You are on page 1of 11

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/260385935

E-manufacturing-fundamentals, requirements and expected impacts

Article · January 2003

CITATIONS READS

5 357

2 authors:

Muammer Koç Jay Lee


Hamad bin Khalifa University Foxconn Technology Group (On Leave from Univ. of Cincinnati)
213 PUBLICATIONS   3,553 CITATIONS    434 PUBLICATIONS   13,655 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Industry 4.0 Interview CCTV View project

Boiling Heat Transfer Enhancement of Colloidal Dispersion and Hybrid Nano-coatings View project

All content following this page was uploaded by Jay Lee on 12 June 2014.

The user has requested enhancement of the downloaded file.


Proceedings of the 2nd International Conference on Responsive Manufacturing (ICRM 2002)
June 24-28, 2002- Gaziantep University, G. Antep, Turkey

e-MANUFACTURING- FUNDAMENTALS, REQUIREMENTS AND


EXPECTED IMPACTS
Muammer Koç and Jay Lee
NSF I/UCRC Center for Intelligent Maintenance Systems (IMS), The University of Wisconsin-Milwaukee,
Milwaukee, WI 53211, USA
koc@uwm.edu

ABSTRACT Planning (ERP), and Manufacturing Execution


Systems (MES). However, e-business without a
This paper outlines an e-manufacturing system profound manufacturing base lacks delivering what
architecture giving details of its required features and customers are asking. Integration of the front
elements after summarizing the recent trends and elements the e-Business has exerted an increaing
needs in the business and manufacturing industry in pressure on the manufacturing industry.
particular. e-Manufacturing is a recent concept being Manufacturers are now under a tremendous pressure
developed aimed at complete integration of all the to improve their responsiveness and efficiency in
elements of a business- including suppliers, customer terms of product development, operations, resource
service network, manufacturing enterprise and plant utilization with a transparent visibility of production
floor assets with connectivity and intelligence and quality control. Lead times must be cut short to
brought by the internet, web-enabling, tether-free their extreme extend to meet need the changing
technologies and computational tools to meet the demands of customers in different regions of the
demands of e-business/e-commerce practices that world. Products are required to be make-to-order
gained great acceptance and momentum over the last with no or minimum inventory, requiring (a) an
decade. e-Manufacturing offers the manufacturers a efficient information flow between customers and
position where they can easily re-configure the manufacturing (i.e. plant floor and suppliers), (b) a
manufacturing capability with agility to respond tight control between customers and manufacturing
quickly for changing demands and opportunities and (c) near-zero downtime of the plant floor assets
in the market by possessing predictive intelligence on as depicted in Figure 2.
the plant floor.
Global market / economy
KEY WORDS Competition

Business
e-Manufacturing, e-maintenance, predictive Automation Predictive Intelligence
(e-Business) • Product Realization
maintenance, e-business • M fg. Operations ??
Velocity • Delivery
• Services

1. INTRODUCTION Outsourcing Products/


Machines ?
Mfg.
Processes

In today’s continuously changing and Government / Regional Regulations


Systems

challenging global market, customized products, fast Loss of Skills / Experience


Aging Workforce
delivery, on-time service, reliability and transparent Challenges
• Short Decision-Making Time & Cannot Afford Downtime
processes are in an increasing demand in addition to • High Demands in Responsive Supply-Chain
• Sell Life-Cycle Reliability and Support as Part of Product

the given near-zero defect quality and continuous Figure 1. Transformation of industry and challenges
cost reduction requirements, Figure 1. Customers, introduced with this transformation
empowered with the internet, select the products not
only based on their brands and sales cost, but also S2
S1
Suppliers
their reliability, efficiency and after-market service Sn
Distribution

options. Consequently, the selection of products is


Marketing

C1
Design & Inventory /
Manufacturing Customers C n
Development Stocks
more and more based on their overall life cycle C2

value. Customer’s access to more vendors and their •



Predictive Intelligence
Seamless Integration

products through Internet unlimit their power of • Responsiveness

purchase, and allow them to ask for better products at


Distribution
Marketing

C1
a less cost with extended service options and Design &
Development
e -Manufacturing Customers C n

guaranteed life cycle performance. C2

S1
e-Business has been evolved during the past S2 Suppliers
Sn
decade to satisfy such needs of customers through
development and improvement of innovative Figure 2. Expected transformation of manufacturing
business enterprise systems such as e-procurement, enterprise. Changes in market conditions force
Supply Chain Management (SCM), Customer manufacturing industry to be agile, responsive and
Relations Management (CRM), Enterprise Resource intelligent with predictive capabilities [after 1, P. Zipkin]
Proceedings of the 2nd International Conference on Responsive Manufacturing (ICRM 2002)
June 24-28, 2002- Gaziantep University, G. Antep, Turkey

A company’s ability to satisfy the customer transformation, prediction, optimization, and


orders and needs primarily depends on their synchronization.
capabilities for fast product realization, uninterrupted e-Manufacturing is a recent concept being
production with defect free products and on-time developed aimed at complete integration of all the
service not on their warehouse capacity, or their elements of a business- including suppliers, customer
quickness to accept and process the orders over the service network, manufacturing enterprise and plant
internet. Lately many companies started outsourcing floor assets with connectivity and intelligence
most of their production activities to (a) focus on brought by the internet, web-enabling, tether-free
marketing and servicing to achieve customer technologies and computational tools to meet the
satisfactions, and increase their profit margins, and demands of e-business/e-commerce practices that
(b) overcome the demands of changing market and gained great acceptance and momentum over the last
customers by leveraging the capacity and capabilities decade, Figure 3. e-Manufacturing includes the
of partners and suppliers. However, it has been seen ability to monitor the plant floor assets, predict the
that this was not an ultimate solution as variation of product quality and performance loss of
manufacturers were then dependent on the any equipment for dynamic rescheduling of
capabilities and responsiveness of their suppliers, production and maintenance operations, and
Figures 1 and 2. In any case that these suppliers can synchronize with related business services to achieve
not achieve the required responsiveness, customer a seamless integration between manufacturing and
satisfaction would be so are the future opportunities higher level enterprise systems. e-Manufacturing
of that business offers the manufacturers a position where they can
Hence, suppliers’ manufacturing easily re-configure the manufacturing capability with
responsiveness and capabilities need to be agility to respond quickly for changing demands and
understood, improved and integrated fully to provide opportunities in the market by possessing predictive
ultimate market solutions and customer satisfaction. intelligence on the plant floor. e-Manufacturing
Since there are broken chains between existing solutions would execute customer orders on the plant
enterprise systems such as ERP, MES, and SCM and floor and across the supply chain, bringing
plant floor assets, real-time production control unprecedented levels of speed, flexibility and
capabilities cannot be achieved. For instance, visibility to the production process reducing
Manufacturing Execution Systems (MES) reside inventory, excess capacity and uncertainty. It also
between plant floor operations and ERP, and are provides efficient configurable information
supposed to perform functions such as allocation and exchanges among manufacturing units, customer
scheduling of resources and production and relationship management (CRM) systems and supply
management of quality, maintenance, process and chain management (SCM) systems, Figure 3.
labor. However, today’s MES still perform on a
reactive mode as information flow from the plant •

Green processes & products
Traceability
floor is handled by the operators and engineers in an 1 •

Scalability
Responsiveness 1

unautomated fashion. In order to meet the demands 2 Suppliers


• M RO
• Standardization
Customers 2
• M fg partners &
of the new economy and global changes, 3
• Raw material P lant 1 P lant 2 P lant n Distributors
• C omponent • Distributors 3
manufacturers need a comprehensive solution that: • Design,
• R&D
SCM e-Manufacturing CRM • E nd C ustomers
• E xchanges
(a) connects customers to the plant floor in real-time, n • etc
n

and provides visibility, (b) empowers plant floor ERP- Enterprise Resource Planning
Sales & Distr., Work orders, Materials, Prod. Plan
operators perform proactive measures promptly • E nv ironmental
requirements
Work Flow, Plants Maintena nce, Quality, Human Reso urces;
Order Status, WIP Status, Quality Data, Customer Orders

based on customer orders, (c) provides reliability, • International, regional


and gov ernmental MES- Manufacturing Execution Systems
Work Instructions, Control Parameters, Prod. Scheduli ng,
dependability and a common platform for all •
regulations
Labor reg ulations
Efficiency, Maintenance; Resource Status, WIP,

Workforce needs
operations, plants, and (d) offers predictability of •
• etc
Control Systems- PLCs, Controllers, etc.
Equi pment, devices, people, processes, Sensors, I/O;
Status and performance of devices, Operation & Work status,
variations and changes on the assets (i.e., equipment, Process values,

devices, people, processes) through use of


computational tools and widely available internet
networking technologies in a secure way as presented Figure 3: e-Manufacturing at the intersection of three
dimensional needs and expectations. Integration of e-
in Figure 3.
Manufacturing into e-Business systems to meet the
increasing demands through tightly coupled SCM, ERP
2. E-MANUFACTURING- RATIONALE and CRM systems as well as environmental and labor
AND DEFINITION regulations and awareness.
Change the information on your document
to contain e-Manufacturing is a system that enables While marketing and supply chain
the manufacturing operations highly compatible with integration efforts of a business utilize the latest web
the remaining elements of the business through the enabled technologies in a successful manner, the
use of internet, tether-free technology (i.e. wireless, backbone of the business- manufacturing- undergoes
web, etc.) and predictive technologies to successfully an increasing pressure to meet the demands using
perform the functional objectives of an enterprise, i.e. their conventional techniques and skills. Accepting
Proceedings of the 2nd International Conference on Responsive Manufacturing (ICRM 2002)
June 24-28, 2002- Gaziantep University, G. Antep, Turkey

the orders through the internet is the easy part of the semiconductor industry (i.e. chip making companies
business, fulfilling the requirements of these orders is and foundry), the power generation industry (i.e.
the tough part. The velocity injected into the power utilities companies), the transportation
business through e-commerce and Internet, leaves industry (i.e. trucking, airlines), the medical
very little room in manufacturing for downtime, equipment industry, and the building system industry
production loss, and delay in realizing newer (i.e. elevators, HVACs), have been strategically
products. In order to leverage the web enabled benchmarking smart predictive maintenance and
marketing side of the business, manufacturers have to customer service practices globally to maintain their
find new methods and techniques to respond to this manufacturing productivity and customer’s intimacy
demand imposed by the e-Business improving the for a better competitive position. For example, the
following approaches and concepts: (a) build-to- latest Pratt & Whitney jet engines have an array of
order, (b) short lead time, (c) flexible and visible sensors and on-board diagnostics that are capable of
production process, and (d) smart asset management. generating information with sufficient detail to allow
Customers’ have varying needs requiring the ground systems to forecast repair needs months in
customized solutions and products. Therefore, advance. This helps the airlines schedule engine
satisfying their needs demands fast response in terms removals so that the cost of downtime is balanced
of design, production and delivery. The system also against the cost of overhaul. The on-board systems,
has to be flexible to adapt significant changes in the in concert with the ground based analysis capabilities,
design, materials, process, etc. Manufacturers have also reduce costly unscheduled events such as delays,
to also provide visibility to their customers so that cancellations, diversion and in-flight shutdowns.
they follow the status of their order through supply,
production and delivery stages. Not only customers 1980 1990 2000
2010

but also plant and maintenance managers, IT and Products


Product
financial officers, etc. of a company would be Product
Intelligent
Mechatronics
That
That Learn,
Grow, Reconfigure,
Thinks and Links
interested in such a visibility as this would help them Focus (data & control
intelligence)
(information &
and Sustain
(knowledge &
Computer intelligence)
to organize and plan for order changes, resource e - Intelligence)

status, equipment failures, scheduling of Factory Enterprise


E-Mfg &
Business Automation
maintenance, etc. There should be a continuous and Manufacturing
Focus
Automation
(flexibility)
Integration
(agility) (velocity )
seamless flow of information in e-Manufacturing.
Real-time access of information at various and Near-Zero-Downtime
Six-Sigma Customer Solutions
remote locations would increase the overall Quality
SPC & TQM
for Manufacturing For
Business Process and Total Asset Optimization
efficiency, Figure 3. Every member of chain- Focus Process
(factory) (enterprise) (customers)

suppliers, distributors, logistic and technology


providers, manufacturers at all levels- should possess Figure 4: Evolution of product design, manufacturing, and
the capability, willingness and technology to quality expectations
implement the elements of e-Manufacturing. If any
member in the chain lacks this requirement, the Today’s automotive industry is also looking
whole chain or the system will not function as for ways to embrace telematics for getting automatic
desired [2, 3]. roadside assistant and route guidance into car. GM
Figure 4 illustrates the evolution of product has set a goal for its OnStar telematics service of one
and manufacturing innovation and the future trend million subscribers by the end of 2000. By 2005,
from 1980s to today and future. Nowadays product GM plans to have all of its cars equipped with the
focus has been changed from information and hardware and software components for OnStar with
computer intelligence to knowledge and e- enhanced features [5]. For the semiconductor
intelligence as a need to fulfill the fickle customer manufacturing industry, one of the major thrusts is e-
demands. Manufacturing focus is shifting from diagnostics. e-Diagnostics is the ability to remotely
factory automation to business automation to trigger service equipment and provide application support
the product focus. Quality focus is addressing e- such as correcting chip-manufacturing errors via the
service for customer solutions and total asset internet. Chip manufacturers use remote diagnostics
optimization where as in the past it was mentioned as software instead of having an engineer come out to a
TQM and six sigma for business process as the site to adjust chip-making equipment. As chips get
manufacturing is not beyond the factory floor. smaller, production errors rise. Moving to smaller
Therefore, the changes of the business strategies are chips can result in 10 times as many errors, even for
needed to stay with the competitive business, which leaders such as Intel Corp. and IBM Corp. These
comes from growing trend of customers demands. errors can force a chip company to shut a production
For example, General Electric Company’s line, causing it to lose thousands of dollars. This is
strategic transformation [4] toward an industry leader one of the biggest problems chipmakers face today.
in six-sigma service in 1995 has inspired many KLA-Tencor foresees that the semiconductor
companies to follow. Many industries, including the industry will move toward e-diagnostics technologies
aerospace industry (i.e. jet engine, aircraft), the for integrated diagnostics monitoring. These
Proceedings of the 2nd International Conference on Responsive Manufacturing (ICRM 2002)
June 24-28, 2002- Gaziantep University, G. Antep, Turkey

enabling technologies will impact the entire industry, operating as a virtual enterprise as an effective part of
not just semiconductor business [6]. Further an extended supply chain.
advancements of these smart industrial applications
will be dependent upon continuous innovation and 2 1
1 2 3
the advancement of next-generation web-enabled 3
Suppliers
P lant 1 Plant 2 P lant n
CRM
Cus tomers & n
SCM Distributors
intelligent systems. n e-Manufacturing

Manual
3. E-MANUFACTURING Information
Autom ated
Info Flow
Flow
ARCHITECTURE AND FUNCTIONS Work Order Work Order
Orders Status / Data Order Status / Data
Currently, Manufacturing Execution
Systems (MES) provide a higher-level view of ?

Gaps & Sol utions


! IMS Manufacturing Exe cution
production but these systems were often inflexible to Hardcopies
Manual ! C omputation
Data ! P rediction
operate to be truly effective. The enterprise resource Manual
Coordination Entry ?
Controls
! O ptimization
planning (ERP) systems have become the financial ! Sy nchronization

backbone of many corporations. However, existing ?


structure of the ERP and MES cannot include the Plant Floor Plant Floor:
S mart, Pr oactive & Networked Devices
Legacy Devices & Systems, Static & Unused Data
dynamics of the factory floor such as maintenance or E mpowered operat ors

unpredictable downtime, variation and reliability of Figure 5: Today’s business system in which supply chain
suppliers and customers as illustrated in Figure 5. and customer relations are tightly coupled with the use of
Currently, a multitude of MES and ERP suppliers are internet and information technology whereas
manufacturing is still the weakest and broken link in the
working with manufacturers in an effort to integrate
chain. On the other hand, in the e-manufacturing
information throughout various parts of the environment, manufacturing is no longer the broken chain
enterprise. However, the crucial link between MES with the real-time monitoring, predictive intelligence ,
and ERP systems is hindered by the lack of smart devices and reconfigurable assets [after 7].
integration between these systems and the
information coming from and flowing to control Today, even with the best implemented lean
systems on the plant floor. manufacturing practices, many companies still face
It is necessary to identify the functional the following problems, which are all interrelated to
elements of e-Manufacturing to exchange transparent each other: 1) Defect parts, 2) High downtime, 3)
and automated information between shop floor and High energy utilization and cost, 4) Long changeover
enterprise. The major functions and objectives of e- and ramp up time, 5) Long lead time for new product
Manufacturing are (a) transparent, seamless and realization (long part and process design, and
automated information exchange process that can be validation periods), 6) Slow decision making (high
also coined as O.H.I.O (only handle information inventory, slow scheduling, etc.).
once), (b) total asset management that aims The proposed e-manufacturing architecture
improving the utilization of plant floor assets using a in this study addresses the above issues including
holistic approach combining the tools of predictive bringing about innovations on smart assessment
maintenance techniques, (c) e-supply chain network methodologies for degradation prediction and
that links entire SCM operation and asset prevention as well as to develop advanced
optimization, and (d) sophisticated customer service maintenance and service technologies to enable
system that serves customers utilizing the latest manufacturers and customers to sustain their assets
predictive intelligence methods and tether-free with near zero-downtime conditions. Briefly, the
technologies, Figure 5. main requirements of the e-Manufacturing can be
The real-time production information should listed as follows:
be made available to the entire organization, 1) Plant-floor, production processes, flow of material,
including managers at the top floors and machine etc. within the manufacturing system should be
operators at the plant floor in order to successfully modeled thoroughly. The model should contain
apply proactive maintenance to achieve near zero the entire flow of information for processes,
downtime of the systems. Then, e-Manufacturing products, equipment, people, and other assets in all
can target to optimize supply chain execution, agile plants involved including suppliers.
and adaptive customer service, improve asset 2) Information should be more than raw data. It
utilization, and implement dynamic preventive should be converted into a useful, understandable,
maintenance. Rockwell Automation [7] published a and intelligent form. Via such information mode,
white paper stating four competencies (design, complete visibility can be achieved. Status of
operate, maintain and synchronize) that are required equipments, orders, resources, products, changes in
for any manufacturer to be a world class the processes across various plants in the enterprise
manufacturing company. e-Manufacturing means can be monitored and provided using such
giving the company agility to react quickly to intelligent information management system.
changes in the market, technology and customers by 3) There should not be any black holes in the real
time flow of information throughout the system
Proceedings of the 2nd International Conference on Responsive Manufacturing (ICRM 2002)
June 24-28, 2002- Gaziantep University, G. Antep, Turkey

including outsourcing suppliers, customers, plant control, optimization of energy/power utilization,


management, machine operators and machine etc., 4) Development of virtual design platform for
itself. collaborative part, process, tooling design among
4) e-Manufacturing should include intelligent suppliers, design and process engineers and as well
maintenance and performance assessment systems as customers for fast validation and decision making.
to provide reliability, dependability and minimum Figure 6 illustrates the proposed e-Manufacturing
downtime with equipment running smoothly at concept as outline above.
their highest performance.
5) The entire system should be flexible enough to 4. MAJOR ELEMENT AND REQUIREMENTS
change with the varying market and demand
conditions in a short lead-time. In order to effectively implement the e-
6) Standard, flexible and enterprise-wide platforms Manufacturing platform described above in various
and protocols should be used to offer such manufacturing environments (food processing, metal
flexibility. Management of the assets both locally fabrication, petro-chemical processing, paper/tissue
and remotely should be provided. production, aircraft industry, etc.), the following
elements need to be accomplished and integrated:
Predictive Intelligence
Integrated w/ tether-free
• Data gathering from various machinery and
Real Factory communication systems
processes. This has been already done at various
... levels. However, massive raw data is not useful
unless it is reduced and transformed into
(1) Data gathering
information and knowledge for responsive actions.
ERP (2) Transformation Hence, data mining tools for data reduction,
System
Dyn amic Database XML
(4) Synchronization
representation and prediction adopted for plant
Production User interface Internet
Plan
Logistics
floor data needs to be developed
(Weekl y
Daily, ..)
Root
Cause
Degradation
prediction
Plan

Production
• Advanced prediction methods and tools need to be
Historical Simulation
Failure / Repair
Distribution
Model
Library
Maintenanc e
Strateg y
Maintenance developed in order to detect degradation,
Shipment

Virtual Factory
Manpower pool Procurement performance loss or trend of failure not faults,
CRM
Maintenance
Planni ng
Manpower
Availability
DSS
Optimiz ation Inventor y
SCM breakdowns, etc. For prediction of degradation on
Algorithm
EAM
Re-scheduling

(3) Analysis / Prediction / Optimization


components/machinery, computational and
D2B Platform
statistical tools should be developed to detect the
Figure 6: Proposed e-Manufacturing system includes degradation through massive computation using
reliable informatics platform (D2B™) thats comprises (1) slight changes and differences in the data coming
data gathering, (2) transformation, (3) analysis, prediction from devices. Figure 7 illustrates a schematic of a
and optimization, and (4) synchronization using dynamic predictive agent based on Cerebellar Model
virtual factory simulation, distributed predictive Articulation Controller (CMAC) technique [8, 9,
intelligence agents integrated with tether-free 10] being improved at the Center for Intelligent
communication systems for real-time decision-making. Maintenance Systems (IMS) in addition to various
other approaches as part of a Watchdog toolbox.
e-Manufacturing can offer comprehensive • For flexibility of data processing, analysis, remote
solutions by addressing the following research and monitoring and control of devices on the plant
development issues: 1) Development of intelligent floor existing tether-free technologies (wireless
agents for continuous, real time, remote and communication, wireless sensor, vision systems,
distributed monitoring and analyses of devices, etc.) should be utilized. Information pipeline- An
machinery and systems to provide the first and most information highway that provides reliable, secure,
needed element of predictive maintenance via scalable and compatible information
offering real time information about machine’s transformation between device and business
performance status (health condition), its capability systems (Device-to-Business platform, D2B™).
of producing quality parts (or completing its tasks), Figure 7 summarizes the use of wireless
etc, 2) Development of remote, distributed and web- technology combined with predictive agents.
based quality control systems and their integration • For wider and easier implementations, new
with intelligent predictive agents described above in standards need to be identified and developed.
order to identify quality variations and their causes in Existing efforts to develop standards should be
real time, 3) Development of a dependable and leveraged via active participation and involvement
scalable information pipeline/platform for complete (i.e. IEEE 802.xx standard committees, MIMOSA,
transformation, optimization and synchronization of etc.)
plant floor problems, issues, and solutions with
higher level production, maintenance and transaction
• Web-enabled decision-making portal- performs
dynamic optimization and synchronization based
scheduling systems, inventory control systems,
on the information obtained through D2B™
supply chain systems and with ERP for dynamic
pipeline, and provides knowledge for real time
scheduling of production, maintenance, human and
other resources, dynamic inventory monitoring and
Proceedings of the 2nd International Conference on Responsive Manufacturing (ICRM 2002)
June 24-28, 2002- Gaziantep University, G. Antep, Turkey

decision making to achieve the ultimate business prognostics for preventive maintenance. Users from
automation, Figure 6 and 7. different factories or locations can also share this
information through these web tools. This will
enable users for high-quality communications since
they all sharing the same set of data formats without
any language barriers.
Moreover, as another example, by knowing
E-Maintenance

1 the degradation of machines in the production floor,


ERP
the operation supervisor can estimate its impacts to
2 Embedded and web-enabled
Performance and
Degradation
Information
Via
Internet CRM
the materials flow and volume and synchronize it
Smart Prognostic Agent
(WATHCDOG) with the ERP systems. The revised inventory needs
5

3
and materials delivery can also be synchronized with
other business tools such as CRM system. When
cutting tools worn out on a machining center, the
4 information can be directly channeled to the tool
providers and update the tool needs for tool
performance management. In this case, the cutting
Figure 7: Various elements of e-Manufacturing: (1) data
gathering and predictive intelligence- D2B™ platform and
tool company is no long selling cutting tools, instead,
Watchdog Agent™, (2) tether-free communication selling cutting time. In addition, when the machine
technologies [7] and (5) optimization and synchronization degrades, the system can initiate a service call
tools, and their position in the architecture through the service center for prognostics. This will
change the practices from MTTR to MTBD (mean
With the further advancement of next- time between degradation) [11].
generation web-enabled tether-free infotronics agent
integrated under an e-Manufacturing system, the 5. INTELLIGENT MAINTENANCE SYSTEMS AND
massive data bottle-necks between the plant floor and E-MAINTENANCE ARCHITECTURE
business systems can be eliminated by converting the Predictive maintenance of plant floor assets
raw machine health data, product quality data and is a critical component of e-Manufacturing concept.
process capability data into information and Predictive maintenance systems, also referred as e-
knowledge for dynamic decision making. Maintenance in this document, provides
The technological advances for achieving manufacturing and operating systems with near-zero
such a highly collaborative design, manufacturing, downtime performance through use and integration
supply chain and service management environment is of (a) real time and smart monitoring, (b)
based on information-based engineering tools and a performance assessment methods, and (c) tether-free
highly reliable communication system. Distributed technologies. These systems can compare a
procedures and tools enable concurrent engineering product’s performance through globally networked
design, remote and distributed operation, monitoring monitoring systems to shift to the degradation
and control of manufacturing processes and systems. prediction and prognostics rather than fault detection
This transition is dependent upon the advancement of and diagnostics. To get maximum performance from
next-generation manufacturing practices to “e-factory plant floor assets, e-maintenance systems can be used
and e-automation ”; the use of information to make to monitor, analyze, compare, reconfigure, and
things happen collaboratively on a global basis. Such sustain the system via internet. In addition, these
a transition from today’s tightly coupled supplier- intelligent decisions can be harnessed through web-
enterprise-customer mode, which has a weak link enabled agents and connect them to e-business tools
with manufacturing, to a mode with a tightly coupled (such as customer relation management systems,
with intelligent manufacturing demands the ERP systems and e-commerce systems) to achieve
development of embedded intelligent machine smart and effective service solutions. Remote and
infotronics and dependable link between the real time assessment of machine’s performance
devices/machinery and business systems that enable requires an integration of many different
products, machinery and systems to (1) learn about technologies including sensory devices, reasoning
their status and environment, (2) predict degradation agents, wireless communication, virtual integration
of performance, (3) reconfigure itself to sustain and interface platforms. Figure 8 summarizes the
functional performance, and (4) informate business predictive maintenance (e-Maintenance) vision in e-
decisions directly from the device itself [2,3]. product development, e-manufacturing and e-service
At this level, many web-enabled goal whereas Figure 9 illustrates positioning of e-
applications can be performed. For example, we can Maintenance elements consisting of predictive
perform remote machine calibration through intelligence (Watchdog Agent™) and device-to-
telescopic Ball bar in a different location. Experts business (D2B™) platforms in a manufacturing
from machine tool manufacturers can assist users to setting.
analyze machine calibration data and perform
Proceedings of the 2nd International Conference on Responsive Manufacturing (ICRM 2002)
June 24-28, 2002- Gaziantep University, G. Antep, Turkey

Enhanced normal operation state, degraded state, maintenance


Six-Sig ma Design Predictive

Design for Reliability Product or


Monitoring Perfor mance state, and failure state [12]. Aging of components
System Sensors & Events
And Serviceability
In Use
• Web-enabled
monitoring,
and machinery usually progresses through a series of
Product Product
Prognostics, and
diagnostics degradation states before failure occurs. If a
Degradation
Center Redes ign Watchdog Agent™ • Innovative Customer degradation condition can be measured and detected,
Diagnostics
Relation Management
-

Self Maintenance Communications


(CRM)
&
then a proactive and corrective maintenance activity
S mart
Design
Redundancy
Active
Passive
• Tether-Free
(Bluetooth)
• Internet
• Prognostics
Agent for Data Mining can be performed before a worse degradation
• TCP/IP
• E-Business
Integration Tools
condition or failure occurs. The degraded states can
Web-enabled
- D2B™ Platform
• Asset Optimization
be defined at a gross level or a detailed level. At a
gross level, a component is described as degraded
Figure 8: Intelligent e-Maintenance System- position and whenever deterioration occurs but does not cause loss
function of degradation prediction and device-to-business of its function. Detailed degradation is associated
(D2B™) platform with a given range of characteristics of the
components, or performance of the components, such
as detailed degraded states for a car battery indicator
Machinery
and temperature indicator. The advantage of defining
more detailed degradation states is that we can
accurately predict impacts on the failure of a
component.
Tether-free
Factory Embedded Infotronics Agent
(EI A)

Embedded Prognostics
System/Plant
Level

Machine/Device
Level

1 5

2
4
Figure 9: Illustrative positioning of Watchdog Agent™
and D2B™ platform in a manufacturing setting 3
Component
Level

In order to perform an effective e-


maintenance system for smart asset management as
Virtual Machine
part of a successful e-Manufacturing practice, overall Virtual Model
Comparison
performance assessment of the entire manufacturing
facility beginning with the system/plant level and
ending at the components on machinery (top-to-
Self
bottom) should be conducted based on the existing Comparison

maintenance, quality and process data collected either


through PLCs, maintenance work sheet, CMMS, or Peer
operators to identify the critical machinery/devices Comparison

that are either the most costly in terms of


maintenance or critical to the operations such that
their downtime may cause serious consequences. Degradation
Furthermore, common failure modes, failure Asses sment of
Machinery/Product/Systems
frequency, time to fix (MTBF- mean time between
failure), cost of repair/exchange, and labor Figure 10: Overall performance of systems is assessed
requirements/difficulty. for the critical components based on the health condition of critical machinery.
and machinery should be identified. Through Performance of machinery/devices on the other hand is
categorization of machinery/devices and component assessed through the understanding of critical components.
types within various industry and applications, Degradation prediction at the lowest level ensures the total
performance assessment of the entire system can be performance assessment and understanding of the entire
methodological and straightforward for repetitive system whether it is a manufacturing plant or a building
implementations. Furthermore, through multiple system
comparison with (a) peers, (b) history and (c) virtual
models, assessment of the performance becomes very 6. DEVELOPMENT OF DEVICE INFOTRONICS-
PREDICIVE INTELLIGENCE FOR DEGRADATION
strong with a high confidence index, Figure 10. DETECTION
The operational performance of The core-enabling element of an intelligent
component/device can be divided into four states: maintenance system is the smart computational agent
Proceedings of the 2nd International Conference on Responsive Manufacturing (ICRM 2002)
June 24-28, 2002- Gaziantep University, G. Antep, Turkey

that can predict the degradation or performance loss if necessary, sensor/transducer type to quantify the
(Watchdog Agent™), not the traditional diagnostics degradation (Vibration- accelerometer sensor,
of failure or faults. A complete understanding and Current- eddy current sensors, Temperature- RTDs,
interpretation of states of degradation is necessary to Speed-encoder, Force- strain gauges, etc.) or PLC
accurately predict and prevent failure of a component data (command-status data), (c) development of a
or machine once their being critical to the overall degradation prediction strategy and method using one
production or operation systems is assessed through or a combination of the above techniques, (d)
the performance assessment methods explained in the Representation of prediction results in terms of mean
previous sections. A product’s performance time between degradation (MTBD), root causes, etc.
degradation behavior is often associated with multi- To realize the aforementioned tether-free embedded
symptom-domain information cluster, which consists Watchdog Agent™, technical breakthroughs are
of degradation behavior of functional components in needed to: (1) effectively correlate and compute
a chain of actions. The acquisition of specific massive and multi-sensor data from various sensors
sensory information may contain multiple behavior to predict the degradation and health condition using
information such as non-linear vibration, thermal or peer groups and virtual machine models, (2) integrate
materials surface degradation, and misalignment. All this dispersed information in a networked
of the information should be correlated for product environment, (3) synchronize with business systems
behavior assessment and prognostics. at an enterprise level utilizing web-enabled and
Degradation prediction of a wireless communication technologies. To effectively
component/device can be accomplished through apply this methodology in various kinds of products
application of various techniques such as Neural and machines, its integration with the working
Networks, Genetic Algorithms, Applied Statistics or environment (i.e. sensors, actuators, controllers, and
Hybrid Approaches. In this study, the development human interfaces) needs to be further developed. In
of a predictive intelligence toolbox (Watchdog the next section, more details are provided about
Toolbox™) is proposed. Such a Watchdog some of the predictive intelligence tools (Watchdog
Toolbox™ would include many of the above Agents™).
predictive techniques in a re-configurable and
modifiable fashion for a wide variety of 6.1. CMAC method for prediction of degradation
implementations as presented in Figure 11. In The output of Watchdog Agent™ represents
addition to the overall performance assessment the quantitative measures of the performance of the
methodologies depicted in the previous section and product or equipment at a given state. By comparing
Figure 10, comparison of results of the multiple performance value at different states, it can predict
Watchdog Agents™ will add another layer of and assess the condition and performance of complex
performance assessment, and provide a profound and machinery and systems. It is known that biological
high confidence prediction of degradation. perceptual systems are often conceptualized as
containing two stages of processing: an early stage
detects individual stimulus features and a later stage
combines the detected features into abstract
representations that facilitate thought, decision
making, or action. Drawing inspiration from this
biological phenomenon, a neural network based
multi-symptom-domains behavior assessment
methodology has been developed [9, 13] as one of
the approaches used in the Watchdog Toolbox™.
This approach uses behavior-based computation
NN-based Statistical CMAC based JTF based Hybrid
approach approach approach approach approach
rather than a model-based computation, which often
involves in using machine condition data. Jim Albus
Watchdog Toolbox™
[8] described the overall methodology of
Figure 11: Several predictive intelligence tools (i.e. performance assessment with an analogy to human
Watchdog Agents™) for degradation detection are sensing and assessment system. As depicted in
developed. Each has unique features and requirements for Figure 12, data from any physical domain (such as
different applications high temperature, rough surface, etc.) is transformed
into a value judgment (knowledge) after sensory
Whatever approach is used, alone or in processing and behavior assessment. The principle
combination with others, the common steps of relies on biometrics computation rather than a model-
degradation assessment can be generalized as based computation so that it can correlate multiple
follows: (a) determination of major, frequent and relevant inputs and hashing them to a set of weight
costly failure types occurring on a particular table. Its output represents the quantitative measures
component or machine through historical failure data of the performance of the product or equipment at a
and maintenance record analysis, (b) determination given state. By simply comparing performance value
of measurable parameters and selection of additional,
Proceedings of the 2nd International Conference on Responsive Manufacturing (ICRM 2002)
June 24-28, 2002- Gaziantep University, G. Antep, Turkey

at different states, it can predict and assess the capability. Information regarding (a) degradation
performance of machinery and products. Figure 13 modes, (b) degradation sections of the machine, (c)
illustrates the watchdog agent remote monitoring and degradation frequency, (d) degradation time and
control system and its working principle as well as place, (e) time required to prevent, (f) cost required
how multiple signals (vibration, temperature and to prevent, (g) suggested and/or applied maintenance
current) from a gearbox and electric motor system practices, etc. is recorded and stored in a database.
are hashed, and transformed into a performance This information is used for behavior learning and
confidence index. rule extraction purposes as a part of intelligent expert
maintenance system. Hence, a knowledge base
Sensors
Sensory Value Worl d system is achieved through intelligent conversion of
Processing Judgme nt Model
temperature warmth
data into information, and information into
knowledge.
pleasure WM
skin pressure softness
skin
area SP VJ Are a
3

s 2.5
deformation Gentle stroking frame p Speed, mps
of skin m
/
d 2
e
e
p
s
x 1.5 maintenance
a
M
Figure 12: Transformation of sensed data into value 1
0 50 100 150 200 250 300 350 400
Time/days
judgment (knowledge) analogous to a human’s sensing 4

system s/
e
3.5
Cy cle time (sec)
maintenance
mi 3
T
el
c
y 2.5
C
2
0 50 100 150 200 250 300 350 400
Time/days

1
Failure Line
prediction2 prediction1
yt e Probability of
iil r
uil 0.5 degradation
b a
a f maintenance
b f
o
P
r o point1 point2
0
0 50 100 150 200 250 300 350 400
Time/days

Historical Data & Statistical Degradation Model


Events

Training and C onvergence


V ibration Real-time Degradation
Tim e D om ain T rainin g Real-life
Prediction
data

V ibration
Fr equency
D om ain
Machine
A ssessm ent
onti
a
T em perature da
gr
e
D

Estimation of
Remaining Life
C urrent
C onfidence V alue: 90
Figure 14: (a) An example of degradation prediction
scheme using applied statistics techniques. Two critical
Figure 13: Web-based Watchdog Agent™ for machine
parameters of a device component (speed and time delay
behavior learning and prediction of degradation on a
between command and status) are measured continuously
gearbox and electric motor system using CMAC approach.
and converted into a performance index (probability of
degradation) after constructing models from historical
This approach intertwines the sensor fusion, data. (b) Statistical degradation model is constructed and
hashing, behavior learning and comparison updated using historical data and events such as
techniques as illustrated in Figure 13. In the next adjustments and reset values. Its comparison with real-
generation of this approach, evaluation of the data time and real-life signal data results is probability of
can be performed either locally on the degradation and an index of estimation for remaining life
machine/product or web-based depending on the
needs of the particular application. The performance 6.2. Applied statistical approach (probability of
index and comparison results are, then, transferred degradation) for prediction of degradation
through the wireless or internet/intranet network to Figure 14a illustrates the results of another
decision making platforms such as production, prediction degradation method on a device where
maintenance and manpower scheduling as well as applied statistics techniques were used. This
supplier and shipment synchronization. Reduction approach involves the following three steps as
and conversion of data into useable health depicted in Figure 14b: (1) use historical data and/or
information through smart prognostics would maintenance adjustments and events to model a
definitely eliminate hurdles regarding the performance index for the device, (2) estimate the
transmission of large data on a given bandwidth real-time performance using real-life sensor data, and
improve the model, and (3) update equipment
Proceedings of the 2nd International Conference on Responsive Manufacturing (ICRM 2002)
June 24-28, 2002- Gaziantep University, G. Antep, Turkey

degradation predictions continuously and estimate machinery and products, (b) relation between product
the remaining life based on comparison with quality variation and machine and process
statistical model. Using the historical sensory signals degradation, (c) data mining, reduction and data-to-
(speed) and controller data (cycle time), probability information-to-knowledge conversion tools, (d)
of degradation model was constructed. This model is reliable, scalable and common informatics platform
then compared with real-time data, obtained through between devices and business including
D2B™ platform, to predict the instantaneous implementation of wireless, Internet and Ethernet
probability of degradation (i.e. performance), networks in the manufacturing environment to
estimate the remaining life, and determine a window achieve flexible and low cost installations and
of appropriate maintenance before product fails. The commissioning, (e) data/information security and
frequency of computation is dependent on the vulnerability issues at the machine/product level, (f)
characteristics of the product/component and the data distributed and web-based computing, (d)
features in the question. For a dynamic component optimization and synchronization systems for
such as the spindle of a CNC machine tool, dynamic decision making, (g) education and training
computations need to be performed in very short of technicians, engineers and leaders to make them
periods of time (i.e. seconds). For an elavator, on the capable of pacing with the speed of information flow
other hand, prediction frequency can be triggered as and understanding the overall structure, and (h) a
the elevator is used. More likely it will be at larger new enterprise culture that has the flexibility of local
time intervals (i.e. minutes). dynamic decision making and strength of global
competition.
Technology

SCM
Infrastructure e-Busine ss References
1. P. Zipkin, 2002, Seminar on the Limits of Mass
CRM
Outsourcing
Trading
Customization, Center for Innovative Manufacturing and
e-Manufacturing
VMI Exchanges Operations Management (CIMOM), April 22, 2002.
Collaborative
e-
2. Patrick Waurzyniak, 2001, “Moving towards e-
Real-time Planning
Data
Technology
Procurement
factory”, SME Manufacturing Magazine, v. 127, n. 5,
e-Maintenance
Condition-based
Infrastructure Nov. 2001
Dynamic
Monitoring Real-time
Information
Decision 3. Patrick Waurzyniak, 2001, “Web tools catch on”,
Making
Predictive
Technologies Information
Asset
SME Manufacturing Magazine, v. 127, n.4, Oct. 2001
pipeline
Management 4. Jack Welch and GE, Business Week, Oct. issue
1996.
5. Telematics to the Rescue, pp93, IEEE Spectrum,
Figure 15: e-Manufacturing as a wholly integrated part of Jan. 2001
the e-Business system. e-Maintenance and its basic 6. Kenneth Schroeder Interviews with Investor’s
elements are part of e-Manufacturing requirements. Business Daily, CEO of KLA-Tencor Corp., August
2000.
7. Conclusions and Future Work 7. Rockwell Automation e-Manufacturing Industry
This paper proposes an e-Manufacturing Road Map, http://www.rockwellautomation.com
architecture, and outlines its fundamental 8. J.S. Albus, 1975, “A new approach to manipulator
requirements and elements as well as expected control: the CMAC”, Journal of Dynamic Systems and
impact to achieve non-stop, efficient and integrated Control, Tans. ASME, Series, G., v.97, n.3, Sept., 1975,
pp. 220
operations. Information technology will play an
9. J. Lee, 1996, “Measurement of machine
indispensable role in supporting and enabling the performance degradation using a neural network model”,
complex practices of design and manufacturing by J. Computers in Industry, v.30, pp 193.
providing the mechanisms to facilitate and manage 10. Y. Wong and S. Athanasios, 1992, “Learning
the integrated system discipline with the higher convergence in the CMAC”, IEEE Transactions on
system levels such as SCM and ERP. Manufacturers Neural Networks, v.3, n.1, Jan. 1992, pp.115
and users will benefit from the increased equipment 11. J. Lee and B. Wang, 1999, Computer-aided
and process reliability with the e-manufacturing Maintenance: Methodologies and Practices, Kluwer
strategies as well as optimal asset management in Academic Publishing.
terms of rapid deployment, rapid time-to-market 12. J. Lee, 1999, “Machine Performance
Assessment Methodology and Advanced Service
cycles, non-stop operations and seamless integration
Technologies, Report of Fourth Annual Symposium on
with suppliers and customer. Frontiers of Engineering, National Academy Press,
In order to accomplish the proposed e- pp.75-83, Washington, DC.
Manufacturing system to complete the e-Business 13. J. Lee, and B.M. Kramer, 1992, Analysis of
matrix, which has evolved as a consequence of global Machine Degradation using a Neural Networks Based
economy demands during the past decade, the Pattern Discrimination Model, J. Manufacturing
followings need to be researched, developed and Systems, Vol. 12, No. 3, pp. 379-387.
implemented as summarized in Figure 15: (a)
predictive intelligence (algorithms, software and
agents) focusing on degradation detection on various

View publication stats

You might also like