
disease mortality. A meta-analysis of prospective cohort studies. Eur. J. Clin. Nutr. 57, 904–
908 (2003)
6. Marko, D., Puppel, N., Tjaden, Z., Jakobs, S., Pahle, G.: The substitution pattern of
anthocyanidins affects different cellular signaling cascades regulating cell proliferation. Mol.
Nutr. Food Res. 48, 318–325 (2004)
7. Jayaprakasam, B., Vareed, S.K., Olson, L.K., Nair, M.G.: Insulin secretion by bioactive
anthocyanins and anthocyanidins present in fruits. J. Agric. Food Chem. 53, 28–31 (2005)
8. Canter, P.H., Ernst, E.: Anthocyanosides of Vaccinium myrtillus (bilberry) for night vision
– a systematic review of placebo-controlled trials. Surv. Ophthalmol. 49, 38–50 (2004)
9. Moriwaki, T., Kato, S., Kato, Y., Hosoki, A., Zhang-Akiyama, Q.M.: Extension of lifespan
and protection against oxidative stress by an antioxidant herb mixture complex (KPG-7) in
Caenorhabditis elegans. J. Clin. Biochem. Nutr. 53, 81–88 (2013)
10. Kim, M., Na, H., Kasai, H., Kawai, K., Li, Y.S., Yang, M.: Comparison of blueberry
(Vaccinium spp.) and Vitamin C via antioxidative and epigenetic effects in human. J. Cancer
Prev. 22(3), 174–181 (2017)
11. Sáyago-Ayerdi, S.G., Arranc, S., Serrano, J., Gońi, I.: Dietary fiber content and associated
antioxidant compounds in Roselle flower (Hibiscus sabdariffa L.) beverage. J. Agric. Food
Chem. 55(19), 7886–7890 (2007)
12. Saftner, R., Polashock, J., Ehlenfeldt, M., Vinyard, B.: Instrumental and sensory quality
characteristics of blueberry fruit from twelve cultivars. Postharvest Biol. Technol. 49, 19–26
(2008)
13. Skrovankova, S., Sumczynski, D., Mlcek, J., Jurikova, T., Sochor, J.: Bioactive compounds
and antioxidant activity in different types of berries. Int. J. Mol. Sci. 16, 24673–24706 (2015)
14. Capocasa, F., Scalzo, J., Mezzetti, B., Battino, M.: Combining quality and antioxidant
attributes in the strawberry. The role of genotype. Food Chem. 111, 872–878 (2008)
15. Milivojević, J., Nikolić, M., Dragišić-Maksimović, J.: Pomological and antioxidant fruit
properties of strawberry cultivars grown in Mačva region. Agroekonomie 15(5), 83–90 (2009)
16. Vračar, L.: Manual for quality control of fresh fruit, vegetables, mushrooms and non-alcoholic
beverages, Manual, Faculty of Technology, University of Novi Sad, Serbia (2001)
17. Prosky, P., Asp, N.G., Schweizer, T.F., De Vries, J.W., Furda, I.: Determination of insoluble,
soluble, and total dietary fiber in foods, food products: interlaboratory study. J. Assoc. Off.
Anal. Chem. 71, 1017–1023 (1988)
18. Waterhouse, A.: Folin - Ciocalteau Micro Method, Department of Viticulture & Enology,
University of California, Davis (2006). http://waterhouse.ucdavis.edu/phenol/
folinmicro.htm. Accessed 2 Dec 2014
19. Wrolstad, E.R., Terry, E.A., Eric, A.D., Penner, M.H., Reid, D.S., Schwartz, S.J., Shoemaker,
C.F., Smith, D.M.: Handbook of Food Analytical Chemistry-Pigments, Colorants, Flavors,
Texture and Bioactive Food Components. Wiley, Hoboken (2001)
20. Ciancaglinia, P., Santosa, H.L., Daghastanli, K.R.P., Thedei, G.: Using a classical method of
vitamin C quantification as a tool for discussion of its role in the body. Biochem. Mol. Biol.
Educ. 29, 110–114 (2010)
21. ISO Sensory analysis, General guidance for the selection, training and monitoring of
assessors-Part 1. ISO 8586-1(E) (1993)
22. Wach, D.: Estimation of growth and yielding of highbush blueberry (Vaccinium corymbosum
L.) cultivated on soil developed from weakly loamy sand. Folia Hortic. 20(2), 47–55 (2008)
23. Leposavić, A.P.: Pomološke osobine novointrodukovanih sorti visokožbunaste borovnice
(Vaccinium corymbosum L.), Disertation, Faculty of Agriculture, University of Novi Sad,
Novi Sad, Serbia (2014)
24. Savić, Z.: Influence of cultivar on quantity and composition of essential elements,
polyphenolic compounds and yield of blueberry cultivars (Vaccinium spp), Disertation,
Faculty of Agriculture, University of Zagreb, Zagreb, Croatia (2011)
25. Skupień, K.: Chemical composition of selected cultivars of highbush blueberry fruit
(Vaccinium corymbosum L.). Folia Hortic. 18(2), 47–86 (2006)
26. Dragović-Uzelac, V., Savić, Z., Levaj, B., Bursać-Kovačević, D., Biško, A.: Evaluation of
phenolic content and antioxidant capacity of blueberry cultivars (Vaccinium corymbosum
L.) grown in the Northwest Croatia. Food Technol. Biotechnol. 48(2), 214–221 (2010)
27. Silva, S., Costa, E.M., Coelho, M.C., Pereira, M.F., Costa, M.R., Pintado, M.M.: Comparison
of bioactive compounds in four blueberry cultivars throughout the years: selection of the best
cultivars to be used in health promoting foodstuffs. In: Congress on Food Structure Design,
Portugal, vol. 1, pp. 15–17 (2014)
28. Hodžić, A., Vranac, A., Kurtović, M., Grahić, J., Memić, S., Gaši, F.: Anthocyanins content
in bilberry (Vaccinium myrtillus L.) fruits from three populations in Bosnia and
Herzegovina), Works of the Faculty of Agriculture and Food Sciencies, University of
Sarajevo, vol. LXII, no. 67/2, pp. 140–145 (2017)
29. Kampuse, S., Šnē, E., Šterne, D., Krasnovam, I.: Chemical composition of highbush blueberry
cultivars. Latv. J. Agron. 12, 53–59 (2009)
30. Alibabić, V., Skender, A., Bajramović, M., Šertović, E., Bajrić, E.: Evaluation of
morphological, chemical, and sensory characteristics of raspberry cultivars grown in Bosnia
and Herzegovina. Turk. J. Agric. For. 42, 67–74 (2018)
31. Skender, A., Ajdinović, T., Bećirspahić, D., Kurtović, M., Hadžiabulić, S., Aliman, J., Ercisli,
S.: The comparison of pomological characteristics of commercial improved and wild
blackberry genotypes,Works of the Faculty of Agriculture University of Sarajevo, vol. 62,
no. 67(1), pp. 17–22 (2017)
32. Official Gazette of the Croatia, National Rulebook of quality of fruit, vegetables, mushrooms
and pectines, NN. 46/07 (2007)
The Effect of Technological Process
on Physico-Chemical and Nutritional Properties
of Sour Cherries Products
Ramzija Cvrk1 ✉ , Azra Begović1, Snježana Marić1, and Nils V. Juul2

( )
1
Faculty of Technology, University of Tuzla, Univerzitetska 8,
75 000 Tuzla, Bosnia and Herzegovina
ramzija.cvrk@untz.ba
2
Sør-Trøndelag University College (HiST), Høgskolen i Sør-Trøndelag,
7004 Trondheim, Norway
nils.v.juul@hist.no
Abstract. The aim of this research was to determine the effect of selected
processing methods on physico-chemical and nutritional properties of sour cher‐
ries (Prunus cerasus L.) products. In laboratory conditions, is applied lyophili‐
zation (freeze drying) of sour cherries, and vacuum concentration to obtain
concentrate of sour cherry. For obtained products were analyzed of vitamin C (L-
ascorbic acid), crude fiber, sugar content, total fruit acids, and color intensity using
by CIELAB system, with coordinates (L, a*, b*). Experimental results in lyophi‐
lized sour cherry showed a significantly higher (p < 0,05) content of vitamin
C(35,76 ± 0,05 mg/100 g), crude fiber (4,80 ± 0,04%), total fruit acids
(7,23 ± 0,05%), and significantly higher levels of red color spectrum a*
(32,09 ± 1,00). This indicates that sour cherries processed by lyophilisation
largely retain nutritionally valuable substances, in comparison with sour cherry
concentrate obtained by the vacuum concentration.
Keywords: Lyophilization · Vacuum concentration · Added nutritional value
1 Introduction
Consumption of some fruits is strongly linked with several health benefits due to their
high nutritional value and medicinal properties [1]. The importance of fruits in the
human diet comes from its specific nutritional value based more on biological and less
energy value. This fact indicates a relatively high technological demands in processing
technologies of fruits. As many studies have shown that there is a direct correlation
between the preservation of human health and the ways of their diet, or the kind of food
you consume. For that reason, researchers are increasingly focusing on the study of
various factors in the process of influencing preserving the original quality of the raw
materials, and thus the quality of the finished product, as well as study and improve the
various processes of drying (lyophilization/freeze drying, spray drying), the procedures
of freezing and storage conditions and storage of raw materials and finished products.

https://doi.org/10.1007/978-3-319-90893-9_59
508 R. Cvrk et al.
Such scientific research is increasingly concerned with the study of the operations
and processes that aim to produce food products with added nutritional value and better
sensory properties. It is important to choose the proper technologies and devices, and
optimize the process conditions in order to be as little as possible disturb the original
quality of raw materials and get of high quality product with sustained nutritients.
During the experiment analyzed of some important quality parameters in fruits –
sour cherries and sour cherry products obtained in the experiment- lyophilized (freeze-
drying) sour cherry and sour cherry concentrate. In this research used sour cherry vari‐
eties “Oblačinska” origin from Bosnia and Herzegovina.
2.1 Description of the Fruits and Obtained Sour Cherry Product Analysis
Fruit were harvested at consumption maturity stage when fruits were fully coloured and
soft. Fresh sour cherries was pitted and ground in a laboratory blender (MOD.CL5,
volume 5 L, 230 V, FIRMAR, Italy) for analysis soluble solids content (SSC), expressed
as percentage (˚Brix) by using digital refractometer RX-5000 (AD-11, AC adapter, serial
number 960721 W). Titratable acidity (TA) was determined by titrating 50 ml of 1:5
diluted sour cherry pulp with 0,1 mol L−1 NaOH. Three titration analyses per sample
were performed. Titratable acidity was expressed in percentage of malic acid. The
soluble solid content/titrable acidity ratio (SSC/TA ratio) was used as an indicator of
taste quality [2]. The pH value were measured by using pH meter (METTLER TOLEDO
FE20). The total dry substance of the cherries and cherry products is determined using
a conventional method by drying at 105 ˚C. The sugar content was determined following
the Luff-Schoorl method [3], and determination of insoluble fibers in samples was
carried out according to KÜRSCHNER-Hanak method [3].
Ascorbic acid content (vitamin C) was determined by HPLC, according to [4,
5]. During analysis of the HPLC system used (Agilent Technologies 1260 series
infinity), with the column Zorbax SB C-18 (INF column no.69), mobile phase of water
at pH 2.2, flow rate 0.4 ml/min, detection 254 nm, and injection volume of 20 μl at room
temperature.
Extraction of sample for analysis of AA was prepared by weighing out appr. 5 g of
fruit or product sample into large centrifuge tubes. Add 50 ml of 1 mg/ml m-phosphoric
acid and homogenize using the POLYTRON homogenizer, and centrifuge at 12000 rpm
for 15 min (SORVALL RC 5B, USA). Filter the extracts through a fluted filter and store
at −80 ˚C.
Measure the pH of the extracts. The pH should be in the range between 2 and 2.5.
Adjust pH with 1M sulphuric acid if necessary. The concentration of vitamin C is deter‐
mined from a standard curve made from fresh prepared solutions. Prepare a standard
solution with concentration 0.8 mg/ml ascorbic acid. Make a serial dilution into: 0.2 mg/
ml, 0.1 mg/ml and 0.025 mg/ml. Run two parallels of each standard. Inject 20 μl of each
sample, 2–3 times. If the response is outside the linear range of the standard curve, dilute
the samples. Calculate the vitamin C concentrations graphically from the standard curve.
The Effect of Technological Process 509
Color measurements for all samples were conducted using The CIELAB colour
system (CIE, Commission Internationale de l’Eclairage 1986) with its values L*, a* and
b* using Minolta Chroma Meter CR-200 (Minolta Co. Ltd, Japan). Interpretation of L*,
a* and b* color system is carried out in accordance with Chroma Meter Instruction
manual (Minolta), where L*- lightness variable -from black (0) to white (100), a* -
chromaticity variable, defining the color on scale from green (−60) to red (+60), and
b*- chromaticity variable, defining the color on scale from yellow (+60) to blue (−60).
All parameters were analyzed for fresh sour cherries, lyophilized sour cherries, and
concentrate of sour cherries obtained in the process of vacuum concentration as
described in the section materials and methods.
2.2 Preparation of Lyophilised Sour Cherry

Lyophilization (freeze drying) of sour cherries (Prunus cerasus L.) was conducted using
of Lyophilizer (LEYBOLD-HERAEUS GT2, Germany). Before the of the lyophiliza‐
tion process, sour cherry was frozen at –20 °C. Sour cherry is then weighed and placed
in the lyophilizer. Lyophilization is carried out at temperature –40 °C, initial pressure
was 2,5 mbar, and at end process of lyophilization pressure was 1,0 mbar. After comple‐
tion of the process cherry promptly weigh and pack due to the hygroscopic properties
of lyophilized product. The calculation moisture as the percent loss in weight after drying
of sour cherries determined in accordance with [6], Eq. (1):
(fresh sample − dried sample) × 100

Water content (%) = (1)
fresh sample
2.3 Preparation of Sour Cherry Concentrate
Sour cherry concentrate is obtained using of laboratory Rotavapor R-210 (BUCHI,

Switzerland). Previously pitted sour cherries was milled and squeezed in a laboratory
blender to obtain sour cherry juice and then concentrated at a temperature of 50 °C and
pressure of 75 bar, to a dry matter min. 45 brix (ref., 20 °C).
2.4 Statistical Analysis
Results obtained it this investigation analysed by different statistical analysis including

T-test, standard deviation and analysis of variance (ANOVA) was carried out using
SPSS software (version 17). Duncan’s test was used to ranking of samples were statis‐
tically different by all the properties and acceptability (P < 0,05).
Results obtained in this research shown that sour cherry fruit is very high quality in all
investigated physical and chemical characteristics (Table 1). Value of the soluble solid
content/titrable acidity ratio were 9.578, and indicating a favorable taste and maturity
510 R. Cvrk et al.
of sour cherry fruit. Also, a high content of ascorbic acid (vitamin C), reducing sugars,
crude fiber, and total organic acids in sour cherry fruit (Table 1), placed this fruit in focus
as fruits suitable for the production of food products with “added nutritional value”.
According to many authors [7–12] it is very important intensity of significantly higher
levels of red color spectrum a*, because the color is very important indicator of quality
of fresh fruit. It also serves for estimating the stage of maturity of fruit. Plant pigments
responsible for the color of some kind of fruits are anthocyanins. Anthocyanins are the
flavonoids which are present in high amounts in sour cherries [7]. In this research red
color spectrum a* for sour cherry fruit is 20,16 (±1,53)b, for sour sherry concentrate
1,76 (±0,37)a, and value of red color spectrum a* for lyophilized sour cherry is 32,09
(±1,00)c. Also, according to [8] better indicator of anthocyanin accumulation are L*
color parameters. According to this claim, in this study it is shown that fresh sour cherry
also has a high value color spectrum L* 22,13 (±1,58)b, for lyophilized sour cherry is
33,64 (±1,13)c, and for sour cherry concentrate this value is 12,46 (±0,90)a. Values of
the color spectrum obtained in this research clearly indicating that the selection proper
technological processing of the sour cherry fruit, possible obtain product with high
values of color, which indicates a high content of anthocyanins in lyophilized sour cherry
and sour cherry fruit.
Table 1. Results of all investigated parameters (Duncan’s test)

Parameters Sour cherry fruit Sour cherry Lyophilized sour
concentrate cherry
pH value 3,46 (±0,01)b 3,37 (±0,02)b 3,10 (±0,10)a
Soluble solids content 17,05 (±0,05)a 45,79 (±0,08)b 97,63 (±0,65)c
(˚Brix)
Total dry matter at 18,05 (±0,19)a 47,86 (±1.03)b 98,09 (±0,72)c
105 °C (%)
Total acidity (g/100 g) 1,78 (±0,02)a 4,16 (±0,07)b 7,23 (±0,05)c
Crude fiber content 0,84 (±0,01)b 0,78 (±0,005) a
4,80 (±0,04)c
(%)
Ascorbic acid content 17,16 (±0,03)b 12,73 (±0,07)a 35,76 (±0,05)c
(mg/100 g)
Reducing sugars 8,57 (±0,00)a 15,96 (±0,02)a 36,64 (±0,38)c
content (%)
Color parameter - L* 22,13 (±1,58)b 12,46 (±0,90)a 33,64 (±1,13)c
Color parameter - a* 20,16 (±1,53)b 1,76 (±0,37)a 32,09 (±1,00)c
b a
Color parameter - b* 8,17 (±0,51) 4,34 (±0,85) 10,04 (±1,27)c
a,b,c
- Mean values in the same rows, with different superscript are significantly different (P < 0.05)
Cherry concentrate has a significantly lower (P < 0,05) values of the red color spec‐
trum a* and color spectrum L* and indicating to the conclusion that it impact of the
processing sour cherry by concentration process.
Also, this research shows comparison of obtained parameters for two selected tech‐
nological process of cherries. The results showed that there was statistically significant
The Effect of Technological Process 511
difference (P < 0,05) in all investigated parameters. All the parameters of sour cherry
obtained by lyophilization (freeze-drying process) have high value, and are more accept‐
able in nutritional aspect, than sour cherry concentrate obtained in the concentration
process.
Comparing obtained parameters from two selected process (Table 1), should high‐
light of several times higher values of ascorbic acid (vitamin C) content, red color spec‐
trum a* and color spectrum L*, total acidity- total content of organic acids, and reducing
sugars content in lyophilized sour cherry and that recommended of lyophilization
(freeze-drying process) in the processing sour cherry.
4 Conclusion
Sour cherry fruit, showed a very high quality for all investigated parameters, and may
acquire new interest, mainly due to the fact that it can be considered as suitable for
production food with “added nutritional value”. This research has indicated that the
selection proper technological processing of the sour cherry fruit, possible obtain
product with high nutritional values. Also, it is shown that process of lyophilization
(freeze-drying) largely keep all the important nutritional properties of sour cherry fruit.
Acknowledgements. This research was supported by the Norwegian government through HERD
(Programme for Higher Education, Research and Development) project “EDUFOOD -
Development of education and transfer of knowledge in the area of food technology”.
References
1. Ames, B.M., Shigena, M.K., Hagen, T.M.: Oxidants, antioxidants and the degenerative
diseases of aging. Proc. Natl. Acad. Sci. USA 90, 7915–7922 (1993)
2. Crisosto, C.H., Crisosto, G.M., Metheney, P.: Consumer acceptance of ‘Brooks’ and ‘Bing’
cherries is mainly dependent on fruit SSC and visual skin color. Postharvest Biol. Technol.
28, 159–167 (2003)
3. Trajkovic, J., Miric, M., Baras, J., Siler, S.: Analysis of foodstuffs. University of Belgrade
(1983)
4. Romero Rodiguez, M.A., Vazquez Oderiz, M.L., Lopez Hernandez, J., Simal Lozano, J.:
Determination of vitamin C and organic acids in various fruits by HPLC. J. Chromatogr. Sci.
30(11), 433–437 (1992)
5. Pineli, L.L.O., Moretti, C.L., Santos, M.S., Campos, A.B., Brasileiro, A.V., Cordova, A.C.,
Chiarello, M.D.: Antioxidants and other chemical and physical characteristics of two
strawberry cultivars at different ripeness stages. J. Food Compos. Anal. 24(1), 11–16 (2011)
6. Wrolstad, R.E.: Handbook of Food Analytical Chemistry: Pigments, Colorants, Flavors,
Components Texture, and Bioactive Food. Wiley-Interscience, Hoboken (2005)
7. Pedisic, S., Levaj, B., DragoviC-Uzelac, V., Kos, K.: Physicochemical composition, phenolic
content and antioxidant activity of sour cherry cv. marasca during ripening. Agriculturae
Conspectus Scientificus 73(3), 259–262 (2009)
8. Mozetic, B., Trebse, P., Simcic, M., Hribar, J.: Changes of anthocyanins and
hydroxycinnamic affecting the skin colour during maturation of sweet cherries (Prunus avium
L.). Lebensm.-Wiss.u.-Technol. 37, 123–128 (2004)
512 R. Cvrk et al.
9. Blando, F., Gerardi, C., Nicoletti, I.: Sour cherry (Prunus cerasus L.) anthocyanins as
ingredients for functional foods. J. Biomed. Biotechnol. 2004(5), 253–258 (2004)
10. Hartmann, A., Patz, C.-D., Andlauer, W., Dietrich, H., Ludwig, M.: Influence of processing
on quality parameters of strawberries. J. Agric. Food Chem. 56, 9484–9489 (2008)
11. Gundogdu, M., Bilge, U.: Determination of organic, phenolic, sugars and vitamin c contents
od some cherry cultivars (prunus avium). Int. J. Agric. Biol. 14, 595–599 (2012)
12. Papp, N., Szilvassy, B., Abranko, L., Szabo, S., Pfeiffer, P., Szabo, Z., Nyeki, J., Ercisli, S.,
Stefanocits-Banyai, E., Hegedus, A.: Main quality attributes and antioxidants in Hungarian
sour cherries: identification of genotypes with enhanced functional properties. Int. J. Food
Sci. Technol. 45(2), 395–402 (2010)
Application of New Technologies in Meat
Processing Industry in the Function
of Improvement of Total Quality of Products
and Consumer Protection
Amir Ganić(&), Munevera Begić, and Enver Karahmet

{a.ganic,m.begic,e.karahmet}@ppf.unsa.ba
Abstract. Just as it does in the world, the meat processing industry in Bosnia
and Herzegovina also represents one of the main branches of the
food-processing sector. In the last twenty years, there have been great devel-
opments and applications of technological solutions, which accompany this
branch of industry. Today, it is unimaginable that any production plant which
deals with meat processing does so without the help of tens of
highly-sophisticated machines (meat crusher, automatic mixers and fillers, micro
and macrocutters, pickl-injectors, tumblers-massagers, smoking chambers,
vacuum packing machines, etc.). Besides the fact that these machines, in their
basic intent, have in a significant deal replaced the human factor, their appli-
cation has other implications as well. Namely, the application of the newest
technological accomplishments in the meat industry not only has it in many
ways influenced the expansion of the palette of products, but also the overall
quality of the products as well. On the other hand, the technology of lasting
dry-cured meat products has even in the most advanced meat processing plants
kept the basic contours of traditional production, in which the human work is
predominantly present (meat cutting, ropes tying, salting/brining, drying,
smoking, etc.). With the intent to try and find modalities of application some of
the technological solutions in this production segment, the use of specific
machines for some technological phases was done experimentally. More pre-
cisely, the phase of salting and/or brining, which is traditionally done by hand,
was attempted in this study to be replaced by injecting machines
(pickl-injectors). In addition, the modification of the technological process of
production was performed in the segment of salting/brining. Namely, in this part
of the production process the use of tumbler-massagers was implemented, for
the purposes of the experiment. The aim of the application of these two con-
temporary machines was to establish the justification of their use from a tech-
nological standpoint. Besides that, their influence on the end quality of the
product was the most important element of the justification of their use.
Keywords: Meat processing industry New technologies

Technology of dry-cured meat products Product quality

https://doi.org/10.1007/978-3-319-90893-9_60
514 A. Ganić et al.
1 Introduction
Within the scope of the meat processing industry, one of the most important segments
is the production of lasting (dry-cured meat) products. In the technological sense, it is
characterized by a dominant or exclusively presence of human factor. The presence of
contemporary technological advancements is minimal. Even in the most advanced meat
processing plants, this activity kept the basic contours of traditional production. The
human factor is still crucial in some of the most important phases in the technology of
production (meat cutting, salting and/or brining, drying, smoking, etc.). Taking into
consideration the developments of new technologies and products, this approach has
been overcome. The market expects new products on a daily basis, which, at their core,
need to preserve some traditional values, but with appropriate innovations. Further-
more, the advices and recommendations of nutritionists, regarding nutrition, all point to
trends which, without a doubt, require a completely different technological approach.
Modifications primarily need to be made in terms of standardization of the final
product. In the production with exclusively traditional technological postulates (where
the human factor is crucial), it is almost impossible to unify the product, from both the
aspects of quantity and quality. In this regard, the problem of inconsistency of the
production and the quality of product is present. Additionally, one of the problems,
which we will face it with in the future, is the lack of work force. It is necessary to
make certain projections in that regard as well. As far as the quality is concerned, as per
recommendations of the nutritionists, certain modifications of the products are set as an
imperative. Before all others, it is necessary to reduce the presence of nitrates in the
finished product, which we mainly consume through NaCl (kitchen salt). Traditional
technology, in a rule implies the excessive use of NaCl. Due to all of the things pointed
out, for the needs of the research, the application of certain machines in some of the
phases of technological process was done experimentally. The aim of the research was
to determine the justification of their application in regarding of standardization of the
total quality of product, achieving certain nutritional properties, as well as to decrease
influence of the human factor in the overall technological process of production. In the
projection of this experiment, the working hypothesis for proving was set, which states:
through the substitution of the human factor by the certain machines in some phases of
the technological process, a significant change in the quality of the finished product will
not occur.
2 Current Traditional Approach of the Production

Technology
The production of dry-cured meat products, along the combined use of salting, drying,
smoking methods, has a very long tradition. Besides the undeniable role of tradition,
during the production of this group of products in one region, the general character-
istics of the region especially the specific climate conditions have a significant impact
on the production process and quality characteristics [1]. In general, a high amount of
kitchen salt is traditionally present in lasting dry-cured meat production the market of
Application of New Technologies in Meat Processing Industry 515
Bosnia and Herzegovina. It is often the case that an amount of over 10% of NaCl [2]
can be found in some of the products from this group. It has been proven that an
excessive intake of kitchen salt negatively influences the health status of the consumer.
These effects can be reflected through the direct risk of heart attack [9], hypertrophy of
the left chamber [3], increase of the hardening of the arteries, or in other words –
decrease of elasticity of the vessel walls [4], etc. Besides that, very often, it comes to
negligence in the technological procedure. These faults are exclusively tied to the
human factor, which is impossible to standardize and unify. As a result of these
anomalies, the flaws in the quality of the finished products (Figs. 1 and 2) are shown
through insufficient salting, color inequality, excessive saltiness of the external layers,
and likewise, insufficient saltiness in the interior of the product, etc.
Fig. 1. Traditional salting method Fig. 2. Inequality of product quality
Due to all this, a need to find certain solutions in the technological process of
production is set as an imperative, for the purpose of reduction of the human factor
influence, which for the most part influences the final quality of the product.
3 Application of New Technological Solutions

in the Production Process
In the attempt to find a substitution model for the human factor with some techno-
logical solutions, in the process of production of beef prosciutto, the use of certain
machines was implemented experimentally. During the salting and brining phase,
which is traditionally done by hand (Fig. 1), the use of a device for injection
(pickl-injector - Fig. 3) was implemented. Additionally, the modification of the tech-
nological process of production was done in the segment of the salting and brining.
Namely, in that part of the production process, the use of tumbler-massagers was
implemented. Injectors and meat massagers are used exclusively in the sausage pro-
duction and the production of half-permanent smoked products. Injectors are devices,
which, with the help of hollow needles that have a vertical gait, perform the intro-
duction and injection of brine in the pieces of meat (Fig. 3). The result of their activity
is a more complete and better quality salting across the entire cross-section of the piece
Fig. 3. Pickl meat injector Fig. 4. Tumbler for massage of meat
of the meat. The tumblers (massagers) are devices, which have a task to massage the
injected pieces of meat. Their primary task is to ensure that the injected meat retains as
much of the injected brine (brine is a mix of spices and additives in water) as possible
(Fig. 4).
For the purposes of this research, the following work parameters were set:
– brine was made in a way so that 100 L of brine consisted out of 90 L of water
(80:20% water:ice), with the addition of 10 kg of kitchen salt and the temperature
from −1 °C to +1 °C,
– during the injection the pressure of the injecting needles was set to 2 bars (because
of the destruction of the muscle tissues), and this was done twice (each piece of
meat was injected on two occasions),
– during the massage of the pieces, the tumblers had a vacuum work regime of 6 work
hours, with 15 min of work and 15 min of break with six spins per minute.
After the massaging, the pieces of meat were hanged on the sticks and the tradi-
tional cold smoking was performed in a classic smokehouse in the duration of seven
days (Fig. 6). For the purposes of the research, one batch of products was produced
(30 sample products–Fig. 5).
Fig. 5. Beef prosciutto Fig. 6. Cross-section of beef prosciutto

4 Qualitative Indicators of Prosciutto with the Use

of Contemporary Technological Solutions
The results of research expressed through the sensory evaluation and the chemical
quality of beef prosciutto were determined in the laboratory of the Faculty of Agri-
culture and Food Sciences of Sarajevo. A five-member expert commission get done the
sensory evaluation. A descriptive grade with the use of non-linear scoring system was
used for the following sensory properties [5]: the external appearance (3), color (2),
consistency (3), cross-section appearance (4), odour (3), and taste (5 points). On the
other hand, for determination of the chemical quality, standard methodology was used.
The content of moisture was determined with drying on up to 105 °C. Quantitative fat
presence was established with the Soxhlet method, proteins with the Kjeldahl method,
NaCl with the Mohr method, and the contents of ashes by burning at up to 580 °C. The
pH value was determined by a pH meter.
sensory of quality of prosciuƩo

the external
appearance
100
80 76,00
60 85.50
taste color
59.20 40
20
0
70.33
odour 74.30consistency
86.75
appearance secƟon
Chart 1. Sensory quality of beef prosciutto
Based on the sensory evaluation of the beef prosciutto (Chart 1) using the techno-
logical modifications, it can be concluded that the appearance of the cross-section
(86,75%) and the color (85,50%) are the best graded sensory properties. On the other hand,
taste with 59,20% of the maximum score was graded the weakest. Generally looking, the
application of technological solutions by using the injectors and massagers instead of the
human factor, gave, with the exception of taste, positive effects on all other sensory
parameters. The most significant segment of the sensory evaluation are the appearance of
the cross-section and the color of prosciutto. Namely, in the classical salting procedure
(human factor), variations from insufficient to extremely large amounts of NaCl added are
common. Both extremes have negative impact on the sensory quality, especially on the
appearance of the cross-section and the color. If there is not enough added kitchen salt,
quality salt diffusion into interior of the product will not be possible. In that case, there
remains unsalted meat on the cross-section (Fig. 2), which, with the bad sensory
impression, represents a very good environment for the growth and multiplication of
microorganisms. On the other hand, in case when the amount of added salt is too great, the
product becomes too salty, the taste is worse, and as such, is very harmful for the health of
the consumer. In this case, the taste received the weakest grades. The results of chemical
analysis show that the total amount of NaCl in the product was 4,19%. The Bosnia and
Herzegovina market is traditionally used to dry-cured meat products with somewhat of a
greater amount of salt. In that sense, the grade of the taste of the prosciutto was possibly a
result of subjective perceptions of the assessors. Also, in future research works, the
mechanism of NaCl diffusion into muscle fibers of meat through injection of brine – as
opposed to classical salting, should be tested, and in that sense it needs to be determined
whether there are differences. In comparison to the results reported by other researchers
[6–12], it is generally possible to conclude that the use of injectors and massagers in the
technological procedure of prosciutto manufacture is justified. Of course, in order to
primarily make improvements regarding the taste of the product, it is necessary to conduct
additional research, especially in terms of the contents of brine, as well as the work regime
of the tumblers (massagers).
Through chemical analysis (Chart 2) of the prosciutto sample, it was established
that the presence of moisture is 52,12%, proteins 25,37%, fats 14,68%, ash 5,21%,
NaCl 4,19% and the pH value was 5,24. If compared to the results that other
researchers reported, it can be concluded that, with the exception of the presence of
Chemical quality of prosciuƩo

60
50
40
30
%
20
10
0
moisture fat protein NaCl ash pH
Series1 52.12 14.68 25.37 4.19 5.21 5.24
parameters
Chart 2. Chemical quality of prosciutto

moisture, which was uncharacteristically high, for this type of meat product, other
parameters had approximatevalues. In the samples of “Visočka pečenica”, produced in
an industrial activity, Ganić et al. (2012) have established a 40,99% of moisture, fat
27,22%, protein 25,82%, ash 5,82%, and NaCl 4,96%. Stamenković (2003) points out
in his research that the content of moisture in beef prosciutto amounted 48,10% and
Radovanović et al. (2003) 44,36%. Čausevic et al. (1986) have also established an
average content of ash of 5, 23% in the samples of the beef prosciutto. Radovanović
et al. (2004) state that the content of ash in beef prosciutto from the town of Užice was
37,32% and fat 2,50%. Ganić et al. (2012) report that the average content of fat in
samples of homemade prosciutto was 20,73% and protein 28, 21%. Ganić et al.
(2016) report similar values of the contents of protein and fat in sheep “stelja”.
Chemical quality of prosciutto

60
50
40
%
30
20
10
0
moisture fat protein NaCl ash pH
Series1 52.12 14.68 25.37 4.19 5.21 5.24
parameters
Through an analysis of the total results of chemical analysis of prosciutto, it can be

said that only the content of moisture had non-standard values for this type of product.
Reasons for this need to be looked for in the fact that the samples were exposed to the
impact of brine, which in its content has over 90% of water. Besides that, reasons for a
moisture value this high need to be looked for in the fact that the samples went through a
massage treatment in tumblers, whose task was that meat retains the added moisture. In
order to reduce the value of moisture in a finished product to an optimal value, in further
research the phase of massaging could potentially be excluded from the technological
production procedure. In addition, the research should be directed towards increasing
the content of salt in brine, which would be transferred to the finished product, which
could in turn affect the reduction of moisture content in the finished product.
5 Conclusions
Based on the results of the research it can be concluded that there is a justifiable use of
advanced technologies, in this case of injectors and tumblers, in the production of
permanent dry-cured meat products. Through the use of contemporary technologies the
results of sensory evaluation show that the samples of prosciutto, has, with the
exception of taste, satisfactory quality. In addition, based on the conducted chemical
research, it can be stated that the finished product, with the exception of the contents of
moisture, has no significant deviations from the classic traditionally produced product.
The advantages of using the aforementioned devices in the production process can be
seen through the possibilities of standardization of the end product in all quality
parameters. Additionally, rationalization of the scope of the work of the machine in
comparison to the human factor is a very important factor. On the other hand, the
biggest disadvantage of using the advanced technologies is the procurement of the
equipment that is extremely expensive. In addition to that, maintenance is expensive as
well. To justify the aforementioned, it would be necessary to conduct research and
follow the technological process closely for a longer period of time in order to properly
judge and grade the use of these devices from the economic standpoint of their usage.
References
1. Dumić, S.: Examination of more important quality properties of “Sjenička stelja” like basis
for the protection of designations of origin. Master’s thesis, University of Belgrade, Faculty
of Agriculture, Belgrade (2008)
2. Ganić, A., Zorlak, A., Begić, M.: The influence of different quantities of the added NaCl on
the sensory properties of “Visočka pečenica”, Meat, Zagreb (2018). (Article in press No
180105)
3. Schmieder, R.E., Messerli, F.H.: Hypertension and the heart. J. Hum. Hypertens. 14, 597–
604 (2000)
4. Avolio, A.P., Clyde, K.M., Beard, T.C., Cooke, H.M., Ho, K.K., O’Rourke, M.F.: Improved
arterial distensibility in normotensive subjects on a low salt diet. Arteriosclerosis 6, 166–169
(1986)
5. Robović, B., Smajić, A., Tahmaz, J.: Sensory analysis of Bosnian prosciutto. Works of the
Faculty of Agriculture and Food Sciences University of Sarajevo, vol. 53, pp. 129–133
(2003)
6. Ganić, A., Lilić, S., Krvavica, M., Čandek-Potokar, M., Pejkovski, Z.: Main properties of
quality of “Visočka pečenica”. Meat Technol. 53(2), 134–139 (2012)
7. Kurćubić, V., Vesković-Moračanin, S.: Comparative testing of quality of beef prosciutto
produced in a traditional and industrial manner. In: Proceedings XXII Symposium on
Biotechnology, vol. 2, pp. 585–590 (2017)
8. Čaušević, Z., Milanović, A., Glogovac, Ž., Velagić-Habul, E., Smajić, A., Lelek, M.:
Attachment to knowing the production of beef prosciutto. Works of the Faculty of
Agriculture Sarajevo, vol. 38, pp. 153–161 (1986)
9. Sinanović, N., Smajić, A., Ganić, A.: Sensory evaluation of the quality of dry-cured meat
products at the market of Sarajevo Canton. Works of the Faculty of Agriculture and Food
Sciences University of Sarajevo, vol. 55, no. 2, pp. 177–187 (2005)
10. Stamenković, T.: Use of kitchen salt in meat products. Meat Technol. 45(5–6), 170–176
(2004)
11. Zorlak, A., Ganić, A., Begić, M.: The influence of different quantities of the added NaCl on
the sensory properties of “Visočka pečenica”. In: Book of Abstracts 28th International
Scientific-Expert Conference of Agriculture and Food Industry Sarajevo, p. 74 (2017)
12. Ganić, A., Radić, I., Stojković, S., Krvavica, M., Ratković, D.: Modification technology in
order to improve sensory quality of beef ham. In: Proceedings 10th Conference of Chemists,
Technologists and Environmentalists of Republic of Srpska Banja Luka, pp. 272–278 (2013)
13. http://es.joseplagares.com/sobre-mi/literatura/articulos-tecnologia/marinado-de-carnes-
frescas-por-efecto-spray/
The Impact of Water on PVC Floor Coverings
Mario Krzyk ✉ and Darko Drev

( )
University of Ljubljana, 1000 Ljubljana, Slovenia

{mario.krzyk,darko.drev}@fgg.uni-lj.si
Abstract. PVC floor coverings are more or less resistant to water. Plasticised
PVC typically has relatively loosely bound molecules of plasticizers and secon‐
dary plasticizers (DOP, BBP, etc.). In gelled PVC layers the durability is mostly
the result of physical binding of plasticiser molecules in between PVC chains.
PVC plastisol ingredients are by design not particularly solidly bound to the
structure, which makes them susceptible to leaching or migration of water mole‐
cules into the structure of PVC floor coverings. The plasticiser molecules and
other PVC paste compounding ingredients leach into the water (pigments, fillings,
stabilisers, foaming agents, inhibitors, etc.). The resistance to water is signifi‐
cantly improved with an upper polyurethane (PU) layer. In the PVC paste produc‐
tion, the PVC types using the emulsion process are highly suitable. The PVC types
so produced contain a significant proportion of emulsifier residue. Emulsifiers
have a strong water-binding ability, which is reflected in the opacity of the PVC
layer. Micro-suspension-type PVCs are much more stable concerning the impact
of moisture, but such PVC pastes are usually less stable.
Keywords: PVC floor coverings · Plasticizers · Gelation · Migration
1 Introduction
Several types of polyvinyl chloride (PVC) floor coverings are known, which differ
substantially one from another. They are classified into the following main groups:
– homogeneous PVC floor coverings (with one or several layers of the same material),
– heterogeneous PVC floor coverings (with one or several layers of different materials),
– composite PVC floor coverings (heterogeneous with an additional sub-base),
– foamed PVC floor coverings with a structured surface.
All these floor coverings use plasticized PVC. In residential buildings foamed PVC
floor coverings with a structured surface are commonly used, while for commercial and
industrial premises other non-foamed types of PVC floor coverings are used.
Foamed PVC floor coverings have, approximately, the following composition:
– 40–45% PVC
– 15–20% plasticiser
– 35–40% fillers
– 2–5% pigments
– 1–2% other additives

https://doi.org/10.1007/978-3-319-90893-9_61
The Impact of Water on PVC Floor Coverings 523
The PVC’s protective coating must allow for great wear-resistance, transparency,
and moisture-resistance and has optionally matt or gloss surface. PVCs with K values
of around 70 and 80 are used for gloss and matt surfaces, respectively. In most cases
these PVC types are combined with 20–30% of extender-type PVC. Rather than used
independently, plasticizer extenders are used in combination with other paste-type
PVCs. No more than 30% of the extender-type is added with regard to the total quantity
of PVC. Extender types improve rheological properties of pastes, as they are composed
of relatively large round particles with a diameter between 2 × 10−2 and 10−1 mm. K
values are somewhat lower than in other paste-type PVCs.
Fig. 1. Illustration of PVC paste gelation (Drev 1986)
The mechanical and, to a large extent, chemical durability of PVC floor coverings
depend on the gelation of the PVC paste (Drev 1986). Gelation of PVC pastes is a
physical process where the plasticiser molecules are inserted between the PVC chains
and bind the PVC particles together. The PVC particles suspended in the plasticizer are
of different shapes and sizes. As the plasticizer molecules migrate into the PVC particles,
the particles start to spread. The plasticizer migrates further inside the PVC particles
524 M. Krzyk and D. Drev
while the phase border between the liquid plasticizer and the solid PVC is getting smaller
until it completely disappears.
The PVC molecules and plasticisers are governed by various forces of attraction that
differently affect the gelation process. The behaviour during gelation is governed by the
forces that are the result of thermodynamic interaction, while static forces are not rele‐
vant here. Strong induction forces act between plasticizer molecules, which are not
temperature-dependent. During the heating of the PVC paste, the bonds between PVC
chains are loosened, because thermal motion overpowers the molecular forces of attrac‐
tion. The plasticizer molecules interpenetrate the loosened bonds to form dipole-dipole
associates, as shown in Fig. 1.
The migration of plasticizer molecules into PVC particles is mostly governed by the
forces between dipoles (Drev 1986), as demonstrated by the Eq. (1):
𝜇2 ( )
D 𝜕𝜀
v⃗ = grad T (1)
k T 12𝜋 𝜀0 𝜀2 r3 𝜕T
Where:
v⃗ is plasticizer migration rate [ms−1]

D is diffusion coefficient [m2s−1]
k is the Boltzmann constant [eVK−1]
T is temperature [K]
𝜇 is dynamic viscosity [kg m−1s−1]
𝜀0 is electric constant
𝜀 is dielectric constant
r is distance from the dipole to the upper point
2 The Impact of Water on PVC Floor Coverings
Water-resistance of PVC floor coverings depends on the composition of PVC coatings

and their gelation. No PVC floor covering is completely impermeable to water mole‐
cules. The presence of water has the following effects on PVC floor coverings:
– bacteria and moulds cause the emergence of colour spots on floor coverings,
– if there are emulsion-type PVCs present in the protected PVC layer the transparent
layer turns opaque,
– in foamed PVC layers it is important to distinguish between open-cell and closed-
cell foam structure:
– chemical foam has closed-cell structure and therefore low water absorption,
– mechanical foam, which is usually used for the back layer, has open-cell structure
and therefore high water absorption,
– the water absorbed in PVC layers can react with various admixtures (fillings, dyestuff,
reinforcing agents, etc.),
– the individual ingredients of PVC floor coverings can leach in the water - hot water
allows for more leaching to occur (e.g. hot steam treatment).
It should be underlined that plasticized PVC has significantly higher water perme‐
ability than non- or less-plasticized PVC, and that the impact of gelation on resistance
of PVC floor coverings against water is significant. In PVC floor coverings resistance
to water is measured using the following tests:
– moulding (Streptoverticillium reticulum strain of bacteria) causing pink decolouri‐
sation of the PVC floor covering (ref. ASTM E1428-99:2004),
– water tightness (standard EN 13553, Resilient floor coverings – Polyvinyl chloride
floor coverings for use in special wet areas – Specification), and
– sliding friction on wet floor surfaces (DIN 51130 Testing of floor coverings – Deter‐
mination of the anti-slip property).
Figure 1 shows that PVC has relatively low permeability to water vapour and oxygen
where plasticized PVC has a significantly higher permeability to water vapour and
oxygen (Figs. 2 and 3).
Fig. 2. The permeability of polymer layers to water vapour and oxygen (Eyerer et al. 2008a, b)
Fig. 3. The permeability of polymer layers to water vapour and oxygen (Domininghaus 2005)
The impact of the presence of water on PVC floor coverings is mostly aesthetic. Two
examples are shown in Figs. 4 and 5.
Fig. 4. Illustration of pink stains on a PVC floor covering as a result of the growth of moulds or
bacteria
Fig. 5. Illustration of the opaque surface of PVC floor coverings due to uptake of water molecules
in the emulsion-type PVC
3 Conclusion
We distinguish between several types of PVC floor coverings. The most commonly used
PVC floor coverings are produced based on PVC pastes, which are the focus of this
paper. These PVC floor coverings typically consist of an upper transparent protective
coating, the middle insulation layer based on chemical foam, and the back layer which
is usually based on mechanical foam. Nowadays, a polyurethane (PU) coating is
commonly applied to the upper PVC layer. If the transparent PVC layer contains the
emulsion-type PVC, it becomes opaque during prolonged water exposure. After several
days of drying the water slowly evaporates from the PVC layer and the layers turn
transparent again. This is much more difficult when, due to moisture, bacteria and
moulds begin to grow in PVC floor coverings – usually pink-coloured stains are
produced. With heavy bacterial growth these stains are also brown and they are not easily
removed.
PVC floor coverings are usually made of strips of a width of 4 m, which are laid on
the floor. The contacts between strips and edges are never completely water-tight. These
contact areas are most at risk of water penetrating the inside of a PVC floor covering.
Even when the joints are completely water-tight, PVC floor coverings are still permeable
to water to a certain degree, as shown in Figs. 2 and 3. Plasticized PVC is much more
permeable to water than non-plasticized PVC.
References
Dominighaus, H.: Die Kunststoffe und ihre Eigenschaften. Springer, Heidelberg (2005)
Drev, D.: Raziskaveželiranja PVC past: magistrskodelo. [samozal.], Ljubljana, 109 f.,
graf.prikazi. (COBISS.SI-ID 340527) (1986)
Eyerer, P., Elsner, P., Hirth, T.: Die Kunststoffe und ihre Eigenschaften. Springer, Berlin (2008a)
Eyerer, P., Hirth, T., Elsner, P.: Polymer Engineering. Springer, Heidelberg (2008b)
Exposure to PM10 Aerosol Particles and Other
Aerial Pollutants in the Capital City
of Montenegro
Jovana Jovanovic1(&) and Svetlana Stevovic2

1
University Union Nikola Tesla, 11000 Belgrade, Serbia
jocka747@gmail.com
2
University of Beograd, 11000 Belgrade, Serbia
svetlanas123@gmail.com
Abstract. Aerial contaminations are of booming interest in the 21st century as

they attenuate and disable favorable living conditions for each human being.
This manuscript represents trajectories on the occurrence of PM10, particulate
matters which are defined as atmospheric aerosol particles and emission den-
sities of other toxic gases such as carbon monoxide, and nitrogen-dioxide.
Trajectory on the occurrence of PM10 aerosol particles for one characteristic,
in-situ spot of capital city of Montenegro, (Podgorica) from January 1st to July
20th (2017) was detected. Trajectories on the occurrence of other airborne
pollutants (CO, NO2) from January 1st to June 30th (2017), were detected. As
plain of these toxic gases, the district Nova Varos in the capital city of
Montenegro was considered. All databases of these emissions were taken from
Montenegrin Environmental Protection Agency (MEPA). Collected temporal
data of aforementioned toxic gases emissions were presented either as daily
average values or as average values for the section of eight hours or hourly. In
attachment are also given the air quality index grids for Europe and USA. The
air quality situation of mentioned urban settlement Nova Varos in accordance
with European air quality index grid was analyzed, since European air quality
index grid encompasses Montenegrin air quality index grid. This manuscript
serves as insight into the degree of local air pollution and limitative, stringent
policies of air purification within one city. It reveals the genuine causes of such
aerial contamination and discovers statistically the occurence frequencies of
these particles in certain district.
In that way, this is a guide through a certain microclimate and the readers can
anticipate how much is a microclimate of a spot influential on a global scale.
Keywords: PM10 aerosol particles Airborne pollutants Trajectories

Air quality MEPA
1 Introduction
Fast-growing cities such as Beijing, New York and Paris produce broad strips and
plumes of aerial particles which are known as aerosols. Aerosols by definition, are
widespread specks of matter, recognizable as particles from soot, ash and smoke which

https://doi.org/10.1007/978-3-319-90893-9_62
530 J. Jovanovic and S. Stevovic
react with vapor and other substances in the air. They move rapidly and are detectable
by satellites over mainland and water. Even, during pre-Monsoon season, an aerosol
swirl can be noticed over Indian Ocean, Bangladesh and Pakistan [1].
Sulphur dioxide SO2, nitrogen oxides NOx, ammonia, and volatile organic com-
pounds VOC can be targeted as main precursors of many inorganic aerosols. They
cause natural disasters such as acidification of soil, smog, greenhouse effect and
deterioration of ground-level, protective ozone layer.
In Holland, the two data sets from tracking simulation meteorological models
LOTOS-EUROS and MARGA were exploited to execute the research on emission time
series of PM3,4, PM10 and PM2,5 aerosol particles [2–4]. In Montenegro, in 2010, 513
people died prematurely from particulate matter (PM) and ozone exposure [5].
Moreover, during 2013, 65% of the citizens in urban or urban-industrial cities in
Montenegro was exposed to the pollutant concentrations above the reference level [6].
The plains of these hazardous gases could be smokestacks, petroleum refineries,
landfills, coal power plants, traffic exhaust gases, forest fires, waste incineration and
any other type of combustion.
There is a vast scientific literature addressing the airborne PM10, PM2.5 aerosol
concentrations and concentrations of other airborne pollutants within various countries,
cities and various increments within one city. In the subsequent articles, the PM
concentrations in the regions of Mexico City were analyzed. The reference [7] shows
monitoring campaign of PM particles with aerodynamic diameters less than 2.5 lm
and 10 lm, over 3 suburbs of Mexico City. Conducted gravimetric and chemical
analysis of 330 samples of PM10 and PM2.5 particles, indicated that PM2.5/PM10
average ratios across Xalostoc, La Merced and Pedregal were 0.42, 0.46 and 0.52
respectively. Reference [8] reported about the changeability in occurrence, intensity
and configuration of PM10 time-series on the semi-rural site of the state Colima in
Western Mexico. Similarly in [9], the compiled atmospheric samples of PM2.5 particles
were chemically categorized as dissolved in water ionic species and carbonaceous
aerosols containing 31% of organic carbon and 16% of elemental carbon. In [10] was
used Multilayer Perceptron Neural Network structures to anticipate average emission
dozes of PM10 particles within next 24 h.
There are many articles which witness the presence of PM10 aerosols across the
other parts of the globe. [11, 12] announced the investigations done on PM10 particles,
their inner composition and impact on human health in India and Iran.
Other pollutants, their mass concentrations and environmental impacts were
recorded in articles [13–17]. In this manuscript the certain hazardous gases emissions
of (PM10 aerosol particles, carbon monoxide and nitrogen dioxide) within one urban
settlement were monitored. The available statistical data of these gases emissions
(PM10 aerosol particles, carbon monoxide and nitrogen dioxide) allow to conduct a
comparative study among real time data recorded in a half year period or 7 months
period and values referenced in European air quality index grid. In addition, precisely
can be seen how many times different average values of these emissions were surpassed
and how air pollution within one urban settlement can be ranked. It can be also
estimated how much each of these gases contributed singularly to the air pollution.
Based on these statistical data of gases emissions, the difference among heating and
Exposure to PM10 Aerosol Particles and Other Aerial Pollutants 531
non-heating season was confirmed. USA air quality index was added to show that
index values from both air quality grids can be brought into correlation.
2 Observations and Measurements of Suspended Aerial

Pollutants Over an Urban Settlement in the Podgorica,
Capital City of Montenegro
The monitoring campaign carried out by Montenegrin Environmental Protection

Agency, (MEPA) comprised systematic filter data and apportionment of pollutant
emissions recorded in a consecutive period from January 1st to July 20th (2017) [18–
24]. The capital city of Montenegro (Podgorica) is urban-industrial city where many
urban settlements are susceptible to aerial pollution.
Monitored urban settlement in the capital city of Montenegro, (Podgorica) was sc.
Nova Varos. During the aforementioned air quality control program, the sampling and
data processing were prescribed by MEST EN referent standard methods and norma-
tive (MEST EN 14212, MEST EN 14211, MEST EN 14626, MEST EN 14625,
MEST EN 12341).
For the urban settlement Nova Varos, the only recorded pollutant emission den-
sities from January 1st to July 20th (2017) stemmed from airborne pollutants PM10
aerosol particles. Carbon monoxide (CO) and nitrogen-dioxide (NO2) were recorded in
a half-year period, from January 1st to June 30th. Records of nitrogen-dioxide (NO2)
were laid out only in a tabular display with monthly statistics of measurements. Carbon
monoxide data were tabular and graphically presented. The incidence of recording were
24 h measurements of PM10 aerosol particles, 8 h measurements of carbon-monoxide
and hourly measurements of nitrogen-dioxide. All data were presented in average and
were valid only for the district Nova Varos [18–24].
2.1 Trackings of PM10 Particles and Other Aerial Pollutants

in the In-Situ Spot Nova Varos
The in-situ spot Nova Varos has geographical coordinates, i.e. geographical longitude
of 6603787.37 m and geographical latitude of 4700417.54 m. The elevation of this
in-situ spot is 41 m. The monitored zone Nova Varos is an urban settlement, in the
capital city of Montenegro (Podgorica) which is susceptible to many traffic loads. As it
was said previously, sampling time was continuous and incidence of data integration
(PM10 aerosol particles, CO and NO2) was 24 h, 8 h and one hour [18–24]. As
equipment for processing suspended PM10aerosol particles was used automatic ana-
lyzer with lenses, whilst used method for processing was analysis of beta radiation
absorption. For monitoring 8 h average level of carbon monoxide was also used
automatic analyzer as instrument and IR absorption as method. For collecting the
hourly average records of NO2 was used hemiluminiscence as method [18–23]. In the
Table 1 are presented daily average values of emitted PM10 aerosol particles in the
district Nova Varos fromJanuary 1st to July 20th (2017) with limit value per daily
average values (50 lg/m3) and limit value per annual average values (40 lg/m3). There
Table 1. Daily average values of emitted PM10 aerosols (January 1st–July 20th, 2017) in the
district Nova Varos, Montenegro [18–24]
Date January February March April May June July Limit Limit
(PM10-lg/m3) (PM10-lg/m3) (PM10 m3) (PM10-lg/m3) (PM10-lg/m3) (PM10-lg/m3) (PM10-lg/m3) daily annual
value value
1 151.33 112.09 38.74 34.21 12.36 21.67 32.98 50 40
2 118.53 72.91 47.41 33.25 15.21 23.04 26.95 50 40
3 147.63 53.89 57.37 31.25 17.21 22.55 12.43 50 40
4 36.6 43.81 42.52 19.25 22.36 18.44 17.48 50 40
5 22.07 34.82 36.26 15.28 25.36 16.76 23.93 50 40
6 7.27 34.05 21.64 16.25 19.25 18.51 38.76 50 40
7 8.55 11.2 13.92 15.22 24.25 26.67 26.92 50 40
8 6.52 21.96 14.71 16.25 26.31 18.85 23.65 50 40
9 13.89 32.94 22.58 20.41 24.22 10.31 23.74 50 40
10 32.97 58.83 19.83 30.21 20.36 16.25 20.87 50 40
11 62.18 74.25 12.98 26.58 19.21 12.15 24.37 50 40
12 100.08 77.07 15.66 22.21 31.21 23.88 26.77 50 40
13 120.94 66.12 15.06 25.36 54.25 30.3 23.08 50 40
14 37.92 58.69 19.29 25.36 34.25 28.18 28.19 50 40
15 14.12 104.04 29.54 24.25 22.36 24.47 27.56 50 40
16 66.3 96.43 31.91 21.35 24.25 17.93 35.31 50 40
17 37.97 86.17 36.54 8.84 20.32 11.93 32.53 50 40
18 20.93 56.51 46.76 11.62 22.36 8.17 18.75 50 40
19 32.61 25.59 51.42 21.2 21.64 12.88 18.99 50 40
20 54.44 47.08 55.67 17.45 25.36 21.53 48.55 50 40
21 115.38 61.74 44.12 29.58 18.21 28.28 50 40
22 127.68 69 44.82 17.55 14.25 19.63 50 40
23 222.84 95.4 43.52 15.24 18.36 26.57 50 40
24 135.6 114 49.86 19.25 27.22 26.26 50 40
25 45.45 16.8 45.87 22.36 22.36 23.73 50 40
26 29.13 62.7 29.86 29.11 17.25 20.9 50 40
27 77.5 115.5 13.52 34.56 16.69 24.96 50 40
28 112.33 85.8 26.78 36.62 21.25 29.26 50 40
29 137.78 35.62 36.79 22.36 25.37 50 40
30 127.08 23.07 31.21 22.36 26.53 50 40
31 115.61 27.8 24.21 50 40
was also calculated a number of exceedings of limit daily value (50 lg/m3) for
emission of PM10 aerosol particles, on monthly basis, i.e. for each month during
observed period. For the PM10 aerosol particles emissions, the allowed number of
exceedings of limit daily value (50 lg/m3) is 35 times per year [18–24]. In the Table 2
are presented average values of emitted carbon monoxide (CO) per the intersection of
8 h in the district Nova Varos, from January 1st to June 30th (2017). The limit value for
the intersection of 8 h of carbon monoxide emission is 10 mg/m3. For the carbon
monoxide emissions, it does not exist the tolerance limit for exceeding of limit daily
8 h value (10 mg/m3) [18–23].
Quantity of emitted PM10 aerosol particles was calculated in lg/m3. The percentage
of valid in-situ measurements of PM10 aerosol particles in the district Nova Varos, for
the observed intersection of 7 months was 100%. According to the Table 1., the
number of exceeding of limit daily value for emission of PM10 aerosol particles in
January was 17, February 19, March 3, April 0, May 1, June 0 and July 0. This is
probably due to the fact that furnaces for fossil fuel combustions and coal power plants
are predominantly more engaged during heating season (in this case January and
February). During non-heating season the consummation of fossil fuels, coal, biomass
is a lot less, and consequently that explains the minor or none emission of PM10 aerosol
particles in April, May, June and July. These circumstances could be also explicable by
the proximity of local aluminum combine to researched area and its reinforced pro-
cessing of metal alloys, emissions of heavy metals especially during winter months.
Many chemical processes and electrolysis in the aluminum combine take place par-
ticularly in winter. Therefore, it can be concluded that this district is susceptible to the
pollution caused by emission of PM10 aerosol particles and that there is an overdose of
such aerosol particles in that district.
From the graphics shown below (Fig. 1), it is also perceptible that emissions of
PM10 aerosol particles in the district of Nova Varos, in January and February were the
highest and consequently above thresholds.
μg/m³
Daily average values of PM10 aerosols

250
200
150
100
50
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
days
January (PM10-μg/m³) February (PM10-μg/m³)
March (PM10-μg/m³) April (PM10-μg/m³)
May (PM10-μg/m³) June (PM10-μg/m³)
July (PM10-μg/m³) Limit daily value
Limit annual value
Fig. 1. Daily average values of emitted PM10 aerosols (January 1st–July 20th, 2017) in the
district Nova Varos, Montenegro
The performed in-situ measurements of carbon monoxide (CO) in the district Nova
Varos, were not 100% successfully executed. In April and May, the percentages of
valid 8 h measurements of carbon monoxide were 53.33% and 93.54% respectively.
Limit daily 8 h value for emitted quantity of carbon monoxide in the district Nova
Varos was not overdrawn, i.e. all monthly values of emitted CO were below boundary
value. Emitted carbon monoxide was calculated in mg/m³. Daily average 8 h values of
emitted carbon monoxide from January 1st–June 30th are given on Fig. 2 and Table 2.
The small numerical values of emitted CO support the fact that this gas is present in air
in traces, (rare in nature), except when it is produced by human kind.
mg/m³
Daily average 8-hour values of CO
20
18
16
14
12
10
8
6
4
2
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
days
January (CO-mg/m³)
February (CO-mg/m³)
March (CO-mg/m³)
April (CO-mg/m³)
May (CO-mg/m³)
June (CO-mg/m³)
Limit daily 8-hour value (mg/m³)
Fig. 2. Daily average 8 h values of emitted CO (carbon monoxide) (January 1st–June 30th,
2017) in the district Nova Varos, Montenegro
Table 2. Daily average 8 h values of emitted (CO) carbon monoxide (January 1st–June 30th,
2017) in the district Nova Varos, Montenegro [18–23].
Date January February March April May June Limit daily
(CO-mg/m3) (CO-mg/m3) (CO-mg/m3) (CO-mg/m3) (CO-mg/m3) (CO-mg/m3) 8 h value
(mg/m3)
1 3.51 2.86 0.71 0.46 / 0.26 10
2 2.91 1.69 1.49 0.38 / 0.29 10
3 3.24 1.45 1.68 0.31 0.28 0.26 10
4 1.09 1.68 1.54 0.39 0.24 0.26 10
5 1.1 1.06 1.15 0.4 0.26 0.24 10
6 0.24 0.74 0.46 0.29 0.25 0.21 10
7 0.24 0.53 0.66 0.35 0.23 0.27 10
8 0.22 0.62 0.88 0.48 0.24 0.28 10
9 0.34 0.56 0.27 0.45 0.33 0.23 10
10 0.71 1.34 0.26 0.52 0.34 0.24 10
11 1.7 1.66 0.29 0.49 0.31 0.15 10
12 2.33 1.84 0.35 0.33 0.32 0.24 10
13 2.56 1.8 0.36 0.35 0.28 0.23 10
14 1.67 2.14 0.29 0.32 0.27 0.23 10
15 1.68 2.64 0.37 0.35 0.25 0.23 10
16 1.45 2.92 0.73 0.39 0.22 0.21 10
17 1.54 1.85 0.75 / 0.23 0.21 10
18 0.4 1.77 0.84 / 0.28 0.14 10
19 0.36 1 0.65 / 0.28 0.17 10
20 2.56 1.44 0.84 / 0.25 0.22 10
21 3.39 2.03 0.94 / 0.21 0.23 10
22 3.75 2.24 0.54 / 0.23 0.26 10
23 4.31 2.04 0.61 / 0.22 0.26 10
24 3.19 1.92 0.59 / 0.23 0.26 10
25 3.39 1.18 0.62 / 0.25 0.24 10
26 0.52 1.48 0.28 / 0.17 0.22 10
27 2.38 1.8 0.23 / 0.16 0.22 10
28 2.59 1.89 0.59 / 0.18 0.21 10
29 3.22 / 0.69 / 0.22 0.2 10
30 3.59 / 0.47 / 0.22 0.24 10
31 2.6 / 0.43 / 0.22 / 10
In the Table 3 are presented minimum and maximum hourly values of emitted
nitrogen dioxide (NO2) in the district Nova Varos, from January 1st to June 30th
(2017). The limit hourly value of nitrogen dioxide emission is 200 lg/m3. Limit hourly
value must not be exceeded over 18 times annually. Limit annual value of nitrogen
dioxide emission is 40 lg/m3.
Table 3. Minimum and maximum hourly values of emitted NO2 (nitrogen dioxide) (January
1st–June 30th, 2017) in the district Nova Varos, Montenegro [18–23].
NO2 January February March April May June
Percentage of 100% 100% 99.86% 52.36% 95.43% 100%
valid hourly
average values
Minimum hourly 0.38 lg/m3 0.3 lg/m3 0.38 lg/m3 0.74 lg/m3 0.25 lg/m3 0.28 lg/m3
value
Maximum hourly 145 lg/m3 143.58 lg/m3 110.7 lg/m3 78.65 lg/m3 68.21 lg/m3 39.91 lg/m3
value
The number of 0 0 0 0 0 0
exceeding limit
hourly value
Emitted nitrogen dioxide was calculated in lg/m3. The percentages of valid hourly
measurements of nitrogen dioxide (NO2) in March, April and May were 99.86%,
52.36% and 95.43% respectively. According to the given statistics of monitored
nitrogen dioxide, all recorded maximum hourly values per each month are below the
limit hourly value of nitrogen dioxide which is 200 lg/m3. All maximum hourly values
of emitted nitrogen dioxide are at least one to three times bigger than limit annual value
of nitrogen dioxide emission, which is 40 lg/m3. It can be observed that maximum
hourly values of emitted nitrogen dioxide, on monthly basis, have a downstream course
from January to June. The highest hourly emitted concentrations of nitrogen dioxide
were recorded in January and February.
2.2 Montenegrinair Quality Index vs. American Air Quality Index

Air quality index (AQI) is a tool which serves as a benchmark of airborne pollution.
Many countries assign air quality index to most of the airborne pollutants. Air quality
index is grouped into air quality categories. This index generally demonstrates the
instant exposure to hazardous gases such as: sulphur dioxide (SO2), carbon monoxide
(CO), nitrogen dioxide (NO2), PM particulate matters (PM2.5, PM10) and ozone (O3)
(Air quality in Europe 2012).
European air quality index, which is also valid for the state of Montenegro, can be
divided into 3 air quality sub-indexes: hourly index, daily index and annual index. The
European air quality index is in compliance with the European directives and norms.
European air quality index legend has 5 air quality categories which are shown in
Table 4 (Air quality in Europe 2012). In Table 5 are presented air quality indexes for a
roadside and urban background in Europe, which are again valid for the state of
Montenegro. Within European air quality index grid, roadside index represents traffic
zones burdened by traffic loads, i.e. roadside monitoring sites. Shown background
index represents other urban entities as monitoring sites. In accordance with European
air quality index grid, the urban settlement subjected to traffic loads sc. Nova Varos was
analyzed. European air quality index does not have specifically developed calculation
pattern such as the case of USA.
Table 4. European air quality index legend (Air quality in Europe 2012)
Pollution Index value

Very low 0/25
Low 25/5
Medium 50/7
High 75/100
Very high >100
Table 5. European air quality index grid (Air quality in Europe 2012)
• NO2, SO2, O3- hourly values/maximum hourly values in lg/m3

• PM10, PM2.5- hourly values/ maximum hourly values or daily average values in
lg/m3
• CO- 8 h average value in lg/m3
Since the monitored urban settlement Nova Varos is subjected to traffic loads, the
roadside index of European air quality index grid can be assigned to it. Comparing the
specifications referenced for mandatory pollutants PM10 aerosols in the European index
grid and emissions of PM10 aerosol particles in the district Nova Varos, on daily basis
(per 24 h) can be concluded that:
• In January, the emission of PM10 aerosol particles in district Nova Varos fluctuates
from very low, to low and very high index values which correspond to very low,
low and very high caused aerial pollution according to the European scheme.
• In February, dominates the medium index value for the emission of PM10 aerosol
particles in district Nova Varos according to the European air quality index grid
which causes the most widespread medium level of aerial pollution. Also the very
high and high rates of aerial pollution are present.
• In March, according to the European air quality index grid, there are no index
values which correspond to very high and high rate of aerial pollution in the
mentioned urban settlement. Observed rates of pollution correspond to very low and
low aerial pollution.
• In April, dominate very low and low rates of aerial pollution in the aforementioned
urban settlement.
• In May, highly dominates the very low rate of aerial pollution in the urban set-
tlement Nova Varos.
• In June, also highly dominates the very low rate of aerial pollution in the urban
settlement.
• In July, the both low rate and very low rate of aerial pollution are present in the
urban settlement.
This numerical assortment of PM10 aerosol particles emission values (in observed
7 months) according to the European air quality index valorization, matches with the
previous conclusion that the greatest rates of PM10 aerosol particles emissions happen
during heating season (in this case January and February) and lowest rates of PM10
aerosol particles emissions during non-heating season.
As for the European air quality index valorization of carbon monoxide emission
values during observed 6 months, carbon monoxide emission values can be marked as
the ones that cause slight aerial pollution or the very low rate of aerial pollution. That
drags the conclusion that carbon monoxide is present in normal dozes in the air,
surrounding this district.
The pollution caused by nitrogen dioxide in the mentioned district can be ranked as
in range from very low to medium regarding maximum hourly emission values of
nitrogen dioxide.
American air quality index has a unique calculation pattern and therefore specifi-
cally developed calculator which converts pollutants’ aerial dozes into coefficient of air
quality (air quality index). It is classified into 6 groups indicating different intensity of
impact on human health and different colour of air. The computation platform for the
USA coefficient of air quality (air quality index) is shown as (USA air quality index
2011):
Ihigh Ilow
I¼ ðC Clow Þ þ Ilow ð1Þ
Chigh Clow
I- coefficient of air quality (air quality index);

C- the pollutant concentration;
Clow - the concentration breakpoint that is C;
Chigh - the concentration breakpoint that is C;
Ilow - the index breakpoint corresponding to Clow ;
Ihigh - the index breakpoint corresponding to Chigh .
Any pollutant concentration from European air quality index grid can be converted
into pollutant concentration valid for USA air quality index grid. It implies calculated
american air quality index and calculated levels of health concern which brings these
two indexes in correlation, as well as many others in correlation.
The USA air quality index grid is shown in Table 6.
Table 6. American air quality index legend (USA air quality index 2011)
3 Conclusion
In this manuscript the statistical data processing of certain hazardous gases’ emission
(PM10 aerosol particles, carbon monoxide and nitrogen dioxide) within one urban
settlement (Nova Varos) was exposed. Under the monitoring campaign of that urban
settlement, the metering of present airborne pollution quantity was done.
The following conclusions can be drawn from the article:
• By statistics of recorded emissions and European air quality index grid and val-
orization, the overdose of PM10 aerosol particles in the considered district (Nova
Varos) during heating season (especially January and February) was confirmed.
• By statistics of recorded emissions and European air quality index grid and val-
orization, the much lower dose of PM10 aerosol particles in the monitored district
(Nova Varos) during non-heating season was observed.
• Generally, the tendency and susceptibility of this district to the aerial contamination
and exceeded emissions of PM10 aerosol particles can be stated.
• Certain humankind sources of PM10 aerosol particles can be concluded and found.
• Carbon monoxide emissions in this district cause a slight and mild aerial pollution
which can be categorized as very low pollution according to the European air
quality index grid and valorization.
• Human activities are not the only activities and triggers that cause exceeded
emissions of carbon monoxide.
• Nitrogen dioxide (NO2) also contributes to aerial contamination of this district
particularly during heating season.
• PM10 aerosols and nitrogen dioxide (NO2) can be assorted among main aerial
pollutants of this district.
• Air quality index grids which exist worldwide help to determine the level of aerial
pollution, kind of an aerial pollution and aerial pollutants concentrations. They help
in finding a solution for aerial pollutants’ remediation and they need to be always
improved regarding air quality regulations.
References
1. Aerosols: TinyParticles, Big Impact: Feature Articles - NASA Earth Observatory (2010).
https://www.earthobservatory.nasa.gov/Features/Aerosols/
2. Weijers, E.P., Sahan, E., Ten Brink, H.M., Schaap, M., Matthijsen, J., Otjes, R.P., Van
Arkel, F.: Contribution of secondary inorganic aerosols to PM10 and PM2.5 in the
Netherlands; measurements and modelling results. Netherlands Research Program on
Particulate Matter Report 500099006 (2010). http://www.pbl.nl/en
3. European Commission - National Emission Ceilings Directive (2016). http://www.ec.europa.
eu/environment/air/pollutants/ceilings.htm
4. European Commission - New Air Quality Directive (2008). http://www.ec.europa.eu/
environment/air/quality/legislation/directive.htm
5. EEA: Air quality in Europe report (2014). https://www.eea.europa.eu/publications/air-
quality-in-europe-2014
6. Montenegrin Environmental Protection Agency-MEPA: Indicator-based State of the
Environment Report of Montenegro (2013)
7. Vega, E., Reyes, E., Ruiz, H., Garcia, J., Sanchez, G., Martinez-Villa, G., Gonzalez, U.,
Chow, J.C., Watson, J.G.: Analysis of PM2.5 and PM10 in the atmosphere of Mexico City
during 2000–2002. Air Waste Manage. Assoc. 54, 786–798 (2004)
8. Campos-Ramos, A., Aragon-Pina, A., Alastuey, A., Galindo-Estrada, I., Querol, X.: Levels,
composition and source apportionment of rural background PM10 in western Mexico (state
of Colima). Atmos. Poll. Res. 2, 409–417 (2011)
9. Vega, E., Ruiz, H., Escalona, S., Cervantes, A., Lopez-Veneroni, D., Gonzalez-Avalos, E.,
Sanchez-Reyna, G.: Chemical composition of fine particles in Mexico City during 2003–
2004. Atmos. Poll. Res. 2, 477–483 (2011)
10. Cortina-Januchs, M.G., Quintanilla-Dominquez, J., Vega-Corona, A., Andina, D.: Devel-
opment of a model for forecasting of PM10 concentrations in Salamanca, Mexico. Atmos.
Poll. Res. 6, 626–634 (2015)
11. Chandra, S., Kulshrestha, M.J., Singh, R.: Temporal variation and concentration weighted
trajectory analysis of lead in PM10 aerosols at a site in Central Delhi, India. Int. J. Atmos.
Sci.-Hindawi Publishing Corporation (2014). https://doi.org/10.1155/2014/323040
12. Bagherian Marzouni, M., Alizadeh, T., Rezaei Banafsheh, M., Khorshiddoust, A.M.,
Ghanbari Ghozikali, M., Akbaripoor, S., Sharifi, R., Goudarzi, G.: A comparison of health
impacts assessment for PM10 during two successive years in the ambient air of Kermanshah,
Iran. Atmos. Poll. Res. 7, 768–774 (2016)
13. Fazlzadeh, M., Rostami, R., Hazrati, S., Rastgu, A.: Concentrations of carbon monoxide in
indoor and outdoor air of Ghalyun cafes. Atmos Poll. Res. 6, 550–555 (2015)
14. Mohammed, M.O.A., Song, W., Liu, L., Ma, W., Li, Y., Wang, F., Ibrahim, M.A.E.M., Qi,
M., Elzaki, A.A., Lv, N.: Distribution patterns and characterization of outdoor fine and
coarse particles. Atmos. Poll. Res. 7, 903–914 (2016)
15. Tyagi, S., Tiwari, S., Mishra, A., Hopke, K.P., Attri, D.S., Srivastava, K.A., Bisht, S.D.:
Spatial variability of concentrations of gaseous pollutants across the National Capital Region
of Delhi, India. Atmos. Poll. Res. 7, 808–816 (2016)
16. Guerranti, C., Benetti, F., Cucciniello, R., Damiani, D., Perra, G., Proto, A., Rossi, F.,
Marchettini, N.: Pollutants monitoring and air quality evaluation in a confined environment:
the ‘Majesty’ of Ambrogio Lorenzetti in the St. Augustine Church in Siena (Italy). Atmos.
Poll. Res. 7, 754–761 (2016)
17. Khanum, F., Nawaz Chaudhry, M., Kumar, P.: Characterization of five-year observation data
of fine particulate matter in the metropolitan area of Lahore. Air Qual. Atmos. Health 10,
725–736 (2017)
18. Montenegrin Environmental Protection Agency: Monthly air quality report of Podgorica
(Nova Varos), January 2017. http://www.epa.org.me/index.php/o-nama/87-azzs/171-
vazduh-kvalitet-vazduha
(Nova Varos), February 2017. http://www.epa.org.me/index.php/o-nama/87-azzs/171-
vazduh-kvalitet-vazduha
(Nova Varos), March 2017. http://www.epa.org.me/index.php/o-nama/87-azzs/171-vazduh-
kvalitet-vazduha
(Nova Varos), April 2017. http://www.epa.org.me/index.php/o-nama/87-azzs/171-vazduh-
kvalitet-vazduha
(Nova Varos), May 2017. http://www.epa.org.me/index.php/o-nama/87-azzs/171-vazduh-
kvalitet-vazduha
(Nova Varos), June 2017. http://www.epa.org.me/index.php/o-nama/87-azzs/171-vazduh-
kvalitet-vazduha
(Nova Varos), July 2017. http://www.epa.org.me/index.php/o-nama/87-azzs/171-vazduh-
kvalitet-vazduha
25. Air quality indices definition: Air quality in Europe (2012). http://www.airqualitynow.eu/
about_indices_definition.php
26. USA Environmental Protection Agency: Air Quality Index (AQI) Basics (2011). https://
www.airnow.gov/index.cfm?action=aqibasics.aqi
New Technological Procedures for Production
of Thioncarbamates as a Selective Flotation
Reagents
Milutin M. Milosavljević1(&), Milan M. Milosavljević1,

Milutin Živković2, and Ljiljana Pecić2
1
University of Priština, 38220 Kosovska Mitrovica, Serbia
milutin.milosavljevic@pr.ac.rs, milan3386@hotmail.rs
2
Bachelor School for Professional Technical Studies, 37240 Trstenik, Serbia
milutinzivkovic@gmail.com,
ljiljana.pecic@vtmsts.edu.rs
Abstract. In this paper an innovative technological procedures for the pro-

duction of carbamates (selective flotation reagents) was presented, as well as,
some advantages over the classical, known process of production. A techno-
logical procedures for the synthesis of thiocarbamate starts with reaction
between alkylxanthogens and amines in the presence of nano-palladium mul-
tiwall carbon nanotube catalysts, and continues with oxidation of the xan-
thogenic acid amine salt and catalytic reaction from isobutanol and amine.
Obtained results are compared with classical ammonolysis in terms of yield and
quality of the obtained products. New technological procedures for the pro-
duction of carbamates exhibit comparative significant advantages such as mild
reaction conditions, higher conversion rate and ecologically justified without the
separation of the product.
Keywords: Thioncarbamates Xanthogenate Xanthogen-acetic acid

Nano-palladium
1 Introduction
Thiocarbamates basically represent thiocarbamic acid derivatives [1] and thio- and thiol
esters. Direct association of thioacyl groups and nitrogen provides a various biological
activity of thioncarbamates [2], so it can be used as fungicides [3, 4], bactericides [3,
5], herbicides [6, 7], germicides [8], pesticides [9–11], insecticides [12, 13], etc. Also,
O-alkylthiocarbamates can be used as polymerization accelerators and selective fleet
reagents [14]. There are a great number of known methods for alkylthiocarbamate
synthesis by ammonolysis of alkaline salts of xanthogenic acid and various catalytic
processes. Synthesis of alkylthiocarbamate can be achieved with reaction between
xanthogens, amines and various catalysts: mangan (II) acetate and nickel zeolite cat-
alyst [15].
Thiocarbamates can also be obtained by the reaction of xanthates and amines in the
presence of nickel (II) sulfate heptahydrate as a catalyst [16], while high yield of

https://doi.org/10.1007/978-3-319-90893-9_63
New Technological Procedures for Production of Thioncarbamates 543
isopropylthiocarbamate can be obtained by the oxidation of xanthogenic acid amines

with hydrogen peroxide or sodium hypochlorite [14].
N-phenylthiocarbamates can be obtained from phenylchloro formation already
gained by the reaction of thiophosgene and phenol (the corresponding amine in dioxane
as a solvent). This method can not be applied for the synthesis of N-aryl-O-
alkylthiocarbamate [17]. Catalyzed reaction of sodium alkylxanthogens and primary
and secondary aliphatic amines with Raney Nickel catalyst gives alkylthiocarbamates
and sodium hydrogensulfide as a by-product., The reaction was carried out at 80 °C for
16 h [18].
N-aryl-O-alkyl thiocarbamates are obtained by addition of alkyl alcohol to isoth-
iocyanates. Increasing the reaction rate is achieved by the addition of catalytic amounts
of triethylamine, diethylamine or pyridine [19]. N-phenyl-O-phenylcarbamate can be
obtained from aniline and phenyl chlorothioformate (phenyl esters of chlortionic acid)
in the presence of carbonate solvents and recrystallization from toluene [20], while
N-aryl-O-alkylthiocarbamate can be made in a solution of methylsulfoxide or
dimethylformamide starting from alkoxide and p-chlorophenyl isothiocyanate, and
recrystallization in hexane [21]. Synthesis of S-aryl (alkyl) thiocarbamates can be
carried out through the reaction of isocyanates and disulfides in the presence of
Zn/AlCl3 as a catalyst [22], as well as, starting from thiols of thiol and isocyanates
using a DBU catalyst in methylene chloride as a solvent [23, 24]. In this paper, an
innovative technological procedure for the production of carbamates (selective flotation
reagents) was presented, as well as, some its advantages over the classical, known
process of production.
2 Technological Procedures for the Synthesis

of Thiocarbamate
2.1 A Classic Procedure for the Synthesis of Thiocarbamate

with Xanthogenic Acetate Amonolysis
This synthesis was carried out through the amonolysis of sodium isobutylxan-
thogenacetate (NaiBXAc) using primary or secondary alkylamines. NaiBXAc was
formed by the reaction of potassium isobutylxantate (KiBX) and sodium
monohloroacetate (NaClAc), obtained in neutralization of monochloroacetic acid and
sodium carbonate. Sodium thioglycolate, obtained as a by-product, was converted to
thioglycolic acid (TGK) by acidification with hydrochloric acid, and finally extracted
with dipropylether and packaged as a commercial product.
S S
RR'NH HSCH2COONa ð1Þ
iBuO SCH2COONa iBuO NRR'
0 0
where: R and R : R = alkyl, isoalkyl or cycloalkyl and R = H in reaction with
0
monoalkyl amines, R = alkyl and R = alkyl in reaction with dialkyl amines.
544 M. M. Milosavljević et al.
The scheme of the technological procedure for the production of isobutylcarbonate

at the industrial plant is shown in Fig. 1.
Fig. 1. Scheme of the technological procedure for the production of isobutyl-thiocarbamate on

industrial plant
2.2 Synthesis of N-alkyl- and N, N-dialkyl-O-ethyl-

and O-isopropylthiocarbamate by Oxidation of the Amine Salt
of Xanthogenic Acid
Process was optimized for N-alkyl-; N, N-dialkyl-O-ethyl- and O-isopropylthiocar
bamate through the oxidation of the amine salts of ethyl and isopropylxanthogenic
acids using sodium hypochlorite and hydrogen peroxide. Alkylamine salts of ethyl and
isopropylsanthoic acids were obtained by reaction between alkylammonium sulfate and
sodium ethyl xanthate or sodium isopropylxanthate. The reactions are carried out
according to the following reaction scheme shown in Fig. 2, and the yield and purity of
the synthesized N-alkyl- and N, N-dialkyl-O-isopropylthiocarbamate are presented in
Table 1. Innovative process for the production of liquid xanthates provides the pos-
sibility of using obtained xanthate in the form of aqueous solutions in this synthesis.
Fig. 2. Synthesis of N-ethyl-O-ethylthiocarbamate in the presence of sodium hypochlorite and

hydrogen peroxide as oxidizing agents.
Table 1. Yield and purity of synthesized N-alkyl- and N, N-dialkyl-O isopropylthiocarbamate

Compound Reaction time (h) Yield (%) Boiling GC purity
temperature (oC/Pa) (%)
H2O2 NaOCl H2O2 NaOCl
iPrOC(S)NHEt 1.5 97.0 96.8 105-6/660 99.0 98.9
iPrOC(S)NEt2 2.0 93.0 90.2 107-8/660 98.7 98.0
iPrOC(S)NHnPr 1.5 95.0 91.0 119-0/760 97.6 97.9
iPrOC(S)N(nPr)2 2.0 90.1 88.9 120-1/790 99.5 99.6
iPrOC(S)NHiPr 1.5 94.3 90.2 114-5/790 99.7 99.6
iPrOC(S)N(iPr)2 2.5 88.2 80.1 126-7/790 99.2 99.3
2.3 Synthesis of N-alkyl- and N, N-dialkyl-

and N-cycloalkyl-O-isobutylthiocarbamate by a Catalytic Reaction
from Isobutanol and Amine
The synthesis of N-alkyl, N, N-dialkyl- and N-cycloalkyl-O-isobutylthiocarbamate can
be carried out using a 1,5-diazabicyclo [5.4.0] undec-5-ene (DBU) catalyst. Thus,
starting from isobutanol, amine and bis (benzotriazolyl) methanthione in the presence
of a DBU catalyst, a series of products of N-alkyl-, N, N-dialkyl- and N-cycloalkyl-O-
isobutylthiocarbamate were synthesized, Fig. 3.
N
R
N
DBU
OH S N O N
H
R R'
N R'
S
N
Fig. 3. Synthesis of N-alkyl, N, N-dialkyl- and N-cycloalkyl-O-isobutyltinecarbamate from

isobutanol, amine and bis(benzotriazolyl)methanthione in the presence of a DBU catalyst
2.4 Synthesis of Thiocarbamate from Xanthogens and Aniline

in the Presence of Hybrid Catalyst Nano-platinum of Multilayer
Carbon Nanotubes (Pt/MWCNT)
The new process for the synthesis of N-(3-and 4-substituted-phenyl)-O-isobutylcarba
mates from O-isobutyl xanthate and 3- and 4-substituted anilines was made using
synthetic hybrid catalysts of nanoparticles of multilayer carbon nanotubes
(Pt/MWCNT).
Synthesized catalysts were obtained by precipitation of platinum nanoparticles on
aminodified MWCNT. Amino-functionalized MWCNTs were prepared using two meth-
ods. The first procedure relates to the oxidation of MWCNTs and transformation of the
present carboxyl groups by direct coupling of 1,4-phenylenediamine to introduce a terminal
functional group on the phenyl nucleus. Pre-activation with N, N-diisopropylethylamine
(DIEA) and the subsequent construction of the ester with O-(7-azabenzotriazol-1-yl)-N,
0 0
N,N ,N -tetramethyluronium hexafluorophosphate (N-HATU) allows the formation of
an amide linkage mild reaction conditions (Fig. 4).
General deposition of nano-platinum particles onto the surface of amino-modified
MWCNTs was performed in ethylene glycol solution and the addition of H2PtCl6
precursors using microwave treatment to provide a homogeneous dispersion of
NH2Ph-MWCNTs and reduction of nano-Pt particles on the surface. The preparation of
the Pt/MWCNTs preparation is given in Fig. 5, whereby a Pt/NH2PDA-MWCNT
catalyst is obtained.
The second procedure was performed in situ by diazotyping with 4-nitroaniline,
resulting in the direct introduction of the diazo component onto the surface of
Fig. 4. Schematic representation of MWCNT functionalization
NO2 NH2 NH 3 Pt
ONO LiAlH 4 , THF EG, H2 PtCl6

NO2 NH2 NH 3 Pt
stir bar, 60o C sonication, 1 h microwave, 60 s
NO2 NH2 NH 3 Pt
raw-MWCNT NO 2Ph-MWCNT NH 2 Ph-MWCNT Pt/NH 2Ph-MWCNT
Fig. 5. Scheme of the preparation of the Pt/MWCNTs catalyst: (a) diazotizing MWCNTs;
(b) reduction of NO2Ph-MWCNTs; (c) precipitation of nano-Pt under the effect of microwave
radiation
MWCNTs; by reduction of the terminal nitro group with lithium aluminum hydride to
give the amino terminal group on the phenyl nucleus.
The synthesis of the nano-Pt/MWCNT catalyst according to the methods presented
here substantially influences their activity during the synthesis of N-(3- and
4-substituted phenyl)-O-isobutylcarbamate. It has been shown that the catalyst obtained
by diazotism contributed to the production of products with higher yields and greater
purity. In addition, a commercial Pt/activated carbon catalyst was used, and the syn-
thesis results were compared to compare the efficiency of the synthesized catalyst with
respect to the commercial (Method C, Table 2).
The characterization of the obtained catalysts was performed, and the results of the
FTIR (Fig. 6) and the HRTEM analysis of the precursors and the resulting catalysts
were presented (Fig. 7).
The results of the synthesis obtained using the catalyst were compared with the
results obtained by the reactionary classical method (Method A). According to this
procedure, the reaction of the thiocarbamate synthesis was performed using sodium
isobutylxanthate and the 3- and 4-substituted anilines, shown in the following scheme:
NH 2
X
S
S
+ + HSCH2COONa
iBu O SCH2COONa iBu O N
H
where X: H, 4-OCH3, 4-OH, 4-CH3, 4-I, 4-F, 4-Br, 4-Cl, 3-Cl, 3-Br, 3-OCH3 и 4-N
(CH3)2 substituent. The results obtained in the synthesis of N-(3- and 4-substituted
Table 2. The yield and purity of N-(3- and 4-substituted phenyl)-O-isobutylthioncarbamates

obtained according to methods A-C
Compound Substituent Method A Method B Method C
Yield Puritya Yield Puritya Yield Puritya
(%) (%) (%) (%) (%) (%)
1 H 32.1 99.2 40.1 99.3 38.2 99.4
2 4-OCHb3 45.9 99.0 63.1 99.2 50.1 98.9
3 4-OH 44.5 99.0 61.0 99.5 48.2 98.8
4 4-CHb3 30.3 98.7 48.9 99.6 36.1 99.1
5 4-Ib 52.1 98.7 72.3 99.4 59.3 99.4
6 4-Fb 18.3 98.7 34.0 99.3 20.8 99.5
7 4-Brb 51.4 98.7 74.5 99.7 61.8 98.7
8 4-Cl 40.6 98.7 56.2 99.7 43.3 99.6
9 3-Cl 18.8 98.8 34.1 99.8 21.3 99.3
10 3-Brb 44.2 98.8 69.1 99.2 48.6 98.8
11 3-OCHb3 41.3 99.2 60.2 98.9 44.8 99.7
12 4-N,N 67.6 99.0 80.9 99.3 70.1 99.6
(CH3)b2
a
a Determined using the GC method; b of compounds 2–7, 10, 11 and 12 are novel compounds
Fig. 6. FTIR spectra (a) of untreated MWCNT, (b) NO2Ph-MWCNT and (c) NH2Ph-MWCNT
modified
phenyl)-O-isobutylthioncarbamate were carried out with the results of the synthesis of

the same compounds in the presence of Pt/NH2PhMWCNT Pt/NH2PhMWCNT
(Method B) and Pt/NH2PDA-MWCNT (Method C) of the catalyst. The synthesis
results obtained using the A-C method are given in Table 2.
Fig. 7. HRTEM analysis of the precursors and the resulting catalysts: (a) and
(b) NH2Ph-MWCNT. (c) diffraction of sample electron with Pt nanoparticles (d) HAADF
figure of Pt/NH2PhMWCNT catalyst
2.5 Synthesis of N-alkyl- and N, N-dialkyl-

and N-cycloalkyl-O-isobutylthiocarbamate by Reaction
of the Oxidation of the Amine Salt Xanthogenic Acid with Potassium
Peroxodisulphate
Figure 8, presents the synthesis of N-alkyl-, N, N-dialkyl- and N-cycloalkyl-O-
isobutylthioncarbamate starting from KiBX and amine in the presence of an oxidation
agent: potassium peroxodisulfate.
S S
2 K2S2O8 2 R' K2SO3

2 N K2SO4 H2O 2S
O SK H O N
R
R'
R
Fig. 8. Synthesis of N-alkyl-, N, N-dialkyl- and N-cycloalkyl-O-isobutylthioncarbamate from

potassium-isobutylxanthate and amine in the presence of potassium peroxodisulfate
The results of the synthesis of N-alkyl-, N-, N-dialkyl- and N-cycloalkyl-O-

isobutylthioncarbamate using three comparative oxidation agents: potassium perox-
odisulfate, hydrogen peroxide and sodium hypochlorite are presented in Table 3.
Obtained yields (Table 3), suggest potassium peroxodisulphate to be the most
effective oxidation agent., The reaction mechanism for the synthesis of thiocarbamate
from KiBuX and amine in the presence of the oxidation agent consists of two steps:
(1) oxidation of xantogenate in dixantogenate - fast process and (2) steps determining
the speed of the overall reaction of the system, i.e. nucleophilic heterolyses of per-
sulfide bonds, when a product is obtained from carbamate. It was found that lower
yields in the presence of sodium hypochlorite are obtained in the synthesis of N-alkyl-
and N, N-dialkyl-O-ethyl- and O-isopropyl thiocarbamates, which is the case in the
synthesis of N-alkyl, N, N-dialkyl and N-cycloalkyl-O-isobutylthiocarbamate. It has
been found that acrylic chlorine produces chloramine derivatives that can not carry out
Table 3. Yields and purity of N-alkyl, N, N-dialkyl and N-cycloalkyl-O-isobutylthioncarbamate

Compound Yield (%) GC purity (%)
H2O2 NaOCl K2S2O8 H2O2 NaOCl K2S2O8
1 iBuOC(S)NHEt 87.2 85.1 90.3 98.5 98.3 98.6
2 iBuOC(S)NHPr 88.6 86.5 91.7 97.3 97.5 97.2
3 iBuOC(S)NHBu 89.5 86.8 92.6 99.0 99.1 99.2
4 iBuOC(S)NHs-Bu 89.3 87.2 92.4 99.4 99.3 99.1
5 iBuOC(S)NHi-Pr 87.1 85.0 90.2 99.2 99.0 99.2
6 iBuOC(S)NHi-Bu 93.6 92.5 96.7 97.0 97.2 97.8
7 iBuOC(S)NHi-Pent 94.1 92.0 97.3 97.5 97.7 97.9
8 iBuOC(S)NHc-Pr 77.4 75.6 80.2 97.8 98.0 98.6
9 iBuOC(S)NHc-Pen 83.6 82.4 86.7 98.0 98.1 98.5
10 iBuOC(S)NHc-Heks 87.7 85.6 90.8 97.0 97.2 97.9
11 iBuOC(S)N(Et)2 93.8 91.7 96.5 97.4 97.6 97.5
12 iBuOC(S)N(Pr)2 95.2 94.0 98.0 98.0 98.1 98.8
13 iBuOC(S)N(Bu)2 95.9 95.1 98.2 97.0 97.2 97.9
the nucleophilic heterolysis of the disulfide bonds of the disintegrants, which affects the
reduction of the yield of thiocarbamate.
3 Conclusion
In this paper, an innovative technological procedure for the production of carbamates

(selective flotation reagents) was presented, as well as, some its advantages over the
classical, known process of production. By innovated synthesis procedures, thiocar-
bamates can be obtained: from alkylxanthate and amines in the presence of
nano-palladium multiwall carbon nanotube catalyst, oxidation of xanthogenic acid
amine, catalytic reaction from isobutanol and amine are shown. Comparative advan-
tages over the classical production method are given, which consists in the
ammonolysis of the alkyxanthene acetic acid sodium salt with the corresponding alkyl
and arylamines, and the obtained results are compared in terms of yield and quality of
the obtained products. New innovative methods have comparative advantages over the
classical, such as a higher degree of conversion of raw materials into products, mild
reaction conditions, non-derivation of the product. All of these methods are experi-
mentally, technically and technologically acceptable for the synthesis of carbamates at
the semi-industrial level, whereby the tendency is to achieve procedures that have
positive techno-economic indicators and satisfy the principles of green technologies.
Acknowledgements. This work was supported by the Ministry of Education, Science and
Technological Development of Serbia (Project Numbers 43007).
References
1. Walter, W., Bode, K.D.: Syntheses of thiocarbamates. Angew. Chem. 79, 281–293 (1967)
2. Glazsrin, A.B., Denisov, A.N., Talzi, V.P., Savin, B.P.: Tiokarbamats, Prommiilennostm po
proizvodstvu mineralmnih udobreniy, seriz, pp. 1–32. Nauchno-isledovalteski institut
Tehniko-ekonomicheskii isledovanii, Moskva, Himicheskie sredstva zaoitmi rasteniy (1988)
3. Chisholm, M.H., Extine, M.W.: Reactions of transition metal-nitrogen r bonds. 3. early
transition metal N, N-dimethylcarbamates. Preparation, properties, and carbon dioxide
exchange reactions. J. Am. Chem. Soc. 99, 782–792 (1977)
4. Savin, V.P., Tolzi, V.P., Bek, N.O.: Org. Hbv. 20, 1984 (1842)
5. Ewing, S.P., Lochson, D., Kencks, W.P.: J. Am. Chem. Soc. 102, 3072 (1980)
6. Milosavljević, M.M., Ražić, S.: Vaprosy himii i himicheskoi tehnologii 3, 50 (2005)
7. Glazsrin, A.B., Denisov, E.N., Talzi, V.P., Imaiev, U.B.: Ghbrk. Hbv. 58(10), 2300 (1985)
8. Kunihiko, F., Kuniak, S.: US patent 4101670 (1978)
9. Rinehart, J.K.: US patent 4059609 (1977)
10. Rinehart, J.K.: US patent 4055656 (1977)
11. Ruminski, P.G.: US patent 5457134 (1995)
12. Fumio, K.: JP patent 58105985 (1983)
13. Kisida, H., Hatakoshi, M.: US patent 4486449 (1984)
14. Milisavljević, S.S., Marinković, A.D., Milosavljević, M.M.: Novi postupak sinteze N-alkil-
iN,N-dialkil-O-etil-.i O-izopropiltiokarbamata oksidacijom aminskih soli ksantogene kise-
line. Hem. Ind. 64(5), 401–409 (2010)
15. Milosavljevic, M., Sovrlić, M., Marinkovic, A.D., Milenković, D.D.: Monats. Chem. 141,
749 (2010)
16. Jonhson, G., Rafferty, M.F.: Amide, sulfonamide, urea, carbamate, thiocarbamate, and
thiourea derivatives of 4-hydroxybenzylamine having anti-inflammatory and analgesic
activity, U.S. Patent 4,980,366 (1990)
17. Albors-Velaco, M., Thorne, J., Wain, R.L.: Agnc. Food. Chem. 43, 2260–2261 (1995)
18. Bishop, M.D., Gray, L.A.: US 5041599 (1991)
19. Reuerin, M.Y.: JP 59055867 (1984)
20. Keun Oh, H., Young Oh, J.: Bull. Korean Chem. Soc. 27(1) (2006)
21. Bauman, R.A.: J. Chem. Eng. Data 11(2), 274–276 (1966)
22. Movassagh, B., Zakinezhad, Y.: A new one-pot synthesis of thiocarbamates from
isocyanates and disulfides in the presence of Zn/AlCl3 System. Chem. Lett. 34, 1330–
1331 (2005)
23. Movassagh, B., Soleiman-Beigi, M.: Monatsh. für Chem. 139, 137–140 (2008)
24. Milosavljević, M.M., Vukićević, I.M., Drmanić, S.Ž., Nikolić, J.B., Marinković, A.D.,
Krstić, S.S., Petrović, S.D.: Simple one-pot synthesis of thioureas from amine, carbon
disulfide and oxidants in water. J. Serb. Chem. Soc. 80, 1–14 (2015). https://doi.org/10.
2298/JSC150831087M
The Content of Heavy Metals in “PET” Bottles
of Drinking Water and Its Electrical Conductivity
Ekrem Pehlić1 ✉ , Aida Šapčanin2, Husein Nanić1, and Adnan Ćehajić1

( )
1
pehlic_ekrem@yahoo.com, {ida,nanic.h,adnan.cehajic}@bih.net.ba
2
Faculty of Pharmacy, University of Sarajevo, Sarajevo, Bosnia and Herzegovina
Abstract. Heavy metals are natural ingredients of the Earth’s crust. Due to their
persistence, high toxicity and the tendency to accumulate in the ecosystem, heavy
metals pose a threat to human health. The aim of the paper is to examine the
content of heavy metals in the PET packaging of water, the pH value as well as
the electrical conductivity of commercially produced bottled water. The analysis
used bottled water purchased in Una-Sana Canton markets from twelve different
manufacturers. PET packaging samples were digested in the microwave oven
Anton Paar Multiwave PRO, and analysed on the PerkinElmer Atomic Absorp‐
tion Spectrum AnalyserAAnalyst 800 with Zeeman correction. Cadmium (Cd),
lead (Pb), cobalt (Co), zinc (Zn), and iron (Fe) were analysed by flame technique.
The measured concentrations of heavy metals in PET packaging were low, and
these values do not deviate from the European Parliament and Council Directive
94/62/EC on packaging and packaging waste. The electrical conductivity for all
samples was in accordance with the Ordinance on table water in B&H. The
analysed PET packaging does not present a danger to human health because the
measured concentrations of heavy metals comply with Directive 94/62/EC.
Keywords: Heavy metals · Toxicity · PET
1 Introduction
Seven from the ten most precious metals in nature belong to heavy metals, six of which
have been classified into a subset of essential microelements. Heavy metals have
multiple relevance. They represent an important raw material for many industrial
branches, some of which are necessary for living organisms, they can work favorably
on productivity of agriculture and most of them are often a significant polluter of the
environment [1]. Plastic is a polymer, a material obtained from petroleum, an ore located
inside the Earth. Using a variety of methods for treating the polymer, a wide range of
plastic materials was obtained. Heavy metals can be found in raw materials for the
production of packaging in which juices and drinking water have been packed [2, 3].
The composition of plastics includes: binders, fillers, plasticisers, stabilisers, catalysts,
paints, materials for the formation of a porous structure [4]. The specific mass of plastic
ranges from 1000 to 2000 kg/m3, while the volume depends on the porosity achieved.
The largest number of plastics is water resistant, as well as on the action of aqueous

https://doi.org/10.1007/978-3-319-90893-9_64
The Content of Heavy Metals in “PET” Bottles of Drinking Water 553
solutions of acids, bases and salts. Plastic masses are usually not stable at elevated
temperatures. PET or PETE (polyethylene terephthalate) - The most common type of
plastic that plastic bottles for drinks and similar packaging are made of. As a rule, PET
packaging is intended for single use, because reusable use is not recommended due to
the possibility of bacterial contamination. This type of plastic can be easily recycled,
giving materials for new PET bottles or polyester fibers that are further used for the
production of textiles. PVC (polyvinyl chloride) - PVC plastic is used for the production
of wide-range items, from packaging, furniture, toys, car parts, medical supplies to
construction materials due to the relatively low cost of production [5]. By-product of
PVC production are dioxins, toxic chemical compounds. It is therefore not recom‐
mended to heat the food packaged in PVC packaging, in order to avoid releasing of these
compounds. This is also the reason why this type of plastic is very rarely recycled.
According to some estimates, less than 1% of PVC is recycled. LDPE (Low density
polyethylene) - LDPE is a relatively safe type of plastic for use, and is used for the
production of plastic bottles, bags, textile and furniture. It can be recycled, although in
practice it has been poorly implemented so far. PP (polypropylene) - PP plastic has good
mechanical properties, as a rule it is solid, lightweight and heat resistant and therefore
has a wide use. It is used for the production of pipes, laboratory dishes, containers,
fittings, cases, disposable napkins, yogurt and margarine containers, chips bags and
straws [6]. Although it can be recycled, data show that today only about 3% of the
produced PP is recycled nowadays. The toxicity of heavy metals is particularly enhanced
through the chelating process and the production of sulphides with biologically active
substances, especially enzymes. This is a procedure called biomethylation. Particular
toxicity is shown by the metallurgical compounds of mercury, lead, chromium and sele‐
nium [7]. An elevated concentration of heavy metals can be the cause of autoimmune
diseases, whereby antibodies are created directed against own organs. Cadmium as
heavy metal has a particularly negative effect on the kidneys [8]. Concentrations of these
heavy metals if they are found in drinking water must be in accordance with the Regu‐
lation on the hygienic correctness of drinking water in B&H and the Ordinance on table
water in B&H [9, 10]. The most common examples are different types of allergies, and
in women, interfering with the yellow body of the ovaries (corpus luteum) that prepares
the mucous membrane to implant the fertilised ovum. It is assumed that heavy metals
also affect the metabolism of zinc, causing its lack. Zinc deficiency can cause interfer‐
ence in the function of the pituitary gland, thyroid, adrenal glands, ovaries and testes,
which can negatively affect fertility.
2.1 Material
For the analysis of heavy metals in PET bottles of drinking water, eight different samples
were used. They were purchased in the markets of the Una-Sana Canton, and they were
produced not only in Bosnia and Herzegovina but in abroad as well (Table 1).
554 E. Pehlić et al.
Table 1. Samples used for determination of heavy metals in PET packing of bottled water
Sample serial Sample name Coexistence of the sample
1 Bottled water “Oaza” PET packing
2 Bottled water “Sarajevski kiseljak” PET packing
3 Bottled water “Mg+ Mivela” PET packing
4 Bottled water “Jana” PET packing
5 Bottled water “Aqua viva” PET packing
6 Bottled water “Evian” PET packing
7 Bottled water “Studena” PET packing
8 Bottled water “Vodavoda” PET packing
2.2 Method
The samples were prepared by wet digestion in the Anton Paar Multiwave 3000 micro‐
wave oven. Destruction was performed in a mixture of nitrogen and hydrogen peroxide.
Samples (0.5 g) were measured to Teflonic Plates and 1 mL H2O2 and 6 mL HNO3 were
added [11]. Microwave digestion was performed by first heating for 10 min, then incin‐
erating for 10 min, and finally cooling operation as shown in Chart 1. After incineration,
the solution became clear, quantitatively transferred to a 50 mL graduated flask and
added to the label with ultra-clear water. Analyses were made on atomic absorption
spectrophotometer “Perkin Elmer” AAnalyst-800 with Zeeman flame correction tech‐
nique [12]. The same procedure was used for blind testing but without sample. Cali‐
bration standards for Cd, Pb, Co, Zn and Fe of 1000 mg/L (Perkin Elmer, USA) have
been used for the calibration of the instrument. Working standards were prepared by
diluting certified standards.
Chart 1. The process of PET packaging digestion

3 Results
See Tables 2 and 3, Graphs 1, 2, 3, 4, 5, 6 and 7.
Table 2. Concentrations of heavy metals in PET packings of bottled water

Sample Concentrations of heavy metals (mg/kg)
serial Cd Pb Co Zn Fe
1 0,0037 3,2838 0,4087 0,0950 0,7147
2 0,0023 3,2526 0,4065 0,0950 0,7115
3 0,0023 3,2563 0,4065 0,0945 0,7118
4 0,0027 3,2874 0,4065 0,0950 0,7203
5 0,0029 3,2834 0,4117 0,0967 0,7207
6 0,0031 3,2685 0,4097 0,0949 0,7149
7 0,0029 3,2755 0,4107 0,0953 0,7184
8 0,0025 3,2755 0,4154 0,0953 0,7183
Table 3. pH value and electrical conductivity of samples

Sample serial pH Electrical conductivity
(μS/cm)
1 6,97 256
2 7,25 220
3 7,16 199,2
4 6,60 293
5 5,75 811
6 7,04 278
7 7,28 244
8 7,25 280
9 8,44 91,1
10 7,42 223
11 7,13 294
12 5,65 1330
0.004
0.0035
Concentration mg/kg
0.003
0.0025
0.002
0.0015 Cd
0.001
0.0005
0
1. 2. 3. 4. 5. 6. 7. 8.
Sample serial
Graph 1. Cadmium concentrations in PET packings of bottled water
3.29
3.28
Concentration mg/kg
3.27
3.26
Pb
3.25
3.24
3.23
1. 2. 3. 4. 5. 6. 7. 8.
Sample serial
Graph 2. Lead concentrations in PET packings of bottled water

0.416
0.414
Concentration mg/kg
0.412
0.41
0.408 Co
0.406
0.404
0.402
1. 2. 3. 4. 5. 6. 7. 8.
Sample serial
Graph 3. Cobalt concentrations in PET packings of bottled water
0.097
0.0965
Concentration mg/kg
0.096
0.0955
0.095
0.0945 Zn
0.094
0.0935
0.093
1. 2. 3. 4. 5. 6. 7. 8.
Sample serial
Graph 4. Zinc concentrations in PET packings of bottled water

0.722
0.72
Concentration mg/kg
0.718
0.716
0.714
0.712 Fe
0.71
0.708
0.706
1. 2. 3. 4. 5. 6. 7. 8.
Sample serial
Graph 5. Iron concentrations in PET packings of bottled water
9
8.6
8.2
7.8
7.4
pH value
7
6.6 pH
6.2
5.8
5.4
5
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.
Sample serial
Graph 6. pH value of bottled water samples

Elektrical conductivity (μS/cm) Elektrical conductivity
1350
1250
1150
1050
950
850
750
650
550
450
350
250
150
50
1 2 3 4 5 6 7 8 9 10 11 12
Sample serial
Graph 7. Electrical conductivity in bottled water samples
4 Conclusion
• Concentrations of heavy metals in PET bottles of bottled water ranged from 0.0023
to 3.287 mg/kg. The lowest measured value for cadmium (Cd) in samples 2 and 3
(0.0023 mg/kg) and highest in lead (Pb) in sample 4 (3.287 mg/kg). Concentrations
of heavy metals in PET packaging for drinking water were not large and these values
did not deviate from the European Parliament and Council Directive 94/62/EC on
packaging and packaging waste.
• Furthermore, the pH value of bottled water as well as the electrical conductivity in
all 12 samples were measured. The pH value measured ranged from 5.65 to 8.44 and
the electrical conductivity of 91.1 to 1330 μS/cm. The lowest measured pH value
was in sample number 12 (5.65) and highest in sample 9 (8.44), while the lowest
electrical conductivity value was measured in sample 9 (91.1 μS/cm) and the highest
in sample number 12 (1330 μS/cm).
• According to the Ordinance on hygiene drinking water quality in BiH, pH value of
drinking water should be within the range of 6.8–8.5. Therefore, the samples 4 (6.60),
5 (5.75), and 12 (5.65) had a pH below the optimum limit, while the other samples
were within the permissible limits.
• The electrical conductivity of water according to the Rulebook on the hygienic
correctness of drinking water in BiH should be <1000 μS/cm. All samples of bottled
water had values below 1000 μS/cm, which means that they met the conditions
prescribed by the Ordinance on hygienic correctness of drinking water and do not
represent a danger to human health.
References
1. Adriano, D.C.: Trace Elements in Terrestrial Environments, pp. 3–24. Springer, New York
(2001)
2. Stieger, G.: Heavy metals in Indian plastic bottles. The Food Packaging Forum, pp. 1–8 (2015)
3. Khan, S., Khan, A.R.: Toxic heavy metal contamination in locally made plastic food
container. Int. J. Sci. Eng. Res. 6(6), 45–47 (2015)
4. Bonato, J., Šabalja, Đ.: Tehnologijski razvoj i prve primjene polimernih materijala, pp. 307–
313. Sveučilište u Rijeci, Rijeka (2012)
5. Ćatović, F.: Nauka o materijalima: polimeri, keramika, kompoziti. Univerzitet ‘‘Džemal
Bijedić’’, Mostar (2001)
6. European parliament and council directive 94/62/EC/1994 on packaging and packaging waste
7. Panijan, R., Pezerović, Dž.: Teški metali i biološki sustavi s posebnim osvrtom naživu. Zavod
za histologiju i embriologiju Medicinskog fakulteta, Zagreb, pp. 157–162 (1979)
8. Drčić, D., Pavlović, G.: Ekotoksikologija kadmija. Tekstilno-tehnološkiFakultet Zagreb 4,
66–76 (2014)
9. Pravilnik o higijenskoj ispravnosti vode za piće BiH, Sl. Glasnik BiH br.40/10
10. Pravilnik o stolnim vodama BiH, Sl. Glasnik BiH br.40/10
11. Rouessac, A., Rouessac, F.: Chemical Analysis Modern Instrumentation Methods and
Techniques. University of Le Mans, France, pp. 285–305 (2007)
12. Atomic Spectroscopy A Guide to Selecting the Appropriate Technique and System,
PerkinElmer, pp. 1–16 (2011)
Determination of Heavy Metals in Hair Dyes
by the Atomic Absorption Spectrophotometry
Ekrem Pehlić(&), Husein Nanić, Huska Jukić, and Aldina Aldžić

pehlic_ekrem@yahoo.com, nanic.h@bih.net.ba,
huskaj037@gmail.com, aldina_aldzic@hotmail.com
Abstract. Heavy metals are present in our environment, and it is therefore

difficult to avoid their presence and impact on the human organism. Some heavy
metals are essential to the human body, but if not in high concentrations. It is
known that heavy metals can be found in many cosmetic products, and one of
these is hair colours, where they can enter the body through the skin and root of
the hair. In this way, if they are in high concentrations in the body, they can
cause various health problems. It is therefore more important to determine the
content of heavy metals (Cd, Pb, Co, Zn and Fe) in hair dyes by using different
types made by different manufacturers. Also, the aim of the paper is to compare
the concentrations of heavy metals in the dyes used to produce brighter and
darker hair shades. For determining the concentration of heavy metals, eight
haircut dyes produced in 2015 were used, and they were purchased in markets
and drug stores in Bosnia and Herzegovina. The analysis of heavy metals was
done by the atomic absorption spectrophotometry method on the atomic
absorption spectrophotometer “Perkin Elmer” AAnalyst-800 with Zeeman cor-
rection, flame technique. The samples were prepared by wet digestion in the
microwave oven Anton Paar Multiwave 3000, in a mixture of nitric acid and
hydrogen peroxide in a ratio of 1 mL H2O2 and 6 mL HNO3. In all the analysed
hair dyes’ samples, heavy metals Cd, Pb, Co, Zn and Fe were detected. The
concentrations of heavy metals in the analysed hair colours were different, but
there was no significant difference between the concentration of the metal in the
same dyes’ shades from different manufacturers. Hair dyes producing the darker
shades of hair have a higher content of heavy metals than those used to produce
brighter shades. Also, low concentrations of toxic heavy metals Pb and Cd were
detected in all samples, however these heavy metals were strictly forbidden to be
used as material (EU Directive No 1223/2009) for the production of cosmetic
products, including hair dyes. Thus, all these analysed cosmetic products should
not be used because of a potential negative health effect.
Keywords: Hair dyes Heavy metals AAS
1 Introduction
Hair colouring is already so common and routine for everyone that sometimes they do
not think about the negative effects. Though our bodies require some of them in
varying amounts, overexposure to heavy.

https://doi.org/10.1007/978-3-319-90893-9_65
metals can cause illness. Metal toxicity depends upon the absorbed dose, the route
of exposure and duration of exposure, i.e. acute or chronic. This can lead to various
disorders and can also result in excessive damage due to oxidative stress induced by
free radical formation [1, 2]. Various hydrogens and hair whiteners remove hair dye
from the hair by the oxidation process. In ancient Egypt more than 4000 years ago, a
black color was used for practically lasting hair dying. The recipe preparation consisted
of a mixture of lead oxide with calcium hydroxide, which made the hair lubricate;
within three days the hair became extremely black [3, 4]. Contemporary investigations
of this hair have shown that in the hair there were nanocrystals of lead sulphide in the
size of about 5 nm, in the structure very similar to galenite (a sulphide mineral PbS that
crystallises in the cubic system). Archaeological evidence suggests that the recipe of
the Egyptians was also used by ancient Greeks and Romans. This recipe created then,
in a slightly modified form, is being used today [5]. We often wonder what is analysed
in hair to reveal heavy metals?!Firstly heavy metals are so present in our environment
and everyone is exposed to them. French chemist Eugene Schuller first created the
formula for painting eyelashes even back in 1909. Initially his company was named
“Harmless”, and a year later, the name was changed to “L’Oreal”. There are also
various types of alcohol in hair dyes, which can also dry out hair. Cd and Pb can be
found above the allowed limits for heavy metals in cosmetic products. There is
therefore a need for direct and regular testing of imported and local made cosmetic
products [6]. High levels of heavy metals and trace elements in children’s hair can have
association with autism in children [7]. Heavy metals in hair samples can also be found
in people living near industrial plants [8, 9]. Cadmium disturbs the synthesis of nucleic
acids, oxidative phosphorylation, and inhibits enzymes containing the -SH groups [10].
By analysing human hair to the content of heavy metals, the degree of poisoning with
heavy metals can be diagnosed [11, 12].
In an effort to ensure that hair dyes contain only safe substances, the European
Commission has adopted Directive 2012/21/EU limiting the use of about 45 chemicals
in hair colours and EU Directive 76/768/EC for make-up and Regulation (EC) no
1223/2009 [13, 14]. Most commercial hair dyes are with complex formulas, with
dozens of ingredients, and formulas vary considerably from producer to manufacturer.
In general, hair dyes contain dyestaff, modifiers, antioxidents, alkalizers, soaps,
ammonia, wetting agents, scents, and a range of other chemicals that have been used in
small quantities to pass special qualities to the hair (such as softening texture) and most
of these substances contain heavy metals. The dyes were most commonly found in:
amino compounds such as 4-amino-2-hydroxytoluene and m-aminophenol, then metal
oxides, such as titanium dioxide and iron oxide.
2 Material and Method
Material: Eight hair dyes samples were taken to determine the concentration of heavy
metals. All of them were produced in 2015 and they were bought in markets and
drugstores in B&H. In this paper, the concentrations of heavy metals in hair colours
originating from manufacturers from the EU and outside the EU were determined, and
Determination of Heavy Metals in Hair Dyes 563
have been found in the markets and drugstores of the Una-Sana Canton. Samples were
numbered from 1 to 8 according to Table 1.
Table 1. Samples of hair colors from different manufacturers

Type of Flavour Syoss Beauty Rubella Garnier Palete deluxe Nutriss Aroma
colour colour colour premium garnier color
sense fashion
Manufacturer Aroma Henkel Bonyplus Bugarska Loreal Schwarczkopf Loreal Aroma
“rubella” group & henkel group
Sample serial 1 2 3 4 5 6 7 8
Method: The following heavy metals were determined in the samples: Cd, Pb, Co, Zn
and Fe. The analysis was done by the method of atomic absorption spectrophotometry
“Analytical Methods” FP-3 Analysis of Meat and Meat Products (2000) [15, 16].
Heavy metal concentrations were measured on the atomic absorption spectropho-
tometer “Perkin Elmer” AAnalyst-800 with Zeeman correction by flame technique [17,
18]. The samples were prepared by wet digestion in the Anton Paar Multiwave 3000
microwave oven. Destruction was conducted in a mixture of nitrogen and hydrogen
peroxide. Samples of 0.5 g were weighed in Teflonic vessels and 1 mL H2O2 and 6 mL
HNO3 were added. Microwave digestion was performed in a way that the first samples
were heated for 10 min, then incinerated for 10 min, and at the end of the cooling, as
indicated by the Digestion Diagram 1.
Diagram 1. Phase processes in combustion of hair dyes

After incineration, the solution was clear and quantitatively transferred to a 50 mL

graduated flasks, and added to the label with ultra-clear water. The same procedure was
used for blind testing, but without sample. Calibration standards for Cd, Pb, Co, Zn and
Fe of 1000 mg/L (Perkin Elmer, USA) have been used for the calibration of the
instrument. Working standards were prepared by diluting certified standards. The
sample digestion flow pattern was shown in Diagram 2.
Diagram 2. The sample digestion flow
After preparation, the sample was introduced into the apparatus by means of a
nebuliser and was read under the recommended conditions. The conditions in which
heavy metal analysis was performed were shown in Table 2.
Table 2. Recommended conditions for analysis on AAS

Name of Flame Wavelength (nm) Burner Calibration Stock stand.
element method solution
Cd Air-acetylene 228,8 10 cm Linear/zero 1,000 µg/ml
Pb Air-acetylene 283,3 10 cm Linear/zero 1,000 µg/ml
Co Air-acetylene 240,7 10 cm Linear/zero 1,000 µg/ml
Zn Air-acetylene 213,9 10 cm Linear/zero 1,000 µg/ml
Fe Air-acetylene 248,3 10 cm Linear/zero 1,000 µg/ml
3 Results
See Table 3, Charts 1, 2, 3, 4 and 5.
Table 3. Concentrations of heavy metals

0.35
Sam. Concentrations of heavy metals (mg/kg)
0.3
serial Pb Cd Zn Fe Co
0.25
concentration mg/kg
1 0,28 0,04 0,02 0,11 0,13
2 0,21 0,03 0,01 0,08 0,10 0.2
3 0,15 0,02 0,01 0,01 0,08 0.15
4 0,29 0,01 0,02 0,10 0,13 0.1
5 0,20 0,03 0,07 0,10 0,06 0.05
6 0,17 0,02 0,01 0,05 0,08
0
7 0,31 0,03 0,02 0,16 0,12 1. 2. 3. 4. 5. 6. 7. 8.
8 0,19 0,02 0,13 0,09 0,07 sample serial
Chart 1. Concentration of Pb in the samples
0.14 0.05
0.12
0.04
0.1
concentration (mg/kg)
0.08 0.03
0.06
0.02
0.04
0.01
0.02
0 0
1. 2. 3. 4. 5. 6. 7. 8. 1. 2. 3. 4. 5. 6. 7. 8.
sample serial sample serial
Chart 2. Concentration of Zn in the samples Chart 3. Concentration of Cd in the samples

0.2 0.14
0.12
0.15
0.1
0.08
0.1
0.06
0.05 0.04
0.02
0 0
1. 2. 3. 4. 5. 6. 7. 8. 1. 2. 3. 4. 5. 6. 7. 8.
sample serial sample serial
Chart 4. Concentration of Fe in the samples Chart 5. Concentration of Co in the samples
4 Conclusion
• Heavy metals of Cd, Pb, Co, Zn and Fe were detected in all analysed samples of
hair dye. The content of heavy metals in hair dyes differed, but there was no
significant difference between heavy metals concentration in the same dyes from
different manufacturers.
• The content of heavy metals in hair dyes was regulated in the EU by Directive;
Regulation (EC) no 1223/2009 of the European Parliament and of the council, on
cosmetic products, while the legislation for heavy metals in hair dyes has not been
adopted in B&H yet.
• Hair dyes that serve to create darker shades of hair have higher content of heavy
metals than those used to obtain lighter ones.
• In all the analysed samples, low concentrations of toxic heavy metals Pb and Cd
were detected, however, these heavy metals were strictly forbidden to be used as
raw materials (EU Directive No 1223/2009) for the production of cosmetic prod-
ucts, which include hair dyes, therefore these cosmetic products should not be used.
References
1. Jaishankar, M., Tseten, T., Anbalagan, N., Mathew, B.B., Beeregowda, K.N.: Toxicity,
mechanism and health effects of some heavy metals. Interdiscip. Toxicol. 7(2), 60–72 (2014)
2. Andrea Zander, M.: eHow Contributor, Heavy Metal Poisoning Signs & Symptoms (2012)
3. Hair Dye. http://www.madehow.com/Vol.3/Hair-Dye.html
4. The Chemistry of Permanent Hair Dyes. http://www.compoundchem.com/2015/05/14/hair-
dyes/
5. How Hair Coloring Works. http://science.howstuffworks.com/innovation/everyday-innovations/
hair-coloring8.htm
6. Amartey, E.O., Asumadu-Sakyi, A.B., Adjei, C.A., Quashie, F.K., Duodu, G.O., Bentil, N.O.:
Determination of heavy metals concentration in hair pomades on the Ghanaian market using
atomic absorption spectrometry technique. Br. J. Pharmacol. Toxicol. 2(4), 192–198 (2011)
7. Al-Ayadhi, L.Y.: Heavy metals and trace elements in hair samples of autistic children in
central Saudi Arabia. Neurosciences 10(3), 213–218 (2005)
8. Nnorom, I.C., Igwe, J.C., Ejimone, J.C.: Multielement analyses of human scalp hair samples
from three distant towns in southeastern Nigeria. Afr. J. Biotech. 4(10), 1124–1127 (2005)
9. Perumal, A.S., Thangamani, A.: AAS determination of heavy metals lead and chromium
levels in human hair of people living in Katpadi and Yelagiri Hills of Vellore district. Int.
J. Res. Ayurveda Pharm. 2(5), 1568–1570 (2011)
10. Chemical properties of cadmium - health effects of cadmium - environmental effects of
cadmium. http://www.lenntech.com/periodic/elements/cd.htm#ixzz3rSgCS3pD
11. Poon, W.T., Ling, S.C., Chan, A.Y., Mak, T.W.: Use of hair analysis in the diagnosis of
heavy metal poisoning: report of three cases. Hong Kong Med. J. 10(3), 197–200 (2004)
12. Watts David, L.: Hair analysis and heavy metals-hidden toxicity detox, vol. 20, no. 2 (2009)
13. Regulation (EC) no 1223/2009 of the European Parliament and of the Council on Cosmetic
Products. Official Journal of the European Union l 342/59 (2009)
14. Commission Implementing Directive 2012/21/EU Amending, for the Purpose of Adaptation
to Technical Progress, Annexes II and III to Council Directive 76/768/EEC Relating to
Cosmetic Products
15. Analytical Metods, Atomic Absorpcion Spectrometry FP-3: Analyzis of Meat and Meat
Product, Perkin Elmer Instruments LLC, Part No, 0303-0152, USA (2000)
16. Schwartz, H.S.: Atomic Absorption Spectrometry. Elsevier, Michigan (1991)
17. Kreuz, B.: Atomic absorption Spectroscopy. Michigan University Press, Dearborn (2002)
18. Rouessac, F., Rouessac, A.: Chemical Analysis Modern Instrumentation Methods and
Techniques, University of Le Mans, France (2007)
Microbiological Analysis of Surface
Waters in the Area of National Park “Una”
Melisa Zulić ✉ , Huska Jukić, Asmir Aldžić, and Adnan Ćehajić

( )

melisa_zulic@hotmail.com, huky_jukic@yahoo.com,
asmiraldzic@hotmail.com, adnan.cehajic@bih.net.ba
Abstract. Natural water quality becomes one of the limiting factors for use and
maintenance of the environment balance. Uncontrolled pollution of surface and
groundwater reduces the quantity of quality water resources. Indicators of surface
water faecal pollution are pathogenic microorganisms that are actively developing
in water, and reaches the rivers by anthropogenic impacts through sewage
systems. The primary aim of this research is to determine the microbiological
analysis of surface waters in the area of National Park “Una”.
Water samples were collected from two places: locality I (Martin Brod) and
locality II (Kulen Vakuf). The selected microbiological parameters analysed: the
number of colonies of aerobic organotrophs, the total number of coliform bacteria
and the total number of faecal streptococcus. These parameters were used to
present realistic image of microbiological water quality and to determine pollu‐
tion hotspots at National Park “Una”.
Keywords: Water quality · Surface water · Microbiological parameters

National park “Una”
1 Introduction
Bosnia and Herzegovina has significant water resources, which is one of the major
factors of general economic development. Surface and ground water springs have a
special value as natural resources. Surface water is more or less polluted in chemical,
bacteriological or biological terms. Apart from large number of microorganisms, surface
water contains infectious (pathogenic) germs that can cause epidemics of various
diseases [1].
The most significant contaminant when it comes to human health are microorganisms
that come into water from human and animal excreta, mainly due to the unhygienic
disposition of wastewater [2]. Water quality classes are mainly determined by physical
– chemical tests. However, lack of insight into microbiological, biological and radio‐
logical characteristics would imply imprecise quality assessment of water and its clas‐
sification and would result in large errors. Therefore, the classification of natural waters
should be based on the integral quality system of the WQI (Water Quality Index) [3].
Rivers in Bosnia and Herzegovina are mostly of poor quality due to large pollution
by untreated wastewater from settlements and industry. Only the headwaters of several

https://doi.org/10.1007/978-3-319-90893-9_66
Microbiological Analysis of Surface Waters 569
rivers (Una, Sana, Neretva) have mostly good water quality [4]. Excessive exploitation
of natural resources such as forests, excessive drainage, mines residues, planned hydro‐
electric power plants and human activities inevitably remain in water, endangering the
unique location of the National Park “Una”, which accounts for about 25% of the
protected area in Bihać municipality [5].
2 Research Goal
The primary aim of this research is to conduct the microbiological analysis of surface
waters in the area of National Park “Una” (Martin Brod, Kulen Vakuf). In order to
achieve the goal, the following tasks are set:
1. Analysis of the extent and variation of faecal pollution;
2. Faecal pollution hotspots identification;
3. Assessment and description of microbiological water quality in the area;
4. Generating the map of microbiological quality of surface waters;
3.1 Material
Water samples were collected in 1 L sterile Schott containers using a sampling rod at
depths of about 20–30 cm. Subsequently, the samples were transferred to the microbio‐
logical laboratory at the Faculty of Natural Sciences for further analysis.
3.2 Methods
The microbiological examination of water involved detection of: the number of colonies
of aerobic organotrophic at 22 °C, the total number of coliform bacteria at 37 °C, the
total number of colonies of faecal origin at 44 °C and the total number of faecal strep‐
tococcus. Escherichia coli was detected by Colilert 18 [6], the technique of most prob‐
able number using three dilutions (100 ml, 1 ml, 0.01 ml). The samples were incubated
at 35 ± 0.5 °C for 18 to 22 h and analysed in the UV - cabinet. Intestinal enterococci
were analysed by MU/SF microtiter plates, by most probable number technique using
four dilutions (200 μl of 1: 2, 1:20, 1: 200; 1: 2000, ISO 7899 - 1988) [7]. The samples
were incubated at 43 ± 1 °C for 32 to 40 h and analysed in the UV - cabinet. Since we
had only two localities with accurate results, predictive modelling of microbiological
parameters was performed for the entire National Park “Una” area using ArcGIS 9.3.
software, in order to gain insight into the condition of water in the National Park
“Una” [8].
570 M. Zulić et al.
4 Results
The microbiological analysis of surface waters in the area of National Park “Una” was
conducted on two localities: locality I (Martin Brod), locality II (Kulen Vakuf). Due to
the fact that microbiological parameters are not included in the Water Framework
Directive, a realistic assessment of water conditions in the National Park “Una” area
cannot be completed. To interprete the obtained results, we used a system which clas‐
sifies faecal indicators into five classes of microbiological quality of water [9].
PROFILE NAME: MARTIN BROD

WATERCOURSE: UNA
PROFILE LABEL: U-1
REQUIRED CATEGORY: I
The analyses at the first location of Martin Brod were conducted in three series. As
presented in Table 1, in the first series, the number of aerobic organotrophic colonies at
22 °C was 128, in the second series 1 200, and in the third series 0. The maximum
allowed values (MDK) for the required category and aerobic organotrophs are 500, the
average value in these three series is 443.
Table 1. Microbiological water quality indicators of the river Una, location Martin Brod
Nr. Quality Unit of I series II series III series Medium MDK
indicators measure value
1. Aerobic N/ml 128 1 200 0 443 500
organotrophic
colonies
number on
22 °C
2. Total number N/100 ml 9 600 7 500 500 5 867 500
of coliform
bacteria at
37 °C
3. Total number N/100 ml 1 500 100 0 533 100
of coliform
bacteria of
faecal origin at
44 °C
4. Faecal N/100 ml 100 100 200 133 0
streptococcus
total number
The coliform bacteria number on 37 °C in the first series was 9 600, in the second 7
500, in the third 500. The maximum allowed values for the required category I, coliform
bacteria at 37 °C is 500, the medium value in these three series is 5 867.
Coliform bacteria number of faecal origin at 44 °C in the first series is 1 500, in
second 100, in third 0. The maximum allowed values for the required category I, coliform
Table 2. Microbiological water quality indicators of the river Una, location Kulen Vakuf
Nr. Quality Unit of I series II III IV V series VI Medium MDK
indicators measure series series series series value
1. Aerobic N/ml 5 1600 120 150 3 200 3 400 1 413 10 000
organotrophic
colonies number
on 22 °C
2. Total number of N/100 ml 100 360 800 1 500 30 000 800 5 593 10 000
coliform
bacteria at 37 °C
3. Total number of N/100 ml 0 180 300 500 600 0 263 1 000
coliform
bacteria of
faecal origin at
44 °C
4. Faecal N/100 ml 20 150 200 100 100 600 195 100
streptococcus
total number
bacteria of faecal origin to 44 °C is 100, the average value for these three series is 533.
Faecal streptococcus number in the first series is 100, in the second 100, in the third 200.
The maximum allowed values for the required category I, faecal streptococcus is 0, the
average value for these three series is 133.
Except for microbiological indicators (coliform bacteria), waters in Martin Brod area
fully meet the demanded category I. Given that it is an open watercourse that receives
water from Unac and the settlements upstream, bacteria presence is not unexpected
(Table 2).
PROFILE NAME: VS KULEN VAKUF

WATERCOURSE: UNA
PROFILE LABEL: UKF
REQUIRED CATEGORY: II
For second location, KulenVakuf, six series of analyzes were performed. In the first
series, the number of aerobic organotrophic colonies at 22 °C was 5, in the second series
1 600, in the third series 120, in the fourth 150, in the fifth 3 200, in the sixth 3 400.
Maximum allowed values for the required category II aerobic organotrophs are 10 000,
average value in these six series is 1 413.
The number of coliform bacteria at 37 °C in the first series was 100, in the second
360, in the third 800, in the fourth 1 500, in the fifth 30 000, in the sixth 800. The
maximum allowed values for the required category II, coliform bacteria at 37 °C are
10 000, the average value in these six series is 5 593.
The number of colonies of faecal origin at 44 °C in the first series is 0, in the
second 180, in the third 300, in the fourth 500, in the fifth 600, in the sixth 0.
Maximum permissible values for the required category II, colonies of faecal origin
at 44 °C are 1 000, the average value for these six series is 263. The number of fecal
streptococcus in the first series is 20, in the second 150, in the third 200, in the
fourth 100, in the fifth 100, in the sixth 600. The maximum permissible values for
the required category II, fecal streptococcus are 100, the average value for these
three series is 195.
Water of river Una in Kulen Vakuf fully corresponds to the required category II. The
presence of coliform bacteria is a normal occurrence for open waterways.
The number of aerobic organotrophs at 22 °C, at selected localities, meets the
requirements for the required category of water, up to 500 for the toughest category I,
up to 10 000 for the requested category II.
The number of coliform bacteria at 37 °C, at Martin Brod locality belongs to the
required category I water, where MDK for these bacteria 500, the average value calcu‐
lated on the basis of the six series analysis is 5 867, which indicates the contamination
of this area with coliform bacteria. The location of Kulen Vakuf meets the MDK of these
bacteria at a given location for the required category. An increased number of coliform
bacteria of faecal origin at 44 °C, at Martin Brod location, indicates faecal contamination
by humans. The location of Kulen Vakuf meets the MDK of these bacteria at a given
location for the required category.
The total number of fecal streptococcus in these two locations (Martin Brod, Kulen‐
Vakuf) has increased in relation to MDK, which also indicates the presence of fecal
contamination by humans.
Picture 1, shows the area of the National Park “Una” with the corresponding water‐
course class. Smaller areas or areas with fewer inhabitants, where there is less
pronounced anthropogenic impact on water quality, such as Gruborski Naslon, Veliko
Očijevo, Malo Očijevo, Očigrije, Veliki Cvjetnić, Mali Cvjetnić, belong to the first class
of watercourse (I), waters which can be used (in the natural state, with possible disin‐
fection) for drinking or in the food industry, while surface waters can be used for fish
farming (Salmonidae) [9].
Settlements such as Bosanski Osredci, Bastaši, Boboljušci, Martin Brod, Palučci,
Veliki Stjenjani, Kalat, Ostrovica, Rajinovci, Doljani (Donji Lapac) belong to the second
(II) water category (class), which means that water, in its natural form, can be used for
swimming and for recreation of citizens, water sports, for the farming of other fish
species (Ciprinidae), or which, with the usual treatment method – conditioning (coag‐
ulation, filtration, disinfection and the like), may be used for drinking or in the food
industry [9].
The settlement of Klisa belongs to the third (III) category of watercourses, waters
that can be used for irrigation, and after the usual methods of processing can also be
used in industry, with the exception of food industry. Going further downstream Una
River already in the area of Kulen Vakuf shows a significant degradation of all relevant
abiotic and biotic parameters. Water in the area of Kulen Vakuf belongs to the fourth
(IV) watercourse class which indicates waters that can be used for other purposes only
after proper treatment. This trend is intensified downstream of Bihać, due to the ever -
increasing wastewater inflow from settlements and smaller industrial plants [9].
Picture 1. General assessment of the microbiological status of surface waters in the area of the
National Park “Una”
5 Conclusion
The basic aim of water protection is the preservation of clean water within the prescribed
categories and to stop the trend of deterioration of the quality of groundwater and surface
water. The area of Una - Sana Canton with its natural resources has the cleanest surface
and underground waters.
The results of the conducted research indicate that water quality in the area of the
National Park “Una” based on microbiological parameters meets the prescribed stand‐
ards that ensure functioning.
The most significant anthropogenic activities leading to the deterioration of micro‐
biological quality at the locations of Martin Brod and Kulen Vakuf are: increasing
urbanization rates, direct discharges of municipal wastewater into open watercourses,
and presence of fish ponds (Martin Brod). Many municipalities in BiH have sewage
systems, but these are only used for collecting and discharging faecal water directly into
open canals or rivers, which is particularly noted at the locality of Kulen Vakuf. At the
locations indicating anthropogenic impacts, it is necessary to take measures in order to
enhance the protection of Una river in order to preserve the microbiological quality of
water.
References
1. Hukić, M.: Bacteriology and Diagnostics, 1st edn., Tuzla (1995)

2. Simičić, H.: Wastewater treatment processes, Tuzla (2002)
3. Duraković, S.: Food Microbiology, Medical edition, Zagreb (1991)
4. Džankić, N., Makić, H., Budimlić, A.: Natural and antropogenic impact on the water quality
in area of river Una, University of Bihać, Faculty of Biotechnology, Bihać (2006)
5. Management Plan for National park “Una”, Electroproject, Zagreb (2011)
6. IDEXX Water Quality Control Laboratory, Ludwigsburg, Germany, ISO/IEC 17025:2005
7. Water quality - Detection and enumeration of intestinal enterococci - Part 1:Miniaturized
method (Most Probable Number) for surface and waste water, ISO 7899-1:1998
8. ESRI., Environmental Systems Research Institute, SAD (2007)
9. Kavka, G.G., Kasimir, G.D., Farnleitner, A.H.: Microbiological water quality of the River
Danube (km 2581–km 15): longitudinal variation of pollution as determined by standard
parameters. In: 36th International Conference of IAD, Austrian Committee Danube Research/
IAD, Vienna (2006)
Research of Antimicrobial Resistance
of Clinical Important Multi-resistent Gram
Negative Bacterial Isolates in the Una-Sana
Canton Area
Asmir Aldžić1 ✉ , Huska Jukić1, Kanita Dedić2, and Amela Dubinović-Rekić2

( )
1
University of Bihac, 77000 Bihac, Bosnia and Herzegovina
asmiraldzic@hotmail.com, huskaj037@gmail.com
2
Cantonal Hospital, “Dr. Irfan Ljubijankić” Bihac, 77000 Bihac, Bosnia and Herzegovina
Abstract. The term “microbial resistance” means the immunity/resistance of

bacteria to antimicrobial drugs. The level of their resistance depends on the anti‐
microbial drug, the bacterium itself and the mechanism of resistance. Microbial
resistance may be a specific feature of a single bacterial isolate, but also of certain
bacterial species and genera. When one bacteria is resistant to three or more anti‐
biotics (i.e., an antibiotic group), there is multiple resistance.
The enormous occurrence of resistance to both pathogenic and non-patho‐
genic bacteria occurred due to uncontrolled use of antibiotics. The occurrence of
resistance of non-pathogenic bacteria disturbs the balance of existing microbial
ecosystems. Since bacteria occupy different ecological niches, the process of gene
transfer between members of the same or different species/genera, the number of
resistant strains increases. Particularly important is the interspecific transfer of
genes from pathogens to non-pathogenic bacteria and vice versa. The main role
in this process is the micro-organisms commensals of human and animal, which
are “vectors” for the transmission of genes of resistance between pathogenic and
non-pathogenic bacteria. Monitoring the occurrence of resistance in non-patho‐
genic bacteria, primarily of human and animal commensals and food micro-
organisms, is of great importance not only from the ecological but also from the
aspect of the treatment of bacterial infections of humans and animals.
Isolates of bacteria from hospitalised patients at the departments of the
Cantonal Hospital “Dr. Irfan Ljubijankić” in Bihac were used for the research and
development of this work.
In the study period, the occurrence and resistance of Gram-negative multi-
resistant types was monitored.
Of the total of 9 species of isolated and identified Gram-negative multire‐
sistant bacteria, the most abundant isolates were: Klebsiella pneumoniae ESBL
(with 16 isolates, or 23,18%), Escherichia coli ESBL, Acinetobacter spp. and
Pseudomonas aeruginosa (with 11 isolates, or 15,94%), Enterobacter spp. ESBL
(with 8, or 11,59% isolates), Proteus mirabilis (with 5 isolates, or 7,24%), then
Serratia spp. (with 3 isolates, or 4,34%), and Citrobacter spp. ESBL and Kleb‐
siella oxytoca (with 2 isolates, or 2,89%).
Keywords: Microbial resistance · Multi-resistance · Bacterial isolates

Gram-negative bacteria

https://doi.org/10.1007/978-3-319-90893-9_67
576 A. Aldžić et al.
1 Introduction
Resistance to bacteria triggered by broad and frequent use of antimicrobial drugs is

currently one of the biggest medical problems, which has a significant impact on the
outcome of patients’ treatment, as well as on health care costs. Thus, the infection with
resistant agents is associated with a worse outcome of treatment. The correlation of the
consumption of antimicrobial drugs with the selection of resistant bacteria has been
confirmed by many studies, however, direct evidence of degree and association and the
different potential of certain antimicrobial drugs or groups in the selection of resistant
bacteria are still lacking. Resistance leads to inadequate empirical therapy, delay in
starting successful treatment, and to the use of less effective, more harmful and more
expensive drugs. Resistance to antimicrobial therapy is associated with failure of treat‐
ment, prolonged or additional hospitalisation, increased cost of treatment and increased
mortality. Additional costs and lost free beds for hospitalisation also arise due to the
need to control the spread of strains resistant to antimicrobial therapy within the hospital.
All of the above has a direct effect on patients, and on the cost-effectiveness ratio. For
this reason, there is a great need for controlling resistance to antimicrobial therapy by
improving the use of antibiotics [2].
The discovery of antibiotics and chemotherapeutics and their successful and wide‐
spread application in the 1940s marked a turning point in the fight against infectious
diseases. However, it soon became apparent that bacteria were able to develop different
defense mechanisms and become resistant - resistant to the effect of the drug. When a
bacterium is resistant to three or more antibiotics, it is known as a multiple resistance.
When resistance can be transmitted from resistant types to the same or other types of
bacteria, it is an infective, portable resistance [5, 6]. The problem of bacterial resistance
to antimicrobial agents is not as simple as in the days of their initial use. Numerous
scientific studies point to the continuous occurrence and spread of resistance in bacteria
and the justified fear of occurrence of fatal infectious diseases without the possibility of
effective antimicrobial therapy. This is why this problem is actual not only at the local
but also at the global level, especially if antibiotics/chemotherapeutics are known to act
on both pathogenic and non-pathogenic bacteria, leading to suppression of growth or
elimination of sensitive types. This feature is not only a health but an environmental
problem due to favoring the growth of few resistant types of bacteria [5, 6]. The resist‐
ance of bacteria to antibiotics does not only endanger the treatment of common infectious
diseases, but also advances in many branches of medicine. Many invasive diagnostic
and therapeutic procedures have developed due to the possibility of prophylaxis and
treatment of infectious complications. It is therefore not surprising that the fight against
antibiotic resistance is one of the priorities of the World Health Organization and one
of the requirements of the Council of the European Union set up for all its members.
Bacteria have developed many different resistance mechanisms so there is no antibiotic
to which resistance has not developed. There is a particular problem recently in the
resistance to carbapenems in Gram-negative bacteria, and the recent onset of resistance
to colistin in acinetobacter and enterobacteria also worries. Such isolates are generally
“pan-drug-resistant”, which means that there is no effective antibiotic for them. It is
Research of Antimicrobial Resistance 577
unlikely that new antibiotic groups will appear in the near future that would have an
effect on such isolates, and it is therefore important to preserve the efficacy of existing
antibiotics by rational use and suppression of the spread of multiresistant types using
hospital hygiene measures [1].
2 Research Objectives
Determine the frequency of multiresistance among different Gram negative bacterial

isolates.
3 Material and Methods of Work
3.1 Material of Work
For the research and creation of this work, isolates of bacteria from hospitalised patients
at the departments of the Cantonal Hospital “Dr. Irfan Ljubijankić” in Bihac were used.
During the research, the resistance of isolated multiresistant bacteria was observed for
a period of two years: 2014–2015. All isolates of certain bacterial species have been
tested for all antibiotics envisaged for this species. Antibiotics envisaged for a particular
species were listed in the forms for monitoring resistance for the years indicated. The
duplicate types were excluded from the data - defined as isolates of the same bacterial
species - isolated in the same patient in any sample for a period of 30 days.
3.2 Methods of Work

An ESBL test was used to demonstrate ESBL types, or types producing β-lactamases
of the enlarged spectrum. It was found that 0.5 ml of the colony was added to the 2 ml
of saline solution to obtain a dilution, followed by the addition of a strip with an impreg‐
nated antibiotic cefepime and amoxicillin + clavulanic acid. If it shows the difference
in the three dilutions, it is a sign that the test was positive or that it is the ESBL’s.
Methods of antibiotic susceptibility testing (disc diffusion method and methods for
determining the minimum inhibitory concentration of MIK, the so-called E-test) were
used according to EUCAST.
Identification of the causative agent and the production of antibiograms is a “gold
standard” in antibiotic therapy because it provides the possibility of selecting the target
antibiotic, making the therapy more efficient, shorter and cheaper. The discovery of
multiresistant bacteria was one of the main tasks of the research study. The selection of
antibiotics for testing took into account the following: clinical drug effectiveness, resist‐
ance (known local resistance), efforts to minimise resistance, costs of therapy, the latest
recommendation on the first choice, and alternative (reserve) antibiotics.
4 Results of Research
Of the total of 9 isolated and identified Gram-negative bacteria, the most abundant
isolates were: Klebsiella pneumoniae ESBL (with 16 isolates, or 23,18%), Escherichia
coli ESBL, Acinetobacter spp. and Pseudomonas aeruginosa (with 11 isolates, or
15,94%), Enterobacter spp. ESBL (with 8 ili 11,59% isolates), Proteus mirabilis (with
5 isolates, or 7,24%), Serratia spp. (with 3 isolates, or 4,34%), then Citrobacter spp.
ESBL and Klebsiella oxytoca (with 2 isolates, or 2,89%) - Table 1.
Table 1. Types of isolated Gram-negative multiresistant bacteria

Isolated cause N %
Acinetobacter spp. 11 15,94
Citrobacter spp. ESBL 2 2,89
Enterobacter spp. ESBL 8 11,59
Escherichia coli ESBL 11 15,94
Klebsiella oxytoca ESBL 2 2,89
Klebsiella pneumoniae ESBL 16 23,18
Proteus mirabilis 5 7,24
Pseudomonas aeruginosa 11 15,94
Serratia spp. 3 4,34
TOTAL 69 100
The results of the antimicrobial resistance test of 16 isolated multiresistant types of

Klebsiella pneumoniae ESBL were shown in Chart 1.
Chart 1. Antimicrobial resistance of isolated multiresistant strains Klebsiella pneumoniae ESBL
The results of antimicrobial resistance testing of 11 isolated multiresistant strains of

Escherichia coli ESBL are shown in Chart 2.
Chart 2. Demonstration of the antimicrobial resistance of isolated multiresistant strains of

Escherichia coli ESBL
Results of antimicrobial resistance testing of 11 isolated multiresistant strains Acine‐

tobacter spp. were shown in Chart 3.
Chart 3. Demonstration of antimicrobial resistance of isolated multiresistant strains of

Acinetobacter spp.
The results of antimicrobial resistance testing of 11 isolated multiresistant strains of

Pseudomonas aeruginosa were shown in Chart 4.
12
10
8
R
6
I
4 S
2
0
Chart 4. Demonstration of antimicrobial resistance of isolated multiresistant strains of

Pseudomonas aeruginosa
5 Discussion
When it comes to Gram-negative multiresistant bacteria, they show a high degree of

sensitivity to carbapenems, that is imipenem, as well as to aminoglycosides and
amikacin, while other antibiotics show resistance in almost 100%, which coincides with
similar research, carried out in Croatia [10].
Isolates of Gram-negative bacteria Klebsiella pneumoniae ESBL were represented
with 23.18% compared to the total number of bacterial isolates, i.e. 16 isolates. It should
be noted that all isolates were multiresistant, i.e. showed 100% resistance to almost all
tested antibiotics, which coincides with a similar study conducted in Croatia [9].
Isolates of Gram-negative bacteria Acinetobacter spp. were represented with 15.94%
compared to the total number of bacterial isolates [11]. It should be noted that all isolates
of Acinetobacter spp. were multidirectional, i.e. showed 100% resistance to 4 out of 7
tested antibiotics. Acinetobacter spp. has become one of the leading causes of hospital
infections in all countries of the world, especially in intensive care units. Recent litera‐
ture describes this microorganism as a hospitalised oxymoron that represents a threat to
resistance to all pan-drug-resistant menace [4]. Hospital-acquired infections caused by
types of Acinetobacter spp. most often affect the respiratory tract, but these types caused
hospital infections of the urinary tract and wound infection, which can cross into septi‐
cemia [8]. The percentage of colonisation with these bacteria is extremely high in hospi‐
talised patients, especially in those who were hospitalised for a long time and who
received wide spectrum antibiotics [3].
Isolates of Gram-negative bacteria Escherichia coli were represented with 15.94%
compared to the total number of bacterial isolates (i.e. 11). It should be noted that all
isolates were multiresistant, i.e. showed 100% resistance to 9 antibiotics, while the
remaining 7 antibiotics also showed high resistance (above 70%), which again coincides
with similar research conducted in Croatia [9].
Isolates of Gram-negative bacteria Pseudomonas aeruginosa were represented with

15.94% compared to the total number of bacterial isolates (i.e. 11 isolates). It should be
noted that no single isolate showed 100% resistance to 7 tested antibiotics, but 2 anti‐
biotics (ciprofloxacin and gentamicin) showed a sensitivity of over 90%. In a related
study conducted in Croatia, the isolates showed a high degree of resistance to almost all
tested antibiotics [7, 10].
The next ones follows Enterobacter spp. ESBL (8 isolates or 11.59%), Proteus
mirabilis (5 isolates or 7.24%), Serratia spp. (3 isolates or 4.34%), Citrobacter spp.
ESBL and Klebsiella oxytoca (2 isolates or 2.89%), which showed a high degree of
resistance, which also coincides with similar research conducted in Croatia [10].
6 Conclusion
The results of the study showed that in relation to Gram’s differential colouring, a total
of 69 isolates belonged to Gram-negative multiresistant bacteria, or 9 isolated types of
Gram-negative bacteria. Of the total of 9 isolated and identified Gram-negative multi‐
resistant bacteria, the following isolates were most common: Klebsiella pneumoniae
ESBL (with 16 isolates or 23.18%), Escherichia coli ESBL, Acinetobacter spp. and
Pseudomonas aeruginosa (with 11 isolates or 15.94%), Enterobacter spp. ESBL (with
8 isolates or 11.59%), Proteus mirabilis (with 5 isolates or 7.24%) Serratia spp. (with 3
isolates or 4.34%), Citrobacter spp. ESBL and Klebsiella oxytoca (with 2 isolates or
2.89%).
Gram-negative multiresistant bacteria have shown a high degree of sensitivity to
carbapenems, that is imipenem, as well as to aminoglycosides and amikacin, while other
antibiotics show resistance in almost 100% form.
References
1. Bedenić, B., Sardelić, S., Ladavac, M.: Multirezistentne bakterije. Medicinski fakultet.
Zagreb, Sveučilište u Zagrebu (2015)
2. Bojić, A.: Praćenje rezistencije sojeva Pseudomonas aeruginosa na antibiotike u JU
Kantonalna bolnica Travnik. Master teza, Prirodno-matematički fakultet Sarajevo (2010)
3. Bonten, M.J., Slaughter, S., Ambergen, A.W., Hayden, M.K., Van Voorhis, J., Nathan, C.I.,
Weinstein, R.A.: The role of “Colonization pressure” in the spread of vancomycin resistant
enterococci. Arch. Int. Med. 158, 1127–1132 (1998)
4. Davey, P., Brown, E., Fenelon, L., Finch, R., Gould, I., Holmes, A., Ramsay, C., Taylor, E.,
Wiffen, P., Wilcox, M.: Systematic review of antimicrobial drug prescribing in hospitals.
Emerg. Infect. Dis. 12, 211–216 (2006)
5. Falagas, M.E., Karageorgopoulos, D.E.: Clinical microbiology and infectious diseases
(ECCMID) - 18th European congress. Drug resistance among gram negative and gram-
positive bacteria. Drugs 11, 409–411 (2008)
6. Fridkin, S.K., Pear, S.M., Williamson, T.H., Galgiani, J.N., Jarvis, W.R.: The role of
understaffing in central venous catheter-associated bloodstream infections. Infect. Control
Hosp. Epidemiol 17, 150–158 (1996)
7. Kalenić, S.: Rezistencija bakterija na antibiotike. Zavod za kliničku i molekularnu

mikrobiologiju, Klinički bolnički centar Zagreb (2000)
8. Karlowsky, J.A., Draghi, D.C., Jones, M.E., Thornsberry, C., Friedland, I.R., Sahm, D.F.:
Surveillance for Antimicrobial susceptibility among clinical isolates of Pseudomonas
aeruginosa and Acinetobacter baumannii from hospitalized patients in the United States,
1998 to 2001. Antimicrob. Agents Chemoth. 47(5), 1681–1688 (2003)
9. Tambić Andrašević, A., Tambić, T., Kalenić, S., Katalinić-Janković, V., Payerl Pal, M.:
Osjetljivost i rezistencija bakterija na antibiotike u Republici Hrvatskoj 2008. g. Akademija
medicinskih znanosti Hrvatske, Zagreb (2008)
10. Tambić-Andrašević, A., Jelić, M., Gužvinec, M., Butić, I., Bukovski, S.: Rezistentne
enterobakterije u Hrvatskoj – uloga praćenja rezistencije na antibiotike na nacionalnoj razini.
Klinika za infektivne bolesti dr. Fran Mihaljević, Zavod za kliničku mikrobiologiju, Zagreb
(2012)
11. Tambić-Andrašević, A.: Otpornost bakterija na antibiotike - vodeći problem medicine u 21.
stoljeću. Klinika za infektivne bolesti dr. Fran Mihaljević, Zavod za kliničku mikrobiologiju,
Zagreb (2007)
New Technologies in Civil Engineering,
Education, Control Quality
Nanotechnology in Civil Engineering
Merima Šahinagić-Isović ✉ , Marko Ćećez, and Fuad Ćatović

( )
“Džemal Bijedić” University of Mostar, 88000 Mostar, Bosnia and Herzegovina

{merima.sahinagic,marko.cecez,fuad.catovic}@unmo.ba
Abstract. Nanotechnology is a way forward, not only in civil engineering, but

in other industries as well. Much effort is being given to these research, and main
focus is in improving existing materials, in terms of their properties. In this paper
brief information of nanotechnology applied and investigated in civil engineering
will be given. Most of the presented research include enhancement of concrete,
as the most used material in civil engineering, thus silica fume, fly ash and red
mud are presented, as well as carbon nanotubes, which are a great leap forward
in this field.
Keywords: Carbon nanotubes · Silica fume · Fly ash · Red mud
1 Introduction
Investigation and exploitation of internal characteristics of materials on nano scale with

the aim to obtain exceptional materials’ performance is today one of the most attractive
scientific fields. In the last 15 years, these investigations gained a lot of attention, and it
is to believe that nanotechnology would be next industrial revolution. Nanotechnology
deals with materials’ manipulation on nano scale and in that level affects material char‐
acteristics that would be explored in practice.
Civil engineering sector was one of first to recognize the promising roll of nano‐
technology in application. But in the last years, civil engineering is falling behind in
these investigations compared to other industries. Most of the research is focused on
new materials, materials characteristics and materials enhancement. Also, the consump‐
tion of natural materials is one of the topic in focus today, and researchers are turning
to nano scale in order to replace natural resources with appropriate replacing materials.
In this paper some brief insight of the research in this field will be given.
2 Carbon Nanotubes (CNT)
Carbon nanotubes are one of the most investigated nano materials today. Discovered in
1991, carbon nanotubes are still considered as wonder material with great potential.
They have increased mechanical, magnetic and electronic properties with great resil‐
ience. CNT have only 17% of steel weight, but they have 100 times greater tensile
strength [1]. With radius of only one nanometer these tubes can be used for strengthening
almost every material. Even though they have small radius, length can be much greater,

https://doi.org/10.1007/978-3-319-90893-9_68
586 M. Šahinagić-Isović et al.
up to thousand microns. Carbon nanotubes can be produced by twisting the graphite

sheet. Sheets can be twisted in many ways. Graphite modulus of elasticity is one of
largest for any known material. Thus, CNT have extremely high modulus of elasticity
as high as 1 TPa (TeraPascal), giving it high tensile strength [2]. Also, high elongation
is observed, as high as 18%, with low coefficient of thermal conductivity.
In civil engineering CNT (Fig. 1) can be used for strengthening of concrete, by
providing tensile strength and slowing crack propagation. It can be used for replacement
of steel where properties of steel are not sufficient. Due to high chemical resistance, it
can be used for constructions exposed to chemical attack.
Fig. 1. Structure of CNT
3 Silica Fume
Silica fume (amorphous electro filter SiO2 dust) is obtained as a secondary product in
the production of ferosilicate and similar materials. The silica fume is an industrial
mineral additive by origin, and according to its characteristics it is a pozzolan. Silica
fume added to concrete is expected to improve hydration and produce CSH gel (calcium
silicate hydrate). Silica fume is most active pozzolan with the highest pozzolanic
activity. It consists mainly of SiO2 with frequency ranges from 85 to 97%. The silica
fume particle size ranges from 0.01 to 0.3 μm (10–300 nm) (Fig. 2). After being added
to the concrete, the silica fume is initially inert. Only when the Portland cement and
water begin to react, i.e. the beginning of hydration, silica fume is activated. The chem‐
ical reaction creates two chemical compounds: calcium silicate hydrate (CSH) and
calcium hydroxide (CH), which is still called free lime, serves only as a coating of
concrete pores as a filer or drains out of concrete. Pozzolanic reaction is taking place
between the CH and silica fume, producing additional CSH in pores around hydrated
cement. This process leads to the creation of a much denser mixtures of concrete, at the
expense of undesirable voids in the concrete.
Nanotechnology in Civil Engineering 587
Fig. 2. Silica fume in solid state (left) and comparation between cement and silica fume particle
size (right)
With the proper use of silica fume not only the mechanical properties can be
improved, but also the durability of concrete. Numerous studies have been carried out
on use of silica fume as an additive for concrete, which confirm the improved properties
of the concrete, primarily an increase in compressive strength and water resistance [3–
5]. Also, small silica fume particles contribute to the fresh concrete behavior by
improving the workability and increasing the cohesiveness and filling the inner structure
because they are considerably smaller than the cement particles.
4 Fly Ash
Flying ashes is an industrial mineral additive, and according to the characteristics it

belongs to pozzolan because it actively participates in the process of hydration of cement.
It is obtained by combusting coal dust in coal-fired power plants and by separating the
electrostatic filter equipment from waste gases. The chemical composition and charac‐
teristics of flying ash depend on a large number of factors, and hence its ability to use.
Actively used in the manufacture of some types of cement and can be used directly in
the concrete mixture [5]. Particle size of fly ash is from 0.5 μm to 300 μm.
In practice, it is known that fly ash improves the performance and quality of concrete.
It has an impact on the improvement of plastic properties of concrete and workability,
reduces the amount of water required, reduces segregation and bleeding and affects the
reduction of the hydration heat. In addition, it increases the strength of concrete, reduces
permeability, reduces corrosion of reinforcing steel, increases resistance, and reduces
the alkaline-aggregate reaction [6].
588 M. Šahinagić-Isović et al.
5 Red Mud
Red mud is a byproduct of industrial aluminum production. Aluminum is commercially

produced from bauxite in two steps. In the first stage, aluminum is obtained by Bayer’s
procedure, and in the second stage, aluminum is electrolysed to obtain metal for use. In
this process, insoluble residue is created. That residue is known as red mud. It is esti‐
mated that at 1 t of produced aluminum, it creates 0.3 to 1.7 t of red mud [7]. Red mud
consists mainly of iron oxide, quartz, sodium aluminosilicate, calcium carbonate, tita‐
nium dioxide and sodium hydroxide. Iron oxide is giving it recognizable red color
(Fig. 3). It has an average pH value of 10 to 12.5 and a particle size of less than 10 μm.
Red mud can not be fully considered artificial pozzolana, because it does not meet all
the requirements. However, its pozzolanic activity index is good.
Fig. 3. Dried red mud
When used as an additive to concrete it can significantly increase durability of

concrete, enhancing characteristics such as permeability, frost resistance etc., but it
requires increased consumption of water and reduces mechanical properties [5].
6 Conclusion
In this paper brief insight of nanotechnology used in civil engineering is presented. Much
effort is being given in the research of material enhancement. Since concrete is most
used material in civil engineering, thus improvement of its’ characteristics is the most
logical way. Materials such as silica fume and fly ash are used in construction, but there
is much more space for improvement in this field. Besides them, a lot of other materials
are being explored, such as red mud, but their application is still on hold.
References
1. Saito, R., Dresselhaus, G., Dresselhaus, M.S.: Transport Properties of Carbon Nanotubes.
Physical Properties of Carbon Nanotubes, pp. 137–162 (1998)
2. Popov, V.N.: Carbon nanotube: properties and applications. Mater. Sci. Eng. R: Rep. 43(3),
61–102 (2004)
Nanotechnology in Civil Engineering 589
3. Jain, A., Pawade, P.Y.: Characteristics of silica fume concrete. In: International Conference
on Quality Up-gradation in Engineering, Science and Technology (ICQUEST2015),
Maharashtra, India, pp. 22–26 (2015)
4. Shihada, S., Arafa, M.: Effects of silica fume, ultrafine and mixing sequences on properties of
ultra high-performance concrete. Asian J. Mater. Sci. 2, 137–146 (2010)
5. Šahinagić–Isović, M., Ćećez, M., Ćatović, F.: Properties of self-compacting concrete with the
addition of silica fume and fly ash. In: 6th International Scientific and Professional Conference
“Civil Engineering - Science and Practice”, GNP 2016, 07–11 March 2016, Žabljak,
Montenegro, pp. 765–772 (2016)
6. Bjegović, D., Štirmer, N., BanjadPečur, I.: By-products of the metallurgical industry as
alternative raw materials in the construction industry. In: 6th International Scientific and
Professional Conference “Civil Engineering - Science and Practice”, GNP 2016, 07–11 March
2016, Žabljak, Montenegro, pp. 679–687 (2016)
7. Sawant, A.B., Kumthekar, M.B., Sawant, S.G.: Utilization of neutralized red mud (industrial
waste) in concrete. Int. J. Inventive Eng. Sci. (IJIES) 1(2), 9–13 (2013)
Support Parametric Vector Analysis
of Available Resources for Minimisation
of Project Duration–Four Varieties
of Conditions
Omer Kurtanović1(&) and Lejla Dacić2

1
adsami@bih.net.ba
2
University of Travnik, 72270 Travnik, Bosnia and Herzegovina
lejla.dacic@gmail.com
Abstract. This paper considers the dependence of the project duration on

available resources through the problem of parametric analysis (PA). Vector PA
has been presented from the aspect of variants of perturbation directions and
intensities for several different categories of resources that belong to the ‘work
type’ category. Four varieties of direction combination and selected change
intensities are illustrated for m = 3 resource categories Pi with constant avail-
0 0
able amounts ni ¼ ni ðtÞ and intensities ni ¼ ni ðtÞ ; i 2 I; t 2 T. A general
formulation of the mathematical model, solution algorithm, and an illustrative
example, are given too. The illustrative examples were solved using the MS
Project software.
Keywords: Project minimisation Available resources Parametric analysis

Variants of conditions Software
1 Introduction
Every optimisation problem requires analysing the sensitivity of the solution to the
change of significant elements. In the problem of linear programming (LP) with one
criterion, two approaches to testing the value of criteria can be implemented [1–3].
(1) A post-optimal analysis or sensitivity analysis of an optimal solution (AO) from
the point of view of: (1) the criteria (the intervals of the coefficients changes with
variables in order to maintain the optimal values of the variable), (2) the free
members of constraints (the intervals of changes are determined in order to
maintain the structure of the optimal solution, that is, the base of each variable),
(3) technological matrices (the change intervals are determined for the charac-
teristic coefficients to maintain the optimal solution structure), and (4) the com-
bination with (1) to (3).
(2) Parametric analysis of the model (PA), observing all the admissible values of the
corresponding elements in cases (1) to (4) mentioned for the analysis (1).

https://doi.org/10.1007/978-3-319-90893-9_69
Support Parametric Vector Analysis of Available Resources 591
Software packages for LP typically support analyses (1) and (2) under (1). More
efficient versions allow (1) and (2) under (1) and (2). For example, the reference [4]
details the application of DecisionSupportSoftware for OR/MS - OperationsResearch
and ManagementScience (WinQSB) in the example of the transport task LP for AO
and PA areas by applying a model with one criterion and more criteria, as well as a
target programming model (CP). The same type of models shows the reference [5] on
production examples.
In the field of project management (UP), the problem of project planning is usually
reduced to minimising the duration, i.e. duration and cost. Of particular importance is
the parametric analysis of project costs in various infrastructure branches (software
development, construction, pharmaceutical industry, etc.). See, e.g. [6]. The minimi-
sation of the project duration and the PA of the available resources was exposed the
reference [7] in analogy to the problem of LP. Further, the vector PA is more readily
interpreted.
2 Types of Resources on the Project
There is no project that does not require adequate resources. For example, construction
projects involve dozens of qualification structures of workers, many types of mecha-
nisation, and hundreds of different materials for the execution of various significant
works (electrical installations, water supply and sewerage, etc. in cases of construction
of a residential or business facility) [8].
Two types of resources and their characteristics were discussed in details by ref-
erence [6] on a general project with activities marked to form a network diagram
(MD) with activities in nodes (EvenOriented Network), j 2 J ¼ f1; . . .; pg:
Non-consumable or permanently present, i.e. renewable resources Pi of “Work-
type”, (are workers of necessary qualifications, working machines with certain char-
acteristics, etc., i 2 I ¼ f1; . . .; mg: They determine the time of the activities and the
project.
Consumable or non-renewable resources Mk “Materialtype” are various types of
materials, parts, and the like, used by Work type resources to perform certain activities,
k 2 K ¼ f1; . . .; qg:
Generally speaking, it is assumed that the acquisition of the resources “Material”
in sufficient quantities is required to carry out the project with the available resources of
“Work”. This is based on the most commonly used standard software packages for UP
in our area (Primavera and MS Project). The resources of “Material” are baseline data
for the total need for each activity that the software deploys during the period of its
implementation by applying appropriate technological rules (uniformly or in some
other way). The analyst plans to procure these resources (inputs) with which he will
provide non-negative stocks (condition) to perform all planned activities on each day of
the project (exit). The CA Super Project software supports the recording of purchases,
inventory calculations and neutralisation of negative inventories (resource leveling of
“Material” at the same time with the leveling of resources “Work” (see, e.g., [9]). If
software is used without the abovementioned capability, it is necessary for the analyst
592 O. Kurtanović and L. Dacić
to shift activities outside the period with negative stocks of “Material” resources and
level the resources level of “Work” by using software [9–12].
Resources of “Work” should be considered in more details, in accordance with the
reference [12], which is partly transmitted below with the replacement of the activity
and resource indexes. The wider aspects of project planning from other references are
taken into account. It is convenient to use the following tags: TP ; t; v; T ¼ f1; . . .TP g-
duration of the project, absolute time units (days) t 2 T for analysis of project elements,
their number and their set; Wij ;Ij -total amount of work (hours) Pi to perform Aj and an
indexes’ set of resources types used, i 2 Ij I; j 2 J; wij ðtÞ; nij ðtÞ; aij ðtÞ-amount of work
(intensity from the time point of view), planned number of units (intensity from the
point of view of resources’ units) and daily working hours (hours) for Pi on Aj in t ; tij ;
tj -duration (days) of engaging Pi on Aj and duration Aj . In order to simplify the
problem, let’s assume constant values for the following elements: wij ¼ wij ðtÞ; nij ¼
nij ðtÞ; aij ¼ aij ðtÞ; i 2 Ij ; j 2 J; t 2 T: The relations (a) and (b) apply. If Aj uses only
one type of Pi ; then it is followed by tj = tij : Otherwise, for several types of resources, tj
determines a longer tij (c).

nij ¼ wij =aij ; given wij and aij
wij ¼ aij nij ) ; i 2 Ij ; j 2 J ðaÞ
aij ¼ wij =nij ; given wij and nij
tij ¼ Wij =wij ¼ Wij =ðaij nij Þ; i 2 Ij ; j 2 J ðbÞ
tj ¼ Max ftij g ¼ Max fWij =wij g ¼ Max fWij =nij =aij g; j 2 J ðcÞ
i2Ij i2Ij i2Ij
The general mathematical model of minimisation Tp ; considering the corresponding

Gant’s diagram (GD), can be formulated by relations (1.1)–(1.3).
min Tp ðn; hÞ ð1:1Þ

GD
With limitations:
X
j2Ji
nij ðtÞhij ðtÞ ni ðtÞ; i 2 I; t 2 T ð1:2Þ

1
hij ðtÞ ¼ ; j 2 Ji ; i 2 I; t 2 T ð1:3Þ
0
(1.2) determines that the total needs of Ri at all Aj displayed in resources’ units, by
observing each t; do not exceed the available quantities ni ðtÞ: Binary variables hij ðtÞ
have the purpose to identify those of Aj on MD that use Pi in t (1 if Aj is done in t and
uses Pi ; 0 otherwise). The set Ji consists of indexes for Aj used by the observed Pi ;
j 2 Ji J; i 2 I: In (1.1) it is noted that TP function from the schedule Aj in time (on
MD), of the available units’ numbers Pi in t (variable-quantity matrixes nm;v or column
vector nm for constant quantities), and binary variables (matrix hm;v Þ: Minimum
duration of a project Tpmin in (1.1) was found by the standard software for UP that
supports the leveling of resources “Work”, that is, reducing them to the available limits.
3 PA of Maximum Resources “Work” in Minimising

the Project
Different values of ni ðtÞ for Pi in t affect the Tpmin model (1.1)–(1.3). A detailed
examination of the consequences is allowed by the PA model (2.1)–(2.3), observing all
the acceptable values of ni ðtÞ and the following elements: n matrix or the original
values vector ni ðtÞ; l factor of permutation (a common indicator of changes);
n 0i ðtÞ; n 0 directions with intensities of change for Ri both their matrix or vector;
i 2 I; t 2 T:
0
min Tp ðn þ l n ; hÞ ð2:1Þ
GD
With limitations:
X 0
j2Ji
nij ðtÞhij ðtÞ ni ðtÞ þ l ni ðtÞ ; i 2 I; t 2 T ð2:2Þ
hij ðtÞ ¼ 1 ili 0; j 2 Ji ; i 2 I; t 2 T ð2:3Þ
The PA algorithm is implemented in two steps: first it is required to determine the

0
quantities ni ðtÞ þ l ni ðtÞ with which the corresponding project plan with the UP
software is formed and determines Tp . Depending on the characteristics of the concrete
project, individual PA (each resource separately) can be performed or/and parametric
PA (all resources or multiple selected resources), observing the project in full or
appropriate periods. Otherwise, the parametric PA allows a consideration of individual
Ps also, by setting up ns ðtÞ0 = 1 i ni ðtÞ0 = 0 for i 6¼ s:
The usefulness of the results of the vector PA depends on the correct definition of
the directions and intensity of the changes in the amount of resources in order to fully
or to a greater extent reflects the particularity of the project under consideration.
Changing the amount of certain resources’ categories can be mutually dependent. For
example, planning to provide more or less of Pi in the coming years may require
appropriate changes of some other Pi . This is remarkable for close degrees of expertise
within the same qualification of workers (electricians, plumbers) and mechanisation
(excavators of smaller and larger work performance). Generally, two shapes for
directions (same or different) and two types of intensities (same or different) can be
combined. Four variants of combining routes and selected change intensities for m ¼
3 resource categories Pi with constant available quantities ni ¼ ni ðtÞ and intensities
n0i ¼ ni ðtÞ 0 ; i 2 I; t 2 T (Table 1).
Remark 1. It is required for ni to be whole numbers, so it is necessary to choose two
sizes appropriately: Dl for growth l at intervals with the corresponding project ele-
ments and the starting value for ni :
Table 1. Variants of vector PA for m ¼ 3

Directions Analysis P1 and P2 ; intensities n0i Analysis P1 to P3 ; intensities n0i
Same Same: [1, 1, 0] Same: [1, 1, 1]
Different: [2, 1, 0] Different: [2, 1, 3]
P1 changes (grows/falls) • P1 grows/falls
2 times faster than P2 2 times faster than P2
P3 grows/falls
3 times faster than P2
Different Same: [1, −1, 0] Same: [1, −1, 1]
• P1 grows/falls • P1 and P3 grows/falls
• P2 inversely (falls/grows) • P2 inversely (falls/grows)
Different: [2, −1, 0] Different: [2, −1, 1]
P1 changes P1 changes:
2 times faster than P2 in the opposite direction • 2 times faster than P2
• in the opposite direction
• 2 times faster than P3
in the same direction
4 Illustrative Examples
A hypothetical project with activities from A1 to A9 and resources “Work” P1 to P3 has

duration between Tpmin ¼ 100 i Tpmax ¼ 150 days. Earliest start dates Aj determine Tpmin
(a)
(b) Cp
Tp
Chart 1. (a) Project plan without resources constraints ½nmin

i = [17, 19, 10] units and Tp
min
= 100
days (b) Project duration and costs for all variants of ni
and maximum needs ½nmax i = [23, 19, 10]. Leveling resources for Tpmin gives optimal
i = [17, 19, 10] (Chart 1a). Further individual PA and the
(minimal) quantities ½nmin
combination of results find all combinations of nmin i (Chart 1b) determine minimal
values [ni ] = [10, 9, 8] za ni ; same with the highest nij on which Pi : n12 = 10, n25 = 9 i
Min
n38 = 8 are used.

Remark 2. Start of the project 01 August 2016 and planned increase in initial unit
resource costs ½ci = [2,00; 2,50; 3,00] currency units for 3% since 01 October 2016 and
for an additional 5% at the beginning of 2017, determine the cost of the project
between CpMin = 28.777,60 currency units and CpMax = 29.543,33 currency units
(Chart 1b). PA is not done for Cp ; but for resources only.
0
Vector PA illustrates with ½ni = [23, 19, 10] i ½ni ¼ ½2; 1; 1, when five
solutions were determined (Table 2).
Table 2. Example of vector PA

l Functions ni n1 ¼ 23 þ 2l n2 ¼ 19 1l n3 ¼ 10 1l Tp Cp
0 ni (nmin
i )
23 (23) 19 (19) 10 (10) 100 28.777,60
1; 2 ni (nmin
i )
23 + 21 = 24 (23) 19 + (−1)1 = 18 (14) 10 + (−1)1 = 9 (8) 110 28.813,60
3; +/ Intolerably 23 + 23 = 29 19 + (−1)3 = 16 10 + (−1)3 = 7 < nmin
3
– –
−1; −3 ni (nmin
i )
23+2(−1) = 21 (17) 19 + (−1)(−1) = 20 (19) 10 + (−1)(−1) = 11 100 28.792,00
(10)
−4; −5 ni (nmin
i )
23 + 2(−4) = 15 (13) 19+(-1)(-4) = 23 (14) 10 + (−1)(−4) = 14 105 28.937,20
(10)
−6; −6 ni (nmin
i )
23 + 2(−6) = 11(10) 19 + (−1)(−6) = 25 (14) 10 + (−1)(−6) = 16 (8) 125 29.143,12
−7; −/ Intolerably 23 + 2(−7) = 9 < nmin
1
19 + (−1)(−7) = 26 10 + (−1)(−7) = 17 – –
5 Conclusion
Two forms of analysis in optimisation problems and two types of resources on the
project were discussed. The hypothetical example was illustrated in detail by a para-
metric analysis of the available amount of resources “Work” in minimising the project.
References
Journal Papers
1. Geoggrion, A.M., Nauss, R.: Parametric and post optimality analysis in integer linear
programming. Manage. Sci. 23(5), 453–466 (1977)
2. Kurtanović, O., Nikolić, N.: Parametarska analiza raspoloživih resursa za problem
minimizacije trajanja projekta. In: Zbornikradova, ICDQM 2016, 19. Međunarodnakonfer-
encija–Upravljanjekvalitetomipouzdanošću, Prijevor, 29–30 June 2016
3. Hendrickson, C.: Project management for construction, Department of Civil and Environ-
mental Engineering, Carnegie Mellon University, Pittsburgh (2008)
4. Nikolić, I., Krčevinac, S., et al.: Upravljanje pro-jektimai korišćenje CA-Super Project 2,
RTB Bor, Institut za bakar Bor (1998)
5. Nikolić, I.: Modeli nivelisanja resursa i planiranja nabavki sirovina na projektu. In: IIPP,
Naučno-stručni časopis Istraživanja i projektovanja za privredu, broj 18 – 2007, godina V,
Institut za istraživanja i projektovanja u privredi, Beograd, pp. 25–28 (2007)
6. Nikolić, I., Kurtanović, O.: Informacioni sistem procesa upravljanja projektom primenom
softvera MS Project. In: On-Line Zbornik radova, YuInfo 2013, XIX konferencija iz oblasti
informacionih i komunikacionih tehnologija, Kopaonik 3. do 6. marta (2013)
7. Kurtanović, O., Nikolić, N.: Determination of activity duration in dependence of ‘Work’
type resources engagement and project time minimization by applying MS Project software.
In: Proceedings of 4th International Conference LifeCycle Engineering and Management,
Belgrade, pp. 77–87, 27–28 June 2013
Books
8. Vanderbei, J.R.: Linear Programming: Foundations and Extensions, Department of
Operations Research and Financial Engineering. Princeton University, Princeton (1976)
9. Крчевинац, С. и др., Операциона истраживања, Факултет организационих
наука, Београд (2001)
10. Nikolić, I., Božičković, R.: Metode optimizacije u zadacima tipa transporta sa jednim i više
kriterijuma- Primena softvera WinQSB, Univerzitet Istočno Sarajevo, Saobraćajno-tethnički
fakultet, Doboj, Republika Srpska, BIH (2007)
11. Nikolić, I., Božilović, S.: Kvantitativne metode i modeli u menadžmentu- Odabrani problemi
u graditeljstvu (Primena softvera WinQSB i ExpertChoice), Fakultet za graditeljski
menadžment, Fakultet za preduzetnički biznis, Beograd (2009)
12. Goordpasture, C.J.: Quantitative Methods in Project Management. J. Ross Publishing, Inc.,
Fort Lauderdale (2004)
New Technologies in Education
at “Džemal Bijedić” University in Mostar
Alim Abazović, Dragi Tiro ✉ , and Fuad Ćatović

( )
“Dzemal Bijedic” University of Mostar, 88000 Mostar, Bosnia and Herzegovina

{alim.abazovic,dragi.tiro,fuad.catovic}@unmo.ba
Abstract. Information and communication technologies (ICT) have become

unavoidable in educational processes. They are used in higher education to a
significant extent today in a complex Bosnian-Herzegovinian reality. This paper
presents an overview of the development and experience in applying ICT in
education with suggestions for further development and improvement at the
Dzemal Bijedic University of Mostar.
Keywords: ICT · Higher education · E-learning · Distance learning
1 Introduction
Information technologies have recently changed human life; the ways we work,
communicate, spend free time, and also the ways of learning [1]. E-learning is a way
to learn using educational materials in the electronic (digital) form using informa‐
tion technology, such as a computer, computer networks and all other infrastructure
needed for educators to provide the educational materials in electronic form to
students. FTP protocol is the most commonly used to upload and download mate‐
rials, as well as e-mail.
Distance learning (d-Learning, DL) existed even before the advent of computers and
information technology. Mail was mostly used. Today, this type of learning uses modern
information technology.
M-learning is a mobile learning environment for using educational materials by
smartphones, the other necessary ICT infrastructure and wireless connection to the
Internet. Mobile learning is a connection of e-learning and distance learning and it is
described as learning “anytime” and “anywhere” [4].
Blended learning refers to learning models that combine traditional classroom prac‐
tice with e-learning solutions [5].
2 Historical Context of Modern Technologies Development

in Educational Process at the Džemal Bijedić Universiy of Mostar
Modern high education in Mostar has a tradition of over six decades long [2]. Informa‐
tion technologies were used At the Dzemal Bijedic University of Mostar before the last

https://doi.org/10.1007/978-3-319-90893-9_70
598 A. Abazović et al.
war (1992–1995), but to a very limited extent, for example there was a computer center
with a great number of personal computers at the Faculty of Mechanical Engineering.
During the war, the University lost its infrastructure (buildings, equipment, etc.), and
after the war it was located at the former military barracks Sjevernilogor, where a great
number of facilities were used. The University campus is located here today.
Writing about new technologies in education is difficult without a historical overview
of the most significant facts related to the development of the University and its capaci‐
ties. We need a brief explanation of the context in which the University survived and
acted in the near past, from the perspective of modern technologies in education.
All the property of the University has become inaccessible to teachers and students,
as well as equipment, libraries, buildings for student accommodation and all other
capacities of the University in 1993. The University reopened In November of 1993.
The 1993/94 academic year begins with the creation of minimum working conditions
(one classroom and one inconvenient office). The teaching process is organized in
Jablanica and Konjic [3].
At the end of the war in 1995, the first IT equipment was purchased and the University
has the computer center with about 10 PCs and uses Mostar Training Center, which was
located outside the campus. Mostar Training Center was used for IT training of profes‐
sors, assistants and students.
In the following years, the University has built several faculty buildings, which are
connected by optical cables and IT equipment has been procured. Fifteen years ago, the
University has made its computer network connected to the Internet. This created the
basic conditions for the application of ICT in teaching on subjects that are not directly
related to computer science.
In the initial stages of ICT development in teaching, since 2002, some of the existing
services such as FTP, e-mail, etc. have been used. At that time, an online education
platform was launched to support distance learning and material distribution.
Students from the Faculty of Information Technology (FIT) have been enabled to
enter the DL mode from the beginning of this study. This set up many requirements for
the development team, and the main of them was support to the processes carried out
within the student service and teaching.
The distance learning system is actively applied at the FIT since the academic
2001/2002 year, which makes it a beginner and a long-standing concept of distance
learning, first in Bosnia and Herzegovina, but also in the region.
The project “Introduction of e-Learning and Distance Learning System at the Dzemal
Bijedic University of Mostar” was conducted from 2005 to 2008 and was funded by the
International Forum Bosnia. Within this project, for the first time, web sites were created
for teachers to upload their materials, and students to download them. Participants from
Agromediterranean Faculty, Civil Engineering Faculty, Mechanical Engineering
Faculty and Teacher Training Faculty participated in the project. The project showed
all the advantages of this teaching method. The students were pleased that it was much
easier to get the teaching materials. It has been noted that continuous improvement of
infrastructure is needed, teacher training for this learning system. It has also been noted
that during the implementation of the project, some teachers and associates did not want
to make available certain teaching materials because of copyright.
New Technologies in Education at “Džemal Bijedić” University in Mostar 599
Within the framework of capacity building, the University introduced the optical
connection and wireless technology to the campus. Dzemal Bijedic Universiy of Mostar
strengthens human and infrastructural capacities in the period from 2008 to 2011.
By engaging in various projects of domestic and international character the Univer‐
sity gets significant equipment, is an active participant in the reform processes: quality,
curriculum reform, mobility and international cooperation, etc.
This period was marked by strategic planning, the adoption of the document
“Basis of the University development strategy”, which served as a guide for the
development of the University for many years. Further work continues moderniza‐
tion of ICT equipment.
The projector and computer with the accompanying broadband optical and wireless
internet connection have become standard equipment in classrooms. It is actively used
in teaching-scientific process and scientific research.
Since 2011, the University has already built capacities that enable and support up-
to-date technologies for modern teaching. Dzemal Bijedic Universiy of Mostar has a
unique computer network that connects 16 university facilities. Most objects are
connected by 1 Gbps optical cables, and all are connected by wireless technology. All
objects have a hub for connectivity. They are equipped with manageable optical
switches. The central university hub has the equipment for connecting these facilities
and connection to the Internet. The connection speed is up to 10/1 Mbps (Flat ADSL).
Figure 1 shows the scheme of physical link of objects.
Fig. 1. Scheme of physical link of objects.

The resources relevant to new technologies in education, from the aspect of support,
the University has: servers (mail servers for employees and students with access through
HTTP, SSL, POP3, IMAP4 and SMTP protocols, web servers, in the DMZ security zone
based on ASP.Net/MS SQL platform through a unique Customer Relationship Manage‐
ment (CMS), which allows controlled content upload to the university and faculties web
site, as well as various other services (newsletter, forum, integration of other modules,
etc.). Figure 2 shows the scheme of logical links of faculties and other parts of the
network.
Fig. 2. Scheme of logical links of faculties and other parts of the network
University uses also php pages based on the “open source” application, as well as
databases (human resources database, student database, database for projects). All web
servers and database servers are necessary for operation of the specified CMS. They are
located on virtual servers.
The education support system at the Dzemal Bijedic Universiy of Mostar is devel‐
oping and constantly improving. In addition to modernizing existing capacities and
building new mechanisms for supporting education and learning, the development
directions of ICT from 2014 aim to create mechanisms that will:
• establish, measure, integrate and monitor key performance indicators within the
information system,
• explore the possibilities of full utilization of existing information system in moni‐
toring and analysis of various indicators of success of studying, i.e. its use as a tool
for formulating strategic decisions of the University management
• make plans for improving the information system and adopt the principles of invest‐
ment in information resources,
• analyze and, if necessary, improve the current capacity of Internet access and connec‐
tivity to the academic network,
• develop new modules for student service software, and integrate a human resources
management module with a student service application,
• to carry out the integration of the module for managing projects and scientific
research,
• integrate the module to create and monitor teaching schedules,
• create and execute module integration for OnLine exam application for students of
all cycles at all faculties
• integrate the module for keeping the register of diplomas and diploma supplements,
• make an automated module for creation of necessary and defined statistical report,
• reactivate the student mail server with the domain @ student.unmo.ba
• implement the UnSIS application that allows students access data regarding study
process and statutory data
• implement the document management system at the University,
• to constantly educate users about new services and changes in ICT.
3 New Technologies in Educational Process
Today, all the study programs of the Dzemal Bijedic University of Mostar use ICT in a
greater or lesser extent in the teaching process, depending on the study program and the
subject itself. The projector and computer (desktop or laptop) are used in almost all
teaching subjects. Smart boards and smart TVs are also used in several faculties (Faculty
of Humanities, Civil Engineering, …). In some faculties, such as Faculty of Law, there
is a practice of online week before the exams where students can consult teachers online
and get a quick answer to their question. Sending and submission of seminar papers,
program assignments, essays and similar works using e-mail has become an everyday
practice at all faculties. Web site communication has also become a practice, so students
do not have to come to the University, but only quickly check the appropriate website.
We should emphasize that most IT equipment is purchased through projects, mostly
TEMPUS and Erasmus+. On study programs, teachers place materials that students
download using file sharing systems, as is the case with the Faculty of Mechanical
Engineering, while in some studies teachers give materials to a student representative
of the year who distributes these materials to students using closed groups within social
networks. Lately, students are frequently using social media groups where they
exchange materials.
The Faculty of Information Technology (FIT) was the first to start with the intro‐
duction of information technology in teaching. FIT was the first to acquire its servers
and make an information system for the entire faculty, day to day work, and also for e-
learning and d-learning. FIT was also first at the University to introduce a distance
learning system, which is with some upgrades still in use today. The distance learning
platform developed at the Faculty of Information Technology is today used by several
higher education institutions in Bosnia and Herzegovina.
The introduction of the distance learning system of education has most attracted
students from the distant parts of Bosnia and Herzegovina and other countries in the
region, as well as those already employed in the ICT sector.
Internal research has shown, and the labor market has recognized, students who
completed the DL concept in the level of acquired competences are no different from
those who completed their studies in the status of a full-time student. Still it is obvious
that this type of study in the concrete case should be further improved and modernized.
The online education platform, which has experienced several versions at FIT and
is still in use, in principle supports a form of study that involves the use of ICT to access
teaching materials, registering, taking and passing exams, etc.
A DL study allows the student to access teaching materials from anywhere and at
any time. DL students receive access information in order to be able to use the DLWMS
(Distance Learning Web Management System) system.
This distance learning system greatly facilitates the teaching process to both teaching
staff and students, as well as administration.
This way of studying also implies the installation of teaching materials with lectures
and exercises on the DL system, which is accessed through a web browser.
The DLWMS software solution also provides online exam application, chat and
forum for students and teaching staff. The development of software for this platform
was initially started as a graduate work of a student who grew into the software that we
have today. Since its first application, this system has been upgraded to date and it has
been added new functionality to monitor and improve the teaching process. Adding new
reports within reporting modules represents common steps in this software’s lifecycle.
The modernization and improvement of the online education platform was condi‐
tioned by the rapid development of information technologies, the new functionalities
that needed to be supported and the change of business processes and rules in line with
internal and external opportunities.
An increasing number of users have also generated a list of new requests that needs
to be implemented.
In addition to improving the system in the administrative part: setting the condi‐
tionality of exam periods according to all criteria, solving the issue of registration fees,
checking and approving student images, setting a timetable, electronic document
processing (confirmation), support for work on several institutions (faculties), improved
security and monitoring of system work, improved reporting system, changes also
include solutions for problems such as: poor reporting system, inability to control data
and manual issuance of certificates to students, which has proved impractical and time-
consuming due to the large number of students. The existing system did not offer an
analysis of the large amount of accumulated data, and these data represented a problem
when it comes to large number of students using the system. However, key improve‐
ments have benefits for teaching staff and students and greatly contribute to the facili‐
tation of everyday work of teaching staff. In addition to redesigning the user interface,
there was a need for some of the following functionalities: creating certificate requests
and issuing them, recording ECTS points, online submission and evaluation of seminar
and other papers, integration with Youtube channel, advanced version of online tests,
etc. With constant improvements, the database is released from redundant student’s and
other data. From the perspective of a student, each new version also brings improve‐
ments. After internal analysis and support by students and teaching staff, the upgrade of
the existing student part of the system is reflected in completing and submitting requests
for certificates, evaluating educational materials on each subject, a detailed review of
success (in addition to the final grade, the overall evaluation of the subject is also
consolidated), an overview of paid tuition fees, an application for an exam that is not
necessarily followed by teacher constraints, selecting only those subjects that are of key
interest to students to actively monitor their materials and information, using improved
material search and notice, taking over graduation and final papers from other students,
attendance, etc. Finally, functionalities that facilitate the work and teaching process to
teaching staff are reflected in the following functionalities added to the latest version of
the system: monitoring the evaluation of published material by students, defining the
evaluation of the subject, performing a final grade on subjects, attendance records,
publication of final and graduate papers, review of their own schedule, improved
reporting. The current version of the system is adapted to work on several different
institutions with a modified and customizable graphic design, which is one of its key
features. Upgrading the administrative module, the student module, and the teaching
staff module as the three basic elements of the system include the transition to the new
ASP.NET and SQL Server technology.
4 Conclusion
The development and proper application of the e-learning system is driving force for
revolutionary changes in education and defining clear vision of the impact of technology
on the development of education in the future. Popularity of open platforms for mass
online education, so-called MOOC (Massive Open Online Courses) today also increases
the opportunity for students from all over the world to jointly participate in acquiring
knowledge and skills that are needed in their midst, offered by the world’s most presti‐
gious institutions.
The application of ICT technology will develop in order to improve the assumptions
for the smooth interaction of students and teaching staff in real time from anywhere in
Bosnia and Herzegovina as well as in the World.
It is obvious that the role of teachers and students is changing. The differences
between various forms of study (regular, extraordinary, inter-university, interstate,
distance learning, etc.) are changing or disappearing at all.
M-learning, mobile phone learning where all content can be used and that will
support all services from real-time activities to access the most common data is a near
future.
The system is as good as its weakest link. In the sea of new technologies, each of
which can be used in some way with or without certain modifications in education,
people will remain a key factor. Developing the awareness of all involved in the higher
education process on the application of ICT should be a priority, because it makes it
easier for them to work or study. Their education on the application of new technologies
and the desire to use it must be clear and permanent.
References
1. Vrhovac, V., Orošnjak, M., et al.: Nove Tehnologije Učenja U Visokom Obrazovanju. In:
XXIII skupTrendovirazvoja, Zlatibor (2017)
2. Ćatović, F., Abazović, A.: ‘Univerzitet ‘Džemal Bijedić’ u Mostaru kao generator razvoja juga
Bosne i Hercegovine, Zbornik radova sa naučnog skupa: Hercegovački naučnici i tradicija
istraživanja u Hercegovini, Mostar, Decembar 2015
3. Grupa autora: Pregled predavanja za 1999–2000. akademsku godinu, Univerzitet “Džemal
Bijedić” u Mostaru, Mostar (1999)
4. Haddad, W., Jurich, S.: ICT for education: potential and potency. Technologies for education:
potentials, parameters, and prospects. UNESCO and the Academy for Educational
Development (2002)
5. Tinio, V.: ICT in Education UNDP Asia-Pacific Development Information Programme,
UNDP-APDIP (2003)
Author Index
A C
Abazović, Alim, 597 Čabaravdić, Malik, 354
Ajvazović, Karlo, 452 Čančar, Sanela, 354
Aldžić, Aldina, 561 Ćatović, Fuad, 585, 597
Aldžić, Asmir, 568, 575 Ćećez, Marko, 585
Alexey, Yakimov, 121 Ćehajić, Adnan, 552, 568
Alibabić, Vildana, 498 Ćuković, Saša M., 271
Arshad, Muhammad, 317 Cvrk, Ramzija, 507
D
B Dacić, Lejla, 590
Babič, Matej, 227 Dašić, Predrag, 386
Balashev, Ivan, 80 Dedić, Kanita, 575
Balashev, Radoslav, 80 Dedić, Luka, 345
Banjanovic-Mehmedovic, Lejla, 309 Dedić, Remzo, 211
Baručija, Anel, 57 Delalić, Zemira, 491
Batinić, Krsto, 397 Delprete, Cristiana, 413
Begić, Munevera, 513 Djukanovic, Milena, 170, 180
Begić, Sabina, 469, 480 Drev, Darko, 522
Begović, Azra, 507 Drstvenšek, Igor, 33
Belegu, Majlinnda, 238 Dubinović-Rekić, Amela, 575
Bešlagić, Ernad, 41
Bičo Ćar, Mirha, 247, 254 F
Bisha, Arian, 413 Fakić, Belma, 113
Blagojević, Jovana, 443 Furdić, Martina, 359
Boshnjaku, Gerti, 238
Boskovic, Vuk, 170 G
Brezočnik, Lucija, 289 Ganić, Amir, 513
Brezočnik, Miran, 33 Gavrilovic, Branislav, 429
Bubanja, Milos, 180 Ghionea, Ionuţ G., 271
Buljan, Stipo, 144 Golubović, Dušan, 397
Bundalo, Zoran, 429 Gotlih, Janez, 33, 156
Burić, Adisa, 113 Gotlih, Karl, 33, 156
© Springer International Publishing AG, part of Springer Nature 2019 605

https://doi.org/10.1007/978-3-319-90893-9
606 Author Index
Gregurić, Martin, 332 Markus, Marina Mijanovic, 180

Grujicic, Rade, 170 Martinčič, Sanda Ipšič, 227
Metović, Sadjit, 377
H Mićić, Vladan, 469, 480
Haskić, Erzad, 128 Milosavljević, Milan M., 542
Hodžić, Nihad, 377 Milosavljević, Milutin M., 542
Horoz, Edib, 113 Mirza, Riaz, 3
Hozdić, Elvis, 262 Muhamedagić, Sulejman, 72
Husak, Ermin, 107, 128 Mujagić, Derviš, 72
Husaković, Anel, 354 Mujić, Ibrahim, 498
Mulahusić, Admir, 452
I Mumović, Marko, 405
Imamović, Aida, 72
Isić, Safet, 101, 135 N
Ivanjko, Edouard, 332 Nanić, Husein, 552, 561
Nezirić, Emir, 135
J
Jarović-Bajramović, Narcisa, 86 O
Jelačić, Zlata, 196, 211 Ognjanović, Ivana, 219
Jokanović, Mirjana, 50 Omer, Salah-Eldien, 163
Jovanovic Dolecek, Gordana, 423 Orašanin, Goran, 397, 443
Jovanovic, Jovana, 529 Orašćanin, Melisa, 498
Jukić, Huska, 561, 568, 575 Oruč, Mirsada, 72
Jukić, Mahmut, 459 Osmić, Selma, 480
Jukic, Zlatan, 317
Jurković, Zoran, 262 P
Juul, Nils V., 507 Pecić, Ljiljana, 542
Pehar, Blago, 369
K Pehlić, Ekrem, 552, 561
Karabegović, Edina, 101, 107 Penev, Mariel, 80
Karabegović, Isak, 3, 135, 227, 423 Petrović, Zvonko, 386
Karahmet, Enver, 513 Pikula, Boran, 196
Karner, Timi, 33, 156 Podgorelec, Vili, 289
Kazagić, Anes, 377 Pop, Bianca Alina, 279
Knezović, Nikola, 65 Porča, Almir, 41
Koprivica, Aleksandra, 50
Košarac, Aleksandar, 443 R
Kovacevic, Ahmed, 18 Radunovic, Luka, 170
Kovalevska, Olena, 202 Rane, Sham, 18
Kovalevskyy, Sergiy, 202 Razboršek, Boštjan, 156
Krzyk, Mario, 522 Romić, Dejan, 443
Kurtanović, Omer, 590 Rupar, Miljan, 211
L S
Lemeš, Samir, 57 Šahić, Elvir, 247, 254
Liubov, Bovnegra, 121 Šahinagić-Isović, Merima, 585
Šapčanin, Aida, 552
M Schatten, Markus, 338
Mahmić, Mehmed, 107 Seferović, Raif, 86
Mandžuka, Sadko, 299, 332, 338, 345 Selmani, Erjon, 413
Manjenčić, Darko, 469 Šemić, Edin, 101
Marić, Snježana, 507 Šendelj, Ramo, 219
Author Index 607
Shaba, Fatjon, 238 Tuno, Nedim, 452

Šibalić, Nikola, 50 Turmanidze, Raul, 202
Simić, Stojan, 397
Šišić, Ifet, 459 V
Skender, Azra, 498 Varga, Gyula, 227
Škorput, Pero, 338, 359 Vladimir, Tonkonogyi, 121
Škrinjar, Jasmina Pašagić, 359 Vojić, Samir, 191
Stevovic, Svetlana, 529 Voloder, Avdo, 135, 149
Stosic, Nikola, 18 Vučina, Adisa, 211
Stoyanov, Ivan, 80 Vujić, Miroslav, 345
Stupar, Savo, 247, 254 Vujošević, Vuk, 405
Šunjić, Darko, 144 Vujovic, Mihailo, 180
Vukčević, Milan, 50
T
Taylor, William R., 271
Terzić, Edin, 86 Z
Tiro, Dragi, 597 Zaimović-Uzunović, Nermina, 41
Titu, Mihail Aurel, 279 Zalihić, Suad, 369
Titu, Stefan, 279 Zanaj, Blerina, 238
Tomović, Aleksandar, 405 Živković, Milutin, 386, 542
Tomović, Radoslav, 405 Zlomušica, Elvir, 369
Topić, Angela, 65 Zukic, Mirzet, 309
Topoljak, Jusuf, 452 Zulić, Melisa, 568

Isak Karabegović - New Technologies, Development and Application-Springer International Publishing (2019)

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Isak Karabegović - New Technologies, Development and Application-Springer International Publishing (2019)

Uploaded by

Copyright:

Available Formats

Lecture Notes in Networks and Systems 42

Isak Karabegović Editor

More information about this series at http://www.springer.com/series/15179

ISSN 2367-3370 ISSN 2367-3389 (electronic)

© Springer International Publishing AG, part of Springer Nature 2019

Printed on acid-free paper

New Technologies in Mechanical Engineering, Metallurgy,

Strength and Deformation Calculation of Flat O-Springs . . . . . . . . . . . . 80

Vibration Analysis of Motorcycle Handles . . . . . . . . . . . . . . . . . . . . . . . 196

Computer Science, Information and Communication

Intelligent Transport Systems, Logistics, Trafﬁc Control

New Technologies in the Field Energy: Renewable Energy,

Gas Escape from Combustion Chamber to Crankcase, Analysis

New Technologies in Agriculture and Ecology, Chemical Processes

The Impact of Water on PVC Floor Coverings . . . . . . . . . . . . . . . . . . . 522

New Technologies in Civil Engineering, Education, Control Quality

Alim Abazović “Dzemal Bijedic” University of Mostar, Mostar, Bosnia and

Majlinnda Belegu Agricultural University of Tirana, Tirana, Albania

Luka Dedić University of Zagreb, Zagreb, Croatia

Safet Isić Faculty of Mechanical Engineering, University “Džemal Bijedić”

Aleksandar Košarac University of East Sarajevo, East Sarajevo, Bosnia and

Ivana Ognjanović University of Donja Gorica, Podgorica, Montenegro

Mihailo Vujovic University of Montenegro, Podgorica, Montenegro

Isak Karabegović1 ✉ and Riaz Mirza2

Abstract. The development of robotic technology, owing to the advancement

Keywords: Industry · Process · Robot · Welding · Arc welding · Spot welding

© Springer International Publishing AG, part of Springer Nature 2019

2 The Representation of Industrial Robots

In order to depict the representation of industrial robots in production processes world‐

Number of units Number of units

3 Automation of Welding Production Processes by Use of Industrial

The tendency of representation of industrial robots in automotive industry is growing

a – Representation by industry b – Representation in welding compared to

Fig. 6. Percentage of representation of industrial robots in 2016 in industrial branches and

Number of units 65.004

Numbeer of units Numbeer of units 28.4331.000

a – Vehiclle production worldwide b – Vehicle production by country

Numberr of units 6.805

Numberr of units Numbeer of units 10

As shown in Fig. 11(a), China demonstrates growing tendency of use of industrial

• The tendency of annual representation of industrial robots in the automation of two

Ahmed Kovacevic1(&), Sham Rane2, and Nikola Stosic1

Keywords: Screw compressor Computational Fluid Dynamics

© Springer International Publishing AG, part of Springer Nature 2019

model for predicting the oil

2 Multi-phase Modelling in CFD

3 Grid Generation for Oil Injected Screw Rotors

3.2 Applying the Elliptic PDE Solver to Improve Algebraic Grid

3.3 Smoothing the Transition of the Boundary Between the Rotor

O Grid TFI Mesh

The smooth rack obtained

3.4 Improvements in Grid Quality

4 Case Study CFD Model of Oil Injected Compressor

Table 1. Operating conditions for comparrison of CFD and measurements

Table 2. Mesh statistics and quality report

4.1 Integral Parameters

4.2 Oil Distribution

4.3 Temperature Distribution

5 Development Trends and Future Work

Janez Gotlih ✉ , Miran Brezočnik, Igor Drstvenšek,

Timi Karner, and Karl Gotlih

Abstract. A case study to determine the inﬂuence of injection molding process

Keywords: Injection molding · Quality prediction · Parameter study

To investigate the inﬂuence of injection molding (IM) process parameters on part