Ferro Final 12052015

FERRO-11
11th International Symposium on Ferrocement

and
3rd ICTRC
International Conference on Textile Reinforced Concrete
Aachen, Germany
07 - 10 June 2015
Edited by Wolfgang Brameshuber
RILEM Publications S.A.R.L.

Published by RILEM Publications S.A.R.L.
157 rue des Blains F-92220 Bagneux - France
Tel : + 33 1 45 36 10 20 Fax : + 33 1 45 36 63 20
http://www.rilem.net E-mail: dg@rilem.net
© 2015 RILEM – Tous droits réservés. ISBN: 978-2-35158-152-0
e-ISBN: 978-2-35158-153-7
Publisher's note: this book has been produced from electronic files provided by the indi-
vidual contributors. The publisher makes no representation, express or implied, with re-
gard to the accuracy of the information contained in this book and cannot accept any legal
responsibility or liability for any errors or omissions that may be made.
All titles published by RILEM Publications are under copyright protection; said copy-
rights being the property of their respective holders. All Rights Reserved.
No part of any book may be reproduced or transmitted in any form or by any means,
graphic, electronic, or mechanical, including photocopying, recording, taping, or by any
information storage or retrieval system, without the permission in writing from the pub-
lisher.
RILEM, The International Union of Laboratories and Experts in Construction Materials,

Systems and Structures, is a non profit-making, non-governmental technical association
whose vocation is to contribute to progress in the construction sciences, techniques and
industries, essentially by means of the communication it fosters between research and prac-
tice. RILEM’s activity therefore aims at developing the knowledge of properties of materi-
als and performance of structures, at defining the means for their assessment in laboratory
and service conditions and at unifying measurement and testing methods used with this
objective.
RILEM was founded in 1947, and has a membership of over 900 in some 70 countries. It
forms an institutional framework for co-operation by experts to:
• optimise and harmonise test methods for measuring properties and performance of build-
ing and civil engineering materials and structures under laboratory and service environ-
ments,
• prepare technical recommendations for testing methods,
• prepare state-of-the-art reports to identify further research needs,
• collaborate with national or international associations in realising these objectives.
RILEM members include the leading building research and testing laboratories around the
world, industrial research, manufacturing and contracting interests, as well as a significant
number of individual members from industry and universities. RILEM’s focus is on con-
struction materials and their use in building and civil engineering structures, covering all
phases of the building process from manufacture to use and recycling of materials.
RILEM meets these objectives through the work of its technical committees. Symposia,
workshops and seminars are organised to facilitate the exchange of information and dis-
semination of knowledge. RILEM’s primary output consists of technical recommenda-
tions. RILEM also publishes the journal Materials and Structures which provides a further
avenue for reporting the work of its committees. Many other publications, in the form of
reports, monographs, symposia and workshop proceedings are produced.
II
FERRO-11 – 11th International Symposium on Ferrocement and
rd
3 ICTRC - International Conference on Textile Reinforced Concrete
07 – 10 June 2015, Aachen, Germany
RILEM Publications
The following list is presenting the global offer of RILEM Publications, sorted by series.
Each publication is available in printed version and/or in online version.
RILEM PROCEEDINGS (PRO)

PRO 1: Durability of High Performance Concrete (ISBN: 2-912143-03-9); Ed. H. Sommer
PRO 2: Chloride Penetration into Concrete (ISBN: 2-912143-00-04);
Eds. L.-O. Nilsson and J.-P. Ollivier
PRO 3: Evaluation and Strengthening of Existing Masonry Structures (ISBN: 2-912143-02-0);
Eds. L. Binda and C. Modena
PRO 4: Concrete: From Material to Structure (ISBN: 2-912143-04-7); Eds. J.-P. Bournazel
and Y. Malier
PRO 5: The Role of Admixtures in High Performance Concrete (ISBN: 2-912143-05-5);
Eds. J. G. Cabrera and R. Rivera-Villarreal
PRO 6: High Performance Fiber Reinforced Cement Composites - HPFRCC 3
(ISBN: 2-912143-06-3); Eds. H. W. Reinhardt and A. E. Naaman
PRO 7: 1st International RILEM Symposium on Self-Compacting Concrete (ISBN: 2-912143-
09-8); Eds. Å. Skarendahl and Ö. Petersson
PRO 8: International RILEM Symposium on Timber Engineering (ISBN: 2-912143-10-1);
Ed. L. Boström
PRO 9: 2nd International RILEM Symposium on Adhesion between Polymers and Concrete
ISAP ’99 (ISBN: 2-912143-11-X); Eds. Y. Ohama and M. Puterman
PRO 10: 3rd International RILEM Symposium on Durability of Building and Construction
Sealants (ISBN: 2-912143-13-6); Eds. A. T. Wolf
PRO 11: 4th International RILEM Conference on Reflective Cracking in Pavements
(ISBN: 2-912143-14-4); Eds. A. O. Abd El Halim, D. A. Taylor and El H. H. Mohamed
PRO 12: International RILEM Workshop on Historic Mortars: Characteristics and Tests
(ISBN: 2-912143-15-2); Eds. P. Bartos, C. Groot and J. J. Hughes
PRO 13: 2nd International RILEM Symposium on Hydration and Setting (ISBN: 2-912143-
16-0); Ed. A. Nonat
PRO 14: Integrated Life-Cycle Design of Materials and Structures - ILCDES 2000
(ISBN: 951-758-408-3); (ISSN: 0356-9403); Ed. S. Sarja
PRO 15: Fifth RILEM Symposium on Fibre-Reinforced Concretes (FRC) - BEFIB’2000
(ISBN: 2-912143-18-7); Eds. P. Rossi and G. Chanvillard
PRO 16: Life Prediction and Management of Concrete Structures
(ISBN: 2-912143-19-5); Ed. D. Naus
PRO 17: Shrinkage of Concrete – Shrinkage 2000 (ISBN: 2-912143-20-9);
Eds. V. Baroghel-Bouny and P.-C. Aïtcin
PRO 18: Measurement and Interpretation of the On-Site Corrosion Rate (ISBN: 2-912143-21-7);
Eds. C. Andrade, C. Alonso, J. Fullea, J. Polimon and J. Rodriguez
PRO 19: Testing and Modelling the Chloride Ingress into Concrete (ISBN: 2-912143-22-5);
Eds. C. Andrade and J. Kropp
PRO 20: 1st International RILEM Workshop on Microbial Impacts on Building Materials (CD
02) (e-ISBN 978-2-35158-013-4); Ed. M. Ribas Silva
III
PRO 21: International RILEM Symposium on Connections between Steel and Concrete
(ISBN: 2-912143-25-X); Ed. R. Eligehausen
PRO 22: International RILEM Symposium on Joints in Timber Structures
(ISBN: 2-912143-28-4); Eds. S. Aicher and H.-W. Reinhardt
PRO 23: International RILEM Conference on Early Age Cracking in Cementitious Systems
(ISBN: 2-912143-29-2); Eds. K. Kovler and A. Bentur
PRO 24: 2nd International RILEM Workshop on Frost Resistance of Concrete
(ISBN: 2-912143-30-6); Eds. M. J. Setzer, R. Auberg and H.-J. Keck
PRO 25: International RILEM Workshop on Frost Damage in Concrete
(ISBN: 2-912143-31-4); Eds. D. J. Janssen, M. J. Setzer and M. B. Snyder
PRO 26: International RILEM Workshop on On-Site Control and Evaluation of Masonry
Structures (ISBN: 2-912143-34-9); Eds. L. Binda and R. C. de Vekey
PRO 27: International RILEM Symposium on Building Joint Sealants (CD03); Ed. A. T. Wolf
PRO 28: 6th International RILEM Symposium on Performance Testing and Evaluation of
Bituminous Materials - PTEBM’03 (ISBN: 2-912143-35-7; e-ISBN: 978-2-912143-77-8); Ed.
M. N. Partl
PRO 29: 2nd International RILEM Workshop on Life Prediction and Ageing Management of
Concrete Structures (ISBN: 2-912143-36-5); Ed. D. J. Naus
PRO 30: 4th International RILEM Workshop on High Performance Fiber Reinforced Cement
Composites - HPFRCC 4 (ISBN: 2-912143-37-3); Eds. A. E. Naaman and H. W. Reinhardt
PRO 31: International RILEM Workshop on Test and Design Methods for Steel Fibre Rein-
forced Concrete: Background and Experiences (ISBN: 2-912143-38-1); Eds. B. Schnütgen and
L. Vandewalle
PRO 32: International Conference on Advances in Concrete and Structures 2 vol.
(ISBN (set): 2-912143-41-1); Eds. Ying-shu Yuan, Surendra P. Shah and Heng-lin Lü
PRO 33: 3rd International Symposium on Self-Compacting Concrete
(ISBN: 2-912143-42-X); Eds. Ó. Wallevik and I. Níelsson
PRO 34: International RILEM Conference on Microbial Impact on Building Materials
(ISBN: 2-912143-43-8); Ed. M. Ribas Silva
PRO 35: International RILEM TC 186-ISA on Internal Sulfate Attack and Delayed Ettringite
Formation (ISBN: 2-912143-44-6); Eds. K. Scrivener and J. Skalny
PRO 36: International RILEM Symposium on Concrete Science and Engineering – A Tribute
to Arnon Bentur (ISBN: 2-912143-46-2); Eds. K. Kovler, J. Marchand, S. Mindess and J.
Weiss
PRO 37: 5th International RILEM Conference on Cracking in Pavements – Mitigation, Risk
Assessment and Prevention (ISBN: 2-912143-47-0); Eds. C. Petit, I. Al-Qadi and A. Millien
PRO 38: 3rd International RILEM Workshop on Testing and Modelling the Chloride Ingress
into Concrete (ISBN: 2-912143-48-9); Eds. C. Andrade and J. Kropp
PRO 39: 6th International RILEM Symposium on Fibre-Reinforced Concretes - BEFIB 2004
(ISBN: 2-912143-51-9); Eds. M. Di Prisco, R. Felicetti and G. A. Plizzari
PRO 40: International RILEM Conference on the Use of Recycled Materials in Buildings and
Structures (ISBN: 2-912143-52-7); Eds. E. Vázquez, Ch. F. Hendriks and G. M. T. Janssen
PRO 41: RILEM International Symposium on Environment-Conscious Materials and Systems
for Sustainable Development (ISBN: 2-912143-55-1); Eds. N. Kashino and Y. Ohama
IV
rd
PRO 42: SCC’2005 - China: 1st International Symposium on Design, Performance and Use of
Self-Consolidating Concrete (ISBN: 2-912143-61-6); Eds. Zhiwu Yu, Caijun Shi, Kamal Henri
Khayat and Youjun Xie
PRO 43: International RILEM Workshop on Bonded Concrete Overlays (e-ISBN: 2-912143-
83-7); Eds. J. L. Granju and J. Silfwerbrand
PRO 44: 2nd International RILEM Workshop on Microbial Impacts on Building Materials
(CD11) (e-ISBN: 2-912143-84-5); Ed. M. Ribas Silva
PRO 45: 2nd International Symposium on Nanotechnology in Construction, Bilbao
(ISBN: 2-912143-87-X); Eds. Peter J. M. Bartos, Yolanda de Miguel and Antonio Porro
PRO 46: ConcreteLife’06 - International RILEM-JCI Seminar on Concrete Durability and
Service Life Planning: Curing, Crack Control, Performance in Harsh Environments
(ISBN: 2-912143-89-6); Ed. K. Kovler
PRO 47: International RILEM Workshop on Performance Based Evaluation and Indicators for
Concrete Durability (ISBN: 978-2-912143-95-2); Eds. V. Baroghel-Bouny, C. Andrade, R.
Torrent and K. Scrivener
PRO 48: 1st International RILEM Symposium on Advances in Concrete through Science and
Engineering (e-ISBN: 2-912143-92-6); Eds. J. Weiss, K. Kovler, J. Marchand, and S. Mindess
PRO 49: International RILEM Workshop on High Performance Fiber Reinforced Cementi-
tious Composites in Structural Applications (ISBN: 2-912143-93-4); Eds. G. Fischer and V.C.
Li
PRO 50: 1st International RILEM Symposium on Textile Reinforced Concrete
(ISBN: 2-912143-97-7); Eds. Josef Hegger, Wolfgang Brameshuber and Norbert Will
PRO 51: 2nd International Symposium on Advances in Concrete through Science and Engineering
(ISBN: 2-35158-003-6; e-ISBN: 2-35158-002-8); Eds. J. Marchand, B. Bissonnette, R. Gagné, M.
Jolin and F. Paradis
PRO 52: Volume Changes of Hardening Concrete: Testing and Mitigation (ISBN: 2-35158-
004-4; e-ISBN: 2-35158-005-2); Eds. O. M. Jensen, P. Lura and K. Kovler
PRO 53: High Performance Fiber Reinforced Cement Composites - HPFRCC5
(ISBN: 978-2-35158-046-2); Eds. H. W. Reinhardt and A. E. Naaman
PRO 54: 5th International RILEM Symposium on Self-Compacting Concrete
(ISBN: 978-2-35158-047-9); Eds. G. De Schutter and V. Boel
PRO 55: International RILEM Symposium Photocatalysis, Environment and Construction
Materials (ISBN: 978-2-35158-056-1); Eds. P. Baglioni and L. Cassar
PRO56: International RILEM Workshop on Integral Service Life Modelling of Concrete
Structures (ISBN 978-2-35158-058-5); Eds. R. M. Ferreira, J. Gulikers and C. Andrade
PRO57: RILEM Workshop on Performance of cement-based materials in aggressive aqueous
environments (e-ISBN: 978-2-35158-059-2); Ed. N. De Belie
PRO58: International RILEM Symposium on Concrete Modelling - CONMOD’08
(ISBN: 978-2-35158-060-8); Eds. E. Schlangen and G. De Schutter
PRO 59: International RILEM Conference on On Site Assessment of Concrete, Masonry and
Timber Structures - SACoMaTiS 2008 (ISBN set: 978-2-35158-061-5); Eds. L. Binda, M. di
Prisco and R. Felicetti
PRO 60: Seventh RILEM International Symposium on Fibre Reinforced Concrete: Design and
Applications - BEFIB 2008 (ISBN: 978-2-35158-064-6); Ed. R. Gettu
PRO 61: 1st International Conference on Microstructure Related Durability of Cementitious
Composites 2 vol., (ISBN: 978-2-35158-065-3); Eds. W. Sun, K. van Breugel, C. Miao, G. Ye
and H. Chen
V
PRO 62: NSF/ RILEM Workshop: In-situ Evaluation of Historic Wood and Masonry Struc-
tures
(e-ISBN: 978-2-35158-068-4); Eds. B. Kasal, R. Anthony and M. Drdácký
PRO 63: Concrete in Aggressive Aqueous Environments: Performance, Testing and Model-
ling, 2 vol., (ISBN: 978-2-35158-071-4); Eds. M. G. Alexander and A. Bertron
PRO 64: Long Term Performance of Cementitious Barriers and Reinforced Concrete in Nu-
clear Power Plants and Waste Management - NUCPERF 2009 (ISBN: 978-2-35158-072-1);
Eds. V. L’Hostis, R. Gens, C. Gallé
PRO 65: Design Performance and Use of Self-consolidating Concrete - SCC’2009
(ISBN: 978-2-35158-073-8); Eds. C. Shi, Z. Yu, K. H. Khayat and P. Yan
PRO 66: 2nd International RILEM Workshop on Concrete Durability and Service Life Plan-
ning - ConcreteLife’09 (ISBN: 978-2-35158-074-5); Ed. K. Kovler
PRO 67: Repairs Mortars for Historic Masonry (e-ISBN: 978-2-35158-083-7); Ed. C. Groot
PRO 68: Proceedings of the 3rd International RILEM Symposium on ‘Rheology of Cement
Suspensions such as Fresh Concrete (ISBN 978-2-35158-091-2); Eds. O. H. Wallevik, S. Ku-
bens and S. Oesterheld
PRO 69: 3rd International PhD Student Workshop on ‘Modelling the Durability of Reinforced
Concrete (ISBN: 978-2-35158-095-0); Eds. R. M. Ferreira, J. Gulikers and C. Andrade
PRO 70: 2nd International Conference on ‘Service Life Design for Infrastructure’ (ISBN set:
978-2-35158-096-7, e-ISBN: 978-2-35158-097-4); Ed. K. van Breugel, G. Ye and Y. Yuan
PRO 71: Advances in Civil Engineering Materials - The 50-year Teaching Anniversary of
Prof. Sun Wei’ (ISBN: 978-2-35158-098-1; e-ISBN: 978-2-35158-099-8); Eds. C. Miao, G.
Ye, and H. Chen
PRO 72: First International Conference on ‘Advances in Chemically-Activated Materials –
CAM’2010’ (2010), 264 pp, ISBN: 978-2-35158-101-8; e-ISBN: 978-2-35158-115-5, Eds.
Caijun Shi and Xiaodong Shen
PRO 73: 2nd International Conference on ‘Waste Engineering and Management - ICWEM
2010’ (2010), 894 pp, ISBN: 978-2-35158-102-5; e-ISBN: 978-2-35158-103-2, Eds. J. Zh.
Xiao, Y. Zhang, M. S. Cheung and R. Chu
PRO 74: International RILEM Conference on ‘Use of Superabsorsorbent Polymers and Other
New Addditives in Concrete’ (2010) 374 pp., ISBN: 978-2-35158-104-9; e-ISBN: 978-2-
35158-105-6;
Eds. O.M. Jensen, M.T. Hasholt, and S. Laustsen
PRO 75: International Conference on ‘Material Science - 2nd ICTRC - Textile Reinforced
Concrete - Theme 1’ (2010) 436 pp., ISBN: 978-2-35158-106-3; e-ISBN: 978-2-35158-107-0;
Ed. W. Brameshuber
PRO 76: International Conference on ‘Material Science - HetMat - Modelling of Heterogene-
ous Materials - Theme 2’ (2010) 255 pp., ISBN: 978-2-35158-108-7; e-ISBN: 978-2-35158-
109-4; Ed. W. Brameshuber
VI
rd
PRO 77: International Conference on ‘Material Science - AdIPoC - Additions Improving

Properties of Concrete - Theme 3’ (2010) 459 pp., ISBN: 978-2-35158-110-0; e-ISBN: 978-2-
35158-111-7; Ed. W. Brameshuber
PRO 78: 2nd Historic Mortars Conference and RILEM TC 203-RHM Final Workshop –
HMC2010 (2010) 1416 pp., e-ISBN: 978-2-35158-112-4; Eds J. Válek, C. Groot, and J. J.
Hughes
PRO 79: International RILEM Conference on Advances in Construction Materials Through
Science and Engineering (2011) 213 pp., e-ISBN: 978-2-35158-117-9; Eds Christopher Leung
and K.T. Wan
PRO 80: 2nd International RILEM Conference on Concrete Spalling due to Fire Exposure
(2011) 453 pp., ISBN: 978-2-35158-118-6, e-ISBN: 978-2-35158-119-3; Eds E.A.B. Koenders
and F. Dehn
PRO 81: 2nd International RILEM Conference on Strain Hardening Cementitious Composites
(SHCC2-Rio) (2011) 451 pp., ISBN: 978-2-35158-120-9, e-ISBN: 978-2-35158-121-6; Eds
R.D. Toledo Filho, F.A. Silva, E.A.B. Koenders and E.M.R. Fairbairn
PRO 82: 2nd International RILEM Conference on Progress of Recycling in the Built Environ-
ment (2011) 507 pp., e-ISBN: 978-2-35158-122-3; Eds V.M. John, E. Vazquez, S.C. Angulo
and C. Ulsen
PRO 83: 2nd International Conference on Microstructural-related Durability of Cementitious
Composites (2012) 250 pp., ISBN: 978-2-35158-129-2; e-ISBN: 978-2-35158-123-0; Eds G.
Ye, K. van Breugel, W. Sun and C. Miao
PRO 85: RILEM-JCI International Workshop on Crack Control of Mass Concrete and Related
issues concerning Early-Age of Concrete Structures – ConCrack 3 – Control of Cracking in
Concrete Structures 3 (2012) 237 pp., ISBN: 978-2-35158-125-4; e-ISBN: 978-2-35158-126-1;
Eds F. Toutlemonde and J.-M. Torrenti
PRO 86: International Symposium on Life Cycle Assessment and Construction (2012) 414
pp., ISBN: 978-2-35158-127-8, e-ISBN: 978-2-35158-128-5; Eds A. Ventura and C. de la
Roche
PRO 87: UHPFRC 2013 – RILEM-fib-AFGC International Symposium on Ultra-High Per-
formance Fibre-Reinforced Concrete (2013), ISBN: 978-2-35158-130-8, e-ISBN: 978-2-
35158-131-5; Eds F. Toutlemonde
PRO 88: 8th RILEM International Symposium on Fibre Reinforced Concrete (2012) 344 pp.,
ISBN: 978-2-35158-132-2, e-ISBN: 978-2-35158-133-9; Eds Joaquim A.O. Barros
PRO 89: RILEM International workshop on performance-based specification and control of
concrete durability (2014) 678 pp, ISBN: 978-2-35158-135-3, e-ISBN: 978-2-35158-136-0;
Eds. D. Bjegović, H. Beushausen and M. Serdar
PRO 90: 7th RILEM International Conference on Self-Compacting Concrete and of the 1st
RILEM International Conference on Rheology and Processing of Construction Materials
(2013) 396 pp, ISBN: 978-2-35158-137-7, e-ISBN: 978-2-35158-138-4, Eds. Nicolas Roussel
and Hela Bessaies-Bey
PRO 91 draft: CONMOD 2014 - RILEM International Symposium on Concrete Modelling
(2014), ISBN: 978-2-35158-139-1; e-ISBN: 978-2-35158-140-7,
VII
PRO 92: CAM 2014 - 2nd International Conference on advances in chemically-activated ma-
terials (2014) 392 pp., ISBN: 978-2-35158-141-4; e-ISBN: 978-2-35158-142-1, Eds. Caijun
Shi and Xiadong Shen
PRO 93: SCC 2014 - 3rd International Symposium on Design, Performance and Use of Self-
Consolidating Concrete (2014) 438 pp., ISBN: 978-2-35158-143-8; e-ISBN: 978-2-35158-
144-5, Eds. Caijun Shi, Zhihua Ou, Kamal H. Khayat
PRO 94 (online version): HPFRCC-7 - 7th RILEM conference on High performance fiber
reinforced cement composites, e-ISBN: 978-2-35158-146-9, Eds. H.W. Reinhardt,
G.J. Parra-Montesinos, H. Garrecht
PRO 95: International RILEM Conference on Application of superabsorbent polymers and
other new admixtures in concrete construction, ISBN: 978-2-35158-147-6; e-ISBN: 978-2-
35158-148-3, Eds. Viktor Mechtcherine, Christof Schroefl
PRO 96 (online version): XIII DBMC: XIII International Conference on Durability of Build-
ing Materials and Components, e-ISBN: 978-2-35158-149-0
PRO 97: SHCC3 – 3rd International RILEM Conference on Strain Hardening Cementitious
Composites, ISBN: 978-2-35158-150-6; e-ISBN: 978-2-35158-151-3, Eds. E. Schlangen, M.G.
Sierra Beltran, M. Lukovic, G. Ye
RILEM REPORTS (REP)

Report 19: Considerations for Use in Managing the Aging of Nuclear Power Plant Concrete
Structures (ISBN: 2-912143-07-1); Ed. D. J. Naus
Report 20: Engineering and Transport Properties of the Interfacial Transition Zone in Ce-
mentitious Composites (ISBN: 2-912143-08-X); Eds. M. G. Alexander, G. Arliguie, G. Ballivy,
A. Bentur and J. Marchand
Report 21: Durability of Building Sealants (ISBN: 2-912143-12-8); Ed. A. T. Wolf
Report 22: Sustainable Raw Materials - Construction and Demolition Waste
(ISBN: 2-912143-17-9); Eds. C. F. Hendriks and H. S. Pietersen
Report 23: Self-Compacting Concrete state-of-the-art report (ISBN: 2-912143-23-3);
Eds. Å. Skarendahl and Ö. Petersson
Report 24: Workability and Rheology of Fresh Concrete: Compendium of Tests
(ISBN: 2-912143-32-2); Eds. P. J. M. Bartos, M. Sonebi and A. K. Tamimi
Report 25: Early Age Cracking in Cementitious Systems (ISBN: 2-912143-33-0); Ed. A. Ben-
tur
Report 26: Towards Sustainable Roofing (Joint Committee CIB/RILEM) (CD 07)
(e-ISBN 978-2-912143-65-5); Eds. Thomas W. Hutchinson and Keith Roberts
Report 27: Condition Assessment of Roofs (Joint Committee CIB/RILEM) (CD 08)
(e-ISBN 978-2-912143-66-2); Ed. CIB W 83/RILEM TC166-RMS
Report 28: Final report of RILEM TC 167-COM ‘Characterisation of Old Mortars with Respect to
Their Repair (ISBN: 978-2-912143-56-3); Eds. C. Groot, G. Ashall and J. Hughes
Report 29: Pavement Performance Prediction and Evaluation (PPPE): Interlaboratory Tests
(e-ISBN: 2-912143-68-3); Eds. M. Partl and H. Piber
VIII
rd
Report 30: Final Report of RILEM TC 198-URM ‘Use of Recycled Materials’ (ISBN: 2-
912143-82-9; e-ISBN: 2-912143-69-1); Eds. Ch. F. Hendriks, G. M. T. Janssen and E.
Vázquez
Report 31: Final Report of RILEM TC 185-ATC ‘Advanced testing of cement-based materi-
als during setting and hardening’ (ISBN: 2-912143-81-0; e-ISBN: 2-912143-70-5); Eds. H. W.
Reinhardt and C. U. Grosse
Report 32: Probabilistic Assessment of Existing Structures. A JCSS publication
(ISBN 2-912143-24-1); Ed. D. Diamantidis
Report 33: State-of-the-Art Report of RILEM Technical Committee TC 184-IFE ‘Industrial
Floors’ (ISBN 2-35158-006-0); Ed. P. Seidler
Report 34: Report of RILEM Technical Committee TC 147-FMB ‘Fracture mechanics appli-
cations to anchorage and bond’ Tension of Reinforced Concrete Prisms – Round Robin Analy-
sis and Tests on Bond (e-ISBN 2-912143-91-8); Eds. L. Elfgren and K. Noghabai
Report 35: Final Report of RILEM Technical Committee TC 188-CSC ‘Casting of Self Com-
pacting Concrete’ (ISBN 2-35158-001-X; e-ISBN: 2-912143-98-5); Eds. Å. Skarendahl and P.
Billberg
Report 36: State-of-the-Art Report of RILEM Technical Committee TC 201-TRC ‘Textile
Reinforced Concrete’ (ISBN 2-912143-99-3); Ed. W. Brameshuber
Report 37: State-of-the-Art Report of RILEM Technical Committee TC 192-ECM ‘Environ-
ment-conscious construction materials and systems’ (ISBN: 978-2-35158-053-0);
Eds. N. Kashino, D. Van Gemert and K. Imamoto
Report 38: State-of-the-Art Report of RILEM Technical Committee TC 205-DSC ‘Durability
of Self-Compacting Concrete’ (ISBN: 978-2-35158-048-6); Eds. G. De Schutter and K. Audenaert
Report 39: Final Report of RILEM Technical Committee TC 187-SOC ‘Experimental deter-
mination of the stress-crack opening curve for concrete in tension’ (ISBN 978-2-35158-049-3);
Ed. J. Planas
Report 40: State-of-the-Art Report of RILEM Technical Committee TC 189-NEC ‘Non-
Destructive Evaluation of the Penetrability and Thickness of the Concrete Cover’ (ISBN 978-
2-35158-054-7);
Eds. R. Torrent and L. Fernández Luco
Report 41: State-of-the-Art Report of RILEM Technical Committee TC 196-ICC ‘Internal
Curing of Concrete’ (ISBN 978-2-35158-009-7); Eds. K. Kovler and O. M. Jensen
Report 42: ‘Acoustic Emission and Related Non-destructive Evaluation Techniques for Crack
Detection and Damage Evaluation in Concrete’ - Final Report of RILEM Technical Committee
212-ACD (e-ISBN: 978-2-35158-100-1); Ed. M. Ohtsu
IX
Conference Chair:
Univ.-Prof. Dr.-Ing. Wolfgang Brameshuber
Institute of Building Materials Research - RWTH Aachen University
Schinkelstraße 3, 52062 Aachen, GERMANY
International Technical Committee

• Aldea, Corina, Canada • Austriaco, Lilia, Philippines
• Bedoya,Daniel, Colombia • Bentur, Arnon, Israel
• Brameshuber, Wolfgang, Germany • Canmorano, José, Uruguay
• Castro, Marcelino, Nicaragua • Chudoba, Rostislav, Germany
• Emilio M. Colón, Puerto Rico • Curbach, Manfred, Germany
• De Andrade Silva, Flavio, Brazil • de Moya, Raul, Dominican Republic
• Delvasto, Silvio, Colombia • Djausal, Anshori, Indonesia
• Dubey,Ashish, USA • Garcia, David, Spain
• Garmendia, Leire, Spain • Gopinath, Smitha, India
• Gries, Thomas, Germany • Guerra, Antonio, Dominican Republic
• Guettu, Ravindra, India • Hamelin, Patrice, France
• Hanai, Joao Bento, Brazil • Hegger, Josef, Germany
• Jesse, Frank, Germany • Jones, John, U.S.A
• Kiattikomol, K., Tailandia • Mäder, Edith, Germany
• Malaga, Katharina, Sweden • Martirena,Fernando, Cuba
• Mechtcherine, Victor, Germany • Melo, Antonio, Spain
• Milenkovic, Milenko, Serbia • Mobasher, Barzin, USA
• Moreno, Paul, Ecuador • Naaman, Antoine, USA
• Nanni, Antonio, USA • Nimityogskul, Pichai, Tailandia
• Ohama, Y. Japan • Papanicolaou, Catherine, Greece
• Paramasivam, P., Singapore • Peled, Alva, Israel
• Quiñones, Javier, Guatemala • Reinhardt,Hans-Wolf, Germany
• Rhyner, Kurt, Switzerland • Sayamipuk, Sun, Thailand
• Shah, Surendra, USA • Sherif, C., Germany
• Si Larbi, Amir, France • Taerwe, Luc, Belgium
• Tan, Kiang Hwee, Singapur • Tatsa,Elisha, Israel
• Toledo Filho, Romildo, Brazil • Triantafillou, Thanasis, Greece
• Uomoto, T. Japan • Vázquez, José, Mexico
• Waldslagel, Owen, USA • Wastiels, Jan, Belgium
Local Organizing Committee

Gabi Foerster, Sigrid Hubatsch, Ute Leonhardt, Dorothée Slaats
Conference Secretariat
Dorothée Slaats
X
FERRO-11 – 11th International Symposium on Ferrocement and 3rd ICTRC - Internation-
al Conference on Textile Reinforced Concrete
Preface
Preface
In the year 1854 it was the first time that something about ferrocement was published by
Coignet and Wilkinson and later by Lambot, respectively. Not forgetting the flower tubs
built by Monier in 1847. A period of about 165 years! From these first developments a
product had been generated which is applied in the building industry worldwide until to-
day. Textile reinforced concrete is much younger, means recently 20 years ago. But the
basics of design and the material behaviour are quite similar. Therefore, the conference in
Aachen took over the task to combine the topics of ferrocement and textile reinforced con-
crete. The proceedings at hand show quite clearly that the complete width of the two com-
posites could been covered.
Préface
En 1854, Coignet et Wilkinson ont publié en premier à titre du ferciment, Lambot leur sui-
vait en 1865. A ne pas oublier les caisses à fleur de Monet en 1847. Une période dénviron
165 ans! Il en résultait un produit, employé mondial en génie civil jusqu´à ce jour. Le bé-
ton renforcé en textile est beaucoup plus récent, à savoir seulement 20 ans. Cependant, ses
principes de calcul et son comportement de matériau sont absolument comparables. Cést
pourquoi la conférence à Aix-la-Chapelle se met à réunir les deux sujets, le ferciment et le
béton renforcé en textile. Les actes de conférence montrent clairement que la totalité des
deux composites a pu être traitée.
Prólogo
Las primeras publicaciones acerca del tema de ferrocemento de Coignet y Wilkinson se
hicieron en el año 1854, seguidas de la publicación de Lambot en 1865, y sin olvidar por
supuesto las macetas de Monier de 1847. ¡Un lapso de 165 años transcurridos desde en-
tonces! Este inicio dio pie al desarrollo de un producto, que ha sido utilizado mundial-
mente en el sector de la construcción hasta la fecha. Mucho más reciente es el concreto
reforzado con textiles, que cuenta con tan sólo 20 años de antigüedad. Este nuevo com-
puesto muestra un comportamiento material similar al ferrocemento y parte del mismo
principio para su diseño de construcción. La conferencia en Aachen se dio por ello a la
tarea de juntar los temas de ferrocemento y concreto armado con textiles. Este compendio
reune así un amplio contenido acerca de estos dos materiales compuestos.
Vorwort
Im Jahr 1854 wurde erstmals etwas zum Thema Ferrozement von Coignet und Wilkinson
und 1865 von Lambot veröffentlicht. Nicht zu vergessen sind dabei die Blumenkübel von
Monier 1847. Ein Zeitraum von etwa 165 Jahren! Daraus hat sich ein Produkt entwickelt,
das bis heute weltweit im Bauwesen zum Einsatz kommt. Viel jünger, nämlich erst 20 Jah-
re her, ist der Textilbeton, wobei die Grundsätze der Bemessung und des Werkstoffverhal-
tens durchaus vergleichbar sind. Daher hat es sich die Konferenz in Aachen zur Aufgabe
gemacht, Ferrozement und Textilbeton thematisch zusammenzuführen. Der vorliegende
Tagungsband macht deutlich, dass die gesamte Breite der beiden Verbundwerkstoffe abge-
deckt werden konnte.
XI
11th International Symposium on Ferrocement and Textile Reinforced Concrete 3rd ICTRC
Contents
1 - Plenary Session
Ferrocement: progress review and a critical need for the future page 9
Antoine E. Naaman, Professor Emeritus
Textile reinforced Concrete – from the idea to a high performance mate-

page 15
rial
Silke Scheerer, Frank Schladitz, Manfred Curbach
Ferrocement in civil construction: Its use in Latin America and The Car-
page 35
ibbeans
Dr. Eng. Hugo Wainshtok Rivas
Classification of approaches to dimensioning and assessment of light-

page 45
weight TRC structures
Josef Hegger, Norbert Will, Rostislav Chudoba, Alexander Scholzen
2 - Load Carrying Behavior

Experimental and numerical analysis of TRC composites as tensile rein-
page 53
forcement in concrete shells
Evy Verwimp, Tine Tysmans, Svetlana Verbruggen and Marijke Mollaert
Textile reinforced refractory concrete at high temperature: mechanical

page 63
properties under uniaxial tensile loading
Dimas Alan Strauss Rambo, Flávio de Andrade Silva and Romildo Dias Tole-
do Filho
Bearing behavior of impregnated textile reinforcement page 71

Sergej Rempel, Christian Kulas, Josef Hegger
Flexural behavior of mesh-and-fiber reinforced cementitious composites page 79

P.B. Sakthivel, A. Ravichandran and N. Alagumurthi
Hysteresis parameter identification and reliability assessment of ferro-

page 91
cement walls
Felipe Uribe, Diego A. Alvarez, Jorge E. Hurtado, Bryan Chalarca and Daniel
Bedoya-Ruiz
Behavior of precast ferrocement walls under cyclic loading page 103

Bryan Chalarca, Daniel Bedoya-Ruiz, Diego A. Alvarez and Jorge E. Hurtado
Flexural behavior of mesh-and-fiber reinforced cementitious composites page 111

Rasiah Sriravindrarajah and Micheal Alvaro
1
Strength assessment of sandwich-type ferrocement structural walls under

page 119
cyclic loading
Bryan Chalarca, Daniel Bedoya-Ruiz, Felipe Uribe, Diego A. Alvarez and
Jorge E. Hurtado
3 - Modelling
Nonlinear analysis of TRC shells using an anisotropic, damaged-based
page 133
material model
Ehsan Sharei, Alexander Scholzen, Rostislav Chudoba
Analysis of tensile response of TRC using digital image correlation tech-

page 141
nique combined with multi-scale stochastic modelling
Yingxiong Li, Rostislav Chudoba, Václav Sadílek, Rostislav Rypl, Miroslav
Vořechovský
Implementation of experimental data in analyses of textile reinforced

page 149
concrete structures
Natalie Williams Portal, Mathias Flansbjer, Karin Lundgren, Katarina Malaga
Ultimate limit state assessment of TRC structures with combined normal

page 159
and bending loading
Alexander Scholzen, Rostislav Chudoba, Josef Hegger
Analysis and geometric design of sealing junctions for precast ferroce-

page 167
ment tanks
Henry Hernández Sotomayor, Hugo R. Wainshtok Rivas, Débora Acosta
Elastic-Plastic Analysis of I- Shaped Normal Strength Ferroce-

page 181
mentBeams
Azad A. Mohammed
Study of earthquake resistant behavior in prefabricated ferrocement of

page 199
deposit. Calculation of materials
Henry Hernández Sotomayor, Hugo R. Wainshtok Rivas
4 - Bond
Improving the bond behaviour of textile reinforcement and mortar
page 215
through surface modification
C. Morales Cruz, U. Gohil, T. Quadflieg, M. Raupach and T. Gries
Bond behaviour and durability of basalt fibres in cementitious matrices page 225
Simone Hempel, Marko Butler, Viktor Mechtcherine
Analysis on pull-out behaviour of continuous multi-filament glass yarns

page 233
embedded in cementitious matrix by using a developed model
H. Aljewifi, X.B. Zhang, J. Li
2
5 - Durability
Towards durability design for structural elements and structures made of
page 243
or strengthened with TRC
Viktor Mechtcherine
The Mechanics of Natural Jute Textile Reinforced Concrete page 253

Maria Ernestina Alves Fidelis, Flávio de Andrade Silva, Romildo Dias Toledo
Filho, Steffen Müller and Viktor Mechtcherine
Corrosion in Ferrocement page 265

Paul J Nedwell, Olatona O Damola & Nicholas Stevens
6 - Maintenance and Sustainanbility

Textile reinforced concrete for sewer rehabilitation page 275
Tobias Walther and Manfred Curbach
Combined Formwork and textile reinforcement System for a mineral

page 285
corrosion protection layer for offshore application
Prof. Dr.-Ing. Ludger Lohaus, Dipl.-Ing. Christoph Tomann, Dipl.-Ing. Han-
nes Weicken
Structural concrete elements reinforced with textiles – sustainability per-

page 293
spective
Nahum Lior, Peled Alva and Gal Erez
7 - Measuring Methods
Test methods for Textile Reinforced Concrete page 305
Elisabeth Schütze, Enrico Lorenz and Manfred Curbach
Electrical resistivity measurements on TRC tensile coupons page 317

Vassilis D. Plamantouras and Catherine G. Papanicolaou
Integrated Monitoring of TRC using carbon fibres page 325

Y. Goldfeld, O. Rabinovitch, T. Quadflieg and T. Gries
8 - Application
Slender façade structures made of textile-reinforced high performance
page 335
concrete
Sergej Rempel, Josef Hegger
Mesh Mould: Robotically Fabricated metal Meshes as Concrete Form-

page 345
work and Reinforcement
Norman Hack, Willi Lauer, Fabio Gramazio, Matthias Kohler
3
Restoration of the historical concrete-glass windows of the town-hall in

page 357
aachen
Sergej Rempel, Stephan Geßner
Ferrocement solutions for affordable housing prototypes in India page 363

Prof. Dr. Anupama Kundoo
The morphogenetic design and influential demiurgic qualities of a hybrid

page 373
ferrocement building system
Owen Waldschlagel
Proposed Joints for a Ferrocement Building Resi-dential System page 385

Msc. Julia R. Alvarez, Dr. Hugo Wainshtok Rivas; Eng. Debora Acosta
32 years with ferrocement - from small prefabricated bridges to pontoons

page 393
for floating houses
Carlos A. Gschwind
Ferro Past and Ferro Future page 405

Michael Pemberton
9 - Production Technologies
Locally adapted biaxial warp knitted textiles as reinforcement of folded
page 417
concrete elements
A. Koch, C. Kerschl, T. Gries and W. Brameshuber
Extrusion of textile reinforced concrete page 425

Wolfgang Brameshuber
10 - Strengthening
Textile-Reinforced Mortar (TRM) as retrofitting material of masonry-
page 435
infilled frames
Lampros N. Koutas, Stathis N. Bousias and Thanasis C. Triantafillou
Confrontation of TRC with carbon and stainless steel grids for strength-
page 443
ening of masonry
R. Contamine, J.P. Plassiard and P. Perrotin
Shear behaviour of masonry walls strengthened by external bonded FRP

page 451
and TRC
L. Bui, A. Si Larbi, N. Reboul, E. Ferrier
Shear strengthening of prestressed concrete beams with textile reinforced

page 463
concrete
Martin Herbrand
4
Evaluation of the FIB bulletin 14 design guidline for external TRC bend-
page 471
ing reinforcement
Svetlana Verbruggen, Jan Wastiels, Tine Tysmans, Sven De Sutter, Maciej
Woźniak
Utilisation of linear traction separation bond law to improve the numeri-

page 481
cal simulation of the bending behaviour of hybrid TRC-concrete beams
Maciej M. Wozniak, Tine Tysmans, John Vantomme, Svetlana Verbruggen
and Sven de Sutter
Shear strengthening of RC beams using textile reinforcement in cement

page 489
or epoxy based matrices
Zoi C. Tetta, Lampros N. Koutas, Dionysios A. Bournas, Brwa H. Salihi
Behaviour of hybrid composite-concrete beams under static flexural load-

page 499
ing: a comparative experimental analysis
Sven De Sutter, Tine Tysmans, Svetlana Verbruggen, Maciej Wozniak and
Matthias De Munck
Reinforced masonry using textiles page 509

Dorothea Saenger; Wolfgang Brameshuber
Applications of ferrocement in strengtheining of heat damaged reinforced

page 521
concrete
Danie Roy A.B., Umesh Kumar Sharma and Pradeep Bhargava
Repair corroded reinforced concrete columns using alkali-activated slag

page 529
ferrocement
Shuai Fang, Eddie Siu-Shu Lam and Wing-Ying Wong
5
6
CHAPTER 1
PLENARY
SESSIONS
7
8
FERROCEMENT: PROGRESS REVIEW AND CRITICAL NEED

FOR THE FUTURE
Antoine E. Naaman, Professor Emeritus, Department of Civil and Environmental Engineering,

University of Michigan, Ann Arbor, USA
Abstract: This paper comprises two main parts, namely: a brief summary of key advance-
ments in the field of ferrocement over the past five decades, and an opinion as to the criti-
cal need for an active dedicated professional society to continue to spur future advance-
ments, particularly the development and maintenance of a widely recognized building code
for ferrocement. Progress over the past five decades covers topics such a better under-
standing of the key mechanical properties of ferrocement, their modeling and prediction;
improved properties of component materials; improved construction methods; improved
practices; and some daring applications.While progress has definitely been commendable,
it is nevertheless hindered by the lack of a focused environment and continually fostered
activities such as generally provided by a dedicated professional society that can address
systematically and in details all issues related to ferrocement. These include the develop-
ment of standards, guidelines and codes, cost-effective housing solutions, durability issues,
etc...While it can be argued that such activities can be addressed by existing concrete so-
cieties such as ACI, RILEM, or a number of other institutes generally dealing with con-
crete, efficiency, sustained progress and leap advances seem only possible when a dedicat-
ed institution is involved. It is unfortunate that the lack of funding or endowed funding has
been very detrimental to the full progress of ferrocement and its potential, and, as a result,
has not served the branch of society that could most benefit from it, the public and more
particularly its less fortunate component.
INTRODUCTION
Thin reinforced cement-based composites such as ferrocement and TRC (textile reinforced
concrete) form a special category within the broader family of reinforced concrete. Such
composites can provide slender, light-weight, durable solutions in many structural applica-
tions and are often cost competitive with respect to their methods of production and mate-
rial/structural properties. Their applications cover both marine (boats, pontoons, docks)
and terrestrial structures, as well as repair-rehabilitation for both types. Marine structures
also cover floating structures such as floating houses on river banks or in areas prone to
flooding. Terrestrial structures include mostly small containment structures (silos, water
tanks, waste containment and treatment vessels), and structural and non-structural compo-
nents for various types of construction (housing, school facilities, agricultural structures,
and water channels). A limited number of larger size structures using ferrocement have
been built and the trend is increasing but, unfortunately, it is constrained by legal construc-
tion issues. Indeed, while there is a building code for ferrocement developed by the Inter-
national Ferrocement Society [1], it is limited by the fact that it is not yet adopted and
sponsored by large worldwide organizations such as ACI (the American Concrete Insti-
tute), Fib, or Rilem.
9
SUMMARY-EVOLUTION OVER THE LAST FIVE DECADES
In a prior publication [2] the author had summarized key advancements in the field of fer-
rocement since its modern development in the1960’s. At that time a shift took place
whereas ferrocement, which was seen then as a marine material for amateur boat building
evolved into a construction materials for terrestrial applications as well, particularly small
scale agricultural structures. Below is a brief summary of some key issues discussed in
[2].
1. Evolution in Applications. The most noticeable progress here relates to a signifi-
cant shift from marine applications to terrestrial applications, particularly at first in
small size agricultural structures such as grain silos, water tanks, and housing units.
Today we can cite examples of large scale applications such as the roof of the Siger
monument in Lampung, Indonesia [3], and under construction at time of this writ-
ing, the roof of the Stavros Niarchos Foundation Cultural Center in Athens, Greece
[4].
2. Evolution in Cement Matrix. Here the most notable progress relates to changes in
the composition and workability of the cement mortar matrix as modified by vari-
ous additives or agents; its compressive strength which was commonly in the range
of 20 to 30 MPa, can today exceed 200 MPa in the case of UHPC (ultra high per-
formance concrete). Furthermore, with UHPC the fineness of the matrix and is
flow-ability can be ideal in ferrocement mold-cast applications and for excellent
surface finish.
3. Evolution in Reinforcement. Here progress has been quite impressive on several

fronts. Initially the reinforcement consisted mostly of chicken wire mesh (also
called hexagonal mesh or aviary mesh) made with relatively low strength steel
wire; it evolved into stronger forms of square woven or welded meshes and more
cost-effective ones such as expanded metal mesh. Today the choices are many and
include the use of very high strength steel wires or stands in mat formats, 3D steel
meshes (limited commercial availability), 2D or 3D textiles with non-metallic rein-
forcements, and hybrid combinations of both steel and polymeric reinforcements
with the possible addition of discontinuous fibers.
4. Evolution in Mechanical Performance. Bending resistance in ferrocement plates

increased from about 3 MPa per 1% total volume of hexagonal steel wire mesh re-
inforcement to 24 MPa per 1% total volume of reinforcement consisting of very
high strength steel mat. For these numbers the bending resistance is equal for posi-
tive or negative moment and for the X or Y directions. As a result the cost-
performance ratio has also improved (decreased) significantly. In conventional fer-
rocement with conventional square steel wire mesh, a modulus of rupture of about
50 MPa can be achieved with 7% total reinforcement by volume. A recent investi-
gation has shown that with an ultra-high performance cement matrix and very high
strength steel strands in a mat format, a modulus of rupture close to 230 MPa can
be achieved; the reinforcement consisted of 6.27% by volume of steel mat and
3.5% by volume of micro steel fibers [5]. Note for ferrocement composites with
uniformly distributed reinforcement, the direct tensile strength can be assumed
equal to about one third the bending resistance.
10
5. Evolution in Equipment and Construction Methods. Here the evolution has

been modest and much less noticeable than in other areas. Most applications of fer-
rocement involve hand trowel-ling over reinforcement armature but the introduc-
tion of super-plasticizers and the use of small particles such as fly ash and fine sand
allows ferrocement elements to be cast in molds (as in typical reinforced concrete)
while guaranteeing penetration of the mesh system by the matrix. However, the
most important piece of equipment needed for everyday ferrocement and for both
contractors and the amateur builder is a mechanical mesh tying instrument; while
rebar tying machines exist for reinforced concrete, they have not been tailored and
optimized for use with wire meshes where the opening between wires is relatively
very small. It is believed that the development of such instrument can cut the cost
of labor in ferrocement structures by half. Since generally labor accounts for more
than half of total cost, such an improvement will have an enormous impact on fer-
rocement competitiveness.
6. Evolution in Guidelines and Code. Here progress has been slow and the subject
is in dire need of a serious effort. Committee 549 of the American Concrete Insti-
tute developed two widely used reports, one titled “State-of-the-Art Report on Fer-
rocement,” initially in 1982 and a modified version in 1997 [6], and the other one
titled “Guide for the Design, Construction and Repair of Ferrocement,” initially in
1988 and a revised version in 1993 [7]. The need to develop a technical Code for
ferrocement design was identified in the early 1970’s by numerous researchers and
practitioners, but little progress was made. Related history is given in the next Sec-
tion.
Other progress areas related to modeling, technical and professional activities, education,
people training, and the like are discussed in [2]. In the next section, what the writer con-
siders to be at the moment the most important driver for the progress of ferrocement is ex-
panded upon.
FERROCEMENT MODEL CODE AND CRITICAL NEEDS AND MISSION OF

THE INTERNATIONAL FERROCEMENT SOCIETY
As mentioned in item 6 the above Section, the need for a building code for Ferrocement
has been identified as early as the 1970’s at the onset of modern development of ferroce-
ment. Although some guidelines were developed, the development of a detailed building
code for ferrocement carrying legal weight encountered enormous difficulties. For in-
stance, from a reinforced concrete code viewpoint, how can a type of reinforced concrete
(ferrocement) have almost zero cover to the reinforcement? It seems easy but then how to
put it in a code in such a way that it cannot be “mistaken” or “mis-interpreted” or even
“abused.” Similarly, how can a ferrocement structure be only 10 mm thick while mini-
mum requirements for fire rating in concrete slabs often leads to thicknesses exceeding
about 100 mm. Legally appointed construction inspectors have difficulties with such issues
and unless legally clear tend to dis-approve them. Reconciling reinforced concrete and
ferrocement within the same code turned out to be a daunting task if carried out through
the numerous committees leading to the reinforced concrete building code committee. So it
was concluded that perhaps an entirely separate code focused on ferrocement and thin ce-
ment based composites with steel reinforcements should be developed.
11
In the late 1980’s, a number of pioneers in the field ferrocement entertained the idea to
form a society dedicated to ferrocement with the primary mission to develop a building
code for ferrocement design, construction, and repair. And thus the International Ferroce-
ment Society (IFS) was born. IFS was officially formed in 1991 and registered as an entity
within the Asian Institute of Technology in Bangkok, Thailand. Ricardo Pama, then at
AIT, became the first president of IFS, and Lilia Robles Austriaco, who was also involved
with the International Ferrocement Information Center at AIT, became the first secretary.
To develop a mission statement for IFS broad enough to cover all future expansions, the
mission statement of the American Concrete Institute (ACI) can be first used as a model:
“..The American Concrete Institute is a leading authority and resource worldwide

for the development and distribution of consensus-based standards, technical re-
sources, educational programs, and proven expertise for individuals and organiza-
tions involved in concrete design, construction, and materials, who share a com-
mitment to pursuing the best use of concrete.”
On a grand scale the above mission statement applies to thin reinforced cement composites
such as ferrocement and textile reinforced concrete (TRC). To make it more specific, it is
sufficient to simply replace the word “concrete” by ferrocement or TRC. In its present
form IFS is not a trade association; such an association is typically founded and funded by
businesses; it is industry and niche specific with the objective to interact to the mutual ben-
efits of all its members. So far, IFS has acted mostly as a learned society for the benefits of
all including the public at large, with particular attention to the less fortunate portion of
that public. Since its inception in 1991, IFS has received very little funding from business-
es; its funding was strictly from membership and publication activities such as the Journal
of Ferrocement. IFS can be viewed as the result of an expression of love, faith, enthusiasm
and commitment by a combination of researchers, academics, amateur builders, and busi-
nesses. Thus IFS is a leaned society with a mixed mission: on the one hand to be the repos-
itory of all resources and information related to ferrocement and make them available to
the public, and on the other hand to help develop resources for ferrocement when they are
missing and needed. The most urgent being to develop and maintain building codes, stand-
ards and guidelines for analysis, design, construction, repair and appropriate materials for
ferrocement.
Although the starting activities and progress of IFS are modest it is useful to expand its
potential mission to simulate that of the American Concrete Institute, that is:
“The mission of the International Ferrocement Society is to provide resources
worldwide for the development and distribution of consensus-based standards,
technical resources, educational programs, and proven expertise for individuals
and organizations involved in ferrocement design, construction, and materials,
who share a commitment to pursuing the best use of ferrocement.”
Note that the above mission statement could be similarly applied to textile reinforced con-
crete (TRC) should a related international society be sought for that material. During the
mid-1990’s, IFS organized several committees to help develop reports and recommenda-
tions for various activities; they included: the Ferrocement Model Code committee, the
Education Committee, the Corrosion and Durability Committee, the Housing and Terres-
trial Structures Committee, the Advanced Materials Committee, and the Seismic Applica-
12
tions Committee. These were considered subjects of most importance at the time and are
still relevant to today’s needs. The Ferrocement Model Code was first published by IFS in
January 2001 [1]. History would likely conclude in the future that the onset of the third
millennium was not kind to ferrocement. IFS encountered several difficulties; key board
members who were working at AIT in Bangkok, retired. Shortly thereafter, AIT cut out
funding and the physical facilities of the IFS and its right arm, the International Ferroce-
ment Information Center (IFIC) were taken over for other activities. All physical docu-
ments, paper prints, reports, photos, and journal archives of IFS and IFIC were then stored
in a room on the floor of the structural engineering laboratories at AIT awaiting further
decision as to their future permanent housing. Then disaster struck in 2011 when Bangkok
was flooded by torrential rains; all the above documents were flooded and stayed in water
for many days. For all practical purposes they were destroyed and later, when attempt was
made to remove all damaged materials, they were too damaged and simply discarded.
Members of the Board of IFS decided then to recreate a repository of information related
to ferrocement by collecting all hard copies documents available to them (personal library
or other) and the membership, and digitize them. This activity is being carried out at the
University of Manchester by Paul Nedwell while simultaneously a web site is being built
where such information will be stored. Since 2011, the IFS society has been acting mostly
virtually awaiting the completion of the web site in Manchester, and the establishment of a
base at the University of Manchester in the UK. In the meantime, the most urgent need
remains most critical: the Ferrocement Model Code, first published in 2001, needs to be
updated and expanded. It is hoped that such activity will be given top priority by the socie-
ty.
CONCLUDING REMARKS
None of the recognized advantages of ferrocement have changed since the onset of its
modern development in the 1960’s, while both its cost-performance and mechanical effec-
tiveness have improved with the introduction of new materials and systems. It seems that
the main deterrent to its progress remains the establishment of a widely accepted and up-
dated building code for its design and construction. Amateur builders and self-help practi-
tioners will continue to use ferrocement in small scale applications, but opportunities for
all other applications are hindered by the absence of such a code. The objective of the pro-
posal for an international technical society (such as IFS) will foster a related discussion at
large among professionals that should lead to some tangible actions. On one hand, dedicat-
ed focus and independence will guarantee the production of a code document for ferroce-
ment; on the other hand, to receive wide acceptance, the adoption of such code by a larger
umbrella organization such as ACI or Rilem will be eventually necessary. This could be
the only route to follow in the future to guarantee the worldwide adoption of a code.
REFERENCES
1. IFS Commitee 10, “Ferrocement Model Code: Building Code
2. Recommendations for Ferrocement” IFS 10-01, International Ferrocement Society, Asian Institute
of Technology, Bangkok, Thailand, 2001, 95 pages.
3. Naaman, A.E., “Ferrocement: Four Decades of Progress,” in Proceedings of International
Symposium on Ferrocement and Thin Cement Products; Edited by P. Nimityongskul, A.E. Naaman,
J.E. Bolander, C. Jaturapitakkul, S. Sujivorakul, and S. Sayamipuk; Thai Concrete Association,
Bangkok, Thailand, February 2006, pp. 3-21. ISBN 974-93905-1-X. Also published in Jounral of
Ferrocement. Journal of Ferrocement, Vol. 36, No. 1, January 2006, pp. 741-756.
13
4. Baysoni, and Djausal, A., “Structural Design of Ferrocement for Lampung’s Siger Monument in
Indonesia,” Proceedings of 9th International Symposium on Ferrocement (Ferro-9), The University
of Lampung, Bandar Lampung, Indonesia, May 2009, pp. 177-182.
5. Martin, B., “Innovating with Ferrocement: Stavros Niarchos Foundation Cultural Center,” a pdf file
accessible at: http://www.concretecentre.com/pdf/StavrosNiarchos_BruceMartin.pdf; ”, mpa - The
Concrete Center.
6. Wille, K, and Naaman, A.E., “Preliminary Investigation on Ultra-High Performance Ferrocement,”
in Proceedings of 10th International Symposium on Ferrocement and Thin Reinforced Cement
Composites, La Havana, Cuba, Edited by H. Wainstock Rivas, L. Prada Seoane, and I. Randa
Castro, ISBN: 978-959-247-098-9, October 2012, pp. 251-260.
7. ACI Committee 549 (1997). “State-of-the-Art Report on Ferrocement.” Concrete International, Vol.
4, No. 8, Aug. 1982, pp. 13-38. Reinstated as ACI 549-R97, in Manual of Concrete Practice,
American Concrete Institute, Farmington Hills, Michigan, 26 pages.
8. ACI Committee 549, (1988, …). “Guide for the Design, Construction and Repair of Ferrocement.”
ACI 549.1R -88, ACI Structural Journal, May - June, 1988, pp. 325 - 351. Also reinstated as
549.1R-93 and published in ACI Manual of Concrete Practice, American Concrete Institute,
Farmington Hills, Michigan, 27 pages.
14
TEXTILE REINFORCED CONCRETE – FROM THE IDEA TO A

HIGH PERFORMANCE MATERIAL
Silke Scheerer, Frank Schladitz, Manfred Curbach, Institute of Concrete Structures, Technische
Universität Dresden, Germany
Abstract: Textile reinforced concrete (TRC) is on everyone's lips. But who had the idea for
this new building material? This question was the intention of this article, however, the
question cannot be answered clearly and the authors do not even claim to be able to give a
complete overview. The article presents early approaches for the reinforcing of building
materials and structures. It presents selected alternatives to conventional reinforcing steel
that has been constantly developed, especially since the 1980s. Then this article takes a
much closer look on research and developments in the run-up to the two major Collabora-
tive Research Centres (CRCs, in German: SFB) for textile reinforced concrete in Dresden
and Aachen. In these two collaborative projects, the research on the composite material
was mainly pushed. The CRCs will be presented in their basic features and it will be
demonstrated how the transition from theory to practice took place. Other aspects of this
last part of the article are the focal points of recent research initiatives and perspectives for
the use and the application of carbon in the building industry.
INTRODUCTION
The use of relatively brittle materials with a low tensile strength such as clay or concrete
has been known for thousands of years. It is almost a philosophical question of when the
first material was produced that is comparable to "concrete" in the modern sense [1]. The
idea of embedding fibres in brittle materials, so that cracking does not lead to failure, is
also not new. Even thousands of years ago straw was incorporated to clay brick (e.g. left in
Figure 1) in order to make it more resistant to cracking e.g. due to strong temperature fluc-
tuations during the day [2]. Later, other methods were devised to strengthen solid concrete
structures. References [3] and [4] for example, report on iron brackets on head joints al-
ready in antiquity, e.g. at Rome’s Colosseum or at Porta Nigra in Trier, and in [4] on first
tensile rings used at brick domes. Best-known early examples are the wooden tensile ring
made of chestnut logs at the dome of Florence Cathedral and the iron tensile ring around
the dome of St. Peter's Basilica in Rome [5]. At the end of the 16th century iron ring an-
chors were already installed in this amazing dome that should take the bow thrust.
“Was the use of iron pieces to reinforce concrete in the middle of the 19th century really
the most obvious? One might suspect it, ... ", Wieland Ramm writes in [4]. The first pa-
tents date back to the year 1854 (Coignet and Wilkinson) and 1865 (Lambot) and were
issued in England, France and Belgium. The beginning and the subsequent history of the
"Building Material of the Century" reinforced concrete as the leading building material of
modern times is well known, see [4]. Progress in design and construction, materials and
construction methods have established reinforced concrete as the leading building material
in almost all construction projects. However, there is one major disadvantage: reinforcing
steel can corrode. Following Wieland Ramm one has to ask oneself the question, whether
the search for alternative reinforcement materials was not a logical consequence of previ-
ous developments.
15
Figure 1: Fibre reinforced materials; on the left side: adobe bricks for house construction in Kyrgyzstan (pho-
to: Vmenkov, CC BY-SA 3.0, Wikimedia commons), middle: UHPC with steel short fibres, on the right side:
crack in a specimen of high ductile concrete (photos: Silke Scheerer)
ALTERNATIVE REINFORCING MATERIALS

In recent decades, various ways have been developed to replace the conventional reinforc-
ing steel. Some will be described briefly. There is no exact date when the development of
short fibre concrete or fibre concrete (Figure 1, centre) was intensified as an alternative to
steel reinforced normal concrete. It is, however, certain that the beginning of this compo-
site material dates back as early as the starting time of reinforced concrete construction. In
reference [6] asbestos cement is described as "the first diverse, popular, and industrial level
cement composite material”. Today, it is mainly alkali-resistant glass fibres (AR glass fi-
bres), steel or plastic short fibres that are added to the matrix of short fibre concrete. In
addition to an increase in the compressive and tensile strength, the fibres mainly have a
positive effect on the flexural strength, which is why the material is often used for highly
loaded industrial floors. Other applications are concrete products or thin-walled precast
elements ([7], paragraph 6). A more recent development is the ultra-ductile concrete, e.g.
[8], right in Figure 1. A disadvantage is that the fibres usually are randomly oriented.
Therefore, short fibres cannot be used for a load transfer in a defined direction.
Another alternative development is the use of fibre reinforced plastics (FRP) in concrete
structures, see references [7], [9], [10], and [11]. In general, fibre reinforced plastics in-
clude materials that are made of plastic-embedded fibres, for example carbon, glass, and
aramid, which have hitherto been mainly used in lightweight construction. Such a matrix
achieves the full utilisation of the high efficiency of the individual fibre material within the
construction. FRPs usually have clearly directional strength and stiffness properties. A
variety of publications in journals and proceedings verifies early research on fibre materi-
als in Japan like the CSCE, the Japan Society of Civil Engineers (JSCE) which established
a “committee on continuous fibre reinforced materials” in 1987 [11]. A State-of-the-Art
report of this topic was already published in 1992 by JSCE. For solid construction, carbon
fibre reinforced plastics CFRP are especially interesting. Currently CFRP can mainly be
found in the form of sheets as well as bars and ties, see references [7], [11], [12], [13] and
Figure 2 in the middle and on the right. The joint research project “C3 – Carbon Concrete
Composite” has the aim to significantly expand the field of application of carbon in con-
struction (see paragraph 6). It should be mentioned that reinforcing bars made of glass fi-
bre reinforced plastics (GFRP rebars left in Figure 2) have been developed and applied, see
reference [14]. Currently there is only one product approved by the building authority in
Germany [15]. Considerations to replace steel reinforcement by the use of continuous fi-
bres or grids that were made from continuous fibres already began in the 1980s, see refer-
16
ence [7]. Among experts this new, innovative composite building material is known today
as textile reinforced concrete or TRC. The development shall be represented in more detail
below, but there is no claim to completeness.
Figure 2: on the left side: ComBar − GFRP reinforcement bars from Schöck
(http://www.schoeck.de/de/produktloesungen/typen-zubehoer-101), middle: carbon tendon [13], on the right
side: strengthening with CFRP strips (http://www.bautas.ch)
THE BEGINNING OF RESEARCH ON AND DEVELOPMENT OF TRC

First activities and patents
Considerations about the use of textiles as reinforcement for concrete have already started
in the early 1980s. Reference [16] refers to activities of the Sächsisches Textilforschung-
sinstitut e.V. STFI1, Chemnitz, and the WTZTT2 Dresden as well as to the first patents.
Patent DD 210102 (December, 1982) offers and patents a possibility for the substitution of
steel reinforcement for transport safety (Figure 3). This applies, for example, for minor
building products such as plates for slops or for tracks for agricultural infrastructure con-
structions. Shortly later, concrete barriers were developed with rope-like reinforcing ele-
ments made of production waste, and combined with pieces of textile or fibre concrete
[17], patent no. DD 253,442 (20/01/1988). Patent no. DD 275 008 "Procedures for fixing
the position of line-shaped flexible reinforcement elements in the vertical machining of
plate-like elements" (also spacers), describes examples of textile reinforcements: "reinforc-
ing strands" with a diameter of 8 mm and a mesh size of 25 mm. It gives a first impression
on textile reinforced concrete.
1: single component
2: reinforcement with
3: heavily perforated, outer sheath
4: slidable sleeve
5: fibrous sheath
6: tension carrying core (here: slidable)
Figure 3: Early textile reinforcement element, taken from DD 210102
1
Saxon Textile Research Institute
2
WTZ Technology Textiles
17
The following discussion will mainly focus on the research and development of the com-
posite building material textile reinforced concrete TRC. This would not have been possi-
ble, of course, without trend-setting developments in other relevant disciplines. To mention
all technological progresses in textile technology, fibre production and fibre processing,
concrete design and modelling – to name just a few – would, at this point, exceed the scope
of this article by far.
Research in the 1990s in Germany

Initial research projects followed first utility models and patents soon. In Saxony, for ex-
ample, there was a project funded by the BMFT at the beginning of the 1990s at STFI
Chemnitz and the ITT Dresden. In this project, the suitability of flax fibres that had been
converted into grid-like knits, was investigated for use in reinforced lightweight concrete
or plaster elements [18]. At that time topics that influence the effectiveness of textile grids
/ fabrics embedded in fine concrete, were already the main subjects of research, e.g. the
interaction between surface and core fibres, the bond or the alkali resistance.
Other projects followed. In order to achieve rapid progress in exploring the performance of
the new building material, it was essential to cluster expertise. Early on researchers of the
universities in Dresden and Aachen worked closely together and are still leading in the
textile reinforced concrete research. The initial cooperation can be demonstrated by two
joint research projects, while the second project in particular was decisive for further re-
search. In 1996 the AiF-Project No 9272 "Knitted reinforcement fabrics for textile con-
struction" was completed [19], which was handled by the Institute of Textile Technology
at RWTH Aachen and the Institute of Textile and Clothing Technology at the Technische
Universität Dresden (now: ITM, Institute of Textile Machinery and High Performance Ma-
terial Technology). Involved in another joint AiF-research project (No. 10378 B) on "Fun-
damental studies on the development of knitted reinforcement scrims for the use in thin-
walled parts made of mineral building materials" were also the Institute of Structures and
Materials (now: Institute of Construction Materials, TU Dresden), the Sächsisches Textil-
forschungsinstitut3 e.V. (STFI, Chemnitz) and the Institute of Textile and Clothing Tech-
nology Aachen [20]. In 1996 -– after an early recognition on a TechTextil Conference in
the 1990s – a construction made of textile reinforced concrete won a price for the first time
and set an example at least among the professional public: the design prize of the
Deutscher Betonverein e.V.4 for a textile reinforced concrete boat that students had been
built for the Concrete Canoe Regatta of 1996 in Dresden. The students of the concrete boat
team were supported by several research institutes of the TU Dresden with technical advice
and working space. These include the Institute of Construction Materials and the Institute
of Concrete Structures with their experience and the latest research results regarding the
building with concrete. Furthermore it was also the textile researchers, who, according to
[21], had recently succeeded to process the world’s first continuous yarns from alkali-
resistant glass to knits or fabrics by using the Malimo-technique, which were then used as
reinforcement in the boats. Thus, two boats emerged for the design competition: "August
the Strong" with a reinforcement of AR glass and the "Tender Gustel" with a reinforcement
of carbon fibres (left in Figure 4). Subsequent to the competition the "Tender Gustel" was a
popular exhibit and travelled to Amsterdam, Atlanta and Ulm [22].
3
Saxon Textile Research Institute
4
German Concrete Association
18
Figure 4: Beginning of textile reinforced concrete research; left: the first Dresden textile reinforced concrete
boats in 1996; right: early multi axial fabric made of AR glass fibres (the left one: WIMAG S3, Liba-
technique, Aachen; the right one: Malimo-technique, Dresden); [7]
At this point it was clear: building with textile fabrics has a large potential. Based on earli-
er research, further and sometimes smaller projects were initiated at various institutions of
both universities. First doctoral dissertations were written, for example [23], [24]. The col-
lective knowledge was then collected in the "Gemeinschaftsforschungskreis Textil-
bewehrter Beton5" of the Deutscher Betonverein e.V.4 and summarized in 1998 into the
"Sachstandbericht zum Einsatz von Textilien im Massivbau6" [7]. Contributors were not
only researchers from Dresden and Aachen but also colleagues from the University of
Stuttgart, TU Darmstadt, Dywidag and from STFI Chemnitz. The report includes a sum-
mary of preliminary research results on the structural composite textile reinforced concrete.
In the following the topics are mentioned briefly (paragraph 7 in [7]):
RWTH Aachen: As preliminary work for a planned Collaborative Research Centre

(CRC), flexural and tensile properties of textile reinforced, flat and U-shaped specimens
were carried out with different reinforcement. A typical early textile can be seen right in
Figure 4, first test specimen in
Figure 5. In addition to the ultimate bearing capacity and basic information on the state of
use it was essential to learn more about the bonding behaviour, the ductility or the use of
theoretically possible tensions.
0: reference without reinforcement 2: short fibres (1.5 mass-%) and 2 textile layers
1: short fibres (1.5 mass-%) 4: 2 textile layers all around
5
Community Research Group Textile Reinforced Concrete
6
Status report on the use of textiles in Concrete
19
Figure 5: First experiments on textile reinforced concrete specimen on RWTH Aachen: specimen for tensile
and bending tests with different reinforcement ratios, [7]
TU Dresden: First experimental tests were carried out on hollow blocks, thin TRC-plates,
and on reinforced concrete slabs that had been reinforced with textile reinforced concrete.
The hollow blocks were reinforced with textiles made of AR glass fibres and carbon fibres,
the reinforcing layers in the bending tests were only reinforced with AR glass fibres
(Figure 6). Impressive is the significantly increased load capacity of reinforced compo-
nents. Since the existing textiles were not as load-bearing as those today, many layers were
applied, but the cross-sectional thickness itself was increased by only 10 ... 15 %. This is
contrasted with an increase of the load bearing capacity by 75 %. In addition, with the
same applied load there were significantly smaller deflections up to the time of failure of
the textile reinforced concrete layer compared to the reference that has not been reinforced.
Figure 6: First specimens and test results; on the left side: window lintels with polystyrene core; on the right:
results of 3-point bending tests on RC-plates without and with TRC-strengthening layer on the bottom side,
according to [25]
Since the mid-1990s there were also research activities in Stuttgart and Darmstadt. Prof.
Reinhardt and colleagues (University of Stuttgart) dealt with aramid textile reinforcements
and later on with the prestressing of TRC, e.g. [26], [27]. The researchers led by Prof.
Wörner (TU Darmstadt) dealt with the use of surface near textile grids to restrict the crack
width, e.g. [28].
International research on textile reinforced concrete

A complete overview of international research activities in the field of textile reinforce-
ment of concrete cannot be given. Though, there are three examples of research groups /
regions that should be noted.
University of Surrey (UK), Prof. D.J Hannant and colleagues: known are various publica-
tions, for example on reinforcement webs made of polypropylene fibres from the 1980s.
The Ohno-Hannant effect is often mentioned with regard to explanations of the behaviour
of fibre bundles in a mineral matrix. The effect describes the interaction of the outer fibres
with direct bond to the matrix and the inner core fibres that can only partially absorb
strains [29].
Ben Gurion University of the Negev and the Technion (Israel), e.g. [30]: At the beginning
of the 1990s Prof. Alva Peled, Prof. Arnon Bentur and colleagues have studied the influ-
ence of the structure of fabrics on the mechanical properties of composite materials made
of textiles in a cement matrix. Until today, there is a close research exchange between re-
20
search institutes in Israel and Germany. Prof. Peled is one of the world's leading scientists
in the field of textile reinforcements for concrete and is so far the only one with first expe-
riences on the use of mineral fillers for textile structures [31]. She already stayed several
times as a guest scientist in Dresden, recently in 2014 at the Institute of Construction Mate-
rials.
Japan also began quite early with very extensive research on the application of alternative
fibre materials in concrete, see paragraph 2. To name just two examples: one patent (1987)
and a research paper (1990). Reference [32] describes the use of fibres embedded in resin,
which could then be used as a grid-like reinforcing element which could be made of vari-
ous fibre materials such as glass, carbon or aramid. In [33] results are presented for the use
with carbon fibre networks in a mineral matrix.
One can only guess the existing research worldwide. For further information please see [7]
or [10] and proceedings of relevant symposia and the according literature. However, inter-
national activities do not exceed the comprehensive and profound research carried out by
the two Collaborative Research Centres7 (CRCs) funded by the Deutsche Forschungsge-
meinschaft8 DFG in 1999. There is not only the combination textile fabrics and concrete
but also the strengthening of masonry (e.g. [34], chapter 22) and of wooden structures [35].
In the 1990s there have been first textile reinforced prototypes and first applications of
textile fabrics used for structures. For further information please see the references, e.g.
[7], [36].
FUNDAMENTAL RESEARCH IN THE DFG-CRC 528 AND 532

Based on the wide preparatory work two Collaborative Research Centres (CRCs, in Ger-
man: SFB) were prepared in the late 1990s in Dresden and in Aachen and were presented
to the German Research Foundation DFG. Despite their formal similarity both projects
were recommended by the experts for funding. This was a unique opportunity to explore a
wide research field in different locations simultaneously with combined efforts. To exclude
a double funding, the individual research fields were strictly defined during the application
stage. This is also reflected in the titles of the respective CRC:
• CRC 528 (Technische Universität Dresden; speaker: Manfred Curbach, Institute of
Concrete Structures): „Textile reinforcements for structural strengthening and re-
pair“ [37],
• CRC 532 (RWTH Aachen University, speaker: Josef Hegger, Institute of Concrete
Structures): „Textile reinforced concrete − basics for the development of a new
technology“ [38].
Both CRCs were funded by the German Research Foundation (DFG) between 1999 and
2011. The main research topics will be presented shortly on the following pages.
CRC 528 in Dresden

The researchers in CRC 528 had set themselves the goal to provide a verified basis for the
use of textile reinforcements for strengthening and repair in construction. The scientists
conducted research regarding materials and their mechanical description, regarding design
and constructive details. Of interest were further the technological application, the tech-
7
In German: Sonderforschungsbereich SFB
8
German Research Foundation
21
nical construction work, and the long-term properties of TRC and thus the safety and dura-
bility [39].
Figure 7: Project structure of CRC 528, according to [39]
There were four planned funding periods with twelve years runtime. In the beginning theo-
retical and experimental studies were conducted to develop basic knowledge of the source
materials themselves, the bond between them and the bond between the old concrete and a
new textile reinforced concrete layer under static loading. In addition, the potential of tex-
tiles to compensate anisotropies in wood and the resulting differences in strength and stiff-
ness were examined. The research fields were later expanded and partially modified. The
transfer of the new material TRC into the construction practice became more and more
prominent since the third funding period.
The project structure in its main features is shown in Figure 6, see also [39], [40]. The ex-
perimentally oriented sub-projects are marked in blue, orange stands for the more theoreti-
cal projects. The capital letters in turn, represent one of the five big research areas:
• Research field A: material-oriented investigations of textile structures and fine
concrete matrices (mesoscopic level). Over the entire period of the CRC various
fine grained concretes were developed, varied and refined. Regarding the rein-
forcement, AR glass, carbon, aramid and polypropylene were used as fibre ma-
terials. Both, the fibres as well as the production of textiles including machine
design and development of coatings were partial aspects. The experiments start-
ed with mostly static loading in period I. Later the project focused on dynamic
and long-time effects.
• Research field B: Composite textile reinforced concrete (macroscopic level). In
the beginning the uniaxial stress-strain behaviour was at the centre. Later biaxial
stress conditions and dynamic loads were considered in addition. In Phase III
and IV of the CRC statistical analyses were performed (in view of a safety fac-
22
tor) and intensive research was done on anchoring and overlapping lengths,
since without this building with TRC would be impossible.
• Research field C: Bond between new and old. In the C projects the theoretical
and experimental groundwork was laid in terms of the interaction between an al-
ready existing old concrete layer and a new TRC layer. Topics examined include
the bond zone, the production technology and the problem of stress transmis-
sion. The combination of reinforcing textiles and wood has also been studied
during the first two periods of the CRC.
• Research field D: Strengthening of structural components. Starting with the sim-
plest example of application – the flexural strengthening of reinforced concrete
slabs – the basis was set for the strengthening of structural components for the
load cases bending, longitudinal forces, shear, torsion and combinations of these
loads. Later, the fire loads and initial tests with respect to cyclic loading were
added. Based on numerous experiments and theoretical considerations, engineer-
ing models were created to calculate the strengthening effect with the help of
which the first applications have been realised, see e.g. [41], [42]. In the project
area D photogrammetric methods have also been developed that were almost
uncharted territory at the beginning of the CRC 528 and are now well known.
• Research field E: Safety and durability. Research field E addressed questions
about the remaining service life of components, which were strengthened with
TRC, as well as about the durability of the materials used. In parallel, the basic
principles of a security concept for the strengthening with textiles as reinforcing
structures were also covered.
In addition, accompanying research projects have been carried out, a part of which is also
shown in Figure 7. First prototypes and practical applications followed (see chapter 5).
Besides that, transfer projects were initiated to connect the basic research with the building
practice more and more. A total of four faculties with 11 institutes and one non-university
institution were involved in the research within the CRC 528. Within the duration of the
CRC 528, 48 doctoral dissertations and several hundred publications has been written [37].
CRC 532 in Aachen

In the Collaborative Research Centre CRC 532 (Aachen) "Textile Reinforced Concrete −
basics for the development of a new technology" the researchers focused their attention −
in addition to basic research similar to the Dresden´s CRC − especially on textile rein-
forced concrete in terms of new components. The number of sub-projects and participating
institutions was comparable. One can also find parallels in the general project structure,
[43] and Figure 8:
• Research field A: Filaments / yarn,
• Research field B: Textile,
• Research field C: Concrete,
• Research field D: Composite,
• Research field E: Structural members and production,
• Research field Y: Technical information system.
23
Figure 8: Main topics of CRC 532, acc. to [38]
Some points, where the research in Aachen is different from that in Dresden, are:
• At the beginning of both CRCs one dealt on both sides essentially with rein-
forcements made of AR glass. Later Dresden´s researchers focused on the mate-
rial carbon. This was partly due to the fallen price of the fibres in the meantime
and on the other hand to the higher performance and durability of carbon for the
planned applications. In Aachen, however, the research on AR glass fabrics was
deepened. A special feature are the by impregnation with resins stabilized, pre-
fabricated reinforcement cages which are optimal as formwork for prefabrication
and precast constructions (Figure 9).
• At RWTH Aachen there was a major focus of research on the production, which
was discussed in Dresden just on the outskirts or in accompanying, smaller pro-
jects.
• Furthermore, in Aachen a technical information system for long-term research
and application of TRC was established. It is based on a cross-institutional and
interdisciplinary database. Based on this a freely accessible internet platform
was created in period IV of the CRC 532 [38], that is currently being completed.
Here one can get information about, for example, textiles, building products or
calculation methods.
Figure 9: AR glass reinforcement cages for precast elements (RWTH Aachen, photo: ibac, [44]) and flexible
carbon textile for variable application during strengthening procedures (TU Dresden, photo: Ulrich van
Stipriaan)
Common activities
During the preparation of the two CRCs there was already a very close collaboration be-
tween the Universities of Aachen and Dresden. Since then the exchange of experiences
was encouraged. Noteworthy are the former six joint technical conferences „Colloquium
24
on Textile reinforced Structures (CTRS)“, which were carried out alternately in Aachen
and Dresden with national and international participation since 2001. 2009, the first “user
conference” on TRC took place directly after the CTRS4. This event will take place in
2016 for the seventh time [45]. Furthermore, joint publications were published (Figure 10),
e.g. the RILEM State-of-the-Art Report about textile reinforced concrete [46], the 2004´s
June issue of the peer-reviewed German journal Beton- und Stahlbetonbau that was com-
pletely dedicated to the topic TRC [47], and a special issue of the same journal to a first
general building approval [48]. Another RILEM report – RILEM 232-TDT „Test methods
and design of textile reinforced concrete“ – is being created currently headed by Prof.
Brameshuber (RWTH Aachen) [49].
Figure 10: Important common publications, [7], [46]–[48]
THE STEP TO PRACTICE

Networking
The transfer of research results into practice has become increasingly important since the
mid-term of the two CRCs. On the one hand there were transfer projects within the Collab-
orative Research Centres (see e.g. T-projects in Figure 7). On the other hand there were
several projects with industrial partners, funded for instance within the funding “Zentrales
Innovationsprogramm Mittelstand ZIM”9. Thus, the way into building practice was opened
for TRC. Another important point was the establishment of the Deutsches Zentrum Textil-
beton DZT10 [50] in 2007. It was part of the model project "Exchange Platform for CRCs
and Clusters of Excellence", initiated by the Stifterverband für die Deutsche Wissen-
schaft11 [51]. For the first time the transfer of new technology from science to industry
was specifically promoted parallel to its basic research. As part of the TU Dresden AG
(TUDAG), the DZT is a competence centre and a focal point for all companies who are
interested in TRC. Expertise and networking of all actors along the value added chain play
an important role for a successful market launch of the textile reinforced concrete. For this
purpose, the TUDALIT e.V. was founded in 2009 [45]. This registered association has
currently 28 members from industry and science. Association goals are networking and
especially the jointly working on general construction approvals (abZ) in the field of textile
concrete to establish the name TUDALIT® as a quality brand. Among others, the texton
e.V. and working groups of the Alliance fibre-based materials Baden-Württemberg e.V.
and the Carbon Composites e.V. deal with the topic TRC today.
9
Central Innovation Programme SME; SME: small and medium-sized enterprise
10
German Centre Textile Reinforced Concrete
11
Donors' Association for German Science
25
Brief presentation of selected projects

A complete overview of the practical activities during both CRCs until today would fill a
book. Therefore, this paragraph will mention briefly some exemplary projects in which the
TU Dresden and the RWTH Aachen were significantly involved. This overview is intend-
ed to give an idea of the various application possibilities. For more information, reference
is made to the numerous existing publications, e.g. the homepages of the CRCs or [41],
[42], [48], [52].
Façade panels. Curtain walls of reinforced concrete are at least between 7 and 10 cm thick
because of the necessary concrete cover. With TRC a thickness of 2 to 3 cm is often suffi-
cient. The much lighter façades find a very wide application spectrum, because, among
other things, the substructure can be executed filigree and transport and assembly are less
expensive. In 2002, 240 m² façade surface of a new building for the Institute of Concrete
Structures of the RWTH Aachen University were covered with small-scale textile concrete
slabs developed in CRC 532. The company HeringBau received the first general construc-
tion approval for such components in 2004 [53]. The pilot application of large-format fa-
çade elements was carried out in 2008 at the construction of a new test facility at RWTH
Aachen [54]. Many other applications followed, e.g. Figure 11, where large façade panels
were used at the cafeteria of St. Leonhard Gymnasium in Aachen.
Balkony slabs. Often balconies are placed in front of existing residential buildings, con-
sisting of a metal frame construction and accessible reinforced concrete plates. The weight
of the 20–25 cm thick reinforced concrete slabs is often decisive for the design of the sup-
porting steel structure, the costs of transport and the choice of a suitable crane. In a ZIM
project, TRC balcony slabs were developed, which had an average thickness of approxi-
mately 7 cm [55], Figure 11. As a result, the steel structure can be designed filigree, mate-
rial can be saved and the transportation costs and the necessary bearing capacity of the
crane can be significantly reduced.
Shell structures. A material that is both lightweight and very load bearing is ideal for shell
structures. Several buildings were realized with Dresden´s and Aachen´s participation,
ranging from small roofing to complete pavilions (Figure 11).
Spacers. Today, TRC is often produced in lamination process, in which layers of fine
grained concrete and textile fabric are built in a horizontal formwork step-by-step until the
required component geometry and number of layers is reached. Much more efficient is the
casting method that is usual for reinforced concrete. Within a ZIM project, unique spacers
have been developed, which are specially adapted to thin and filigree textile fabrics, [56]
und Figure 11.
Pontoons. Pontoons for working platforms, floating homes or boat launches are usually
made of steel, in some cases made of reinforced concrete. The Institute of Concrete Struc-
tures of TU Dresden and the company Dyckerhoff developed textile reinforced concrete
pontoons (Figure 11) and evaluated their abilities in a practical test. The Pontoons are thin
walled, not susceptible to corrosion and still very viable. They consist of plane plates
which have been glued together. The pontoons are a good alternative to existing solutions
in terms of price.
26
Heating with carbon reinforcement. Carbon fibres are electrically conductive. In a ZIM
project, a reinforcement-heating has been developed which takes over supporting and heat-
ing functions simultaneously [57]. This system can possibly save a separate heating at TRC
houses in the future. After extensive testing, a first element with a wall heating was in-
stalled in a laboratory building at HTWK Leipzig. Figure 11 shows the installation of this
element.
Figure 11: New components made of textile reinforced concrete (from top left to bottom right): TRC façade
panels at a high school building in Aachen[48] (photo: Robert Mehl); balcony slab on the site of Metallbau
Guke GmbH in Seerhausen, Saxony (photo: Michael Frenzel); 14 × 14 large roofing made of TRC prefabri-
cated elements, RWTH Aachen [52]; special spacers for textile reinforced concrete DistTEX (photo: fil-
maton); pontoon at the first practical test (photo: Frank Schladitz); textile reinforced concrete wall element
with integrated heating (photo: Elisabeth Schütze)
Strengthening of structures. In addition to the previously presented new components

made of TRC, textile-reinforced concrete is mainly used for the strengthening and rehabili-
tation of existing reinforced concrete structures. On this topic, the researchers have special-
ized at the Technische Universität Dresden, see e.g. [41], [42], [59]. Columns, roofs, floor
slabs, but also a monument, a sugar silo (Figure 12) and balcony structures were already
strengthened. Other applications are currently in planning and execution, e.g. the rehabili-
tation of first sewage structures with textile reinforced concrete [58]. Meanwhile, TRC is
an option even in the more unusual projects in conservation. An impressive example is the
restoration of glass block windows from the 1960s. These have been installed after the
Second World War in the Gothic St. Mary's tower of Aachen City Hall [52]. Later, the
steel reinforcement used at that time was corroded locally. Spalling was the result. The
solution was the use of carbon textiles impregnated with epoxy resin embedded in fine
concrete (on the right in Figure 12).
27
Figure 12: strengthening and restauration with TRC; application of the pre-tailored carbon textile in a twin
compartment silo Uelzen (photo: Silvio Weiland) [59] and reconstructed historical windows inside of the
Mary's tower in Aachen [52]
Approvals
TRC is not yet a generally approved building product and its workmanship is still no gen-
erally approved construction method. Approvals for individual cases (ZiE) were thus re-
quired in previous applications in building practice. To avoid that one need separate ap-
provals for each single building project, general technical approval (abZ) are necessary.
First general approvals were given for small façades panels (see chapter 5.2). In June 2014,
there was a first abZ for the flexural strengthening of reinforced concrete structures [60],
[48]. This approval is the result of extensive evidence about the characteristics of fine
grained concrete matrix, the special carbon reinforcement and the bond between these two.
The approval is currently limited to flexural strengthening of indoor components under
predominantly static loading. Nevertheless, the document is an important milestone. Ex-
tensions to the existing approval should now be possible with much less effort. In addition,
one can refer to this existing document for future approvals. This approval is linked to the
quality brand TUDALIT® and was financed by the members of TUDALIT e.V. The ap-
proval holder is the TUDAG. The TUDAG is also responsible for the granting of licenses
that are needed to use the approval. A further requirement is that the user can show a certif-
icate, which can be purchased in a two-day TRC-approval training organized by the
TUDAG. More information gives the Deutsches Zentrum Textilbeton as part of TUDAG
[50]. Figure 13 shows the strengthening of a slab as part of one of the TUDALIT training.
Figure 13: Strengthening of a slab at a TRC-approval training (photo: TUDAG)
28
PROSPECT
The introduction of textile reinforced concrete onto the market has been accompanied with
an extensive expansion of research on TRC and its components. For example, almost every
institute of concrete structures in Germany deals with the topic TRC today. Even adjacent
areas of research begin to consider textile reinforced concrete e.g. the lightweight construc-
tion. The DFG-Priority Programme SPP 1542 "Concrete Light" is just one example. The
program is coordinated by the Institute of Concrete Structures at the TU Dresden. In sever-
al sub-projects, alternative reinforcements are used [61]. In addition, a wide variety of pub-
lications demonstrate the worldwide interest in this innovative building material.
Due to its comprehensive and profound research Germany has emerged as today’s world-
wide leader in the research on textile reinforced concrete and is able to set up quality
standards. However, it has to be said: despite very successful research results and despite
the excellent intrinsic material properties of textile reinforced concrete, the construction
method is so far applied only to a small extent. Thus, it is vital to keep and expand the ad-
vanced knowledge, to pass this knowledge on to all professionals and establish it on the
market, so that TRC can be applied on a large scale in the future. Nonetheless, is it suffi-
cient to obtain more approvals and patents for the material and the construction method, to
increase the bearing capacity of the reinforcement, to improve the utilisation of the materi-
als and to reduce the retail price? Experts – whether planners, contractors, engineers and
research institutions – agree that this is not enough. The design and construction method
has to be reconsidered and a new construction design has to be introduced. Above all: fu-
ture constructions have to be environmentally conscious, resource-efficient, durable, aes-
thetic and flexible. There are still open aspects like the need for a worldwide standardiza-
tion of material testing and construction as well as questions about demolition, recycling,
industrial health and safety as well as gaps in the value chain that have to be closed. Con-
sequently, comprehensive studies are required now!
The large-scale project “C3 – Carbon Concrete Composite” funded by the German Federal
Ministry for Education and Research (BMBF) tackles exactly that problem since 2014
[62]. This project is part of the funding program "Twenty20 – Partnership for Innovation"
[63] and was initiated by the Technische Universität Dresden and is funded until the year
2020 with approximately 45 million euros. The project is carried out by a consortium with
currently 130 partners from industry and science who have merged into a joint organiza-
tion: the C3 e.V. As the partners contribute own financial resources the project has a total
sum of nearly 70 million euros at its disposal. This is a great opportunity for TRC, but also
a huge commitment for those involved. The vision of the project is nothing less than to
find a new design to construct with carbon reinforced concrete and to put it on the market.
The project will focus on strategic considerations until the middle of 2015. The TU Dres-
den developed a concept for the general orientation of the whole project C3 until 2020 in
accordance with superordinate boards. The result was a roadmap as shown in Figure 14.
Almost all basic projects have already started; first individual projects are currently in mo-
tion. The basic projects mainly develop materials, establish standardizations and consider
newest ideas on their feasibility. In 2016, the project will address the elimination of so-
called market entry barriers and will consider issues such as production optimization, recy-
cling and industrial health and safety aspects of textile reinforced concrete or its compo-
nents. The following years will be characterized by technology implementation and the
further development of applications.
29
Figure 14: Roadmap of the project C3 – Carbon Concrete Composite (graphic: Matthias Tietze)
In addition to this major research project funded by the BMBF, there are and will be more
projects that will deal with textile reinforced concrete and its practical application for the
construction industry. It is also certain that the number of construction projects with textile
reinforced concrete as well as its users will continue to rise constantly as more and more
professionals will be convinced of the benefits of the new design.
ACKNOWLEDGEMENT
An acknowledgement often closes a publication. However, this article opens up the ques-
tion: whom should the authors thank? As the presented facts have already shown it is al-
most impossible to offer thanks to all true pioneers of textile reinforced concrete. Still,
there are a few that can be singled out. Peter Offermann, Hilmar Fuchs and Burkhard
Wulfhorst – all textile experts – have already set the basis for TRC in the early 1990s. The
authors especially thank them for this promising idea and are thankful to Peter Offermann
and Hilmar Fuchs in particular for contributing their knowledge of facts and their memo-
ries from the time before the beginning of the CRCs. Among others the authors also thank
the German Research Foundation DFG very much for the financial support of the basic
research in CRC 528 and CRC 532, which laid the foundation for the new design of textile
/ carbon reinforced concrete. More thanks go to all committed partners in science and
business that have contributed significantly through a variety of follow-up projects for the
spreading of textile reinforced concrete and to develop it to a marketable commodity. The
authors would also like to thank the "courageous" contractors, who gave and will give
TRC a chance even though a lot of research and development will still be necessary to es-
tablish a universal standard or a Eurocode.
Finally, the authors thank their colleagues Dajana Musiol und Dominik Schlüter for their
support while writing and translating this article and Mrs. Slaats for her patience.
30
Figure 15: Aesthetic and efficient concrete structures should shape our future! (graphic ai:L, HTWK Leipzig)
LITERATURE
[1] Jahren, P., ′Concrete – History and accounts′, (Tapirc academic, Trondheim, 2011).
[2] Rienecker, F. and Maier, G., ′Lexikon zur Bibel′, 6th Edn (Brockhaus, Wuppertal, 2006).
[3] Lamprecht, H.-O., ′Opus Caementitium. Bautechnik der Römer′ (5th ed., Beton-Verlag, Düsseldorf,
1996).
[4] Ramm, W., ′Über die faszinierende Geschichte des Betonbaus′, in ′Gebaute Visionen. 100 Jahre
Deutscher Ausschuss für Stahlbeton 1907–2007′ (DAfStb, Beuth, Berlin · Wien · Zürich, 2007) 27–
130.
[5] Heinle, E. and Schlaich, J., ′Kuppeln aller Zeiten – aller Kulturen′ (DVA, Stuttgart, 1996).
[6] Mobasher, B., ′Mechanics of Fiber and Textile Reinforced Cement Composites′ (CRC, Boca Raton,
2012).
[7] Curbach, M. et al., ′Sachstandbericht zum Einsatz von Textilien im Massivbau′ (DAfStb, Heft 488,
Beuth, Berlin, 1998).
[8] Mechtcherine, V. and Schnell, J. (eds.), ′Hochduktile Betone mit Kurzfaserbewehrung / Ultra-
ductile concrete with short fibres′ (ibidem, Stuttgart, 2005).
[9] Rostásy, F.S., Budelmann, H. and Hankers, C., ′Faserverbundwerkstoffe im Stahlbeton- und
Spannbetonbaubau′, Beton- und Stahlbetonbau 87 (5) (1992) 123-128.
[10] Burgoyne, C.J., ′Advanced Composites in Civil Engineering in Europe′, Structural Engineering
International 4 (1999), 1-7.
[11] fib Task Group 9.3, ′FRP reinforcement in RC structures′, in: fib (ed.), ′technical report no. 40′
(Stuttgart, 2007).
[12] Curbach, M. and Scherer, S., ′Carbon im Brückenbau′, Carbon im Brückenbau. Bauingenieur –
VDI-Jahresausgabe 2014/2015 (2014) 109–118 (second reprint, Proc. of 24. Dresdner
Brückenbausymposium 2014).
[13] Meier, U., ′Carbon Fiber Reinforced Polymer Cables: Why? Why Not? What If?′, Arab J Sci Eng 37
(2012) 399–411
[14] ′U-Bahn-Bau mit spezieller GFK-Bewehrung′, BauPortal (12) (2010) 742.
[15] Zulassung Schöck ComBAR vom 4.6.2014: http://www.schoeck.de/upload/files/
download/Zulassung_ComBAR_Z_1.6_238[1085].pdf (Zugriff: 20.3.2015)
[16] Fuchs, H., ′Ausarbeitung zu Vorarbeiten des WIZTT / STFI vor 1991 für das FK Textil′
(unpublished doc., 2001).
31
[17] Fuchs, H., Arnold, R., Bartl, A.-M. and Hufnagl, E., ′Einsatz technischer Textilien im Straßenbau –
von textilarmierten Leitplanken bis zu Heizmatten für Verkehrswege′, Chemiefasern/Textilindustrie
41/93 (7) (1991) T202–T204
[18] Fuchs, H., Bartl, A.-M., Arnold, R. and Hufnagel, E., ′Grundlegende Untersuchungen zum Einsatz
gitterartiger Gewirke aus Garnen mechanisch aufgeschlossener Flachsfasern als Bewehrung von
Leichtbeton- und Gipselementen′, Melliand-Textilberichte (11) (1993) 1125.
[19] Wulfhorst, B., Bischoff, T., Offermann, P. and Franzke, G., ′Verwirkte Verstärkungsgelege für das
textile Bauen′ (Abschlussbericht AiF-Vorhaben Nr. 9272 B, Inst. f. Textiltechnik (RWTH Aachen),
Inst. f. Textil- und Bekleidungstechnik (TU Dresden), Aachen, 1996).
[20] Offermann, P., Franzke, G., Fuchs, H., Bartl, A.-M., Wulfhorst, B., Bischoff, T., Graße, W. and
Hempel, R., ′Grundlegende Untersuchungen zur Entwicklung gewirkter Verstärkungsgelege für den
Einsatz in dünnwandigen Formteilen aus mineralischen Baustoffen′ (Abschlussbericht AiF-
Vorhaben 10378 B, Inst. f. Bekleidungstechnik (TU Dresden), Sächs. Textilforschungsinst. e.V.
(Chemnitz), Inst. f. Textiltechnik (RWTH Aachen), Inst. f. Tragwerke und Baustoffe (TU Dresden),
Dresden, 1997).
[21] Hempel, R., ′Bauingenieurstudenten der TU Dresden erfolgreich bei der 6. Deutschen
Betonkanuregatta – ein Beispiel für gelungene Verknüpfung von Forschung und Lehre′,
Jahresmitteilungen 1996, Schriftenreihe des Inst. f. Tragwerke und Baustoffe, TU Dresden, 76−79.
[22] Butler, M., http://www.betonkanu-regatta.de/fileadmin/
betonkanu/tu_dresden_betonboote_05.pdf, 132−133.
[23] Jesse, F., ′Textilbewehrter Beton′ (Diplomarbeit, Inst. f. Massivbau, TU Dresden, 1997, unpubl.)
[24] Kleingries, P., ′Tragverhalten von Bauteilen aus textilbeton unter reiner Biegebeanspruchung′
(Diplomarbeit, RWTH Aachen, 1997, unpubl.).
[25] Curbach, M., ′Entwicklung eines innovativen Verbundwerkstoffes − Textilbewehrter Beton′, lecture
hold at Deutscher Betontag 1999, see also: DBV E. V. (ed.): Proc. of Deutscher Betontag 1999,
Berlin, April, 1999 (Ernst & Sohn, Berlin ,1999) 443−451.
[26] Reinhard, H. W., ′Vorteile einer Hauptbewehrung aus Aramid in Stahlbetonkonstruktionen′, in:
′Textilbewehrter Beton, Aktuelle Einsatzmöglichkeiten′, Proc. of 8th Int. Techtextil Symp.,
Frankfurt/M., Band 4.2 (1997), Paper 4.22.
[27] Reinhardt, H.-W., Krüger, M. and Grosse, C. U., ′Thin plates prestressed with textile reinforcement′,
in: Balagura, P., Naaman, A. and Weiss, W. (ed.), ′Concrete: Materials Science to Application. A
Tribute to Surendra P. Shah′ (ACI SP-206, Farmington Hills, 2002) 355−372.
[28] Wörner, J.-D., ′Oberflächennahe Textilbewehrung im Stahlbetonbau′, Bauen mit Textilien 1 (1)
(1998) 18−22
[29] Ohno, S., Hannant, D.J., ′Modelling the Stress-Strain Response of Continuous Fibre Reinforced
Cement Composites′, ACI Materials Journal 91 (1994) 306−312.
[30] Peled, A., Bentur, A. and Yankelevsky, D., ′Woven fabric reinforcement of Cement Matrix′, Nat.
Building Research Inst. Technicon, Haiva, (Elesevier Science Inc., 10.1993) 8 pages.
[31] http://tu-dreden.de/exzellenz/zukunftskonzept/tud_people/
ddfell/ddfellows2014.
[32] ASAHI GLASS MATEX et al., ′Patent no EP0227207 et alii′, 1987-07-01.
[33] Hayashi, R.; Yamada, K.; Inaba, S.; Tsubouchi, S., ′Tensile properties of carbon fiber mesh
reinforced mortar with various wavings′, Proceedings of the Japan Concrete Institute (12-1) (1990)
1043−1048.
[34] Mobasher, B., ′Mechanics of fiber and textile reinforced cement composites′ (CRC, Boca Raton,
2012).
[35] Haller, P. and Chen, C. J., ′Textile reinforced timber joints in timber construction′, Struct. Eng. Int. 9
(4) (1999) 259−261.
[36] Hirai, T., ′Use of Continuous Fibers for Reinforcing′, in: ACI Compilation 28 (ACI, Detroit, 1994)
58–60.
[37] http://tu-dresden.de/sfb528
[38] http://www.imb.rwth-aachen.de/sfb-53/and http://www.textilbeton-aachen.de/
[39] Curbach M. et al., ′Sonderforschungsbereich 1706 [later: 528] “Textile Bewehrungen zur
bautechnischen Verstärkung und Instandsetzung“ − Finanzierungsantrag′ (TU Dresden, Dresden,
1998).
[40] Curbach, M. and Ortlepp, R. (ed.): ′Sonderforschungsbereich 528 – Textile Bewehrungen zur
bautechnischen Verstärkung und Instandsetzung – Abschlussbericht (shortened)′, http://nbn-
resolving.de/urn:nbn:de:bsz:14-qucosa-86425
32
[41] Frenzel, M.; Ehlig, D.; Schladitz, F.; Curbach, M., ′Textile Reinforced Concrete – Practical
Applications′, in ′Engineering a Concrete Future: Technology, Modeling & Construction′,
Proceedings of the fib Symposium 2013, Tel-Aviv, April, 2013 (Dancygier, A. N. (ed.), Israeli
Association of Construction & Infracstructure Engineees (IACIE) & Faculty of Civil and
Environmental Engineering, Technion – Israel Institute of Technology, 2013).
[42] Ehlig D., Schladitz F., Frenzel M. and Curbach M., ′Textilbeton – Ausgeführte Projekte im
Überblick′, Beton- und Stahlbetonbau 107 (11) (2012) 777−785.
[43] Reports of the SFB 532, unpublished, different years
[44] http://www.db-bauzeitung.de/db-themen/db-archiv/ein-sandwich-aus-textilbeton/
[45] http://tudalit.de/.
[46] Brameshuber, W. (ed.), ′Textile Reinforced Concrete − State-of-the-Art Report of RILEM TC 201-
TRC′ (RILEM Publications S.A.R.L., Bagneux, 2006).
[47] Beton- und Stahlbetonbau, ...
[48] Beton- und Stahlbetonbau Spezial (1) (2015) – Verstärken mit Textilbeton
[49]http://www.rilem.org/gene/main.php?base=8750&gp_id=247.
[50] http://www.textilbetonzentrum.de/.
[51] http://www.stifterverband.info/index.php.
[52] Rempel S., Will N., Hegger J. and Beul P., ′Filigrane Bauwerke aus Textilbeton′, Beton- und
Stahlbetonbau 110 (S1) (2015) 83−93.
[53] Deutsches Institut für Bautechnik (DIBt), ′Allgemeine Bauaufsichtliche Zulassung: Fassadenplatte
aus Betonwerkstein mit rückseitig einbetonierten Befestigungselementen zur Verwendung bei
hinterlüfteten Außenwandbekleidungen, Zulassungsnummer Z-33.1-577′, 2004.
[54] Hegger, J.; Schneider, M.; Kulas, C., ′Dimensioning of TRC with application to ventilated facade
systems, RILEM, 2010′, in Proceedings of the International RILEM Conference on Materials
Science (MatSci), Vol. I: 2nd ICTRC Textile Reinforced Concrete, Aachen, September, 2010
(Brameshuber, W. (Ed.)) 393−403.
[55] Frenzel M., Lieboldt M. and Curbach M., ′Leicht Bauen mit Beton: Balkonplatten mit
Carbonbewehrung′, Beton- und Stahlbetonbau 109 (10) (2014) 713−725.
[56] Walther T., Schladitz F. and Curbach M., ′Textilbetonherstellung im Gießverfahren mit Hilfe von
Abstandhaltern′, Beton- und Stahlbetonbau 109 (3) (2014) 216−222 and http://www.disttex.com/.
[57] Schütze, E.; Tietze, M., ′smarttex – klimaneutrale Strahlungsheizung in Textilbeton′, TUDALIT-
Magazin (9) (2013) zur 5. Anwendertagung Textilbeton, Friedrichshafen, September, 2013, 25.
[58] Walter, T; Curbach, M., ′Textile reinforced concrete for sewer rehabilitation′, in: Proceedings of
ferro-11, Aachen, June, 2015.
[59] Erhard, E.; Weiland, S.; Lorenz, E.; Schladitz, F.; Beckmann, B.; Curbach, M.,
′Anwendungsbeispiele für Textilbetonverstärkung; Instandsetzung und Verstärkung bestehender
Tragwerke mit Textilbeton′, Beton- und Stahlbetonbau Spezial (2015) – Verstärken mit Textilbeton,
74–82.
[60] Deutsches Institut für Bautechnik (DIBt), ′Allgemeine bauaufsichtliche Zulassung “Verfahren zur
Verstärkung von Stahlbeton mit TUDALIT® (Textilbewehrter Beton)” ′, Zulassungsnummer: Z-
31.10-182, 06.06.2014.
[61] http://spp1542.tu-dresden.de/.
[62] http://www.bauen-neu-denken.de/.
[63] http://www.unternehmen-region.de/de/6829.php.
33
34
FERROCEMENT IN CIVIL CONSTRUCTION: ITS USE IN LATIN

AMERICA AND THE CARIBBEAN
Dr. Eng.Hugo Wainshtok Rivas, Civil Engineering Department. Emeritus Professor, Higher
Polytechnic Institute ¨Jose Antonio Echeverría¨ (CUJAE)
Abstract: Some of the most important applications of Ferrocement in Latin America and
the Caribbean between the 70´s and 90´s and still in use are described. In Cuba the exten-
sive construction of Ferrocement fishing boats starting in the 1970’s: the construction of a
big prefabricated barge, 30 m long and 400 ton of capacity, water tanks, sculptures, houses
and especially two swimming pools, one of Olympic dimensions and another of irregular
form with a length of 100 m, the largest in Ferrocement at the time. In Mexico Prof. Al-
fonso Olvera constructed, using different building technologies, hundreds of prefabricated
ferrocement houses, among them a group of prefabricated buildings used for cool storage.
From Brazil, one of the American countries where Ferrocement has a broader use, schools,
water tanks and large span roofs by the Sao Carlos Group, especially architect Joao Figuei-
ras Lima and Professors Joao Bento de Hanai and Lafael Petroni are described. Also small
and medium capacity water tanks by Prof. Rubén Jerves and prefabricated water tanks,
drainage channels and sanitation water tanks by engineers Joao Bento de Hanai and Savio
Nuñez are presented. From after the 90´s new applications built by Prof. Alonso Fernan-
dez, Margarito Ortiz and others Mexican engineers of CIITDIR in Oaxaca Mexico are de-
scribed, including the first time use of ferrocement in a dam wall. Also the work of ecol-
ogist architect Javier Senosiain in housing construction is described. Similarly modern
houses, urban furniture and the biggest ferrocemento sculpture in America, built in Bolivia
by ecologist architects Mario and Javier Moscoso are presented. From Chile, the work of
Prof. Hernan Arnés and others building low cost ferrocement houses in seismic zones and
useful floating barges are presented. The newest use of ferrocemento in Cuba in swimming
pools, housing, sculptures and in equipment are described. Other noteworthy applications
include: from Ecuador, irregular swimming pools by engineers Jaime and Javier Landivar;
from Haiti, the Dominican Republic, and Nicaragua, prefabricated ferrocement houses by
architects Pedro Galiano and Dr. Kurt Rhyner and finally from Honduras, treatment water
tanks and rural water tanks of different uses and capacity are presented.
INTRODUCTION
A first stage of Ferrocement development in our region until the 90s resulted in many
works that endure today and are paradigms which can be constructed using this material. In
Cuba under the guidance of the author boats, swimming pools, warehouses, apartments and
sculptures were built (Figs.1, 2, 3, 4 and 5). In Brazil, under the guidance of Architect
Joao Figueiras Lima and Engineers Joao Bento de Hanai and Lafael Petroni, schools, water
tanks and large roofs (Figs 6, 7) were built.
In Mexico Prof. Alfonso Olvera built prefabricated houses and warehouses, the last used
for refrigeration purposes with remarkable success (fig. 8). Thereafter and over 40 years
the author, has been involved as a designer, consultant, partner or just being in contact
with the authors of the most important examples built in Latin America and the Caribbean.
35
He has carried on national and internationally, many courses and conferences about the
issue and written some books and many articles. He has had and maintains a personal rela-
tionship with the most important Professionals of Latin America and Caribbean on the is-
sue.
Fig.1 and 2: Fishing and passenger ships. Cuba.
Fig. 3: Prefabricated barge 30 m long. Cuba.
Fig.4: Man and animals from the prehistory. Baconao, Cuba.
36
Fig. 5: First houses built with ferrocement, Santiago de Cuba.
Fig. 6: Olympic swimming pool, Cuba
Fig.7: Prefabricated School and Florianopolis Terminal, Brazil. Ing. Joao B. de Hanai
Fig.8: Cold Storage 60x50 m, México. Prof. Alfonso Olvera
37
DEVELOPMENT IN THE LAST 25 YEARS
In the 90´s a new and significant boost took place in Mexico, Bolivia, Chile, Cuba and
other countries in the region where new professionals joined the civil construction field
using Ferrocement. In Oaxaca, Mexico, Prof. Alonso Fernández and colleagues from the
Interdisciplinary Technology Research Center for Integrated Regional Development
(CIITDIR) of the National Polytechnic Institute developed houses, bridges, storages and a
local auditorium applying the mortar directly by hand. (Fig. 9)
Fig. 9: Storage and Auditorium CIITDIR .Prof. Alonso Fernández
They have also successfully applied ferrocement in construction of medium tank capacity
and for the first time in the construction of dams screens. (Figs. 10 and 11)
Fig. 10: Water tank 100 M3 Fig. 11: Screen dam (both Ing. M. Ortiz)
Meanwhile Arq Javier Sinesiain an ecological Architect designed and built homes of an
undeniable aesthetic value (Fig. 12).
Fig.12: Echologic house, México. Arq. Javier Sinesiain
38
In Chile with the participation of Prof. Hernán Arnes , professionals related to the cement
factory BioBio have built following Cuban experience , prefabricated housing adapted to
the climate and seismicity of the country with very satisfactory results. They have built
also barges used as stores in calm water. (Fig.13 and 14).
Fig.13: Barge with storage on deck Fig. 14: Houses, Prof. Hernán Arnés.
In Bolivia Architect Mario Moscoso with the collaboration of the Author, finished in 1993,
a work that had begun years before in reinforced concrete and had to be paralyzed due to
its high cost and technical difficulties: the Christ of Concordia (fig.15). Today the big
sculpture is representative of the city of Cochabamba. In Oruro was built the Jurassic Park,
with very good technical quality and aesthetics, in the same place where found, decades
ago ,the original footprints of prehistoric animals that lived there (Fig. 16). It is to signify
the high art and beauty, that Ecological Architects Mario Moscoso and his son Javier have
shown building homes and other works of curved shapes. (Fig. 17 and 18)
Fig.15: Cristo de la Concordia Fig. 16: Prehistóric animal (both Bolivia)
Fig.17: House ,Architect. M. Moscoso Fig. 18: Front of shopping center. Arch. Javier Moscoso.
39
In Brazil, although the retirements of several of its members has decreased the outputs of
San Carlos Group, important applications of ferrocement continues to take place in that
country, like the tanks for water treatment constructed by Eng. Savio Núñez and sanitation
urban works in Rio de Janeiro designed by the Group, demonstrating the current strength
of Ferrocement in that country. (Fig 19 and 20)
Fig. 19:- Water tank and prefabricated drainage conduit, Brazil
Fig.20: Water tanks. Brazil. Ing. Sabio Nuñes.
In Honduras Ferrocement has been developed by the Honduran Fund for Social Interest
(FHIS) in water supply treatment and small tanks for social use, with excellent results (Fig.
21)
Fig. 21: Water treatments tanks, fhis. Honduras.
40
In La Habana, Cuba semi-prefabricated square small deposits were built, to replace those
built before with asbestos and cement. In Guantánamo Eng. Miguel Moreno designed a
prefabricated tank of 3000 m3 capacity (Fig. 22 and 23). Furthermore the Ministry of Sug-
ar (MINAZ), under the guidance of the author is refurbishing old steel tanks of 500 m3,
with an inside layer of ferrocemento using projecting mortar
Fig. 22: Prefabricated tank of small capacity. Fig. 23: Large prefabricated tank, Cuba.
Similarly, a group of young sculptors continued to use Ferrocemento in this field. It is to

signify the complexity and beauty of the works accomplished in this endeavor, as the crab
or the city of Cardenas or specific sculptures for different hotels in the country. (fig. 24-26)
Fig. 24: Woman with mirror, Cuba. Fig.25: Crab, city of Cárdenas
Fig. 26: Rescue. Civil Defense National Center
The production of prefabricated ferrocement panels in Cuba for housing became more
technical, from the use of fixed molds to the use of portable steel ones, where the mortar
set on place using a vibrating table. It markedly improved the quality of the results. Apart-
41
ment Buildings in Pinar del Rio and the hole small town of Minas 1 in the Escambray
Mountains in the center of the country, are good examples (Fig. 27 and 28 ).
Fig. 27: Prefabrícate building, Pinar del Rio Fig. 28: Pueblo Minas 1, Cienfuegos, Cuba
In the 90th was developed in the Higher Polytechnic Institute ¨ Jose Antonio Echeverria¨
(ISPJAE) in Havana, by architect Emilio Loret de Mola and the author, the Ferrocement
Residential Buildings System (SERF). This System uses modulated prefab panels for use
in walls, floors and roofs. (Fig.29). Also a technology named PRELAB, was designed us-
ing an slide vibrating-compactor machine, allowing greater quality and productivity in
panels construction. This technology, designed by Dr. Eng. Sergio Marrero with the partic-
ipation of the author, presents many technical advantages and has been exported to other
countries. All Ferrocement houses built in Cuba in the last years have been built using this
system (Fig. 30, 31 and 32)
Fig. 29: SERF. Fig. 30: PRELAB (Tecnology. Dr. S. Marrero)
Fig. 31: Two stories high building Fig. 32: Building in La Havana.
42
In Dominican Republic, Haiti and Nicaragua, Architects Kurt Rhyner and Pedro Galiano,
have built with the author advice earthquake-withstand ferrocement building, of great
beauty and functionality. (Fig. 33 and 34)
Fig. 33 and 34: Building in Dominican Republic Arch. Pedro Galiano, Haiti house Arch. Kurt Rhyner
Starting in 90th, large pools of Ferrocement, both prefabricated and concreted in place
have continued to be built in Cuba, reaching what Architect Joao Figueiras Lima and Eng.
Lafael Petroni from University of San Pablo, named a Cuban school for making Ferroce-
ment pools (Fig.35 and 36). In all cases, 2-3cm thick ferrocement sheets were used for the
bottom and walls in swimming pools up to 2m deep and 2000m2 surface area, leading to
costs between 4 or 5 times lower than those achieved with reinforced concrete pools.
Fig.35: Swimming pool, Varadero Fig.36: Pool, Punta Arenas Hotel
Ferrocemento swimming pools have also built in Brazil and Bolivia, usually for family
use. In Ecuador under the author advise, Eng. Jaime and Javier Landivar and Marco
Estupiñan have started to build free-form pools with very good results (Fig. 37 and 38).
Fig. 37: Esmeralda Pool, M. Estupiñan Fig. 38: Guayaquil Pool, Ing. J. Landívar
43
Also in Ecuador Prof. Rubén Jerves, from the University of Cuenca, has developed an in-
teresting technique for the construction of medium capacity circular tanks (Fig. 39).
Fig. 39: Water tanks. Ecuador Prof. Rubén Jerves
CONCLUSIONS
Ferrocement has been widely used in civil construction in Latin America and the Caribbe-
an, and have demonstrated to be an important, appropriate and sustainable alternative
construction for developing countries and mainly for those buildings whose form can be
very difficult and expensive to build in reinforced concrete. Ferrocement development in
the region has allowed the International Ferrocement Society (IFS for its acronym in Eng-
lish) to organize in Cuba, two International Symposium on Ferrocement, the first FER-
RO 4, in October 1991 and the second FERRO 10 in October 2012. Additionally the au-
thor has been appointed as the current President of the IFS
BIBLIOGRAPHY:
1. Wainshtok Rivas, Hugo. Ferrocemento: Diseño y Construcción, Riobamba, Ecuador 2010, 4ta.
Edición.
2. Olvera López, Alfonso. El Ferrocemento y sus aplicaciones .Instituto Politécnico Nacional, México,
1985
3. Bento de Hanai, Joao. Construcciones de Argamasa Armada (in portugués). Editora Fini LTA 1992,
Sao Paolo
44
CLASSIFICATION OF APPROACHES TO DIMENSIONING AND

ASSESSMENT OF LIGHTWEIGHT TRC STRUCTURES
Josef Hegger, Norbert Will, Rostislav Chudoba, Alexander Scholzen, Institute of Structural
Concrete, RWTH Aachen University
Abstract:Textile reinforced concrete (TRC) has a high potential for lightweight, load-
bearing concrete structures as demonstrated by recent applications including large-scale
ventilated façade systems, pedestrian bridges, self-supporting sandwich panels and shells
structures. The available types of textiles exhibit a wide range of mechanical properties
depending on the fiber material, the roving size and binding type. This range of properties
is even broader for epoxy- or polymer-impregnated yarns. As a consequence, specific me-
chanical behavior of TRC structures reinforced with impregnated and non-impregnated
fabrics call for different approaches to dimensioning and ULS assessment. Besides the
mechanical characteristics, also technological properties of a particular type of fabrics (e.g.
form stability) need to be considered in view of the feasibility of the manufacturing process
(e.g. casting, spraying, and lamination). The present paper gives a brief summary of the
major issues to be considered in approaches to design and construction of TRC structures.
INTRODUCTION
Textile reinforced concrete (TRC) has a high potential for construction of light-weight,
load-bearing structures. Due to its non-corrosive textile reinforcement, made of alkali-
resistant glass or carbon fibers, a minimum concrete cover is necessary. The possible ap-
plication field of this type of composite ranges from slender façade panels [1,2], TRC
sandwich elements [3,4] to load-bearing structural members such as pedestrian bridges [5]
and thin shell structures [6,7] (Figure 1). Depending on the demands of the specific appli-
cation, i.e. the manufacturing method and the geometric boundary conditions, the charac-
teristics of the composite vary on a broad scale depending on the applied textile fabrics and
fine-grained concrete mixture. The present paper gives a brief summary of the major de-
pendencies between material properties, fields of applications and available dimensioning
approaches.
(a) (b)
45
(c) (d)
Figure 1: Application examples of lightweight TRC structures: ventilated façade panels (a) (picture: Hering
Bau); sandwich panels with polyurethane core (b); pedestrian bridge (c) (picture: solidian) (right); hypar shell
structure (d)
CHARACTERIZATION OF THE TEXTILE FABRICS
Mechanical behavior of TRC is primarily affected by the characteristics of the textile fab-
rics, in particular by the fiber material, cross-sectional area of the rovings (quantified as
“tex=g/km”), binding type of the warp-knitted fabrics (e.g. tricot, plain stitch bond) and,
eventually, by the type of the applied impregnation material (e.g. epoxy-resin or styrol-
butadiene). Impregnation of the fabrics leads to a homogenization of the stress profile
within the multifilament yarn cross section resulting in a better activation of the inner fila-
ments compared to non-impregnated yarns and, thus, leads to a higher strength and reduc-
tion of the necessary anchorage length. Furthermore, yarn impregnation can be used to
achieve form stability required for casted structural components. At the same time, it re-
duces form flexibility needed for curved geometries. Material characteristics of the differ-
ent types of textile reinforcement are qualitatively summarized in Table 1.
STRENGTH CHARACTERISTICS
In order to provide a qualitative orientation within the design space of TRC material struc-
ture let us discuss some examples of achievable values of strength in dependence on the
used fabric types with different fineness, binding type and impregnation treatment. For
example, for 3300 tex carbon yarns with an impregnation using epoxy resin (EP) the max-
imum textile stress reached in composite tensile tests evaluates to approximately 3000 MPa
[8]. This stress level corresponds to the material strength of carbon fibers reflecting an al-
most complete material activation of the yarn including the inner filaments without direct
contact to the concrete matrix. In comparison, an impregnation of 1650 tex carbon yarns
with styrol-butadiene (SBR) leads to strength in composite of up to 2000 MPa. The lower
yarn efficiency for the SBR-impregnation is due to the longer anchorage length within the
composite resulting in a partial pull-out of the inner filaments.
For non-impregnated fabrics an acceptable strength values of up to 1500 MPa can be
achieved with small, flat and open yarn cross section, e.g. using 800 tex carbon yarns and
warp knitted fabrics with plain stich bond [6,7]. Other binding types in combination with
larger yarn cross section lead to a significant decrease of yarn strength due to the insuffi-
cient activation of the inner filaments within the compacted yarn cross section.
46
FORM STABILITY AND FLEXIBILITY

Besides the strength of the fabrics the type of impregnation also significantly affects their
form flexibility. For applications in straight, rectangular structural components such as
beams and slabs (Figure 1a,b) manufactured by casting of fresh concrete into the mold a
high form stability of the fabrics is desired. The EP-impregnation makes it also possible to
produce "prepreg" reinforcement components, e.g. for the T-beam cross-sectional shapes
for the pedestrian bridge shown in Figure 1c. These structural elements exhibit high load-
bearing capacities due to the positive effect of an EP- and/or SBR-impregnation on the
strength of the fabrics.
However, TRC structures with curved geometries such as the hypar shell structure shown
in Figure 1d require a high form flexibility so that the application of EP-impregnated fab-
rics is not feasible. Therefore, non-impregnated fabrics were used for manufacturing the
depicted hypar shell using a spraying and lamination procedure. Another example of a
curved shell structure described in [9] consisting of casted shell elements was constructed
using SBR-impregnated fabrics which retain a certain form flexibility, but require a closed
form-work for curved structures.
The mentioned application examples demonstrate the diversity of the design space availa-
ble for existing versions of TRC. This fact can be seen as a challenge and chance at the
same time. The design of the composite including the type of fabrics in combination with
the particular type of cementitious matrix and manufacturing technique needs to be tailored
for a particular field of application and required structural properties. Table 1 summarizes
the relevant material characteristics of textile fabrics used in applications realized so far.
Table 1: Characteristics of common textile fabric types
type of textiles
impregnated with EP impregnated with SBR non-impregnated
efficiency high medium medium / low

fineness-independent fineness-dependent fineness-dependent
anchorage length short medium medium / long

fineness-independent fineness-dependent fineness-dependent
form stability high medium low

applicable for prepregs less applicable for prepregs not applicable for prepregs
form flexibility low medium high
straight, planar curved, spatial

possible geometries
beams, slabs shells
47
EVALUATION OF CROSS-SECTIONAL STRENGTH CHARACTERISTICS
Availability of cross-sectional strength characteristics of frequently used fabric types in

combination with frequently used cross-sectional forms and reinforcement layouts is the
key requirement for a wide application of TRC in engineering practice. As indicated in
Section 2 a systematic approach to evaluation of strength characteristics has to consider the
phenomenological differences in the material behavior of different reinforcement types:
While impregnated yarns exhibit only a low effect of matrix cracking on the fabric
strength, non-impregnated yarns show a strong interaction of the cracking process and ef-
fective fabric strength achievable in the composite [10,11].
As a result, for impregnated yarns the ultimate tensile and bending strength can be derived
from the material behavior of the components, i.e. fabrics and fine grained concrete, in
analogy to steel reinforced concrete [8]. If further validation tests are provided for particu-
lar material combinations standard ω-tables in analogy to steel reinforced concrete can be
envisioned in the future for such material combinations.
For non-impregnated yarns, the direct derivation of cross-sectional strength characteristics
based on the equilibrium of normal force and bending moment at the critical cross-section
is no longer possible. The reason for that is the interaction of the bond behavior with the
matrix cracking leading to different yarn strength for different reinforcement ratios and/or
cross-sectional layouts.
Table 2: Approaches to the evaluation of cross-sectional strength characteristics
cross-sectional strength characteristics
impregnated with EP impregnated with SBR non-impregnated
influence of matrix
cracking on yarn low moderate high
strength
cross-sectional derived from the material laws experimentally determined

strength characteris- of the components composite strength for given layout
tics yarn tension, concrete compression tensile and bending tests
explicit
cross-sectional di- ω-tables not possible
based on yarn/concrete components
mensioning
cross-sectional
strength envelope n-m interaction diagram
based on composite strength
for ULS design
For this reason, the cross-strength characteristics for TRC structures reinforced with non-
impregnated fabrics must be determined using test on composite with a cross-sectional
layout consistent with the structural design at hand. Approaches to the experimental evalu-
48
ation of the composite tensile and bending strength and construction of n-m interaction
diagrams are presented in [6,7].
APPROACHES TO DIMENSIONING AND ULS ASSESSMENT OF TRC STRUCTURES
As the impregnated fabrics are mostly used for straight beam or slab elements, it is possi-
ble to formulate an ULS assessment approach based on the concept of the a-priori known
critical cross section (e.g. at mid-span for simple supported beams or at the middle support
of a two-span beam [8]).
On the other hand, in case of slab and shell structures exhibiting interaction of cross-
sectional stress resultants, the critical cross sections cannot be identified a-priori. As a
consequence, the dimensioning approach for TCR shell structures must be formulated for a
general cross-sectional point of the structure including the interaction between normal
force and bending moment. Moreover, the anisotropy of cross-sectional strength of a TRC
cross section due to the nonaligned direction of principal stresses and fabric orientation
must be considered. Due to the higher complexity of the assessment procedure for such
structures a computer aided ULS assessment is necessary [6,7]. An example of the pro-
posed ULS assessment is provided for a barrel-vault shell in [12]. Detailed experimental
and numerical analysis of the nonlinear behavior of the shell serving for the validation of
the assessment is provided in [13].
Classification of issues relevant for systematic ULS assessment of structural components
depending on their form and type of loading is provided in Table 3.
Table 3: Classification of approaches to ULS assessment
structural components
beams slabs and shells
interaction of normal no yes

combined loading scenarios need to be consid-
and bending loading generally not
ered
critical cross section yes no

known corresponds to location of maximum depends on interaction of combined loading and
a-priori bending moment material anisotropy
check utilization for critical cross check utilization ratio for

ULS assessment
section only entire structure
SUMMARY
In order to provide a systematic approach to the dimensioning and assessment of TRC
components and structures a classification of possible cross-sectional designs is necessary.
In particular, phenomenologically different mechanical behavior of the composite is
achieved depending on the choice of the type of fabrics. At the same time, constraints on
manufacturing procedures and achievable forms of components affect the choice of mate-
rial. Despite of the large size and complexity of the design space of TRC structures, di-
mensioning and assessment methods for TRC need to retain transparency and simplicity.
49
The present paper illuminates the elementary correspondences between the choice of mate-
rial components, manufacturing procedures and possible shapes and loading conditions of
a structural component. Details and examples of dimensioning and assessment procedures
classified here are provided in other contributions to this conference [8,12].
ACKNOWLEDGMENT
The financial support of the German Research Foundation (DFG) within DFG project CH
276/2-2 is gratefully acknowledged.
REFERENCES
1. Hegger, J.; Kulas, C.; Horstmann, M.: Realization of TRC façades with impregnated AR-glass
textiles, Key Engineering Materials, 2011, 466:121-130.
2. Rempel, S.; Kulas, C.; Hegger, J.: Slender façade structures made of textile-reinforced high per-
formance concrete, In: Brameshuber, W. (Ed.): 11th International Symposium on Ferrocement
FERRO-11 in connection with the 3rd ICTRC, Aachen, Germany, 7–10 June 2015.
3. Shams, A:, Horstmann, M:, Hegger, J.: Experimental investigations on textile-reinforced concrete
(TRC) sandwich sections. Composite Structures, Vol. 118, December 2014, pp. 643-653.
doi:10.1016/j.compstruct.2014.07.056.
4. Shams, A:, Horstmann, M:, Hegger, J.: An analytical model for sandwich panels made of textile-
reinforced concrete. Construction and Building Materials, Vol. 64, August 2014, pp.451-459.
doi:10.1016/j.conbuildmat.2014.04.025.
5. Hegger, J.; Kulas, C.; Raupach, M.; Büttner, T.: Load-bearing behavior and durability of a slender
textile reinforced concrete bridge (in German), Beton- und Stahlbetonbau, 2011, 106(2): 72-80.
6. Scholzen, A.; Chudoba, R.; Hegger, J.: Thin-walled shell structure made of textile reinforced
concrete; Part I: structural design and construction. Structural Concrete, 16(1), 2015, doi:
10.1002/suco.201300071 (in press).
7. Scholzen, A.; Chudoba, R.; Hegger, J.: Thin-walled shell structure made of textile reinforced
concrete; Part II: experimental characterization, ultimate limit state assessment and numerical
simulation. Structural Concrete, 16(1), 2015, doi: 10.1002/suco.201400046 (in press).
8. Rempel, S.; Kulas, C.; Hegger, J.: Bearing behavior of impregnated textile reinforcement, In:
Brameshuber, W. (Ed.): 11th International Symposium on Ferrocement FERRO-11 in connection
with the 3rd ICTRC, Aachen, Germany, 7–10 June 2015.
9. Ehlig, D.; Schladitz, F.; Frenzel, M.; Curbach, M.: Textilbeton – Ausgeführte Projekte im Über-
blick (in German). Beton- und Stahlbetonbau, 2012, 107 (11): 777-785.
10. Li, Y.; Chudoba, R.; Sadílek, V.; Rypl, R.; Vořechovský, M.: Analysis of the tensile response of
textile reinforced concrete using digital image correlation technique combined with multi-scale
stochastic modelling, In: Brameshuber, W. (Ed.): 11th International Symposium on Ferrocement
FERRO-11 in connection with the 3rd ICTRC, Aachen, Germany, 7–10 June 2015.
11. Rypl, R., Chudoba, R., Scholzen, A., Vořechovský, M.: Brittle matrix composites with heteroge-
neous reinforcement: Multi-scale model of a crack bridge with rigid matrix, Composites Science
and Technology, 2013, 89:98-109.
12. Scholzen, A.; Chudoba, R.; Hegger, J.: Ultimate limit state assessment of TRC shell structures
with combined normal and bending loading, In: Brameshuber, W. (Ed.): 11th International Sym-
posium on Ferrocement FERRO-11 in connection with the 3rd ICTRC, Aachen, Germany, 7–10
June 2015.
13. Sharei, E.; Scholzen, A.; Chudoba, R.; Hegger, J.: Nonlinear analysis of textile reinforced con-
crete shells using an anisotropic, damage-based material model, In: Brameshuber, W. (Ed.): 11th
International Symposium on Ferrocement FERRO-11 in connection with the 3rd ICTRC, Aachen,
Germany, 7–10 June 2015.
50
CHAPTER 2
LOAD
CARRYING
BEHAVIOR
51
52
EXPERIMENTAL AND NUMERICAL ANALYSIS OF TEXTILE

REINFORCED CEMENT COMPOSITES AS TENSILE
REINFORCEMENT IN CONCRETE SHELLS.
Evy Verwimp1,2, Tine Tysmans1, Svetlana Verbruggen1 and Marijke Mollaert2 , 1 Department of
Mechanics of Materials and Constructions (MeMC), Vrije Universiteit Brussel. 2 Department of
Architectural Engineering (ARCH), Vrije Universiteit Brussel.
Abstract: Textile reinforced cement (TRC) composites are increasingly studied as a mate-
rial for structural stay-in-place formwork for concrete structures, because elements with
low thickness and relative high mechanical capacities can be produced. Current envisioned
formwork applications - beams, slabs, columns - do not exploit another advantage of
TRC’s, namely that the flexible textile reinforcement allows an exclusive freedom of form.
With TRC, freeform moulds are easily prefabricated to serve as structural formwork for
shells, facilitating the construction on site. This paper investigates the contribution of TRC
formwork to the loadbearing behaviour of concrete shells. Tests are performed on a TRC
reinforced spherical shell of 2m span and 200mm height, subjected to a distributed load.
The experiment is compared to a numerical simulation performed in the finite element
software Abaqus. The test shows promising results for TRC as reinforcement as the TRC
layer contributes significantly to the concrete section. The finite element model shows sim-
ilar behaviour as monitored during the test. This model allows to further develop the de-
sign methods for concrete shells with a stay-in-place TRC formwork and reinforcement.
With the use of TRC composites, a more efficient and faster construction method for con-
crete shells is just around the corner.
INTRODUCTION
The construction of concrete shells is one of the challenges in civil engineering in western
regions. Firstly, the current formwork methods for concrete shells, i.e. timber moulds [1]
and foam blocks [2], are labour intensive and/or material wasting. Currently, research is
performed on alternative formwork methods like flexible fabrics [3], [4] pneumatic form-
work [5] or hybrid cable-net and fabric formwork [6]. These systems are flexible enough to
easily realize curved surfaces, however they still experience relatively large deformations
when casting concrete. Secondly, the placing of the rigid traditional steel reinforcement
limits the curvature, is labour intensive and therefore increases the cost. Recently, alterna-
tive reinforcement methods based on fibre textiles are investigated in concrete shells,
namely fibre and textile reinforced cements (FRC / TRC), like demonstrated in [7] and [8].
However, these alternative methods do not solve the formwork problems discussed before.
Considering these issues, we developed an inventive formwork solution, which exploits the
properties of textile reinforced cement (TRC) composites. These TRC composites are
composed of a continuous fibre system, a textile, which is impregnated with a cementitious
matrix. While this cementitious matrix is wet, the impregnated textile remains flexible
(Figure 1a) and can be shaped onto any (reusable) mould, like e.g. a pneumatic formwork
(Figure 1b). The low weight of the thin TRC layer compared to the concrete, which is
normally directly cast onto these moulds, prevents large deformations and thus ensures a
final shape closely approximating the initially designed shape.
53
a b
Figure 1: While the cementitious matrix is wet, the textile remains flexible (a), and can be shaped on e.g.
(reusable) pneumatic formwork (b).
After hardening of the cement, a thin composite is obtained which is rigid enough to with-
stand the casting of the concrete. When the concrete is hardened the stay-in-place form-
work gets an additional structural function, i.e. as (partial) tensile reinforcement of the final
concrete structure. The structural capacity of TRC stay-in-place formwork was already
demonstrated for beam elements in [9] and [10]. This paper presents the experimental and
numerical analysis of a TRC reinforced concrete dome. Firstly, the plain concrete dome,
reinforced by the stay-in-place TRC formwork, is subjected to a point load. Secondly, the
experiment is numerically simulated within the finite element software Abaqus. Finally,
the experiment and the model are compared and conclusions are drawn.
EXPERIMENTAL ANALYSIS OF A TRC REINFORCED DOME

Materials and experimental setup
During the test, a TRC composite consisting of a textile mat of 300g/m² random E-glass
fibres impregnated with Inorganic Phosphate Cement - IPC (18 % fibre volume fraction) -
[11] is used in combination with concrete. Their mechanical properties are found in table 1.
Table 1: Properties of the concrete and TRC composite used for the experiment.
Properties concrete Properties TRC
Compressive characteristic 50 MPa Compressive characteristic strength 80 MPa

strength
Tensile characteristic strength 4.1 MPa Tensile characteristic strength 40 MPa
Young modulus 35.7 GPa Young modulus stage 1 18 GPa
Young modulus stage 3 4 GPa
The mechanical behaviour of glass fibre textile reinforced IPC (GFTR-IPC) differs signifi-
cantly in compression and tension because of the brittle matrix. In compression, the com-
posite is assumed to be linear elastic until failure [12], but in tension, it shows already a
nonlinear behaviour at low tensile stresses - around 7 MPa - (Figure 2) due to the low ten-
sile failure strain of the matrix relative to that of the fibres.
54
Figure 2: GFTR-IPC is linear in compression and nonlinear in tension [13]
The geometry and experimental setup of the concrete dome are shown in Figure 3. The
concrete dome has a span of 2 m, a height of 200 mm, a concrete thickness of 20 mm and
it is reinforced with a 5 mm thick TRC. The shell is subjected to a point load (ϕ100 mm)
on the top. To restrict the horizontal displacement the edge is encircled with wooden pan-
els.
Figure 3: The spherical dome is subjected to a point load, strains are measured by strain gauges at five differ-
ent locations.
Figure 4 shows the production process of the TRC reinforced concrete spherical dome. The
glass fibre mat is placed in a foam mould (a) and impregnated with a cementitious matrix
(IPC) (b). To obtain a TRC formwork of 5 mm, ten layers of mats are needed. After hard-
ening, the composite is turned and a strong and stiff formwork is obtained (c). A rougher
contact surface - by adding a layer of small stones - is created to assure the bond (d). Rub-
ber rulers are placed on the formwork to control the concrete thickness during casting (e
and f). The shell is tested after 31 days.
a b c
55
d e f
Figure 4: Production process of the dome with mixed TRC-concrete section.
The experiment is monitored by strain gauges and Digital Image Correlation (DIC). For the
DIC, two areas of approximately 400 x 400 mm (light blue squares in Figure 3) were
monitored by two 3D camera systems. Strain gauges are placed on both the TRC layer (so
in between the TRC formwork and concrete before, blue lines Figure 3) and on top of the
concrete (red lines Figure 3), both in meridional and hoop direction. The nomenclature of
the strain gauges is defined by three parameters:
1 M: meridional direction 2 Position (see Figure 3) 3 C: on concrete layer

H : hoop direction R: on TRC layer
S : 45° direction
Results during testing of the TRC-concrete shell

The vertical deformation of the top centre, measured by the load cell, of the TRC rein-
forced concrete dome is shown in Figure 5. The curve is linear up to 17.5 kN, where a first
drop occurs. In an initial assumption this could indicate that the shell remains intact up to
this drop, where after cracks initiate within the concrete material until the ultimate load of
18.7 kN. Finally, the load decreases to a constant value of 2 kN, where the test was
stopped.
Figure 5: Load progress shows two dropping points (at 17.5 kN and at 18.7 kN).
Figure 6 shows the shell after testing. During the experiment visible cracks initiated from
the top along radial lines, exactly where the rubber rulers were placed while casting. Here-
after, a circular crack initiated around the point load. When the constant load of 2 kN is
reached, no other visible cracks were formed. After the test, the concrete layer proved to be
56
completely debonded from the TRC shell, indicating the loss of composite action between
both (Figure 6b).
a b
Figure 6: Cracks formed along the radial lines were rubbers were placed (a) and debonding of TRC and con-
crete (b).
The strains monitored by the strain gauges are shown in Figure 7a, b, c and d.
a b
c d
Figure 7: Strain gauges show similar trends as long as no cracks are initiated.
Figure 7a shows the meridional strains on positions 2 and 4, on a quarter span of the dome.
The strains on the TRC layer (M2R and M4R) are both in compression and have the same
57
curve up to 13 kN. From here on the axisymmetric behaviour clearly vanishes, indicating
the initiation of damage, probably by concrete cracks. This phenomenon was however not
observed in Figure 5, where the initial constant stiffness is retained until a load of 17.5 kN.
A possible explanation for this phenomenon can be found in the crack bridging capacity of
the TRC stay-in-place formwork, as is already observed for TRC external beam reinforce-
ment [14] and [15]. Surprisingly, the strains on the concrete layer (M2C and M4C) do not
exhibit a similar behaviour, not even before the initiation of the cracks. Considering the
similar evolution of M2C to M2R and M4R, it can be assumed that a measurement error
occurred for M4C.
The hoop strains on position 2 and 4 are shown in Figure 7b. Until the first crack, the
strains on the TRC layer (H2R and H4R) are in tension and behave similar. The concrete
layer also appears in tension (H2C), however it shows a breakpoint at 8 kN, which is again
not noticed in load-deflection behaviour (Figure 5). The tension in the hoop direction ex-
plains the first concrete cracks, induced along the radial lines. Comparison of Figure 7a
and b indicates the expected biaxial stress state of the shell, as the meridional strains are in
compression and the hoop strains are in tension.
Figure 7c shows the strains on the centre top of the TRC layer. S3R and M3R behave op-
posite and H3R shows no similarities. However, a similar behaviour is expected on the top
in an axisymmetric geometry. This discrepancy can be explained by imperfections: i.e.
load not perfectly in the middle, strain gauges not perfectly placed at the centre, etc. How-
ever these strains remain limited in comparison with the ones obtained at a quarter span.
The meridional and hoop strains near the edge (position 5) are shown on Figure 7d. The
meridional strains on both layers (M5C and M5R) are in tension until 10 kN, where after
they switch to compression. For the hoop strains on both layers (H5C and H5R) the oppo-
site phenomenon occurs. However up to 10 kN this observation is opposite to the expecta-
tions, which are similar to Figure 7a and b, namely meridional strains in compression and
hoop strains in tension. This can possibly be explained by the lateral supports of the shell;
the zone near the edge moves up due to the central loading. This effect leads to tension in
the radial direction and compression in the hoop direction until the first concrete crack.
Up to concrete cracking the strains are as expected. Here after the axisymmetry is lost and
the behaviour becomes unpredictable. The initiation of cracks is however well noticed in
the strain curves by several breakpoints at 10 and 13 kN, indicating that the cracks are
probably initiated at a lower load than 17.5 kN, which was initially assumed based on the
constant stiffness in the load-deflection curve. The tensile strains in the TRC remain lim-
ited up to shell failure and do not exceed 400 με, which falls within the first linear part of
the TRC stress-strain diagram (Figure 2). Thus, no stresses higher than 7 MPa are ob-
tained, which is only 17.5 % of the tensile strength of the TRC (40 MPa). The DIC obser-
vations show qualitative results concerning the displacements over the surface of the shell.
The vertical displacement of several points during loading is shown in Figure 8. All points
follow the same trend, namely a downward displacement followed by a sudden upward
displacement. The upward movement indicates the debonding of concrete and TRC, in-
duced by the cracks in the concrete. These breakpoints approximately correspond to the
two main drops in the load-deflection curve at 17.5 kN (400s) and 18.7 kN (800s) (Figure
5).
58
a b
Figure 8: Displacement of the concrete layer is first downwards and then upward
NUMERICAL ANALYSIS OF A TRC REINFORCED DOME

The experiment is simulated using the finite element (FE) software Abaqus and the exper-
imental results are compared with the numerical ones.
Model build-up
The numerical model of the dome is built up in two parts, which are placed on each other
as two layers; the TRC layer is modelled as a shell part with a continuous shell section and
the concrete layer is modelled as a solid part with continuous solid section.
For the TRC layer two types of linear elements are used, namely the 3-node S3 and the 4-
node S4R elements, which are elements for doubly curved thin or thick shells with reduced
integration. For the concrete layer, the 8-node C3D8R elements are used, which are linear
bricks with reduced integration. Mesh convergence was checked for this geometry; in the
end, elements with a seed of approximately 30 mm were used, chosen as a compromise
between accuracy and calculation time.
The material behaviour for both the concrete and TRC layer are modelled with the Con-
crete Damaged Plasticity (CDP) model, which is inbuilt in Abaqus. Wozniak proved in
[13] that the CDP model predicts the TRC material’s behaviour the best. As the concrete is
cast on the TRC layer, the bond between the two material layers is not perfect and is there-
fore taken into account in the numerical model. The bond is modelled by the inbuilt cohe-
sive behaviour in Abaqus, which has the possibility to include the damage of the bond. The
parameters needed for modelling the bond are described in [16]. The dome’s edge is re-
59
stricted in three directions. The model is loaded by its self-weight and a point load, which
gradually increases to 20 kN, at the centre top with a diameter of 100 mm.
Numerical results
The comparison of the measured strains during the experiment with the numerically calcu-
lated strains are shown in Figure 9 a and b. The numerical results are represented by a dot-
ted line. The numerical model predicts general trend well, namely the biaxial behaviour of
the shell, which means the meridional strains are in compression and the hoop strains in
tension. Moreover, the order of magnitude is also well predicted. However, individually
some differences are observed.
a b
Figure 9: Comparison of the experimental and numerical results shows a similar trend and order of magni-
tude is well predicted.
Figure 9a shows the meridional and hoop strains on position 2, at a quarter span of the
dome. For both strains the general trend and order of magnitude is well predicted, but the
progress of the experimental and numerical curves are different. The hoop strains on the
TRC layer are larger than those on the concrete layer, which proves opposite in the
experiment.
Figure 9b compares the numerical and experimental meridional and hoop strains near the
edge, on position 5. Up to the concrete cracking the numerical simulation fails to predict
the strain distribution, as the inverted strains are observed in the experiment. This deviation
can be attributed to local phenomena near the edges of the shell (as discussed above),
which are not taken into account in the numerical simulation.
Due to the presence of nodes, in the numerical model near the load application zone, i.e.
top centre, which encounter stress concentrations, these results are not represented. The
comparison of the numerical and experimental vertical deformation of the top centre is
shown in Figure 10. The numerical calculated deformation corresponds well to the first
part of the load-deflection curve of the experiment, however after 17.5 kN the numerical
curve linearly continues and shows no dropping point as was observed during the experi-
ment.
60
Figure 10: Numerical calculated deformation of the top centre corresponds well to the first part of the load-
deflection curve of the experiment
The numerical model did not predict debonding of the two material layers, as the contact
shear stress never exceeded the allowable shear stress of 2.8 MPa (Figure 11) [16]. How-
ever, note that the bond model is very complex and currently is intensively studied by
Wozniak [16], this is possibly also the reason because of the discrepancy.
a b
Figure 11: Contact shear stress on the contact surface is smaller than the allowable shear stress both in the x-
direction (a) and y-direction (b).
CONCLUSIONS
This paper studied the experimental and numerical analysis of a TRC reinforced concrete
dome. The dome - 2 m span, 0.2 m height and 20 mm thick - is reinforced with a 5 mm
thick TRC and subjected to a point load on the top. The results of the experiment showed
interesting results concerning both the production and the behaviour during loading. First-
ly, visible cracks formed along radial lines during the test exactly where the rubber rulers
were located while casting the concrete. Obviously, this phenomenon must be taken into
account and every adjacent segment of concrete must be casted directly next to each other.
In this way a good adhesion and a homogenous behaviour of the concrete shell is assured.
Secondly, debonding between the concrete and TRC layer occurred sooner than expected.
However, all measurements clearly indicate a composite action between the TRC and con-
crete before the crack initiation. As longs as debonding does not occur, the forces can be
transferred through the contact surface and the TRC form work works as a tensile rein-
forcement.
The experiment was simulated with the finite element software Abaqus. In general the nu-
merical model predicts the same trends and order of magnitude for the measured strains,
except for those near the edge, where exactly the opposite phenomenon occurs. The verti-
cal deformation of the top centre is also relatively well predicted up to debonding of both
constituents. However, some notable differences are seen between the experimental and
numerical analysis. These discrepancies can be explained by the experiment which failed
61
much faster than expected. The early debonding of the two layers was not foreseen and
also not predicted by the numerical model. For this reason, it is difficult to validate the
numerical model, thus the production and experiment must be improved. Simultaneously,
the numerical model needs further development to exclude the stress concentrations near
the load application zone and to make sure the model corresponds to the experiment.
These interesting and promising results encourage the further research on TRC reinforced
concrete shells, where after a setup amelioration and model verification, other shell geom-
etries can be examined.
REFERENCES
[1] A. Weilandt, Rolex Learning Center in Lausanne: From Conceptual Design to Execution, Proceedings
of the International Association for Shell and Spatial Structures, Valencia, Spain, 2009, 640-653.
[2] Nedcam, Spencer Dock Bridge Dublin (March 2014).
URL http://www.nedcam.com/spencer-dock-bridge-dublin.htm
[3] N. Cauberg, T. Tysmans, S. Adriaenssens, J. Wastiels, M. Mollaert, B. Belkassem, Shell Elements of
Textile Reinforced Concrete Using Fabric Formwork: A Case Study, Advances in Structural Engi-
neering 15 (4) (2012) 677-689.
[4] M. West, Thin-shell concrete from fabric moulds (March 2014).
URL http://www.umanitoba.ca/cast_building/resources.html
[5] P. van Hennik, L. Wagemans, Revival of pneumatic formwork for the construction of monolithic
(irregular) shells, Proceedings of the International Association for Shell and Spatial Structures, Mont-
pellier, France, 2004.
[6] D. Veenendaal, P. Block, Design process for prototype concrete shells using a hybrid cable net and
fabric formwork, Engineering Structures 75 (2014) 39-50.
[7] T. Tysmans, S. Adriaenssens, J. Wastiels, O. Remy, Textile reinforced cemencomposites for the de-
sign of very thin saddle shells: A case study, Proceedings of the 18th International Conference on
Composite Materials, Jeju Island, South Korea, 2011.
[8] D. Ehlig, F. Schladitz, M. Frenzel, M. Curbach, Textile concrete an overview of executed projects,
Beton- und Stahlbetonbau 107 (2012) 777-785.
[9] S. Verbruggen, O. Remy, J. Wastiels, T. Tysmans, Stay-in-Place Formwork of TRC Designed as
Shear Reinforcement for Concrete Beams, Advances in Materials Science and Engineering (2013).
[10] S. De Sutter, O. Remy, T. Tysmans, J. Wastiels, Development and experimental validation of a light-
weight stay-in-place composite formwork for concrete beams, Construction and Building Materials 63
(2014) 33-39.
[11] EP 0 861 216 B1, Inorganic Resin Compositions, Their Preparation And Use Thereof.
[12] H. Cuypers, Analysis and Design of Sandwich Panels with Brittle Matrix Composite Faces for Build-
ing Applications, Doctoral Thesis, Vrije Universiteit Brussel, Brussel, Belgium, 2001.
[13] M. Wozniak, T. Tysmans and J. Vantomme, Finite element modelling of glass fibre reinforced com-
posites with Inorganic Phosphate Cement matrix: Comparison of inbuilt Abaqus concrete models.
Proceedings of the 19th International Conference on Composite Materials, Montréal, Canada, 2013.
[14] S. Verbruggen, T. Tysmans and J. Wastiels, TRC or CFRP strengthening for reinforced concrete
beams: An experimental study of the cracking behaviour. Eng Struct, Vol.77, p.49-56; 2014.
[15] S. Verbruggen, Reinforcement of concrete beams in bending with externally bonded textile reinforced
cementitious composites. PhD thesis, Vrije Universiteit Brussel; 2014
[16] M. Wozniak, T. Tysmans, J. Vantomme and S. De Sutter, Numerical simulation of the bond behav-
iour of high-performance fibre-reinforced cement composites during the double-lap shear test. Pro-
ceedings of International RILEM Conference on Strain Hardening Cementitious Composites
(SHCC3), Dordrecht, Netherlands, 2014.
62
TEXTILE REINFORCED REFRACTORY CONCRETE AT HIGH

TEMPERATURE: MECHANICAL PROPERTIES UNDER UNIAXIAL
TENSILE LOADING
Dimas Alan Strauss Rambo1, Flávio de Andrade Silva2 and Romildo Dias Toledo Filho3,
1Department of Civil Engineering, Universidade Federal do Rio de Janeiro (COPPE/UFRJ), P.O.
Box 68506, 21945 – 970, Rio de Janeiro, RJ, Brazil, dimasrambo@gmail.com, 2Department of
Civil Engineering, Pontifícia Universidade Católica do Rio de Janeiro (PUC-Rio), Rua Marques de
São Vicente 225, 22451-900 - Rio de Janeiro - RJ, Brazil, fsilva@puc-rio.br, 3Department of Civil
Engineering, Universidade Federal do Rio de Janeiro (COPPE/UFRJ), P.O. Box 68506, 21945 –
970, Rio de Janeiro, RJ, Brazil, toledo@coc.ufrj.br
Abstract: In the present work the thermo-mechanical properties of a textile refractory

composite reinforced with basalt fibers under tensile loading was investigated. A series of
uniaxial tensile tests was performed under temperatures ranging from 25 to 1000°C. The
composites were produced as a laminate material using five basalt bi-directional fabric
layers as reinforcement. A high alumina cement matrix was used in the matrix composition
which was designed using the compressible packing method. Thermogravimetry analysis
were used to study the deterioration/phase changes as a function of the studied tempera-
tures. Scanning electron microscopy (SEM) was used to investigate the hydrated products
formed in the matrix and the damage processes in the fiber–matrix interfaces after expo-
sure to high temperatures.
INTRODUCTION
TRC´s (textile reinforced concrete) are used, actually, in a wide range of applications that
include: strengthening and repair in structural elements, protective linings, thin-walled el-
ements, facade elements, bridges and also freeform and lightweight structures. It´s im-
portant to consider, however, that in many of these applications the concrete and the textile
undergoes thermal effects, becoming the study of thermo-mechanical performance of TRC
absolutely indispensable [1, 2, 3]. The aim of this article is to investigate the effect of ele-
vated temperatures on the mechanical properties of a textile refractory composite rein-
forced with basalt fibers submitted to tensile loading. At first, the refractory composites
were produced with a cementitious matrix, made of calcium aluminate aggregates and cal-
cium aluminate cement, reinforced with low-cost basalt fabrics. The composites were test-
ed under temperatures ranging from 25 to 1000°C. The processing method used for making
bi-directional reinforced composites is described and then, tensile responses of the compo-
sites are reported.
MATERIALS AND MIX DESIGN

The matrix used in this research (compressive strength of about 45MPa) was designed
following the compressible packing model (CPM) routine [4, 5] and then adapted to the
rheology necessary to produce laminated TRC´s. The materials used in the TRC
composition were a Brazilian calcium aluminate cement (Secar 51) with alumina content
of approximately 50%, a synthetic calcium aluminate aggregate (with an alumina content
of approximately 40%) ranging from 0.001mm to 1.18mm and a polycarboxilate
63
superplasticizer in powder. The water/cementitious material ratio of the refractory concrete

was 0.35. Table 1 gives the composition of concrete matrix.
Table 1: Mix composition.
Composition Amount
Dense aggregate (kg/m³) 1416.2
Cement (kg/m³) 750
Superplasticizer (kg/m³) 4.87
Viscosity modifier agent - VMA (kg/m³) 0.562
Water (kg/m³) 262.6
Superplasticizer content (%) 0.65
Water/cementitious material ratio 0.35
A basalt textile commercialized by the chinese company Zhejiang GBF Basalt Fiber Co.
Ltd. was used as reinforcement for the TRC specimens. The basalt textile was produced
with a styrene-acrylic latex coating (43g/m²). The warp as well as the weft is formed by
about 800 monofilaments with average diameter of 13µm. The refractory concretes were
produced in a room with controlled temperature of 24°C ± 1°C using a planetary mixer
(previously moistured) of 5 liters capacity. The cementitious materials were homogenized
by dry mixing for 60 seconds prior to the addition of water. The mixture was blended for 5
minutes. The VMA was added after 4 minutes of mixing. The specimens were covered in
their molds for 24 hours prior to moist curing for 7 days in a cure chamber with 100% RH
and 24±1°C. Rectangular plates measuring 400 mm x 250 mm x 13 mm (length x width x
thickness) were produced for direct tensile tests using a lamination technique.
For the production of the plates, the concrete mixtures were placed in the acrylic molds.
The process started with a thin concrete layer placed on the bottom surface of the mold.
The second phase consisted on positioning the first fabric of basalt textile reinforcement
over the fresh concrete. The basalt textile was then pre-stressed manually, pressed and
smoothed in order to regularize and align the surface of the layer.
After the second phase, the procedure was repeated until reaching the desired number of 5
textile layers. Samples without textile layers were also produced. The samples suffered
manual vibration. After 7 days, the rectangular plates were cut, resulting in 4 specimens of
400 mm x 60 mm x 13 mm (length x width x thickness).
64
EXPERIMENTAL TESTING PROCEDURE

Heating treatment
The TRC specimens were heated up to 75, 150, 200, 300, 400, 600 and 1000°C and
subsequently cooled by a natural process inside the furnace. Experiments on specimens
stored at room temperature were also carried out. An electric oven with radiant heating
(metal alloy type Kanthal ‘‘A1’’ in alumina tubes) internally protected by perforated
ceramic plates was used. All samples were subject to a heating rate of 10°C/min in the
furnace. The temperature level was kept constant for 60 min in each target temperature.
The furnace was cooled down before the specimens were removed in order to prevent
thermal shock. The rate of cooling was not controlled.
Mechanical Testing
The direct tensile tests on the TRC plates were performed in a Shimadzu universal testing
machine model AGX – 100 kN and controlled by the actuator displacement at a rate of 0.4
mm/min. The force was transferred to the specimens via rotatable steel plates screwed to
the TRC plates. Deformation of TRC´s was measured using two LVDT´s. Four rectangular
shaped specimens measuring 400mm x 60mm x 13mm (length x width x thickness) were
tested using a gage length of 200mm with fixed–hinged boundary conditions. The tensile
load and actuator displacement were recorded.
DISCUSSION AND ANALYSES

The produced TRC showed to be strongly affected by thermal processing when submitted
to a tensile loading. Likewise, the mechanical properties of the basalt yarn as well as the
basalt fabric were also influenced. Figure 1 presents one representative curve obtained
from the TRC in the direct tensile test for each target temperature investigated in the
present work. The stress-strain curves were depicted in 3 distinct stages using roman
numerals, each one with its elastic modulus. The stage I corresponds to the elastic-linear
range where both matrix and fiber behave linearly. The post BOP stage is characterized by
formation of distributed cracking (stage II). In this phase, as the applied strain increases,
more cracks form culminating in a multiple cracking pattern along the sample. The stage
III, is characterized by a crack-widening phase that starts at the end of phase II (εt, II),
leading to a stretching of the fabric and, posteriorly, failure of the composite. Figure 1 also
presents the temperature ranges where occurs the decomposition of stable and metastable
hydrates present in refractory concrete matrix. The results of evaluation of all curves are
given in Table 2. As expected, strain-hardening behavior was observed for the TRC´s in
the majority of target temperatures, excluding only 600°C and 1000°C which presented,
respectively, strain-softening and fragile behavior. The positive influence of the preheating
up to 75°C and 150°C on the mechanical performance of TRC was clearly noticeable along
the entire course of stress–strain curves, mainly with relation to the tensile strength at the
bend-over point (BOP) and, specially, to the multiple cracking pattern (Figure 1 (b-c)).
This behavior was more pronounced for the temperature of 150°C where the TRC showed
σBOP 92% greater than the TRC tested at room temperature. With relation to the ultimate
tensile strength (σUTS), the contribution of the preheating up to 75°C and 150°C was not so
significant, generating increases of, respectively, 10.7% and 13.4%. At these temperatures
the main process induced in the matrix by the thermal regime is the evaporation of free
water, lost of structural water of the AHx gel and the dehydration of metastables CAH10
(CaO.Al2O3.10H2O) and C2AH8 (2CaO.Al2O3.8H2O). For temperatures greater than 150°C,
65
the stress level at the BOP and the ultimate tensile strength decreased gradually with
increasing temperature.
This decrease was accompanied by a significant reduction of the elastic modulus from
200°C. The TRC heated to 200°C showed a multiple cracking pattern, similar to the TRC
heated to 150°C, however significant sings of strength deterioration were noticed above
this temperature.
(a) (b) (c) (d) (e) (f) (g)
Figure 1: Influence of temperature on the (a) tensile stress curves of TRC´s and in the cracking pattern at (b)
room temperature and after heating process at (c) 75°C, (d) 150°C, (e) 200°C, (f) 300°C and (g) 400°C.
Table 2: Average results of the direct tensile test.
First crack values Post crack values

Target
PBOP σBOP σBOP, I Et,I PUTS σUTS Et,II Et,III εt,II εUTS,III
temperature
(kN) (MPa) (%) (GPa) (kN) (MPa) (GPa) (GPa) (%) (%)
25°C 2.85 3.45 0.011 34.64 11.13 13.49 0.45 0.67 0.42 1.58
75°C 3.73 4.85 0.021 28.57 11.48 14.94 0.31 0.79 0.71 1.66
150°C 5.10 6.65 0.024 31.63 11.74 15.30 0.12 0.73 0.62 1.64
200°C 4.19 5.09 0.036 23.82 10.22 12.40 0.15 0.64 0.39 1.50
300°C 3.47 4.31 0.033 20.52 9.64 11.97 0.19 0.75 0.50 1.59
400°C 1.87 2.42 0.025 10.93 3.79 4.98 0.36 0.90 0.47 0.68
600°C 1.56 1.98 0.044 5.37 - - - - - -
1000°C 1.10 1.33 0.037 3.96 - - - - - -
The thermogravimetric analysis (TGA) showed that the loss in strength ranging from
150°C to 200°C, can be explained by the end of the dehydration of metastables CAH10,
C2AH8 and by the start of dehydration of the stables AH3 (Al2O3.3H2O) and C3AH6
(3CaO.Al2O3.6H2O) (see Figure 2). For temperatures between 200°C and 300°C, remain-
ing hydrates AH3 and C3AH6 are almost completely dehydrated (also detected by TGA
66
analyses) causing more losses in the elastic modulus and in the tensile strength. Between
the target temperatures of 300°C and 400°C severe strength loss was observed.
(b) (c)
(a)
Long prisms Plates
(d)
Cubes
Figure 2: Thermal analysis of the refractory concrete (7 days of age) on initial sample mass basis (a) and
SEM images of the hydrated products CAH10 (b), C2AH8 (c) e C3AH6 (d)
For temperatures greater than 400°C the specimens failed abruptly and large cracks oc-
curred. At these temperatures, both the refractory matrix and basalt fabric showed signifi-
cant signs of deterioration causing pronounced brittle behavior beyond the peak stress. The
tensile strength of basalt fibers, as well as similar fibers (S-Glass fibers and E-Glass fibers)
[4], decrease rapidly at temperatures greater than 400°C. As a consequence, the thermo-
mechanical behavior of the TRC was largely affected above this temperature. As reported
previously, the TRC heated at 600°C showed a strain softening behavior. At this tempera-
ture, the minor crack-bridging action of the fibers could be neglected if compared to the
results at the lower target temperatures. The composite heated at 1000°C, however, showed
fragile failure. This brittle behavior is attributed, in part, to the melting of the filaments in
the yarns that occurred for lower temperatures than those defined by the manufacturer
(about 1350°C). The average tensile strength for the composite heated at 1000°C at the
first crack was 1.33MPa, 38.5% of the σBOP at room temperature. There was no visible
concrete spalling for any one of the used target temperatures (75, 150, 200, 300, 400, 600 e
1000°C), even with samples in the saturated condition. This can be attributed to the small
thickness of the samples and to the textile reinforcement that creates relieve channels to the
internal vapor pressure present in the concrete pores.
A micro-structural investigation using the SEM was carried out to confirm the damage
process on the TRC discussed previously. Figure 3 shows fibre–matrix interface micro-
graphs of the basalt yarns in its natural condition (Figure 3 (a)) and after exposure to
200°C, 400°C and 1000°C. At 75°C the styrene-acrylic latex coating starts to transforms
from visco-elastic to plastic, thus, impregnating more strongly the matrix and the internal
filaments. At this temperature a pronounced merging of polymer towards the matrix oc-
curs. Between 75°C and 150°C, as the temperature is not able to decompose the polymer,
after cooled, the styrene-acrylic latex becomes stiffer and visco-elastic again. The newly
generated polymer interlocks change the interface mechanisms on the TRC increasing the
cross links between the fibers and matrix. As reported in previous researches [6, 5], the
penetration of coating in the yarn, in this case due to the heating regime, makes pro-
67
nounced changes in the composite behavior. This occurs because the filaments are stressed
more equally, taking part in the load support and, consequently, increasing the tensile
stress.
Furthermore, the irreversible shrinkage of the hardened paste (between 25°C and ~200°C)
[7, 8], that remains after cooling down the material, can lead to an improve in the bond
between the fibers and matrix. This microstructure evolution occurs due to the evaporation
mechanism of free water and the dehydration of calcium aluminate hydrates and is accom-
panied by micro-cracking in the paste. The first signs of degradation of the polymer started
to appear at 200°C. From the Figure 3 (b) (TRC preheated at 200°C) it can be seen that the
coating between the fibers is porous and presents less interlocks than the coating at the
room temperature. At 200°C the copolymer take the form of thin-films or scales deposited
over the filaments. As a result of the less amount of coating between the fibers and in the
fiber–matrix interfaces, significant strength losses were observed. This means that even
though the fiber content remain constant the presence of coating can become a determinis-
tic factor in the tensile response of the composite. At 300°C, the polymer is still present,
however, the thin layers of polymer almost disappear. Figure 3 (c) shows that the coating
surrounding the fibers was entirely lost at 400°C. As a consequence, the tensile behavior of
the TRC change leading to a reduced crack-widening phase (stage III) and a sudden drop
in the serviceability limit of the composite. After preheating up to 600°C, both matrix and
textile shown to be weakened by the heating regime resulting in a strain softening behavior
due to the sliding of the fibers. Figure 3 (d) shows the fiber–matrix interface micrographs
highlighting the volumetric variation caused by the target temperature of 1000°C to the
filaments.
(a) (b) (c) (d)

Figure 3: Fiber-matrix interface: (a) in its natural condition and after exposure to temperatures of (b) 200°C,
(c) 400°C and (d) 1000°C.
CONCLUSIONS
The experimental investigation shows that the mechanical residual properties of the TRC
exposed to high temperatures depend on the loss in strength of the matrix and of the fabric.
Furthermore, the coating was strongly affected by the heating regime changing the fiber-
matrix bonding properties at certain temperatures. At temperatures up to 150°C the soften
and re-stiffen mechanism of copolymer helped in generating more efficient cross links be-
tween the fibers and matrix. At higher temperatures (400 and 600°C), the remaining coat-
ing was completely deteriorated and the behavior of TRC became brittle.
68
ACKNOWLEDGEMENTS
The authors gratefully acknowledge the Brazilian Agency CNPq for its partial financial
support.
REFERENCES
1. Krüger M, Reinhardt HW. (2006) Chapter 6: Fire resistance. In: Report 36: Textile Reinforced Con-
crete - State-of-the-Art Report of RILEM Technical Committee 201-TRC. Ed: Wolfgang Brameshu-
ber.
2. Silva FA, Butler M, Hempel S, Toledo RD, Mechtcherine V. Effects of elevated temperatures on the
interface properties of carbon textile-reinforced concrete. Cement & Concrete Composites. 48 (2014)
26–34.
3. Reinhardt HW. On the Biaxial Testing and Strength of Coated Fabrics. Experimental Mechanics. 16
(1976) 71-74.
4. De Larrard F. (1999) Concrete mixture proportioning: a scientific approach. London. E&FN SPON.
5. Sedran T. Rhéologie et rhéométrie des bétons: application aux bétons autonivelants. Doctoral disserta-
tion. Ecole Nationale des Ponts et Chaussées. (1999) 484p.
6. Köckritz U, Offermann P, Jesse F, Curbach M. Influence of textile manufacturing technology on load
bearing behavior of textile reinforced concrete. In: Proceedings of the 13th International Techtextile-
Symposium. (2005).
7. Innocenti MDM, Cardoso FA, Akyioshi MM, Pandolfelli VC. Drying Stages during heating of high-
alumina. Ultra-low-cement refractory castables. Journal of the American Ceramic Society. 86 (2003)
1146–1148.
8. Bazant ZP, Kaplan MF. Concrete at high temperatures. material properties and mathematical models.
concrete design and construction series. England: Longman Group; 1996
69
70
BEARING BEHAVIOR OF IMPREGNATED TEXTILE

REINFORCEMENT
Sergej Rempel (1), Christian Kulas (2), Josef Hegger (1), (1) RWTH Aachen University, Institute
of Structural Concrete, Germany, (2) solidian GmbH, Germany
Abstract:The current trend in the building industry shows a growing need for high-
performance material with high tension and pressure strength. An innovative building
material, which satisfies the wishes of the construction designer, is textile-reinforced
concrete (TRC). The combination of high pressure strength concrete and corrosion-
resistant reinforcement, which furthermore has high tension strength, enables slender
concrete structures. The realized applications prove the existing potential. The further
development of the textile reinforcement offers new possibilities for bearing structures.
One main progress was the impregnation of the textiles with styrol-butadiene rubber or
epoxy-resin. This process enables extensive higher failure stresses. Additionally, the
durability and the strength improve. Consequently the efficiency rises. Within the scope of
this paper, the bending behavior of slabs and I-beams are presented. Furthermore, a
calculation approach for bending moments is introduced.
INTRODUCTION
Current trends in the building industry are increasingly heading towards high-performance
materials with high tensile or compressive strengths. One of those innovative materials is
textile-reinforced concrete (TRC), which uses mesh-like, non-corrodible reinforcements in
combination with a fine-grained concrete [1], [2]. Since the reinforcements do not corrode,
it is possible to significantly reduce the concrete cover compared to ordinary steel-
reinforced constructions. This leads to thin-walled and slender concrete members with
high-quality concrete surfaces meeting the needs of modern architecture Figure 1. For
complex constructions, like T-beams, shaped reinforcement structures are necessary, which
cannot be realized with non-impregnated textiles due to a lack of form stability (Figure 1
right). Therefore, by impregnating the textiles with, for example, epoxy-resin or styrene-
butadiene, inherently stable and manageable reinforcement structures (prepreg structures)
can be fabricated. Another important advantage of those structures is the improved bond-
behavior of the inner filaments of a roving leading to tensile stresses that are two or three
times higher compared to non-impregnated textiles [3]. At last, the resin improves the du-
rability of the reinforcement, which is especially important for reinforcements made of
glass-filaments [4].
71
Figure 1: View of the Textile- Reinforced Façade in Nimwegen (picture: Ben Vulkers, LIAG Architect-
en) and a cross-section of the TRC-bridge with shaped textiles (picture: solidian)
Textile-reinforced concrete is often used as construction material for façade structures with
concrete covers of about 10 mm to 15 mm (Figure 1, left). Hegger et al. give detailed in-
formation on the applicability of TRC for new constructions, especially on ventilated fa-
çade structures and sandwich panels [5], [6]. Furthermore, TRC is also applied for
strengthening issues, e.g. strengthening of concrete shell structures [7]. Another field of
application is in the construction impact from chlorides, e.g. due to de-icing salt, like
bridges or maritime constructions. An actual example can be seen in a pedestrian bridge
made of TRC located in Albstadt, Germany (Figure 1, right). The bridge has an overall
length of 97 m and is subdivided in six parts, with a maximum length of 17.20 m. Since the
superstructure is made of TRC, an extreme slenderness ratio of H:L = 1:35 (height to
length) is realized [8].
This article deals with the bending load-bearing behavior of impregnated textiles and pre-
sents test-results along with a calculation approach. While the bending moment of TRC
members with non-impregnated textiles can only be calculated under consideration of em-
pirical reduction factors, it is shown that those factors are not necessary for impregnated
textiles.
MATERIALS
Textile reinforcement
Actually alkali-resistant glass (AR-glass) or carbon are the main materials used for current
TRC applications. The basic materials are hair-thin filaments with diameters of 14 μm for
AR-glass and about 7 μm for carbon. A bundle of hundreds, up to ten thousands, of these
filaments shape a roving, which will be processed to make flat textile-reinforced grids. The
textile grids can be used directly without impregnation or impregnation with Styrene-
Butadiene or Epoxy-Resin as reinforcement elements for concrete structures. The impreg-
nation is responsible for higher failure stresses and consequently for an increase in the ef-
fectivity of the textile materials. Also the loss of strength of the AR-Glass, because of the
alkalinity, has to be considered. The impregnation counteracts it and reduces the losses.
72
The failure stresses is dependent on the material, cross-section of the roving, the binding-
type and impregnation and can be seen in Table 1.
Table 1: Properties of the selected textiles

Properties Units Textile 1 Textile 2 Textile 3
Material - Carbon AR-Glass AR-Glass

Penetration - Epoxy-Resin Epoxy-Resin Styrene-Butadiene
Titer tex 1650 3600 2400
Roving dis- 0° mm 21 15/ 5 21
tance 90° mm 23 15/ 2 X 7,5 23
0° mm 0,92 1,34 0,9
Cross section
90° mm 0,92 1,34 0,9
0° N/mm² 2423 1332 1173
Tensile
stress1) 90° 3054 1494 1191
N/mm²
1)
determined in tensile tests on rovings, which have been cut out of the textile
The grids Textile 1 and Textile 2 are impregnated with epoxy resin and differ only in the
distance of the rovings and the filament material. Unlike Textile 1 and Textile 2, the Tex-
tile 3 is impregnated with styrene-butadiene. Every impregnated textile is available as pla-
nar reinforcement and as a U-shaped reinforcement.
Concrete
The use of a typical concrete with its standard grain size was not possible because of the
slight openings in the textile reinforcement grids. By using grains with diameters that are
too large, the concrete would not flow through the meshes properly and the textile would
act as a sieve, resulting in a separation layer within the member. For this reason, the Insti-
tute of Building Material Research of the RWTH Aachen University (ibac) developed dur-
ing the SFB 532 (Collaborative Research Center) a recipe for a concrete with a maximum
grain size of 0.6 mm [2]. Due to the high cement content, this concrete tends to shrink and
creep more than a concrete with a larger grain-size. Thus, the investigations presented in
this paper have been done with a concrete with a maximum grain-size of 5.0 mm (Concrete
1, Table 2), developed by Hering Bau, a construction company specializing in architectural
concrete.
73
Table 2: Properties of the applied concrete
Properties Units Concrete 1

Type of cement - CEM II A-LL 42,5/R
CEM I 52,5 R
Cement content kg/m³ 410
Maximum grain size mm 5,0
Aggregate content kg/m³ 1540
Compression strength
- cubes 40x40x40 mm³ MPa 100
- cylinder Ø150 mm/h=300 mm MPa 87
Flexural strength MPa 10,6
YOUNG’s Modulus MPa 45420
The concrete offer high compression strengths of about 100 MPa and can be classified as
high-strength concretes according to Eurocode 2 [9]. Since TRC members are often de-
signed as fair-faced concrete structures, a high flexural strength is an important property.
Concrete 1, which is used within this paper, has a flexural strength of 10.6 MPa. The main
mechanical properties of Concrete 1 are shown in Table 2. The stress-strain behavior of the
fine-grained concrete was determined in compression tests on cylindrical specimens with a
diameter of 150 mm and a height of 300 mm. According to Eurocode 2 the stress-strain
curve can be calculated with formula (1), which gives an adequate match with the test re-
sults shown in Figure 2. The Young’s Modulus is determined at a stress of 0,4×fcm, where
linearity reaches its limit.
100
fcm = 87 MPa
compression stress σc [MPa]
90
80
70
60 2
ε ⎛ε ⎞
50 1,3 ⋅ c − ⎜ c ⎟
0,4·fcm 2.4 ⎝ 2.4 ⎠
40 σ c (εc ) = 87 ⋅ (1)
= 35 MPa ⎛ε ⎞
1 − 0.7 ⋅ ⎜ c ⎟
30 ⎝ 2.4 ⎠
20 εc1 = 2,4 ‰
10 Eurocode 2 [13]
test results (n=6)
0
0,8
0 1 2 3
strain εc [‰]
Figure 2: Stress-strain behavior under compression of concrete C1: test-results and calculation according
to Eurocode 2 [9].
74
BENDING BEHAVIOR
Experimental investigations
Typical applications for TRC are façade panels with common slab thicknesses of about 30
mm [10], thus, slab specimens with rectangular cross-section were considered for investi-
gating the bending behavior. Additionally, I-beams with an overall height of 120 mm, as
depicted in Figure 3, according to Hegger and Voss [11] were tested, since with this cross-
section it is possible to study the load-bearing behavior of shaped reinforcement structures
like U-shaped textiles. In both members the reinforcement ratio in the tensile zone was
varied by adding planar textile layers. The specimens were tested in four-point bending
tests with a span of 600 mm (slab) or 1,300 mm (I-beam) respectively. The members were
tested displacement-controlled with a loading-rate of 3 mm/min.
a) b)
Figure 3: Specimen for investigating the bending-behavior: left) slab; right) I-beam
Prediction of the ultimate bending moment
The procedure for calculating the bending moment is shown in Figure 4 on the example of
an I-beam and follows the calculation methods of steel-reinforced concrete.
Figure 4: Distribution of strain, stress and forces in a bending member
The basic assumption is a linear strain distribution with the concrete compression εc and
the strain εt,max of the roving with the inner level zi. The stresses and forces can be calcu-
lated under consideration of the material laws of textile and concrete. Therefore, not only
the rovings in the lower flange but also the rovings in the web have to be considered. To
recalculate a test, the specimen was cut after the test and the real effective depths di of each
roving were measured. The compressive force can be calculated by integrating the stress
75
distribution σc. Together with the tensile forces σt,i of each roving the bending moment
can be calculated with the formula (2).
Mu,cal = ∑ Ft ,i ⋅ zi = ∑ ( At ⋅ σ t ,i ) ⋅ zi (2)
i i
In Figure 5 the ultimate bending moment achieved in the tests divided by the calculated
bending moment is plotted over the reinforcement ratio ρl. In the diagram results from both
the four-point bending tests on I-beams and on slabs with a rectangular cross-section are
considered. The predicted bending moment gives an adequate match to the experimental
results. A difference of only 5% is due to test scattering.
a) b)
Figure 5: Bending tests: a) specimen in failure state; b) Comparison of calculation and test results (each
data point is one single test)
The results in Figure 5 show that for impregnated textiles the main input parameter in his
calculation model is the roving strength σt,max.
76
CONCLUSION
Impregnated textiles have certain advantages compared to the ones that are not impregnat-
ed, like higher tensile stresses, improved durability and a sufficient dimensional stability.
For calculating the bending behavior of concrete members which are reinforced with im-
pregnated textiles, the ultimate bending moment can be calculated with a simple iteration
of the strain distribution. The simple calculation model was used for structural analyses of
several TRC applications. Due to the fact that the model is based on the well-known calcu-
lation methods of steel-reinforced concrete, already many engineers were able to realize
TRC constructions.
ACKNOWLEDGEMENT
The authors acknowledge the German Research Foundation (DFG) for financing the
Transfer Project T08 within the scope of the Collaborative Research Center (SFB) 532 at
RWTH Aachen University. Furthermore, the authors thank the partners Groz-Beckert KG,
V. Fraas Solutions in Textile, SGL Carbon, Hering Bau, the Institute of Materials Building
Research and the Institute of Textile Technology at RWTH Aachen University for their
support and the efficient collaborative relationship.
REFERENCES
[1] J. Hegger, S. Voss, Investigations on the bearing behaviour and application potential of textile rein
forced concrete, Engineering Structures, Issue 30 (2008). pp. 2050-2056.
[2] B. Banholzer, T. Brockmann, W. Brameshuber, Material and bonding characteristics for dimension
ing and modelling of textile reinforced concrete (TRC) elements, Materials and Structures 39, Issue
8 (2006), pp. 749-763.
[3] M. Raupach, J. Orlowsky, T. Büttner, U. Dilthey, M. Schleser, Epoxy-impregnated textiles in con
crete – load bearing capacity and durability. In: W. Brameshuber (Ed.): Textile Reinforced Con-
crete, RILEM Report 36. Bagneux, France. ISBN 2-912143-99-3 (2006), pp. 77-88.
[4] T. Büttner, J. Orlowsky, M. Raupach, M. Hojczyk, O. Weichold, Enhancement of the durability of
alkali-resistant glass-rovings in concrete. In: W. Brameshuber (Ed.): 2nd ICTRC – Textile Rein-
forced Concrete. Proceedings of the International RILEM Conference on Material Science (MatSci).
International RILEM Conference on Material Science (MatSci), Aachen, 06.-08.09.2010. -ISBN
978-2-35158-106-3. pp. 333-342.
[5] J. Hegger, C. Kulas and M. Horstmann, Realization of TRC façades with impregnated AR-glass
textiles, Key Engineering Materials, Issue 466 (2011), pp. 121-130.
[6] C. Kulas, M. Schneider, N. Will and R. Grebe, Ventilated structures made of textile reinforced con
crete – structural behavior and construction, Bautechnik, Vol. 88, Issue 5 (2011), pp. 271-280.
[7] D. Jesse, F. Jesse, High Performance Composite Textile Reinforced Concrete –Definitions, Proper
ties and Applications. In: 3rd International fib Congress, Washington, D.C., 29.05-02.06.2010, paper
no 157 on CD-ROM.
[8] J. Hegger, C. Kulas, M. Raupach, T. Büttner, Load-bearing behavior and durability of a slender
textile reinforced concrete bridge. Beton- und Stahlbetonbau 106, Issue 2 (2011), pp. 72-80.
[9] Eurocode 2 (DIN EN 1992-1-1): Design of concrete structures - Part 1-1: General rules and rules for
buildings; DIN (Deutsches Institut für Normung), Beuth Verlag GmbH, Berlin, 01.2011.
[10] C. Kulas, M. Schneider, N. Will and R. Grebe, Ventilated structures made of textile reinforced con
crete – structural behavior and construction, Bautechnik, Vol. 88, Issue 5 (2011), pp. 271-280.
[11] J. Hegger, S. Voss, Application and dimensioning of textile reinforced concrete, In: T. C. Trianta
fillou (Ed.): Fibre Reinforced Polymer Reinforcement for Concrete Structures. Proceedings of the
8th International Symposium for Fiber-Reinforced Polymer Reinforcement for Concrete Structures;
Patras, Greece, 16.-18.07.2007, pp. 632-633. -ISBN 978-960-89691-0-0.
77
78
FLEXURAL BEHAVIOR OF MESH-AND-FIBER REINFORCED

CEMENTITIOUS COMPOSITES
P.B. Sakthivel (1), A. Ravichandran (2) and N. Alagumurthi (3), (1) Department of Civil
Engineering, Pondicherry Engineering College, Puducherry, India, (2) Department of Civil
Engineering, Christ College of Engineering and Technology, Puducherry, India, (3) Department of
Mechanical Engineering, Pondicherry Engineering College, Puducherry, India
Abstract: The present study is an attempt to analyze the flexural behavior of mesh and fiber
reinforced cementitious composites. For this purpose, four-point bending tests were con-
ducted on slabs (700 mm × 200 mm × 25 mm thick) with hybrid reinforcing mix combina-
tions of steel mesh (SM - 3, 4, 5 layers) and stainless steel (SS) fibers (0-2.5% of volume
of specimens, with 0.5% interval). Higher performance is seen in hybrid steel mesh and
stainless steel fiber reinforced cementitious composites (HSMSSFRCC) compared to Steel
Mesh Reinforced Cementitious Composites (SMRCC). Strain-hardening behavior of
HSMSSFRCC was evidenced by multiple cracking behavior with excellent crack distribu-
tion. The crack width was found to reduce on increase in mesh layers and fiber percentage.
Also, the crack spacing in the moment region was found to decrease on increase in rein-
forcement volume fraction. Statistical regression models were developed for predicting the
flexural strength of SMRCC and HSMSSFRCC.
INTRODUCTION
Steel Mesh Reinforced Cementitious Composites (SMRCC) (traditionally known as ferro-

cement) are thin shell-like elements (10-25 mm thick) which are developed as an alterna-
tive material to conventional cement concrete and produced with hydraulic cement mortar
(Ordinary Portland Cement, river sand and potable water) and wire mesh reinforcement.
SMRCC is increasingly used as a major construction material as prefabricated slab/ roof
elements and wall paneling systems, sunscreens, curtain walls and exterior cladding panels.
Some innovative methods such as adding discontinuous fibers in SMRCC have been sug-
gested by Wang et al. 2004 [1], Shannag and Ziyyad 2007 [2], Sakthivel et al. 2014 [3] to
improve its mechanical strength; and such hybrid composites demonstrate multiple crack-
ing behavior upto the maximum post-cracking stress, while the multiple cracks appear be-
fore failure for a strain-hardening composite [3-6]. The addition of fibers to cement-based
matrix has little or no effect on its pre-cracking behavior but is very effective in its post-
cracking ductility [6]; and the higher load carrying capacity of hybrid composites after first
cracking is an indication of deflection-hardening behavior [3].
Shannag and Ziyyad (2007) [2] have assessed the flexural behavior of fiber reinforced
SMRCC plates of size 300 mm × 75 mm × 12.5 mm mixing woven steel square wire mesh
(2 and 4 layers) and discontinuous glass and brass coated steel fibers (with 2% volume
fraction) in hybrid form, and there was a substantial increase in flexural strength to failure
when steel mesh layers were increased from 2 to 4 layers. Shannag and Ziyyad (2007) [2]
have found that there was also a significant increase in flexural strength of 2.6 times when
brass coated steel fibers were added to mesh layers (compared to SMRCC, control speci-
mens without fibers), and numerous fine and well distributed cracks of smaller width were
noticed (when compared to control specimens). They have also reported that the addition
79
of glass fibers to steel meshes (with smaller mesh grid spacing) have improved the flexural
strength, and also there was a crack reduction in the specimens compared to their steel
counterparts. In another study, Lin et al. (2011) [7] have conducted experiments on Poly-
ethylene (PE)-fibrous SMRCC specimens combining various volume fraction (Vf) of PE-
fibers at 0.5%, 1.0% and 1.5% and 0-2 layers of steel mesh reinforcement; and it was
found that the PE-fibrous SMRCC have exhibited superior performance over conventional
SMRCC specimens in terms of both tensile stress and strain capacities; short-discontinuous
PE-fibers to the cement binding matrix have led to multiple micro-cracking and ductile
behavior; and PE-FRCC matrix with Vf of 1.5% have demonstrated the highest stress and
strain capacities under tension. The objective of this experimental investigation is to under-
stand the flexural and cracking behavior of Hybrid Steel-Mesh and Stainless-Steel Fiber
Reinforced Cementitious Composites (HSMSSFRCC) with varying volume fraction of
hybrid reinforcing mix of mesh and fibers, and compare the results with Steel Mesh Rein-
forced Cementitious Composites (SMRCC).
EXPERIMENTAL PROGRAM
Materials and Methods

Slab size of 700 mm X 200 mm X 25 mm (3 replicate specimens) were cast for test speci-
mens (HSMSSFRCC) and control specimens (SMRCC). Ordinary Portland Cement (OPC-
53 Grade) and natural river sand (passing through 2.36 mm sieve) are used with a sand-
cement (s/c) ratio of 2:1 and water-cement (w/c) ratio 0.43 in line with previous studies
[3]. Galvanized steel mesh and non-corrosive stainless steel (SS) fibers (crimped type)
manufactured by Stewols India (P) Ltd., and the properties are given in Table 1. Table 2
shows the steel mesh reinforcement used in SMRCC under Type A3, A4 and A5 (with
steel mesh of 3, 4 and 5 layers) and hybrid steel mesh and stainless steel fibers in
HSMSSFRCC (Type C3A to C3E with 3 layers of steel mesh and SS fiber - 0 to 2.5% re-
spectively, Type C4A to C4E with 4 layers of mesh and fiber - 0 to 2.5% respectively and
Type C5A to C5E with 5 layers of mesh and fiber - 0 to 2.5% respectively).
Table 1: Material Properties

Mesh Fibers
Mesh Type Galvanised Steel Mesh Fiber Type Stainless Steel Fibers
Diameter 0.7 mm Base Material Modified Olefin
Tensile Strength 512.36 N/mm2 Length 12.5 mm
Yield Stress 406.51 N/mm2 Diameter 0.45 mm
Elongation 7.12% Aspect Ratio 27.77
Weld Shear 250 N/mm2 Fiber Shape Crimped
Ultimate Tensile Strength 1353 N/mm2
Table 2: Hybrid Reinforcing Mix Combinations

Mes Fiber- Fiber- Fiber- Fiber- Fiber- Fiber-
Type Type Type Type Type Type
nos. Vf=0.0% Vf=0.5% Vf=1.0% Vf=1.5% Vf=2.0% Vf=2.5%
3 A3 M3SSF0.0 C3A M3SSF0.5 C3B M3SSF1.0 C3C M3SSF1.5 C3D M3SSF2.0 C3E M3SSF2.5
Vr= Volume fraction of mesh reinforcement; M3 - 3 layer Steel Mesh at Vr=0.74%; M4 - 4 layer Steel Mesh at Vr=0.98%; and M5 - 5 layer Steel Mesh at
Vr=1.23%; SSF- Stainless Steel Fibers; Vf=Fiber Volume fraction (0%, 0.5%, 1%, 1.5%, 2.0%, 2.5%)
80
Casting of Specimens
SMRCC (control specimens) are cast with hydraulic plain cement mortar (with sand and
cement using the required s/c and w/c ratios) and steel mesh reinforcement. The test spec-
imens were cast with Stainless Steel (SS) fibrous mortar and steel mesh. SS fibrous mortar
is prepared by adding discontinuous fibers of crimped type (12.5 mm length and 0.45 mm
diameter) (Fig.1A) to dry cement mortar and thoroughly mixing the ingredients; sube-
quently, water is added to dry SS fibrous mortar and mixed well (Fig.1B). The SS fibrous
mortar thus prepared (Fig.1C) is to used to cast the test specimens, along with steel mesh,
as shown in Fig.1D.
The first base layer of 3 mm was laid on the bottom of the slab moulds (using removable
3mm thick glass spacers) and evenly spread with mason's trowel and top surface properly
levelled for proper placement of steel mesh over it. Now the first layer of mesh is placed
over this base cover mortar. Over this first layer of mesh is laid the next layer of mortar
(Fig.1D), and evenly spread and finished, and then the second mesh layer was placed over
the mortar. The same process is repeated till the desired number of mesh layers (3,4,5 nos.)
are laid. Finally, the top mesh layer is provided with reinforcing cover of 3 mm using re-
movable glass spacers and properly finished. At each stage, the mortar is compacted man-
ually by using hand casting pressure with mason's trowel and evenly finished at the time of
laying mortar and mesh. Demoulding of the specimens was done after 24 hours. The spec-
imens were cured with water for (28 days) using gunny sacks. Before testing, the speci-
mens were white-washed to monitor the cracking patterns during the flexure tests. Unique
identification marks are given for the specimens using indelible ink.
The cylinder compressive strength of SS fibrous cement mortar (at 28 days) (using 100
mm X 200 mm height cylindrical moulds) was found to be 30.14, 31.00, 33.54, 36.52 and
38.22 N/mm2 (for fiber volume fraction Vf %=0.5, 1, 1.5, 2, 2.5%) when compared to plain
cement mortar specimens of 25.48 N/mm2. The split-tensile strength of SS fibrous mortar
cylindrical specimens (of size 100 mm × 200 mm ht.) (at 28 days) was found to be 5.63,
6.79, 7.21, 7.64 and 8.28 N/mm2 (using SS fibers of Vf %=0.5, 1, 1.5, 2 and 2.5%), when
compared to plain cement mortar specimens of 4.56 N/mm2. The flexural strength of SS
fibrous mortar (at 28 days) using prismatic beam specimens of 40 mm X 40 mm X 160
mm) have shown 5.94, 6.39. 7.27, 7.84 and 8.09 N/mm2 (Vf %=0.5, 1, 1.5, 2 and 2.5%)
when compared to plain cement mortar specimens of 5.31 N/mm2.
Figure 1: A) SS Figure 1: B) Water in Figure 1: C) SS Fibrous Figure 1: D) Mortar laid

Fibers Dry SS Mortar Mortar over mesh
81
Flexural Testing
The flexural performance of HSMSSFRCC (slab elements) was determined by four point
bending test at 28 days; and the loading set-up for the bending test was inverted in accord-
ance with provisions given in standards, JCSE-SF4 [8] and ASTM C78/C78M-10 [9] for
the purposes of studying the cracks (visible to the naked eye) on the tension side of the
specimens. The loading frame has been designed taking into account the extraneous de-
formations relating to seating or support settlements, twisting or rotations [3]. The flexure
loading test was conducted on slab specimens in simply supported condition with support
span of 600 mm and loading span at one-third points (200 mm), allowing for pure bending
in the central moment region of the slab [3,6]. To measure the flexural deflection of the
specimen during loading, linear variable differential transducer (LVDT-100 mm) was
mounted on the top side of the frame at mid-span (touching the slab specimen). The load
was manually applied through a spreader beam using hydraulic jack, and for every small
incremental load (measured by the load cell), the corresponding deflection was recorded by
computerized data-acquisition system. The first crack and ultimate crack width were meas-
ured using a crack measuring microscope, with high level of accuracy.
RESULTS AND DISCUSSION
Flexural Performance
The flexural strength results obtained from the experimental work are presented in Table 3,
crack distribution and width in Table 4, and crack distribution and numbers in Fig.6. For
each load increment, the first crack, multiple cracks and ultimate crack/s were formed
along the width in the bending moment region at the tension side of the specimens. From
the experimental results shown in Table 3, the load carrying capacity of HSMSSFRCC is
higher than SMRCC. Table 4 shows that the crack width and crack spacing of
HSMSSFRCC are lower than SMRCC, demonstrating higher performance of
HSMSSFRCC. From Table 3 and Fig.2A and 2B, it is seen that the control specimens have
shown first crack flexural load (Pcr-Exp.) of 1.45 kN, 1.65 kN and 1.75 kN, and ultimate
flexural load (Pu) of 1.6 kN, 2.05 kN and 2.65 kN respectively for Type A3, A4 and A5
respectively. From Table 4, Figs. 3A and 3B, Figs.6A, 6G and 6M (for Type A3, A4 and
A5), the first crack width of 0.063 mm, 0.045 and 0.031 mm, and ultimate crack width of
0.208 mm, 0.205 mm and 0.161 mm (respectively) is observed. The crackwidth has re-
duced from Type A3 to A4 and Type A4 to A5. The average number of cracks for SMRCC
were 1, 2 and 5 cracks for Type A3, A4 and A5, with a high crack spacing (of about 200
mm) for Type A3 and A4 but with substantial reduction from 192 mm (for Type A4) to
51.25 mm (for Type A5), when steel mesh layers were increased from 4 to 5 layers.
The experimental results of HSMSSFRCC shown in Table 3 and Fig.2A bring out that the
flexural first cracking load (Pcr) is 1.5 kN (lowest) for Type C3A and 2.30 kN (highest) for
Type C5C. From Table 4, the flexural strength improvement ratio for first crack load
(FSIR-Pcr) between the test and control specimens are 27.59% between Type C3E and
Type A3, 30.30% between Type C4C and Type A4, 31.43% between Type C5C and Type
A5. Also, from Table 4 and Fig.3A, the first crack width of 0.059 mm (highest) is observed
for Type C3A and 0.013 mm (lowest) for Type C3E, Type C4E, Type C5D and Type C5E.
Lower crack width of 0.013 shows excellent cracking performance of HSMSSFRCC.
Similarly, from Table 3 and Fig.2B, the ultimate flexural load (Pu) of HSMSSFRCC is 1.80
kN (lowest) for Type C3A and 3.55 kN (highest) for Type C5C. The flexural strength im-
provement ratio for ultimate crack load (FSIR-Pu) between the test and control specimens
82
are 62.50% between Type C3A and Type A3, 48.78% between Type C4C and Type A4,
and 33.96% between C5C and Type A5.
From Table 4 and Fig.3B, the ultimate crack width of 0.186 mm (highest) is noticed for
Type C3A and 0.030 mm (lowest) with remarkable performance of Type C5E. Overall, the
crackwidth has reduced when the volume fraction of hybrid mesh and fiber reinforcement
is increased. From Table 4 and Figs.6B to 6F, Figs.6H to 6L and Figs.6N to 6R, it is seen
that the number of flexural cracks for HSMSSFRCC have increased on increase in hybrid
mesh and fiber reinforcement from 2 cracks for C3A to 6 nos. for C3D, 4 cracks for C4A
to 8 cracks for C4E, 7 cracks for C5A to 12 cracks for C5E. Also, from Table 4 and Fig.4,
the average crack spacing (ACS) has reduced from 109 mm to 38 mm, 70.33 mm to 30.29
mm, and 32.50 mm to 20.00 mm between Type C3A and C3E, C4A and C4E and C5A
and C5E respectively. Lower crack spacing (between any two cracks) show higher flexural
capacity and cracking efficiency of HSMSSFRCC.
Table 3: Experimental Results of Flexural Load and Deflection of HSMPFRCC

Slab SSF- M Pcr FSIR σcr-Exp. σcr- Pred. Pu FSIR σult.- σult- Pred. σult.-Exp./
Type Vf L (Pcr) Pred. Error (Pu) Exp. Pred. Error σcr-Exp.
% No kN % (N/mm2) % kN % (N/mm2) % (Ratio)
A3 0.0 3 1.45 -- 6.96 7.18 +3.16 1.60 -- 7.68 7.71 +0.39 1.10
C3A 0.5 3 1.50 3.45 7.20 7.50 +4.17 1.80 12.50 8.64 8.48 -1.89 1.20
C3B 1.0 3 1.60 10.34 7.68 7.83 +1.95 1.85 15.62 8.88 9.25 +4.17 1.16
C3C 1.5 3 1.80 24.14 8.64 8.16 -5.88 2.05 28.12 9.84 10.02 +1.83 1.14
C3D 2.0 3 1.80 24.14 8.64 8.48 -1.89 2.30 43.75 11.04 10.80 -2.22 1.28
C3E 2.5 3 1.85 27.59 8.88 8.81 -0.79 2.60 62.50 12.48 11.57 -7.86 1.40
A4 0.0 4 1.65 -- 7.92 8.04 +1.51 2.05 -- 9.84 10.51 +6.81 1.24
C4A 0.5 4 1.70 3.03 8.16 8.36 +2.45 2.15 4.88 10.32 11.28 +9.30 1.26
C4B 1.0 4 1.80 9.09 8.64 8.69 +0.58 2.75 34.15 13.20 12.06 -9.45 1.53
C4C 1.5 4 2.15 30.30 10.32 9.02 -14.41 3.05 48.78 14.64 12.83 -14.11 1.42
C4D 2.0 4 1.85 12.12 8.88 9.35 +5.29 2.55 24.39 12.24 13.60 +11.11 1.38
C4E 2.5 4 1.90 15.15 9.12 9.67 +6.03 2.70 31.70 12.96 14.37 +10.88 1.42
A5 0.0 5 1.75 -- 8.40 8.94 +6.43 2.65 -- 12.72 13.43 +5.58 1.51
C5A 0.5 5 1.85 5.71 8.88 9.26 +4.28 3.05 15.09 14.64 14.20 -3.10 1.65
C5B 1.0 5 2.25 28.57 10.80 9.59 -12.62 3.20 20.75 15.36 14.98 -2.54 1.42
C5C 1.5 5 2.30 31.43 11.04 9.92 -11.29 3.55 33.96 17.04 15.75 -8.19 1.54
C5D 2.0 5 2.00 14.29 9.60 10.24 +6.67 3.50 32.07 16.80 16.52 -1.69 1.75
C5E 2.5 5 2.05 17.14 9.84 10.57 +7.42 3.40 28.30 16.32 17.29 +5.94 1.66
Absolute Error 3.06 4.96
SSF-Vf%-Stainless Steel Fiber Volume Fraction (%); ML-Steel Mesh Layers (3, 4, 5 nos.) with Volume Fraction, Vf=0.74%, 0.98% and 1.23% (respective-
ly); Pcr=First Crack Load; FSIR=Flexural Strength Improvement Ratio; σcr-Exp. =Experimental First Cracking Flexural Strength; σcr-Pred. =Predicted First
Cracking Flexural Strength, σult.-Exp.=Experimental Ultimate Flexural Strength; σult.-Pred.=Predicted Ultimate Flexural Strength
83
Figure 2A: First Crack Flexural Load Figure 2B: Ultimate Flexural Load
Figure 3A: First Crack Width Figure 3B: Ultimate Crack Width
Table 4. Crack Distribution and Width

Specimen SSF- ML No. of Pcr Pu TCD ACS CWcr-Exp CWu-Exp.
Type Vf Cracks
% Nos. Mm mm mm Mm
A3 0.0 3 1 1.45 1.60 200.00 200.00 0.063 0.208
C3A 0.5 4 2 1.50 1.80 109.00 109.00 0.059 0.186
C3B 1.0 5 4 1.60 1.85 186.00 62.00 0.026 0.136
C3C 1.5 3 5 1.80 2.05 180.00 45.00 0.019 0.090
C3D 2.0 4 4+2*=6 1.80 2.30 197.00 39.40 0.017 0.090
C3E 2.5 5 4+1*=5 1.85 2.60 190.00 38.00 0.013 0.065
A4 0.0 3 2 1.65 2.05 192.00 192.00 0.045 0.205
C4A 0.5 4 3+1*=4 1.70 2.15 211.00 70.33 0.036 0.184
C4B 1.0 5 4+4*=8 1.80 2.75 240.00 34.29 0.020 0.116
C4C 1.5 3 5+2*=7 2.15 3.05 198.00 33.00 0.015 0.065
C4D 2.0 4 5 1.85 2.55 160.00 40.00 0.014 0.053
C4E 2.5 5 6+2*=8 1.90 2.70 212.00 30.29 0.013 0.038

A5 0.0 3 4+1**=5 1.75 2.65 205.00 51.25 0.031 0.161
C5A 0.5 4 6+1*=7 1.85 3.05 195.00 32.50 0.030 0.136
C5B 1.0 5 6+2*=8 2.25 3.20 206.00 29.43 0.014 0.112
C5C 1.5 3 6+3*=9 2.30 3.55 215.00 26.87 0.014 0.038
C5D 2.0 4 8+3*=11 2.00 3.50 227.00 22.70 0.013 0.032
C5E 2.5 5 8+4*=12 2.05 3.40 220.00 20.00 0.013 0.030
*Multiple cracks formed at same load; TCD-Total Crack Distribution; ACS-Average Crack Spacing; ML-Steel Mesh Layers; SS-FVf -SS Fiber Volume
Fraction; Pcr-First Crack Load; Pu-Ultimate Load; CWcr-Exp.- Experimental First Crack Width; CWu-Exp.- Experimental Ultimate Crack Width
84
Higher number of cracks at increased flexural loads demonstrate higher multiple cracking
capacity of HSMSSFRC, showing the strain-hardening behavior of the composite. Type
C5E with steel mesh-5 layers and SS fiber-2.5% shows the highest number of 12 cracks
(see Fig.6R) with first crack at 2.05 kN (Crack No.1 - 1 crack), multiple cracking at vari-
ous loads, viz., 2.25 kN (Crack No.2 - 3 cracks), 2.30 kN (Crack No.3 - 1 no.), 2.50 kN
(Crack No.4 - 2 nos.), 2.55 kN (Crack No.5 - 2 nos.), 2.85 kN (Crack No.6 - 1 no.), 2.9 kN
(Crack No.7 - 1 no.) and ultimate load at 3.40 kN (Crack No.8 - 1 no.). Also, for Type
C5E, the least crack spacing between any two cracks is 20 mm, distributed over 220 mm.
Table 3 shows the ratio of experimental ultimate flexural strength to first crack flexural
strength (σult.-Exp./ σcr-Exp.) with the highest value of 1.80 for Type C5D. From Table 3 and
Fig.2A and 2B, it was seen the flexural loading at first crack and ultimate load stages drops
down beyond SS-1.5% in Type C3C (3 mesh layers + SS 1.5%), Type C4C (4 mesh layers
+ SS 1.5%), and Type C5C (5 mesh layers + SS 1.55), and therefore, the optimum SS fiber
percentage (along with 3, 4 and 5 mesh layers) is 1.5%. The average first crack deflection
was found to be less than 3 mm and average ultimate deflection to be about 15 mm; and an
average deflection ductility ratio of 5 is achieved.
Figure 4: Distributed Average Crack Spacing Figure 5: Deflection Ductility of HSMSSFRCC
Statistical Prediction Models

From the experimental data (Table 4), the predicted models are developed by conducting
statistical regression analysis (using IBM SPSS Statistics 22 Software) on first crack stress
(Model 1) and ultimate stress (Model 2) and presented in Table 5. Models 1 and 2 repre-
sent the prediction of flexural strength at first crack and ultimate stages with co-efficient of
determination (R2) values of 0.687 and 0.900 (both at p<0.01); and R2 values suggest that
the independent variables, M-Vr % (volume fraction of steel mesh reinforcement) and SS-
Vf % (volume fraction of stainless steel fibers) jointly predict the first cracking strength in
Model 1 and ultimate flexural strength in Model 2 explaining 68.7% and 90% of the varia-
tions in first crack stress and ultimate stress respectively.
The F values (overall model) of the regression equation for Models 1 & 2 were 16.426 and
67.342 (all at p<0.01) respectively. From Table 4, the absolute error for first cracking and
ultimate strength were 3.06 and 4.96 respectively showing strong correlation of experi-
mental and predicted results.
85
Paired t-test was also conducted on 18 hybrid combinations (N=18) (see Tables 2 and 4) to
check the correlation between the first cracking flexural strength and ultimate flexural
strength (Pair 1). The results in Table 6 show multiple correlation coefficient, R=0.890 (at
P<0.01) for Pair 1, demonstrating strong correlation of first cracking and ultimate stress.
Table 5: Statistical Prediction Models
Model
Predicted Models for SMRCC and HSMSSFRCC R R2 F Sig. Predicted Regression Equations
No.
1 Predicted First Cracking Strength (σcr-Pred.) (N/mm2) 0.829 0.687 16.426 P<0.01 4.517+3.592 M-Vr+0.654 SSF-Vf
2
2 Predicted Ultimate Flexural Strength (σult.-Pred.) (N/mm ) 0.949 0.900 67.342 P<0.01 -0.940+11.684 M-Vr+1.545 SSF-Vf
HSMSSFRCC-Hybrid Steel Mesh and Stainless Steel Fiber Reinforced Cementitious Composites; SS-Vf=Stainless Steel Fiber-volume
fraction (0 to 2.5%); M-Vr= Volume fraction of Steel Mesh Reinforcement (0.74, 0.98, 1.23) for 3, 4 and 5 layers
Table 6. Paired t-test
Pair No. Pairing N Correlation Sig. t df Sig. (2-

Variables tailed)
1 Experimental 18 0.890 P<0.01 7.657 17 P<0.001

First Crack
Flexural
Strength &
Ultimate
Strength
86
Figure 6A: Type A3 Figure 6G: Type A4 Figure 6M: Type A5
Figure 6B: Type C3A Figure 6H: Type C4A Figure 6N: Type C5A
Figure 6C: Type C3B Figure 6I: Type C4B Figure 6O: Type C5B
Figure 6D: Type C3C Figure 6J:Type C4C Figure 6P:Type C5C
Figure 6E: Type C3D Figure 6K: Type C4D Figure 6Q: Type C5D
87
Figure 6F: Type C3E Figure 6L: Type C4E Figure 6R: Type C5E
Fig. 6 Crack Distribution of SMRCC and HSMSSFRCC
CONCLUSIONS
Flexural tests conducted on Hybrid Steel Mesh and Stainless Steel Fiber Reinforced Ce-
mentitious Composites (HSMSSFRCC) show higher flexural strength than Steel Mesh
Reinforced Cementitious Composites (SMRCC). Multiple cracking behavior at various
flexural loads indicates the strain hardening behavior of HSMSSRCC. Since the optimum
SS fiber percentage is 1.5%, and the hybrid mix combination of 5 mesh layers and 1.5%
has shown the highest first crack and ultimate flexural loads, this study recommends the
use of this effective combination of hybrid mesh and fibers. Also, the reduction in crack
spacing and width is attributable to the stainless steel fibers added in SMRCC. The deflec-
tion ductility index is found to be about 5. Statistical Models to predict the flexural strength
and crackwidth of SMRCC and HSMSSFRCC will be quite useful to the practitioners.
ACKNOWLEDGEMENTS
P.B. Sakthivel, Ph.D. Research Scholar in Civil Engineering (Part time-External), Pondi-
cherry Engineering College (P.E.C.), Puducherry, India, and Associate Professor of Civil
Engg., Agni College of Technology, Chennai, India thank Dr. D. Govindarajalu, Principal,
P.E.C. for his technical suggestions for this project.
REFERENCES
[1] Wang, S., Naaman, A.E. and Li VC. (2004), "Bending response of hybrid ferrocement plates with
meshes and fibers", Journal of Ferrocement, 34(1), 275-288.
[2] Shannag, M.J and Ziyyad, T.B. (2007), "Flexural response of ferrocement with fibrous cementitious
matrices", Construction and Building Materials, 21, 1198-1205.
[3] Sakthivel, P.B., Ravichandran, A. and Alagumurthi, N. (2014c), "Flexural Strength and toughness of
Polyolefin Fiber Reinforced Cementitious Composites embedded with Steel Mesh", Proceedings of the
3rd RILEM International Conference on Strain Hardening Cementitious Composites (SHCC3), Delft
University of Technology, The Netherlands, 3-5th November, 441-448.
[4] Ahmed, S.F.U. and Mihashi, H. (2011), "Strain Hardening Behavior of light weight hybrid polyinyl
alcohol (PVA) fiber reinforced cement composites", Materials and Structures, 44, 1179-1191.
[5] Tosun K., Felekoglu, B. and Baradan, B. (2012), "Multiple cracking response of plasma treated polyeth-
ylene fiber reinforced cementitious composites under flexural loading, Cement & Concrete Composites,
34, 508-520.
[6] Tosun-Felekoglu, K. and Felekoglu, B. (2013), "Effects of fibre hybridization on multiple cracking
potential of cement-based composites under flexural loading", Construction and Building Materials, 4,
15-20.
[7] Lin, V.W.J, Quek, S.T. and Maalej, M. (2011), "Static and dynamic tensile behaviour of PE-fibrous
ferrocement", Magazine of Concrete Research, 2011, 63(1), 61-73.
88
[8] JCSE-SF4, Methods of tests for flexural strength and flexural toughness of steel fiber reinforced con-
crete, Concrete Library of JSCE, Japan Society of Civil Engineers, Tokyo, 1984, 3, 58-61.
[9] ASTM C78/C78M-10, Standard Test Method for flexural strength of concrete (using simple beam with
third point loading), ASTM International, 100 Bart Harbor Drive, PO Box C700, West Conshohocken,
PA 19428-2959, USA, 2010.
89
90
HYSTERESIS PARAMETER IDENTIFICATION AND RELIABILITY

ASSESSMENT OF FERROCEMENT WALLS
Felipe Uribe, Diego A. Alvarez, Jorge E. Hurtado, Bryan Chalarca and Daniel Bedoya-Ruiz,
Universidad Nacional de Colombia, Sede Manizales. 170004 Manizales, Colombia, e-mail:
{furibec,daalvarez,jehurtadog,brchalarcaec,dabedoyar}@unal.edu.co
Abstract: An essential step towards a more realistic response and reliability prediction of
structures, involves the consideration of their nonlinear behavior. This paper focuses on an
hysteresis parameter identification procedure and a reliability assessment framework for
the modeling of ferrocement structural walls. First, an experimental investigation on the
hysteretic characteristics of a ferrocement wall is presented. Next, the nonlinear hysteretic
behavior of the wall is simulated using the Bouc-Wen-Baber-Noori model, where its pa-
rameters are identified through a multi-objective optimization technique known as NSGA-
II. This methodology is able to identify the structural system showing a closer agreement
between the experimental and numerical results. Finally, the Monte Carlo and subset simu-
lation techniques are implemented in order to obtain the corresponding reliability curves of
the structural system. The research indicates that the proposed configuration of ferroce-
ment walls is able to provide large inelastic deformation capacity while ensuring a satisfac-
tory seismic performance.
INTRODUCTION
Ferrocement has emerged as an alternative material which contributes to cost-effective,

long term repair and retrofit solutions [9]. Ferrocement can be used for diverse applica-
tions, given its low weight and maintenance cost, versatility to built different kinds of
shapes, among other interesting advantages. However, effective modeling procedures on
the behavior of ferrocement wall are still required. These methods should be able to model
not only the dynamic performance of the construction but also its reliability.
When designing structural systems, engineers must cope with the evaluation of its reliabil-
ity, which plays a key role in decisions associated with cost, maintenance and safety of
structures. The most important problem in reliability engineering is the determination of
the probability of failure, i.e., the probability in which the response of a system exceeds a
predefined admissible level. In this case, the simulation of the nonlinear characteristics of
constitutive materials deserves special attention, since its consideration is an essential step
when analyzing dynamically excited structures in order to obtain a more realistic response
for reliability prediction. In those cases the hysteresis phenomenon emerges as a natural
mechanism of materials to supply restoring forces and to dissipate energy. Hence, various
methods have been proposed to represent this behavior, one of the most important method-
ologies are the Bouc-Wen (BW) type models, which have been applied in different fields
as reported in Ref. [8].
91
These and other considerations have fostered the research on the simulation of the nonline-
ar hysteretic behavior and reliability assessment of ferrocement wall structures. For these
tasks, the present paper adopts a multi-objective optimization method known as the
NSGA-II algorithm [7] for the identification of the parameters of a BW-type model used to
describe the behavior of the structure. This method simultaneously minimizes not only the
discrepancies between the estimated and real displacements, but also between the estimat-
ed dissipated energy and the one dissipated by the structure [10]. Thereafter, the fitted
model is used to represent the structural system so that the Monte Carlo and subset simula-
tion methods can be applied to obtain the associated reliability curve of the ferrocement
wall.
This document is organized as follows: Section 2, presents the experimental investigation
carried out on the ferrocement wall; Section 3, describes the nonlinear BW-type model
used to simulate the response of the ferrocement structure; Section 4, illustrates the identi-
fication process of the BW-type model parameters by means of the NSGA-II method; Sec-
tion 5, describes the methodologies used to estimate the reliability curves of the ferroce-
ment wall. The paper finishes with some conclusions and acknowledgments.
TEST SETUP, MATERIAL PROPERTIES AND LOADING HISTORY
Among its diverse applications, ferrocement has been widely used for the construction of
precast housing elements, domed structures, wall and sandwich panels [9]. This research
focuses on the later, thus, a ferrocement sandwich panel test unit (FSW-01) was construct-
ed using mortar with compressive strength MPa. It had mm height and a
x mm rectangular cross-section, which was built with a wood-cement infill in
order to reduce the wall's weight, and also to provide acoustic and thermal isolation.
The flexural reinforcement distributed along the cementitious matrix consisted of

reinforcing bars in the web region, and of reinforcing bars at the end regions of the
cross-section; all the steel bars had a yield resistance of MPa. The shear and the con-
finement reinforcement were provided by hexagonal wire mesh layers with yield re-
sistance ( ) of MPa and opening of mm; no additional horizontal reinforcing
bars were placed. Figure 1 shows the reinforcement details of the wall. The FSW-01 wall
was anchored to a rigid floor by means of a concrete foundation beam. The lateral load was
applied by a kN, mm servo-controlled hydraulic actuator, which was mounted
on a reaction frame in line with the strong direction (North-South) of the test unit and posi-
tioned mm above its footing. In order to prevent out-of-plane displacements a lateral
bracing system was used at the top of the wall. Thereafter, the displacements and force
induced by the actuator were measured using LVDTs and load cells placed at the top of the
test unit.
92
Figure 1: Schematic representation of the ferrocement wall (dimensions in mm).
THE BOUC-WEN-BABER-NOORI HYSTERETIC MODEL
In order to represent as well as possible the nonlinear characteristics of the ferrocement

wall exposed in the previous section, the Bouc-Wen-Baber-Noori (BWBN) model [4, 5]
was used due to its mathematical tractability and suitability to represent different kinds of
hysteretic patterns [8]. This model represents a single-degree-of-freedom (SDOF) system
with hysteretic restoring force and viscous damping as the differential equation:
with initial conditions, and . Here, and represent the initial ve-
locity and acceleration of the system, denotes the displacement at time , is the vis-
cous damping ratio, is the pseudo-natural frequency of the system, is the
mass of the system, is the ratio of post-yield ( ) to pre-yield elastic stiffness
( ), is the initial yield force, is the initial yield displacement, is the

external excitation, and stands for the hysteretic displacement which is modeled by
the differential equation:
93
with initial condition . The parameters , and control the shape of the hyste-
resis loops, the parameters , and are associated to the amplitude, strength
and stiffness degradation effects, and they are determined as linear functions of the hyster-
etic energy as: , , and ;
where , and are constant parameters which characterize the degradation rate. Fur-
thermore, the function which controls the pinching effect is defined as:
where, , , and
; with the constants , , , , , defining the

form and hardness of the pinching.
Finally, the hysteretic energy per unit mass representing the energy dissipated by
the hysteretic displacement is estimated as [5]:
SYSTEM IDENTIFICATION
A key concern for the implementation of BW-type models lies in the identification of a set
of parameters, i.e., from a set of input-output data, the parameters must be fitted so that the
output of the model approximates as well as possible the observed data. Thereafter, this
model is used to predict the behavior of the structural system under different kind of exci-
tations, such as earthquakes. Several methods have been developed with the aim of esti-
mating those parameters, which are based on experimental input and output data. Some of
them are for example evolutionary algorithms [6], particle swarm optimization [12],
Bayesian optimization techniques [11], among others. In this paper, a multi-objective op-
timization (MOO) approach proposed in Ref. [10] is implemented for this purpose.
The general MOO problem is stated as follows: “Minimize subjected to

and ”; where, is the vector of parameters, are the objec-
tive functions, is the set of inequality constraints, is the set of equality con-
straints and is the parameter space. It should be noted that in most cases, there are con-
flicting scenarios between the objective functions during the optimization process, there-
fore, it is not possible to characterize a single optimum solution which satisfies all func-
tions at the same time; instead several optimal solutions are identified forming a set known
as the pareto optimal front. The main idea behind the multi-objective evolutionary algo-
rithms (MOBEAs) is the preservation and diversification of good solutions in the whole
94
parameter space . The importance of preservation lies in the fact that if good solutions
are found, they must be preserved during the optimization process; and diversity is needed
in order to spread the solutions in the parameter space to completely identify the pareto
optimal front. Some of the most popular MOBEAs are the SPEA, SPEA-II [13] and the
NSGA, NSGA-II [7]. The non-dominated sorting genetic algorithm (NSGA-II), is used in
this research given its good performance in finding multiple solutions, and its versatility in
dealing with uni-modal and multi-modal, continuous and discontinuous functions. The
reader is referred to Ref. [7] for a complete explanation of the algorithm and to Ref. [10]
for its application in the identification of the BWBN model parameters.
Identification procedure
According to Ref. [10], in order to assess the quality of the identification, the set of param-
eters must minimize at the same time the following set of four objective functions
:
1. The weighted error between the real displacements and the estimated displace-
ments with the BWBN model:
where, is a weighting function used to normalize the displacements between

, this procedure is carried out so as to give the same importance to small and
large displacements.
2. The maximum error between the real displacements and the estimated dis-
placements : .
3. The difference between the total dissipated energy per unit mass calculated
from the load-displacement records, and the total dissipated energy per unit mass
estimated with the BWBN model at time for a set of parameters :
. Where, is the number of elements in the time se-
ries. The total dissipated energy per unit mass representing the area enclosed by the
hysteresis loops is the sum of the hysteretic dissipated energy per unit mass (Eq.
(1)) and the elastic dissipated energy per unit mass , i.e.,
. Here, is derived using the elastic restoring force
as:
4. The maximum error between the dissipated energy per unit mass computed from
records ( ) and the estimated dissipated energy per unit mass
( ): .
This formulation requires the restriction of the parameter space to the region
; where and denote the side
95
constraints. Thus, the parameter vector is limited to a region where the initial population
can be generated, and also determines the feasible values that they can take.
Identification results
The full set of BWBN parameters was considered for the identification process ( ),
hence the vector of parameters in this case is given by
. The NSGA-II algorithm was set
up with the following values: parent population size , offspring population size
, number of generations , crossover probability of , crossover distribution
index , mutation probability of and mutation distribution index .
The result of the identification of the BWBN parameters is reported in Table 1. Figure 2
shows the experimental and the estimated responses of the ferrocement wall with the set of
identified parameters; it can be observed that the estimated response is in good agreement
with the real one.
Table 1: Identification results for the BWBN model.
Estimated Estimated
Parameter [ , ] value Parameter [ , ] value
RELIABILITY ANALYSIS
A reliability analysis considers the basic variables that define and characterize the target
problem by means of their corresponding probability distributions. The probability of fail-
ure is expressed as the integral over the uncertain parameter space, .
Where, represents the uncertain parameters used to define the dynamic
96
model of the system; is the joint probability density function of ; and is the
failure domain in the parameter space usually determined by a limit state function
and a prescribed critical threshold level , as . General numerical
integration methods are not appropriate for the evaluation of the failure probability inte-
gral; therefore, simulation methods based on the Monte Carlo simulation (MCS) method
are usually applied. However, from a computational viewpoint the reliability problem be-
comes excessively costly when the dimension increases ( ), since the estimation of
very small probability of failure requires the exploration of a few samples in high dimen-
sional spaces [15]. Due to this shortcoming, different stochastic simulation methods have
emerged, among which are subset simulation [2], horseracing simulation [14], reliability
bounds method [1], among others.
Figure 2: Experimental and estimated response of the ferrocement wall.
Monte Carlo and subset simulation methods

One of the main problems in applying a probabilistic approach for reliability assessment
lies in both, the computation of the reliability integral and the dimensionally of the varia-
bles associated with the problem. Over the past years, the Monte Carlo simulation (MCS)
method has been the reference method to evaluate the failure probability. It estimates the
probability integral by simple statistical averaging of an indicator function which sets
whether the failure is reached or not. Thus, the failure probability estimate by MCS is
computed as,
97
The main advantage of MCS is that its performance does not rely on the dimension of the
parameter space; however, when the method is dealing with the estimation of small failure
probabilities it becomes very inefficient, since the number of samples needed to achieve an
acceptable accuracy is considerable ( ) [15]. The subset simulation (SS) method
deals with this kind of problem by converting a problem involving rare event simulation
into a sequence of problems involving more frequent events [2].
The SS method expresses the failure probability as a product of large values of conditional
probabilities by introducing several intermediate failure events. During the simulation pro-
cess, conditional samples are generated from Markov chains, which are estimated using a
modified Metropolis-Hastings (MMH) algorithm, so that they gradually populate each in-
termediate failure region until they cover the complete failure domain [3]. The essential
concept of SS is to separate the parameter space into a decreasing sequence of nested sub-
sets called intermediate failure regions, starting from the whole space and narrowing to the
failure domain , i.e., , such that
. As a result, the failure probability can be estimated as the
product of conditional probabilities,
The intermediate failure regions are represented as, ., where

form an increasing sequence of intermediate threshold val-
ues, which are chosen so that the conditional failure probability of each level corresponds
to a common specified value . Since it is difficult to know a priori which optimum in-
termediate threshold values adopt, in Ref. [3] the authors show that is a rea-
sonable choice. The reader is referred to Refs. [3, 15] for a complete explanation of the
algorithm.
RESULTS
The failure of the system is defined as the response displacement exceeding a prescribed
threshold level mm. As input for the BWBN model were used random seismic
excitations generated from the Clough-Penzien (C-P) filter as:
here, the first two parameters correspond to the characteristic ground frequency and damp-
ing respectively ( rad/s, ), the last two parameters (
98
rad/s, ) define the desired filtering of the very low frequencies;

is the Gaussian white noise input, with a sequence of
standard Gaussian variables, spectral intensity m2/s3, sampling interval
s, duration of the seismic signals , and number of uncertain param-
eters in the problem ; is a modulating function used to at-
tenuate the input noise and to give a nonstationary character to the generated seismic sig-
nals. The Amin-Ang function is used for this purpose, i.e., for
, for and for ; here,
s corresponds approximately to the arrival time of the shear waves,
s is associated with the final time of strong motion phase, and con-
trols the exponential decay that describes the fading phase of the earthquake.
The MCS method was implemented using a total number of samples of . By con-
trast, the SS algorithm was set up with the following values: Uniform distribution centered
at the current sample with width as proposal PDF; no grouping of variables was used;
probability for each level ; total number of samples for each level . The
results of the reliability analysis are shown in Figure 3, where the reliability curves of the
ferrocement wall calculated by both methods are in good agreement, except in the low
probability region where the error in the MCS estimate is significant. The SS curve was
obtained as an average of simulations, it can be appreciated the standard deviation
range of the estimator, which also demonstrates the variability of the SS method (see Ref.
[15] for further details). Moreover, for the MCS the elapsed time of the simulation was
hours, whereas for each SS was hours, which verifies the lower computational
cost of SS.
Figure 3: Failure probability estimate for the ferrocement wall.
99
CONCLUSIONS
An hysteresis parameter identification procedure and a reliability assessment framework

for the modeling of ferrocement structural walls were presented. An experimental and nu-
merical investigation on the hysteretic characteristics of a ferrocement wall were carried
out. Based on the experimental results, the nonlinear hysteretic behavior of the structure
was simulated using the Bouc-Wen-Baber-Noori model, where its parameters were identi-
fied using the NSGA-II method. The identification algorithm was able to approximate the
parameters describing the hysteresis curves of the ferrocement wall. From the fitted model,
the Monte Carlo and subset simulation techniques were implemented in order to obtain the
corresponding reliability curve of the structural system. As can be appreciated, both meth-
ods are in agreement despite the computational cost is significantly higher in the MCS
method, evidencing the great potential of the SS method. It was also noticed that the prob-
ability levels are very small for the target threshold level, which indicates that the proposed
configuration of ferrocement wall is able to ensure an appropriate seismic performance,
while providing large inelastic deformation capacity, high energy dissipation and low
strength degradation; therefore it has a great potential of being applied as a reliable low-
cost construction system.
ACKNOWLEDGEMENTS
The financial support for the realization of the present research has been received from
COLCIENCIAS under the program: ``Jóvenes Investigadores 2013'', and the Universidad
Nacional de Colombia, Grant: ``Comportamiento experimental y análisis estocástico de
viviendas de mampostería en seco postensada bajo sismicidad simulada''. The support is
gratefully acknowledged.
REFERENCES
[1] D. A. Alvarez and J. E. Hurtado. “An efficient method for the estimation of structural reliability
intervals with random sets, dependence modeling and uncertain inputs”. In: Computers & Structures
142 (2014), pp. 54–63.
[2] S.-K. Au and J. L. Beck. “Estimation of small failure probabilities in high dimensions by subset
simulation”. In: Probabilistic Engineering Mechanics 16.4 (2001), pp. 263–277.
[3] S.-K. Au and J. L. Beck. “Subset Simulation and its Application to Seismic Risk Based on Dynamic
analysis”. In: Journal of Engineering Mechanics 129.8 (2003), pp. 901–917.
[4] T. T. Baber and M. N. Noori. “Random Vibration of Degrading, Pinching Systems”. In: Journal of
Engineering Mechanics 111.8 (1985), pp. 1010–1026.
[5] T. T. Baber and Y.-K. Wen. “Random Vibration of hysteretic degrading systems”. In: Journal of
Engineering Mechanics 107.6 (1981), pp. 1069–1087.
[6] A. Charalampakis and V. Koumousis. “Identification of Bouc-Wen hysteretic systems by a hybrid
evolutionary algorithm”. In: Journal of Sound and Vibration 314.3-5 (2008), pp. 571–585.
[7] K. Deb et al. “A Fast and Elitist Multiobjective Genetic Algorithm: NSGA-II”. In: IEEE transactions
on Evolutionary Computation 6.2 (2002), pp. 182–197. ISSN: 1089-778X.
[8] M. Ismail, F. Ikhouane, and J. Rodellar. “The Hysteresis Bouc-Wen Model, a Survey”. In: Archives
of Computational Methods in Engineering 16.2 (2009), pp. 161–188.
[9] A. E. Naaman. Ferrocement & laminated cementitious composites. Ann Arbor, Michigan, USA:
Techno Press 3000, 2000.
[10] G. A. Ortiz, D. A. Alvarez, and D. Bedoya-Ruiz. “Identification of Bouc–Wen type models using
multi–objective optimization algorithms”. In: Computers & Structures 114-115 (2013), pp. 121–132.
[11] G. A. Ortiz, D. A. Alvarez, and D. Bedoya-Ruiz. “Identification of Bouc-Wen type models using the
transitional Markov chain Monte Carlo method”. In: Computers & Structures 146 (2015), pp. 252–
269.
100
[12] M. Ye and X. Wang. “Parameter estimation of the Bouc-Wen hysteresis model using particle swarm
optimization”. In: Smart Materials and Structures 16.6 (2007), pp. 2341–2349.
[13] E. Zitzler, M. Laumanns, and L. Thiele. SPEA2: Improving the Strength Pareto Evolutionary Algo-
rithm.Technical report 103. Zurich, Switzerland: Swiss Federal Institute of Technology (ETH),2001.
[14] K. M. Zuev and L. S. Katafygiotis. “Horseracing simulation algorithm for evaluation of small failure
probabilities”. In: Probabilistic Engineering Mechanics 26.2 (2001), pp. 157–164.
[15] K. M. Zuev et al. “Bayesian post-processor and other enhancements of Subset Simulation for esti-
mating failure probabilities in high dimensions”. In: Computers & Structures 92-93.0 (2012),pp.
283–293.
101
102
BEHAVIOR OF PRECAST FERROCEMENT WALLS UNDER

CYCLIC LOADING
Bryan Chalarca, Daniel Bedoya-Ruiz, Diego A. Alvarez and Jorge E. Hurtado, Universidad
Nacional de Colombia, Sede Manizales. 170004 Manizales, Colombia, e-mail:{brchalarcaec,
dabedoyar, daalvarez, jehurtadog}@unal.edu.co
Abstract: This research proposes two different configurations of precast ferrocement walls
tested under cyclic load conditions, so as to evaluate their seismic performance and poten-
tial application to the mass production of dwelling houses. The ferrocement precast walls
have been built in a sandwich-type configuration made-up of layers of welded wire mesh
or reinforced bars, hexagonal wire mesh, along with an empty core. A total of four full-
scale walls were tested under cyclic loading conditions; from those results the sufficiency
of the system was evaluated in terms of shear strength, ductility, energy dissipation, damp-
ing and crack patterns. The study shows that these novel configurations of ferrocement
walls can be applied in susceptible earthquake regions because of their high energy dissi-
pation, stiffness and low strength degradation properties.
INTRODUCTION
The western and northern parts of Latin America have suffered the effects of an intense
seismic activity. Recent earthquakes (Colombia 1995 and 1999, Peru 2007, Haiti 2010,
Chile 2011) have demonstrated the poor seismic performance of some material and struc-
tural systems used in the region [1], specially, in those urban areas where people with low
income have built their own houses following traditional techniques.
Additionally, these Latin American countries are considered as developing countries. It

means that most of them have high unemployment rates with low industrial capacities;
furthermore, there is a lack of dwelling houses. These issues represent a great opportunity
to implement novel constructions methods that allow having better living conditions, and
also, reducing unemployment rates by means of constructions companies focused in mass
production of dwelling houses.
This research has explored the usage of ferrocement precast walls, looking for a structural
system that can be implemented in mass productions programs and uses materials well
known by local builders. Four full-scale walls with two configurations of Ferrocement pre-
cast walls (FPW) were tested under severe cyclic load conditions following the directions
of the ASTM norm E 2126-11 [2], load method B. From this results were evaluated the
performance of the system under dynamics loads, shear strength, ductility, energy dissipa-
tion, damping, and crack patterns.
The plan of this article is as follows: The first section gives a brief description of each con-
figuration, and also provides characteristics of the used materials, then, in the next section,
103
some structural parameters are computed from the data obtained from the cyclic test, final-
ly, the last section gives the conclusions and other important notes.
Figure 1: Test setup.
TEST SPECIMENS
In order to develop this research, four full-scale ferrocement precast walls (FPW) were
built to test their seismic performance under severe cyclic load conditions. All the materi-
als used in this project are easily findable in urban zones of Colombia. The constructive
process was thought so as to be replicable under industrial parameters. These precast sys-
tems were composed with a set of columns as boundary elements and panels that form the
wall; those parts were built looking for the characteristics of fibrocement systems. The
panels had the properties of fibrocement with various layers of thin reinforcement in sever-
al directions, they were 25 mm thick, and, the mortar used was a mix of sand and Portland
cement without coarse aggregate. The columns were built with the same materials in a fi-
brocement-like way.
Materials
All the materials used in this project fulfill the requirements of the Colombian seismic
building code (NSR-10) [3]. Reinforced bars, welded wire mesh, hexagonal wire mesh
(See Fig. 2) and mortar were used in different configurations to create the components of
the system. These materials are well known by local builders and can be found easily, it
helps apply the structural system in industrial process of mass production.
104
Figure 2: Rebars, weldmesh, and hexagonal mesh
The used mortar obtained compressive strength of 44 MPa at the age of 28 days. The yield
strength of the reinforcing bars was 420 MPa, and the yield resistance of the welded wire
mesh (150x150 mm, 4 mm) was 485 MPa. The hexagonal mesh had an aperture of 31.75
mm.
Full-scale walls configuration

Four full-scale ferrocement precast walls with two different configurations were tested in
this research. The followed idea was to create a structural system that was easily imple-
mented on industrial process. Every wall was composed by two precast columns, between
them, two precast panels, one per face, were located using an embedding system. As
shown in Fig. 2, the system had two empty cores that allow the usage of thermal or acous-
tic isolators; also, it can be used as duct for hydro-sanitary systems or for electrical lines as
well. These walls had the following dimensions: 2100 mm height, 1000 mm width, and
130 mm thickness.
The differences between both configurations were the setup of the steel reinforcement of
panels (See Table 1). In both configurations, precast panels had eight layers of hexagonal
wire mesh surrounding the steel reinforcement.All the specimens were built and tested at
the Laboratory of Structures of the Universidad Nacional de Colombia at Manizales, Fig. 3
shows the different configurations tested.
Test setup and procedure

A lateral load was applied with a calibrated actuator, which was fixed to a steel reaction
frame. The lateral displacements were recorded using LVDTs embodied in the actuator.
Data were measured on the top of the walls and stored in a data acquisition system. In ad-
dition, photography and video equipment was used during the tests [1]. Test setup is shown
in Fig. 1.
105
Table 1: Characteristics of each configuration
Thickness Column rein-

FPW Panel reinforcement Welded wire mesh
(mm) forcement
2 bars diameter 2 bars diameter 9.53 mm 4
1 130 --
9.53 mm bars diameter 6.35 mm
2 bars diameter One layer XX-131
2 130 2 bars diameter 9.53 mm
9.53 mm rotated 45°
TEST RESULTS
Following the procedures of the ASTM E2126-11 [2], hysteresis curves of every configu-
ration were obtained from a cyclic load test (See Fig. 4). From these data several structural
parameters were calculated.
Structural parameters
In order to assess the seismic response of the ferrocement precast system, were calculated
EEEP curve [2], obtaining several structural parameters shown in Table 2.
Figure 4: Hysteresis curves: FPW-01, FPW-02.
Table 2: Structural parameters

Elastic Failure Area
Maximun Shear Yield
shear limit under Shear modulus G' Yield load Ductility
load strength displacement
FPW stiffness state envelope
Ke Ppeak ∆u νpeak 0.4 Ppeak Ppeak Pyield ∆yield µ = Δu/ Δyield
kN/mm kN mm kN.mm kN/mm kN/mm kN/mm kN mm
1 9.4821 131.7 58.41 6258.6 0.1317 20.8607 4.9605 120.1918 12.6756 4.61
2 9.2425 101 47.83 4723.3 0.101 20.3335 4.6465 85.85 9.2886 5.15
Additionally, Equivalent viscous dumping and energy dissipation were calculated for every
FPW wall. Table 3 shows the obtained results.
Energy dissipation coefficient

The energy dissipation coefficient is a concept related to structural design. This coeffi-
cient is different for every structural system, depends on the material and its capacity of
energy dissipation. Its purpose is to approximate the non-linear behavior when the struc-
106
ture response surpasses the elastic range. Table 4 shows the energy dissipation coefficient
[4] [5].
200 200 200 200 200
15
Hexagonal
0
wire mesh
Hexagonal
wire mesh
0
15
45°
Rebar XX-131
Ø 6.35 mm W eldmesh
Rebar
Ø 9.53 mm
PVC Tube
Ø 19.05 mm
Figure 3a: FPW configurations: Left: FPW-01, right: FPW-02, down: Typical reinforcement
25
25
200 200 200 200 200

80
80
H exagonal H exagonal
25
25
R ebar w ire m esh X X -131 w ire m esh

Ø 6.35 m m W eldm esh
1000
800
200 280 80 260 200
130
R ebar
P V C Tube P V C Tube
Ø 9.53 m m
Ø 19.05 m m Ø 19.05 m m
Figure 3b: FPW cross sections: Left: FPW-01, right: FPW-02, down: Typical reinforcement
FPW wall. Table 3 shows the obtained results.
107
Table 3: Equivalent viscous dumping and accumulated dissipated energy

FPW 01 FPW 02
Equivalent Acummulated Equivalent Acummulated
Drift viscous disipated Drift viscous disipated
damping energy damping energy
% % kN.mm % % kN.mm
0 - 0.000 0.00 - 0.000
0.15 16.79 149.536 0.20 10.31 70.555
0.31 13.81 401.963 0.34 9.36 327.575
0.54 11.31 1673.310 0.64 8.09 1243.188
0.80 8.41 3512.180 0.85 8.38 2574.560
1.05 8.85 6337.222 1.10 7.87 4850.119
1.24 10.46 10361.673 1.24 8.29 7808.793
1.46 10.75 16664.514 1.51 8.91 12117.115
1.65 12.26 24871.387 1.77 10.37 17404.909
1.91 13.22 35830.738 1.99 9.96 23549.818
2.11 14.13 48672.577 2.19 9.53 29312.424
2.38 14.43 63569.160 2.42 8.47 34590.599
2.61 11.68 78093.185 2.67 10.19 40288.903
2.80 13.30 91290.418 2.86 9.55 45945.225
[4] [5]. To calculate the energy dissipation coefficient was used the method proposed by
Takeda et al. This method is independent of the period of the structure and is based on syn-
thetic terrain movements and with Monte Carlo simulations.
(1)
Were is a value obtained from Monte Carlo simulations that ranges from 1.05 to 1.34;
and is the ductility. Both configurations tested in this research are accepted by the Co-
lombia building code (NSR-10) [3] as a dwelling house building system, because the ener-
gy dissipation coefficient of every configuration was above 1.5.
Table 4: Energy dissipation coefficient
FPC Ductility ε min ε max Rµ min Rµ max

1 4.61 1.05 1.34 3.009699 3.840949
2 5.15 1.05 1.34 3.201836 4.086153
108
Observed damages and crack patterns

All the configurations of FPW were tested until failure. In both configurations were ob-
served detachments of mortar in the area of contact between the columns and the panels.
The crack pattern was compounded by horizontal fissures in both columns and fissures
with an inclination of 45° in the panels, also, several cracks were spotted in the points of
contact between columns and panels. There were crumblings of mortar at both extreme
bottom parts of the columns where the rebars presented plastic hinges. FPW-01 presented a
continuous crack pattern along almost its whole surface; FPW-02 presented a continuous
crack pattern mainly at top and bottom parts.
Figure 5 Crack patterns: FPW-01, FPW-02
CONCLUSIONS
The qualities of the system to be used in process of mass production were satisfactory. Al-
so the different tested configurations had features that allow them to be used as dwelling
houses construction method under the directions of the Colombian building code (NSR-
10). Crack patterns formed under severe cyclic loadings were as expected, it was mainly
composed by fissures with fissures less than 1 mm thick, in the columns the fissures were
horizontal and along the panels they had a 45° inclination. In contact areas between col-
umns and panels there were a detachment of mortar and in both bottom extremes due to
plastic hinges, mortar was crumbled.
It is important to notice that columns presented in both configurations worked as were
thought as boundary elements. They increased the overall strength and avoided a collapse
of the system. The configuration FPW 01 showed bigger resistance, better energy dissipa-
tion and also, it had higher maximum load than FPW 02.
Comparing the final damage, the hysteresis curves and energy dissipation is possible to
confirm that FPW 01 showed better behavior under cyclic loads, most of the material
worked in order to contribute with the final resistance, and specially, with the energy dissi-
pation at large displacements. Due to the method used to obtain the energy dissipation only
takes into account the ductility of the system, FPW 02 had bigger energy dissipation coef-
ficient. This result is not the expected respect from the other parameters obtained. It is a
future task to compare these results with other methods in order to obtain a better assess-
109
ment. In general, the performance of the tested ferrocement precast configurations under
severe cyclic load was satisfactory in terms of load-lateral deflection, shear strength, ener-
gy dissipation, ductility, dumping and lack of fragile failure.
ACKNOWLEDGMENTS
COLCIENCIAS under the program: “Jóvenes Investigadores 2013”, and the Universidad
Nacional de Colombia, Grant: “Comportamiento experimental y análisis estocástico de
viviendas de mampostería en seco postensada bajo sismicidad simulada”. The support is
REFERENCES
[1] D.A. Bedoya-Ruiz, C.A. Bermudez, D.A. Alvarez, G.A. Ortiz, J.V. Escobar, “Cyclic behaviour of
prestressed precast concrete walls”, Proceedings of 15th WCEE, Lisbon, 2012.
[2] ASTM International, “Standard Test Methods for Cyclic (Reversed) Load Test for Shear Resistance
of Vertical Elements of the Lateral Force Resisting Systems for Buildings”, 2011.
[3] Asociaciación Colombiana de Ingeniería Sísmica, “Reglamento Colombiano de Construcción Sismo
Resistente, NSR-10”, 2010.
[4] Shaingchin, S., Lukkunaprasit, P., Wood, S. L., “Influence of diagonal web reinforcement on cyclic
behavior of structural walls”, Engineering Structures (2007) 29:4, 498-510.
[5] Priestley, M. J. N., “Seismic design and retrofit of bridges”, John Wiley & Sons (1996).
110
FLEXURAL BEHAVIOUR OF FERROCEMENT WITH PVA

FIBRE REINFORCED HIGH-STRENGTH MORTAR
Rasiah Sriravindrarajah and Micheal Alvaro, Centre for Built Infrastructure Research, University
of Technology Sydney, Australia, Alvaro Bros Builders, Strathfield, Australia
Abstract: Ferrocement is a composite material, consisting of mortar reinforced with con-

tinuous steel-wire mesh layers. The strength, stiffness, deflection and cracking characteris-
tics of the ferrocement composite in flexure are affected by a number of factors, namely
member thickness, span length, quantity and distribution of the wire-mesh reinforcement
and mortar quality. This paper reports the results of an experimental investigation into the
flexural behaviour of 100mm by 500mm by 19mm ferrocement slabs with four layers of
welded wire mesh reinforcement embedded in a high-strength (28-day compressive
strength of 80MPa) mortar. Uncoated PVA fibres, having either 6mm or 12mm in length,
with a fixed volume fraction of 0.25% were incorporated in the mortar matrix. The results
showed that ferrocement with high-strength mortar is a highly ductility composite material
and the addition of PVA fibres in mortar had improved the failure load of ferrocement but
significantly modified the flexural behaviour of the ferrocement by reducing the post-
failure load ductility. Although the increased in the PVA fibre length from 6mm to 12mm
had improved the failure load of the ferrocement, it had reduced the post-failure load duc-
tility.
INTRODUCTION
Ferrocement composite consists of multiple layers of closely spaced either woven or weld-
ed steel wire-mesh reinforcement embedded in mortar. Research had shown that the ferro-
cement has significant crack resistance and suitable to produce thin structural elements [1-
3]. It is widely accepted as a suitable construction material for the construction of water
storage tanks, silos, bus shelters and sunshades due to easy fabrication, using locally avail-
able materials [4]. Similarly, steel fibre reinforced mortar showed improved resistance to
cracking [5-6].
The quality of mortar matrix influences the behaviour of ferrocement through its tensile
strength, stiffness and bond strength with steel. Traditionally, the compressive strength of
mortar is about 30 to 40 MPa and with the availability of high-range water reducing ad-
mixtures, production of highly workable high-strength mortar of 100MPa or more is feasi-
ble. However, due to reduced ductility of mortar with the increase in its strength, the use of
ductile fibres in mortar matrix is considered as a solution to produce high-strength ferro-
cement with high ductility and energy absorption capacity.
Polyvinyl alcohol (PVA) fibre is considered as a suitable fibre in cementitious composites,
in place of hazardous asbestos fibres, due to its high aspect ratio, high tensile strength and
chemical durability [7]. However, its performance in PVA fibre-reinforced composites
(volume fractions of 0.4% and 1.2%), with respect to strength, energy absorption and
111
cracking, was found to be not significant when compared with steel fibre reinforced com-
posite [8]. However, research had shown that a high performance ductile fibre composite,
known as engineered cement composite (EEC), could be developed with the use of high
volume (volume fraction of 2%) PVA fibres. EEC found to have very high ductility, low
permeability and self-healing capability of cracks [9, 10]. The PVA fibres used was coated
with a proprietary hydrophobic oiling agent 1.2% by weight to tailor the interfacial proper-
ties between fibre and cement matrix for strain-hardening performance. Recent research at
the University of Technology Sydney showed that uncoated PVA fibres (volume fractions
of 0.25% and 0.50%) had improved the mechanical properties if fibre-reinforced concrete
with no significant effect on damping characteristics [11] of concrete.
In this investigation, the performance of ferrocement slabs, having the dimensions of
100mm by 500mm by 19mm, was studied under increasing flexural loading until failure.
All the test slabs had four layers of welded-wire steel mesh reinforcement, with or without
uncoated PVA fibres. The 28-day compressive strength of mortar component in the ferro-
cement was about 80MPa and the PVA fibre length was either 4mm or 8mm with a fixed
fibre volume fraction of 0.25%.
EXPERIMENTAL INVESTIGATION
Materials, mix proportion and mixing of mortar

The high-strength mortar mix (80 MPa) was produced by mixing a fixed proportions of
general purpose cement, river sand, silica fume, water and superplasticiser. The mix pro-
portion was chosen from the mixes used by Toutanji and El-Korrchi [12] in their study on
mortar. The mix proportion was 1 : 1.40 : 0.16 : 0.32 : 0.02 (cement : river sand : silica
fume : water : superplasticiser). Fine river sand, passing 2.36mm sieve and condensed sili-
ca fume (silica content over 92%) were used. Modified polycarboxilic ether based super-
plasticiser used was complying with EN 934 : Part 2, to achieve workable mortar mix.
Monofilament uncoated PVA fibres, either 6mm or 12mm in length, was used in two fibre
mortar mixes, having the volume fraction of the fibre of 0.25%. The specific gravity and
diameter of the PVA fibres were 1.29 and 0.014mm, respectively. The tensile strength of
the fibre was 1500MPa with the elongation of 7%. The modulus of elasticity of the PVA
fibre was 41.7GPa which is close to that for the mortar.
Both the control and fibre reinforced mortar mixes were produced in a power driven Ho-
bart 7-litre mixer. The required quantity of superplasticiser for the mortar mixes were pre-
mixed with water before adding to the dry mix. The mixing of mortar was continued until a
uniform mixing was achieved.
Reinforcement
Welded steel wire mesh used in the preparation of ferrocement specimens had 12mm spac-
ing and the diameter of wire was 1mm. Young’s modulus of elasticity of steel was
200GPa.
Making and curing of test specimens

For each mortar mix, three 100mm diameter by 200mm high cylinders and one 150mm
diameter by 300mm high cylinders were cast in steel moulds. A vibrating table was em-
ployed to achieve full compaction in cast test specimens. In addition, three ferrocement
slab specimens, having the dimensions of 100mm by 500mm by 19mm, were cast in
112
wooden moulds. The ferrocement slabs had four layers of welded steel wire mesh which
was cut to the required size, before placing within the mould. The bundled wire mesh rein-
forcement was placed on skeletal steel rods measuring 5mm in diameter. The moulded
specimens were demoulded after 24 hours of casting and stored in water until testing at the
age of 28 days.
Testing of mortar and cast specimens

Freshly mixed plain and fibre reinforced mortar mixes were tested for their flow in accord-
ance with ASTM C1437 Mortar Flow test for their workability. At the age of 28 days, the
small cylinders were tested in direct compression and the large cylinder was used to deter-
mine the static modulus of elasticity. The relevant testing procedures were in accordance
with the procedures described in AS1012: Testing of concrete.
Ferrocement specimens were tested in flexure over a span of 450mm in a four-point test
under controlled rate of loading until failure. The mid-span deflection was continuously
monitored using an LVDT.
Table 1: Properties of mortar mixes
Mortar Flow Compressive Modulus

Strength
(%) (GPa)
(MPa)
Plain 104 83.5 41.6
PVA fibres (6mm) 109 76.3 40.7
PAV fibres (12mm) 113 77.2 40.8
Properties of PVA fibre mortar

Table 1 shows the workability (flow), compressive strength and modulus of elasticity of
the mortar mixes, with and without PVA fibres. The flow of the mortar mix was marginal-
ly increased by the addition of 6mm and 12mm PVA fibre with the volume fraction of
0.25%. The results showed that no noticeable improvement in the workability with the
addition of PVA fibres was observed. While mixing some difficulty was experienced with
fibre mortar mixes, due to the water absorption of hydrophilic uncoated PVA fibre.
Considering the 28-day compressive strength results, it can be said that the PVA fibres had
reduced the mortar strength by 8.6% and 7.5% with the addition of 6mm and 12mm PVA
fibres, respectively. However, the reduction for the modulus of elasticity was less than 3%
even though the modulus of PVA fibre is comparable with that of mortar. The decreases
for compressive strength and modulus of elasticity could be due to increased internal po-
rosity of the mortar, as the consequence of high water absorption of the uncoated PVA
fibres.
113
Figure 1: Load-deflection of PVA (6mm) fibre mortar in flexure
Flexural behaviour of PVA fibre reinforced mortar

Figure 1 shows the mid-span deflection for the fibre reinforced high-strength mortar with
6mm PVA fibres. The maximum failure flexural load was 0.625kN. The mid-span deflec-
tion was increased with the increase in the load. The non-linear behaviour of load-
deflection curve was due the formation of multiple fine cracks. The deflection at the failure
load was about 0.4mm. When the failure load was reached, the load suddenly dropped due
to the formation of multiple large cracks. The mid-span deflection was increased even
though the load had dropped significantly.
Figure 2: Load-deflection of ferrocement with plain mortar in flexure
114
Figure 3: Load-deflection of ferrocement with fibre mortar (6 mm PVA fibres)
Flexural behaviour of ferrocement

Figure 2 shows the flexural behaviour of ferrocement with plain mortar. The maximum
failure load was 0.75kN and the mid-span deflection at this load was 4mm. The mid-span
deflection was increased with the increase in the applied load and ductile behaviour was
clearly observed, due the formation of multiple cracks. The load-carrying capacity of fer-
rocement was significantly reduced after reaching the failure load as seen from the de-
scending portion of the load-deflection curve. The mid-span deflection at 50% of the fail-
ure load was about 10mm. These results clearly showed a highly ductile behaviour of fer-
rocement with high-strength mortar.
Figure 4: Load-deflection of ferrocement with fibre mortar (12 mm PVA fibres)
115
Flexural behaviour of ferrocement with PVA fibre-reinforced mortar

Figures 3 and 4 show the flexural behaviour for the ferrocement with fibre- reinforced high
strength mortar with 6mm and 12mm length PVA fibres, respectively. The failure load for
ferrocement with 6mm long PVA fibre mortar was 0.8kN while that for with 12mm long
PVA fibres was 1.15kN. The improvement of 44% for the failure load is probably due to
the increased effectiveness of PVA fibre with increased aspect ratio.
The results also showed that that load-carrying capacity of ferrocement with PVA fibre
mortar had fluctuated noticeably before the failure load was reached. This could be due to
the formation of multiple cracking of fibre mortar which had substantially reduced the
stiffness of the ferrocement composite. However, the strong bond strength between the
steel wire-mesh reinforcement and high-strength mortar had contributed to maintain the
load-carrying capacity in the cracked ferrocement slabs. The deflection at the failure load
was about 7mm and 8mm with 6mm and 12mm long PVA fibres, respectively.
Figure 5: Cracking of ferrocement with PVA (12mm long) fibre mortar
The results also showed that once the major cracks were developed at the failure load, the
load-carrying capacity dropped suddenly similar to a brittle material. The formation of
wide crack in the ferrocement with 12mm long PVA fibre reinforced mortar is shown in
Figure 5. Therefore, it can be said that the ferrocement with PVA fibre reinforced high-
strength mortar showed improved performance up to the failure load and reduced post-
failure load ductility behaviour compared to ferrocement with plain mortar (Figure 2).
116
CONCLUSIONS
Based on the experimental investigation on the flexural behaviour of ferrocement with high
strength mortar with and without uncoated PVA fibres, the following conclusions could be
made:
1. Ferrocement with high-strength mortar is a ductile material and brittleness of high
strength mortar does not influence the ductility of ferrocement.
2. Addition of PVA fibres to the high strength mortar had significantly modified the flex-
ural behaviour of ferrocement up to the failure load by increasing the ductility with the
loss in the stiffness.
3. PVA fibres in high strength mortar caused the ferrocement to reduce its ductility beyond
the failure load.
4. Increasing the PVA length from 6mm to 12mm had marginally increased the failure
load of ferrocement but reduced the post-failure load ductility more significantly.
REFERENCE
[1] Sri Ravindrarajah, R. and Tam, C. T., ‘Dimensional stability of ferrocement’, J. of Ferrocement 13 (1)
(1983) 1-12.
[2] Sri Ravindrarajah, R. and Tam, C. T., ‘Watertightness in ferrocement’, J. of Ferrocement 14 (1)
(1984) 11-20.
[3] Paramasivam, P. and Sri Ravindrarajah, R., ‘Effect of arrangement of reinforcement on ferrocement
properties’ ACI Structural J. 85 (1988) 3-11.
[4] Lee, S. L., Tam, C. T., Paramasivam P., Das Gupta, N. C., Sri Ravindrarajah, R. and Mansur, M. A.,
‘Ideas tested on Ferrocement applications at the University of Singapore’, Concrete International: De-
sign & Construction 5 (11) (1983) 12-16.
[5] Sri Ravindrarajah, R., ‘Behaviour of fibre-reinforced thin slabs under short and long-term loading’,
Proceedings of an International Symposium and Workshop on Ferrocement and Thin Reinforced
Cement Composite, (FERRO-8), Bangkok, Thailand, February, 2006
[6] Sri Ravindrarajah, R., ‘Fibre-reinforced and Ferrocement Car-park Pavers’, Proceedings of the X
International Symposium on Ferrocement and Thin Reinforced Cement Composites (FERRO 10),
Havana, Cuba, October, 2012.
[7] Akers, S. A. S., Studinka, J. B., Meier, P., Dobb, M. G., Johnston, D. J. and Hikasa, J., ‘Long term
durability of PVA reinforcing fibres in a cement matrix’, The International J. of Cement Composites
and Lightweight Concrete 11 (2) (1989) 79-91.
[8] Kim, D. J., Naaman, A. E. and El-Tawil, S., ‘Comparative flexural behaviour of four fibre reinforced
cementitious composites’ Cement & Concrete Composites 30 (2008) 917–928
[9] Li V.C. ‘On engineered cementitious composites (ECC) - a review of the material
and its applications’. J. Advanced Concrete Technology 1(3) (2013) 215–30.
[10] Lepech M. D. and Li, V. C., ‘Water permeability of engineered cementitious composites’
Cement & Concrete Composites 31 (2009) 744-753.
[11] Noushini, A., Samali, B., and Vessalas, K., ‘Effect of polyvinyl alcohol (PVA) fibre on dynamic and
materials properties of fibre reinforced concrete’ Construction and Building Materials 49 (2013) 374-
383.
[12] Toutanji. H. A. and El-Korchi. T., ‘The influence of silica fume on the compressive strength of ce-
ment paste and mortar’, Cement and Concrete Research, 25(7) (1995) 1591-1602.
117
118
STRENGTH ASSESSMENT OF SANDWICH-TYPE FERROCEMENT

STRUCTURAL WALLS UNDER CYCLIC LOADING
Bryan Chalarca, Daniel Bedoya-Ruiz, Felipe Uribe, Diego A. Alvarez and Jorge E. Hurtado,
Universidad Nacional de Colombia, Sede Manizales. 170004 Manizales, Colombia, e-
mail:{brchalarcaec, dabedoyar, furibec, daalvarez, jehurtadog}@unal.edu.co
Abstract: This study introduces several configurations of ferrocement walls tested under
severe cyclic load conditions, in order to assess their seismic features to be used in the con-
struction of dwelling houses. The ferrocement walls have been built in a sandwich-type
configuration composed of layers of hexagonal wire mesh, welded wire mesh and reinforc-
ing bars, along with a wooden-cement core. A total of eight full-scale walls were tested
under severe cyclic loading conditions; from those results the capacity of the system was
evaluated in terms of shear strength, ductility, energy dissipation, damping and crack pat-
terns. The research evidences that these novel configurations of ferrocement walls have a
great potential of being applied as a low-cost construction system on earthquake prone
regions due to the high energy dissipation, stiffness and low strength degradation proper-
ties, presented at large displacement rates.
INTRODUCTION
Colombia is a developing country with two major issues that give this research a wide ap-
plication field. First of all, Colombia has several regions with high seismic activity in
which the most part of population is located. In addition, there is a lack of proper dwelling
houses for low income population located in new urban areas that grew in a very disor-
ganized way. The constructions techniques in those places have very poor seismic perfor-
mance. Its materials are not exploited in all its features and do not provide good conditions
to live on.
The alternatives offered by ferrocement to solve these issues are highly exploitable [1],
because its materials are widely known by the informal builders, and are also common in
local stores. Ferrocement has excellent seismic performance with low cost. Additionally,
the sandwich-type configuration of this research used an agro-industry waste; this waste
currently does not have a clear use, so, most of the times it just is dumped in a landfill.
119
Figure 1 Test setup.
This research used eight full-scale walls with four configurations of reinforcing bars, hex-
agonal wire mesh, welded wire mesh and a wooden-cement core. This core works as a
weight reducer and as a thermal and acoustic insulator as well. These walls were tested
under cyclic load conditions following the directions of the ASTM norm E 2126-11 [2],
load method B. From these results were evaluated the performance of the system under
dynamics loads, shear strength, ductility, energy dissipation, damping, and crack patterns.
The plan of this article is as follows: The first section gives a brief description of each con-
figuration, and also provides characteristics of the used materials, then, in the next section,
some structural parameters are computed from the data obtained from the cyclic test, final-
ly, the last section gives the conclusions and other important notes.
TEST SPECIMENS
Several characteristics were taken into account to develop this research. The sort of materi-
al, the final cost of the construction and the capacity of the Structures laboratory were the
main features to select the configurations and the geometry of the specimens.
120
Materials
The materials used in this project are easily findable in any urban area of Colombia. Also,
the traditional construction workers are familiarized with them, therefore, implement ferro-
cement houses and teaching this system to these communities will be a simple process.
Reinforcing bars, hexagonal wire mesh, welded wire mesh (See Fig. 2); concrete and a
wooden-cement core were used as part of different configurations tested in this research. It
is important to note that all the materials used during the experimental campaign fulfill the
Colombian seismic building code (NSR-10) [3]. The used mortar obtained compressive
strength of 44 MPa at the age of 28 days. The yield strength of the reinforcing bars was
420 MPa, and the yield resistance of the welded wire mesh (150x150 mm, 4 mm) was 485
MPa. The hexagonal wire mesh had an aperture of 31.75 mm
Figure 2 Rebars, weldmesh, and hexagonal wire mesh
Full-scale walls configuration

The idea of this research was to test different configurations of rebar and welded wire
mesh along with a base of mortar, hexagonal wire mesh and wooden cement core.
These walls had the following dimensions: 2100 mm height, 1000 mm width, thickness
ranges from 120 mm until 145 mm depending on the configuration. In Table 1 is possible
to compare every configuration. The thickness of the wooden-cement core is 80 mm, sur-
rounding it there were four layers of hexagonal wire mesh, then, the different configura-
tions of rebars and weldmesh were set, and finally, there were another four layers of hex-
agonal wire mesh. All the specimens were built and tested at the Laboratory of Structures
of the Universidad Nacional de Colombia at Manizales, Fig. 3 shows the different configu-
rations tested.
121
Test setup and procedure

A lateral load was applied with a calibrated actuator, which was fixed to a steel reaction
frame. The lateral displacements were recorded using LVDTs embodied in the actuator.
Data were measured on the top of the walls and stored in a data acquisition system. In ad-
dition, photography and video equipment was used during the tests [4]. Test setup is shown
in Fig. 1.
Table 1: Characteristics of each configuration
FSW (Wall) Thickness (mm) Rebars Welded wire mesh
4 bars diameter 9.53 mm 8

01 120 --
bars diameter 6.35 mm
One layer XX-131
02 140 --
surrounding the core
4 bars diameter 9.53 mm 8 One layer XX-131

03 140
bars diameter 6.35 mm surrounding the core
One layer XX-131

4 bars diameter 9.53 mm 8
04 145 rotated 45° sur-
bars diameter 6.35 mm
rounding the core
TEST RESULTS
Fig. 5 shows the resulting hysteresis curves of every configuration, these curves were ob-
tained from the cyclic load test. Several structural parameters were calculated from this
data following the procedures of the ASTM E 2126-11 [2]
122
80
940 940
80
100 250 270 250 100 120 150 140
150
Ø 2/8 XX-131
Ø 3/8
940 940
80
80
100 250 270 250 100 140 100 250 270 250 100 145
0
15
Ø 2/8 Ø 2/8
15
0
Ø 3/8 Ø 3/8
XX-131 XX-131
150
45°
150
Figure 3 FSW configurations (Left to Right): FSW-01, FSW-02, FSW-03, FSW-04
123
30 30 30 30
120
140
940 940
80
80
Hexagonal Hexagonal
wire mesh wire mesh
Rebar XX-131
Rebar Ø 6.35 mm Rebar Weldmesh
Ø 9.53 mm Ø 9.53 mm
30 30 30 30
140
145
940 940
80
80
Hexagonal Hexagonal
wire mesh wire mesh
Rebar Rebar
Rebar Ø 6.35 mm Rebar Rebar Ø 6.35 mm Rebar
Ø 9.53 mm Ø 9.53 mm Ø 9.53 mm Ø 9.53 mm
XX-131 XX-131
Weldmesh 0° Weldmesh 45°
Figure 4b FSW cross sections (Left to Right): FSW-01, FSW-02, FSW-03, FSW-04
Figure 5 Hysteresis curves (Left to Right): FSW-01, FSW-02, FSW-03, FSW-04
Following the procedure of the ASTM E2126-11 norm [2], were calculated the envelope
curve and the EEEP curve. From these data, were evaluated several structural parameters
showed in Table 2.
124
Table 2: Structural parameters

Elastic Failure Area Yield
Maximun Shear Yield
shear limit under Shear modulus G' displacemen Ductility
load strength load
stiffness state envelope t
FSW
0.4 µ = Δu/
Ke Ppeak ∆u νpeak Ppeak Pyield ∆yield
Ppeak Δyield
kN/mm kN mm kN.mm kN/mm kN/mm kN/mm kN mm
01 9.0787 135.2 33.9 4774.2 0.1352 19.0654 8.3752 114.92 12.6581 2.68
02 11.7601 90.4 14.03 1125.9 0.0904 24.6962 13.531 76.84 6.534 2.15
03 7.55 160.4 34.39 5709.6 0.1604 15.84 9.79 136.34 18.07 1.90
04 10.32 153.8 29.09 5091.4 0.1538 21.67 11.1 130.73 12.67 2.30
FSW wall. Table 3 shows the obtained results.
125
Table 3: Equivalent viscous dumping and accumulated dissipated energy

FSW-01 FSW-02
Drift viscous disspated Drift viscous disspated
% % kN.mm % % kN.mm
0 - 0 0 - 0
0.29 7.380 155.943 0.22 7.093 93.855
0.52 7.909 1331.263 0.44 10.682 564.201
0.75 9.038 3252.322 0.68 7.866 1761.817
0.96 10.915 6822.711 0.93 16.864 3584.914
1.21 13.756 13054.753 1.14 7.683 4228.702
1.42 15.888 21179.489 - - -
1.68 12.846 27400.234 - - -
1.91 10.598 32839.444 - - -
2.15 10.491 38599.994 - - -
2.40 10.451 43837.279 - - -
2.63 9.886 48501.999 - - -
2.86 8.998 53187.668 - - -
3.09 9.571 57784.934 - - -
FSW-03 FSW-04

Drift viscous disspated Drift viscous disspated
% % kN.mm % % kN.mm
0 - 0 0 - 0
0.21 5.750 48.206 0.22 8.070 115.073
0.44 5.630 481.479 0.41 8.830 673.130
0.66 6.570 1615.918 0.66 8.340 2200.729
0.88 5.960 3701.861 0.87 7.770 4793.448
1.09 5.350 6598.504 1.06 13.130 10678.789
1.34 11.480 13001.292 1.32 17.170 19585.225
1.56 14.390 18450.315 1.54 15.730 29066.284
1.81 14.950 32498.116 1.8 13.430 37557.728
2.03 12.350 41616.517 2.02 14.130 45827.502
2.27 11.970 50175.610 2.25 11.870 51826.271
2.47 10.410 56590.687 2.47 10.860 57412.933
2.71 12.120 62383.340 - - -
2.94 11.730 65398.730 - - -

[5] [6].
126
To calculate the energy dissipation coefficient was used the method proposed by Takeda et
al. This method is independent of the period of the structure and is based on synthetic ter-
rain movements and with Monte Carlo simulations.
(1)
Were is a value obtained from Monte Carlo simulations that ranges from 1.05 to 1.34;
and is the ductulity.
Table 4: Energy dissipation coefficient
FSW Ductility ε min ε max Rµ min Rµ max

01 2.678127049 1.05 1.34 2.191522 2.7968
02 2.147229875 1.05 1.34 1.905818 2.432187
03 1.9031544 1.05 1.34 1.758964 2.244774
04 2.295974743 1.05 1.34 1.990006 2.539627
All the configurations tested in this research are accepted by the Colombia building code
(NSR-10) [3] as a dwelling house building system, because the energy dissipation coeffi-
cient of every configuration are above 1.5.
Observed damages and crack patterns

All the configurations were tested until failure. They presented crack pattern from it middle
part until the bottom part. Along all the wall face, there were fissures less than 1 mm thick,
they started horizontal but at the center had an inclination of 45° to the bottom. In FSW 1,
3 and 4, at the base at both extremes, there were a crumbling of the mortar and the rebars
there formed plastic hinges. FSW 2 did not present crumbling of the mortar at the ex-
tremes, the weldmesh failed and the wall was torn from the foundation beam. Fig. 5 shows
the crack patterns of all the configurations [7].
CONCLUSIONS
The construction of eight full-scale walls allowed observing and evaluating the versatility,
speed and cleanliness of the construction process; its potential to be used in urban areas
with a lack of dwelling houses was observed as well. The different tested configurations
had features that allow them to be used as dwelling houses construction method under the
directions of the Colombian building code (NSR-10) [3].
Under severe cyclic load test, specimens showed a characteristic failure mode. It was
formed by crack pattern with a millimeter of thick or even less, the fissures started at the
edge horizontal and tended to have a 45° inclination near the center of the wall face [7].
The missing of reinforced bars in FSW 02 did not allowed the formation of plastic hinged,
hence, this lack reduced its capacity at displacements beyond 15 mm. In comparison, the
other configurations resisted bigger load after 15 mm of displacement. The lack of plastic
127
hinges in FSW 02 did not allowed good energy dissipation. Comparing the energy dissipa-
tion coefficients and the respective hysteretic curves, it is possible to notice that after 15
mm of displacement this configuration did not provide enough resistance.
The density of the wooden-cement core was three times fewer than concrete; it means a
56% fewer weight and a 46% fewer volume of concrete inside the wall without a detriment
of the stiffness, global strength and ductility. Additionally, the wooden-cement provides
the system with great thermo acoustic insulator properties [7]. In general, the performance
of the tested ferrocement configurations under severe cyclic load was satisfactory in terms
of load-lateral deflection, shear strength, energy dissipation, ductility, dumping and lack of
fragile failure [7].
Figure 5 Crack patterns (Left to Right): FSW-01, FSW-02, FSW-03, FSW-04
ACKNOWLEDGMENTS
COLCIENCIAS under the program: “Jóvenes Investigadores 2013”, and the Universidad
Nacional de Colombia, Grant: “Comportamiento experimental y análisis estocástico de
viviendas de mampostería en seco postensada bajo sismicidad simulada”. The support is
REFERENCES
[1] A. E. Naaman. Ferrocement & laminated cementitious composites. Ann Arbor, Michigan, USA:
Techno Press 3000, 2000.
128
[2] ASTM International, “Standard Test Methods for Cyclic (Reversed) Load Test for Shear Resistance
of Vertical Elements of the Lateral Force Resisting Systems for Buildings”, 2011.
[3] Asociaciación Colombiana de Ingeniería Sísmica, “Reglamento Colombiano de Construcción Sismo
Resistente, NSR- 10”, 2010.
[4] D.A. Bedoya-Ruiz, C.A. Bermudez, D.A. Alvarez, G.A. Ortiz, J.V. Escobar, “Cyclic behaviour of
prestressed precast concrete walls”, Proceedings of 15th WCEE, Lisbon, 2012.
[5] Shaingchin, S., Lukkunaprasit, P., Wood, S. L., “Influence of diagonal web reinforcement on cyclic
behavior of structural walls”, Engineering Structures (2007) 29:4, 498-510.
[6] Priestley, M. J. N., “Seismic design and retrofit of bridges”, John Wiley & Sons (1996).
[7] D.A. Bedoya-Ruiz., “Estudio y comportamiento sísmico de estructuras de materiales económicos de
construcción, tales como el ferrocemento y otros materiales, por medio del análisis experimental y el
análisis numérico determinista y estocástico”, Final report of research, Universidad Nacional de
Colombia, March, 2012.
129
130
CHAPTER 3
MODELLING
131
132
NONLINEAR ANALYSIS OF TEXTILE REINFORCED CONCRETE

SHELLS USING AN ANISOTROPIC, DAMAGED-BASED
MATERIAL MODEL
Ehsan Sharei (1), Alexander Scholzen (1), Rostislav Chudoba (1), (1) Institute of Structural
Abstract: Thin shell structures made of Textile Reinforced Concrete (TRC) exhibit a phys-
ically nonlinear response due to the strain-hardening behavior of the composite. For the
simulation of thin walled TRC structures, an anisotropic damage-based material model of a
microplane type has been used to reflect the propagation of fine, oriented cracks during the
loading. In this paper a method for calibration and validation of a TRC cross section is
proposed and demonstrated. After providing a brief summary of key assumptions of the
material behavior, the calibration procedure using tensile test data is presented. Then the
model is validated by simulation of a test on a TRC vault shell. The comparison is per-
formed both for the load-displacement response and for the observed and calculated distri-
butions of cracks. Finally, the feasibility and validity of the computational model is dis-
cussed.
INTRODUCTION
Textile Reinforced Concrete (TRC) can fulfill the demands of innovative structural designs
by offering a high load bearing capacity and durability. A TRC cross section consists of
fine-grained concrete layers and reinforcing layers of carbon or glass fabrics. Due to the
non-corrosive reinforcement thin cross sections with a thickness of only a few centimeters
can be realized. The textile reinforcement has a high form flexibility that enables also the
construction of lightweight shell structures with curved shapes [1,2]. Its material
characteristics make the composite material also highly attractive for various other fields of
application such as ventilated facade elements or sandwich panels with TRC facings [3].
TRC as an inhomogeneous composite material exhibits an anisotropic crack formation
which depends on the direction of the principle stress in a shell structure. Except of this
damage-induced anisotropy, initial anisotropy due to the orientation of the fabrics can be
observed as well. The tensile stress-strain response of the composite material exhibits a
pronounced strain-hardening effect. In order to cover the anisotropic and nonlinear
behavior of TRC, an anisotropic damage model of microplane type with smeared
representation of finely distributed matrix cracks has been used in this paper. The model is
based on the microplane damage model described in [4] and [5] that has been utilized for
the application to TRC shells [6].
Figure 1 shows a procedure of calibration and validation of the model. On the left hand
side a typical damage function is shown. This function determines the evolution of damage
based on the direction of loading and is identified by an iterative calibration procedure
which is described in [7]. Using this function, the experimental stress-strain curve from a
133
tensile test can be fully reproduced in the numerical simulation. The model has been
validated by comparing the prediction with the TRC bending tests and slab tests with the
same cross-sectional thickness and reinforcement ratio (Figure 1, right). In the present
paper the ability of the anisotropic damage model to reproduce complex stress states within
TRC shell structures is investigated using a large scale test of a TRC barrel-vault shell.
The material model is implemented as a user subroutine for 4-node bilinear isoparametric
finite shell elements with reduced integration in the commercial finite element software
ABAQUS [8]. This implementation corresponds to a former implementation of the model
for solid shell elements [9].
Figure 1: Schematic description of the procedure for calibration and validation of the nonlinear
computational model
The aim of this paper is to demonstrate the feasibility of the model in simulation of
complex TRC shell structures. The capability of the model is discussed in terms of its
ability to predict the crack formation and to estimate the load bearing capacity of the shell
structure.
CALIBRATION OF DAMAGE FUNCTION USING A TRC TENSILE TEST

In order to calibrate the material model for TRC, a series of uniaxial tensile tests has been
performed. Figure 2 right shows the TRC specimen with six layers of textile reinforcement
made of carbon bonded with the fine-grained concrete matrix. Dimensions of the specimen
are shown in Figure 2 left. Strain is measured by displacement gauges on both side of the
specimen with a measuring length of 25 cm.
134
Figure 2: Geometry of tensile test specimen and position of the displacement gauges (left); cross-sectional
layout with six layers of carbon textile fabrics (right)
Figure 3: Damage function calibrated for the tensile test of TRC specimen (left) and stress-strain curves of
tensile test corresponding to experiment and simulation (right)
Figure 3 (left) shows the calibrated damage function with elasticity modulus
E = 18709 MPa and Poisson’s ratio ν = 0.2. The simulation of the tensile test shown in
Figure 3 (right) performed with the calibrated damage function, showing the match
between the stress-strain curves from the test and from the simulation.
VALIDATION OF THE MODEL USING A TRC VAULT SHELL

The TRC barrel-vault shell shown in Figure 4 was fabricated with the cross-sectional
layout corresponding to Figure 2 (right)
135
Figure 4: Dimensions of the tested and simulated TRC barrel-vault shell
The thickness of the shell was 2 cm containing 6 layers of equidistantly placed fabric
layers of carbon. The shell was placed on four steel supports providing a radial and
tangential constraint at each support.
This structure was tested under a radial loading applied by means of a steel stripe which
was pulled using a hydraulic cylinder on each side as shown in Figure 5. Two different
types of supports have been used during a loading-unloading-reloading scenario. At the
first loading step up to a force of F = 60 kN in the hydraulic cylinders, the radial
displacement was allowed at the supports. After that, the shell was unloaded and the radial
displacements were fixed by means of screwed bolts in order to prevent uplifting of the
shell during the test. With the fixed radial displacements the shell was reloaded up to the
failure load (Figure 6).
136
Figure 5: Test setup for the measurement of the load bearing capacity of the investigated TRC barrel-vault
shell (up), tested prototype in the laboratory (down left) and detail of load introduction via hydraulic
cylinders (down right)
The chosen loading scenario led to finely distributed crack pattern throughout large zones
of the shell as intended by the design of the test setup. The highest tensile stresses
developed along the shell edges in the longitudinal direction where also the ultimate failure
of the shell occurred. The ultimate breaking strain of the composite in these regions was
measured using two displacement gauges as shown in Figure 5 (top).
The numerically obtained damage of the shell at the ultimate load is shown in Figure 6. In
agreement to the experiment, largest values of the damage indicator occur in the middle of
the longitudinal shell edges. Moreover, due to the chosen loading and boundary conditions
the structure was also exhibiting damage zones at the front and at the back side of the
structure. Also this result agrees well with the observation in the test and documents the
ability of the model to reproduce the matrix fragmentation process within a complex
structural geometry and boundary conditions.
137
Figure 6: Damage contour in the TRC vault shell under load F and the loading-unloading path with the
corresponding change in the support conditions
Figure 7: Load-tensile strain curve from the simulation of the TRC barrel-vault shell compared to the
experimental results for the left and right side of the shell (unloading branch is not shown)
A qualitative comparison between test and simulation is provided in Figure 7 showing the
load F and the corresponding strain measured by the two displacement gauges at the
longitudinal edges of the shell is plotted in Figure 7. The applied loading history is
sketched in the diagram in Figure 6 containing the loading, unloading and reloading
branches. The calculated prediction slightly underestimates the measured response as
shown in Figure 7. Obviously, the test results are not perfectly symmetric. Further
improvements and detailing of the numerical model including the sensitivity of the
response with respect to the slight perturbations with respect to the geometry will be
138
included in the next phase of analysis. Yet, the shown results demonstrate the ability of the
model to capture the trend of the strong nonlinear structural response for the described
loading-unloading-reloading scenario.
SUMMARY
In this paper a calibration of an anisotropic damage model for the simulation of TRC shell
structures with a given material properties, cross-section and textile fabric reinforcement is
presented. This model is based on microplane damage model and cares for the anisotropy
and physical nonlinearity of textile reinforced concrete. After the calibration step, model
was validated by simulation of a test on a TRC vault shell. The model has shown a good
suitability in representation of fine and discrete cracks. Results from the simulation show
also a good correspondence to the experimental results in the estimation of load bearing
capacity.
ACKNOWLEDGMENT
REFERENCES
1. A. Scholzen, R. Chudoba, J. Hegger. Thin-walled shell structure made of textile reinforced con-
crete; Part I: structural design and construction. Structural Concrete, 16(1), 2015, DOI:
10.1002/suco.201300071 (in press).
2. A. Scholzen, R. Chudoba, J. Hegger. Thin-walled shell structure made of textile reinforced con-
crete; Part II: experimental characterization, ultimate limit state assessment and numerical simula-
tion. Structural Concrete, 16(1), 2015, DOI: 10.1002/suco.201400046 (in press)
3. A. Shams, M. Horstmann, J. Hegger. Experimental investigations on textile-reinforced concrete
(TRC) sandwich sections. Composite Structures, Vol. 118, pp.643-653, December 2014.
4. I. Carol and Z. P. Bažant. Damage and plasticity in microplane theory. International Journal of
Solids and Structures, 34(29), pp.3807 - 3835, 1997.
5. M. Jirásek. Comments on microplane theory. Mechanics of Quasibrittle Materials and Structures,
Hermes Science Publications, pages 55-77, 1999.
6. E. Sharei, A. Scholzen, R. Chudoba and J. Hegger. Anisotropic Damage Model for the Numerical
Simulation of Textile Reinforced Concrete Shell Structures", The Twelfth International Confer-
ence on Computational Structures Technology, B.H.V. Topping and P. Iványi, (Editors), Civil-
Comp Press, Stirlingshire, United Kingdom, paper 9, 2014.
7. R. Chudoba and A. Scholzen. Modeling of reinforced cementitious composites using the micro-
plane damage model in combination with the stochastic cracking theory. Proceedings of EURO-C
2010, Rohrmoos/Schladming, Austria, 15.-18. March 2010.
8. ABAQUS User Subroutines Reference Manual, Version 6.11, Dassault Systèmes Simulia Corp.,
Providence, RI, USA, 2011.
9. A. Scholzen, R. Chudoba, J. Hegger, “Calibration and validation of a Microplane damage model
for cement-based composite applied to Textile Reinforced Concrete”, International Conference
on Recent Advances in Nonlinear Models - Structural Concrete Applications, H. Barros, R. Faria
and C. Ferreira (Editors), pp.417-428, 2011.
139
140
ANALYSIS OF THE TENSILE RESPONSE OF TEXTILE

REINFORCED CONCRETE USING DIGITAL IMAGE
CORRELATION TECHNIQUE COMBINED WITH MULTI-SCALE
STOCHASTIC MODELLING
Authors: Yingxiong Li1, Rostislav Chudoba1, Václav Sadílek2, Rostislav Rypl2, Miroslav
Vořechovský2, 1Institute for Structural Concrete, RWTH Aachen University, Germany, 2Institute
of Structural Mechanics, Brno University of Technology, Czech Republic
Abstract: Textile reinforced concrete (TRC) components exhibit multiple cracking leading
to strain hardening behavior under uniaxial tension. In this paper, the method for character-
ization of the tensile response of TRC capturing the interaction between the matrix crack-
ing and local bond behavior between the matrix and textile fabrics is presented. By com-
bining advanced experimental and modelling approaches the tensile response of TRC spec-
imens can be reproduced in terms of elementary damage mechanisms in the heterogeneous
material structure. In particular, the data obtained by digital image correlation technique is
used for automated detection of crack initiations in the tested zone of the specimen during
the loading. The data describing the history of cracking including the quantitative charac-
teristics (positions and crack widths) serve as an input for the numerical simulation of the
fragmentation process using a multi-scale stochastic cracking model. The micro-scale ef-
fects occurring during the debonding process in the vicinity of a crack are described using
the representative crack bridge model. As a result, the crack bridge behavior is calibrated
in the configuration relevant for the behavior of structural components. Validation of the
model has been performed by comparing the model prediction with test results obtained for
specimens with modified reinforcement ratio.
INTRODUCTION
It is a common practice in civil engineering to combine plain concrete with reinforcement

with high tensile strength to achieve better performance under tensile or bending load.
Components consisting of such composite materials usually exhibit multiple cracking and
strain hardening behavior under uniaxial tension. The debonding of the reinforcements in
TRC and the multiple cracking of the matrix leads to its quasi-ductile tensile response.
Aveston et al. [1] and Aveston and Kelly [2] proposed the well-known ACK model to in-
vestigate the behavior of composites under tension. However, the stochastic nature of the
matrix strength was not considered in the ACK model. Curtin et al. [3] adopted the
Weibull distribution to describe the initial flaws in the matrix such that the influence of
matrix strength on multiple cracking was implicitly considered. In this paper, a numerical
model which explicitly addresses the issue of stochastic matrix strength along the whole
specimen is proposed, a one-dimensional stationary random field is employed to represent
matrix strength field. By assuming the periodicity of stress and strain fields over the whole
specimen length, the model is formulated in such a way that it can be flexibly combined
with a variety of local crack bridge models. In order to validate the numerical model, uni-
141
axial tensile tests of thin TRC specimens were carried out. Using digital image correlation
techniques, the displacement fields in the composite specimens were accurately captured.
The performed tests provided the data on the crack positions and corresponding crack initi-
ating load levels that served for calibration and validation of the numerical model. The
present paper is organized as follows, firstly the numerical model for tracing the crack pat-
tern is presented in section 2, and then the crack detection scheme using ARAMIS is intro-
duced in section 3. The validation of the model is given in section 4 and finally in section 5
major conclusions are drawn and outlooks are given.
NUMERICAL CRACK TRACING ALGORITHM
In the formulation of the semi-analytical model we assume, that in a multiply cracked

composite specimen, the stress and strain fields are symmetrical between two adjacent
cracks. Thus, the reinforcement can be regarded as clamped at the middle of each segment.
A crack bridge is the part of specimen between two symmetry axes, which is shown in
Fig.1(b). A variety of crack bridge models assuming different bond laws have been estab-
lished [4]. Given the external load level and boundary conditions, those models are able to
return the matrix stress profile and reinforcement strain profile, namely
σ m ( z ) = Σm (λ ; z , L↑ , L↓ )
(1)
ε f ( z ) = ε f (λ ; z , L↑ , L↓ ),
where z is the local coordinate which is defined as the distance from the crack position, λ
is composite stress which is defined as the external load level divided by the cross section-
al area, L↑ and L↓ are the boundary conditions, Σ m and ε f are the functions given by the
crack bridge model to evaluate the matrix stress field and reinforcement strain field, re-
spectively. As an example, the strain profiles obtained with crack bridge model assuming
constant bond stress in the bond interface are shown in Fig.1(c).
As shown in Fig. 1(a), a multiply cracked specimen can be regarded as a series of crack
bridges. Then, the response of a tensile specimen can be modeled by repeated usage of the
crack bridge model Eq. (1) in order to evaluate the stress and strain field along the speci-
men. Let us consider an intermediate state of loaded specimen covering the domain
x ∈ [0, L ] with a set of k = 1...K existing cracks at the positions X k( K ) , in order to provide
a flexible link between the domain of the tensile specimen with the crack bridge model we
introduce an index function identifying the crack κ ∈ {1, 2,..., K } with the shortest dis-
tance to a material point x
142
Fig. 1: (a) The multiple cracked specimen; (b) A crack bridge; (c) Matrix and reinforcement strain profiles
of the crack bridge.
κ ( x) = arg min x − X k( K ) . (2)

k =1... K
Then, a function returning the signed distance from a point x to its closest crack can be
written as
z ( K ) ( x) = x − X κ( K( x)) . (3)
In analogy, the boundary conditions L↑ and L↓ defining the domain of the crack bridge
can be formulated as functions of x , namely L(↑K ) ( x) and L(↓K ) ( x) , respectively. Substitut-
ing those three functions into Eq. (1), the strain profile of the reinforcement can be written
as
ε f( K ) ( x) = ε f (σ c( K ) , z ( K ) ( x), L(↓K ) ( x), L(↑K ) ( x)), (4)
where σ c( K ) is the composite stress initiating the K -th crack. The corresponding composite
strain ε c is obtained as an average strain along the specimen using the integral
1 L (K )
L ∫0
ε c( K ) = ε f ( x)dx. (5)
In order to reproduce the stress-strain diagram corresponding to the whole loading process,
the cracking history, i.e., the crack positions X k( K ) and corresponding crack initiating com-
posite stresses σ c( K ) , need to be determined. Assuming that the matrix tensile strength is a
random stationary function of the global coordinate x given as
σ mu = f ( x), (6)
the condition for a new matrix crack to appear at a point is that the matrix stress reaches
the matrix strength, which can be written as
Σm (λ ; z( K ) ( x), L(↓K ) ( x), L(↑K ) ( x)) = σ mu ( x). (7)
Fig. 2: Incremental evaluation of the cracking history
143
By solving Eq. (7) the load factor needed to initiate a matrix crack at a point x can be ob-
tained as
Λ ( K ) ( x) := {λ | Σm (λ; z( K ) ( x), L(↓K ) ( x), L(↑K ) ( x)) = σ mu ( x)}. (8)
The sought load σ c( K +1) initiate the new crack K + 1 corresponds to the minimum of
Λ ( K ) ( x) evaluated along the whole specimen:
σ c( K +1) = min Λ ( K ) ( x) (9)
x∈[0, L ]
with the corresponding crack position expressed as
X K( K+1+1) = arg min Λ( x). (10)

x∈[0, L ]
The history of cracks with associated state variables is obtained by repeated evaluation of
Eqs. (8) – (10), for an incremented crack counter K = K + 1 as shown in Fig. 2. In the initial
state K = 0 without any cracks, the matrix stress is constant along the specimen and can be
obtained explicitly using the mixture rule. The whole cracking history is determined when
either of the following two conditions is met: a) with a fine crack pattern, the whole speci-
men is covered by the shielded zone of the cracks such that no solution can be found for
Eq. (7) at any point x . b) the minimum of Λ ( K ) ( x) is greater than the composite specimen,
which indicates that the ultimate failure of the specimen has been reached. By combining
the cracking history tracing algorithm with the established stochastic crack bridge model
[5] we are able to reproduce the stress-strain diagrams of uniaixally tensioned TRC speci-
mens with statistically represented local properties of the fabrics, bond and matrix.
OPTICAL CRACK DETECTION SCHEME

In order to accurately capture the developments of crack patterns in the tensile tests, we
employed the digital image correlation technique to record and analyse the displacement
field along the tensile specimen. A commercial Aramis 3D system developed by the Ger-
man company GOM mbH [6] was used to measure the history of displacement fields on a
surface grid as indicated in Fig. 3.
144
Fig. 3 The grid of representative points
Fig. 4 (a) The non-local strain field along x direction and its derivatives; (b) Detected crack pattern.
Besides the initial coordinates of the grid points X j ,k ( j = 1, 2,..., n , k = 1, 2,..., m ) Aramis
delivers their displacements u ij ,k in the following time steps i . Based on these displace-
ments, the non-local strains along the specimen were calculated using the difference
scheme
u ij ,k + r − u ij ,k −r
ε i
j ,k = , (11)
X j ,k + r − X j ,k −r
where r represents the interactions radius that was introduced to factor out the noise in the
measured data. The non-local strain profile along the specimen together with the strain
derivatives at the ultimate state is exemplified in Fig. 4(b). With these calculated fields, the
criteria to determine a crack at a point x is formulated as
dε x d 2ε x
=0 ∧ ≤δ, (12)
dx dx 2
Fig. 5 Test setup for validation

where δ is the threshold filtering the possible noise. After the crack positions have been
determined at the ultimate state, the strain field was traced back through the time history in
order to identify the time step and load level at which each individual crack appeared.
145
COUPLED NUMERICAL AND EXPERIMENTAL ANALYSIS

Tensile tests of ten thin TRC specimens, five with 1.0% and the other five with 1.5% rein-
forcement volume fraction, were carried out. The test setup is shown in Fig. 5, the tensile
load was applied through the steel plates [7].
The frictional bond stress in the interface between reinforcement and matrix is assumed
random, prescribed by the Gamma distribution with the scale parameter 2.276 and shape
parameter 0.050. Furthermore, the fiber breaking strain is described by the Weibull distri-
bution with scale parameter 0.0134 and shape parameter 8.806. These parameters were
calibrated using a single crack bridge test [8]. With the bond and fiber parameters at hand
and with the crack positions and load levels delivered by the crack detection scheme de-
scribed in section 3, the stress-strain response of the composite can be recalculated as ex-
emplified in Fig. 6(a) for two selected specimens, one with vf = 1.0% and the other one
with vf = 1.5% .
The crack detection scheme also delivered the load level initiating the first crack in the
matrix, based on which the minimum of the matrix strength field can be calculated. It is
found that the minimum matrix strength grows with increasing fiber volume fraction
( vf = 1.0% → σ mu, min = 3.05 MPa; vf = 1.5% → σ mu, min = 3.71 MPa). This conforms with
the phenomenon that the fiber volume fraction influences the matrix crack stress, which
has been observed experimentally and studied theoretically [9][10].
The first matrix cracking stress can be used to derive the parameters of the matrix strength
profile, which was assumed a 1d auto correlated random field [11]. For the studied cases,
the shape parameter was characterized as mm = 45 , and the scale parameters for
vf = 1.0% was sm = 3.599 ; for vf = 1.5% , it was increased to sm = 4.384 considering the
reinforcing effect of higher fiber volume ratio. Using those parameters, the cracking histo-
ries were numerically reproduced by the crack tracing algorithm described in section 2.
Then according to Eq. (4) and Eq. (5), the stress-strain diagram shown in Fig. 6(b) can be
obtained.
Fig. 6 Stress-strain relation evaluated with cracking history delivered by: (a) optical crack detection scheme;
(b) numerical crack tracing algorithm.
As shown in Fig. 6, the stress-strain curves agree well with the experimental responses.
Also, the average crack spacing predicted by the crack tracing algorithm with vf = 1.5%
146
(9.23mm) is close to the experimental value 10.2mm, this serves as a validation of the cali-
brated model. A minor discrepancy between the cracking load levels delivered by the crack
detection scheme and those predicted by the crack tracing algorithm can be observed. As
shown in Fig. 7, especially for vf = 1.5% , the cracks predicted by the crack tracing algo-
rithm appear slightly later compared to those detected in the experiment. The reason lead-
ing this discrepancy might be we only considered the final crack density with vf = 1.0% in
the calibration procedures while the cracking process was ignored.
Fig. 7 Crack density development: (a) vf = 1.0% ; (b) vf = 1.5% .
CONCLUSIONS
In this paper, we described a numerical model reproducing the tensile response of TRC
specimens. Using the digital image correlation technique, the development of the crack
patterns during the test can be accurately captured. The crack detection tools delivers the
crack positions and cracking load levels. It was observed that larger fiber volume ratio
leads to higher matrix cracking stresses. The bond and fiber breaking strain parameters
were calibrated in a systematic manner. Also, according to the minimum cracking load
levels, the statistical characteristics of the matrix strength could be identified. With these
parameters, the stress-strain diagrams and the average crack spacing predicted by the nu-
merical model agree well with the experimental data.
ACKNOWLEDGEMENT
The financial support of this work by the Deutsche Forschungs Gemeinschaft (DFG) with-
in the project CH 276/2-2 and by the European social fund within the framework of realiz-
ing the project „Support of inter-sectoral mobility and quality enhancement of research
teams at Czech Technical University in Prague“, CZ.1.07/2.3.00/30.0034 is gratefully
acknowledged.
REFERENCES
[1] Aveston, J., Cooper, G. A. and Kelly A, ‘Single and multiple fracture. The Properties of
Fibre Composites’, Proc. Conf. National Physical Laboratories, London, (1971) 15–24.
[2] Aveston, J. and Kelly, A., ‘Theory of multiple fracture of fibrous composites’, J. Mat. Sci. 8
147
(1973) 411–461.
[3] Curtin, W. A., Ahn, B. K. and Takeda, N., ‘Modeling brittle and tough stress-strain behav-
iour in unidirectional ceramic composites’, Acta. Mater. 10(1998):3409–3420.
[4] Rypl, R., and Chudoba, R., et al., Brittle matrix composites with heterogeneous reinforce-
ment: Multi-scale model of a crack bridge with rigid matrix. Composites science and tech-
nology, 89(2013):98-109.
[5] Rypl, R. and Chudoba, R., ‘Multi-scale model for tensile response of brittle matrix compo-
sites: Part I Microscopic representative crack bridge model’, Manuscript prepared for sub-
mission.
[6] GOM mbH, ‘ARAMIS manual Software Aramis 6.3 and higher’, (www.gom.com, 2011).
[7] Scholzen, A., Chudoba, R. and Hegger, J., ‘Ultimate limit state assessment of TRC shell
structures with combined normal and bending loading’, Manuscript submitted to FERRO-11
and 3rd ICTRC
[8] Rypl, R. and Chudoba, R., et al., ‘Multi-scale model for tensile response of brittle matrix
composites: Part II Stochastic cracking model and experimental validation’, Manuscript pre-
pared for submission.
[9] Marshall, D. B. and Cox, B. N., ‘Tensile fracture of brittle matrix composites: influence of
fiber strength’, Acta Metallurgica, 35 (1987):2607-2619.
[10] Marshall, D. B., Cox, B. N., and Evans, A. G., ‘The mechanics of matrix cracking in brittle-
matrix fiber composites’, Acta Metallurgica, 33(1985):2013-2021.
[11] Vorechovsky, M., ‘Simulation of simply cross correlated random fields by series expansion
methods’, Structural safety, 30 (2008):337-363.
148
IMPLEMENTATION OF EXPERIMENTAL DATA IN ANALYSES OF

TEXTILE REINFORCED CONCRETE STRUCTURES
Natalie Williams Portal (1), Mathias Flansbjer (2), Karin Lundgren (3), Katarina Malaga (1), (1)
CBI Swedish Cement and Concrete Research Institute, Sweden, (2) SP Technical Research
Institute of Sweden, Sweden, (3) Chalmers University of Technology, Sweden
Abstract: Textile reinforced concrete (TRC) is an innovative high performance composite

material which has revealed many promising attributes in various applications but has not
yet reached its recognition due to a lack of available design tools. To be able to reach this
stage, standardized test methods and reliable numerical models need to be established to
reduce uncertainty and the need for extensive experimental studies. The aim of this paper
is to present an approach which links experimental methods to simplified numerical anal-
yses. The applied experimental methods are categorized according to the mechanical be-
haviour features they encompass: material, interaction and global levels. A 2D-macro-scale
FE-model was developed, as a case study, using experimental data from material and inter-
action levels as input. The outcome of this analysis was compared to the global level data
to identify the effects of key parameters in FE-modelling of TRC. Comparison of the FE-
analysis results and the experiments indicated that the developed modelling method de-
scribes certain mechanisms adequately while posing certain limitations. It was shown that
the bond between the textile and the concrete is a very critical input parameter which af-
fects crack development, deflections, and failure mechanisms.
INTRODUCTION
In fibre composite materials, such as textile reinforced concrete (TRC), the existing com-
plex heterogeneous structure poses challenges concerning modelling and design. The un-
derlying complexity is related to the bond between reinforcement and the fine-grained con-
crete matrix but also to the bond between the individual filaments found within a yarn.
Various types of numerical methods have been developed for the purpose of deepening the
understanding of the bond behaviour and interaction between the filaments [1-4], yet these
can be computationally expensive when linked to other levels of simulation, i.e. multi-
scale, and the required input parameters are also difficult to come by. Simplified numerical
models could however be developed using detailed experimental input parameters which
describe the complex bond behaviour along with important material and mechanical be-
haviours.
An approach which links experimental methods to simplified numerical models is present-
ed in this work, similar concepts have also been applied by others [5-7]. Macro-scale mod-
elling is investigated using experimental inputs obtained at various levels of detail (see
Figure 1). These experimental inputs capture different features of the mechanical behav-
iour of TRC, which are defined in this work as material, interaction and global. The mate-
rial level consists of the quantification of the material characteristics of the concrete and
149
reinforcement, e.g. tensile and compressive behaviour. The composite behaviour of these
two materials is thereafter taken into account through experimental testing at the interac-
tion level, i.e. pull-out tests. As for the global level, TRC is tested in the form of a compo-
site structure under prescribed loading conditions, i.e. tensile or flexural loading. Thereaf-
ter, the outcome of the macro-scale level is applied to validate and gain insight on the sen-
sitivity of input parameters related to the associated numerical model.
Figure 1: Conceptual figure of the implementation of experimental data in a macro-scale mechanical model
A comprehensive experimental test program based on the defined experimental levels,

shown in Figure 1, is presented in this paper. The resulting outputs of the experiments at
the material and interaction level were included as input parameters in a case study of a 2D
macro-scale model of a TRC panel undergoing four-point bending. By doing so, one can
gain a better understanding of the complex heterogeneous behaviour and the effects of dif-
ferent key parameters used in the macro-modelling of TRC.
EXPERIMENTAL WORK
Material level
The material level includes the measurement of the material characteristics pertaining to
both hardened concrete and the selected textile reinforcement. Concerning the concrete, for
simplicity, the properties of interest here are categorized as compressive and tensile nature
at an age of 28 days.
The mechanical behaviour of concrete in compression was evaluated based on EN 12390-3

[8] and EN 12390-13 [9] to obtain the stress-strain relation, σ(ε), compressive strength, fc,
ultimate strain, εcu, modulus of elasticity, Ec, and Poisson’s ratio, νc. The tests were carried
out in a GCTS servo-hydraulic machine with a stiff load frame and load measurement ac-
curacy of 1 %. The axial deformation of the specimen was obtained by two inductive dis-
placement transducers measuring the distance change between two aluminium rings at-
tached to the specimen. The transducers had a measuring range of ± 2.50 mm and a relative
error less than 1 %. The circumferential deformation was obtained by using a chain mount-
ed around the specimen at half height. The change in chain-opening gap was measured by
an inductive displacement transducer. The instrumentation for axial and circumferential
deformation measurements is illustrated in Figure 2 (left).
150
Figure 2: Example of compressive stress-strain results (left) and tensile stress-deformation results (right)
The material characteristics in tension, including tensile strength, ft, stress-crack opening
relation, σ(w), and fracture energy, GF, were evaluated using uniaxial tensile tests on
notched cylinders (see Figure 2 (right)) based on RILEM TC 187-SOC [10] and RILEM
TC 162-TDF [11]. The tests were conducted using the same test apparatus used for the
compressive tests. The deformation was measured locally over the notch with three induc-
tive displacement transducers, whereas the deformation, δ, was calculated as the mean val-
ue of the three displacement gauges. The accuracy of the load measurement as well as
measuring range and relative error of the transducers were the same as for the compressive
tests.
As a part of the evaluation, the stress-deformation curve shown in Figure 2 (right) was
used for deriving the softening behaviour of the concrete material. The tensile stress, σ,
was derived by dividing the load by the effective cross-section, Aeff, at the notch. The ten-
sile strength, ft, is defined as the peak stress and deformation at peak stress, δtu, which takes
place at the onset of macro cracking.
General material and mechanical properties are commonly provided by the textile rein-
forcement producers, which include geometric parameters of the mesh/grid, weight, ap-
plied coating, tensile strength, modulus of elasticity and elongation, but are not limited to
these parameters. However, the methods used to obtain the mechanical properties vary
based on the source and are most often not documented, thus the soundness of the available
data could be questioned. The quantification of the general material properties of the textile
reinforcement was conducted using tensile testing according to the standard method stated
in ISO 10406-1 [12]. The ultimate tensile capacity, Fu, tensile rigidity EA, and ultimate
strain, εu, of the textile reinforcement are the general outputs of this test method, yet these
can also be converted to tensile strength and Young’s modulus of the material. The tests
were carried out using a universal testing machine (Sintech 20D) illustrated in Figure 3 and
the force was recorded by a load cell with an accuracy of 1 %. The deformation was meas-
ured by a Messphysik Videoextensometer ME46 with backlight technique in the back-
ground and digital camera in the foreground as per Figure 33. Further details pertaining to
this experiment are documented in Williams Portal, Flansbjer, et al. [13].
151
Figure 3: Example of tensile test results (left) and tensile test setup with video extensometer (right).
Interaction level
The interaction level comprises the quantification of the complex heterogeneous behaviour
between the cementitious matrix and the textile reinforcement. This behaviour was meas-
ured by means of pull-out tests and was characterized as a calibrated bond stress-slip rela-
tionship of TRC with carbon textile reinforcement. The pull-out tests consisted of a dou-
ble-sided unsymmetrical configuration, whereby the specimen was divided into a shorter
embedment length zone and a longer anchorage length zone. The specimens had dimen-
sions of 400 x 100 x 15 mm and were reinforced by a centrally cast layer of reinforcement
mesh. The varying of embedment lengths was studied in order to characterize both pull-out
and rupture of the textile. The experimental setup developed to conduct the pull-out tests is
shown in Figure 4. A detailed account of the experimental program, and the evaluation of a
bond stress-slip behaviour, is reported in Williams Portal, Perez Fernandez, et al. [5] and
Williams Portal, Lundgren and Malaga [14]. It should be noted that the concrete and textile
reinforcement applied in these particular tests slightly differ from the other presented tests.
This bond stress-slip behaviour is however applied in the case study as an expected behav-
iour.
Figure 4: Bond stress-slip curve (left) derived from the pull-out tests (right)
152
Global level
The global level includes experiments that describe the global behaviour of TRC struc-
tures. These obtained data can be used in a further step to verify the composite behaviour
of the given developed macro-mechanical model. For instance, from the tensile tests of
TRC, it is possible to evaluate the stress-deformation relation, first crack stress and maxi-
mum load resistance, type of failure, crack spacing and crack widths. Similar outputs can
be retrieved from flexural tests, with the addition of load-midspan deflection relation and
flexural resistance.
To determine the tensile load bearing behaviour of TRC, uniaxial tensile tests were per-
formed according to the draft recommendations given in RILEM TC 232-TDT [15] (see
Figure 5 (right)). In this study, the specimens were 700 x 100 x 20 mm with two layers of
carbon textile reinforcement (see Figure 6). These tests were carried out in an electro-
mechanical universal testing machine (Sintech 20D) and were displacement controlled.
The load and machine displacement were recorded in a data acquisition system (sampling
rate of 10 Hz).
In order to investigate the flexural properties of TRC, four-point bending tests were per-
formed based on EN 12390-5 [16], yet other comparable provisions could also be applied.
The tests were performed using a servo-mechanical testing machine (Instron 1195). The
sample panels with the same geometry as those used in the tensile tests were placed on
roller supports with a span of 600 mm as depicted in Figure 5 (left). The samples were
loaded by two upper point loads each applying a load of P/2 with a distance 200 mm apart
from each other. The loading was applied using displacement control. On opposite sides of
the panels, two linear displacement transducers (LVDT) were placed at midspan.
Figure 5: Flexural test (left) and uniaxial tensile test setup with crack measurements from DIC measuring
system (right)
To obtain precise measurements of the resulting deformations and crack developments, an

optical full-field deformation measurement system ARAMISTM 12M by GOM was used
for tensile tests, but could also be applied to flexural tests. The system uses a measurement
technique based on Digital Image Correlation (DIC) with a stereoscopic camera setup,
consisting of two CCD-cameras. Over the course of the uniaxial tensile testing in this
study, this system was used to record the displacement of the specimen during testing by
tracking the deformation of a speckle pattern in a series of digital images acquired during
the loading. The displacement of the pattern within discretized facets of the image was
153
thereafter analysed. It is possible to capture the initiation of the first crack as well as close
up images of the local crack measurements by means of virtual extensometers as shown in
Figure 5 (right).
NUMERICAL MODEL
A case study is presented here consisting of a non-linear 2D macro-scale model which was
developed based on the presented four-point bending tests of a TRC specimen reinforced
by two layers of carbon textile reinforcement grid (see Figure 6). The commercial software
DIANA (DIsplacement ANAlyser) with pre- and post-processor Midas FX+ was used to
perform the analyses. The main purpose of this model was to verify the experimentally
observed flexural behaviour and failure mode, as well as to identify difficulties and uncer-
tainties in the modelling of TRC.
A mesh size of 2 x 2 mm was chosen to fit element rows below or above the reinforcement
while ensuring that the textile reinforcement was located at a row of nodes. Interface ele-
ments are included here because the bond between the textile reinforcement and matrix
cannot be assumed to be perfect due to the heterogeneous structure of the yarns.
The element types used for the various model components are provided in Table 1 along
with a description of the associated data inputs. In the model, the point loads were applied
as fixed deformations. In each step, equilibrium was calculated using secant iterations,
which was found to yield the most suitable solution. This deformation controlled loading
means that the behaviour, especially during the cracking state, can be followed more accu-
rately which can also lead to improved comparison with the experimental results.
Figure 6: Overview of developed macro-scale model
154
Table 1: Model inputs from material and interaction levels.

Component Element type Material Level Interaction Level
input input
Concrete 2D quadrilateral - Compressive behaviour —

isoparametric (72 MPa , Figure 2)
plane stress; 4
- E-modulus (36 GPa)
nodes (Q8MEM)
- Tensile behaviour
(3.2 MPa, Figure 2)
- Fracture energy
(96 N/mm)
- Poisson’s ratio (0.2)
Textile rein- 1D truss bar; 2 - Tensile behaviour —
forcement nodes (L2TRU) (Figure 3)
grid - E-modulus (250 GPa)
Interface 2D line-interface; — - Bond stress-slip

4 nodes (L8IF) curve (Figure 4)
The numerical simulation results were compared to the Global experimental results related
to the abovementioned TRC specimen undergoing four-point bending. The load versus
midspan deflection relationship of two simulations, denoted as 1 and 2 are compared to the
experimental data in Figure 7. It should be noted that the main difference between Simula-
tions 1 and 2 is the defined contact perimeter of the reinforcement which has been adjusted
to visualize the need for matching Interaction level input parameters, which was similarly
implemented in Williams Portal, Lundgren, et al. [6]. As can be observed, first cracking of
Simulation 1 takes place at a comparable load level to that of the experiments, yet it is de-
layed presumably due to the dissimilar Interaction level input. An attempt to yield an im-
proved correlation was made with Simulation 2, wherein first cracking takes place nearer
to the actual deflection value and slightly lower load. Also, the load drop after the first
crack is smaller while the post-crack stiffness is comparable to the experimental curves.
The state of stress and strain at first cracking for Simulation 2 is depicted in Figure 7.
155
Figure 7: Load versus mid-span deflection relationship. A: shift between first cracking for Simulations 1+ 2.
Furthermore, in previous work by Pettersson and Thorsson [17], it has been shown that an
adequate comparison between experimental and numerical results can be achieved by us-
ing a bond stress-slip curve corresponding to the given Global experimental material com-
bination. Accordingly, it is important to note that these presented analyses can be further
enhanced by incorporating corresponding Interaction level input, reduction of concrete
tensile strength at yarn cross-thread locations, as well as more refined computation using
more adept solvers able to capture the vertical load drops at crack occurrences.
CONCLUSIONS AND OUTLOOK
An approach linking experimental methods to simplified numerical models has been pre-
sented in this paper. The implementation of experimental input data obtained from various
levels of detail (Material and Interaction) in macro-scale modelling was shown to be a
promising approach to compare simulation results to those from the experimental Global
level. One main observation which can be drawn is such that the bond stress-slip relation-
ship from the Interaction Level is a critical input parameter in macro-scale analyses. Ac-
cordingly, it is paramount to have detailed experimental methods which can characterize
input data at the various stated levels for the sake of yielding accurate numerical analyses.
Lastly, it should be highlighted that the geometries applied in this work are based on a giv-
en application, testing apparatuses and test requirements, but can be varied according to the
intended design scenario.
ACKNOWLEDGEMENT
This research was funded by Formas BIC project Tekocrete II - Energy efficient thin fa-
çade elements for retrofitting of Million Programme housing: TRC façade elements, and
also with the support of the European Union’s Seventh Framework Programme under grant
agreement no. 608893 (H-House, www.h-house-project.eu/).
156
REFERENCES
[1] Hartig, J., U. Häußler-Combe, and K. Schicktanz, Influence of bond properties on the tensile behav-
iour of textile reinforced concrete. Cement and concrete composites, 2008. 30(10): p. 898-906.
[2] Lorenz, E. and R. Ortlepp, Bond behavior of textile reinforcements-development of a pull-out test and
modeling of the respective bond versus slip relation, in High Performance Fiber Reinforced Cement
Composites 6. 2012, Springer. p. 479-486.
[3] Richter, M., I. Lepenies, and B.W. Zastrau. On the influence of the bond behaviour between fiber and
matrix on the material properties of textile reinforced concrete. in International symposium of aniso-
tropic behaviour of damaged materials. 2002.
[4] Chudoba, R., M. Vořechovský, and M. Konrad, Stochastic modeling of multi-filament yarns. I. Ran-
dom properties within the cross-section and size effect. International Journal of Solids and Structures,
2006. 43(3): p. 413-434.
[5] Williams Portal, N., I. Perez Fernandez, L. Thrane Nyholm, and K. Lundgren, Pull-out of textile rein-
forcement in concrete. Construction and Building Materials, 2014. 71: p. 63-71.
[6] Williams Portal, N., K. Lundgren, A.M. Walter, J.O. Frederiksen, and L. Nyholm Thrane. Numerical
modelling of textile reinforced concrete. in Proceedings of VIII International Conference on Fracture
Mechanics of Concrete and Concrete Structures. 2013.
[7] Schladitz, F., M. Frenzel, D. Ehlig, and M. Curbach, Bending load capacity of reinforced concrete
slabs strengthened with textile reinforced concrete. Engineering structures, 2012. 40: p. 317-326.
[8] EN 12390-3: Testing hardened concrete - Part 3: Compressive strength of test specimens. 2009.
[9] EN 12390-13: Testing hardened concrete - Part 13: Determination of secant modulus of elasticity in
compression. 2013.
[10] RILEM TC 187-SOC. Experimental determination of the stress-crack opening curve for concrete in
tension: Final report. 2007.
[11] RILEM TC 162-TDF. Test and design methods for steel fibre reinforced concrete: Uni-axial tension
test for steel fibre reinforced concrete. Materials and Structures (RILEM), 2001. 34: p. 3-6.
[12] ISO 10406-1, Fibre-reinforced polymer (FRP) reinforcement of concrete - Test Methods, in Part 1:
FRP bars and grids. 2008.
[13] Williams Portal, N., M. Flansbjer, K. Tammo, and K. Malaga, Alkali Resistance of Textile Rein-
forcement for Concrete Facade Panels, in XXII Concrete Research Symposium. 2014: Reykjavik
[14] Williams Portal, N., K. Lundgren, and K. Malaga. Evaluation of Pull-out Behaviour in Textile Rein-
forced Concrete. in 10th fib International PhD Symposium in Civil Engineering. 2014.
[15] RILEM TC 232-TDT. Test methods and design of textile reinforced concrete: Uniaxial tensile test -
Test method to determine the load bearing behavior of tensile specimens made of textile reinforced
concrete. . Draft recommendations., 2014.
[16] EN 12390-5: Testing hardened concrete - Part 5: Flexural strength of test specimens. 2009.
[17] Pettersson, M. and P. Thorsson, FE-modelling of Textile Reinforced Concrete Facade Elements, in the
Department of Civil and Environmental Engineering, Division of Structural Engineering - Concrete
Structures. 2014, Chalmers University of Technology: Gothenburg, Sweden. p. 88.
157
158
ULTIMATE LIMIT STATE ASSESSMENT OF TRC SHELL

STRUCTURES WITH COMBINED NORMAL AND BENDING
LOADING
Alexander Scholzen (1), Rostislav Chudoba (1), Josef Hegger (1), (1) Institute of Structural
Abstract: For the dimensioning of textile reinforced concrete (TRC) shell structures as-
sessment rules are needed to successfully introduce the innovative composite material into
engineering practice. The present paper describes a dimensioning approach for spatial TRC
structures capturing the interaction between normal forces and bending moments. The ad-
missible loading range herein is delimited by a linear approximation of the n-m interaction
diagram constructed by connecting the values of tensile, compressive and bending
strength. In order to validate this approximation, a test setup for combined normal and
bending load has been developed. The n-m interaction diagram forms the basis for the nu-
merical assessment of the ultimate limit state of TRC shell structures including the effect
of oblique loading on the composite strength. For this purpose a numerical assessment tool
has been implemented that automatically evaluates the utilization ratios of all TRC shell
cross-sections for all loading cases based on linear-elastic finite-element analysis of the
structure. The formulated dimensioning approach has been successfully applied for the
assessment of two TRC shell structures recently realized in Aachen.
INTRODUCTION
High strength, ductility and non-corrosiveness make TRC very suitable for construction of
thin-walled, light weight shell structures [1]. As discussed in [2] one of the major condi-
tions for wide acceptance of this type of structures with complex geometry and stress dis-
tribution in practice is the availability of safe and transparent approaches to dimensioning
and assessment. A systematic procedure for the ultimate limit state (ULS) assessment of
TRC shells has been proposed and successfully applied by the authors for the construction
of two shell structures, a hypar shell structure as described in [3,4] as well as the barrel-
vault shells shown in Figure 1. In the present paper the dimensioning approach is briefly
summarized (Section 2) and its computer-based evaluation exemplified for the depicted
barrel-vault shells. The underlying cross-sectional strength characteristics of the composite
material are determined experimentally as described in Section 3 with the focus on the
novel tensile test setup for thin TRC elements. Due to the applied kinematics in the design
of the clamps the tensile force can be flexibly combined with an independently controllable
additional bending moment. The results of the test series covering the n-m interaction
range between pure tension and pure bending are described and discussed in Section 4.
159
Figure 1: Application examples of lightweight TRC shell structures at RWTH Aachen University: double
curved hypar shells (left); barrel-vault precast elements (right)
UTILIZATION RATIO FOR COMBINED NORMAL AND BENDING LOADS
With reference to the design codes for steel reinforced concrete structures let us remark
that cross-section based strength characteristics provide an efficient means for dimension-
ing and assessment of composite cross sections exposed to combined load action. In analo-
gy to these design codes, a simplified n-m interaction diagram for TRC cross-section was
n
constructed using the experimentally obtained strength values for uniaxial tension t,Rd
n m
[kN/m], uniaxial compression c,Rd [kN/m] and pure bending Rd [kNm/m] assuming a
linear interaction of the cross sectional-strength between these values [4]. For convenience,
the linear n-m interaction diagram has been transformed into a normalized form as shown
in Figure 2 (left) using the normalized stress resultants
nEd n |m |
ηntd = ηncd = − Ed ηmd = Ed (1)
nt,Rd nc,Rd mRd
The normalized stress resultants represent the utilization ratio of the cross-section with
respect to the design values of the strength characteristics for tension ( nt,Rd ), compression
( nc,Rd ) and bending ( mRd ). Considering the fact that the values of the strength characteris-
tics are all positive, the normalized stress resultants attain their ultimate strength values for
η (.) = 1 .0 . For a combined loading of a cross-section with normal force and bending mo-
ment, the total utilization ratio η nmd is defined as:
ηnmd = max(ηntd ,ηncd ) + ηmd (2)
The ultimate limit state assessment of a TRC shell must prove that for all load case combi-
nations the total utilization ratio ηnmd lies within the admissible range η nmd ≤ 1.0 in any
cross section of the shell (dashed area in Figure 2, left).
In the described dimensioning procedure we have assumed isotropic material properties

with cross-sectional strength characteristics independent of the loading direction. However,
a TRC shell cross-section with orthotropic reinforcement structure exhibits an anisotropic
160
behaviour due to the misfit between the direction of the principal tensile stresses and the
orientation of the fabrics. The deviation of the flexible textile reinforcement leads to an
increased stress concentration in the outer filaments of the roving at the crack edges, there-
by reducing the total load bearing capacity. This was taken into account through the reduc-
tion factor kα depending on the deviation angle α [3,4]. Since the design values of the
composite strength depend on the orientation of the textile fabrics, it is not possible to
identify a priori the critical shell cross-sections relevant for the ULS assessment. While for
each load case combination the cross section with maximum principal tensile stresses can
be determined, this cross section does not necessarily become decisive for the dimension-
ing because the largest deviation of the fabric orientation may occur at some other area of
the shell. As a consequence, the design values of the stress resultants must be evaluated for
all possible load case combinations constructed in accordance with DIN EN 1990 and for
all points of the TRC shell taking into account the directional-dependency of the strength
values. The computer based evaluation for the ULS assessment of the TRC barrel-vault
shells shown in Figure 1 is depicted in Figure 2 (right). As can be seen, all utilization ratios
lie within the admissible limits.
Figure 2: Dimensioning approach for TRC shell structures based on a simplified n-m interaction diagram
(left); computer-based evaluation of the utilization ratios for all elements of the finite element discretisation
and all load case combinations (right)
DETERMINATION OF THE CROSS-SECTIONAL STRENGTH CHARACTERISTICS
The cross-sectional strength characteristics of the TRC barrel-vault shell needed for the
ULS assessment were determined experimentally using tensile, compression and bending
tests. The thickness and cross-sectional layup of the TRC specimens hereby corresponded
to the characteristics of the considered barrel-vault shells [5]. All specimens had a thick-
ness of t = 2 cm and were reinforced with 6 layers of carbon textiles [3]). The cross-
sectional strength values for tension were determined using the test setup shown in Fig-
ure 3. It uses buttstrap clamps to introduce the tensile force into the straight TRC speci-
161
men. In contrast to other existing test setups [6,7] the designed setup allows for a free
length at the ends of the buttstraps by adjusting the length of the treaded bars running along
the specimen. By this means the necessary anchorage length of the textile reinforcement
can be provided in order to avoid the pull-out effect. A further advantage of the test setup
is the fact that spurious bending effects can be totally prevented due the kinematic degrees
of freedoms provided by the ball hinges at the ends of the threaded bars (Pos. 1 in Fig-
ure 3). On the other hand, the test setup is limited by the maximum achieavable frictional
force at the buttstrap clamping. In order to increase the frictional force an additional
clamping was added to the setup as depicted in Figure 3. For thicker specimens and very
high reinforcement ratios waisted tensile tests are more suitable if the necessary anchorage
length can be achieved [4]. A thorough analysis of the test results is necessary to verify the
quality of the measured data as described in [4].
Figure 3: TRC tensile test setup with buttstrap clamps and variable anchorage length
The bending strength can be determined in conventional bending tests [4]. In the given
case three-point bending was conducted on TRC stripes (t/b/L = 2/10/50 cm) with a span of
l0 = 46 cm). The composite strength for tension and bending directly depends on the used
textile reinforcement. Therefore, the strength characteristic need to be evaluated both for
the 0°- and 90°-orientation of the textile fabrics. For the considered carbon fabric the spac-
ing of the rovings and therefore the resulting strength characteristics were almost identical
in both directions so that only the smaller values obtained for the 0°-orientation were con-
sidered in the ultimate limit state assessment for the sake of simplicity. In the given test
162
series the average measured composite tensile strength was σ t ,0° = 22.9 MPa . The com-
pressive strength of the TRC shell cross-section corresponds to the compressive strength of
the fine grained concrete with an average value of fcm = 68 MPa .
Based on the conducted test results the characteristic strength values (5% quantiles) were
obtained using the statistical evaluation of the measured ultimate loads in accordance with
DIN EN 1990 leading to a reduction factor due to scatter. Considering a partial safety fac-
tor γ = 1 .5 the design values of the cross-sectional strength characteristics for tension,
compression and bending evaluated to:
nt,Rd = 257 kN m nc,Rd = 750 kN m mRd = 1.8 kNm m (3)
VALIDATION OF THE LINEAR INTERACTION DIAGRAM FOR COMBINED LOADING
In order to validate the simplified linear interaction assumption underlying the proposed n-
m interaction diagram a test setup for combined loading with tensile force and bending
moment has been developed (Figure 3). It is based on the tensile test setup described in
Section 3. For the introduction of an additional bending moment a horizontal load is ap-
plied at the middle of the specimen via a hydraulic cylinder. The cylinder is fixed to a self-
supported steel construction suspended from the testing frame at one side of the tensile test
setup. The steel construction is connected to the buttstrap clamps of the tensile test setup
via steel rings in order to avoid spurious bending moments or additional constraints at the
clampings. The hinge joints at the buttstrap clampings enable a free rotation at the speci-
men ends due to bending. In order to achieve a possibly smooth curvature profile along the
whole tested length the original setup in Figure 3 was modified by moving the position of
the hinges connecting the threaded bars with the clamps to the beginning of the buttstraps
whereas the additional clamping was moved to the center of the buttstraps. Due to the in-
dependent control of the tensile and bending load, the developed test setup can be used to
define arbitrary combinations of tensile force and bending moment. In the performed test
series the normal force was applied at first and kept constant through-out the test. Then, the
bending load was applied at a constant loading rate.
163
Figure 4: Test setup for combined loading with tensile force and bending moment
Due to their slenderness and ductile material behaviour, specimens exposed to bending
load exhibit large deflection. In connection with the applied normal force this has a signifi-
cant influence on the measured ultimate bending moment and needs to be considered in the
interpretation of the test data. The second-order moment MII can be calculated based on the
applied normal force N and the measured deflection w at the mid-span position of the ap-
plied horizontal load F. The deflection was measured via a horizontal displacement gauge
(Figure 3). The effective bending moment M at failure then amounts to the difference be-
tween the first-order and second-order moments M0 and MII, respectively, as:
M = M0 - MII = F· l0 / 4 - w · N (4)
The results of the experimental investigations are plotted in Figure 5 for five different rati-
os of applied normal force N and effective bending moment M ranging from pure tension
(Nmax) to pure bending (Mmax). It must be noted that a low normal force is also necessary
for the bending test conducted with the test setup in Figure 4 in order to apply a slight pre-
tension of the specimen. The three intermediate levels of tensile load (I, II and III) corre-
spond to approximately 25%, 50% and 75% of the expected normal strength. For each lev-
el three test replications were performed.
164
Figure 5: Test results for combined tensile force and bending moment
As apparent from the diagram, the obtained test results show that the cross-sectional
strength for combined loading is relatively well approximated by the assumed linear inter-
polation between tensile and bending strength (corresponding to the dashed line in Figure
4) underlying the proposed n-m interaction diagram.
SUMMARY
For the ULS assessment of TRC shell structures a simplified linear n-m interaction dia-
gram has been proposed based on the experimentally obtained cross-sectional strength
characteristic of the composite. To determine the tensile strength a novel test setup has
been designed providing a sufficient anchorage length of the textile fabrics yielding repro-
ducible strength characteristics without spurious pull-out effects.
The developed test setup also provides the basis for the validation of the described linear n-
m interaction diagram as it can easily be extended with an additional bending load adapter.
The test results for combined loading show good agreement with the underlying assump-
tions of the proposed dimensioning approach for TRC shell structures.
ACKNOWLEDGMENT
165
REFERENCES
[1] Tysmans, T.; Adriaenssens, S.; Cuypers, H.; Wastiels, J.: Structural analysis of small span textile
reinforced concrete shells with double curvature, Composites Science and Technology, 2009, No.
69, 1790-1796
[2] Cassinello, P.; Schlaich, M.; Torroja, J.A.: Félix Candela. In memorian (1910-1997). From thin
concrete shells to the 21th century light weight structures, Informes de la Construcctión, 2010,
62(519): 5-26
[3] A. Scholzen, R. Chudoba, J. Hegger. Thin-walled shell structure made of textile reinforced con-
crete; Part I: structural design and construction. Structural Concrete, 16(1), 2015, DOI:
10.1002/suco.201300071 (in press).
[4] A. Scholzen, R. Chudoba, J. Hegger. Thin-walled shell structure made of textile reinforced con-
crete; Part II: experimental characterization, ultimate limit state assessment and numerical simula-
tion. Structural Concrete, 16(1), 2015, DOI: 10.1002/suco.201400046 (in press)
[5] Sharei, E.; Scholzen, A.; Chudoba, R.; Hegger, J.: Nonlinear analysis of textile reinforced con-
crete shells using an anisotropic, damage-based material model, In Brameshuber, W. (Ed.): 11th
International Symposium on Ferrocement FERRO-11 in connection with the 3rd ICTRC, Aachen,
Germany, 7–10 June 2015
[6] Contamine, R.; Si Larbi, A.; Hamelin, P.: Contribution to the direct testing of textile reinforced
concrete (TRC) composites. Materials Science and Engineering A, 2011, 528: 8589-8598
[7] Hartig, J.; Jesse, F.; Schicktanz, K.; Häußler-Combe, U.: Influence of experimental setups on the
apparent uniaxial tensile load-bearing capacity of Textile Reinforced Concrete specimens. Materi-
als and Structures, 2012, 45(3): 433-446
166
ANALYSIS AND GEOMETRIC DESIGN OF SEALING JUNCTIONS

FOR PRECAST FERROCEMENT TANKS
Henry Hernández Sotomayor1, Hugo R. Wainshtok Rivas2, Débora Acosta3, 1Instituto de

Investigaciones Porcinas (IIP), carretera Guatao Km 1 ½, Punta Brava, La Lisa, La Habana, Cuba,
2
Instituto Superior Politécnico José Antonio Echeverría (CUJAE), 1hhernandez@iip.co.cu,
2
hugow@tesla.cujae.edu, 2cudacostaa@civil.cujae.edu.cu2
Abstract: In the paper the geometric design of pre-cast ferrocement deposits sealing joint
is done, where the central vertical joints between two wall border panels and the vertical
stiffener called counter fort is analyzed, corner joints between two wall panels and a
vertical edges stiffener with a 45 degrees angle and the joint between the wall elements
and bottom slab. A variation in height was taken from 1 to 2m, obtaining a joint design for
each depth; these are evaluated and designed structurally, where cracking is checked from
the allowable stress and parameters to be met by ferrocement structures. Specific Surface
Reinforcement (Sr), which spans from (0.5 to 2 cm-1) and the Volume Fraction of
Reinforcement (Vf), which is between (1 and 8%). Once designed geometrically and
structurally joints at different depths, these are evaluated and compared with each other to
obtain a unique design that is valid for any of the evaluated depths and can provide a
rational design.
INTRODUCTION
Nowadays in the development of mankind the evolution of construction technology plays a
key role, the most widely used materials are concrete and steel, both precast concrete and
in site casted concrete. Today we face rationalization problems and cost savings in engi-
neering works, the introduction of thin-walled structures built with ferrocement, may be
the answer to these questions. Theoretical and experimental investigations carried out in
connection with ferrocement as a building material, confirms the great economy and the
high degree of safety and durability offered by the ferrocement structures. Using of the
advantages of this material, in this work the analysis and design of the sealing joint for
precast ferrocement deposits is proposed.
DEFINITION OF DEPOSIT
Deposits are engineering works for the storage of liquids or gases, whether for industrial,
domestic or recreational uses. They are classified according to their use in:
• tanks
• cisterns
• pools
The latter is the study subject of this investigation.
167
Ferrocement pools in Cuba.

The first ferrocement pool built in the country was the pool located in Jibacoa Beach in
Santa Cruz del Norte, in 1983.From that time on in the country began to be widely used
this material to build countless pools with various shapes and depths, with excellent results
and minimal maintenance. The pools built in Cuba with ferrocement have small wall
thicknesses ranging between 20 and 25mm for pools deeper than 2m.
MAIN EKUATIONS USED
M
σ ft = , Acting Tension (not bigger tha n 5 MPa.)
W …………………………………………1
Mu = 0,8R /b* ⋅ b ⋅ h2 ⋅ k (1 − 0,4k ) , Moment Equation, LSM…………………..…..……2
σ
W = η mϕ m S , Crack width (not bigger than 0.050 mm)…………………………..3
1E
m
Vre inf orcement

Vr = , Volume Fraction . (1% - 8%)
Vcomposite …………………………………..…..4
S reinforcem ent surface

Sr = , Specífic Surface . (0,5 cm - 1 - 2cm - 1 )
Vcomposite ……………………...…5
ANALYSIS AND DESIGN OF THE JOINT
The engineer often faces problems that cannot be addressed by classical methods, hence
they need to simplify and optimize, then modeling becomes the choice for solving a com-
plex engineering problem, usually modeling includes the following:
• A physical modeling includes: defining the geometry, material, boundary conditions
(supporting) and the loads to which they will be subjected.
• A mathematical model which must establish equations governing the behavior of the
joints.
Among the methods used more frequently used the following can be mentioned:
¾ The Finite Difference Method (FDM)
¾ The Method of Boundary or border
¾ The Finite Element Method (FEM).
• The Finite Element Method is used to predict the behavior of objects under the action of
physical phenomena such as mechanical tensional states,(Stress Analysis), mechanical
vibrations, heat transfer, fluid mechanics, electric and magnetic phenomena and also
acoustics ,FE M consists in dividing the object of study in a finite number of elements that
are connected together by points called nodes, where the degrees of freedom are defined,
168
resulting in what is known as a mesh; its behavior is governed by a finite number of pa-
rameters. The accuracy achieved reached in the results is determined by the maximum
number of degrees of freedom that the mesh is capable of assimilating in the analysis, and
the computing power that can be used.
Understanding the behavior of laminar structures can also be approached by FEM. In re-
cent years an increase of modeling programs available in the market, analysis and design
of such structures. Among these stands the Autodesk software Robot®™Structural Analy-
sis Professional, and its 2012 version was used throughout this work. In the modeling sys-
tem there are variables considered as independent: load, geometric characteristics such as
the dimensions of the panels and the counterforts, the material properties (soil filler and
ferrocement) and the boundary conditions. As dependent variables we have deformation,
internal shear stresses, bending moment and local stress due to bending.
Fig. 1 Panels
Fig. 2 Counterforts.
Figure 1 and 2: Schemes panels and counterforts
These dimensions are selected because they are easy to handle requiring only 3 or 4 men.
Three heights of elements are analyzed in such a way as to cover the standard heights for
Olympic and recreational swimming pools.
In order to determine the acting internal loads acting on the panels and the counterforts the
Rankine method was chosen, as floor wall friction is neglected, as filling material frictional
soil is used, which has a higher density and a smaller internal friction angle and will there-
fore generate more thrust on the panels which the worst case scenario, establishing the 2
charge states can act in the pool.
Load state 1: Appears for the empty pool, when only the push of the terrain surrounding
the pool is acting without the hydrostatic load compensating, analyzing the two states of
lateral pressure that could appear in this case, (active and passive states)
169
Fig. 3 Load state 1.
Load state 2: Occurs when the pool is full, and then the hydrostatic pressure of the water
acts in the opposite direction compensating the push of the surrounding terrain, thus dimin-
ishing considerably the strain on the walls.
Fig. 4 Load state 2
Being the load state1 the worst 1, this is the one to be used in the design as stresses in load
case 2 are smaller.
Table 1. Push existing by depth.

Height (m) Active push (kN/m2) Repose push (kN/m2) Hydrostatic push
(kN/m2)
1 8,24 11,72 10
1,5 11,89 16,92 15
2 15,55 22,11 20
170
Fig. 5. Thrust variation vs. height.
Once concluded the study of the soil for the different heights, an analysis of the thrust vari-
ation vs depth was carried out. The parameters for the meshes to be used in the modeling
were established, introducing the physical properties of ferrocement, the boundary condi-
tions are also set, and the loads that the panels are to bear, thus obtaining the bending mo-
ment, and the shear and tensions acting on the sealing joints.
Fig. 6 Parameters of the mesh selection.
171
Fig. 7 Meshing and load graphic for a height of 1,5 m.
For a height of 1,50 m.
Fig. 8 Graphic of bending Moment M, in X direction X, a) Front view, b) 3D, c) lateral view, d) Scale
172
Fig. 9 Graphic of tension σ in X direction: a) Front view, b) 3D, c) lateral view, d) Scale.
Fig. 10 Shear force graphic V in X direction: a) Front view, b) 3D, c) lateral view, d) Scale.
173
Panel deformation.
Fig. 11 Bending Scheme for a height of 1.50 m
Junction types
Taking into account mainly constructive advantages 2 types of joints are analyzed, as
shown in the following figure .In all cases the element thickness is taken as 3 cm
Fig. 12 Type I Section. First criteria: The junction between panels and counterforts are T shaped
Fig. 13 Section type II. Second criteria: T shaped Junction, softening the section transition of the panels to
the counterfort with a 45º angle
174
Dimensions
a=10cm b=12cm c=15cm x=3cm y=5cm
a, b y c: correspond to the overlapping lengths
x; y: Gusset.
Table 2: Modeling Results

Depth(m) No Type Sections De- Area M(kNm) σ(MPa)
W( )
nominations
( )
1 1 Ia 60.00 109.84848 0.2720 2.47583
1.5 2 Ia 0.7920 7.20993
2.0 3 Ia 1.5807 14.38967
1 4 Ib 72.00 150.92308 0.2720 1.80202
1.5 5 Type I Ib 0.7920 5.24771
2.0 6 Ib 1.5807 10.47344
1 7 Ic 90.00 225.00000 0.2720 1.20874
1.5 8 Ic 0.7920 3.52000
2.0 9 Ic 1.5807 7.02526
1 10 IIax 69.00 109.62760 0.2720 2.48082
1.5 11 IIax 0.7920 7.22446
2.0 12 IIax 1.5807 14.41866
1 13 IIay 72.25 109.96035 0.2720 2.47332
1.5 14 IIay 0.7920 7.20260
2.0 15 IIay 1.5807 14.37503
1 16 IIbx 81.00 150.52247 0.2720 1.80682
1.5 17 IIbx 0.7920 5.26167
2.0 18 IIbx 1.5807 10.50131
1 19 Type II IIby 92.50 151.61390 0.2720 1.79381
1.5 20 IIby 0.7920 5.22380
2.0 21 IIby 1.5807 10.42572
1 22 IIcx 99.00 224.69776 0.2720 1.21037
1.5 23 IIcx 0.7920 3.52473
2.0 24 IIcx 1.5807 7.03471
1 25 IIcy 115.00 225.34886 0.2720 1.20687
1.5 26 IIcy 0.7920 3.51455
2.0 27 IIcy 1.5807 7.01438
175
Fig. 14 Tension [MPa] before the section cracks. (Type I)
Fig. 15 Tension [MPa] before the section cracks. (Type II).
Analyzing the behavior of the tensions before the junction shape is modified it was shown
that:
1. The higher tension values were obtained for type I.
2. For an equal height of the structural elements the tension values show a better behavior
in type II, as it has a better load dissipation.
3. For a 2m depth, neither type can be able to reduce the acting tension to a lesser value
than theadmissible for cracking (7MPa). Seer Fig. 16
Fig. 16 Tension ( ) by don´t crack sections.
176
Optimal section selection for each height.

Type I: For 1and 1.5 m height the best type is the Ib
Type II: For a 1and 1.5 m height it is better to use the Iibx
In order to determine the optimal section for a height of two m, the tension was fixed in 7
MPa, clearing from the formula in Strength of Materials equation, we can figure out the
Section Module or bending resistance factor the required section.
A section Module increase can be achieved increasing the thickness of the elements or the
overlapping lengths. Assuming 20 cm overlap and using the 45 degrees gusset transition
with 5 cm, the following section is obtained.
Fig. 17 Section for 2,0 m height.
Table 3 Sections date by 2,0m height
Depth (m) No Sections De- σ(MPa)

Area ( ) W( )
nomination
2.00 28 IId 145.00 381.92 4.13882
The following summarizes the characteristics of the best sections for each height.
Table 4 Characteristics of selected sections.
Depth (m) No Sections De- Area M(kNm) σ(MPa)

W( )
nomination ( )
1.00 4 Ib 72.00 150.92308 0.2720 1.80202
1.50 5 0.7920 5.24771
1.00 16 IIbx 81.00 150.52247 0.2720 1.80682
1.50 17 0.7920 5.26167
177
Joint between wall panels and bottom slab

For the design of the joint between the wall panels forming the bottom slab, this will have
a splicing length of 10 cm because the bottom of the pool simply acts as an impermeable
layer; so between the wall panel and the bottom of the pool, a1m slope conformer at 45
degrees will be the transition between bottom and wall, in order to dissipate the lateral
thrust and the tensions caused by the change in section, therefore the junction will be con-
structive and not structural.
CONCLUSIONS
1. No previous work on the study case of junctions for precast ferrocement tanks.
2. After the analysis of the structural behavior of the types of sealing junctions for pre-
cast deposits construction, we can assure that for lesser depths tan 2 m the suitable
types of junctions are Ib and IIbx, because they provide a better performance and re-
liability.
3. From 2 m to 3 m the recommended junction is type II 20 x 20cm with a 5cm gusset.
LITERATURE
1. A.S.Kalmanok, “Manual para cálculo de placas, 1994.
2. Braja M. DAS, Libro: Fundamentos de Ingeniería Geotécnica. International Tomson Editores, 2001.
3. Wainshtok Rivas. Hugo R, Libro: Ferrocemento Diseño y Construcción del Editorial: Offset Abad Cía.
Ltda., 3era edición – 1998. Guayaquil – Ecuador.
4. http://www.canalconstruccion.comferrocemento.com, Galiano Ortiz. Amado F, Artículo: El Ferrocemento
y sus Posibilidades para el Desarrollo de las Comunidades Pesqueras para el Continente, Empresa de Proyec-
tos Navales (CEPRONA).
5. http://www.canalconstruccion.comferrocemento.com,Wainshtok Rivas. Hugo R, Artículo: Ferrocemento
en Cuba.
6. http://www.canalconstruccion.comferrocemento.com, Artículo: Guía de Construcción para Estructuras de
Ferrocemento, publicado en el 2003.
7. http://www.canalconstruccion.comferrocemento.com, Artículo: Tesina de un colombiano, publicado en el
2009.
8. http://190099-piscina-olímpica-de-pekin-2008-water-cube.com
9. http://www.canalconstruccion.comferrocemento.com
10. http://www.construnario.com
11. http://www.construnario.com/diccionario_de_la_construccion
12. http://Calculo de Piscinas Semi Olímpicas General Ingeniería Civil.com
13. http://Cubo+de+Agua+en+Beijing+de+piscina+olimpica+a+parque+acuatico.com
14. http://¿Cuánto mide una piscina olímpica-Saberia.com
15. http://www.construcción de piscinas olímpicas, las más grandes en el rubro.com
16. http://www.EL DEPORTE Y LA CULTURA EN DUITAMA.com
17. http://ferrocemento,com
178
18. http://FERROCEM Cia. Ltda. Prefabricados y Construcciones en Ferrocemento y derivados. Bahía de

Caráquez – Ecuador.com
19.http://www.file:///E:/TESIS/DOCUMENTOS%20INTERNET-MARZO-
2011/tematicos_res%20acero%20galavanizadoultado.asp
20. http://www.h_natacion.com
21. http://ISon21 » Blog Archive » Watercube La piscina olímpica que nació de una pompa de jabón.com
22. http://inciarco.info/comunidades/showthread.php?p=2482.com
23. http://www.mioruro.com/libros/arquitectura/ferrocemento/ferrocemento.com
24. http://piscinas-olimpicas_4917.com
25. http://Piscinas_Bernat_Picornell.com
26. http://www.Partes-piscina-olimpica.com
27. http://www.Piscinas olímpicas. medidas, información semi olimpicas.com
28. http://partes de una piscina olímpica.com
29. http://www. Piscina - Wikipedia, la enciclopedia libre.com
30. http://quito-estrenara-piscina-olimpica-en-2004-161982-161982.com
31. http://www.sitioferrocemento.com/pagina0003.php
32. http://sitioferrocemento.com/blog/ferrocemento/
33. http://www.Tipos de piscinas y materiales de construcción de piscina, conveniencias y fabricacion para
cada necesidad.mht.com
34. http://Tipos de piscinas y materiales de construcción de piscina, conveniencias y fabricacion para cada
necesidad.com.
179
180
ELASTIC-PLASTIC ANALYSIS OF I- SHAPED NORMAL

STRENGTH FERROCEMENTBEAMS
Azad A. Mohammed is with Civil Engineering Department, Faculty of Engineering, University of

Sulaimani, Kurdistan Region / Iraq.(e-mail:azadkadr@yahoo.com)
Abstract― In this paper flexural analysis of ferrocement I-section beam is presented. Elas-
tic- plastic behavior of ferrocement constituent materials was incorporated in the analysis.
The analysis is based on deriving moment- curvature relationship obtained from the con-
cept of equilibrium of forces and compatibility of strains. Later from the deflection equa-
tion of elastic beam, complete load- deflection response was obtained. Comparison with
the test data on load- deflection relationship of ferrocement I- section beam showed a good
agreement. A simple method for calculating ultimate load capacity of I- section ferroce-
ment beam was proposed. The calculated ultimate load was found to be accurate when
compared with the available test data of the average test/ calculated ratio equal to 1.05.
NOMENCLATURE
a shear span
AS area of steel reinforcement in tension zone
AS' area of steel reinforcement in compression zone
bf width of beam section
c depth of compressive zone
C compressive force
d' distance of steel reinforcement from extreme fiber in tension or in compression
D flexural rigidity
Ec elastic modulus of ferrocement
Ef elastic modulus of ferrocement wire mesh
Em elastic modulus of mortar
Es elastic modulus of steel
Fc compressive force acting on section
Ft tensile force acting on section
fc compressive stress in general
ft tensile stress in general
ftm cracking stress of mortar
ftul ultimate tensile stress in longitudinal direction
ftut ultimate tensile stress in transverse direction
h height of beam section
ht depth of elastic zone in tension
hep depth of elastic- plastic zone in tension
hc depth of elastic zone in compression
Me moment acting on elastic section
181
Mep moment acting on elastic section

tf flange thickness
tw web thickness
Vf ratio of ferrocement wire mesh in the section
α compressive stress reduction factor
εc compressive strain in general
εo strain at peak compressive stress
εt tensile strain in general
εtm cracking strain
εtul strain related to ftul
εtut strain related to ftut
Φ curvature
ψ slope of tensile stress- strain relationship in longitudinal direction
INTRODUCTION
Ferrocement considered as one of the nonconventional types of concrete known by its high
strength to weight ratio and superior crack control. Structural applications of ferrocement
can be seen in different type of shell structures, small water tanks, silos and in different
types of decoration purposes suitable for architectural purposes.Toward more extended us-
ing of ferrocement, researches [1,2] were carried out for gaining information about the suit-
ability of hollow cores and sandwich panels. Other researches[3,4] focused to check the
suitability of I-section ferrocement beams. In low cost housing projects different types of
ferrocement beams have a promised future due to the acceptable economical cost.
The present study aims to learn more about I-beam ferrocement beam through studying
strength and deformation of such types of beams. Ferrocement beams to be analyzed rein-
forced with skeletal steel bars provided to the flanges. Flexural analysis lead to calculating
the complete load- deflection response is carried out in which a nonlinear ferrocement mate-
rials behavior fairly incorporated in the analysis. Later, an equation for calculating ultimate
load capacity obtained from the results of flexural analysis is proposed. Later, predictions of
load- deflection relationship and ultimate load capacity are compared with test results of I-
shaped ferrocement beams to check the accuracy of predictions.
FLEXURAL ANALYSIS
A. Moment- Curvature Relationship

A.1 Materials Stress- Strain Relationship
For ferrocement I - section deriving moment - curvature relationship is more complex

compared with rectangular section, because there is different section widths for different
cracking stages. Furthermore, the case approaches that of composite section if the ratio of
182
ferrocement wires provided to the web is different from that provided to the flange. Other
complexity is in the existence of skeletal reinforcement if provided to the flanges.
Fig.1 shows the idealized compressive stress-strain relationship for ferrocement. This
shape of relationship first used by Walraven[5], in which the actual curve is approximated
by the bilinear one and the effect of strain softening in compression is neglected. It is as-
sumed that the elastic modulus of ferrocement in compression is equal to that in tension.
Ec is the elastic modulus of the composite given by[6]
Ec = Em(1 + Vfn) -----------------------------------------(1)
Em is the elastic modulus of the matrix, can be calculated from the equation given be ACI
Code as follows
Em = 4730 fc ' ------------------------------------------(2)
and n is the ratio of elastic modulus of reinforcement to that of matrix ( Ef/Em ), and Vf is
the ratio of wire mesh reinforcement in the section given by
Af
Vf = ----------------------------------------------------(3)
bh
α is compressive stress reduction factor taken as 0.85. εo is strain corresponding to αfʹc

given by
αf ' c
εo = --------------------------------------------------(4)
Ec
Fig. 2 shows the idealized tensile stress-strain relationship for ferrocement. At the absence
of test data, the cracking stress of matrix ftm can be calculated from the following expres-
sion given by Oluokun[7]
0. 7
ftm = 0.2( f c′ ) --------------------------------------------(5)
After that matrix cracks, the tensile strength of ferrocement ftu is that stress obtained from
the result of dividing the force carried by wire mesh by the section area. The cross section-
al area vanishes from the relationship and ftu becomes
ftu = Vf fy --------------------------------------------------(6a)
It should be noted that in the case of different wire mesh properties in the web and the
flange, there is different values of ftu and Vf. Also fyf may also be different and such case
is similar to that of composite section.
Hence, there are longitudinal direction (flange) and transverse direction (web). For the
flange ftu becomes ftul and is calculated as follows
183
ftul = Vfl fyl -----------------------------------------------(6b)
For the web the tensile strength is ftut and given by
ftut = Vft fyt -------------------------------------------------(6c)
Similarly, instead of strain corresponding to ftu ( i.e. εtu ) there are εtul and εtut. The later
two values are obtained from dividing the stress by elastic modulus, and can be obtained
from the following relationships
ftul
εtul = ------------------------------------------------------(7a)
Ef
ftut
εtut = ------------------------------------------------------(7b)
Ef
Fig. 1 Compressive stress- strain relationship of ferrocement
Fig. 2 Tensile stress- strain relationship of ferrocement
In addition to the above simplifications for the materials further assumptions should be
made for facilitate the analysis
184
i- plane sections of a beam remain plane after bending and accordingly there is a linear
strain distribution.
ii- failure mode of a beam is flexure, and other types of failure like shear failure will not
occur.
iii- there is a complete bond between cement mortar and all types of reinforcement in the
ferrocement section, and
iv- local buckling of the section web and flange will not occur.
A.2. Moment- Curvature Relationship for Elastic Stage
For this stage the curvature is small. Ferrocement materials in compression and tension
behave elastically. Strain and stress distribution for this stage are as shown in Fig. 3a. From
equilibrium of forces acting on section after integrating a small portion of force the follow-
ing relationships for compressive and tensile forces are obtained
c − tf c
C = tw
∫0 ∫
EcΦydy + bf EcΦydy + As ' EsΦ (c − d ' ) -------------(8a)
c − tf
h −c −tf h −c
T = tw
∫0 ∫
EcΦydy + bf EcΦydy + AsEsΦ (h − c − d ' ) -------(8b)
h −c −tf
On equating the two forces and simplifying on obtains
Ef
( h − 2tf )(bf − tw) h − bfh 2 − 2 [ As ( h − d ' ) + As ' d ' ]
c= Ec
Es
2[(h − tf )(bf − tw) − bf ( h + tf ) + tftw − ( As + As ' )]
Ec --------(8c)
For the elastic stage moment- curvature relationship after integrating a moment of small
portion of force about neutral axis lead to the following relationship
Me = tw ∫0c−tf EcΦy 2 dy + bf ∫cc−tf EcΦy 2 dy + As ' EsΦ (c − d ' ) 2 +
tw ∫0h−c−tf EcΦy 2 dy + bf ∫hh−−cc−tf EcΦy 2 + AsEsΦ (h − c − d ' ) 2 − − − −(9a )
On simplifying one obtains
M = m1c3+m2c2+m3c+m4 ---------------------------------------------------(9b)
m1, m2, m3, and m4 are moment parameters given by
185
2 EcΦbf
m1 =
3
m 2 = EcΦ (tw − bf )(h − tf ) − EcΦtwtf + EsΦ ( As + As ' )
m3 = EcΦtwtf 2 − EcΦ (tw − bf )(h − tf ) 2 − 2 EsΦh
EcΦ EcΦtwtf 3
m4 = (tw − bf )(h − tf ) 3 − + EsΦ ( As (h − d ' ) 2 + As ' d ' )
3 3
This stage is valid for all curvature value smaller that that makes from tensile strain reach-
es εtm at which the section cracks indicating the end of the elastic stage. The corresponding
curvature is
εtm
Φe =
(h − c) ----------------------------------------------------------------(10)
A.3. Moment- Curvature Relationship for Elastic- Plastic Stage(1)
Strain and stress distribution for this stage are as shown in Fig. 3b. Compressive force act-
ing on the section is as given by Eq. 8a. Tensile force for this stage is given by
T = tw ∫0h−c−tf EcΦydy + bf ∫hht−c −tf EcΦydy + bf ∫hth−c [ ftm + ψ (Φy − εtm)]dy
+ AsEsΦ(h − d '−c) − − − − − − − − − − − − − − − − − − − − − − − − − −(11a)
Equating the two forces and simplifying, depth of compression zone can be calculated as
follows
c = f 1c 2 + f 2c + f 3 ----------------------------------------------------------(11b)
The force parameters f1, f2 and f3 for this stage is as follows
bfΦ
f1= (ψ − Ec )
2
f 2 = EcΦ(h − tf )(bf − tw) − ftmbf + bfψ (εtm − Φh) + EcΦtf (tw − bf )

− EsΦ( As + As ' )
E cΦ h Ecbfεtm 2 b fψ Φ 2
f3= (tw − bf )( h − 2tf ) + + ftmbfh + h
2 2Φ 2
εtm ftmbfεtm
+ bfψε tm ( − h) − + EsΦ[ As ( h186
− d ' ) + As ' d ' ]
2Φ Φ
For this stage the moment- curvature relationship is calculated from an equation similar to
Eq.9b with the following moment parameters
Φbf
m1 = ( Ec − ψ )
3
ftmbf εtm
m 2 = EcΦ (tw − bf )(h − tf ) + + b fψ ( Φ h − ) + EcΦtf (bf − tw)
2 2
+ EsΦ ( As '+ As )
m3 = − EcΦ (tw − bf )( h − tf ) 2 − bfftmh + bfψh(εtm − Φh)
− EcΦtf 2 (bf − tw) − 2 EsΦ[ As ' d '+ As ( h − d ' )]
EcΦ bfh 2 bfψΦh 3 bfftmεtm 2
m4 = (tw − bf )[(h − tf ) 3 − tf 3 ] + ( ftm − ψεtm ) + −
3 2 3 2Φ 2
bfεtm 3 ψ
+ 2
( E c+ ) + EsΦ[ As ' d ' 2 + As ( h − d ' ) 2 ]
3Φ 2
es εtm at the web face under tension . The corresponding curvature is
εtm
Φep1 =
( h − c − tf ) ----------------------------------------------(12)
(a) Elastic Stage
(b) Elastic- Plastic Stage(1)
(c) Elastic- Plastic Stage(2)
187
(d) Elastic- Plastic Stage(3) and

(4)
(e) Elastic- Plastic Stage(5)
Fig. 3 Stress and Strain Distributions across Ferrocement Section
Strain and stress distribution for this stage are as shown in Fig. 2c. For this stage, depth of
compression zone is calculated from Eq. 11b. The force constants f1, f2 and f3 for this stage
are as follows
tw Φ
f1= (ψt − Ec )
2
f 2 = ( EcΦtf +ψεtm − ftm)(tw − bf ) − Φ (h − tf )(twψt − bfψ ) − bfftm

+ bfψ (εtm − Φh) − EsΦ ( As + As ' )
E cΦ t f 2 Φ
f 3 = (bf − tw)( ) + ( h − tf ){(twψ t − bfψ )[ ( h − tf ) − εtm ] +
2 2
hΦ
ftm (tw − bf )} + EsΦ[ As ( h − d ' ) + As ' d ' ] + bfh[ ftm + ψ ( εtm )] +
2
twεtm Ecεtm εtm
+ ( + ψt − ftm )
Φ 2 2
For elastic-plastic stage(2), moment- curvature relationship yet can be calculated from
188
tw Φ
m1 = ( Ec − ψt )
3
tw
m 2 = EcΦtf (bf − tw) + EsΦ ( As '+ As ) + ( ftm − ψεtm ) + twψtΦ ( h − tf ) + bfψΦtf
2
m3 = EcΦtf 2 (tw − bf ) − 2 EsΦ[ As ' d '+ As ( h − d ' )] + ftm (bf − tw)( h − tf ) − bfftmh
+ bfψh(εtm − Φh) + (bfψ − twψt )(h − tf )[Φ ( h − tf ) − εtm ]
EcΦtf 3 twεtm 2 εtm ψt ftm
m4 = (bf − tw) + EsΦ[ As ' d ' 2 + As ( h − d ' ) 2 ] + 2
[ ( Ec + ) − ]
3 Φ 3 2 2
εtm b fh 2 bfψΦh 3 ftm
+ (bfψ − twψt )( h − tf ) + 2
( ftm − ψεtm ) + + (tw − bf )( h − tf ) 2
2 2 3 2
Φ
+ (twψt − bfψ )( h − tf ) 3
3
es εtul at the extreme fiber of the flange in tension. The corresponding curvature is
εtul
Φep 2 =
(h − c) ---------------------------------------------------(13)
Strain and stress distribution for this stage are as shown in Fig. 3c. For this stage depth of
are as follows
tw Φ
f1= (ψt − Ec )
2
f 2 = ( EcΦtf + ψεtm − ftm)(tw − bf ) − Φ (h − tf )(twψt − bfψ ) − bfftm

+ bfψ (εtm − Φh) − EsΦ ( As + As ' )
EcΦtf 2 Φ
f 3 = (bf − tw)( ) + ( h − tf ){(twψt − bfψ )[ ( h − tf ) − εtm ] + ftm (tw − bf )}
2 2
hΦ twεtm Ecεtm εtm
+ EsΦ[ As ' d '+ As ( h − d ' )] + bfh[ ftm + ψ ( − εtm )] + ( + ψt − ftm )
2 Φ 2 2
For elastic-plastic stage(3), moment- curvature relationship yet can be calculated from
189
twΦ
m1 = ( Ec − ψt )
3
ftmtw twψεtm
m 2 = EcΦtf (bf − tw) + EsΦ ( As + As ' ) + + twψ Φ ( h − tf ) − + bf ψ Φ t f
2 2
m3 = EcΦtf 2 (tw − bf ) − 2 EsΦ[ As (h − d ' ) + As ' d ' ) + ftm(bf − tw)(h − tf ) − bfftmh

+ bfψh(εtm − Φh) + (bfψ − twψt )(h − tf )[Φ(h − tf ) − εtm]
EcΦtf 3 twεtm 2 εtm ψt ftm

m4 = (bf − tw) + EsΦ[ As (h − d ' ) 2 + As ' d '2 ) + 2
[ ( Ec + ) − ]
3 Φ 3 2 2
εtm bf h 2 bfψΦh3 ftm
+ (bfψ − twψt )(h − tf ) 2 + ( ftm − ψεtm ) + + (tw − bf )(h − tf ) 2
2 2 3 2
Φ
+ (twψt − bfψ )(h − tf )3
3
This stage is valid for all curvature values smaller than that lead to yielding steel rein-
forcement provided to tension zone. Limit curvature for this case is
fy
Φep 3 =
( h − c − d ' ) ----------------------------------------------(14)
Strain and stress distribution for this stage are as shown in Fig. 3d. For this stage depth of
are as follows
tw Φ
f1= (ψt − Ec )
2
f 2 = ( EcΦtf +ψεtm − ftm)(tw − bf ) − Φ(h − tf )(twψt − bfψ ) − bfftm

+ bfψ (εtm − Φh) − EsΦAs '
EcΦtf 2 Φ
f 3 = (bf − tw)( ) + ( h − tf ){(twψt − bfψ )[ ( h − tf ) − εtm ] + ftm (tw − bf )}
2 2
hΦ twεtm Ecεtm εtm
+ EsΦAs ' d '+ Asfy ( h − d ' ) + bfh[ ftm + ψ ( − εtm )] + ( + ψt − ftm )
2 Φ 2 2
For elastic-plastic stage(4) moment- curvature relationship is calculated from Eq.9b with
the following moment parameters
190
twΦ
m1 = ( Ec − ψt )
3
ftmtw twψε tm
m 2 = EcΦ tf (bf − tw) + EsΦ As '+ + tw ψ Φ ( h − tf ) − + b fψ Φ tf
2 2
m3 = EcΦ tf 2 (tw − bf ) − 2 EsΦ As ' d '+ ftm (bf − tw)( h − tf ) − bfftmh − Asfy
+ bfψh(εtm − Φ h) + (bfψ − twψt )( h − tf )[Φ ( h − tf ) − εtm ]
E cΦ t f 3 twεtm 2 εtm ψt ftm
m4 = (b f − t w ) + E s Φ A s ' d ' 2 + 2
[ ( Ec + ) − ] + Asf y ( h − d ' )
3 Φ 3 2 2
εtm b fh 2
b fψ Φ h 3 ftm
+ (bfψ − twψt )( h − tf ) 2 + ( ftm − ψε tm ) + + (tw − bf )( h − tf ) 2
2 2 3 2
Φ
+ (twψt − bfψ )( h − tf )3
3
This stage is valid for all curvature value smaller than that makes from tensile strain reach-
es εtul at the face of web in tension. The limit curvature is
εtul
Φep 4 =
(h − c − d ' ) ----------------------------------------------(15)
Strain and stress distribution for this stage are as shown in Fig. 3e. For this stage depth of
are as follows
Φ
f1= − twEc
2
f 2 = (tw − bf ) EcΦtf − twftut − As ' EsΦ
Ectwεtm 2 twεtut εtut twεtm εtm
f3= + [ ftm + ψt ( − εtm)] − ( ftm − ψt )
2Φ Φ 2 Φ 2
εtut Φ
+ twftut (h − tf − ) + bfftultf + (bf − tw) Ec tf 2 + Asfy + As ' EsΦd '
Φ 2
For elastic-plastic stage(5) moment- curvature relationship is calculated from Eq.9b with
the following moment parameters
191
twEcΦ
m1 =
3
twftut
m 2 = (bf − tw) EcΦtf + + As ' EsΦ
2
m3 = EcΦtf 2 (tw − bf ) − 2 EsΦAs ' d ' 2 −( h − tf )twftut − bfftultf − Asfy
E cΦ t f 3 Ectwεtm 3 twεtm 2 ψtεtm
m4 = (bf − tw) + As ' EsΦd ' 2 + 2
+ A sfy ( h − d ' ) −
2
( ftm − )
3 3Φ 2Φ 3
twεtut ftm εtul εtm twftutεtut 2 twftut bfftultf
+ 2 ( +ψ ( − )− + ( h − t 2
f) + ( 2 h − tf )
Φ 2 3 2 2Φ 2 2 2
This stage is valid for all curvature values smaller than that lead to yielding steel rein-
forcement provided to tension zone. Limit curvature for this case is
0.85 f ' c
Φep5 =
Ecc -------------------------------------------------(16)
It should be noted that there are other cracking stages followed the elastic- plastic stage(5).
These stages related to the plastic compressive stress- strain relationship (Fig. 1) and yield-
ing of steel reinforcement in compression. Such stages are not important, because they
affect no the ultimate load capacity and used only to obtain a deflection for a nearly con-
stant load to draw the complete load- deflection response reached by test measurements. In
the present study an analysis for moment- curvature relationship are presented for the sim-
plicity sake.
B. Load- Deflection Relationship
For deriving complete load- deflection response of ferrocement I- shaped beam there is a
need for deflection equation. Deflection equation of elastic beam for any load is available
in the theory of plate textbooks. For deriving load- deflection relationship it is assumed
that the equation is valid for all cracking stages and the effect of cracking is incorporated
via the change in flexural stiffness only. Therefore, in the present study only the material
nonlinearity behavior is incorporated to the analysis.
In order to check the accuracy of the analysis the predictions must be compared with the
test data. Test data on ferrocement I- section beams are available, in which the beams were
tested under two central point loads[4]. Therefore, the analysis for such types of beams are
presented here for the comparison sake.
For a ferrocement beam subjected to two central points load each P/2 with a shear span a
the maximum deflection at the center is given by
192
Pa
ω=
48D(4a 2 − 3L2 ) ----------------------------------------(17)
Differentiating the basic deflection equation of the beam two times lead to the following
equation for load - curvature relationship
2 DΦ
P=
a ----------------------------------------------------(18)
Where D is the value of flexural rigidity of the beam section which varies with the varia-
tion of curvature for different cracking stages.
VALIDITY OF THE ANALYSIS
A series of ferrocement I- shaped beams were tested for the load- deflection relationship by
Mansur and Ong[4]. In their published paper load- deflection relationship exists for four
beams ( Series B ) in which two beams were failed in shear. Here comparison between
predicted load- deflection relationship and test load- deflection relationship are made for
those beams failed in flexure, because the present model of analysis based on the fact that
the beam is strong in shear and failed in flexure.
Fig.3 shows the two load- deflection relationships for those beams failed in flexure ( Beam
B2.0 and Beam B2.5 ). The limits of cracking stages are illustrated in the figure. It is clear-
ly shown that the predicted load- deflection relationship moderately underestimates test
data. Comparison also indicates that the calculated deflection corresponding to ultimate
load is considerably lower than that of test one for the two beams. The difference between
the two deflections is quite low for elastic stage but later the difference becomes higher.
The reason for this appears in the use of constant deflection equation for a ferrocement
beam for all stages of cracking. Therefore, material nonlinearity to the analysis through
using deteriorated flexural rigidity is not sufficient to obtain accurate prediction. In the best
of author’s idea there is a need for incorporating the effect of geometrical nonlinearity into
the analysis to obtain better results. Such effect can be solved by incorporating the role of
plastic hinge rotation to the basic deflection equation used. The author thinks that better
results for the whole load- deflection relationship can be obtained. Here, attempts are not
made to solve such problem and can be done separately in another study.
ULTIMATE LOAD CAPACITY
In the following section a simplified model for calculating ultimate load capacity is pre-
sented based of the proposed flexural analysis. The last cracking stage is analyzed for the
ultimate load capacity. For such stage the ratio of depth of elastic zone, ht , ( Fig. 2e ) to
the section depth is close to 1%.
193
Fig.3 Test and calculated load- deflection relationship
Such depth is quite small and can be neglected for drawing a stress distribution at ultimate
load. The ratio of depth of elastic- plastic zone , hep, is close to 20% and for simplicity this
depth is also neglected. It was found from flexural analysis that the depth of compression
zone is close to the flange depth, tf, and it is assumed here that such depth is equal to tf for
all wire mesh ratios and skeletal steel reinforcement in the flanges. It is also found that the
compressive stress in steel reinforcement is close to 0.5As’fy and tensile stress in wire
mesh is equal to ftul. Based on such approximation the stress distribution at ultimate load
capacity is as shown in Fig.4. From the given stress distribution the moment capacity of
the ferrocement I- section is given by
αf ' cbftf 2 3
Mn = + 0.5 A' sfy (tf − d ' ) + Asfy (h − tf − d ' ) + ftulbftf (h − tf )
3 2
ftultw
+ (h − 2tf ) 2 − − − − − − − − − − − − − − − − − − − − − − − − − (19)
2
194
Fig.4 Stress distribution for Ultimate Load Capacity
In order to check the accuracy of the proposed model for calculating ultimate load capacity
the predictions should be compared with the test data available in the literature. Here, test
data reported by Mansur and Ong[9] are used for comparison sake. Table I shows the re-
sults of test and calculated ultimate load capacity of ferrocement I- section beams. From
the obtained results one can find that the calculated ultimate load capacity moderately un-
derestimates test data of reasonable factor of safety for all beams. Therefore the case is
similar to the results of rigorous flexural analysis presented in this study. However; the
benefit of the proposed model for calculating ultimate load capacity of ferrocement I-
shaped beams is its simplicity for calculation and can be followed for design of such types
of beams.
CONCLUSIONS
From the research study outlined in this paper the following conclusions can be drawn:
1- Using the concept of equilibrium of forces and strain compatibility combined with the
deflection equation of elastic beam able to obtain a model for analysis of ferrocement I-
shaped beams for the complete load- deflection relationship.
2- There may a problem in using constant deflection equation for elastic beam for all stag-
es of cracking and incorporating the deteriorated flexural rigidity is not sufficient for ob-
tain the accurate deflection values for those load values related to elastic- plastic and plas-
tic stages.
195
REFERENCES
[1] S. L. Lee, M. A. Mansur, P. Paramasivam, K. C. G. Ong, and C. T. Tam, Study of Sandwich Wall
Panels, J. Ferrocement (Bangkok), vol.16, no.3, pp. 295-313, 1986.
[2] M. Sarid, E. Z. Testa, and F. Bljuger, Ribbed Slabs made of Ferrocement, J. Ferrocement (Bangkok),
vol.9, no.4, pp. 185-189, 1979.
[3] G. J. Al Sulaimani, and S.F. Ahmad, Deflection and Flexural Rigidity of Ferrocement I- beams, J.
Ferrocement (Bangkok), vol.18, no.1, pp. 1-12, 1988
[4] M. A. Mansur, and C. G. Ong, Shear Strength of Ferrocement I- beams, ACI Structural Journal, vol.
88, no. 4, pp.458-464, 1991.
[5] J. C. Walraven ,and S. E. J. Spierenburg, Behavior of Ferrocement with Chicken Wire Mesh Rein-
forcement, J. of Ferrocement, vol. 15 ,no. 1, pp.3-13, 1985.
[6] S. P. Shah, and P. N. Baluguru, Ferrocement, New Reinforced Concrete , R. N. Swamy , Ed., Surrey
Univ. Press, UK, pp. 1-51A, 1984.
[7] F. A. Oluokun, Prediction of Concrete Tensile Strength from Compressive Strength: Evaluation of Ex-
isting Relations for normal weight concrete, ACI Materials Journal, vol. 88, no. 3, pp. 302-309, 1991.
[8] S. P. Timoshenko, and S. Woinowsky- Krteger, Theory of plates and shells. New York, McGraw-
Hill, 1959
196
TableI Test and calculated ultimate load capacity of ferrocement I- section beam
Beam Pn,test Pn,calc. Pn, test

Pn, calc.
A2.0 60.5 55.67 1.13
A2.5 48.4 44.54 1.09
B2.0 58.4 50.72 1.15
B2.5 46.4 40.57 1.14
C1.5 69.8 52.56 1.33
C2.0 52.1 39.42 1.32
C2.5 39.4 31.54 1.25
D2.0 64.0 56.29 1.14
D2.5 50.4 45.04 1.12
E2.0 80.3 73.74 1.09
E2.5 64.8 58.99 1.10
F1.5 91.5 83.26 1.10
F2.0 72.8 62.44 1.17
F2.5 53.0 49.95 1.06
G1.5 98.0 77.42 1.27
G2.5 57.5 46.45 1.24
Mean 1.17
197
198
STUDY OF EARTHQUAKE RESISTANT BEHAVIOR IN

PREFABRICATED FERROCEMENT OF DEPOSIT. CALCULATION
OF MATERIALS.
Henry Hernández Sotomayor1, Hugo R. Wainshtok Rivas2, 1Instituto de Investigaciones Porcinas

(IIP), carretera Guatao Km 1 ½, Punta Brava, La Lisa, La Habana, Cuba, 2Instituto Superior
Politécnico José Antonio Echeverría (CUJAE), 1hhernandez@iip.co.cu, 2hugow@tesla.cujae.edu.cu
Abstract: The paper presents the study of the earthquake-resistance of a precast ferroce-
ment Olympic pool. The structural behavior is analyzed under various loading conditions
that can act in such a structure. In order to determine the lateral terrain pressure the Ran-
kine method is applied, because the friction between the wall and the soil is neglected.
Having assessed the state of the acting forces, the design is made for the worst case, after-
wards cracking and bending are checked, and the parameters of Specific Surface Rein-
forcement(Sr), which is between (0.5cm-1 -3cm-1) and Reinforcement Volume Factor(Vf),
ranging from 1% to8%, conditions that must be met by ferrocement structures. After the
pool was designed, the amount of material to be used was calculated, and then compared
with the reinforced concrete Olympic pool built in Boyaca, Colombia, for the Pan Ameri-
can games in 2006 taking into consideration only the material cost.
INTRODUCTION
Within the entire set of problems in the economy constructions, play an important troll e,
as they are one of the main branches in the development of a country. They guarantee the
growth rates of national GDP, creating new productive capacities in all branches of the
economy and the expansion of capacity inn on productive areas, as well as in the im-
provement of working conditions and the life of the population.
Currently, among all building materials most used is concrete, both precast and cast in situ,
as it meets the basic requirements for industrial and civil construction in general The gen-
eral trend in the use of reinforced concrete consists in the development of an active and
deep work to improve their technical-economic indicators, as well as the creation and im-
plementation of, new structural elements which are lighter, cheaper, and all on higher per-
centage of industrial production, and have intrinsic better safety and durability. The solu-
tion of these tasks is in perfect correspondence with the introduction in everyday practice
of construction, the use of thin-walled structures built with ferrocement.
Historically ferrocement was invented first than reinforced concrete, and today they differ
mainly in the scale. One of them utilizes bigger diameter reinforced steel rods, and coarse
aggregates to form the matrix, as is the case of reinforced concrete ,while ferrocement uses
fine aggregates and wire, in this case in meshes either waved or welded,
Theoretical and experiment all investigations related to ferrocement as a building material,

experiences in the design and construction of structures of this material for many different
types of industrial and social buildings, ships, tanks, bridges, swimming pools, etc., and the
199
years in the and utilization of ferrocement in many countries, confirm the great economic
advantage of using ferrocement structures and the aright record of safety, reliability and
durability.
This papery a about the design of a pool of ferrocement, a very convenient material, as its
essentially advantageous in spatial structures of thin walls, where the stiffness and strength
are developed through form, having the advantage of being malleable and built in one
piece.
ADVANTAGES AND DISADVANTAGES TO SEISMIC ACTIONS

¾ Advantages
• Structures with high internal damping.
• Adequate control horizontal deformation
• Structures with a large reserve of energy that allow for structural recovery even af-
ter being subjected to severe actions. Also, very easy to repair frequently
• They are generally low period’s structures with little influence of the side effects P
-Δ.
• Their low vibration period allows structures make them affect only slightly emo-
tional behavior of individuals.
• While the specific gravity of ferrocement is similar to that of concrete, it stops be-
ing critical of a suitable dimensioning of its elements with small thicknesses
• Its tensile string this higher than that of reinforced concrete
¾ Disadvantages
• Concentration of armor at the junctions tending to have low resistance of the mor-
tar
• Structures in which it is difficult to achieve high levels of ductility are obtained
• Unreliable mathematical modeling. For elastic-plastic seismic analysis of a real
building
• There is a need for earthquake-resistant belts at floor leveling order to provide ade-
quate strength and rigidity to the floors and roofs of buildings
• To guarantee mechanically that the panels forming the vertical eardrums can work
together in building sunder lateral actions of consideration, primarily when the
building as more than two levels, bolting between panels is required.
• The study subject pool will have the following general characteristics: Conceived to
beburiedwith50m long, 25mwide and a maximum height of 2.5mlocatedin the cen-
ter (see Figure 1).For proper water drainage will possess a slope of 2%, this leads to
the ends where lanes are located to have2m of height.
200
Figure 1: Pool 3D View
The vessel walls are formed with prefabricated ferrocement panels of a 2m length, 1m
width and 3.5cm thickness, In order to achieve an adequate stiffness in the working plane,
counterforts will be fitted having the same thickness and length shall with a height 2m.
The vessel walls are formed with prefabricated panel’s ferrocement length 2m a 1m wide
and 3.5cm thick; to achieve adequate stiffness in the working plane buttresses that have the
same thickness and length shall be fitted with a height 2m.
The wall panel’s width was set also 1m in order to match the mesh as they usually come in
rolls 1.10m wide. To prefabrication the counterfort the same mold as for the wall will be
used, two counter forts can be built with one mold, which is why the thickness of the coun-
terfort was taken equal to that of the wall and the width was set as 50cm. . Because both
the walls and buttresses are 2m high, the space left to reach the bottom, which varies from
0 at the ends up to 50cm in the center, will be covered by the bottom slab, which will be
shaped properly in that area.
The pool was designed for seismic loads. A well-graded gravel, (pure frictional soil) was
used as fill material with the following characteristics: volumetric weight (γf) of 22 kN/m3,
internal friction angle (Φ) of 35 ° and without cohesion (c =0).
All the structural analysis of this work was performed using the Software, "Autodesk Ro-
bot Structural Analysis Professional 2010". After obtaining the solicitation loads, the de-
signed was carried out using the method ultimate limit states ULS), in which the loads are
overestimated and decrease the resistance capacity of the material considered lesser, thus
obtaining the amount of mesh fabrics required for positive and negative reinforcement. It
was double checked by means of the Method of Allowable Stresses and volume factor,
specific surface area, arrow and cracking of the ferrocement.
For the bottom slab design as hypothesis of was assumed that the level of maximum rise of
the water table is at least50cmbelow the bottom of the pool, so the bottom slab works alone
as an impermeable layer, since water pressure is compensated by the soil reaction, is there
is a balance of forces that ensures that the bottom slab will not been forced.
201
DETERMINATION OF RESTING PRESSURE RANKINE

To calculate the upward thrust of the land the Rankine method was applied since it is as-
sumed that there is not friction between the floor and the wall. It is considered that the side
walk surrounding the pool around its perimeter serves as bracing for the wall at its upper
limit and therefore the pressure transmitted by the ground will be static and that the wall
will not move.
For earthquake-resistant design is necessary to take in to account the dynamic in crease

caused by the earth quake, both in field thrust and hydrostatic pressure. In calculating the
dynamic earth pressure increase, the book Analysis and design of reinforced concrete re-
taining walls, for which the Standard Venezolana 1753-2006, which is very clear topic,
was consulted. In the case of dynamic increase of hydrostatic pressure 350.3-01 the ACI
which is focused to the study of earthquake-resistant design for a concrete tank was con-
sulted. Once obtained the dynamic load in gin creases, the following load combinations
were employed: the first combination will consider the time when the water mass ap-
proaches the wall (U1*) and the second time simulates the water mass tends to go away
from the wall (U2*).U1*=1.2 (PH) +1.4 (IDph)-0.9 (ET) U2*=1.6 (ET) +1.4 (IDet) - (PH)
+1.4 (IDph)
Where:
HP: Hydrostatic pressure.
TT: Terrain thrust.
DIhp: Dynamic increase of the hydrostatic pressure.
DItt: Dynamic increase of the terrain thrust.
Determination of the dynamic increase of thrust field (IDET)

The effects of earthquake the resting pressure on the structure. According to the Venezue-
lan standard, a trapezoidal pressure diagram can adopted, whose heights are: at the upper
end of the wall(σxs) and base wall (σxi). 8 (See figure 2).
Figure 2: Increment dynamic of repose push
σxs = 1,5 A0 γ H, Tension in the pool´s upper part

σxi = 0,5 A0 γ H, Tension in the pool´s lower part
202
Where:
A0: The ground acceleration as the zoning map. In this case we take A0 = 0.3 which is the
largest value that exists in Cuba and corresponds to zone 3 in Santiago de Cuba.
γ : The specific weight of the filling material that generates thrust in rest (= 22 kN / m3).
H: is the wall height (H = 2m).
In the diagram in figure 2.12 is shown how to plot and calculate the resulting pressure and
also indicates that is concentrated to 0.6 M, but in this case modeling will be performed
with a more realistic pressure distribution.
Determination of the dynamic increase in hydrostatic pressure (IDHP)

The dynamic increase of hydrostatic pressure (IDHP) depends on four key components:
pressure due to the impulsive side force (Piy), the pressure due to the convertiva lateral
force (Pcy), the lateral pressure due to the inertia of the wall (Pwy), and the pressure due to
the effect of vertical acceleration (PVY) 0.7
¾ Cálculo de la fuerza lateral impulsiva (Piy)
: Liquid design depth (m).

: Altura height Piy as measured from the Wall base (m).
For deposits with L/h_l≥1,333, hi is calculated by the following expression: H_i/h_l=0.375
Where L is the length of the tank inside in the direction parallel to the force of the earth-
quake, or in our case either L or L=25m=50mby dividing by HL= 2mwill always give
morethan1,333. Y:It's the distance from the base of the wall to the point at which you want
to calculate Piy, in this case the function y =0 and y =HL be evaluated so that the two ex-
treme values of the trapezoidal pressures diagram are obtained of Piy. Pi: Total impulsive
side force (kN)
P i=ZSIC _i (w_i /R_Wi)
Z: seismic zone factor. Z=0.3for Zone 3which is the largest that exists in Cuba (see Figure
3). S: coefficient depending on the soil characteristics. As in this case soil characteristics
are not known as it is intended that the pool is constructible anywhere and as directed by
the norm is as summed S= 1.5(see Figure 4).
I: Factor of importance. Because it is intended that the pool can still be used after the quake
took I =1.25 (see Figure 5).
RW: response modification factor. RW= 4 because the pool is buried and fixed on the base
(see Figure 6).
203
Figure 3: Table 4a from ACI 350.3-01 Figure 4: Table 4b from ACI 350.3-01 to determine S.to
determine Z.
Figure 5: Table 4c of ACI 350.3-01for determiner I.
Figure 6: Table 4d of ACI 350.3-01 for determiner Rw.
204
Wi: Impulsive component of WL (kN).

WL: Total liquid mass born by the wall (kN).
Note: Whenever it is necessary to determine anything that is a function of L, this will be
carried out using 2 methods and the one that yields a worst condition will be adopted, for
L=25m and L=50m, as the earthquake can occur in any of the 2 directions.
: Water Specific Gravity (kN/m3).

VL: Water Volume born by the wall (m3).
B: Wall width. B=1m.
Ci: Spectral amplification Factor caused by horizontal movements of the impulsive com-
ponent Depends directly of Ti.
Ti: Tank fundamental oscillation period (s).
requency (rad/s).
: Mass (kNs2/m4).
: Mass impulsive component (kNs2/m4).
: Relative mass to the water density (kNs2/m4).
: Gravity acceleration =9,807m/s2.

The value of is the same for L=25m as to L=50m, as the ratio L/WL is the same
Mass per wall width unit (kNs2/m4).
205
: Wall height (m).

: Wall thickness (mm).
: Masss relative to the wall material density (kNs2/m4).
: Specific gravity of the Wall material (kN/m3).

: Stiffness to bending perwal width unit (kPa).
For open based-fixed deposits as the case .
EC: Elasticity Module for the wall material (MPa). EC=23500MPa
: Wall gravity center (m).

For deposits having , the spectral amplification factor (Ci) can be calculated by:
Lateral convertive force calculation (Pcy)
: Centroide of Pcy height measured from the wall base(m).

PC: Convertive lateral Force (kN).
WC: Convertive component of WL (kN).
: Spectral amplification Factor caused by the horizontal movement of the convertive

component. Depends directly on TC.
TC: Natural period of the first impact mode (convertive) (s).
206
: Oscillation Frequency of the convertive mode (rad/s).
For deposits having , Spectral amplification (CC) can be figured out by means
of:
The value is calculated with value, assuming that the quake will act in a
direction to the 25m side, where the values are higher and thus preventing the worst case
scenario.
Calculo de la fuerza de inercia lateral de la pared (Pwy)
: Effective mass coefficiente .
Note: is an uniformly distributed pressure and its value is not a function of y.
Note: , y are vertically distributed pressures per length unit, in order to distribute
them horizontally they shall be divided by the wall thickness, (B=1m), their value stays the
same but the units will be kN/m2.
Cálculo de la fuerza por el efecto de aceleración vertical (Pvy)
: Effective Spectral Acceleration of an inelastic vertical response spectrum.
: Spectral amplification Factor caused by the vertical movement .For rectangular depos-
its as our case .
: Vertical / horizontal acceleration ratio. When unknown it is recommended, .
: Unitary hydrostatic pressure as y function.
207
Dynamic hydrostatic pressure increase calculation (DHPI)
Structural Design under seismic actions.
After determining the acting stresses in the wall of the pool and on the counterfort, the de-
sign is performed by the method of ELU, fissuration, deformation, specific volume sur-
face, volume reinforcement factor, are checked and the actinging tension is compared to
the permissible. Once finished designing the amount of materials used is determined and
subsequently compared with a similar reinforced concrete swimming pool.
MAIN ECUATIONS USED
M
σ ft = , Acting Tension (no bigger tha n 5 MPa .)
W …………………...…1
Mu = 0,8R /b* ⋅ b ⋅ h2 ⋅ k (1 − 0,4k ) , Moment, Equation MELU……… ….……2
σm
W = η mϕ S , Crack width (no bigger than 0.050 mm)………………....3
1 Em
Vre inf orcement

V = , Volumen Factor . (1% - 8%)
r Vcompound … …………………. 4
, Specific surface. (0,5 cm - 1 - 2cm - 1 )

Asuperfici al reinforcem ent
Sr =
Vcompound ….. 5
Table 3: Material costs for the ferrocement swimming pool.
Element Mortar Volume (m3) Reinforcement Weight (T)
Counterforts 5,39 0,51
Wall Panels 10,50 2,01
208
Bottom slab 31,88 3,92
Total 47,77 6,44
QUANTITATIVE COMPARISON BETWEEN THE OLYMPIC POOL OF BOYACÁ AND DESIGNED

FERROCEMENT.
The comparison is done by elements, as follows: Counterforts ferro vs reinforced concrete

columns, ferrocement panels that make up the wall vs reinforced concrete panels that make
up the wall, bottom slab ferro vs bottom slab reinforced concrete and finally a comparison
between the totals. (See graphic 1 and 2).
Figure 7: Comparison based on mortar volume.
209
Figure 8: Comparison based on reinforcement weight.
As shown the savings in the amount of materials is considerable, about nine times the
amount of mortar and eight times the amount of reinforcement steel.
CONCLUSIONS
After having completed this research has reached the following conclusions:
1. A seismic-resistant design is achieved in the Olympic pool ferrocement which meets all
the required d parameters.
2. The amount of materials to be used is 8.5 times less compared to reinforced concrete.
Whereas the mesh fabric is more expensive than steel reinforcement of reinforced concrete
in Cuba as it is an imported item, the cost savings will ultimately be between 5 and 6 times.
LITERATURE
1. http://upcommons.upc.edu/pfc/bitstream/2099.1/3333/5/34063-5.pdf.
http://www.sitioferrocemento.com/pagina0003.php.
2. Colectivo de Autores, Fundamentos para la aplicación del ferrocemento. Lima, 2000.
3. http://www.mioruro.com/libros/arquitectura/el%20ferrocemento.doc.
4. http://www.araosguzman.org/2011/06/el-ferrocemento.html.
5. http://el-mag.biofutur.org/2010/09/el-ferrocemento/.
6. http://universidaddelmedioambientesurco.blogspot.com/2009/05/estructuras-de-ferro-cemento-
para.html.
7. http://www.ftc.uni.edu.ni/dc/FOLLETOS/Ferrocemento/1_introduccion.ppt.
8. Dr. Ing. Hugo Wainshtok, Ferrocemento. Diseño y construcción. Riobamba-Ecuador 2010 4ta
Edición.
9. http://www.wordreference.com/definicion/alberca.
10. http://www.definicionabc.com/general/piscina.php.
11. http://www.arquigrafico.com/las-piscinas-definicion-partes.
12. http://www.canalconstruccion.comferrocemento.com.
13. Artículo: Guía de Construcción para Estructuras de Ferrocemento, publicado en el 2003.
14. http://www.canalconstruccion.comferrocemento.com.
15. Wainshtok Rivas. Hugo R, Artículo: Uso del Ferrocemento en Cuba.
210
16. Rafael Angel Torres Belandria, Análisis y Diseño de Muros de Contención de Concreto Armado.
17. Colectivo de Autores, Seismic Design of Liquid-Containing Concrete Structures (ACI 350.3-01) and
Commentary (350.3R-01).
18. Colectivo de Autores, Guide for the Design, Construction,and Repair of Ferrocement. ACI 549.
19. Braja M. Das. Principio de Ingeniería de Cimentaciones. 1999.A. E. Naaman, Evolution in Ferroce-
ment and Thin Reinforced Cementitious Composites. USA 2009.
211
212
CHAPTER 4
BOND
213
214
IMPROVING THE BOND BEHAVIOR OF TEXTILE REINFORCE-

MENT AND MORTAR THROUGH SURFACE MODIFICATION
C. Morales Cruz(1), U. Gohil(2), T. Quadflieg(2), M. Raupach(1) and T. Gries(2), (1) Institute of

Building Materials Research (ibac) of the RWTH Aachen University, Germany, (2) Institute
of Textile Technology (ITA) of the RWTH Aachen University, Germany
Abstract:The bond between mortar and textile reinforcement is of crucial evidence for Tex-
tile Reinforced Concrete (TRC). Often, SBR (Styrene-Butadiene Rubber) and EP (Epoxy)
impregnated carbon and AR-Glass fibre textiles are used. However, the bonding behavior
within these reinforcements for TRC has shown potential for improvement, especially
when testing the adhesive tensile strength. Therefore, a preliminary trial was undertaken to
investigate the influence of a surface coating with EP and sand as a bonding agent to im-
prove the bond characteristics of TRC. A SBR-impregnated carbon textile was modified
with silica sand EP ratios of 0.10, 0.20 and 0.30. In the conducted studies tensile tests on
single rovings and on rectangular TRC samples have been carried out. No reduction of the
tensile strength of the single rovings was reported. The first results of this study on TRC
specimens with a surface modified textile reinforcement and a sand EP ratio of 0.10 al-
ready indicate an improvement of the bond behavior. Next steps include tensile strength
tests on TRC specimens with further variations of the surface modification and two silica
sand EP ratios of 0.20 and 0.30. Pull-out tests in order to quantify the “slip” behavior be-
tween textile reinforcement and mortar as well as surface adhesion tests will be undertak-
en.
INTRODUCTION
The load bearing behaviour of Textile Reinforced Concrete (TRC) is determined by the
interaction between the textile reinforcement and the mortar (or fine grained concrete) ma-
trix, especially the load transfer from mortar to textile reinforcement. The toughening
mechanisms in a TRC are directly affected by the stiffness of the fibre-interface-matrix
system. In the brittle cement based composite, the strain capacity of the fibre by far ex-
ceeds that of the mortar leading to a crack bridging effect [Mob12]. Hence, the interface
regions need to be properly studied in terms of the load transfer mechanisms.
Figure 1 schematically shows how the overall strength of the reinforcing material decreas-
es from filament level to the textile actually embedded in the TRC. This loss of strength
can be due to multiple reasons such as the damage caused during winding and transporta-
tion for the stage between filaments to yarn. In addition, the parameters of warp knitting
also affect the strength of the textiles as well as the damage occurring during the concret-
ing process due to the abrasiveness of the mortar or fine grained concrete in the wet state.
215
Fig 1: Sketch of the reduction of strength from filament to textile in TRC [Sch11]
Hence, textiles are coated to activate the whole roving including its internally located fi-
bres. Such a coating achieves superior load transfer from the concrete to all the fibre fila-
ments and minimizes the impact of damage during the concreting process [Rau12]. AR-
Glass textile reinforcements for concrete applications are often coated (mostly with EPs) to
increase the durability of the fibre material by protecting it against the highly alkaline pore
solution of the concrete. The impregnated fibres show less loss of strength when compared
to the normal fibres [Büt12]. Additionally, coating increases the strength of the textile rein-
forcement to compensate the loss of strength during the textile process chain [Büt12].
Within the framework of investigations of eight almost four year old textile reinforced
sprayed mortar sample areas the adhesive tensile strength on two piers of the Horkheim
Weir in Germany was tested. In some cases the bond between textile and mortar has been
the limiting factor for the adhesion strength [Mor14]. Figure 2 shows exemplary the adhe-
sive tensile strength failure site between non surface modified SBR-impregnated textile
reinforcement and sprayed mortar.
Fig 2: Adhesive tensile strength failure site between textile reinforcement and sprayed mortar
216
Previous research already proved the positive effect of surface modification of yarns with
respect to the textile performance in TRC. Schöne et al. [Sch11] showed how surface mod-
ification of yarn can be achieved through sanding or dry coating resulting in an increase of
the textile performance. Plasma handling of the textile surfaces as well as chemically po-
larising can also help in increasing the textile performance. Figure 3 shows how sanding
brings the maximum value to the textile performance [Sch11]. The current work takes this
previous research on yarns forward to find out the best ways to conduct sanding on textile
structures.
Fig 3: Effect of surface modification on textile performance [Sch11]
MATERIALS AND METHODS

In this study epoxy coatings with addition of sand filler materials have been used. The fill-
er material used is silica sand which can be added to the polymer with different ways. One
reason for using silica sand was its high availability in high purity. Figure 4 below shows
the textile used for the current research. A 3200 tex SBR-impregnated carbon textile with a
grid opening of 20 mm by 20 mm was used. The textile had tricot binding as evident from
the figure. The epoxy resin used for coating in this research was EPIKOTE Resin 827
[NN07]. The recommended mixing ratio for the epoxy was 100 parts by weight of the resin
and 14.5 parts by weight of the curing agent. The individual components are well blended
at the indicated mixing ratio at 50 °C. Some of the typical properties of resin and curing
agent are presented in table 1.
Table 1: Resin specification [NN07]
Property Units Value
1 2 3
Delivery state - Liquid
Viscosity at 25 °C, DIN 53015 mPa·s 9000 ± 1000
Density at 20 °C g/ml 1.16
Curing Temperature °C 90 - 120
Crystallization Temperature °C < 15
217
The carbon textile was coated using a dip coating process. The textile was dipped into a
liquid coating bath inside a foulard and was subsequently squeezed by a pair of squeegee
rollers outside the foulard. There was only one pair of squeegee rollers in contact with the
coating compound. The disadvantage with the foulard arrangement was the non-
availability of pressure adjustments since it was only based on the own weight of the
squeegee rollers [Hoj10].
Two different methods to add sand were experimented within this research. They are: sand
mixed in the resin (EP) and sand sprinkling method. A schematic sketch in figure 5 de-
scribes these two methods.
Fig 4: Carbon textile used in this study
Fig 5: Schematic sketch of the coating process. Top: sand sprinkling method; bottom: sand mixed in EP method
The sand sprinkling method, as the name suggests, is a normal coating process followed by
the sprinkling of the sand over the coated, but uncured textile sample. As the sprinkling
was carried out manually, it was difficult to control the actual percentage of sand to be
sprinkled in an open mesh structure. Lastly, the coated textile is kept in the oven for 3 hrs.
at 100 °C for curing.
218
Within the sand mixed in resin method, a mixture of sand and resin was prepared in a sepa-
rate container. The textile was fixed on a wooden frame to maintain its alignment. The
resin was applied on both sides of the textile. It was uniformly coated over the textile and
also kept in the oven for 3 hrs. at 100 °C.
Fig 6: SEM pictures. Left: sand sprinkling method; right: sand mixed in EP method
Initial studies on both sand application methods were conducted with a high sand to EP
ratio (approximately 0.50). A Scanning Electron Microscope (SEM) was used to generate
images to view the surface modification of the textile. Sand distribution on the textile sur-
face is clearly visible in figure 6. Table 2 describes some key parameters of both sand ap-
plication methods as observed during and after the sand surface modification.
Table 2: Comparison between sand mixed in EP and sand sprinkling method
Sand mixed in EP Sand sprinkling method

1 2
Only one coating step High surface roughness
Higher adherence of sand on tex-
Advantages Sand sedimentation avoided
tiles
- Even sand distribution
The images clearly show the even sand distribution and a higher surface roughness of the
sand sprinkling method. However, a low adherence of sand on the EP in the sand sprin-
kling method was reported. This however can be improved by reducing the sand to EP ra-
tio. Therefore, this research considered using 0.10, 0.20 and 0.30 sand ratios relative to the
amount of EP with both methods to do a sensitivity analysis of the performance at different
sand contents.
219
TEST SPECIMEN
Fig 7: Rectangular tensile test specimen - sketch
Rectangular specimens according to the latest RILEM Recommendation [RIL14] with a

length of 500 mm, width of 60 mm and a thickness of 10 mm with one textile layer (3
Rovings in warp direction) were created (see figure 7) for the tensile tests. For this study a
fine grained concrete mixture developed at ibac (PZ-07-0899) with a maximum grain size
of 0.6 mm and a high first crack stress was chosen [Bro06]. In the preliminary studies, two
test series were investigated in order to establish whether the surface modification has an
influence on the bonding behavior (see table 3). These two series have the same amount of
reinforcement as well as the same sand surface modification (sand sprinkling method) and
differ only in the sand to EP ratio.
The specimens where created using the laminating technique. After casting, all specimens
were compacted for 60 sec. Subsequently, these were cured in a mold for 24 hours at 20 °C
and 65 % RH and then stored for four days in a climatic chamber at 20 °C and 95 % RH.
The specimens where then stored at 20 °C and 65 % RH until the testing age of 15 days.
Table 3: Test series
Surface modification
Sand sprinkling Sand mixed in resin Fine grained con-
Series Reinforcement
method method crete
(sand EP ratio)
1 2 3 4
1 - -1)
2 SBR- impregnat- 0.10 0.101)
PZ-07-0899
3 ed carbon 0.201)
4 0.301)
Two strain transducers (DD1) with a measuring length of 250 mm were attached to the
middle of both sides of the specimen. The load introduction on both sides was carried out
by 100 mm long steel plates glued to the fine grained concrete specimen. The tensile tests
220
were carried out at a strain rate of 5 mm/min. A preload of approximately 100 N was ap-
plied.
RESULTS AND ANALYSIS
Figure 8 illustrates the tensile strength and the elongation at break of the different textile
samples. It is evident from the results that surface modification of textiles does not signifi-
cantly change the tensile strength or the elongation at break of the textile.
Fig 8: Left: Comparison mean tensile strength of different specimens, right: Comparison mean elongation of
different specimens
In figure 9, the results of the uniaxial tensile strength of the TRC are presented in one dia-
gram per test series. The mean first crack stress of the mortar was reported between 4.6 and
5.0 N/mm² and is in good agreement in both test series. Although the uniaxial tensile test
presented here is not usually used to determine the bond properties of TRC, it allowed a
quantitative comparison of the bonding behavior of these two series. In figure 9, right, a
greater amount of peaks in the stress-strain diagram, especially in the area between 0.2 and
4 mm/m than in figure 9, left, was recorded. Additionally, lower stress drops were meas-
ured in the surface modified series in figure 9, right. This suggests not only a lower mean
crack spacing but also at better bond behavior. However, in order to prove this assumption,
pull out test must be performed.
Fig 9:Stress-strain behavior of the tensile strength specimens. Left: reference; right: sand EP ratio 0.10
221
Due to the high mortar splitting danger, the strain transducers were removed before the
sample failed in order to prevent damaging the measuring equipment (see figure 10).
Therefore the stress-strain diagrams in figure 9 are only presented up to the strain measur-
ing point of approximately 10 mm/m. A possible explanation for the mortar splitting is the
low thickness of the rectangular specimens. Due to the compressive strength failure of the
mortar at the clamping area without the rupture of the textile, the mean ultimate strength of
the TRC specimen could not be evaluated. Therefore, no interpretation of influence on the
tensile strength of the textile reinforcement due to the surface modification is given. Con-
sequently, the thickness of the specimens will be increased or a high strength mortar mix-
ture with short discrete fibers will be used in future studies. The influence of other sand EP
ratios (for e.g. 0.20 and 0.30) on the bonding behavior of TRC will be tested in further ten-
sile strength tests.
Fig 10: Specimen (sand EP ratio 0.10) shortly before concrete

failure
CONCLUSION AND OUTLOOK
In this paper two primary methods of sand surface modification were discussed. First re-
sults show no reduction of the tensile strength of the surface modified single rovings. The
preliminary findings on the tensile strength test on rectangular TRC specimens point to an
increase of the bond behavior of the surface modified textile reinforcement with the sand
sprinkling method. However, the ideal sand EP ratio still has to be explored. Additionally,
more peaks which suggest a higher amount of cracks and lower mean crack spacing has
been observed.
Further research scope lies in experimentation with different low viscosity high adhesive
polymers to attain higher degrees of impregnation. Although the obtained improvement of
the hand coated textiles was reasonably higher than the uncoated textile, further investiga-
tions have to be done. In further studies more uniaxial tensile tests as well as adhesive ten-
sile strength tests and pull-out tests will be performed.
222
ACKNOWLEDGEMENT
We would like to acknowledge the support provided by the Deutsche Forschungsgemein-

schaft (DFG) through the financing of this work under the project titles T08 SFB 532: Pre-
cast textile reinforcement for concrete buildings, and T09 SFB 532: Textile Reinforced
Sprayed Mortar Layers for the Surface Repair of Waterway Engineering Structures. In
addition, we would like to appreciate the support from our partners Massenberg GmbH,
SGL Technologies GmbH, STO AG, TAG Composites & Carpets GmbH, w + s Bau-
Instandsetzung GmbH, Gunther Kast GmbH, Hering Bau GmbH & Co. KG., Groz-Beckert
KG and Institut für Massivbau (IMB at RWTH-Aachen University).
REFERENCES
[Mob12] Mobasher, B.; Mechanics of fibre and textile reinforced cement composites, CRC Press,
New York, 2012
[Rau12] Raupach, M.; Orlowsky, J.; Büttner, T.: Verbesserung der Dauerhaftigkeit von Textilbeton
mittels Polymeren – Materialauswahl und Langzeitprognose, in 6th Colloquium on Textile
Reinforced Structures (CTRS6) - Textilbeton in Theorie und Praxis, Berlin, 2011
[Büt13] Büttner, T.: Zur Dauerhaftigkeit polymergetränkter AR-Glas Bewehrungen in Textilbeton,
Doctoral Thesis, Faculty of Civil Engineering, RWTH Aachen University, Germany, 2012
[Mor14] Morales Cruz, C.; Raupach, M., Westendarp, A.; Bruder, S.: Untersuchungen an einer vier
Jahre alten Schutzschicht aus textilbewehrtem Spritzbeton am Wehr Horkheim, Beton 10-
2014, pages 390-395, 2014
[RIL14] RILEM TC 232-TDT: Round Robin Tensile Test for Textile Reinforced Concrete. RILEM
Recommendations for Uniaxial Tensile Test, 2014
[Bro06] Brockmann, T.: Mechanical and fracture mechanical properties of fine grained concrete for
textile reinforced composites, Doctoral Thesis, Faculty of Civil Engineering, RWTH Aa-
chen University, Germany, 2006
[Sch11] Schoene, J.; Reisgen, U.; Schleser, M.; Jeřábek, J.; Theoretische und experimentelle Unter-
suchung des spaltungsinduzierten Versagens von TRC Prüfkörpern, 6th Colloquium on Tex-
tile Reinforced Concrete, 19th-20th
September, 2011, Berlin.
[NN07] Technical Information Sheet, Epikote 827 – Momentive, 2007
https://www.momentive.com/Products/TechnicalDataSheet.aspx?id=3603 last visited on
10.04.2015
[Hoj10] Hojczyk. M.; Weichold. O.; Glowania. M.; Gries. T.: Cement-in-polymer dispersions as
coatings for glass rovings and Textiles, Aachen, Germany, 2010
223
224
BOND BEHAVIOUR AND DURABILITY OF BASALT FIBRES IN

CEMENTITIOUS MATRICES
Simone Hempel, Marko Butler, Viktor Mechtcherine, Technische Universität Dresden, Institute of
Construction Materials
Abstract: This contribution presents selected results of investigations performed in study-

ing the time-dependent changes in the mechanical performance of basalt fibre, multifila-
ment yarns in cementitious matrices. Two basalt fibre materials both in their original con-
dition and after treatment with an additional coating were investigated. The fibres were
embedded in tensile test specimens which then were stored for one year in a fog chamber
to accelerate the aging process. After specific storage times, the pullout behaviour of the
embedded fibres was tested and, in parallel, the condition of the basalt fibres investigated
by means of electron microscopy in an ESEM. A good correlation between the loss of me-
chanical performance and the condition of the basalt fibres after accelerated aging was
observed.
INTRODUCTION
Since the mid-90s, many research and development projects have been carried out in the
field of the textile-reinforced concrete (TRC). This compound material consists of a high
performance, fine-grained concrete matrix and one or more layers of textile [1]. Under
tensile loading, after crack formation, the textile reinforcement transfers the forces [2]. In
most cases textile reinforcements have consisted of alkali resistant glass (AR glass) or car-
bon fibres, but the use of basalt filament yarns was considered as material for producing
textiles as well. In particular, the good temperature resistance of basalt fibres and relatively
low fabrication costs could make the resulting basalt yarns attractive in a wide range of
applications [3, 4]. These positive features and, not least, the wide availability of raw mate-
rials make this type of fibre a highly promising material for the textile reinforcement in
concrete.
Basalt is a worldwide accessible, alkaline lava stone consisting primarily of the oxides of
silicon, aluminium, iron, calcium, and magnesium [5, 6, 7]. In the production of the fibre,
the melting of the basalt, and its subsequent rapid cooling during the spinning process, the
crystalline properties of the base material are largely lost. Thus formed, the microstructure
of the basalt fibres is chiefly amorphous and similar to a glass network. However, this low-
ers the resistance to attack by the hydroxyl ions present in the cement-based matrices. This
reduced alkali resistance often results in damage to the fibres and consequently to loss of
fibre strength [8, 9, 10, 11, 12]. Furthermore, for the resistance and durable effectiveness of
the fibre material, glass as well as basalt, the quality of the sizing and of the additional
coatings applied are of special importance. The investigations at hand were performed to
enhance the knowledge of bond behaviour of basalt fibres to and their durability in cement-
based matrices.
225
MATERIALS AND TEST METHODS
Fibre material
Two commercial, multifilament yarns made of basalt fibres (Basalt 1 and Basalt 2) were
used in the investigations as reinforcing material. Both materials had a fineness of 2400 tex
(g/1000 m); the filaments were treated with a polymer sizing applied during fabrication.
Filament sizing is applied mainly for technological reasons in order to optimise the pro-
cessing of the filaments into multifilament yarns. However, it has been shown that the siz-
ing acts also as a protective layer over the fibre bulk, hence improving its resistance to ag-
gressive chemical agents [13]. For enhanced protection of the fibres and for activating the
entire yarn cross section, an additional coating based on Styrolbutadien was applied.
The powder-XRD measurement of the fibres showed that both basalt fibres under investi-
gation were completely amorphous. The chemical composition of the basalt fibres meas-
ured by EDX-analysis is displayed in Table 1.
Table 1: Composition of basalt fibres
Oxide Basalt 1 Basalt 2

Na2O 5.2 4.6
MgO 6.3 4.2
Al2O3 18.9 20.4
SiO2 47.5 52.8
K2O 1.2 1.9
CaO 9.5 7.7
TiO2 1.8 1.5
Matrix material – fine-grained concrete

In the present investigations a fine-grained cement matrix with medium alkalinity was
chosen. As binder, a Portland cement CEM I 32.5 R (Bernburg) was used. The pozzolanic
binder components were fly ash and micro silica (slurry). Table 2 displays the matrix
composition.
Table 2: Matrix composition

Component Mass [kg]
CEM I 32.5 R 548.8
Fly ash 245.6
Micro silica suspension 54.6
Sand 0/1 mm 191.5
Water 245.6
226
TEST PROCEDURES
To assess the performance of the basalt fibres when embedded into the cement-based
matrix, double-sided pullout tests on dumbbell-shaped specimens were carried out; see
Figure 1.
The specimens were produced in plastic forms closed on all sides in order to avoid
differing shrinkage between their tops and bottoms. One multifilament yarn was embedded
along the long axis of the sample. For the setup given, formation of just one crack is
expected, more specifically at the notch in the middle of the specimen. One day after
casting, the samples were demoulded and stored in water for 6 days. Afterwards they were
stored in standard climate conditions (20°C, 65% RH). To accelerate their aging the
specimens were stored in a fog chamber (40°C, 99% RH) starting at a concrete age of 7
days. After storage in the fog chamber, over varying periods, the specimens were
conditioned in the standard climate for 24 h before testing. The pullout tests were
performed in a universal, servo-hydraulic testing machine under a deformation-control
regime. The strain rate was 0.01 min-1, as measured by means of two LVDT fixed to the
specimen and oriented in the direction of loading; see Figure 2, so that the load-
displacement curves could be recorded [14, 15].
Figure 1: Geometry of the dumbbell specimens with indication Figure 2: Testing machine with fixed
of arrangement of the yarn, gauges and load-introducing steel specimen
adapters (geometrical data in mm)
On completion of each pullout test, a strip approximately 5 to 10 mm in width was cut

from the uncracked area of the specimen (red marks in Fig. 1). The ends were cut off, so
that only the area with the roving embedded remained. Afterwards the roving was exposed
by splitting the sample in the middle; see Figures 3 and 4.
227
Figure 3: Sample stripe with embedded basalt roving Figure 4: Sample split with exposed basalt roving
The samples prepared in this manner were examined in a scanning electron microscope
XL30 ESEM in LowVac mode to investigate fibre damages as a result of the accelerated
aging.
RESULTS
Pullout tests – standard climate
Figures 5 and 6 show the results of the pullout tests performed on samples reinforced with
rovings, absent additional coating, and stored in the standard climate until testing at an age
of 28 days. The curves plot pullout force versus crack opening. The specimen containing
the yarn Basalt 1 yielded maximum pullout forces in a range between 0.28 kN and 0.38 kN
with crack opening widths of 0.3 to 0.6 mm. Both the relatively low pullout forces and the
flat curve shape indicate a low bond intensity between filament and matrix with the
domination of the filament pullout; see Figure 5. For the specimens with Basalt 2 yarns,
maximum pullout forces of 0.38 kN to 0.59 kN were attained. The corresponding crack
widths varied from 0.08 mm to 0.23 mm. The mostly steeper increase of the curve
indicates that for this type of basalt fibre there is a more intensive bond in the fibre-matrix
interphase. Thus, higher loads can be transmitted across the crack until fibre damage and/or
fibre pullout occur; see Figure 6.
1.0 1.0
0.9 0.9
0.8 0.8
pull-out force F [kN]
0.7 0.7
0.6 0.6
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0.0 0.0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
crack opening w [mm] crack opening w [mm]
Figure 5: Force versus crack opening curve for Figure 6: Force versus crack opening curve
specimens with uncoated Basalt 1 yarns stored for 28 specimens with uncoated Basalt 2 yarns stored for
d in the standard climate 28 d in the standard climate
228
The main causes of the different performance of the both yarn types could be differences in
the interaction between the fibre surfaces and hydration products of the cementitious ma-
trix. Both to compensate the insufficient surface quality and to activate all filaments as
completely as possible, the yarns were coated with styrolbutadien. Figures 7 and 8 show
the results of the pullout tests performed on the coated multifilament yarns made of Basalt
1 and Basalt 2. With respect to maximum pullout force, a distinct increase could be ob-
served due to the coating. Besides activating the filaments in the cross-section of the yarn,
the better adjustment of the surface quality of the fibres to the cementitious binder may
more probably have led to the positive result.
1.0 1.0
0.9 0.9
0.8 0.8

0.7 0.7
0.6 0.6
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0.0 0.0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
crack opening w [mm] crack opening w [mm]
Figure 7: Force versus crack-opening curve for Figure 8: Force versus crack-opening curve for
specimens with coated Basalt 1 yarns, 28 d standard specimens with coated Basalt 2 yarns, 28 d standard
climate climate
Pullout tests after accelerated aging

To investigate the long-term performance of fibres in cementitious matrices, samples
subject to accelerated aging (40°C, 99% RH) were tested. Both the increased temperature
and humidity lead to a more intensive attack of the alkalis on the amorphous network of
the fibres. Figures 9 and 10 show the representative pullout curves determined on the
specimens with basalt yarns after the fog chamber storages for 28, 56, 180 and 360 days.
Figure 9: Representative pullout curves for Figure 10: Representative pullout curves for
229
specimens with uncoated Basalt 1 yarns after various specimens with uncoated Basalt 2 yarns after various
durations of accelerated aging durations of accelerated aging
With increasing duration of the accelerated aging, both fibre types showed a distinct
decrease in both pullout force and work-to-pullout. Comparing the results after 28 days
storage in the standard climate (cf. Figures 5 and 6), it becomes clear that even a short-term
intensification of the alkaline attack led to a pronounced worsening of the mechanical
performance. The changes occurred with a slightly different dynamic for specimens with
Basalt 1 and Basalt 2 yarns, but at the end of the investigating period, the behaviour of the
samples was similarly unfavourable. The representative curves show that the uncoated
basalt fibres were distinctly damaged in the alkaline matrix. At the end of the observation
period, the fibres were not able to transfer loads after crack formation; rather they failed
already at low crack openings.
The subsequent application of an additional coating increased not only the mechanical per-
formance, but the coating polymer acted as a further barrier against alkaline attack as well.
Figures 11 and 12 show the pullout force - crack opening curves obtained from the tests
performed on the samples reinforced with the coated basalt yarns. With increasing duration
of the aging process, dramatic losses in maximum pullout force and work-to-fracture could
be observed. After a storage duration of 180 days, pullout force was reduced by 50% and
work-to-fracture by 80% when related to the level at 28 days.
In summary, the coating led to a distinct improvement of the bond between the basalt yarns
and the cement-based matrix. However, the alkaline attack on the fibre could not be com-
pletely averted. Due to the polymer coating and the hindered diffusion, the damage devel-
opment was slowed, but over a longer period, considerable damage could accumulate,
leading to premature fibre failure.
Figure 11: Representative pullout curves for Figure 12: Representative pullout curves for
specimens with coated Basalt 1 yarns after various specimens with coated Basalt 2 yarns after various
durations of accelerated aging durations of accelerated aging
MICROSCOPIC INVESTIGATIONS
The accompanying microscopic investigations of the basalt yarns after the respective
storage durations show that the corrosion of the fibres' surfaces are chiefly responsible for
230
the losses of the mechanical performance observed. Figures 15 to 18 display the condition
of the fibres after 180 days' accelerated aging. The uncoated as well as the coated fibre
surfaces are completely damaged, characterised by the dissolution of the fibres and the
formation of thick layers of corrosion. These layers drop away from the remaining fibre
core, leading to a reduced fibre cross-section.
Figure 15: Uncoated Basalt 1 yarn after 180 days' Figure 16: Uncoated Basalt 2 yarn after 180 days'
storage in the fog chamber, filaments are covered by storage in the fog chamber, extensive delamination of
thick layers of corrosion products corrosion layers
Figure 17: Coated Basalt 1 yarn after 180 days' Figure 18: Coated Basalt 2 yarn after 180 days'
storage in fog chamber, thick corrosion crusts on storage in fog chamber, large corrosion shells,
fibre surfaces delamination of the fibre
Regarding both the results of the pullout tests and the microscopic investigations of the
damage process due to accelerated aging, it can be concluded that the coating applied sub-
sequently to the fibres surface delayed the attack on the fibres. However, no long-term pro-
tection against the alkalis was provided with the given coating, so that after 180 days' stor-
age a high degree of damage was suffered by both basalt fibre types.
CONCLUSION
From the findings of the pullout tests presented and microscopical investigations
conducted, the following conclusions can be drawn:
231
Testing after 28 days' storage in the standard climate showed different bond properties
depending on the type of basalt fibre when they had been treated only with the original
sizing. It may be assumed that the sizing applied had not been sufficiently adjusted to the
fibre material with respect to its application in alkaline concrete environment.
Accelerated aging in a fog chamber led in the case of the specimens with the uncoated
basalt fibres to a dramatic decrease in mechanical performance due to the inadequate
resistance of the sizing and the basalt fibres to the alkaline attack. This resulted in
progressive fibre damage and the fibres' failure at relatively low force levels.
The styrolbutadien coating applied to basalt yarns improved the strength of the basalt yarns
and the bond behaviour distinctly. However, with the ongoing aging caused by storage in
the fog chamber, the decrease in mechanical performance was pronounced. The damage of
the basalt fibres could be slowed down by the coating but in the long term not prevented.
To improve basalt multifilament yarns as regards their application in concrete structures

and to protect the fibres in the alkaline matrix sufficiently, the development of suitable
sizings and coatings designed specifically for this material is necessary. Subject to this
condition, it will be possible to use basalt multifilament yarns as a means of reinforcement
for cement-based materials.
REFERENCES
[1] Köckritz, U., 'In-Situ Polymerbeschichtung zur Strukturstabilisierung offener nähgewirkter Gelege‘,
Technische Universität Dresden, Dissertation 2007
[2] Molter, M., ‘Zum Tragverhalten von textilbewehrtem Beton‘, Rheinisch-Westfälische Technische
Hochschule Aachen, Dissertation 2005
[3] http://www.swiss-composite.ch/pdf/i-Werkstoffdaten.pdf (Stand 03.06.2014)
[4] http://www.basfiber.com/1 (Stand 03.06.2014)
[5] Liu, J., ‘Untersuchung von Verbundwerkstoffen mit Basalt- und PBO-Faser-Verstärkung‘, Technische
Universität Dresden, Dissertation 2007
[6] http://www.asatec.at/ (Stand 03.06.2014)
[7] http://www.deutsche-basalt-faser.de/ (Stand 03.06.2014)
[8] Scheffler, C; Förster, T; Mäder, E.; Heinrich, G.; Hempel, S.; Mechtcherine: V., ’Aging of alkali-
resistant glass and basalt fibers in alkaline solutions: Evaluation of the failure stress by Weibull distri-
bution function’, In ‘Journal of Non-Crystalline Solids’ 355 (2009) 2588-2595
[9] Rybin, V. A.; Utkin, A. V.; Baklanova, N. I., ‘Alkali resistance, microstructural and mechanical per-
formance of zirconia-coated basalt fibers’, In ‘Cement and Concrete Research’ 53 (2013) 1-8
[10] Sim, J.; Park, Ch.; Moon, D. Y., ‘Characteristics of basalt fiber as a strengthening material for con-
crete structures’, In ‘Composites’ Part B 26 (2005) 504-512
[11] Förster, T.: Mäder, E., ‘Performance of modified basalt fibres’ In 18th International Conference of
Composite Materials, 2009, Edinburgh, Scotland
[12] Wei, B.; Cao, H., Song, S., ‘Environmental resistance and mechanical performance of basalt and glass
fibers’, In, ‘Materials Science and Engineering’ A 527 (2010) 4708-5715
[13] Scheffler et al., ‚Interphase modification of alkali-resistant glass fibres and carbon fibres for textile
reinforced concrete I: Fibre properties and durability‘, ‘Composite Science and Technology’ 69
(2009) 531-538
[14] Butler, M., Mechtcherine, V., Hempel, S., ‘Experimental investigations on the durability of fibre-
matrix interfaces in textile reinforced concrete’, ‘Cement & Concrete Composites’ 31 (2009) 221-231
[15] Butler, M., Mechtcherine, V., Hempel, S., ‘Durability of textile reinforced concrete made with AR
glass fibre: effect of matrix composition, ‘Materials and structures’ (2010)43: 1351-1368
[16] [Sch11] Schoene,J.; Reisgen, U.; Schleser, M.; Jeřábek, J.; Theoretische
232
[17] und experimentelle Untersuchung des spaltungsinduzierten Versagens von TRC Prüfkörpern, 6th
Colloquium on Textile Reinforced Concrete, 19th-20th
[18] September, 2011, Berlin.
[19] [NN07] Technical Information Sheet, Epikote 827 – Momentive, 2007
https://www.momentive.com/Products/TechnicalDataSheet.aspx?id=3603 last visited on 10.04.2015
[20] [HWG10] Hojczyk. M.; Weichold. O.; Glowania. M.; Gries. T.: Cement-in-polymer dispersions as as
coatings for glass rovings and Textiles, Aachen, Germany, 2010
ANALYSIS ON PULL-OUT BEHAVIOUR OF CONTINUOUS MULTI-
FILAMENT GLASS YARNS EMBEDDED IN CEMENTITIOUS
MATRIX BY USING A DEVELOPED MODEL
H. Aljewifi(1), X.B. Zhang(2), J. Li(3), (1)University of Omar, Almukhtar Elbida, Libya,

(2)Univercité Blaise Pascal de Clermont-Fd, Institut Pascal, Montluçon, France, (3)Univercité Paris
XIII, LSPM, CNRS UPR 3407, France
Abstract: Pull-out test is a reliable tool to investigate the failure mechanism of continuous
fibre reinforced composites. It has been found that the interaction between fiber and matrix
is characterized by different bond parameters that condition the quality of the bond be-
tween fiber and matrix. In this work, a series of pull-out tests is performed with the speci-
mens made from continuous multi-filament glass yarns embedded in cementitious matrix.
By using an analytical model developed recently by authors [1], the bond parameters are
determined from pull-out experiment curves for different types of fiber. The comparison of
these parameters for different types of fiber allows us to analyse the pull-out behaviour of
the fiber/matrix system.
INTRODUCTION
Multi-filament yarns as principal reinforcement have been widely used because of their
numerous advantages compared to traditional reinforcement. The study of the interaction
between multi-filament yarn and cementitious matrix can give a good comprehension of
the failure mechanism of this type of composite. This study may be carried out by means
of pull-out test which is an applicable tool to investigate the shear bond properties. Numer-
ous authors have investigated the bond between filaments and cementitious matrix by dif-
ferent methods. Most of studies are based on the analysis of the shear stress at the interface
such as those of Namure and Naaman [2], Li and Stang [3] and Stang et al.[4]. We have
also studied analytically, numerically and experimentally the failure mechanism of contin-
uous yarn reinforced cementitious composites by analyzing the pull-out behaviour of a
multi-filament yarn reinforced cementitious composites. In order to improve the model
used before, we have developed recently an analytical model by taking the assumption
more approach to the real pull-out behavior [1]. The object of this work is to analyze the
pull-out behaviour of continuous multi-filament yarn reinforced cementitious composites
by using the model recently developed by authors [1]. A series of pull-out tests is per-
formed with the specimens made from different types of fiber. By using the developed
model, the bond parameters are determined from pull-out experiment curves for different
233
types of fiber. The comparison of these parameters for different types of fiber allows us to
analyse the bond behaviour of the fiber/matrix system.
PRESENTATION OF THE ANALYTICAL MODEL
The analytical model developed recently by authors [1] is based on the analysis of the
stress τ and displacement u of the yarn at the interface between yarn and matrix. The fail-
ure mechanism is analyzed by devising the pull-out curve into three phases: perfect bond-
ing phase, debonding phase and pure friction phase. When the uniaxial tensile force F ap-
plied on the free end of a yarn reaches a critical value Fcri, the debonding of the yarn be-
gins to take place. It is the beginning of the second phase. The length of the debonding part
increases as the applied force increases. When this length reaches the embedded length,
the applied force reaches its maximum value Fmax, The pure friction occurs. It is the third
phase which is the pure friction between the yarn and the matrix. The level of applied force
decrease greatly. These three phases are shown in Figure 1.
Fmax pure friction τ perfect

bonding
debonding pure friction
Fcri τmax
perfect
k1 k2 τ0
b di
u0 umax u u
umax
Fig. 1 Pull-out curve Fig 2. Shear stress versus displacement
Fig.2 shows the assumption taken in the model : The shear stress τ at the interface be-
tween yarn and matrix increases proportionally to the displacement u of the yarn (τ = k1 u)
during the perfect bonding phase. It reaches its maximal value τmax at the outset of debond-
ing phase and then it decreases proportionally to the displacement u of the yarn (τ = k2 u)
during the debonding phase. In the end, during the pure friction phase, it becomes constant
τ0 along all length of the yarn. k1 and k2 are the bond modulus corresponding respectively
to the perfect bonding phase and the debonding phase. τmax And τ0 are bond parameters
obtained from experimentation.
Perfect bonding phase

At this configuration the bond between the yarn and the matrix is perfect (Fig. 3). The rela-
tionship between the displacement of the yarn and the applied force is written as follows:
4F chλ ( L + x) with λ =
u ( x) = 4 k1 / DE
λ π D2E shλ L
(1)
234
Where λ is a material parameter depending on the bond quality of the interface between
yarn and matrix. It is function of the bond modulus of the perfect bonding phase k1, the
yarn diameter D and Young’s module E.
matrix
F F
yarn yarn
0 x, u
L
0 x, u
L
Fig. 3 Yarn/matrix system at perfect Fig. 4 Yarn/matrix system at debonding phases

bonding phase
Debonding phase
When the force F applied on the free end of yarn reaches a critical value Fcri, the debond-
ing of the yarn takes place (cf Fig. 1). The length of the debonding part increases as the
applied force increases. Two parts of the yarn are distinguished (Fig. 4): bond part
(− L ≤ x ≤ −l ) and debonding part (−l ≤ x ≤ 0) .
The displacement of the bond part has been shown in Equ. (1). By replacing L by L-l and
u (−l ) = τ max / k1 , the displacement of the bond part of yarn (− L ≤ x ≤ −l ) and the force become
as follows:
τ max chλ ( L + x )
u( x ) =
k1 chλ ( L − l )
(2)
τ maxπD
F ( −l ) = thλ ( L − l )
λ
(3)
In the debonding part (−l ≤ x ≤ 0) , the shear stress τ at the interface between yarn and ma-
trix decreases proportionally to the displacement u of the yarn: τ = k2u. The displacement
of the debonding part and the force can be written as follows:
τ max sin (ω x + ϕ ) k1 + k 2
u ( x) = + τ max
k 2 sin (ω l − ϕ ) k1k 2
(4)
235
πD 2 Eτ max ω λ
F ( x) = cos(ω x + ϕ ) with ω = 4 k 2 / DE and ϕ = ω l − arctg
4k 2 sin (ω l − ϕ ) ω thλ (L − l )
(5)
As λ, w is also a material parameter depending on the bond quality of the interface be-
tween yarn and matrix.
Pure friction phase

When the length of the debonding part of the yarn reaches the embedded length, the ap-
plied force reaches its maximum value Fmax, the pure friction occurs (cf Fig. 1 and Fig. 5).
the displacement u and the pull-out force F can be written as follows:
4τ 0 ⎛ 1 2 ⎞ k +k 2τ L2 τ
u( x ) = ⎜ x + Lx ⎟ + 1 2 τ max + 0 − 0
DE ⎝ 2 ⎠ k1k 2 DE k2
(6)
F = π Dτ 0 ( L + x )
(7)
Loading batch
F=0 matrix Fext
yarn
Multi-filament yarn
L
L u Cementitious matrix
Free length
Embedded length
Fig.5 Pure friction phase Fig.6 Specimens for pull-out tests
EXPERIMENTAL STUDY
Specimens and materials

In order to get a yarn/matrix system studied, in this work the pull-out tests are performed
[6] [7] by using specimens made of a cementitious matrix in cylinder form in which a mul-
ti-filament glass yarn is embedded along the whole length of the cylinder (The cylinder
length is equal to the embedded length, noted as L). The diameter of the cylinder is 34 mm
and two different lengths L are used: L = 100 mm and L = 150 mm.
236
The free length (fibre no embedded in matrix) is constant (100 mm) and its end is glued
between two batches of epoxy resin for loading (Fig. 5). The loading batch is clamped on
the grip of a special tension device. The load is applied on the free end of the yarn under
constant speed of 0.01 mm/min and continues to increase until the yarn is fully pulled of
the matrix. The applied force F and the displacement u are recorded during the test. The
pull-out curves are then drown. Cementitious matrix is made of an ordinary Portland ce-
ment (CEM I 52.5). The maximum aggregates size of micro concrete in mixtures is 0-
1.25mm. A small quantity of admixture is added in order to improve certain chemical or
physical properties. A polycarboxylate based superplasticizer type of high water reducing
CIMFLUID 2002 with solid content of 35.6% is used according to the French standard NF
P18-380. Consequently, the mortar compositions are as follows: W/C= 0.48 % (wa-
ter/cement ratio), S/C = 1.4% (sand/cement ratio), SP/C = 0.0035% (Superplasticiz-
er/cement ratio). The composition of matrix is listed in Table 1.
Table 1. Composition of cementitious matrix
Mixture ingredients Superplasticizer (kg/m3)
CEMI 52.5 736.3
Sand 1031.0
Water 366.8
Superplasticizer 0.917
Package density 2135
Five kinds of fibres were used to produce multi-filament yarn: Owens Corning 359A-AA
multi-end continuous E-glass fibres (called OC1), Owens Corning 111A*11 type 30 single
end continuous E-glass strand (called OC2) and three types of Saint Gobain Vetrotex AR-
glass fibres consisting from thousands of filaments with the same diameter but their sizing
(number of filament in a yarn) is different, called SG1, SG2 and SG3 respectively. Yarn
mechanical proprieties have been determined according to standard ISO 3341 and shown
in Table 2.
Table 2. Data of filament and yarns proprieties
Yarn name OC1 OC2 SG1 SG2 SG3
Glass type E E AR AR AR
Roving type assembled direct assembled assembled assembled
Diameter of filaments (μm) 12 17 14 14 14
Estimated number of filaments 8385 4180 5940 5940 7272
237
Number of strands 40 / 55 30 35
Fineness (tex, g/km) 2400 2400 2450 2450 2500
Ultimate strength (N/tex) 0.465 0.375 0.291 0.349 0.332
Yarn stiffness (GPa) 54.8 55.9 56.5 44.9 45.5
Estimated yarn diameter (mm) 1.099 1.099 1.079 1.103 1.079
Determination of the bond parameters

For each kind of fibre material and each matrix length L, the pull-out tests are performed
with three specimens. Experimental pull-out curves obtained from different kinds of fibre
are used to calibrate the bond quality parameters used in the analytical model : the bond
parameters τmax, τ0 and the bond modulus k1 and k2. These parameters are evaluated ac-
cording to the experimentally obtained values of the maximum applied force, the maxi-
mum pull-out displacement and the slopes of the pull-out curves. Table 3 shows the aver-
age values of the bond quality parameters for different kinds of fibre.
Table 3. Average values of bond parameters
Material τmax (MPa) τ0 (MPa) k1 (MPa/mm) k2 (MPa/mm)
OC1
1.20 0.096 2.0 3.0
OC2
0.65 0.130 1.8 3.24
SG1
1.13 0.147 2.0 3.4
SG2
1.10 0.110 0.6 2.1
SG3
1.60 0.128 1.8 2.34
Figures 7 to 11 show the experimental pull-out curves of different specimens with different
kinds of fibre and different embedded lengths L. The pull-out curves obtained according to
the analytical model are also presented in the figures. It can be observed that the analytic
results agree globally well to experimental results. From the curves obtained according to
the analytical model, one can see the logic fact that the maximum loading force Fmax in-
creases as the embedded length L increases. It can be also observed from Table 3 that the
bond parameter τmax of OC2 fiber is the smallest due to its roving type different of the oth-
er fibers. The bond modulus k1 of SG2 fiber is the smallest because its Yarn stiffness is
smaller than that of the other fibers. There is no great difference for the values of τ0 for all
kinds of fiber.
238
Fig. 7 Pull-out curves for specimens with OC1 fibre
Fig. 8 Pull-out curves for specimens with OC2 fibre
Fig. 9 Pull-out curves for specimens with SG1 fibre
239
CONCLUSIONS
In this work, a series of pull-out tests is performed with the specimens made from continu-
ous multi-filament glass yarns embedded in cementitious matrix. By using an analytical
model developed recently by authors [1], the bond parameters are determined from pull-
out experiment curves for different types of fiber. The comparison of these parameters for
different types of fiber allows us to analyse the pull-out behaviour of the fiber/matrix sys-
tem.
REFERENCES
[1] ZHANG X.B., ALJEWIFI H., LI J., ‘Failure mechanism investigation of continuous fibre reinforced
cementitious composites by pull-out behaviour analysis’, Procedia Materials Science, 3, (2014)
1377–1382.
[2] Namure G., Naaman AE., ‘Bond stress model for fiber reinforced concrete based on bond stress-slip
relationship’, ACI Mater. J., 86(1), (1989) 45-47.
[3] Li VC., Stang H., ‘Interface property characterization and strengthening mechanics in fiber rein-
forced cement based composites’, J. Advanced cement based materials, 6 (1), (1997) 1-20.
[4] Stang H, Li Z, Shah SP., ‘Pullout problem – stress versus fracture mechanical approach’, Journal of
Engineering Eechanics-ASCE,116 (10), (1990) 2136 – 2150.
[5] ZHANG X.B., ALJEWIFI H., LI J., ‘Failure behaviour investigation of continuous yarn reinforced
cementitious composites’, Construction and Building Materials, 47 (October), (2013) 456–464.
[6] ALJEWIFI H., ‘Etude du comportement mécanique à l’arrachement de fils multi-filamentaires enro-
bés dans une matrice cimentaire et influence de l’imprégnation’, Thèse de l’Université de Cergy-
Pontoise, (2011)
[7] Aljewifi H., Fiorio B, Gallias JL. ‘Pull-out behaviour of a glass multi-filaments yarn embedded in a
cementitious matrix’, EURO-C 2010, Austria
240
CHAPTER 5
DURABILITY
241
242
TOWARDS DURABILITY DESIGN FOR STRUCTURAL ELEMENTS

AND STRUCTURES MADE OF OR STRENGTHENED WITH
TEXTILE REINFORCED CONCRETE (TRC)
Viktor Mechtcherine, Institute of Construction Materials, Technische Universität Dresden,
Germany
Abstract: Textile-reinforced concrete (TRC) has great potential for application in structures
exposed to severe mechanical or environmental loading. This paper presents an overview
of the current knowledge on the durability of this composite and its components. An addi-
tional focus is centred on the protection of steel reinforcement as in the case of the
strengthening or repair of RC structures using TRC. In doing this, transport properties in
the cracked state, long-term tensile strength and strain capacity, and resistance to aggres-
sive environments have been identified as critical parameters. Current knowledge indicates
that TRC can exhibit high long-term mechanical performance and favourable transport
properties in the cracked state. While also a superior resistance of TRC to aggressive envi-
ronments may be expected, as compared to ordinary concrete, there is little information
available on the effects of aggressive environments on the mechanical properties of the
material. To be able to utilise the superior qualities of TRC fully, it will be necessary to
develop a realistic and reliable performance-based durability design concept for structures
made of or strengthened by TRC. The article at hand provides a framework for such dura-
bility design.
INTRODUCTION
TRC is a composite construction material in which textile reinforcement consisting of pol-
ymer, glass, or carbon fibres is embedded into a fine concrete matrix. Since textile rein-
forcement generally needs no minimum concrete covering as a protection against corro-
sion, it is possible to manufacture thin-walled construction components and layers. The
fibre material can be exploited very effectively by adjusting the volume proportion and
orientation of the long fibres in the direction of tensile stress. Textile reinforced concrete
can be employed in the manufacture of thin-walled façade elements, load-bearing integrat-
ed formwork, tunnel linings, or in the strengthening of existing reinforced concrete (RC)
structures [1, 2].
Most current applications of TRC are non-structural. Textile reinforcement is primarily

used in controlling shrinkage cracks as well as cracks induced by temperature changes.
Furthermore, the control of cracks in elements of TRC due to overloading during transport
and mounting is an accomplished fact. However, considerable laboratory research has been
performed in the last few years on the behaviour of structural members made of TRC and
there are also a number of examples of the successful use of TRC for building new pedes-
trian bridges and strengthening existing RC structures [3]. Based on these first insights,
some major tendencies may be deduced already.
243
In some instances, textile reinforcement is combined with conventional steel reinforcement

such that in case of the strengthening or repair of RC members by means of TRC, e.g., in
cases when a TRC layer is used to build an integrated formwork. In some other instances, a
combination of textile reinforcement positioned near surface of a structural member and
steel reinforcement (also pre-stressed one) positioned deeper inside may be used to achieve
optimum solutions in new construction with respect to load-carrying capacity, durability,
serviceability and aesthetics. Thus, the protection of steel from corrosion must be consid-
ered with regard to the durability of such structures as well.
Depending on the target applications, durability requirements for different structures made
of TRC may vary significantly. To meet these requirements best, the material composition
of TRC may, in turn, vary considerably as well. This implies that a general framework for
dealing with the problem of durability of such structural elements and structures is needed.
This article is an attempt to provide such a framework.
CHARACTERISTIC LOADS AND EXPOSURES

The wide variety of potential applications for TRC means that the range of mechanical and
environmental loads, as well as their possible combinations, may be very broad. Particular
performance requirements may vary even for different parts of one structural member.
In considering durability of structures, the mechanical loads are of particular interest with
regard to cracking since cracks dramatically increase the transport of fluids and gases in
concrete and consequently the corrosion process. Apart from the mode of loading, e.g.,
tension, shear, and/or bending, all of which affect the crack pattern and the crack width, the
loading type should be considered since, for instance, single overloading, fatigue loading,
or high sustained loading can affect the cracking behaviour of TRC (crack width and crack
distribution).
In most cases, cracking is inevitable in concrete structures and is generally essential for the
effective use of reinforcement, both by steel bars and textile. A great advantage of TRC is
that this group of materials display well-controlled crack width, which is very small in
comparison to conventionally reinforced structures. This enables, at least theoretically, the
full decoupling of the steel reinforcement if it used in combination with textile reinforce-
ment from crack control. Steel reinforcement is used only to ensure adequate load capacity,
while TRC controls cracking behaviour. Obviously, TRC can be simultaneously used for
both purposes: increasing the load-carrying capacity of structural members and the crack
control. These constellations open new opportunities for structural design, thereby allow-
ing a very effective use of such hybrid reinforcements.
BASICS OF DURABILITY ESTIMATION AND DESIGN

General remarks
While discussing the basics of durability design for structures with textile or hybrid rein-
forcement (steel and textile), reference is made to existing durability design rules for steel-
reinforced concrete structures. The prescriptive design rules of current standards are un-
likely to be suitable for taking full advantage of the new class of materials presented in this
book. A performance-based durability design approach will likely be much more appropri-
ate. While such an approach exists for crack-free ordinary concrete (DuraCrete 2000 [4],
244
fib bulletin 34 [5]); significant efforts will be required to develop a similar approach suita-
ble for cracked and crack-free TRC and RC/TRC structures.
Recently Altmann et al. [6] and Altmann & Mechtcherine [7] took a first step towards such
a design concept by developing a fuzzy-probabilistic approach to forecast chloride ingress
into another novel cementituous material, strain-hardening cement-based composite
(SHCC). This approach allows the quantification of parameters based on limited data and
expert knowledge. The same approach can be followed with regard to the durability of
TRC as well as RC structures strengthened by TRC.
In order to ensure structural durability, the durability of the applied materials should first
be assured. Since we are considering also a combination of steel reinforcement and TRC
here, the following requirements arise with regard to the durability of individual system
components and consequently of the system as a whole:
1) Protection of the steel reinforcement from corrosion
2) Durability of the TRC matrix
3) Durability of fibre and polymer yarn coating
4) Durability of the fibre-matrix bond properties.
Protection of steel reinforcement from corrosion

Ordinary steel reinforcement requires anti-corrosion protection, which the alkaline envi-
ronment of the cementitious matrix provides. If aggressive agents such as chloride or car-
bon dioxide penetrate the matrix and reach the reinforcement, however, the steel may cor-
rode. In established codes and regulations (e.g. Eurocode 2), three measures are demanded
to protect steel reinforcement: limitation of the maximum crack width, sufficient concrete
cover, and appropriate concrete quality as defined by corresponding threshold values (min-
imum compressive strength, maximum water-to-cement ratio, minimum cement content).
In principle, these requirements also hold true for RC structures with a repair or strength-
ening layer made of TRC as well as for structures with hybrid (steel bars and textile); how-
ever, the weighting and the thresholds values change accordingly.
Comparing this consideration to existing codes, the following observations can be made:
- Intrinsically small crack widths in TRC or RC/TRC structures are much smaller than
the threshold values allowed by codes for RC structures. The question arises as to
whether or not adjustments should be made as regards the other two protection
measures of the steel reinforcement, namely the thickness of the concrete cover, and
concrete quality.
- The requirement of a particular minimum value of concrete cover is based on numer-
ous considerations, including the carbonation of concrete, the transport of aggressive
agents through the concrete layer to steel reinforcement, and resistance of the concrete
to spalling. In the case of RC/TRC, a much higher spalling resistance than that of ordi-
nary concrete can be achieved with only a very thin TRC overlay.
- The lack of coarse aggregates means that there are no large aggregate-matrix interfac-
es with higher porosity than aggregate-free, hardened cement paste, which might serve
as “highways” for the transport of fluids and gases. However, due to their higher binder
content, the overall porosity of TRC may be more elevated than that of ordinary con-
245
crete. Certainly, this depends on the mix’s water-to-binder ratio, which is usually rela-
tively low for TRC. Furthermore, for TRC it was shown that multifilament yarns pro-
mote water transport actively [8].
- TRC usually possesses higher binder content than ordinary concrete. The associated
higher ability to bind aggressive substances such as chlorides and carbon dioxide is ad-
vantageous with regard to preventing steel corrosion.
In conclusion, it can be stated that for the same exposure classes (e.g., XC carbonation, XD
de-icing salt, or XS seawater according to CEN EN 206-1) the threshold values for TRC
would probably differ significantly from the corresponding values for ordinary concrete if
the established approach for durability design would be applied to RC/TRC structures.
Durability of the matrix

Frost attacks, chemical attacks by aggressive substances, and, in some cases, abrasion must all
be considered with regard to the durability of the TRC matrix. The basic effects of matrix
composition on concrete resistance against such exposure are reasonably well known for or-
dinary concrete. However, they cannot be simply extrapolated to predict the behaviour of
TRC because of its very particular mixture composition, i.e., the absence of coarse aggre-
gates, the high binder content, and the presence of fibres. Relatively fine textiles, especially
used in combination with short fibre, see e.g. [9] might hinder the deterioration of the compo-
site to some extent even if the matrix is severely loaded. The first results on the durability of
TRC subject to the freeze-thaw exposure showed tendencies basically known for ordinary
concrete, e.g., the frost resistance of TRC decreases with increasing substitution of Portland
cement by fly ash and particularly blast furnace cement.
A chemical attack might very well affect the mechanical properties of the matrix, which
would then lead to a change in TRC strength and, especially, in their ductility. A substan-
tial amount of research is still required to investigate the behaviour of these materials under
various exposures.
Durability of fibre and polymer coating

Alkali-resistant (AR) glass fibre and carbon fibre are materials used most often in the man-
ufacture of TRC. There is also research on the use of polymer and natural fibre for this
purpose. Carbon fibre is well known for its high chemical resistance; no degradation of
properties could be observed due to usual chemical environments. The situation with fibre
made of AR-glass is however much more complex. Standard AR-glass contains 16-20%
zirconium dioxide by mass and thus shows significantly enhanced resistance in highly al-
kaline environments as compared to conventional glasses. The corrosive damage of AR-
fibre occurs non-uniformly due to different element concentrations near the filament sur-
face. Organic polymer sizes applied to filament surfaces during the production of glass
yarns can delay such corrosion significantly. Scheffler et al. [10] showed that the corrosion
of AR-fibres in a cement solution in accelerated tests is characterized by the formation of
holes of different size and depth. According to Butler et al. [11] flaws in the size and in-
consistencies in the composition of the bulk glass as well as high pH values of the pore
solution are necessary prerequisites for local damage to fibre made of AR-glass. Corrosion
of glass filaments embedded in fine-grained concrete could be observed only as rare excep-
tions. The authors conclude that stress corrosion (delayed failure) at nanoscopic defects on
the filament surface becomes an important mechanism for damage if, at least at some areas
246
of filament surface, the size is widely removed. However, in the multifilament yarns used
in TRC the failure of individual filaments due to corrosion can likely be compensated by
the redistribution of the load to neighbouring filaments. This means that the failure of a
few filaments cannot affect the load-bearing capacity of the multi-filament yarn signifi-
cantly [12].
Fibre-matrix bond durability

For TRC with textiles made of AR-glass it was found [12] that the extent of the observed
performance losses with increasing aging duration depends primarily on the matrix’s alka-
linity, which influences the formation of solid phases in the fibre-matrix interface and on
the filament surface. The thickness and brittleness of these shell-like crusts increase with
increasing potential for portlandite formation in the matrices. Also amplified precipitation
of Ca(OH)2 between the filaments can be observed with increasing alkalinity of the matrix,
cf. Figure 1a.
a) b)
Figure 1: ESEM images of AR-glass filaments in matrix made of a) Portland cement (here after 28 days of
accelerated aging, and b) blast furnace cement and pozzolana (here after 360 days of accelerated aging) [12].
These filament incrustations reduce the slip of filaments in the vicinity of the cracks, so
that failure strain of filaments is reached at smaller crack widths. Furthermore, straining
the filaments causes local spalling in and on the shells as well as the splitting of the
Ca(OH)2 crystals. The resulting singularities and lateral pressure caused by the wedged
crystals act as notches, leading to premature failure of the filaments. However, when the
Portland cement clinker content of the binder is low and pozzolanic additives are used (low
alkalinity), a stratum of thin CSH phases leading to good bonding characteristics domi-
nates the fibre–matrix interface, cf. Figure 1b. Such thinly walled coatings of the outer
filaments enable good, flawless bonding with the matrix. In the vicinity of matrix cracks a
partial debonding and slip of filaments is possible, leading to ductile overall behaviour of
crack bridging yarn [12].
Furthermore, during continued hydration and aging a slow and gradual process of deposi-
tion of hydration products between the filaments may change the nature of bonding by in-
creasing the sleeve/core ratio [13]. This results in a stronger bond, which is a favourable
effect but may also lead to embrittlement, mainly when brittle fibres such as glass are used.
247
This means that the bonding changes over time and thus also the composite properties and
durability.
CHARACTERISTIC MATERIAL PROPERTIES TO PREDICT LONG-TERM DURABILITY AND

SERVICE LIFE
General remarks
Based on the considerations regarding the durability of individual components outlined in
Section 3, the following characteristic properties of the composite material seem to be de-
cisive with regard to long-term durability and service life of TRC and RC/TRC structures:
- Transport properties of TRC in uncracked and cracked states
- Effects on tensile strength and strain capacity of TRC
- Resistance of TRC in aggressive environments.
On the material level these issues are addressed in previous sections. The following sec-
tions provide a brief overview as well as special considerations for elements and structures
made of TRC.
Transport properties
The key question is what effect a considerable number of fine, well distributed cracks have
on the transport properties of TRC. Average crack width has been used routinely as a ref-
erence parameter. However, for materials with multiple cracks of self-limiting width other
approaches seem to be more appropriate, e.g. using the cumulative crack length [14] or
crack density functions with specific consideration of larger cracks. Where this is not prac-
tical, average strain may serve as a useful surrogate. However, it can be assumed that the
fine reinforcement in combination with a fine grain matrix also affect the condition of
crack profiles and the continuity of the entire crack system, which of course influences the
transport of fluids and gases through the material. There is significant need for research
into reliable and practical methods to describe the crack system in TRC both in laboratory
as well as in field applications.
There is very little information available on the air permeability of TRC. The coefficient of
oxygen permeability in crack-free TRC specimens was found to be influenced clearly by
the polymer coating of the textile but was only marginally dependent on the fineness of
multifilament yarns in the textile [8]. In-situ permeation tests on cracked TRC subjected to
uniaxial tensile loading revealed a pronounced non-linear increase in the transport rates of
oxygen through the TRC (perpendicular to the composite surface) with increasing strain,
corresponding to larger crack widths [8]. A higher degree of fineness in the textile rein-
forcement, i.e., a higher degree of reinforcement, and the polymer coating of the multifila-
ment yarns led to a larger number of cracks in the TRC specimens at given strain levels,
but these cracks were considerably finer. As a result, the permeability for oxygen through
the TRC made with such reinforcement was lower.
Lieboldt and Mechtcherine [15] investigated the transport mechanisms of water in and
through composite concrete specimens made of a cracked ordinary concrete (OC) as a sub-
strate and textile reinforced concrete (TRC) as a cover layer for its strengthening and re-
pair. The TRC cover layer was assessed with regard to its efficiency as a protective layer
against the ingress of water. Since in real applications such TRC layers may actually be or
248
presumed to be cracked, thereby activating the load-carrying function of the textile rein-
forcement, the TRC layer was cracked for purposes of this study. The water transport in
the OC specimens without TRC layer was used as reference. Neutron radiography served
as the main testing technique. In ordinary concrete, quick and deep ingress of water
through relatively wide macro-cracks (100 to 200 µm), followed by transport through the
capillary pore system, caused saturation of large areas in a rather short time. TRC applied
to the OC surface reduced the ingress of water to a large extent. Its small crack widths of
approximately 20 µm modified the suction behaviour fundamentally. In the cracked sub-
strate of ordinary concrete, capillary suction was obviated, and transport through the pore
system of the matrix became the prevailing transport mechanism, based on the suction
force characteristic of capillary pores.
So far, very little information is available on the chloride permeability of TRC. The ob-
served higher resistance of cracked and uncracked TRC against the ingress of chloride ions
can be attributed to a higher content of cementitious materials, a low w/c, the self-limiting
crack width and, as a result, self-healing of the cracks. For sound, ordinary concrete nu-
merous correlations were experimentally and in part also theoretically derived between
different transport mechanisms (diffusion, permeation, absorption) and for different
transport media (e.g. water, air, CO2) [16]. For cracked concrete, however, such compari-
sons have not yet been performed, as is the case for TRC. The knowledge of possible cor-
relations is essential for the prediction of the durability of structures made of or strength-
ened with TRC, if it is based on the results of just one or two particular test methods.
Strain capacity of TRC

Since high strain capacity is a decisive property with regard to the durability of TRC, these
materials as well as members or structures made of them can only be considered as truly
durable if their strain capacities do not change negatively at a substantial rate over time.
Butler et al. [12] investigated the effect of matrix composition by varying the hydration
kinetics and alkalinity of the binder mix on the aging of TRC. The results of tensile tests on
accelerated aged specimens made of TRC showed a pronounced decrease in the tensile
strength and strain capacity for TRC whose matrix was most alkaline (Portland cement was
used exclusively as binder in this composition); see Figure 2a. The performance of TRC
made with a modified, alkali-reduced matrix composition was to a great extent unaffected
by exposure to accelerated aging, cf. Figure 2b. The authors showed that reductions in the
toughness of TRC could be attributed mainly to the observed disadvantageous new for-
mation of solid phases in the fibre–matrix interface, cf. also Section 3.5.
a) b)
249
Figure 2: Tri-linear representations of stress-strain curves for TRC made with a) Portland cement matrix, and
b) blast furnace cement matrix with addition of pozzolana; after demoulding all specimens were stored for 6
days in water (20°C); subsequently, the reference specimens were subjected to storage at 20°C/65% RH until
the testing in the age of 28 days; the other part of the specimens was accelerated aged in a fog room
(40°C/99% RH); the duration of the fog room storage was 28, 56, 90, 180 and 360 days, respectively [12].
Self-healing of cracks
The self-healing of cement-based materials is a well-known phenomenon. In ordinary con-
crete the crack widths are rather large and the water/cement ratio is relatively high; the
deposit of CH and calcite dominates the self-healing process. Larger cracks cannot be
sealed completely. In contrast, TRC possess very small crack widths because of their ten-
dency to multiple cracking. For these kinds of cementitious composites a complete sealing
of cracks and even a partial restoration of stiffness seem to be attainable to a great extent.
The restoration of the stiffness is due to the bridging of fine cracks by the newly developed
C-S-H phases.
In order to investigate the effect of self-healing on the transport properties of TRC, water
permeation measurements were carried out by Mechtcherine & Lieboldt [8] in the unload-
ed, saturated state of TRC specimens at time intervals of 7 days. After storing the cracked
specimens in water for 14 days, the flow rate was reduced to less than 50% of the value
obtained on the same specimen before storing in water. No measurable volume flow oc-
curred after continued water exposure of 21 to 35 days, the particular duration depending
on the type of the textile reinforcement used. The samples containing uncoated AR glass
yarns with low yarn fineness (1280 tex) needed a longer time until a negligible volume
flow was reached, which can be traced back to the initially wider cracks in these speci-
mens. The specimens containing coated carbon textile layers showed the fastest self-
healing, which could be expected since they had the finest cracks.
SUMMARY
There is great potential for the application of textile reinforced concrete in structural ele-
ments and structures exposed to severe mechanical or environmental loading. Present
knowledge indicates that TRC can exhibit a high long-term performance and favourable
transport properties in the cracked state. Beyond that, TRC might show in many instances
superior resistance to aggressive environments when compared to ordinary concrete. How-
ever, with the exception of exposure to alternating freezing and thawing, there is no infor-
mation available on the effects of aggressive environments on the mechanical properties of
the materials.
Since TRC is new material, there is no information available on the long-term field per-
formance of elements and structures made from or strengthened with TRC. However, first
field applications indicate that the superior performance in a laboratory environment as
discussed in previous Sections and summarised above, does indeed translate to superior
field performance.
There is significant research needed to understand the long-term behaviour of TRC and
RC/TRC elements under complex combined loads as encountered in the field. To be able
250
to utilise the superior qualities of this new material fully, it will be necessary to develop a
realistic, reliable, performance-based durability design concept for such structures.
ACKNOWLEDGEMENTS
A considerable part of the work presented in this paper was initiated in the Collaborative
Research Centre SFB 528 ‘‘Textile Reinforcement for Structural Strengthening and Retro-
fitting’’ financed by the German Research Foundation (DFG). The author would like to
acknowledge with gratitude the foundation’s financial support.
REFERENCES
[1] Brameshuber, W., editor. Textile Reinforced Concrete. RILEM State-of-the-Art Report. rep036;
RILEM Technical Commitee 201-TRC; 2006.
[2] Mechtcherine, V.: Novel cement-based composites for the strengthening and repair of concrete struc-
tures. Construction and Building Materials 41 (2013) 365–373.
[3] Weiland, S., Ortlepp, R., Hauptenbuchner, B., Curbach, M.: Textile Reinforced Concrete for Flexural
Strengthening of RC-Structures - Part 2: Application on a Concrete Shell. In: ACI SP-251-3 "Design
& Applications of Textile-Reinforced Concrete", 2008, 41-58.
[4] European Brite-Euram Programm. Probabilistic Performance Based Durability Design of Concrete
Structures. DuraCrete Final Technical Report; 2000 Report No.: BE95-1347/R17.
[5] fib bulletin 34. Model code for Service Life Design. International Federation for Structural
Concrete (fib); Lausanne, 2006.
[6] Altmann, F., Sickert, J.-U., Mechtcherine, V., Kaliske, M.: A fuzzy-probabilistic durability concept
for strain-hardening cement-based composites (SHCCs) exposed to chlorides: Part 1: Concept devel-
opment. Cement and Concrete Composites 34 (2012) 754-762.
[7] Altmann, F., Mechtcherine, V.: Durability design strategies for new cementitious materials. Cement
and Concrete Research 54 (2013) 114–125.
[8] Mechtcherine, V., Lieboldt, M.: Permeation of water and gases through cracked textile reinforced
concrete. Cement and Concrete Composites 33 (2011) 725-734.
[9] Barhum, R., Mechtcherine, V.: Influence of short dispersed and short integral glass fibres on the me-
chanical behaviour of textile-reinforced concrete. Materials and Structures (2013) 46:557–572
[10] Scheffler, C., Förster, C., Mäder, E., Heinrich, G., Hempel, S., Mechtcherine, V.: Aging of alkali-
resistant glass and basalt fibres in alkaline solutions: Evaluation of the failure stress by Weibull distri-
bution function. Journal of Non-Crystalline Solids 2009;355(52-54):2588-2595.
[11] Butler, M., Mechtcherine, V., Hempel, S.: Experimental investigations on the durability of fibre-
matrix interfaces in textile-reinforced concrete. Cement and Concrete Composites 2009;31(4): 221-
231
[12] Butler, M., Mechtcherine, V., Hempel, S.: Durability of textile-reinforced concrete made with AR
glass fibre - Effect of the matrix composition. Materials and Structures 2010;43:1351–1368.
[13] Cohen, Z., Peled, A.: Controlled telescopic reinforcement system of fabric–cement composites – Du-
rability concerns. Cement and Concrete Research 2010;40: 1495–1506
[14] Mechtcherine, V., Lieboldt, M., Altmann, F.: Preliminary tests on air-permeability and water absorp-
tion of cracked and uncracked strain hardening cement-based composites. Proceedings of the Interna-
tional RILEM Workshop on Transport Mechanisms in Cracked Concrete, Ghent, Belgium. RILEM
Publications S.A.R.L.; 2007, p. 55-66.
[15] Lieboldt, M., Mechtcherine, V.: Capillary transport of water through textile-reinforced concrete ap-
plied in repairing and/or strengthening cracked RC structures. Cement and Concrete Research 52
(2013) 53-62.
[16] Hilsdorf, H.K., Schönlin, K., Tauscher, F.: Dauerhaftigkeit von Betonen (in German). Düsseldorf:
Beton-Verlag GmbH; 1997.
251
252
THE MECHANICS OF NATURAL JUTE TEXTILE REINFORCED

CONCRETE
Maria Ernestina Alves Fidelis (1), Flávio de Andrade Silva (2), Romildo Dias Toledo Filho (1),
Steffen Müller (3) and Viktor Mechtcherine (3), (1) Department of Civil Engineering, Universidade
Federal do Rio de Janeiro (COPPE/UFRJ), P.O. Box 68506, 21945 – 970, Rio de Janeiro, RJ,
Brazil, (2) Department of Civil Engineering, Pontifícia Universidade Católica do Rio de Janeiro
(PUC-Rio) Rua Marques de São Vicente 225, 22451-900 Rio de Janeiro, RJ, Brazil, (3) Institute of
Construction Materials, Technische Universität Dresden, 01062 Dresden, Germany
Abstract: This paper presents the preliminary results of an experimental program on the
mechanical characterization of textile cementitious composites reinforced with natural jute
fabrics. A matrix with a low calcium hydroxide content was used. A fiber surface treatment
with styrene butadiene polymer was carried out in order to improve the fiber-matrix inter-
face. Fiber-matrix pullout tests were performed inside an environmental scanning electron
microscope to investigate the bond strength. The newly developed composites reinforced
with 3 and 5 layers of jute fabrics were tested under tensile load. The pullout results
showed that the polymer coating increased the fiber-matrix chemical affinity. The compo-
sites showed a strain hardening behavior accompanied by the formation of multiple cracks.
INTRODUCTION
High performance cementitious composites can be obtained by using continuous fibers as

reinforcement. These composites can be produced with fiber in the fabric form which is
impregnated by the cementitious matrix. This technology has been applied, with limitation,
in the production of thin cement based panels and in applications such as repair and reha-
bilitation [1] [2] [3].
Textile reinforced concrete (TRC) with fibers of glass, carbon, basalt, polyethylene, poly-
propylene, Kevlar and aramid has been widely studied. TRC reinforced with those fibers
presents good mechanical performance with high ultimate tensile stress, strain hardening
and multiple cracking behaviors [4] [5] [6] [7] [8]. Regarding the use of natural fiber TRC
there are only a few studies published [9].
Some requirements are demanded for the matrices used for TRC: chemical compatibility
with the fabric reinforcement, suitable consistency for the appropriate penetration into the
fabric and adequate mechanical properties. Different projects and applications may require
special preparations [10]. Furthermore, in the case of natural fibers, the matrix should have
a low content of calcium hydroxide in order to prevent fiber degradation. The water pre-
sent in the matrix pores is highly alkaline and it is responsible for hemicellulose and lignin
dissolution and alkaline hydrolysis of the cellulose molecules, reducing the degree of
polymerization and fiber strength. Moreover, the hydration products, mainly calcium hy-
droxide, migrate to the walls, lumens and voids of the fibers, causing their mineralization.
Modifications in matrix such as partial replacement of cement by pozzolanic materials,
have been performed in order to obtain a matrix with a low calcium hydroxide content. The
253
degradation of the composites can also be attributed to damages occurred in the fiber-
matrix interface due to volume changes of the fibers [11] [12] [13] [14] [15].
Besides modification in the matrix, fiber treatments are performed in order to remove im-
purities onto the fiber surface, improve the volumetric stability and fiber-matrix stress
transfer. Several treatments can be applied, such as fiber hornification [16] [17], alkaline
treatment [18] [19], and impregnation with blocking agents and water repellents [20]. In
the present research a polymer coating with styrene-butadiene was used. This polymer is
commonly used in the impregnation of glass and carbon fibers [21] [22]. It forms a film
that protects the fiber, preventing its contact with the hydration products.
This research program aims to investigate the fiber-matrix interface through pullout tests
performed inside an environmental scanning electron microscope, and to study the
mechanical behavior of textile cementitious composites using the jute fabric as
reinforcement. The fibers were treated with styrene butadiene polymer. In addition to the
fiber treatment, a matrix with a low calcium hydroxide content was used in the composites
production. Direct tensile tests were performed on specimens reinforced with 3 and 5
layers of uncoated and coated fabrics.
MATERIALS AND PROCESSING
Jute Fibers
Jute can be commercially found in the form of fabric. It is extracted from the stem of the
plant Corchorus capsularis by a combination of processes comprised of the following
steps: cutting, retting, shredding, drying, packing, and classification. The Styrene-
butadiene polymer was used to impregnate the fibers. The fibers were impregnated at the
laboratory for 10 and 50 minutes. Table 1 presents the properties of coated and uncoated
jute fiber. The uncoated and coated fibers were used for preparation of the specimens for
pullout tests (filament and yarn) and for the textile composites.
Table 1: Physical properties of the jute fiber

Filament Yarn Fabric
Properties Uncoated Coated Uncoated Coated Uncoated Coated
Fineness [tex*] - - 326 639 - -

Number of filaments - - 141 141 - -
Number of yarns - - - - 16 16
Area [mm²] 0.0034 0.0036 0.48 0.51 7.68 8.16
*Mass in g of 1 km yarn (tex = g/km).
Matrix
The materials employed in the development of the matrix were: Portland cement, fly ash,
metakaolin, river sand with maximum diameter of 1.18 mm and density of 2.68 g/cm³, a
naphthalene superplasticizer and styrene-butadiene polymer for impregnation with a con-
tent of solids of 48.5%. The Portland cement CPII F-32 defined by the Brazilian standard
[23] as composed with filler (in mass: 85% < clinker < 91%; 3% < gypsum < 5%; 6% <
254
filler < 10%) with a 28 days compressive strength of 32 MPa was used. The metakaolin
used in this work was obtained from Metacaulim do Brasil Industria e Comercio Ltda and
the fly ash is produced commercially by Pozo Fly. The superplasticizer used was the Hor-
mitec SP 430, from Anchortec Quartzolit with content of solids of 41%, density around
1.21 kg/dm3 and action time of 45 minutes. Table 2 summarizes the matrix composition.
The mortar matrix used in this study presented a mix design of 1:1:0.4 (cementitious mate-
rial:sand:water by weight), with 2.5% of superplasticizer and partial cement replacement
(40% metakaolin, 10% fly ash).
Table 2: Matrix composition
Mix ingredients Mix composition [kg/m³]
Cement CP II F32 447
Metakaolin 358
Fly ash 89
River sand 895
Water 358
Superplastisizer 22
PREPARATION OF SPECIMENS AND TEST SETUP
Pullout tests
The pullout tests were performed at the TU Dresden laboratories. Polymer coated and nat-
ural jute filaments and yarns were tested. The preparation of the specimens with filaments
are described as follows. First the fibers were separated in monofilaments and inserted into
the cannulas. The fibers were then fixed at the cannulas with wax and inserted into the
mold (see Figure 1a). The studied embedment lengths of filament specimens were 3, 5 and
10 mm. For the yarn preparation, they were inserted directly in the mold as illustrated in
Figure 1b and fixed with modeling clay. The studied embedment lengths of the yarn were 5
and 10 mm. A PVC mold was used, which allowed the production of 10 specimens at a
time. The specimens were demoulded after 24 hours ant stored in a climate-controlled
room at 20°C and RH 65 % up to a testing age of 28 days.
255
Yarn
Modelling
clay
Filament
(a) (b)
Figure 1. The pullout test molding process: (a) filaments and (b) yarns.
The pullout tests were performed inside an environmental scanning electron microscope
(ESEM). This allowed the observation of the fiber-matrix interface microstructure during
the test. The ESEM model XL30 (FEI Philips Company, Netherlands) was operated under
a low vacuum. All the micrographs were taken under the gaseous secondary electron
(GSE) detector mode at accelerating voltage of 20 kV and chamber pressure of 0.6–0.8
Torr. No coating with carbon or gold, as is customary for a high-vacuum SEM, was re-
quired. For testing, the specimen was fixed to the clamps, and the fiber was glued using a
commercial fast adhesive to the upper mounting plate of a testing machine with a 5 N or 50
N load cells, used for filament and yarn specimens, respectively. The tests were carried out
with a displacement rate of 0.36 mm/min and the deformations were measured by means of
the displacement of the machine cross-head.
Textile composites
Cementitious composites reinforced with 3 and 5 layers of uncoated jute fabric, and com-
posites reinforced with 5 layers of fabric impregnated with polymer were produced in or-
der to verify the efficiency of the treatment.
Before production, the jute fabrics were impregnated with polymer. The fiber treatment
consisted of immersing the fabric in the styrene butadiene polymer for 10 and 50 minutes
and subsequently stretching them on a wooden frame with nails. The fabrics were then
dried at a temperature of 34 °C for 24h. The impregnation time of 50 minutes was chosen
based on the result of absorption tests.
The mixtures were produced using a bench-mounted mechanical mixer of 20 L. The ce-
mentitious materials and sand were dry mixed during 3 min for homogenization at 125
rpm; then the superplasticizer diluted in water was added and mixed for two minutes. Sub-
sequently, the homogenization of the mixture was carried for two minutes at the speed of
220 rpm and during 1 min at speed of 450 rpm.
256
A mold specially developed for this work was produced with dimensions of 440 mm x 580
mm, for producing the textiles composite systems. To ensure the thickness of the compo-
site it was used the value of the density of the fresh mixture for calculating the mass per
layer. The reinforcement was fixed in the screws of the mold. The demoulding was per-
formed 24 hours after molding with the help of scissors and a knife. The composite plate
was kept in a cure chamber with 100 % RH and 21 ± 1 ° C, until one day before the date of
the mechanical tests.
The direct tensile tests were performed on an electro-mechanical testing machine with dis-
placement control at a rate of 0.5 mm/min. The plates were cut and three specimens with a
geometry of 450 mm x 40 mm x 12 mm (length x width x thickness) were tested at the age
of 28 days. The displacements obtained during the test were measured by two LVDT´s
positioned as shown in Figure 2. The specimens were fixed in steel plates and a camera
was attached to the test for acquisition of images which were used to correlate the spacing
of cracks with deformation and applied force.
LVDT
Steel plates
Figure 2. The setup used for the direct tensile tests in the TRC.
Pullout tests results

Figure 3 shows the typical force vs. slip curves for coated and natural filaments, for the
embedment lengths of 3, 5 and 10 mm. The Polymer coating increased the pullout bond
strength for all lengths. The treatment causes a strong chemical adhesion, promoting rup-
ture of the fibers: the fracture energy is reduced while the resistance increases, making the
material strong, but brittle.
257
Figure 3. Typical pullout load vs. slip curves for filaments at embedment lengths of 3, 5 and 10mm – uncoat-
ed and coated fibers.
Figure 4 shows the micrographs for the uncoated filament inside the microscope. The im-
ages were taken at different moments during the test. Figures 4a and 4c are images ob-
tained at the start of the pullout tests and Figures 4b and 4d at the end of the tests. For all
embedment lengths, it was observed fiber fracture for coated specimens (Figures 4c and
4d).
The maximum bond of uncoated jute fiber presented similar behavior to that of sisal fibers,
as shown by Silva et al. [24]. The authors investigated the influence of the embedment
length on the fiber-matrix bond. They observed that the adhesional bond remained con-
stant, about 0.36 MPa, at 3 days of cure when increasing the embedment length. The sisal
fiber presented an adhesional bond of 0.67 MPa at 28 days.
(a) (b)
258
(c) (d)
Figure 4. Micrographs obtained during the fiber-matrix pullout tests, inside the ESEM, for filaments with
embedment lengths of 3mm (a) uncoated fiber – start of the test (b) uncoated fiber – end of the test (a) coat-
ed fiber – start of the test (b) coated fiber – end of the test.
Figure 5 presents the force versus slip curves of the yarn. The polymer treatment increased
the pullout bond strength from 0.99 MPa to 1.25 MPa, for embedment length of 5 mm and
from 1.05 MPa to 1.08 MPa for 10 mm. The adhesion of the fiber in the filaments form is
completely different of the multifilaments (yarn / fabric). In the yarn only the outer fila-
ments have contact with the matrix, then friction occurs between the inner filaments. The
polymer treatment increases the chemical fiber-matrix affinity and causes activation of the
inner filaments, improving the fiber-matrix adhesion [10].
(a) (b)
Figure 5. Typical pullout load vs. slip curves of yarn at the embedment length of (a) 5mm (b) 10mm – un-
coated and coated fibers.
Figure 6 shows the micrographs of the yarn, inside the ESEM, at the end of the tests, for
uncoated and coated yarns. It can be noticed that even without treatment, there was a good
fiber-matrix bond, due the presence of matrix on the fiber surface. It is observed an in-
crease in the force values with increasing of embedment length, and an increment due to
259
the treatment. Also the pullout energy increased substantially for the coated fibers indicat-
ing an increase in the frictional bond. These results are close to that of hemp yarnns, pre-
sented in the study of Asprone et al. [25]. The adhesional bond strength of untreated hemp
fibers was 1.02 and 2.20 MPa for embedment lengths of 40 and 10 mm, respectively, while
for the treated fibers with latex, the adhesion was 1.68 and 2.06 MPa (40 and 10 mm)..
MÄDER et al. [22] observed an increase in the fiber-matrix bond in composites with glass
and carbon fibers, in the form of yarn, impregnated with styrene-butadiene. The authors
report that the improvement in the adhesion may be due to the strong interaction between
the polymer and the matrix, causing variations in the crystalline structure of the interface.
(a) (b)
Figure 6. Micrographs obtained during the pullout tests, inside the ESEM, for the yarn with embedment
length of the 5 mm (a) uncoated fiber – end of the test (b) coated fiber – end of the test.
Composites tensile tests results

Figure 7 shows the stress-strain curves and images of the fracture mode for the natural Jute
TRC reinforced with 3 and 5 layers. The composites showed a high strain capacity and
multiple cracking behavior. The composite reinforced with 3 layers presented 6 cracks
while the composites with 5 layers showed the formation of 11 cracks. The composite rein-
forced with 5 layers of uncoated fabric supported and ultimate tensile stress 8% higher and
strain capacity 8.7% lower than the composite with 3 layers of fabric.
260
M ti 5 layers
3l
(a) (b)
Figure 7. Effect of the reinforcement ratio on the tensile behavior of natural Jute TRC: (a) typical stress vs.
strain curves and (b) cracking pattern.
Figure 8 presents the typical stress vs. strain curves obtained in the tensile tests for the
coated Jute TRC. The fibers were treated by 10 and 50 min. It can be seen that the compo-
sites with treated fabric presented more cracks and smaller cracking spacing compared to
the composite reinforced with 5 layers of uncoated fabric.
The composite with 5 layers of fabric impregnated by 50 min presented the ultimate tensile
stress 46.7% higher than the composite with 5 layers of uncoated fabric as well as strain
capacity 107.6% higher.
5 layers
5 layers 5 layers
treated
without treated
by
treatment by 50
Matrix
min
(a) (b)
Figure 8. Effect of the polymer treatment on the tensile behavior of Jute TRC: (a) typical stress vs. strain
curves and (b) cracking pattern.
261
The toughness of the composites was calculated through the area under the force versus
displacement curves. It is related to the energy absorption capacity of the composite, espe-
cially in the post-cracking region where the performance of the fiber is predominant. The
composite with 5 layers of fabric treated by 50 min presented toughness 155.7% higher
compared to the composite with layers 5 of uncoated fabric. The polymer was responsible
for increasing the ultimate stress and strain capacity. This type of treatment appears as a
solution for application in composites reinforced with vegetable fibers in matrices with a
low content of calcium hydroxide.
CONCLUSIONS
The fiber-matrix interface was investigated by pullout tests, in order to analyze the fiber-
matrix bond. For filament, the increase in the bond, due to the treatment, was related to
strong chemical adhesion with the matrix. For yarn, in addition to the chemical affinity
there was an activation of internal filaments.Textile composites reinforced with jute fibers
showed a strain hardening behavior with the formation of multiple cracks.
The results from direct tensile tests showed that the treatment with polymer presented an
effective cracking control: it increased the number of cracks and reduced the crack open-
ing. Besides that, the treatment increased the tensile strength and strain capacity of the
composites as well as the composite´s toughness. It has also improved the fiber-matrix
stress transfer mechanisms. Based on the results of direct tensile tests, the use of jute fabric
as reinforcement in cementitious composites showed a high potential for the development
of thin walled elements for the construction industry.
ACKNOWLEDGMENT
The authors gratefully acknowledge the CNPq and Capes (Brazilian National Science
Foundations) for partial financial support for this work; the Technische Universitat Dres-
den, Germany, for providing the polymer and enabling the realization of the pullout tests,
through the PROBRAL cooperation.
REFERENCES
1. Faella, C., Martinelli, E., Nigro, E., Paciello, S., ‘Shear capacity of masonry walls externally
strengthened by a cement-based composite material: an experimental campaign’, Construction and
Building Materials 24 (2010) 84-93.
2. Gopinath, S., Kumar, V.R., Sheth, H., Murthy, A.R., Iyer, N.R., ‘Pre-fabricated sandwich panels us-
ing cold-formed steel and textile reinforced concrete’, Construction and Building Materials 64
(2014) 54-59.
3. Mechtcherine, V., ‘Novel cement-based composites for the strengthening and repair of concrete
structures’, Construction and Building Materials 41 (2013) 365–373.
4. Silva, F.A., Butler, M., Mechtcherine, V., Zhu, D., Mobasher, B., ‘Strain rate effect on the tensile
behaviour of textile-reinforced concrete under static and dynamic loading’, Materials Science and
Engineering: A 528 (2011) 1727-1734.
5. Hegger, J., Kulas, C., Hosrtmann, M., ‘Spatial textile reinforcement structures for ventilated and
sandwich facade elements’, Advances in Structural Engineering 15 (2012) 665.
6. Butler, M., Mechtcherine, V., Hempel, S., ‘Experimental investigations on the durability of fibre–
matrix interfaces in textile-reinforced concrete’, Cement and Concrete Composites 31 (2009) 221-
231.
262
7. Peled, A., Mobasher, B., Cohen, Z., ‘Mechanical properties of hybrid fabrics in pultruded cement
composites’, Cement and Concrete Composites 31(2009) 647–657.
8. Rambo, D.A.S., Silva, F.A., Toledo Filho, R.D., Gomes, O.F.M., ‘Effect of elevated temperatures
on the mechanical behavior of basalt textile reinforced refractory concrete’, Materials and Design 65
(2015) 24–33.
9. Hakamy, A., Shaikh, F.U.A., Low, I.M., ‘Characteristics of hemp fabric reinforced nanoclay–
cement nanocomposites’, Cement and Concrete Composites 50 (2014) 27-35.
10. Brameshuber, W., ‘Textile Reinforced Concrete’. State of the art report of Rilem Technical Com-
mittee 201-TRC: textile reinforced concrete. Rilem publications S.A.R.L.2006
11. Gram, H.E., ‘Durability of natural fibres in concrete’, Swedish Cement and Concrete Research Insti-
tute, Stockholm 1983: 255 p.
12. Toledo Filho, R.D., Ghavami, K., England, G.L., Scrivener, K., ‘Development of vegetable fiber–
mortar composites of improved durability’, Cement and Concrete Composites 25 (2003) 185–196.
13. Claramunt, J., Ardanuy, M., García-Hortal, J.A., Toledo Filho, R.D., ‘The hornification of vegetable
fibers to improve the durability of cement mortar composites’, Cement and Concrete Composites 33
(2011) 586–595.
14. Silva, F.A., Toledo Filho, R.D., Melo Filho, J.A., Fairbairn, E.M.R., ‘Physical and mechanical prop-
erties of durable sisal fiber–cement composites’, Construction and Building Materials 24 (2010)
777-785.
15. Savastano Jr, H., Santos, S.F., Radonjic, M., Soboyejo, W.O., ‘Fracture and fatigue of natural fiber-
reinforced cementitious composites’, Cement and Concrete Composites 31 (2009) 232–243.
16. Brancato, A.A., ‘Effect of progressive recycling on cellulose fiber surface properties’. Thesis,
School of Chemical and Biomolecular Engineering, Georgia Institute of Technology. 2008.
17. Ferreira, S.R., Lima, P.R.L., Silva, F.A., Toledo Filho, R.D., ‘Effect of sisal fiber hornification on
the adhesion with portland cement matrices’. Revista Matéria 17 (2012) 1024-1034.
18. Saha, P., Manna, S., Chowdhury, S.R., Sen, R., Roy, D., Adhikari, B., ‘Enhancement of tensile
strength of lignocellulosic jute fbers by alkali-steam treatment’,.Bioresour Technology 101 (2010)
3182-3187.
19. Ray, D., Sarkar, B.K., ‘Characterization of alkali-treated jute fibers for physical and mechanical
properties’, Journal of Applied Polymer Science 80 (2001) 1013–1020.
20. Canovas, M.E., Kawiche, G.M., Selva, N.H., ‘Possible Ways of Preventing Deterioration of Vegeta-
ble Fibres in Cement Mortars in Vegetable Plants and their Fibres as Building Materials’ - Proceed-
ings of the Second Internacional RILEM Symposium, 1990.
21. Scheffler, C., Gao, S.L., Plonka, R., Mäder, E., Hempel, S., Butler, M., Mechtcherine, V., ‘Inter-
phase modification of alkali-resistant glass fibers and carbon fibers for textile reinforced concrete I:
Fiber properties and durability’ Composites Science and Technology 69 (2009) 531-538.
22. Mäder, E., Plonka, R., Schiekel, M., Hempel, R., ‘Coatings on alkali-resistant glass fibres for the
improvement of concrete’, Journal of Industrial Textiles 33 (2004) 191-207.
23. NBR 11578. Cimento Portland Composto. Associação Brasileira de Normas Técnicas (ABNT),
Julho; 1991. [in Portuguese].
24. Silva, F.A., Mobasher, B., Soranakom, C., Toledo Filho, R.D., ‘Effect of fiber shape and morpholo-
gy on interfacial bond and cracking behaviors of sisal fiber cement based composites’, Cement and
Concrete Composites 33 (2011) 814-823.
25. Asprone, D., Durante, M., Prota, A., Manfredi, G., ‘Potential of structural pozzolanic matrix–hemp
fiber grid composites’, Construction and Building Materials 25 (2011) 2867-2874.
263
264
CORROSION IN FERROCEMENT
Paul J Nedwell, Olatona O Damola & Nicholas Stevens, University of Manchester
Abstract:The small cover of ferrocement has been the subject of discussion for many years.
Following a visit to a bridge constructed in 2000 and testing of samples exposed for 20
years which showed that there was virtually no carbonation a test was set up under acceler-
ated corrosion to investigate the different corrosion rates for mesh and traditional bar. The
results showed that after a month of testing the rate of corrosion for the smaller diameter
mesh was reducing whereas that for the traditional steel reinforcement was increasing.
INTRODUCTION
The small cover traditionally used in ferrocement has been a discussion point for many
years. Much of the work on corrosion in ferrocement has been anecdotal and visual rather
than having a rigorous backing [1]. The University of Manchester in 1992 [2] first devel-
oped a method of corrosion potential measurement which was further developed in studies
by Ramesht [3], Nakassa [4] and Vickridge [5]. This work led to the introduction of Silica
Fume to recommended matrices. In 2000 the Colliers Moss Bridge was completed in the
north of the United Kingdom Fig 1[6]. This was visited by the author in 2012 where it was
noted that though the bridge had fallen into a state of disrepair Fig 2 a small wire which
had been at the surface was corroded but that the corrosion had not produced any cracking
even at very small cover depths Fig 3. This contradicted the accepted wisdom that corro-
sion of rebar caused spalling of concrete and that greater cover provided greater protection.
Fig 1 Colliers Moss Bridge Fig 2 Bridge in 2012
During 2012, samples of ferrocement that had been left exposed on the roof of one of the
buildings of the University of Manchester for twenty years were retrieved and tested
both for strength and for carbonation depth. These showed no noticeable increase in car-
bonation over this time.
265
Fig 3 Corroded wire detail Fig 4 Carbonation test on roof sample
A study was therefore devised to investigate the corrosion rates of ferrocement having
small diameter (1.6mm) wire compared to larger diameter (6 mm) ribbed steel reinforcing
bar. Measurements were taken on samples cast and placed in an acceleration tank which
provided an aggressive environment of 3.5% salt solution at 60oC wet and dry cycling eve-
ry 2 hours.
METHODOLOGY
Matrix
A standard mix comprising sand sieved at No 7 (2.63mm) and Ordinary Portland Cement
at a ratio of 2:1 with a water cement ratio of 0.5 was used for the experiment.
Reinforcement
Galvanised square weldmesh having a wire diameter of 1.6mm and mesh opening of
12.5mm was used for the small diameter reinforcement and ribbed rebar of 6 mm diameter
was selected for the larger diameter. These were held in at a frame which set cover at 0, 1
and 2 mm. Fig 4
Fig 4 Frame holding samples Fig 5 Cast sample showing cracks
266
Once cured and de-moulded the samples were inspected for imperfections and it was no-
ticed that there were small shrinkage cracks between the wires in the mesh sample Fig 5 In
order for electrochemical measurements to be taken copper wires were spot welded to each
of the reinforcements. These were shielded whilst in the accelerator tank.
Accelerator tank
A pair of plastic tanks one above the other ware provided with immersion heaters to heat
the liquid and electric fan heaters to maintain the dry temperature. A control system al-
lowed for the change from wet to take place at timed intervals, with liquid being pumped
from bottom to top and falling by gravity from top to bottom. For this experiment a cycle
time of 120 minutes was chosen. To increase the aggressiveness of the environment a 3.5
% salt (NaCl) solution was used. The samples were measured at intervals during a 4 week
period.
MEASURING TECHNIQUES
Due to the relative shortness of time for this experiment it was decided that DC measure-
ments would be more appropriate. Linear Polarization Resistance (LPR) and Potentiody-
namic Polarization (PP) were the techniques available.
Linear polarization resistance

“Polarization resistance, Rp, of a reinforcement embedded in concrete is defined as the
ratio between the applied voltage ∆E (shift in potential from Ecorr) and the step current ∆I,
when the metal is slightly polarised from its free corrosion potential.” [7]
Rp (Wcm2)=∆E/∆i (1)
Where ∆I is the current density = ∆I/A, ∆E is the applied voltage range, ∆I is the step cur-
rent change and A the total surface area of reinforcement.
Measurement is carried out using three electrodes. The working electrode (WE), rein-
forcement, a platinum counter electrode (CE) and a copper/ copper sulphate reference elec-
trode (RE) connected to a potentiostat to apply the voltage and a computer for control and
logging. Voltage was cycled from -0.015v to 0.015v at a rate of 0.001v/s. This produced a
curve, the slope of which is the Rp. Measurements were taken daily.
Potentiodynamic Polarization
As this changes the condition of the specimen due to the large range of potential applied it
was only carried out towards the end of the experimental period. The aim of the test is to
find the Tafel coefficient B. With a technique similar to linear polarization the difference
lies in the potential applied (+/- 200mv compared to +/- 10mv). The value of the Tefal
slope B is derived from
(2)
Where is the slope of the anodic portion of the graph and that of the cathodic por-
tion. B is usually between 26mv for active to 52 mv for passive corrosion
267
Corrosion rate
A corrosion rate is calculated from both values
Icorr = B/Rp (3)
RESULTS AND ANALYSIS

A Saturated Calomel Electrode (SCE) is used as reference to measure the open current
potential.
Fig 6 Mesh reinforcement
Fig 7 Bar reinforcement
268
Fig 6 above for the mesh shows a steady increase as the zinc of the galvanising reacts with
the high level of alkali in the matrix. Fig 7 shows, with the exception of an apparent anom-
aly at 6 six days, a steady state indicating some passivation of the reinforcement.
Polarisation resistance
Mesh
4500
4000
3500
Polarization Resistance
3000
2500
Surface
2000
1 mm
1500 2 mm
1000
500
0
0 5 10 15 20 25
Time
Fig 8 Mesh reinforcement
Bar
25000
20000
Polarization Resistance
15000
Surface
10000 1 mm
2 mm
5000
0
0 5 10 15 20 25
Time
Fig 9 Bar reinforcement
269
-400
-500
-600 mesh surface

mesh 1mm
-700
Potential
mesh 2mm
-800 Bar surface
-900 Bar 1mm

Bar 2mm
-1000
-1100
-1200
-4 -3 -2 -1 0 1 2
Log i (mA)
Fig 10 Polarisation curves for both materials
Figure 10 shows the result of the potentiodynamic technique over a range of +200mVSCE
from the open circuit potential. It shows two distinct regions: the upper region representing
the result for the bar reinforcement and the lower region for the galvanised mesh rein-
forcement.
Table 1 Results
Reinforcement βa βc B (mV) Rp Icorr

(Ωcm2) (mA/cm2)
Mesh(Surface) 138.85 114.65 27.1850 552.055 0.0493
Mesh (1mm) 150.85 129.15 30.121 547.925 0.0550
Mesh (2mm) 211.62 165.85 40.2511 566.02 0.0712
Bar (Surface) 303.56 118.86 36.976 950.77 0.0389
Bar (1mm) 156.54 127.36 30.400 1929.412 0.0157
Bar (2mm) 158.74 128.07 30.6851 9187.008 0.0033
From table 1 above, the corrosion rate of the mesh reinforcement is higher than the single
rod reinforcement. This can be attributed to the stability of the materials in the high alka-
270
line environment with zinc having the least stability. The corrosion rate of the surface rein-
forcements for mesh and single bar showed very close values because of their close prox-
imity to the chloride environment. In the case of both reinforcements 1mm below the sur-
face, the corrosion rate showed a 0.04mA/cm2 difference highlighting the effect of the
concrete cover. The 2mm below surface reinforcements showed the largest difference with
a value of 0.0667mA/cm2.
CONCLUSION
Though the mesh shows a higher corrosion rate the work of Nakassa dicussed the role of
the zinc galvanising on the corrosion rate and concluded that in the initial stages this was
more rapid than would be expected from normal steel.
With the trends of polarisation resistance the mesh showed an upward trend indicating re-
duced rate of corrosion and increasing passivation whereas the steel bar indicated increased
corrosion and reduced passivation.
From this study it may clearly be seen that the subdivision of the reinforcement present in
Ferrocement produces a reduced rate of corrosion and therefore more protection.
REFERENCES
[1] Austriaco L., Ferrocement Corrosion , IFIC, Bangkok, 1987
[2] Nikkhah N, Corrosion Monitoring in Ferrocement Slabs, MSc Thesis, UMIST, Manchester, 1992
[3] Ramesht MH, Vickridge IG & Jafar MI, Electrochemical Corrosion Measurements on Reinforcement in
Ferrocement, Proceedings of the Fifth International Symposium on Ferrocement, pp461-469, UMIST
Manchester Sept 1994 E& FN Spon, London, 1994
[4] Nakassa AS, Development of High Durability Ferrocement, PhD Thesis, UMIST, Manchester, UK 1999
[5] Vickridge IG, Nakassa A & Turner H, High Durability Ferrocement, Proceedings of the Sixth Interna-
tional Symposium on Ferrocement pp 297- 312, University of Ann Arbor, Michigan, June 1998
[6] Nedwell PJ, Adcock IJ & Hope TE, Colliers Moss Millennium Bridge – In Situ Testing, Proceedings of
the Seventh International Symposium on Ferrocement, NUS Singapore, 2001
[7] Martinez C.A and I, Techniques for measuring corrosion rate (polarization resistance) and the corrosion
potential of reinforced concrete structures, Elsevier, 2010
271
272
CHAPTER 6
MAINTENANCE
AND
SUSTAINABILITY
273
274
TEXTILE REINFORCED CONCRETE FOR SEWER

REHABILITATION
Tobias Walther and Manfred Curbach, Institute of Concrete Structures, Faculty of Civil
Engineering, Technische Universität Dresden, Germany
Abstract: A large percentage of the sewer network in Germany is more than 50 years old.
With a rising construction age, the vulnerability to damage also increases. These deficien-
cies have to be repaired as soon as possible to prevent higher secondary damages. Often,
the repair of such deficiencies cannot be appropriately and efficiently carried out with cur-
rent common techniques. The application of textile reinforced concrete (TRC) is a valuable
addition to the existing rehabilitation methods. The advantages of textile reinforced con-
crete are, for example, the good adaptability to many forms, the fine crack pattern and the
high quality and density of concrete. Therefore, it can be used successfully in sewer reha-
bilitation. The paper describes the applicability and technical aspects of a first sewer reha-
bilitation project with textile reinforced concrete. Furthermore, it goes into detail to de-
scribe extensive evaluation tests on the influence of aggressive substances on the strength
of the composite material.
INTRODUCTION
The German Association of water, wastewater and solid waste (DWA e.V.) determines in a
regular cycle of surveys the condition of the sewer-systems in Germany. The results of the
surveys were published to show the importance of rehabilitation and investment in the pub-
lic sewer system. With these surveys and data collection, Germany has taken a European
leading role. The last extensive data collection of the sewer condition in Germany was tak-
en in 2009. The results are published in the DWA-survey 2009 [1].
It is stated in the abstract of the DWA-survey that: “The results underline the need to attach
much greater importance to sewer rehabilitation” [1]. Furthermore, the DWA-survey [1]
gives an overview on the structure age of the sewer. About 28 % of the used sewage sys-
tem in Germany has an age of over 50 years. According to this, repair or rehabilitation
work has to be done. The expected remaining life of the structures at this age is assessed to
be between 10 and 23 years. Additionally, the DWA-survey shows that the trend is going
from replacement to repairs. It should be pointed out that the financial resources are
shrinking, while the quantity of the sewer constructions in need of repair is growing. Also,
the importance of the public sewage system to dispose sewage from private households is
included in the survey. According to the Federal Statistical Office, the connection rate of
private households to public sewage systems amounts to 96 % in Germany [2]. A greater
breakdown of the sewerage could be of unforeseeable consequences to the standard of liv-
ing.
275
In Germany, 72 % of the sewers that can be accessed by walking through (tubes with a
diameter over 800 mm) are made out of concrete, reinforced concrete or a combination of
both.
A large part of the damages in this group of sewers are crack formations (20 %) and sur-
face damages (11 %); damages which could be easily fixed with a thin textile reinforced
concrete (TRC) layer. The following aspects justify the use of TRC in sewer rehabilitation:
• Repair of a concrete canal with a new concrete layer guarantees the compatibil-
ity of materials. Undesirable effects between the rehabilitation layer and the ex-
isting structure due to material incompatibility can be nearly excluded.
• TRC consists of a high-performance concrete matrix. High compressive strength
is characteristic. This performance is due to a dense concrete matrix, which
makes the concrete resistant against penetration of pollutants.
• Textile reinforced concrete results on a finer crack distribution than that of steel
reinforced concrete [3], [4]. Many fine, uncritical cracks arise, instead of a few
critical cracks. Conversely, this results on a lower permeability of the rehabilita-
tion layer.
• Due to the fine crack distribution, a self-healing process is possible (under war-
ranty of further conditions).
• With textile reinforced concrete, many geometric constraints can be accommo-
dated, e.g. [5] - [7]. The TRC layer can be flexible customized to the many
shape requirements of sewers.
The potential applications of textile reinforced concrete in sewer rehabilitation is discussed
in great detail in the following text.
LABORATORY INVESTIGATIONS
SPEPECIMEN
The use of textile reinforced concrete for sewer rehabilitation is a new field of application.
A basic suitability of the material for this purpose was first investigated in the laboratory.
First of all, the eligible materials were scientifically selected. The sewer rehabilitation mor-
tar Ombran MHP from MC Bauchemie (Bottrop) was chosen as the concrete component.
The mixture and characteristics of such mortar are very similar to the products which are
commonly used for TRC, for example, the fine concrete mixture TF 10 from Pagel (Es-
sen). As shown by preliminary investigation, the processing properties, as well as the me-
chanical properties, were comparable with those of the TF 10.
A Low-Tow-textile made out of Carbon was chosen as reinforcement. The yarns in the
warp direction and the fill direction were 800 tex strong with an gap between the yarns of
7,2 mm in every direction. The textile was coated with styrene-butadiene (15 % of mass).
Strain body tests (uniaxial tensile tests) and pull-out test were done. First, big textile rein-
forced concrete slabs, with a size of 1200 mm x 700 mm x 9 mm, were cast. The slabs
were cured after casting in their formwork covered with a film and afterwards in water. At
the 8th day after casting, small specimens to be used in the investigation were cut off. Fig-
276
ures 1 and 2 show the geometry of the samples. More information about the casting of the
specimen are given in [7] and [8].
700 mm
175 mm 350 mm 175 mm
60 mm
test area
load transfer area
1. and 2. layer
cross-section:
textile
3 mm
fine concrete 3 mm
3 mm
Figure 1 : Test specimen for Figure 2 : Test specimen for

uniaxial tensile tests [9] textile pull-out test [9]
Afterwards, the specimens were stored under standard conditions, in water and in an ag-
gressive chemical sulphate solution (29 800 mg SO4/l) until the testing date (see table 1).
The sulphate solution was chosen, because it is a damaging medium for concrete, which is
workable under easy conditions. The measuring of the sulphate concentration was done
according to the research by Wittekindt. The intent of the high concentration of the sul-
phate solution was to simulate a long time exposure to this pollutant with lower concentra-
tion, within the short testing time. The storage time of 28 days was chosen, to compare it to
the standard storage. Additionally, a storage time of 56 days was chosen to investigate the
trend of the behaviour of textile reinforced concrete under sulphate solution. Eventually, it
will be necessary to investigate the long term influence of exposure time on the load-
bearing behaviour of the specimens. An investigation of specimens, which involved longer
time in a sulphate solution, was not possible because of the limited period granted for the
research project. Nevertheless, it would be very interesting and important to evaluate the
long-time behaviour of textile reinforced concrete under extreme conditions.
277
Table 1 : Storage conditions of specimen
Specimen Conditions of storing Time of storing
Standard conditions 20°C, 60 % rel. humidity 20 days
Water storage 20°C, water storage 20 days
Sulphate solution, 28 days 20°C, storage in sulphate solution 20 days
Sulphate solution, 56 days 20°C, storage in sulphate solution 48 days
20°C, 60 % rel. humidity 7 days

Sulphate solution (reference), 28
days
20°C, storage in sulphate solution 13 days
20°C, 60 % rel. humidity 7 days

Sulphate solution (reference), 56
days
20°C, storage in sulphate solution 41 days
TENSILE TESTS
The loading capacity of textile reinforced concrete is determined by an uniaxial tensile test,
known as strain body test. The test setup is described in detail in [8] and [9]. The speci-
mens are fixed between massive steel plates, followed by a consistently, path-controlled
stress until failure of the specimen. The research results, namely the stress-strain-behaviour
of the specimens, are shown in figure 3. Each of the curves shown is an average value of
five tests.
278
Figure 3: Stress-strain relationships of tested specimens, that were stored under different conditions be-
fore the uniaxial tensile tests
The stress-strain-behaviour of the specimens corresponds to well-known and expected be-
haviour. At the beginning, the specimen is not cracked, and the concrete is mainly involved
in load transfer. The textile reinforcement takes over, more and more of the load, after the
first crack. At an ultimate state, when the sample is fully cracked, the textile reinforcement
alone is responsible for load transfer. As it is seen in figure 3, the specimens stored in wa-
ter and in sulphate solution show no variation in behaviour at the serviceability and ulti-
mate limit state as the standard stored specimens. Only the force for the initial crack is a
little bit higher when the specimens were stored in water or sulphate solution. The maxi-
mum strength stress gets lower under sulphate solution storage.
BOND TESTS
An important part of the investigations for textile reinforced concrete is the determination
of the bond parameter. The theory behind the bond behaviour between yarn and concrete
allows calculating anchorage and lapping length, which is needed in every practical work-
manship. The algorithm to calculate anchorage and lap length is described in detail in [11].
It is already known that different storage - and environmental conditions - have extended
influence on the bond behaviour.
The test setup for the investigation is described in [8] and [9]. The specimen, which is
shown in figure 2, is fixed between steel plates. Then, it is set in the testing machine and a
path-controlled force is applied. An extensometer records the crack opening at the prede-
termined breaking point of the specimen. With the help of the length lE,0 (figure 2) of the
yarn and the crack opening, it is possible to visualize the bond flow. The bond flow is cal-
culated as the quotient of the applied force divided by the length lE,0. The basis to deter-
mine the anchorage and lap length is the bond flow-crack opening diagram. Each curve, in
figure 4, represents the average value of five tests.
279
Figure 4 : Bond flow-crack opening relationships of tested specimens

At this early stage, the best results of bond behaviour were obtained for the specimens un-
der standard storage. The graph shows a clear and high adhesive bond. The friction bond is
also at a high level. The water stored specimen shows the worst bond behaviour. An adhe-
sive tensile bond is visible, but at a low level. The friction bond is at same level as the ad-
hesive bond. The specimen stored in sulphate solution shows a better behaviour as the wa-
ter stored specimen, but the behaviour is not as good as that of the specimen under stand-
ard storage. If the specimen is stored longer in a sulphate solution, the bond behaviour is
changing, so an adhesive bond is visible. In other cases, only the friction bond is active.
The reason for the improved bond behaviour of the specimen stored in sulphate solution
compared with that of the specimen stored in water is not exactly clear until now. Due to
sulphate attack, an expansion effect in the concrete is initialised. The yarns of the textile
form a canal in the concrete matrix, where the yarn is laying. Due to the expansion effect,
the diameter of the canal minimizes. The yarn is exposed to a minimal transverse compres-
sion which could maybe improve the bond between the concrete and the yarn. A verifica-
tion of this mechanism is not possible at the moment.
HANDLING-TEST IN A SEWER
To validate the practical application of the method, the urban drainage in Dresden provided
an instance of a sewer section in need of rehabilitation. The section was 70 m long and had
a clearance height of 1.7 m. The profile of the canal had an oval shape with a straight hori-
zontal crown. In figure 5, the deteriorated condition of the sewer is shown. The canal car-
ries mixed water sewerage, which means wastewater and rainwater flow together through
the sewer. At the crown of the sewer section, there were some locations where the built-in
steel reinforcement had corroded. The corrosion products were visible at the surface of the
280
concrete (figure 6). Furthermore, areas of the concrete surface show evidence of heavy
attack by the aggressive components of the waste water. There were no concerns about the
bearing capacity of the structure. The goal of the rehabilitation was to repair the surface of
the sewer and to decrease its permeability. In addition, it was intended for unexpected
stresses to be absorbed by the textile reinforcement, resulting in a finer, denser, and harm-
less crack distribution. Furthermore, the construction workflow, the construction process
and the implementation of details were conducted under realistic conditions.
Figure 5 : initial situation of the sewer Figure 6 : corrosion at the crown

The work took 14 days to be completed. Several interruptions took place because of the
high concentrations of hydrogen sulphide intensive rainfalls, which made working in the
sewer impossible. After cleaning the sewer, the surface was roughened and the locations of
steel corrosion were intensively treated in an appropriate manner before the work with
TRC began.
Two layers of textile reinforcement were built-in by hand lamination. The thickness of the
TRC-layer was 9 mm. The concrete was mixed at the entrance of the manhole and trans-
ported along with the textile reinforcement to the working place by hand.
The result of the work was an excellent surface. The construction process was optimized in
an iterative manner to minimize idle times during construction and to guarantee a smooth
workflow.
281
Figure 6 : work proceeding in the sewer
SUMMARY AND OUTLOOK
As preliminary investigations show, the bearing behavior of textile reinforced concrete

does not change dramatically under chemical influence, for short exposure times. At a
maximum, the bearing capacity is reduced by 20 %. But the storing conditions have a sig-
nificant influence in the bond behavior. Under standard conditions (20 °C, 65 % humidity),
the best bond properties could be achieved, so this was found to be the ideal case. The
storage of samples in water produced the worst bond properties in these investigations.
In a first practical test, the suitability of TRC for sewer rehabilitation was demonstrated.
The material was applied properly, but under hindered conditions. The result of the work
likewise satisfied the installer, persons responsible for this research, and the owner. To
evaluate the behavior of the rehabilitation layers, long time monitoring is already planned.
Consequently, the next step is to use textile reinforced concrete as a strengthening layer in
sewer rehabilitation. But for that, more characteristic values of the behavior of TRC and of
the influence of harsh chemicals are necessary to guarantee the long-time bearing behavior
of the material and the repair.
ACKNOWLEDGMENT
The authors thank the staff of the Otto-Mohr-Laboratory (TU Dresden), Kanal-Service
Dresden GmbH, the coordinator of the research project Mrs. Dr. Gaitzsch and the Federal
Ministry for Economic Affairs and Energy.
REFERENCES
[1] Deutsche Vereinigung für Wasserwirtschaft, Abwasser und Abfall e.V. (Hrsg.): Zustand der Kanalisa
tion in Deutschland – Ergebnisse der DWA-Umfrage 2009
[2] Statistisches Bundesamt: Fachserie 19 Umwelt, R. 2.1 Öffentliche Wasserversorgung und Abwasser
beseitigung 2004, Wiesbaden, 2006
[3] Jesse, F.; Curbach, M.: Verstärken mit Textilbeton. Beton-Kalender 2010, Berlin: Ernst & Sohn, 2009,
S. 457-565.
[4] Schladitz, F; Lorenz, E.; Curbach, M.: Biegetragfähigkeit von textilbetonverstärkten Stahlbetonplat
ten. Beton- und Stahlbetonbau 106 (2011), H. 6, S. 377-384 - doi: 10.1002/best.201100002
282
[5] Frenzel, M.; Ehlig, D.; Schladitz, F.; Curbach, M.: Textile Reinforced Concrete - Practical Applica
tions. In: Dancygier, A. N. (Hrsg.): Engineering a Concrete Future: Technology, Modeling \& Con-
struction - Proceedings of the fib Symposium 2013, Tel-Aviv, 22.-24. April 2013. Israeli Association
of Construction & Infracstructure Engineees (IACIE) & Faculty of Civil and Environmental Engineer-
ing, Technion - Israel Institute of Technology, 2013.
[6] Ehlig, D.; Schladitz, F.;Frenzel, M.; Curbach, M.: Textilbeton-Ausgeführte Projekte im Überblick.
Beton- und Stahlbetonbau 107 (2012), H. 11, S. 711-722 – DOI: 10.1002/best.201200034
[7] Scheerer, S.; Michler, H.: Freie Formen mit Textilbeton. Beton und Stahlbetonbau Spezial 2015 –
Verstärken mit Textilbeton (2015), S 94–100 – DOI: 10,1002/best.201400113
[8] Lorenz, E.; Schütz, E.; Schladitz, F.; Curbach, M.: Textilbeton – Grundlegende Untersuchungen im
Überblick. Beton und Stahlbeton 108 (2013), H. 10, S. 711-722 – DOI: 10.1002/best.201300041
[9] Lorenz, E.; Schütze, E.; Curbach, M.: Test methods for TRC. In: Brameshuber, W. (Hrsg.): Proceed
ing of Ferro 11 and 3rd ICTRC, Aachen, 08.-10.06.2015
[10] Lorenz, E.; Schütze, E.; Weiland, S.: Textilbeton – Eigenschaften des Verbundwerkstoffs. Beton und
Stahlbetonbau Spezial 2015 – Verstärken mit Textilbeton (2015), S 29–41 –
DOI: 10.1002/best.201400099.
[11] Lorenz, E.: Untersuchungen zur Endverankerung und Übergreifung textiler Bewehrungen in Verstär
kungsschichten aus Textilbeton. TU Dresden, Dissertation (in Vorbereitung).
283
284
COMBINED FORMWORK AND TEXTILE REINFORCEMENT

SYSTEM FOR A MINERAL CORROSION PROTECTION LAYER
FOR OFFSHORE APPLICATION
Prof. Dr.-Ing. Ludger Lohaus (1), Dipl.-Ing. Christoph Tomann (1), Dipl.-Ing. Hannes Weicken
(1), (1) Institute of Building Materials Science, Leibniz Universität Hannover, Appelstr. 9A, 30167
Hannover, (www.baustoffe.uni-hannover.de)
Abstract: Functional corrosion protection systems are absolutely essential for ensuring the
durability of offshore wind energy converters. As an alternative to conventional corrosion
protection systems, a thin layer of high performance mortar could be applied around the
turbine tower to protect it from the harsh maritime conditions. A flowable mortar with a
high density is required in order to prevent chloride penetration that could potentially dam-
age the steel structure. However, such mortars have a relatively high tendency to shrink,
which might lead to the formation of cracks in the mineral layer. These cracks increase the
possibility for the ingress of water and chlorides penetrating through the mortar, thereby
raising the risk of corrosion. To reduce these cracks respectively, the maximum crack
width and to ensure an adequate layer thickness, the high-performance mortar (HPM) has
to be applied by using textile reinforcement and suitable formwork. This paper presents
concepts for a possible realization of a textile reinforced mineral corrosion protection sys-
tem for offshore wind energy converters (OWEC). To this end, different textile reinforce-
ments were investigated with regard to their suitability for application and crack width
limiting properties by using a model steel tower.
INTRODUCTION
Maritime conditions provide the current corrosion protection systems for offshore wind
energy converter (OWEC) for a major challenge. By the combined effects of seawater,
alternating cycles of wet and dry and the mechanical abrasion, present corrosion protection
systems are, especially in the area of the “splash-zone” (cf. Fig.1), exposed to high stress-
es. The currently used organic corrosion protection systems are manufactured in a multi-
stage procedure, in a thin layer thickness with tight tolerances. In consequence of this
complex manufacturing process, a constant quality is very complicated to ensure. As a
result of the thin layer thickness, these protection systems are prone to transport damages
and to the assembly process. In addition, maintenance measures of these systems are very
complex and expensive. As an alternative to these conventional systems, a concept for a
mineral corrosion protection system was developed at the institute of building materials
science (IfB) at the Universität Hannover. In such a system, a high performance mortar
(HPM) is applied to the foundation structure of an OWEC, in the highly stressed “splash-
zone”. Next to the initial corrosion protection, such system suits, as an offshore mainte-
nance measure due to the hydraulicity of the cementitious binder. In addition to the various
concrete technology problems concerning the durability, which were tested in literature [1]
and verified for an undamaged case, cracks within the HPM layer must be analyzed due to
285
the part-specific requirements. Therefore, crack width limiting reinforcements are exam-
ined for their suitability by using a model steel tower is discussed in this paper.
Fig. 1 schematic representation of the corrosion rate [2]
MINERAL CORROSION PROTECTION SYSTEM
To realize such a mineral corrosion protection system, a suitable formwork and textile rein-
forcement has to be applied around the steel tower and filled with a HPM [3]. The passiv-
ating properties of cementitious building materials are well known from the classical rein-
forced concrete design. However, since the planned layer thickness of the mineral corro-
sion protection system (10 mm) is significantly lower than the required minimum concrete
cover in literature [4] for normal hydraulic concrete structures, the used mortar has to be
very dense and durable.
Due to shrinking processes, cracks can arise in the mineral corrosion protection layer be-
cause of the integrated steel tower (cf. Fig. 2). Since separation cracks can be expected in
this case, in the area of cracks an increasing ingress of chlorides can occur penetrating up
to the steel surface. As a result of self-healing effects of cementitious systems, where oc-
curring cracks can be sealed, by the reaction of unhydrated cement, a precipitate CaCO3
and the plugging of tiny components, cracks don’t necessarily imply a failure of the whole
corrosion protection. The results of literature [5], in which the water permeability and the
self-healing effects of separated cracks in concrete were
investigated, show an entire sealing of dynamically
stressed cracks after a duration of at least 15 weeks up to a
crack width of 0.2 mm. In order to guarantee a reliable and
durable corrosion protection, the maximum crack width,
has to be limited by using suitable textile reinforcement.
Fig. 2 cracks in the protec. layer
286
EXPERIMENTAL PROGRAMM
TEST FACILITY
In order to limit the crack width, different systems of different reinforcement materials
with different positions in the mortar layer are applied. To investigate such systems, a
model steel tower, with a diameter of about 90 cm and a circumference of about 275 cm is
used (cf. Fig. 3). In this study, five different reinforcement systems with a thickness of 10
mm and a height of 20 cm are applied around the steel tower and filled with HPM. For the
application a three dimensional spacer fabric (System S-I), a polymer-coated textile rein-
forcement (System S-II, S-III, S-IV) and a conventional reinforcement mesh, known from
plaster applications, were used. The three dimensional spacer fabric of system S-I repre-
sented the reference system and simulates an un-reinforced situation. Furthermore the tex-
tile reinforcements of the other systems are applied in different position in the mortar layer.
A schematic representation of the system structures the positioning of the textile rein-
forcement and their technical data is shown in Fig. 4. To guarantee a sufficient flow path,
which is necessary for self-compacting mortars, the specimens were filled from the bot-
tom-up, by using an inlet which is integrated into the formwork. In order to analyze the
cracking behavior and to monitor the filling process, a transparent formwork surface is
used. As an alternative to the closed formwork surface, a permeable stainless steel mesh
was applied in system S-V. The background of this exchange was, to create an additional
possibility for deaerating the mortar and to integrate the outer surface of the formwork as
an additional reinforcement layer.
Due to the reduced stiffness of the stainless steel mesh,
the formwork has bulged during the filling process. For
this reason, a defined layer thickness could not be
maintained. In addition, the other systems show an
adequate deaerating process of the mortar, therefore
this system (S-V) is not considered hereafter. The wa-
ter curing of the mortar was guaranteed by the closed
formwork for 72 hours. The formed cracks were visu-
ally detected with a crack magnifier and documented
directly after stripping of the formwork in regular in-
tervals up to a supposed end of the crack initiation.
Fig. 3 test facility
287
central outer
zone zone technical data of the
textile reinforcement (TR)
formwork surface
S-I
TR1 TR2
steel tower
S-II
carbon fiber
TR1
alkali-resistant
material reinforced
glass fiber
S-III plastic
TR2 TR1
S-IV Coating epoxy resin -
S-V mesh width

10 5
TR1
[mm]
thickness
mortar layer 10 [mm] 2 1
[mm]
Fig. 2 positioning and characteristic data of the textile reinforcement
RESULTS
INTEGRATION OF THE TEXTILE REINFORCEMENT
The surfaces of the systems show, that an entire filling process and therefore a complete
and spotless integration of the textile reinforcement is possible (S-I, S-II). Only those sys-
tems, in which the textile reinforcement is positioned in the outer surface zone (S-III, S-V),
show tiny isolated defect spots in the area of the reinforcement meshes (cf. Fig. 5).
System S-I System S-II System S-III System S-IV
Fig. 3 surface of the specimens
Due to the fact, that the surface of the formwork is directly positioned on the textile rein-
forcement, air voids were entrapped and for this reason the surface of the mortar layer is
not homogeneously closed. Consequently the entrapped air produces voids into the layer
surface, which partially reduces the cross-section of the mineral corrosion protection sys-
tem and leads to a reduced durability of the whole protection system. In conclusion, the
systems S-I and S-II could guarantee a complete integration of the textile reinforcement. In
contrast, the systems S-III and S-IV could not archive this entire integration, but with a
modification of the system structure, a significantly improved result is expected.
288
CRACK DISTRIBUTION
The crack distribution of the specimen is shown in Fig. 6. Presented are the number of
cracks (diagram A), the overall summed crack width (diagram B) and the maximum crack
width (diagram C) for each applied system. The x-axis of all diagrams is, for reasons of
clarity, scaled up using the root square of time. The test results of these investigations were
documented over a period of 36 days after stripping of the formwork.
System S-I, which represents the reference specimen, is the only system which showed
cracks directly after stripping of the formwork (cf. Fig.6). Consequently, the integrated
three dimensional spacer fabric was not able to carry the stresses in the early phase, which
probably resulted from the autogenous shrinkage process of the HPM. This could be due to
the low strength of the used spacer fabric. Furthermore, system S-I shows only a small
increase in the number of cracks over the observed period. Systems S-II to S-IV, however,
features up to 65 cracks in the first 6 days after stripping of the formwork. As a result of
the high number of cracks, the integrated textile reinforcement is activated in these sys-
tems, thereby limiting the crack width. Furthermore, diagram A shows that system S-III
with a positioning of the textile reinforcement in the outer zone, leads to a lower number of
the cracks in comparison to system S-II where the same textile reinforcement was posi-
tioned in the central zone. The descending graph in system S-IV reveals the coalescence of
different cracks. After 12 days, the number of cracks is constant in all systems which can
be explained by the end of the bulk shrinkage process.
Diagram B in Fig. 6 showed the overall summed maximum crack width of all cracks for
each system. System S-I features, as the only system, an overall summed crack width of
1.4 mm, directly after stripping of the formwork. The other systems, however, have their
first summed crack width value, one day after stripping of the formwork. Generally all
systems show a potentially similar trend. Corresponding to the number of cracks, the value
of the summed crack width increases within the first 12 days in all systems, up to a maxi-
mum value of 6 mm in system S-IV. After 12 days the summed crack width remains con-
stant in all systems over the documented time period, which means, that the leading crack
distribution is practically completed. In this context, system S-III, which has a polymer
coated textile reinforcement, positioned in the outer zone, leads, in regard to the overall
summed crack width, to the lowest value of all documented systems. System S-II and S-III
shown in diagram C reveal that the used polymer coated textile reinforcement is able to
limit the crack width, irrespective of its positioning in the mortar layer (central or outer
zone), up to a limit of 0.1 mm over the whole time period of 36 days. Consequently, the
maximum tolerable crack width required in literature [5] is maintained, which is necessary
for a positive self-healing effect. System S-I with a maximum crack width of 1.3 mm and
S-IV with a maximum crack width of 0.4 mm, 12 days after stripping of the formwork,
however, exceeded the tolerable crack width limit.
Generally, the results of crack distribution in these first tests demonstrate that the position-
ing of the textile reinforcement in the outer zone of the specimen is preferable. Since this
application, however, may leads to voids in the mortar surface (cf. results of the integration
of the textile reinforcement), a modification of the spacer bars of the textile reinforcement
should be conducted to create a homogeneous closed mortar surface and to further develop
the mineral corrosion protection system. In this case, a thin gap between the formwork
surface and the textile reinforcement could improve the result.
289
SUMMARY AND OUTLOOK
Since the maritime conditions, require particular measures for the corrosion protection,
especially in the “splash-zone”, a durable protection system is required. To ensure these
high durability properties with a mineral corrosion protection system, a closed and dense
mortar surface is necessary. Due to the structural and geometric boundary conditions,
cracks are not preventable. However, crack widths could be reduced through the use of
textile reinforcement and self-healing effects of the used cementitious materials. In order to
guarantee a durable corrosion protection, the textile reinforcement should be homogene-
ously integrated and the maximum crack width should be limited.
The investigations have shown, that the textile reinforcement can be completely integrated
into a thin mortar layer and as expected, the crack width can be significantly reduced. The
most significant effect was found in relation to the maximum crack width. In this context,
the experimental study demonstrated that a maximum crack width of 0.1 mm could be real-
ized. This crack width is according to literature [5], in the range of a self-healing effect
with a potential to a complete crack sealing. Since the study in literature [5], were carried
out with tap water, the scope of the self-healing effects, needs further evaluation under
offshore conditions.
In conclusion the first tests shown, that the mineral corrosion protection system might
work. Therefore, it offers a high potential for offshore applications either for an initial cor-
rosion protection than for maintenance purposes. However, for the realization of such a
system, the application techniques and the self-healing effects of the high performance
mortar should be further examined and optimized.
In this context, the Institut of Building Materials Science in Hannover is currently doing
some research works with regard to the influence of chloride penetration into cracks, self-
healing mechanisms and the effects of self-healing supportive additives, especially under
offshore conditions.
290
Fig. 4 crack distribution after stripping of the formwork note: x-axis is scaled up using the root square of time
291
REFERENCES
[1] Weicken, H. ; Lohaus, L. (2014): Measures for autogenous shrinkage compensation and their influ-
ence on selected durability properties, International RILEM Conference on Application of Super-
absorbent Polymers and Other New Admixtures in Concrete Construction, RILEM Proceedings
PRO95, pp. 245-256, Dresden, 2014
[2] S. Steppeler; G. Haake (2008): Forschung im ersten deutschen Offshore-Windpark, Stahlbau 77,
Heft 9, S. 681-684
[3] Lohaus, L. ; Huhn, H. ; Weicken, H. ; Zockoll, A. ; Peters, N. C. et al. (2011): Ganzheitliches Di-
mensionierungs-konzept für OWEA-Tragstrukturen anhand von Messungen im Offshore-Testfeld
alpha ventus, BMU GIGAWIND alpha ventus, Jahresbericht 2010, Hannover, 2011
[4] DIN 1045-1 (2008): “Concrete, reinforced and prestressed concrete structures - Part 1: Design and
construction“; German version
[5] Edvardsen, Carola Katharina: Wasserdurchlässigkeit und Selbstheilung von Trennrissen in Beton.
DAfStb 455, Rheinisch-Westfälische Technische Hochschule Aachen, 1996.
292
STRUCTURAL CONCRETE ELEMENTS REINFORCED WITH

TEXTILES – SUSTANABILITY PERSPECTIVE
Nahum Lior (1), Peled Alva (2) and Gal Erez (2), (1) Energy Engineering Department, Ben Gurion
University of The Negev, Beer Sheva, (2) Structural Engineering Department, Ben Gurion
University of The Negev, Beer Sheva
Abstract: In industrialized societies, buildings account for a large fraction of the overall
energy use and a similar proportion of greenhouse gases emitted. Quantifying the
embodied energy, that is the direct and indirect carbon emission in the production and
utilization process, is an important component in the effort to reduce the carbon footprint.
As well known, because of its many qualities, concrete is the most widespread building
material, particularly concrete reinforced with conventional steel. Yet it has historically
shown disadvantages in terms of durability and vulnerability to corrosion attack.
Furthermore, these materials are energy abundant. Textile Reinforced Concrete (TRC), a
combination of fine-grained concrete and textile fabrics, can offer a promising new
alternative for its corrosion resistance and high performance properties, resulting
lightweight and thinner structures, thus holds great potential in reducing the element
embodied energy. This work studied the potential energy saving of concrete structural
elements for massive building applications, by reinforcing them with layers of TRC, made
of carbon instead of steel bars. Structural design calculations were performed to find the
optimal structural configuration and validated with experimental data along with embedded
energy calculation. The vast potential of reinforcing elements with epoxy impregnated
carbon fabrics was clearly shown.
INTRODUCTION
In industrialized societies, buildings account for a large fraction of the overall energy use:
residential and commercial buildings are responsible for approximately 40% of the total
energy consumed [1], and for a similar proportion of greenhouse gases emitted [2] (and
this total is projected to increase by up to 56% between 2010 and 2040). These propor-
tions, however, only account for the energy consumed in buildings during the period of
their active usage. The share of energy used by buildings increases significantly when the
energy used in their production and demolition is included as well. Thus any comprehen-
sive assessment of the total energy consumed by buildings must in fact consider the entire
life cycle of the building, which can be divided into three phases: (i) Pre-use phase: Em-
bodied Energy (EE), the energy required for the production of a building and its compo-
nents which includes material extraction, production and transportation, and the building
construction itself; (ii) Use phase: Operational Energy (OE), the energy required for the
use and maintenance of a building during its useful life which is dominated by heating,
cooling and lighting, and is affected by the thermal properties of materials; and (iii) Post-
use phase: the energy needed after the building’s useful life, which includes demolition and
disposal, or possible recycling and reuse. Depending on the expected lifetime of the build-
293
ing and its operational energy efficiency, the pre-use energy typically represents between
10% and 60% of the total energy used during the life time of the building [3]. Therefore
nowadays, new technological advances making use of non-traditional types or amounts of
material and energy are required to meet the demand for a sustainable industry [4] that will
allow minimal impact on the environment, both at the time of production as well as
throughout the life cycle of the building.
An innovative technology which has sprung from this notion is textile-reinforced concrete
(TRC), which is a combination of fine-grained concrete and textile fabrics. Research find-
ings show that TRC significantly improves the tensile strength, ductility, and energy ab-
sorption capacity of the concrete element [5]. It has been found that TRC can be utilized to
build slender, lightweight, modular and freeform structures and eliminate the risk of corro-
sion. A number of structural elements made of TRC have been successfully produced, in-
cluding a slim TRC footbridge, large-sized TRC facade slabs [6] and sandwich panels [7],
showing the high potential for using textile fabrics as reinforcements for structural compo-
nents. Furthermore, several Life Cycle Assessment (LCA) approaches have been per-
formed on TRC materials showing their high potential in terms of environmentally sustain-
able buildings [8].
The goal of this research was to develop environmental friendly structural elements made
of concrete for massive building applications, with reduced embodied energy and envi-
ronmental impact by reinforcing them with layers of Textile fabric Reinforced Concrete
(TRC), made of carbon instead of steel bars. Structural design calculations were performed
to find the optimal structural configuration and validated with experimental data along with
embedded energy calculation.
The energy efficiency of concrete beam reinforced for TRC was examined under pure
bending behavior only, due to the complex integrated behavior under bending and shear
stresses of the TRC which requires greater understanding and a more complex model.
However, this does not reduce the possible contribution of fabrics as shear reinforcement
and the fact that it is the complementary part of the energy saving vast potential.
MATERIALS
The TRC specimens have been reinforced with a two-dimensional, tricot (warp pattern)
carbon fabric. Carbon was chosen as the reinforcement material due to its superior me-
chanical properties and long-term durability. The fabric consists of 1600 tex (g/1000 m) in
the warp and weft directions. Concrete with compression strength of 50 MPa as well as
fine grained concrete (mortar) mixtures were used to produce the beams (Tables 1 and 2,
respectively).
Table 1: Concrete mixture
CEM II Water Aggregate Aggregate Sand (≤1.18) w/c
[kg/m^3] [kg/m^3] (1.18-4.75) (4.75-19) [kg/m^3]
[kg/m^3] [kg/m^3]
382.2 225.5 503.8 435.6 754.7 0.59
294
Table 2 : Mortar mixture

CEM II Water Sand (0.6) Silica Fly ash Superplasticizer w/c w/b
[kg/m^3] [kg/m^3] [kg/m^3] fume [kg/m^3] [kg/m^3]
[kg/m^3]
600 270 1400 41 30 5.2 0.45 0.40
BEARING BEHAVIOUR UNDER TENSILE LOAD
In order to determine the bending behavior of an element, one must first comprehends the
load bearing behavior under uniaxial tension of the reinforcement material. For that, epoxy
(EP520) impregnated and non-impregnated carbon textiles were tested under direct tension
using a 100KN Instron machine with a displacement rate of 0.5 mm/min. The fabric length
was 200 mm with a mesh width of 40 mm containing 5 bundles along the loading direction
of the fabric. Figure 1 presents stress vs. strain curves of the carbon fabric with and without
polymer impregnation, normalized by the cross-sectional area of the textile reinforcement
which in this case was of a single bundle. The results show that there is indeed a significant
improvement in the modulus of elasticity and tensile strength of the epoxy coated fabric
compared to the reference without the polymer.
Figure 1: Stress vs strain curves of carbon fabrics without and with (EP520) epoxy rasin.
Bearing behavior under uniaxial tensile load of a fine grained cement-based composite
reinforcedwith the epoxy coated fabric was investigated on specimen with a length of 26
cm, 4 cm wide, and thickness of 1 cm (Figure 2). The specimen was with a reduced cross-
section in mid-length, made by two 15 mm long notches. The specimens were made with 1
layer of carbon fabric with five bundles in the warp direction, while four out of five longi-
tudinal (warp - 0°) bundles in the notch area were pre-cut so there is only one continuous
bundle bearing the load. Also here Instron type of machine with a displacement rate of 0.5
mm/min was used as for the fabric alone. The average stress vs. strain curve that has been
chosen to represent the composite reinforcement tensile behavior used in the mathematical
model of the TRC bending behavior is presented in Figure 3. The presnted values are
normalized by the cross-sectional area of the textile reinforcement.
295
Figure 2 : Geometry of a tensile test specimen
Figure 3 : Tensile stress vs. strain curve of fine grained concrete (mortar) reinforced with one layer of epoxy
impregnated carbon fabric, with mid-length reduced cross-section (notched specimen).
BEARING BEHAVIOR UNDER FLEXURAL LOAD
The flexural behavior of textile reinforced concrete differs from that of steel reinforcement,
and therefore according to Brameshuber (2006) [9] analytical models typically used for
steel reinforced concrete structures are invalid. Several analytical models for TRC have
been proposed over the last decade e.g. Brameshuber (2006) [9], Hegger et al. (2008) [6]
and Mobasher (2008) [10]. Conversely, based on Yin et al. (2013) [11], the common steel
reinforced concrete models were shown to agree with experimental results of thin-walled
TRC elements with a maximum error of 10%. In this research, the analytical model is fun-
damentally based on the common steel reinforced concrete model, with several adjust-
ments.
Flexural experimental program and results

Flexural bearing behavior tests were conducted in order to validate the compatibility of the
analytical model and the reinforcement's characteristics required for calculating the TRC
element's flexural load capacity. Four point bending tests were conducted on 1 m long
beams so that pure bending will be obtained in mid-span area, using 100KN Instron ma-
chine with a displacement rate of 0.5 mm/min. As shown in Figure 4, the beams were cast
continuously from two types of concrete; top 7 cm were made of conventional concrete
(Table 1) with compression strength of 50 MPa while the bottom 3 cm were made of fine
grained concrete (Table 2) with compression strength of 70 MPa. All beams were prepared
296
with epoxy impregnated carbon textile layers (of varied number of layers) located 4 mm
from the bottom of the beam, i.e., at the fine grained concrete region, without any steel
bars for bending reinforcement. The beams were made each with different number of car-
bon fabric layers: one layer (1L), two layers (2L) and 4 layers (4L) with steel rebar for
shear stresses (Figures 5a-6a). Another beam type was prepared with two layers of "U"
shaped impregnated epoxy carbon fabrics (2F) providing reinforcement for both shear and
bending without the need for shear steel rebar as presented in Figures 5b-6b. Table 3 pre-
sents all the prepared beam types.
Figure 4 : Side view and cross sectional view (A-A) of beam made of conventional concrete (top 7 cm) and
fine-grained concrete (bottom 3 cm) reinforced with epoxy impregnated textile layers, along with the bending
set up
(a) (b)
Figure 5 : (a) shear steel rebar and (b) "U" shaped carbon textile for shear and bending reinforcement.
Figure 6 : Schemes of the side and cross-sectional view of the textile reinforced concrete beam with (a) the
shear steel rebar (B-B is the cross-sectional view with the steel rebar), (b) "U" shaped fabrics (D-D is the
cress-sectional view with the "U" shape of the textile).
297
Table 3 : Beams types.

Name No. Bundles Reinforcement Shear reinforcement
of layers per layer cross-sectional area
(mm^2)
1L 1 11 4.97 steel rebar
2F 2 7 6.33 carbon “U” shape
During the bending tests, load and mid-span deflection were measured. Concurrently, pic-
tures were taken in 15 seconds intervals in order to study the crack formation and type of
failure. Four specimens were examined for each beam type. Figure 7 presents typical load
versus deflection curves of each tested system. The beam reinforced with four layers of
carbon fabric exhibited the highest load bearing capacity, following with the “U” shape
fabric beam (2F) that also shows the greatest deflection up to failure. However, all the
beams with the straight layers of carbon fabric and the shear steel rebar exhibited intensive
delamination, yet not such a delamination was observed for the beam reinforced with the
“U” shape fabric. Lack of continuity between the shear rebar and the textile reinforcement
prevents full stress transmission which resulted the severe delamination between the
beam's bulk to the textile reinforcement and consequently an early failure; while the spec-
imens with the two layers of ”U” fabric shape (without steel rebar) failed only when the
textile reinforcement failed. Furthermore, if taking into account the reinforcement ratios,
the load capacity of the specimens with the steel rebar is much lower than that with the two
layers of the “U” shape carbon fabrics. This clearly indicates the potential benefit of using
textile reinforcement as steel bar replacement under bending condition; however this
should be further explored for better utilizing this technology.
Figure 7: Typical load-deflection of the beams bending behavior.
ANALYTICAL MODEL
An analytical model was performed to calculate the maximum bending load capacity of
conventional steel reinforced concrete (SRC) under four point bending, of two different
span lengths, 1 m and 4 m. Based on this calculation the required amount of bundles (B)
and textile layers (L) of beams reinforced with TRC was calculated, aiming to reach the
same load capacity of the conventional concrete with the two different span lengths. This
298
was for Life Cycle Assessment (LCA) calculation. TRC with full rectangular cross-section
(TRC – F) similar to that of the SRC beam, and a beam in configuration of a box shape
reinforced with TRC (TRC – B), aiming to reduce the beam cross-sectional area were ex-
plored. Table 4 presents the geometry of the beams and the required calculated
reinforcements. For the TRC beam calculation, the textile strain at failure was determined
as the maximum strain of the textile reinforcement. An elastic behavior of the cross-section
was assumed, the stress of each layer of the textile reinforcement was driven from the
curve in Figure 3 as a function of the strain. Because of the relatively high strain of the
reinforcement, there is a much smaller compressed area in the TRC concrete beam then in
steel reinforced concrete, therefore the force of the compressed area wasn’t made by the
rectangular proximal calculation but by direct use of the known stress-strain parabolic be-
havior of the concrete. The tensile force of a single reinforcement layer and the ulti-
mate moment capacity of the reinforced concrete element were calculated by equations
(1) and (2) respectively as follows:
(1)
Where: = number of bundles in a layer, = cross-sectional area of a single textile rein-
forcement bundle, = tensile stress of the bundle as a function of the strain.
(2)
Where: = compression force of the compressed area, = internal lever arm between
tension center of mass and compression center of mass, = number of textile reinforce-
ment layers, = layer index, = tensile Strength of the textile reinforcement layer i as a
function of the strain, = lever arm to layer i.
For model validation, analytical calculation of the reinforced concrete beam moment ca-
pacity was made for an element with the exact same properties as in the experiment. From
those results the maximum load capacity of 17.9 KN was calculated and compared to ex-
perimental value of 16.9 KN obtained for the tested “U” shape fabric concrete beam (2F).
The calculated value is within the range of variability of the experimental results of 1.2 KN
(7.1%), i.e., clearly validated the analytical model. These results indicate an excellent cor-
relation between the calculated value by the analytical model and the experimental results.
Table 4: Examined beam elements geometry and its reinforcement.
SRC TRC - TRC -
F B
Beam length [cm] 100 400 100 400 100 400

Beam height [cm] 10 40 10 40 10 40
Beam width [cm] 10 10 10 10 10 10
Cross-sectional area
100 400 100 400 75 275
[cm^2]
Steel reinforcement 2Φ 8 3Φ 16 ---- ---- ---- ----
Carbon reinforcement ---- ---- 9Bx3L 12Bx11L 9Bx3L 12Bx11L
* B stand for bundle, L stand for layers.
299
LIFE CYCLE ASSESMENT (LCA)
The embodied energy was calculated by Cradle to gate perspective approach for conven-
tional steel reinforced concrete beam (SRC) and the two TRC reinforced concrete beams,
TRC – F, and TRC – B discussed above (Table 4). The energy consumption values of the
different materials used to prepare the beams were based on Pearlmutter et al [12]. Figures
8-9 present the calculated embodied energy and the relative energy savings (relative to
SRC) of 1 m and 4 m beams respectively for each beam type of the different cross sections
with and without taking into account durability (service life) aspects. In the cases where
durability aspects were taken into consideration the total life cycle of the element was as-
sumed to be 50 years for the conventional steel reinforced beam and 150 years for the TRC
reinforced beams duo to the TRC corrosion resistance.
(a) (b)
SRC TRC - F TRC - B SRC TRC - F TRC - B

Textile 5.36 5.36 5.36 5.36
Steel 13.73 41.19
Concrete 20.06 22.57 16.93 60.17 22.57 16.93
Energy Saving 0% 17% 34% 0% 72% 78%
Figure 8 : Embodied energy for 1 m long beam: (a) regardless to durability, (b) with reference to
durability.
(a) (b)
SRC TRC - F TRC - B SRC TRC - F TRC - B

Textile 104.76 104.76 104.76 104.76
Steel 329.56 988.67
Concrete 320.89 337.68 239.41 962.66 337.68 239.41
Energy Saving 0% 32% 47% 0% 77% 82%
Figure 9: Embodied energy for 4m long beam: (a) regardless to durability, (b) with reference to durability.
Rather the durability aspect is taking into account or not, it can be seen that there is a con-
siderable amount of energy saving which is becoming more significant as the beam length
grows. As high as 47% saves in the embodied energy was recorded for the TRC beams of 4
m span when durability aspects were not considered for box shape beam. If considering
300
life cycle of 150 years for the TRC this value was even more significant with embodied
energy saving of 82%. These results clearly show the vast potential in energy saving when
structural elements are reinforced with epoxy impregnated carbon fabrics.
CONCLUSIONS
This work studied the potential energy saving of concrete structural elements for massive
building applications, by reinforcing them with layers of Textile Reinforced Concrete
(TRC), made of carbon fabric instead of steel bars. Structural design calculations were
performed to find the optimal structural configuration and validated with experimental data
along with embodied energy calculation. The effect of the TRC Embodied Energy (EE)
consumption was measured under bending behavior only. The vast potential of reinforcing
elements with epoxy impregnated carbon fabrics was clearly shown. As high as 47% or
even 82% of embodied energy saving was recorded for beams reinforced with TRC of
open box shape, without and with taking into account durability aspects. However even for
full rectangular cross section beam a significant saving in the embodied energy was calcu-
lated. Further research takes into account both bending and shear reinforcement is requires
in order to fully comprehend the total influence of TRC on the element's EE. .
REFERENCES
[1] DOE/EIA, 'Annual energy review 2005', in, 'Energy Information Administration', (2006).
[2] Laurenzi, M. P., editor, 'Building Energy Efficiency', An Asia Business Council Book, (2007).
[3] Huberman, N. Pearlmutter, D., 'A life-cycle energy analysis of building materials in the Negev desert',
in 'Energy and Buildings', (2008), 40(5), 837-848.
[4] Vanegas, J., DuBose, J. & Pearce, A., 'Sustainable Technologies for the Building Construction Indus-
try', in 'Proceedings, Symposium on Design for the Global Environment', Atlanta, GA, (1996).
[5] Brameshuber, W., editor, 'International RILEM Conference on Material Science – MatSci', Vol. I, (IC-
TRC, Aachen, Germany, 2010).
[6] Hegger J. and Voss S., 'Investigations on the bearing behaviour and application potential of textile rein-
forced concrete'. Engineering Structures, Vol. 30 (7), (2008) 2050-2056.
[7] Horstmann, M., Hegger, J. 'Sandwich façades made of Textile Reinforced Concrete – Experimental
investigations' (in German; 'Sandwichfassaden aus Textilbeton - experimentelle Untersuchungen'),
Bautechnik, 88(5), (2011), 281-291.
[8] Tomoscheit, S., Gries, T., Horstmann, M., and Hegger, J., 'Project Life INSUSHELL: Reducing the
Carbon Footprint in Concrete Construction', International Journal of Sustainable Building Technology
and Urban Development, 2 (2), (2011), 162-169.
[9] Brameshuber W., 'Textile Reinforced Concrete-State-of-the-Art Report of RILEM TC 201-TRC.36',
(RILEM publications, 2006).
[10] Soranakom C, Mobasher B, ' Correlation of tensile and flexural responses of strain softening and strain
hardening cement composites', Cement & Concrete Composites 30 (2008), 465–477.
[11] Xu S., Yin S. and Cai X., 'Investigation on the flexural behavior of reinforced concrete beam strength-
ened with textile-reinforced concrete', Tumu Gongcheng Xuebao/China Civil Engineering Journal, Vol.
44 (4), (2011), 23-34.
[12] Pearlmutter D., Meir I. and Huberman N., 'The Embodied Energy of Building Materials in Israel', Re-
port RD-40-2012, for the ministry of energy and water resources, State of Israel, (2013).
301
302
CHAPTER 7
MEASURING
METHODS
303
304
TEST METHODS FOR TEXTILE REINFORCED CONCRETE
Elisabeth Schütze, Enrico Lorenz and Manfred Curbach, Institute of Concrete Structures,
Technische Universität Dresden, Germany
Abstract: The standardization of test methods for determining the material properties of
textile reinforced concrete (TRC) becomes increasingly important, as the composite mate-
rial is gaining significance. While conducting fundamental research on the material as well
as realizing several practice projects, substantial characteristic values and the correspond-
ing experimental tests became evident. This publication gives an overview of this topic,
detailing the tests and values that were used in order to obtain several approvals in individ-
ual cases as well as a general technical approval for the flexural strengthening of steel rein-
forced concrete slabs with a TRC system. The authors describe test setups suitable for the
determination of the bond and strength parameters of TRC and suggest methods for the
evaluation of the experimental results, while demonstrating common mistakes in testing as
well as ways to prevent them. Typical results from the determination of the characteristics
are given for an exemplary TRC system.
INTRODUCTION
In reinforced concrete construction, alternative reinforcement materials have been searched

for some time since steel is susceptible to corrosion. Apart from materials such as alloy
steel or natural fibers, over the last 20 years industrially manufactured fiber materials with
significantly higher load bearing capacities, such as alkali-resistant glass and carbon, were
investigated increasingly. These high performance fibers are already successfully used in
the construction of light and highly stable structures in the aircraft and automobile indus-
try. Now the advantages of these fiber materials can also be used in civil engineering, if
technical textiles are applied as noncorrosive concrete reinforcement, thus creating the
composite material textile reinforced concrete (TRC), e. g. [1]. By combining the material
properties of concrete with those of high performance reinforcements, extremely thin TRC
elements and strengthening layers with superior load-bearing capacity can be realized. In
two DFG collaborative research centers at TU Dresden (SFB 528) and at RWTH Aachen
(SFB 532) intensive research on this topic has been pursued beginning in 1999. Ever since
2005, textile reinforced concrete has been used increasingly in construction practice, [5].
Consequently, approvals for individual cases as well as general technical approvals are
constantly gaining in importance. In the meantime various other research centers in Ger-
many and abroad have started to study textile reinforced concrete. Scientists who research
textile reinforced concrete for the first time as well as numerous regulatory authorities face
the challenge to identify the substantial characteristic values as well as to determine and
verify them. While a number of publications already show answers to individual problems
(e. g. [3]), a comprehensive survey is still hard to find.
305
Faced with the growing demand for standardization, this publication shall give an over-
view of the main characteristics of TRC and show how to determine them. Numerous ap-
provals for individual cases (ZiE) as well as the first general technical approval for the
flexural strengthening of steel reinforced concrete slabs with a TRC system [4] have al-
ready been based on the authors’ suggestions presented in this paper. As there will not be a
uniform standard in the near future, this article shall be a guide for scientists and users.
ESSENTIAL CHARACTERISTICS
Textile reinforced concrete consists of concrete and textile reinforcement. As is common in

construction, the load bearing capacity, serviceability and durability of the individual com-
ponents and their bond have to be verified before the composite can be used. Current rule-
books are used to test and classify the concrete. However, such normative rules do not ex-
ist for the textile reinforcement and the composite TRC yet. Consequently, the determina-
tion of material properties is based on experience from basic research, practice projects,
numerous approvals in individual cases and the general technical approval [4].
The textile reinforcement in TRC is used to take up tension forces exclusively. The tension
test described in section 4 allows for a reliable determination of the tensile strength as well
as a simple characterization of the tensile load bearing behavior of the composite. For a
safe application of textile reinforced concrete, in addition to the tensile strength of the ma-
terial, the transmission of forces between textile reinforcement and concrete matrix is cru-
cial. The bond test presented in section five serves to determine the magnitude and pro-
gression of the relevant forces. Based on this, the bond stress-slip relationship can be cal-
culated. This specific bond characteristic allows for the evaluation of the bond behavior of
the textile reinforced concrete system. The tests required to determine the load bearing
capacity of the end anchorage and overlap areas are described in section six.
The number of necessary test samples depends on the respective purpose. Accordingly, at
least three samples for tensile tests and six samples for bond tests are recommended for the
determination of the general load bearing behavior and for parameter studies. In order to
find design values and for statistic analyses it is, however, advisable to examine at least 10
to 20 test samples.
The characteristics determined with the test methods presented here make it possible to
establish characteristic and design values subsequently. A design concept for flexural
strengthening and the required partial safety factors are shown in [5].
SAMPLE PRODUCTION
The TRC samples tested in the standard test setups are produced in hand lamination pro-
cess in smooth, non-absorbent steel formworks as large-format textile reinforced concrete
slabs. The required number of textile and fine grained concrete layers are placed in the
formwork alternately with a concrete layer on bottom and top. The reinforcement layers
are arranged symmetrical to the sample thickness, level and parallel to the slab surface.
After the last concrete layer has been applied, the surface is leveled. According to [1], the
required concrete cover is defined mainly by the yarns’ diameter, the maximum aggregate
size and the manufacturing precision. To avoid spalling and, thus, a reduced load bearing
306
capacity of individual yarns, it is recommended to produce a concrete cover of at least

1.5 mm. The suggested layer thickness is 3 mm. However, this value can be varied depend-
ing on the textile reinforcement. The samples are stored according to DIN 18555-3, Table
1. For the first three days, the slabs remain in the formwork at about 20°C, protected
against drying out. Then, they are stripped and stored in water until the seventh day. After
eight days, the samples are cut from the textile reinforced concrete slab with a water-
cooled diamond saw. Especially for samples with a small number of yarns it is important to
set the cuts so the yarns are distributed symmetrically across the sample’s width and no
yarns are damaged. Until the time of the test, the samples are stored in a climate chamber
at 20°C and 65 % relative humidity, standing and exposed to air.
TENSION TESTS
As textile reinforcements in TRC primarily serve as tensile reinforcement, the analysis of

the tensile load bearing behavior is essential for the characterization and safe application of
this composite material. Most of the design models developed in the past years are based
on this. To be able to predict the failure load of the textile reinforced concrete layers, the
uniaxial tensile strength of the textile reinforcement and the respective strain at rupture
have to be known. According to [1] and [6], the effective tensile strength of the yarns in
the textile reinforced concrete is influenced by a number of geometric parameters and the
manufacturing process. As a result, it can usually not be directly deduced from tests on
individual filaments or on individual unprocessed yarns. Consequently, the uniaxial tensile
tests described in [6] and [7] are carried out to characterize the tensile load bearing behav-
ior of the textile in concrete. That way, the uniaxial tensile strength of the textile rein-
forcement in the composite as well as the corresponding stress-strain curve can be deter-
mined.
Test Specimens
Normally, specimens with the dimensions 700 mm × 60 mm × 9 mm and two layers of
textile reinforcement are used in the uniaxial tensile tests, see figure 1.
They are cut from the textile reinforced concrete slabs described in section 3. However,
depending on the specific application, samples with a deviating number of layers can be
examined.
In the course of the advancement of textile reinforcements, so-called heavy-tow yarns –

reinforcement textiles made of carbon with large yarn cross sections – are used increasing-
ly. Particularly for yarns with compact cross sectional shape, the increase of the cross-
sectional area leads to a disproportionally high increase of the yarn tensile force compared
to the composite surface. That is why very large end anchorage lengths can be required. To
guarantee tensile fracture – the objective of the test –, premature failure of the sample
caused by pull-out of the textile reinforcement from the fine grained concrete matrix dur-
ing the tension test has to be avoided, hence an adjusted sample structure may be neces-
sary. For instance, an additional bond improving coating made of epoxy resin and sand can
be applied to the coated textile fabric in the anchorage areas of the specimens before con-
creting [9]. Alternatively, the load transfer areas can be molded in epoxy resin instead of
concrete.
307
To maximize the forces that can be anchored, the sample length and thus the length of the
load introduction areas can be increased. Additionally, arranging extra layers of reinforce-
ment in the end anchorage areas allows for a better distribution of the load [9]. Figure 2
shows the structure of such a sample, which was developed based on [6] und [7]. The addi-
tional reinforcement can also be applied in combination with the additional coating of
epoxy resin and sand described above.
Figure 1 : Test specimen for Figure 2 : Test specimen with a geometry adapted for
uniaxial tensile tests, acc. to [8] heavy tows, acc. to [8]
Test Setup
As shown in figure 3, the samples are evenly clamped between steel plates in the upper and
lower load transfer area on a length of 175 mm respectively. These clamping constructions
are flexibly connected to the testing machine. Thus the load introduction is free of shear
forces and moments. To avoid local stress peaks, caused e.g. by unevenness of the concrete
surface, a leveling layer made of 0.5 mm thick rubber is arranged between sample surface
and steel plates. In case individual specimens show greater unevenness, it may become
necessary to apply an additional leveling layer of a suitable mineral mortar. The contact
pressure has to be set at such a level as to prevent the samples slipping out of the load in-
troduction. At the same time, the compressive strength of the sample must not be exceeded
in the load transfer area between the steel plates. If the applied loads are very high the lev-
eling layer of rubber can be replaced by a layer of abrasive cloth to further improve grip
and avoid slipping of the specimen.
In case samples with additional reinforcement in the end anchorage area (Figure 2) are
tested, longitudinal splitting cracks may occur in the area of the jump in height, particularly
if an additional bond improving coating has been applied to the textile in the end anchor-
age area. To prevent this, it is necessary to arrange a small additional clamping consisting
of steel sheets (100 mm x 30 mm x 5 mm) screwed together on both sides of the specimen.
308
Figure 3 : Test setup for uniaxial tensile tests and typical stress-strain relationship of TRC for rupture and an-
choring failure, acc. to [8]
Experimental Procedure and Measurements

After fixing the clamping, the sample is mounted in a servo-hydraulic testing machine with
accuracy class 1 according to DIN EN 12390-4 and a load capacity of 100 kN (Figure 3).
A preload of 0.2 kN is then applied so that the sample straighten. The subsequent loading
is applied displacement-controlled at a loading rate of 1 mm/min until the sample fails due
to tensile fracture. The registered measurements are force F, machine path s and elongation
Δl in the center of the specimen measured with extensometers with a gauge length of
l = 200 mm on both sides of the sample. At this loading rate, a measuring rate of 5 Hz is
recommended.
Results of the Tension Tests

Direct results of the tests are the ultimate force Fu and the corresponding elongation Δlu in
the area of the extensometers. The strain ε is calculated by relating elongation Δl to meas-
uring length l as shown in equation (1):
Δl
ε= (1)
l
The measured force can be related to the textile cross section At as well as to the cross
sectional area of the concrete Ac. The concrete gross section can be determined from the
width b and the height h of the tensile specimen with equation (2).
Ac = b ⋅ h (2)
Subsequently, the stress σc in the concrete cross section can be calculated with equation
(3).
F
σc = (3)
Ac
309
These values can be used to find the stress level at the first crack σcr as well as for a general
comparison of the tensile stresses that can be carried by the textile reinforced concrete lay-
er.
However, the tensile strength of TRC depends on the number and cross section of the
yarns. Consequently, the characteristic values relevant for the design are determined based
on the stress σt which relates to the respective textile cross section At. According to equa-
tion (4), the yarns’ cross sectional area At of the textile results from the number n of yarns
in the TRC cross section as well as the manufacturer’s information concerning the textile
fineness f and the density ρ of the yarn material.
f
At = n ⋅ (4)
ρ
According to equation (5), the stress σt in the yarn cross section is the quotient from the
force F and the yarn cross sectional area At of the sample in test direction.
F
σt = (5)
At
Based on this, the ultimate strength regarding the yarn cross section can be calculated
through equation (6)
Fu
σtu = (6)
At
Furthermore, the corresponding stress-strain curve can be derived. According to [6] and [7]
the stress-strain curve typically consists of three distinct sections, which are all approxi-
mately linear (figure 3).
In the first section (state I), the textile reinforced concrete sample – similar to steel rein-
forced concrete – is still non-cracked. The course of the curve is accordingly steep. The
textile reinforcement does not yet carry a considerable part of the load. When the first
cracks occur, the sample passes into state IIa (multiple crack formation). Due to the low
stiffness in this state, the increase of the curve is usually very low here. With the increase
of load and corresponding strain, further cracks develop while the crack distance decreases,
similar to steel reinforced concrete. As a result, an almost horizontal plateau develops in
the stress-strain curve. The course of the curve steepens again in the following state IIb
(final crack state), fairly close to indicating the stiffness of the textile reinforcement. When
the tensile strength of the textile reinforcement is met, a brittle tensile failure occurs. Ten-
sile fracture can be recognized easily by a characteristic tearing noise in connection with a
rapid decrease of force towards zero. Further elaboration regarding the tensile load bearing
behavior can be found in [6].
In contrast to this, premature and unplanned failure may also occur when the textile is not
sufficiently anchored in the end anchorage areas of the specimen. In this case, the textile
reinforcement may be pulled out from the anchorage areas on each end of the specimen. As
figure 3 shows, this leads to a gradual decrease of force before the ultimate force Fu has
310
been reached. In contrast to tensile failure, the textile is pulled out from the sample slowly
and without a tearing noise.
BOND TEST
A safe load introduction and transfer is essential for a fully functional TRC-system that
uses the load bearing capacity of the textile reinforcement to its full extend. The basis for
the evaluation of the bond behavior is the determination of the effective bond strength.
That is why an asymmetrical pull-out test was developed in [10], that serves as a standard
test for the bond behavior of coated textile reinforcement structures in TRC. In contrast to
bond tests on single filaments or single yarns, the bond properties of textile processed
structures in fine grained concrete are directly verified in this test.
Test Specimen
The specimens presented in figure 5 have the dimensions 270 mm × 60 mm × 9 mm. They
are reinforced with one textile layer. The textile reinforcement inside is arranged symmet-
rical and parallel to the edges. During the textile pull-out test, only one individual yarn of
the structure is tested. It is anchored sufficiently in the lower long part of the sample on a
length of 180 mm. In the upper short section of the sample, yarn pull-out is forced by chos-
ing a short but well-defined anchorage length lE,o.
Figure 4: Test setup for textile pull-out tests, acc. to [10]
This anchorage length lE,o is defined through a saw cut right below the clamping device in
which the tested yarn is cut. Toward the bottom, the upper anchorage area is limited by a
predetermined breaking point, which is defined through saw cuts on both sides of the spec-
imen. The size of the upper anchorage area is set to the same length as the simple distance
of the weft yarns. The cross-yarn should be located in the center of the upper anchorage
area. This way, the bond behavior of the yarn can be analysed in sections taking into ac-
count the influences of the textile processing. The length lE,o should not be smaller than
14 mm in order to ensure safe handling of the specimen. If the distance of weft yarns is
smaller than that, lE,o should be chosen as a multiple of this distance.
311
Test setup
As figure 4 shows, the load is applied through a clamping device on the upper and lower
end of the specimen. The clamping construction is flexibly connected to the testing ma-
chine at the upper end of the specimen. At the lower end, the specimen is clamped directly
into the tensile testing machine. The contact pressure is set in such a way that the specimen
doesn’t slip out of the clamping and the sample’s compressive strength is not exceeded.
The test area remains unaffected by the resulting lateral pressure, as the sample is clamped
above the anchorage area. Thus an influence on the bond can be avoided. To avoid damag-
ing the bond between yarn and fine-grained concrete during the installation of the upper
load introduction and mounting the specimen in the testing machine, the saw cuts at the
predetermined breaking point are stabilized with PU foam until the test.
Testing Method and Measurement

After the upper clamping has been installed, the specimens are mounted into a servo-
hydraulic tensile testing machine of accuracy class 1 according to DIN EN 12390-4 and a
suitable (low) load capacity. The force transducer is at the top of the testing machine. The
pull-out tests are carried out displacement-controlled at a loading rate of 1 mm/min until a
crack opening of 3 mm is reached. Subsequently, the tested yarn is pulled from the upper
anchorage area entirely at a rate of 10 mm/min to visually check the yarn’s surface and
determine the anchorage length lE,o.
The registered measurements are force F, machine path s and the crack opening w in the
area of the predetermined breaking point, measured with extensometers on both sides of
the sample (figure 4). The suggested measuring rate is 5 Hz. The dead weight of the upper
clamping and sample part has to be taken into account when determining the actual pull-
out force FG. Hence, the measured force F has to be corrected according to equation (7).
FG (w ) = F(w ) − Fkorr (7)
The corrective force Fkorr is defined as the remaining force due to the dead weight of the
test setup’s upper section measured after the complete pull-out of the yarn.
Results of the Textile Pull-Out Tests

The result of a textile pull-out test is a force-crack opening-curve as shown in figure 5. The
force FG correlates to the pull-out-resistance of the examined yarn. The corresponding
crack opening w equates to the sum of the relative displacements in the crack from the up-
per and the lower anchorage area.
The remaining concrete cross section in the area of the predetermined breaking point
cracks while the pull-out force FG is still at a low level, before the examined yarn is fully
activated. This effect does not have any influence on the examined bond behavior. That is
why the ascending branch of the FG-w-relation can be connected linearly from the point
were this crack forms at the force Fcr to the origin of the coordinate system, as shown in
figure 5.
312
Figure 5 : Results of the of the textile pull-out test
In order to evaluate and compare the bond behavior of various textile reinforced concrete
systems independently of the respective bond length, the results of the textile pull-out tests
are described as bond flow-crack opening curves (figure 5). The bond flow TG in N/mm is
calculated by relating the pull-out force FG to the anchorage length lE,o of the filament yarn
(equation (8)).
FG (w)
TG (w) = (8)
lE,o
The test results allow an evaluation of the bond forces of different textile reinforcements in
TRC. The experimentally determined bond flow-crack opening-curves typically consist of
three distinct sections that are characteristic for textile reinforcements in a fine grained
concrete matrix (figure 5): an ascending (a), a descending (b) and a subsequent slightly
ascending or descending branch (c). The sections show the activation of the adhesive bond
(a), the destruction of the adhesive bond (b) and the friction bond (c).
If, however, the results of the textile pull-out tests are to be used for further investigations
such as the determination of end anchorage- und lap lengths of textile reinforcements in
textile reinforced concrete, it is necessary to determine the bond stress-slip-relation as the
textile specific bond characteristic. [10] describes a way to determine the bond stress-slip-
relation and [8] gives a simplified approach to the problem.
END ANCHORAGE AND OVERLAP TESTS
According to analyses by [11], various failure mechanisms can apply for the end anchorage
of a TRC strengthening of structural components. Thus, the old concrete or the bonding
joint between old and fine grained concrete may fail. Apart from this, longitudinal matrix
splitting (delamination) within the textile layer or yarn pull-out before reaching the ulti-
mate load of the textile reinforcement may occur. In [11] comprehensive research regard-
ing the three plane failure modes by delamination, old concrete failure and the destruction
of the bonding joint between fine grained and old concrete has been conducted. The end
anchorage test which was developed based on [11] is shown in figure 6. With this test set-
up the end anchorage length required for the end anchorage of a specific textile tensile
force can be determined, taking a possible delamination failure of the TRC system into
313
account. Furthermore, the test can prove that the strength and the pre-treatment of the con-
crete as well as the roughness of the concrete surface that is to be strengthened can guaran-
tee an adequate bond between old concrete and TRC strengthening. Based on a design ap-
proach described in [11], the end anchorage lengths of TRC-strengthening layers for spe-
cific reinforcement textiles can be calculated in such a way that delamination and old con-
crete failure can be avoided. Premature failure by yarn pull-out can be avoided if the inter-
dependences presented in [10] are taken into account.
Figure 6 : TRC end anchorage test acc. to [11] Figure 7 : Overlap tests with photogrammetric measurement,
acc. to [10]
When applying textile reinforced concrete, lap joints of the textile reinforcement layers
usually cannot be avoided. That is why research concerning the load-carrying behaviour
and the dimensioning of lap joints was conducted in [10]. A uniaxial tension test with a
symmetrical lap joint at the sample’s centre, which was developed to test the joint areas, is
presented in figure 7. This overlap test takes into account the limit state of a delamination
failure or longitudinal matrix splitting as well as premature failure by yarn pull-out before
the ultimate strength of the textile reinforcement has been reached.
[10] proves that the required lap lengths in the limit state of yarn pull-out correlates to 1.0
times the value of the end anchorage lengths. Consequently, the lap tests can also be used
as a simplified method to experimentally investigate the pull-out failure in end anchorage
areas. Based on the bond characteristics of textile reinforced concrete determined accord-
ing to section 5, the lap joint areas can be dimensioned and the failure loads predicted
through the calculation approach presented in [10].
DURABILITY, TEMPERATURE, LONG-TERM AND FATIGUE TESTS
The tests described in this publication are usually carried out to investigate the material
behavior of TRC under predominantly static and short time loads at indoor climate, mean-
ing approximately 20°C and 65 % relative humidity. However, when TRC is applied in
practice, it is exposed to sustained and dynamic loads as well as to varying climatic condi-
tions. As a result, the tests described above have to be supplemented by temperature tests,
durability tests and tests regarding the long-term and fatigue strength of the composite.
Typically, these tests are conducted based on the specimens and test setups described in
section 5 and 6 in order to provide comparable test results and material characteristics.
Further details can be found in [5], [12], [13] and [14].
314
SUMMARY AND OUTLOOK
The increasing dissemination of textile reinforced concrete entails the increased im-
portance of a standardized way to determine material characteristics. This publication gives
an overview of test and evaluation methods that are used to determine and assess the deci-
sive strength and bond parameters. Additionally, experimental methods to test the load
bearing capacity of end anchorage and lap joint areas are presented – further parameters
that are relevant for design. The tests described above have been the basis for several ap-
provals for individual cases and a general technical approval. Experiences from basic re-
search, numerous practice projects and procedures regarding the granting of approvals will
be part of the process of standardizing textile reinforced concrete in the years to come.
ACKNOWLEDGEMENT
The authors would like to thank the Deutsche Forschungsgemeinschaft (DFG) for funding
the research regarding textile reinforced concrete as well as the TUDALIT e. V. group for
supporting the investigations necessary in order to obtain the General type approval for the
flexural strengthening of steel reinforced concrete slabs with a TRC system. Further thanks
go to the staff of the Otto-Mohr-Laboratory at TU Dresden for their numerous contribu-
tions to developing the test methods described in this paper.
REFERENCES
[1] Jesse, F.; Curbach, M.: Verstärken mit Textilbeton. Beton-Kalender 2010, Berlin: Ernst & Sohn, 2009,
S. 457-565.
[2] Frenzel, M.; Ehlig, D.; Schladitz, F.; Curbach, M.: Textile Reinforced Concrete - Practical Applica-
tions. In: Dancygier, A. N. (Hrsg.): Engineering a Concrete Future: Technology, Modeling & Con-
struction - Proceedings of the fib Symposium 2013, Tel-Aviv, 22.-24. April 2013. Israeli Association
of Construction & Infracstructure Engineees (IACIE) & Faculty of Civil and Environmental Engineer-
ing, Technion - Israel Institute of Technology, 2013.
[3] Frenzel, M.; Ortlepp, R.; Scheerer, S.; Curbach, M.:Textile Reinforced Concrete. A novel, light and
shapeable material. Hormigón y Acero, 2013, H. 64, S. 47-74.
[4] Deutsches Institut für Bautechnik (DIBt), ‘Verfahren zur Verstärkung von Stahlbeton mit TUDALIT
(Textilbewehrter Beton), Zulassungsnummer Z-31.10-182’, 2014.
[5] Beton- und Stahlbetonbau Spezial (1) (2015) – Verstärken mit Textilbeton.
[6] Jesse, F.: Tragverhalten von Filamentgarnen in zementgebundener Matrix. TU Dresden, Dissertation,
2004.
[7] RILEM TC 232-TDT (2010): Uniaxial tensile test - test method to determine the load bearing behav-
iour of tensile specimens made of Textile Reinforced Concrete. Entwurf vom Juli 2011, Veröffentli-
chung in Vorbereitung.
[8] Lorenz, E.; Schütze, E.; Schladitz, F.; Curbach, M: Textilbeton – Grundlegende Untersuchungen im
Überblick. Beton- und Stahlbetonbau 108 (2013) 10, S. 711-722.
[9] Lorenz, E.; Ortlepp, R.: Basic research on the anchorage of textile reinforcement in cementitious ma-
trix. In: Oehlers, D.J.; Griffith, M.C.; Seracino, R. (Hrsg.): 9th International Symposium on Fiber-
Reinforced Polymer Reinforcement for Concrete Structures (FRPRCS-9), Sydney, 13.-15.07.2009.
Book of Abstracts – ISBN 978-0-9806755-0-4, 2009, S. 136.
[10] Lorenz, E.: Untersuchungen zur Endverankerung und Übergreifung textiler Bewehrungen in Verstär-
kungsschichten aus Textilbeton. TU Dresden, Dissertation (in Vorbereitung).
[11] Ortlepp, R.: Untersuchungen zur Verbundverankerung textilbewehrter Feinbetonverstärkungsschich-
ten für Betonbauteile. TU Dresden, Dissertation, 2007.
[12] Ehlig, D.; Jesse, F.; Curbach, M.: High Temperature Tests on Textile Reinforced Concrete (TRC)
Strain Specimens. In: Brameshuber, W. (Hrsg.): 2nd International Conference on Textile Reinforced
Concrete (ICTRC2010), Proceedings of the International RILEM Conference on Material Science
315
(MatSci), Aachen, 6.-10.9.2010, Vol. 1, pp. 141-151.

[13] Butler, M.; Mechtcherine, V.; Hempel, S.: Auswirkungen der Matrixzusammensetzung auf die Dauer-
haftigkeit von Betonen mit textilen Bewehrungen aus AR-Glas. Beton- und Stahlbetonbau 104 (2009),
H. 8, S. 485-495 - doi: 10.1002/best.200900003.
[14] Schütze, E.; Lorenz, E.; Curbach, M.: Static and Dynamic Fatigue Strength of Textile Reinforced
Concrete. In: IABSE Conference Nara 2015, Nara, 13.-15.05.2015.
316
ELECTRICAL RESISTANCE MEASUREMENTS ON TRC TENSILE

COUPONS
Vassilis D. Plamantouras (1) and Catherine G. Papanicolaou (1), (1) Civil Engineering Department
University of Patras, Greece
Abstract: This paper presents the preliminary results of an ongoing experimental program
aiming at assessing the piezoresistivity of carbon textile reinforced concrete dumbbell
specimens under monotonic tensile loading, along the direction of loading. During testing
both longitudinal strain and longitudinal electrical resistivity were recorded; electrical re-
sistivity measurements were realized using a high-precision multimeter. The results of this
experimental campaign may be used for setting the ground for establishing appropriate
correlation factors between damage progression and piezoresistivity properties for TRC
elements.
INTRODUCTION
Textile-reinforced concrete (TRC) is a high performance composite material comprising

layers of a textile (a fiber grid structure) embedded in an inorganic matrix, (e.g. a cement-
or lime-based mortar or even a geopolymer). Structural applications of TRC in various
fields such as the precast sector (e.g. for stay-in-place participating formworks) or the up-
grading/seismic retrofitting of structurally deficient reinforced concrete or masonry mem-
bers are well-established and sufficiently documented in the academic literature.
There is a current marked trend for enriching structural materials with capabilities other
than structural rendering them multifunctional while in service. Probably, the most promi-
nent such capability is a sensory one. The latter refers to the ability of the material to sense
its own condition (physical or mechanical). In particular, strain or damage sensing is valu-
able for structures for the purpose of damage origin identification and is related to structur-
al health monitoring (SHM) and smart structures. Material sensory abilities can be attained
by either embedding/attaching sensors in/on the material or using the material itself as a
sensor. In the first case, the interventions entail a toll related to the high cost of the devices
used and – when sensors’ embedment is the method of choice – to their relatively low du-
rability which is often accompanied by degradation of the mechanical properties of the
modified matrix material. In the second case, self-sensing is achieved by exploitation of
certain physical phenomena and material properties. Piezoresistivity is an electromechani-
cal phenomenon in which the electrical resistivity of a material changes with strain, which
relates to stress. Since the early ’90s (see [1]) the high electric conductivity (volumetric
impedance) of carbon fibers has fueled numerous studies towards the investigation of the
piezoresistive behavior of carbon fiber composites (both polymer- and cement-based) and
its exploitation for the development of smart composites. In the field of short or continuous
unidirectional carbon fiber epoxy-and cement-matrix composites fundamental work has
317
been conducted by [2]-[7]. Especially for carbon fiber cement-based composites the work
of [8]-[12] and [13] must also be cited.
The scope of this study is to assess the piezoresistivity of carbon textile reinforced concrete
dumbbell specimens under tensile loading, along the direction of loading, in an attempt to
correlate piezoresistivity and damage progression in this type of composites. The results
presented herein are preliminary as the work is ongoing.
EXPERIMENTAL PROCEDURE
Materials and specimens
Three dumbbell-type TRC specimens were used for tensile testing to accommodate con-
figurations with one, two and three carbon fiber textile layers; specimens’ designation is
1L, 2L and 3L, respectively, whereas specimen geometry is given in Figure 1 for the 2L
case. This type of specimen layout (introducing an area of reduced thickness) ensures (to a
certain extend) cracking development within the part of the specimen extending between
its loaded ends (see [14]). For specimen 3L the thickness of the load introduction area and
the intermediate part of the specimen was 25 mm and 19 mm, respectively. The inter-layer
spacing was 2 mm and 5 mm for 2L and 3L specimens, respectively. In the single-layer
specimen the textile was placed at mid-thickness. Specimens were cast in a high precision
open steel mould. Textiles were secured in place under a slight tension by the use of spac-
ers bolted at the ends of the mould.
Figure 1. Specimen 2L layout (based on [14]).
A commercial textile with equal quantity of high-strength dry carbon fiber rovings in two
orthogonal directions was used. Mid-tow spacing and grid opening were equal to 7 mm
and 10 mm, respectively. The weight of carbon fibers in the textile was 348 g/m2 and the
nominal thickness of each layer (based on the equivalent smeared distribution of fibers)
was 0.097 mm. The guaranteed tensile strength of the fibers in each direction was taken
from technical sheets of the producer equal to 3800 MPa. The elastic modulus of the fibers
was 225 GPa.
318
The textile layer(s) were embedded in a mortar consisting of a commercial ready-mix ce-
ment-based thixotropic binder containing redispersible polymers at a water-to-binder ratio
of 0.2 by weight, resulting in plastic consistency and good workability with a retention
period of approximately half an hour in ambient temperature (20ºC). The binder’s 28-day
flexural and compressive strengths were 4.0 MPa and 17.9 MPa, respectively.
Specimens were removed from the mould 24 hours after casting and moist-cured for 28
days; testing took place soon after. For direct-current (DC) electrical resistance measure-
ment in the loading direction the specimens had to undergo special preparation. As the
four-probe method was opted for the above measurement, four shallow grooves (approx. 1
mm deep) were formed at specified cross-sections of each specimen (symmetrical about
specimen’s mid-length) around their whole perimeter using a hand-held grinder. Next, the
grooves were covered with silver conductive paint using a brush cap. The specimens were
then put in the oven for heat curing at 125°C for 10 minutes so that maximum silver paint
conductivity was achieved. After that, the specimens were left at room temperature to cool.
A small quantity of soldering paste was then applied along the grooves each of which re-
ceived an electrode (an uncoated copper wire embedded in the paste-filled groove). The
copper wires were kept in place by soldering. The above-described procedure was the re-
sult of a trial-and-error procedure that aimed at the elimination of wire-to-mortar contact
resistance phenomena. The two outer contacts (electrodes) were used for passing current
and were 220 mm apart whereas the two inner ones were used for measuring voltage dis-
tancing 200 mm. Snapshots of the contacts’ application procedure are given in Figure 2.
Figure 2. Snapshots of the contacts’ application procedure.
Test set-up
Tests were carried out in lab conditions (T = 20 ± 2 °C, RH = 65 ± 5 %) using a universal
testing machine furnished with a 250 kN MTS actuator. The load introduction system
comprised steel wedges adjusted to the radius of the load introduction zone of the speci-
mens (curved parts); each wedge could rotate in the plane of the specimen so that perfect
wedge-to-specimen contact was ensured throughout each test. Thin rubber sheets with a
thickness of 0.5 mm were applied in between the wedges and the specimens to avoid stress
concentrations phenomena. In order to ensure self-alignment of the load introduction sys-
tem the latter was equipped with hinges having in-plane rotational capacity in regards to
the specimens.
319
Tests were performed in a displacement-controlled mode at a rate of 0.008 mm/sec. De-

formation measuring devices (LVDTs) were mounted on both specimen’s sides. The DC
electrical resistance in the loading direction was measured using a Keithley 2002 multime-
ter. The measured resistance accounted for the bulk resistance between the inner contacts.
Load and deformation data acquisition was automatically synchronized with the one for
electrical resistance at a rate of 5 Hz using a programmable data acquisition system. Each
test was run in a fully computerized manner and was completed when specimen failure
occurred. Tensile test set-up details and instrumentation is given in Figure 3a, whereas the
resistance measurement configuration is given in Figure 3b.
(a) (b)
Figure 3. (a) Tensile test set-up details and instrumentation and (b) Resistance measurement configuration.
Prior to tensile testing, each specimen – while at rest – was connected to the multimeter
(supplying DC power) in order to reach a fractional resistance (ΔR/R0) plateau (fractional
resistance being the ratio of the change in electrical resistance, ΔR – in regards to the ini-
tial one, R0 – over the initial electrical resistance). As has also been shown by [15] and
[16], continuous DC resistance measurement of carbon fiber reinforced cement-based ma-
terials for an extended time causes electric polarization, which results in increase in the
measured resistance. Electric polarization refers to the phenomenon in which the centers of
positive and negative charges do not coincide. Carbon fiber reinforced cement composites
are dielectric materials (insulators); the electrical resistivity of carbon fibers is 10-4 Ω·cm
(generally, accounting for volume fractions less than 10% in the composites), whereas that
of cement paste (more than 90% by vol. of the composite) is 105 Ω·cm. When a dielectric
material is placed in an electric field, electric charges do not flow through the material as
they do in a conductor, but only slightly shift from their average equilibrium positions.
As a result, positive charges are displaced toward the field and negative charges shift in the
opposite direction (i.e. the polarization-induced electric field in the material is opposite in
direction from the applied electric field causing material resistance increase). Polarization
effect on cement-based materials reaches a plateau (polarization saturation) after different
time durations for each specimen (depending on ambient conditions, e.g. temperature,
among others). From test results on dumbbell TRC specimens (not shown herein) the po-
larization saturation time threshold ranges between ½ h and 3 h. In Figure 4, the evolution
of fractional resistance (ΔR/R0) with time is given for specimen 2L. For this case, polariza-
tion saturation occurred after 1500 s at which point tensile testing could start; in this way,
320
electrical resistance measurements were not influenced by the polarization effect but were
solely dependent upon strain development.
Figure 4. Fractional change of resistance versus time for the 2L specimen.
TEST RESULTS
Before monotonic tests were conducted the uniaxial tension loading protocol was restricted
to the elastic regime of the composite materials and consisted in a series of repeated and
reversed tension at increasing strain amplitudes. During these tests the variation of frac-
tional resistance change, ΔR/R0, with strain (or stress) was unclear. Therefore, it was de-
cided to explore the ΔR/R0 to strain relationship for the entire strain regime of the ten-
sioned specimen up until failure.
Figure 5 shows the relationship between stress and strain and that between fractional re-
sistance change, ΔR/R0, and strain during monotonic tensile testing up to failure for the
three specimens referenced herein (1L, 2L and 3L). It is noted that the fiber volume frac-
tions in the longitudinal direction were equal to 0.97%, 1.94% and 1.53% for specimens
1L, 2L and 3L, respectively. A general trend is revealed from Figure 5: the fractional re-
sistance change is negligible throughout the material’s elastic regime (while the mortar
matrix remains uncracked and the fibers do not participate as much to the load uptake).
First crack occurrence initiating the multiple cracking stage of the material behavior is ac-
companied by an abrupt increase in fractional resistance change. This change seems to be
sensitive to individual crack formation (instant increase in ΔR/R0 for each load drop at
crack opening). The fractional resistance change continues to increase after the end of the
multiple cracking stage (stage IIa) for the entire part of the material’s response correspond-
ing to the stabilized crack pattern stage (stage IIb); the inclination of the ΔR/R0 vs. strain
curve differentiates between stage IIa and IIb (depending upon the degree of stages’ differ-
entiation of the stress-strain curve itself).
321
Figure 5. Stress and fractional change of resistance versus strain curves.
It is hypothesized that fractional resistance change increase during stage IIa is attributed to
the loss of mortar continuity (though mortar’s contribution to the conductivity of the com-
posite is small), to the mortar-to-longitudinal rovings progressive bond degradation and to
fibers’ strain built-up; during stage IIb ΔR/R0 increase is assumingly owed to further fi-
bers’ strain built-up (fibers are the sole carriers of the load at this stage) and to rupture of
individual carbon filaments.
The fractional change in resistance per unit strain (gage factor of the carbon fiber “TRC
sensor”) at stage IIb for specimens 2L and 3L is approximately 28, whereas for specimen
1L is equal to 7. Nevertheless, due to the preliminary nature of these results the computed
gage factors must be considered as boxed values. More tests are needed in order to derive
322
reliable correlation factors between fractional change in resistance and strain for carbon
fiber TRC undergoing tension.
CONCLUSIONS
In this paper preliminary results of an ongoing experimental program aiming at assessing

the piezoresistivity of carbon fiber TRC dumbbell specimens undergoing tension are pre-
sented. The electrical conductivity of carbon fibers and the slight conductivity of the ce-
ment-based matrix allow for the measurement of the DC electrical resistance of this type of
TRC which provides a means to detect the onset of the material’s different stages of tensile
behavior. The detection of stage IIa or IIb onsets are of particular interest for real-life ap-
plications and could trigger respective alarm signals if recorded by a SHM system that
takes advantage of the inherent carbon fiber TRC self-sensing abilities.
REFERENCES
[1] Crasto, A. S., & Kim, R. Y. (1993, October). Using carbon fiber piezoresistivity to measure residual
stresses in composites. In: Proceedings-American Society for Composites (pp. 162-162). Technomic
Publishing Ag.
[2] Wang, X., & Chung, D. D. L. (1995). Short-carbon-fiber-reinforced epoxy as a piezoresistive strain
sensor. Smart Materials and Structures, 4(4), 363.
[3] Wang, X., & Chung, D. D. L. (1996). Continuous carbon fibre epoxy-matrix composite as a sensor of
its own strain. Smart materials and structures, 5(6), 796.
[4] Wang, X., & Chung, D. D. L. (1997). Real-time monitoring of fatigue damage and dynamic strain in
carbon fiber polymer-matrix composite by electrical resistance measurement. Smart materials and structures,
6(4), 504.
[5] Wang, X., & Chung, D. D. L. (1997). Sensing delamination in a carbon fiber polymermatrix composite
during fatigue by electrical resistance measurement. Polymer Composites, 18(6), 692-700.
[6] Wang, X., Fu, X., & Chung, D. D. L. (1999). Strain sensing using carbon fiber. Journal of materials
research, 14(3), 790-802.
[7] Wen, S., Wang, S., & Chung, D. D. L. (2000). Piezoresistivity in continuous carbon fiber polymer-
matrix and cement-matrix composites. Journal of materials science, 35(14), 3669-3675.
[8] Wen, S., & Chung, D. D. L. (1999). Piezoresistivity in continuous carbon fiber cement-matrix
composite. Cement and concrete research, 29(3), 445-449.
[9] Wen, S., Wang, S., & Chung, D. D. L. (2000). Piezoresistivity in continuous carbon fiber polymer-
matrix and cement-matrix composites. Journal of materials science, 35(14), 3669-3675.
[10] Wen, S., & Chung, D. D. L. (2000). Uniaxial tension in carbon fiber reinforced cement, sensed by
electrical resistivity measurement in longitudinal and transverse directions. Cement and concrete research,
30(8), 1289-1294.
[11] Wen, S., & Chung, D. D. L. (2001). Carbon fiber-reinforced cement as a strain-sensing coating. Cement
and Concrete Research, 31(4), 665-667.
[12] Wen, S., & Chung, D. D. (2006). Effects of strain and damage on strain-sensing ability of carbon fiber
cement. Journal of materials in civil engineering, 18(3), 355-360.
[13] Azhari, F., & Banthia, N. (2012). Cement-based sensors with carbon fibers and carbon nanotubes for
piezoresistive sensing. Cement and Concrete Composites, 34(7), 866-873.
[14] Andreadi A.-I. (2011). Experimental investigation of the mechanical behavior of TRC under uniaxial
tension. M.Sc. Thesis (in Greek), University of Patras.
[15] Wen S, Chung DDL. Electric polarization in carbon fiber reinforced cement. Cem Concr Res
2001;31(2):141–7.
[16] Cao J, Chung DDL. Electric polarization and depolarization in cement-based materials, studied by
apparent electrical resistance measurement. Cem Concr Res 2004;34(3):481–5.
323
324
INTEGRATED MONITORING OF TRC USING CARBON FIBRES
Y. Goldfeld 1, O. Rabinovitch 1, T. Quadflieg2 and T. Gries2, 1 Faculty of Civil and

Environmental Engineering, Technion – Israel Institute of Technology, Israel, 2 Institut fuer
Textiltechnik, RWTH Aachen University, Germany.
Abstract: The development of self-sensory capabilities for structural health monitoring of

concrete structures in general, and, particularly, of textile reinforced concrete (TRC) struc-
tures opens a new spectrum of opportunities. Carbon fibre tows that are embedded within
the textile can serve as the reinforcing component for the structural action as well as the
self-sensing component. The main advantage of this technology is that there is no need for
additional sensors and that the monitoring agent is located within the load bearing element
serving as an inherent part of its structural form. Preliminarily results quantitatively
demonstrated the feasibility of the sensing concept in two applications: detecting strains in
a mechanically loaded structural element and monitoring the interaction of the structural
element with a wet environment. This paper aims to present an overview of this innovative
technology. The paper addresses aspects ranging from the production process of the hybrid
carbon based TRC structural element to its mechanical performance and its structural, en-
vironmental, and functional self-sensory capabilities. The paper also aims to highlight new
directions for future research into the emerging technology.
INTRODUCTION
Textile Reinforced Concrete (TRC) is an excellent building material with advantages rang-
ing from sustainability to applicability to retrofitting of existing structures. Its advantages
in terms of reduced greenhouse emissions due to the minimized use of cement, the ability
to cast thin-walled structural elements that can easily adhere to complex geometries, and
the flexible design possibilities designate this system as an excellent candidate for many
applications. These advantages can be greatly advanced by integrating monitoring func-
tionalities into the reinforcement textile. Since TRC elements are commonly precast, the
integration of sensors does not cause inefficiency in production and mitigates the risk of
damage to the sensors during fabrication. This study aims to provide an overview on the
development of a TRC element with self-sensory capabilities, to discuss some aspects of
its feasibility, and to designate present and future research directions towards the ad-
vancement of this emerging technology.
Carbon fiber reinforcement is characterized by high load bearing capacities. In addition, it

is highly resistant to corrosion. This makes it perfectly suitable for the use in TRC for vari-
ous types of structural components [1]. Production methods of TRC components include
laminating, casting, spraying, spinning, and injection molding. By optimizing the arrange-
ment of the carbon fiber tows in the textile grid, the amount of reinforcement can be con-
siderably reduced and cost-effective structures are achievable. As an example, the pedes-
trian pridge in Albstadt (Germany), which is made of TRC, weighs only a quarter of what a
traditional reinforced concrete bridge would have weighed, see Figure 1 [2]. The reduction
of the environmental impact associated with the construction of such facility and the poten-
325
tial contribution to the sustainability of the built environment cannot be overestimated. In

the near future, it is very plausible that TRC will be found in applications such as roofs and
balconies; noise barriers; precast elements; lost formwork for slabs, beams and columns;
sandwich structures, and water tanks and pipelines. Along with new concrete elements,
strengthening of existing concrete structures by thin textile-reinforced concrete layers is
another field of applications already possible. In the particular case of pipelines and tanks,
the use of textile based reinforcement allows the reduction of wall thickness, mainly by
means of the superior corrosion resistance of the reinforcement material, see Figure 2.
Since such applications are inherently susceptible to leakage due to concrete cracking, they
become a natural candidate to the application of the concept of a structural system that
combines structural health monitoring and functionality monitoring capabilities.
Monitoring techniques for reinforced concrete elements range from traditional ones such as
visual inspection or tap tests, to modern techniques based on embedded or surface mounted
strain sensors, either localized or distributed. When the damage is substantial, these meth-
ods can determine if damage has occurred. At incipient stages of damage, however, some
of the existing methods are not so successful. In addition, the requirement of physical ac-
commodation in the load-bearing element and the potentially degradation of its effective
properties, as well as the clear distinction between the structural system and the sensory
one (which is definitely not a part of the load bearing system) are additional drawbacks of
these methods. The sensory system itself is commonly expansive and its implementation is
costly, time-consuming, and it requires highly qualified workers.
Figure 1: Pedestrian bridge in Albstadt (Germany) - a cooperative project of RWTH Aachen University with
GrozBeckert. (Ref. [2])
Figure 2: Prototype of textile reinforced pipe, (ITA group).
326
A TRC element with an inherent sensing capability, combines the advantages and the fa-
vorable features of the TRC technique and uses them to overcome some of the challenges
associated with the development of a smart structure. In the concept overviewed here,
glass/carbon fiber based textiles are used as the reinforcement for the TRC element and the
electro-structural characteristics of the carbon tows are used as a basis for its sensory fea-
ture.
Records of the use of the electrical conductivity features of continuous carbon fibers em-
bedded in concrete or cement paste for sensing purposes are found in the literature, see
Refs. [3-4]. However, in these cases, the investigation was limited to the component level
where the tensile strain and the corresponding changes to the electrical properties are
equally distributed along the fiber. A recent investigation that move from the fundamental
single tow scale to the integrative TRC structural element scale has been presented by
Goldfeld et al. [5]. In that study, a glass/carbon based TRC structural beam element was
investigated. The same array of carbon tows was used as the main reinforcement required
for the load resisting system and, at the same time, as the basis for the sensory system. This
combination aims to provide the structure with the self-sensory features. In this configura-
tion, the carbon tows were embedded in a glass textile net, as part of the production of the
textile. By utilizing the self-sensing capacity gained by means of the piezoresistive proper-
ties of the continuous carbon fibers, the cost and labor inputs needed for converting the
component into a sensory one are significantly reduced or even eliminated. Ref. [5]
demonstrated the feasibility of the multi-functionality of the carbon based TRC. The exper-
imental investigation outlined in [5] demonstrated the reinforcement capability of the car-
bon based textile, as well as the sensory feature of the carbon tow in two applications: de-
tecting strains in a mechanically loaded TRC beam and monitoring the interaction of the
structural element with a wet environment. This paper aims to present an overview of the
concept outlined in [5] and to designate discuss and further explore additional new direc-
tions of research into the self-sensory TRC technology.
INTEGRATED SENSORY-STRUCTURAL TEXTILES
The structural-sensory textile used in [5] combines two types of tows: tows made of glass
fibers and tows made of carbon fibers. Optionally, it can also include embedded yarns
made of metallic filaments, see Ref. [1, 6-8], or other materials. The glass tows provide the
main reinforcement platform. The carbon tows are also part of the main reinforcement and,
at the same time, their electromechanical coupling of strain and electric resistivity (piezo-
resistivity) are used as the sensory agent. This feature designates the main advantage of the
glass/carbon system. Metallic yarns embedded in the glass-based textile, [7-8], also use the
electromechanical features of the material for sensory capabilities but they do not contrib-
ute to the load carrying capacity.
As a representative example of the spectrum of designs, a warp knitted grid structure with
a mesh size of 7-8 mm is selected for the textile. The carbon tows have been inserted into a
glass fiber grid in the warp knitting process as warp yarn. In the weft direction, glass fiber
tows are used. The reinforcing tows are then combined with warp- knitting yarns made of
polypropylene. The stitch type of the warp knit is pillar. The production of the multi-
material textile is illustrated in Figure 3. The production process is not involved with any
modifications in terms of tribology or tension control of the fibers. This demonstrates the
327
suitability of the selected fiber materials to be used in standard processes without any cost
intensive modifications. The use of standard processes and conventional materials as well
as the other hand, the design flexibility gained by the well-established textile production
technology paves the way to a variety of designs. These may range from dense textiles for
fine mortars or grouts to sparse ones that can be used with coarse aggregates. More appli-
cations including reinforced soil, reinforced mortars for cladding applications, or reinforc-
ing materials for rapid and versatile retrofitting of seismically damaged structures can also
be considered.
SELF-SENSORY STUCTURAL ELEMENTS
Concrete beams reinforced with the sensory textile can have several textile layers. From
the construction point of view, it is recommended that the beam is reinforced with at least
two layers of textile, one close to the upper face of the beam and the other close to the low-
er faces of the beam. In order to maintain a preliminary low level of pre-tensioning of the
textile and yet to allow penetration of the concrete mix without segregation, special atten-
tion should be drawn to the design and construction of the mold. In Ref. [5], it was sug-
gested to use spacers that are placed at the sides of the mold keeping the layers of textile at
the desired height. The layers could then be manually tensioned while the wooden spacers
are pressed against the mold. An example of the mold used in [5] is illustrated in Figure 4.
Each textile layer should include longitudinal (0o) tows of sensory carbon tows and glass
fibers, while the transverse (90o) tows should be made of only glass fibers. This feature, as
well as the aforementioned demands regarding the molds, define interesting and important
challenges associated with the scaling up of the coupon scale beams examined in [5].
While the geometry scales up easily, the properties of the sensory textile reinforcement
cannot be changed much. This necessitates special techniques for holding the textile layers
in place and for providing them with the proper level of pre-tensioning during casting. It
also draws the attention to the type of equipment to be used for casting and vibrating the
concrete and for the means to be taken in the curing phase.
SENSING WITH TRC ELEMENTS
The fundamental physical phenomena upon which the monitoring capabilities of the TRC
beam are based are derived from the piezoresistive changes to electrical resistance of the
carbon tows in the textile reinforcement.
Figure 3: Warp knitting machine MALIMO during production (ITA group)
328
Figure 4: Mold of TRC beam after placing the layers of sensory textile and before casting.
Taking the advantage of the continuous configuration of the carbon tow a Wheatstone
bridge apparatus is used for measuring the voltage and the voltage change across the
bridge, [5]. By means of the Wheatstone bridge layout and by using the known properties
of all other resistors, the voltage change across the bridge can be converted to changes of
resistance along the carbon tow. In this method, the two ends of the carbon tow are con-
nected to the DAQ and there is no need for additional external connectors along the tow.
This simplifies the installation of the tow in the concrete element. Therefore, it enables to
use this method for a spectrum of concrete structures without changing their structural ge-
ometry. This versatility in terms of geometry is particularly relevant to shell and shell-like
structures. In this case, the combination of the TRC technique with the self-sensory fea-
tures define an interesting spectrum of potentially smart and effective thin walled struc-
tures with the following capacities:
Strain Sensing: The detection of strain is based on the basic relation between the electric
resistance and the ratio between the length of the resistor and its cross-sectional area. Ten-
sion straining of the tow yields elongation and reduction of the cross sectional area. Both
effects increase the integrative electrical resistance of the tow. Since the strain is not uni-
formly distributed along the beam (and therefore along the tow), and since the resistance
changes are measured from one end of the tow to another, the measured quantities are con-
sidered to be an integrative indexes. This feature designates the localization of the strain
measurement as one of the most challenging tasks in the further development of this tech-
nology. In [5], the sensing capability was indeed focused on the integrated value. Thus, the
ability to capture the localized piezoresistive effect associated with localized straining or,
on the other hand, the segmentation of the monitored beam and the definition of an optimal
spatial resolution are still open questions. These questions define essential direction where
future research is called for.
Monitoring Wet Environments: The wetting of the carbon tow and the surrounding con-
crete body has several types of potential effects on the localized electrical scheme of the
wetted region [5]. First, the aqueous solution itself is conductive. Although the specific
ionic conductivity of the solution is smaller than the electronic conductivity of the carbon
fibers, it may still affect the localized properties of the wetted tow. Damage in the form of
cracking yields localized straining of the carbon tows that bridge over the crack. If the
structure is exposed to a wet environment, cracking also allows the penetration of water
through the crack and the wetting of the tow. Both effects impact the electrical resistance
of the tow and allow the detection of such damage. Yet, the issue of localization and the
329
temporal/spatial resolution of the detection apply to this case as well. Due to the continu-
ous nature of the carbon tows, the system can sense a wetting event but cannot specifically
trace its location along the tow. Technological solution that can face this challenge, e.g.
segmentation, modulation in the time domain, dynamic methods, etc. define another set of
future research directions.
HYBRID CARBON BASED TRC STRUCTURAL ELEMENTS
Smart textile can be effectively used as a multi-functional sensory reinforcement for TRC
elements. According to Ref. [5], the structural behavior of the concrete element reinforced
with the smart textile reveals a satisfactory structural behavior in terms of the initial pre-
cracked linear behavior, the post cracking behavior, the distribution and width of cracks,
the behavior at ultimate stages, and the mode of failure. At the same time, the smart senso-
ry textile provides continuous electrical readings. These readings can be quantitatively cal-
ibrated and correlated with the mechanical state of the structural element [5]. The readings
of the smart textile also provide an effective indication of wetting and of the coupling of
wetting and cracking. These features highlight the possibility of using smart and sensory
TRC elements for the continuous monitoring of the structural health of TRC elements. Fur-
thermore, it highlights the potential of using smart TRC rapid or even automatic assess-
ment of structures after extreme events as part of a natural disaster control enterprise.
An example of the normalized electrical response of such smart TRC beam is illustrated in
Figure 5. The details of the tested beam are outlined in Ref [5] but opposed to the beams
studied in [5], the sensory behavior examined here focuses on a beam that was manufac-
tured with two types of sensory textile tows. One made of carbon fibers and one made of
stainless steel filaments (see Ref [5] for the mechanical performance of such beam). Figure
5 presents the response of the beam before and after it was subjected to the mechanical
loaded. The curves shown in Figure 5 refer to the voltage response of the beam to wetting
with tap water. The curves are then normalized to fit the non-dimensional range of -1 to 1
using the same normalization factor. In other words, the normalization factor is determined
based on the post-loading response of the tow. The same scaling factor is applied to the
pre-loading curve as well.
The results reveals that after the exposure to load and the formation of flexural cracks, the
sensory beam effectively detects the wetting event (also see [5]). Before the mechanical
loading, the electrical response to wetting is insignificant and the detected values are 3-4
orders of magnitude smaller than the ones observed after cracking. The implications of this
type of behavior are two-fold. First, it allows to directly detect wetting, penetration of wa-
ter through cracks, and leakage. Second, it allows to detect cracking that takes place be-
tween one wetting event and another. The combination of these two capabilities is a no-
ticeable advantage in the design and realization of monitoring systems for TRC structures
that may be exposed to a wet environment.
330
Figure 5: Response to wetting before and after exposure to mechanical loading of beam CM1. (a) Carbon
tow; (b) Metallic yarn.
POTENTIAL APPLICATIONS OF TRC STRUCTURAL ELEMENTS
In the light of the possibility of using carbon based TRC as a multifunctional structural-
sensory system, future research in this field will pursue studying the possibility of devel-
opment and the realization of the technology of smart TRC members. Specifically, the
planned research effort will further study the application of the smart TRC for the simulta-
neous structural and functional health monitoring of load bearing elements. As a particular
example, future research will focus on the detection of leakage due to structural failure in
wet infrastructural systems such as concrete pipelines in drinking water supply systems.
All these infrastructure systems are commonly buried way below the urban ground level,
and, in most cases, they are difficult to access and inspect. The implementation of a com-
bined structural and functional health monitoring sensory network system and the ability to
detect either structural or functional failure (or both) will impact the built environment in
four aspects:
(a) It will provide an online measure of the structural condition of the infrastructure system
and will potentially provide warning before failure. This will minimize damage and down
time of critical systems and support better allocation of maintenance resources.
(b) It will avoid waste of drinking water and diminish the risk of environmental pollution
due to leakage of sewage.
(c) It will support the early and effective detection of manmade sabotage to critical infra-
structures and therefore help to avoid the chain reaction that such sabotage may trigger.
(d) It will provide effective guidance for emergency and rescue teams, right after a large
natural disaster event. For example, in the event of a high magnitude earthquake striking a
densely populated urban area, a well-monitored water supply system can effectively and
immediately point at the highly damaged regions. This information can then be immediate-
331
ly used for guiding search and rescue teams to the relevant regions of the urban environ-
ment.
CONCLUSIONS
This paper has presented an overview of the emerging technology of self-sensory TRC
elements. This technology combines structural features with sensory ones and potentially
provides a self-sensory structural element made of TRC. Together, the various capabilities
of the sensory element define an appealing direction in the development of smart concrete
structures. This paper has discussed these capabilities and outlined essential research direc-
tion that are needed in order to convert this concept into a feasible technology.
ACKNOWLEDGEMENTS
This research was supported by the Umbrella Cooperation Program of the Technion- Israel
Institute of Technology, RWTH Aachen University, and JULICH Forschungszentrum, and
the New York Metropolitan research fund. Their financial support is gratefully acknowl-
edged. The authors are also grateful for the help of Barak Ofir, Elhanan Yitzhak, and the
technical and administrative stuff of the National Building Research Institute at the Tech-
nion.
REFERENCES
[1] Berger, U.,Hausding, J., Cherif, C., Quinonens, V., Plaggenborg, B., Zeidler, G. and Weiland, S. 2011.
"Production concepts for textile reinforcements made of 50K carbon filament yarn for civil engineering
applications / Herstellungskonzepte für die Textilverstärkung mit 50K Kohlenstoff-Filamentgarn für
Bauingenieuranwendungen" New Material Characteristics to Cover New Applications Needs, SEICO,
SAMPE Europe International Technical Conference & Forum, Paris, FR, Mar 28-29, 2011.
[2] Weiland S.; Engler, T.: Superlightweight / Superleichte Brücke. 2nd Aachen-Dresden In-ternational Tex-
tile Conference - Aachen-Dresden Textiltagung, ITB der TU Dresden, Dres-den, DE, 4.-5. Dez, 2008.
[3] Wen, S. and Chung D.D.L. 1999. "Piezoresistivity in continuous carbon fiber cement-matrix composite",
Cement and Concrete Research, 29, 445-449.
[4] Wen, S., Wang, S. and Chung D.D.L. 2000. "Piezoresistivity in continuous carbon fiber polymer-matrix
and cement-matrix composites", Journal of Materials Science, 35:3669-3675. [5] Goldfeld, Y. Rab-
inovitch, O, Fishbain, B, Quadflieg, T., and Gries, T. "Sensory carbon fiber based textile reinforced con-
crete for smart structures”, Journal of Intelligent Material Systems and Structures, in print 2015.
[6] Hegger, J., Golraski, C. And Kulas M. C. 2011. "Schlanke Fußgängerbrücke aus Textilbeton –
Sechsfeldrige Fußgängerbrücke mit einer Gesamtlänge von 97 m", Beton und Stahlbau, 106(2), 64-71.
[7] Quadflieg, T., Tomoscheit, S. and Gries, T. 2013. "Humidity and Strain Monitoring for Textile Rein-
forced Concrete", Second Conference on Smart Monitoring, Assessment and Rehabilitation of Civil Struc-
tures, Istanbul, Turkey, 9-11 September 2013
[8] Goldfeld, Y., Rabinovitch, O., Quadflieg, T., Fishbain, B. and Gries, T. (2014) "Smart Textile Reinforced
Concrete Sensory Structures", Proceedings of the 7th European Workshop on Structural Health Monitor-
ing, Nantes, France, July 8-11, 2014.
332
CHAPTER 8
APPLICATION
333
334
SLENDER FAÇADE STRUCTURES MADE OF TEXTILE-

REINFORCED HIGH PERFORMANCE CONCRETE
Sergej Rempel, Josef Hegger, RWTH Aachen University, Institute of Structural Concrete, Mies-
van-der-Rohe Str. 1, 52074 Aachen, Germany
Abstract: The innovative composite material textile-reinforced concrete (TRC) opens up

new application areas in structural concrete engineering and has advantages especially in
the field of façade engineering. By using non-corroding technical textiles made of alkali-
resistant glass (AR-glass) or carbon, the concrete covers can be minimized resulting in
extremely slender concrete members. This leads to material savings, which add up to 70 %
compared to common solutions for façades. The use of fine-grained concrete with maxi-
mum grain-sizes of 8 mm in combination with special curing techniques enables high-
quality fair-faced concrete surfaces. This paper summarizes basics material properties as
well as calculation models and focuses on the presentation of several projects in the field
of façade engineering. Here, TRC panels with areas up to 14 m² and a thickness of only 30
mm have been realized within industrial-scale projects.
INTRODUCTION
Steel-Reinforced (RC) façade-systems became less attractive due to architectural design in
the 1960s and 70s. Additionally, concrete spalling, caused by the corrosion of the steel and
the small concrete cover, added to the unpopularity. The development of the innovative
composite material Textile Reinforced Concrete (TRC) provides a new way of construc-
tion with non-corroding reinforced elements, such as alkali-resistant glass (AR-Glass) or
carbon. Consequently the necessary concrete cover can be reduced significantly. Only a
few millimeters are required to guarantee a good bond-behavior between the concrete and
the textile reinforcement. As a result, thin construction-elements can be realized, which are
impressive due to their little weight and slenderness. Also, the use of concrete with a max-
imum grain size of 8 mm allows sharp edged parts and architectural high-quality surfaces.
335
Figure 1: Façade of the St.-Leonhard school in Aachen (picture: Robert Mehl)
During the collaborative research center SFB 532 research institutes of the RWTH Aachen
and industrial partners developed architectonical TRC façade panels with impregnated tex-
tiles. Although the size reached up to 14 m², the thickness was reduced to 30 mm, for ex-
ample the façade of the St.-Leonhard school in Aachen (Figure 1). The combination of the
size and the low weight of the panels leads to low transport and production costs.
MATERIALS
Actually alkali-resistant glass (AR-glass) or carbon are the main materials used for current
TRC applications. The basic materials are hair-thin filaments with diameters of 14 μm for
AR-glass and about 7 μm for carbon. A bundle of hundreds, up to ten thousands, of these
filaments shape a roving, which will be processed to make flat textile-reinforced grids 1.
The textile grids can be used directly without impregnation or impregnation with Styrene-
Butadiene or Epoxy-Resin as reinforcement elements for concrete structures. The impreg-
nation is responsible for higher failure stresses and consequently for an increase in the ef-
fectivity of the textile materials 2. Also the loss of strength of the AR-Glass, because of the
alkalinity, has to be considered 4. The impregnation counteracts it and reduces the losses.
The failure stresses are dependent on the material and impregnation and can be seen in
Table 1.
336
Table 1: Properties of the selected textiles
Properties Units Textile 1 Textile 2 Textile 3
Project - Community College Laboratory hall „Schiefer-

Leiden (NL) IMB Erlebnis“ Dormet-
tingen
Material - Carbon AR-Glass Carbon
Roving: Producer - Chomarat North OCVTM Rein- solidian
America forcements
Penetration - Epoxy-Resin Epoxy-Resin Epoxy-Resin
Roving distance mm 46 /41 17 ; 8 / 17 38 / 38
Cross section mm²/m 38 / 42 108 / 75 110 / 110
Tensile strength N/mm² 2267 1160 / 1390 3099
Tensile strain ‰ 2,3 12,6 / 12,0 19,1
Table 1 lists the properties of the applied textiles which were used for the TRC façade pro-
jects. The failure stresses and strains were investigated for the main load-bearing directions
of the façade elements. The highest stresses (3099 N/mm²) were reached by the carbon
textile 3, which is impregnated with epoxy-resin. Furthermore, the impregnated textile of-
fers simplified handling, to ensure an easy installation and provide inherently stable rein-
forcements during the casting of concrete. Shaped reinforcements, e.g. for T-beams are
necessary and can only be produced with impregnated textiles 6.
Concrete
The use of a typical concrete with its standard grain size was not possible because of the
small openings in the textile reinforcement grids. For this reason, the Institute of Building
Material Research of the RWTH Aachen University (ibac) developed during the SFB 532 a
recipe for a concrete with a maximum grain size of 0.6 mm 8. During the realization of the
TRC façade projects the industrial partners asked for a higher diameter to optimize the
workability by reducing the cement amount at the same time. By now, concrete mixtures
with maximum grain sizes between 2 mm and 8 mm (Table 2) have shown the most practi-
cal properties. In the case that the surface of the façade needs to be acid-washed, polished,
or blasted, a minimum grain size of 5 mm is advisable. The openings of the textile rein-
forcement grids were adjusted accordingly.
337
Table 2: Properties of the applied concrete
Properties Units Concrete 1 Concrete 2 Concrete 3

Project - Community College Laboratory hall „Schiefer-
Leiden (NL) IMB Erlebnis“ Dormet-
tingen
Maximum grain mm 5 5 8
size
E-Modulus N/mm² Not known 35700 Not known
Compression N/mm² 64,4 80,0
70,9
resistance
Central
N/mm² 4,8 4,0 4,2
Tensile strength
Concreting tech- - Lamination Lamination / cast- Lamination
niques ing
Table 2 lists the properties of the applied concrete which were used for the TRC façade
projects. The compression resistance reaches up to 80 N/mm² and the central tensile
strength up to 4,8 N/mm², which is perfect for requirements as crack free surfaces.
APPLICATION
Small-sized façade-elements
Small façade-elements made out of TRC, with a size of 2 m² and thicknesses of 20 mm, are
used for ventilated façades. The small weight, which led to money-savings due to the less
expensive transport and installation-work, and the architectonic possibilities were used for
numerous applications. Especially the first general construction approval of the TRC fa-
çade-element betoShell® 10 supports the applicability (Figure 2).
Figure 2: Façade panels of the Community College in Leiden (Netherlands): structural details (left); elevation
of the Community College (right)
338
One of the first applications was the 10,000 m² façade at the Community College in Leiden
(Netherlands) (Figure 2). A single element has a size of 1.14 m² (1.780 m x 0.642 m) and a
thickness of 30 mm (Figure 2).
On one hand, the size played an important role for the calculation and design of the panel.
On the other hand, the location close to the North Sea and the height of the building of
about 50 m had to be considered. Consequently an extremely high wind pressure of wk =
3.0 kN/m² has to be carried by the 30 mm thick panel. Therefore an efficient carbon-textile
(textile 1 in Table 1) with an epoxy-resin impregnation was used. The tensile stress
reached 2267 N/mm² during a bond-test under tensile load. Another important detail were
the fixing devices of the panel.
Large-sized façade-elements
Small-sized façade-elements need numerous anchorage elements per square meter, howev-
er this will increase cost. Additionally such elements have a high percentage of gaps,
which is not preferred by architects. For this reason, a team of three parties, Hering Bau
company, Institute of Building Materials Research (ibac) and Institute of Structural Con-
crete (IMB) of RWTH Aachen University, supported by the Federal Ministry of Econom-
ics and Technology (BMWi), developed a large-sized façade-element (betoShell®XXL)
with dimensions of 2.51 m x 4.88 m (=12.2 m²) (Figure 4). At the same time a thickness of
the panel of only 30 mm was realized (Figure 4).
Figure 3: Pilot project: Large-sized façade-panel of the laboratory hall of the Institute of Structural Concrete,
RWTH Aachen University; a) elevation, b) cross-section and the shaped textile grid
On the backside of the panel two concrete webs were arranged to stiffen the panel and to
accommodate the anchorage system. Ordinary steel-reinforced concrete anchors were used
for fixing the panels to the load-bearing wall of the hall. On the contrary, new textile rein-
forcement made out of AR-Glass (textile 2 in Table 1) was installed in the panels, which
was impregnated with Epoxy-Resin. The impregnation allowed a shaping of the textile
339
grids. After the drying process of the Epoxy-Resin was completed, the shape of the grids
persisted (Figure 4) 12.
A further development of the large-sized panels is the façade of the “SchieferErlebnis”

building in Dormettingen (Figure 4). Although the size of the panel is 1,5 m x 4,1 m (A =
6,15 m²), the thickness was reduced to 50 mm. Even the typical concrete webs on the back-
side were not needed. Especially the production process profits from the development of
theses panels by saving time and costs.
Figure 4: Façade of the SchieferErlebnis in Dormettingen (picture: solidian)
A high-performance carbon textile was used for the façade in Dormettingen (textile 3 in
Table 1), which tensile stress reaches up to 3099 N/mm². The anchorage system is a typical
façade application, which was modified for thin concrete panels.
Load-Bearing behavior
Small-sized and large-sized specimens were necessary to examine the maximum load bear-
ing and anchorage capacity of the façade-elements (Figure 5 and 6). Under service loads
the outer facing panel has to stay uncracked, which was a requirement by the building
owner and proofed during the tests. Low deformation and uncracked surface is typical for
TRC façade panels.
340
Figure 5: Setup of a four-point-bending test (left); view of the specimen during the test (right)
As soon as the first cracks appear the textile reinforcement gets activated, which guaran-
tees the carrying capacity at the ultimate limit state. The good bond behavior and the high
textile strength are responsible for the small crack size and a high increase of the loading.
The tests show that the load can be nearly quadrupled.
The dimensioning tools, developed by Voss 14 and Kulas 16 during the transfer project
T08, were used to design the façade panels. Altogether an existing safety level could be
reached, that was nearly twice as much as required.
Figure 6: Anchorage test: Pull-Out (left); combined test (pull and shear) (right)
The anchorage test showed even a higher safety capacity. The investigated resistance was
nearly four times higher than the load caused by wind or weight.
PROCESS
Small-sized façade-panels prefer the lamination process, which is divided in three parts.
After the first concrete-layer is poured into the concrete formwork, the textile was laminat-
ed into the fresh concrete and the second concrete layer was poured in.
341
There are many advantages, especially the easy installation without reinforcement spacers
and the lack of extra work to prevent the textile from floating. Here is to say, that flat tex-
tile reinforcement grids simplify the production compared to rolled up textiles which like
to bulge out.
A combined laminating and casting process is beneficial for large-sized façade panels with
backsided webs (Figure 7). The concrete technique for the panel is the same as for the
small-sized façade-panels. Just the concreting of the web needs an additional step. A
formwork element, which is placed on the fresh concrete and fixed up to the side after the
concreting of the panel is finished, forms the web.
Three days after the production of the element, the surface was treated with acid dissolving
the fine particles and uncovering the grain matrix. This results in a rough surface with a
natural stone character.
Figure 7: Concreting techniques: lamination of the panel (left); casting of the concrete webs (right)
CONCLUSION
The realized projects during the last years demonstrate the applicability and the advantages
of TRC for façades. This paper presented both small-sized as well as large-sized façade-
panel projects. Industrial research programs supported the realization of TRC façade-
panels. Because of the missing standards for TRC, research had to be done to examine the
load bearing capacity. The results are the crack-free visible areas of the panels for the SLS,
and a global safety level of more than 4 could be reached for the ULS. That is up to 40 %
more than the required global safety level.
ACKNOWLEDGEMENT
The authors thank the DFG and the BMWi for the financial support during SFB 532 period
and the transfer project T08. A special gratitude goes to Hering Bau and solidian for their
commitment to support the development of the new composite material TRC.
REFERENCES
1. Offermann, P., Engler, T., Gries, T., Roye, A.: Technische Textilien zur Bewehrung von Betonbau-
teilen. Beton- und Stahlbetonbau 99 (2004) Heft 6, S. 437-443.
2. Raupach, M., Orlowsky, J., Büttner, T., Dilthey, U., Schleser, M.: Epoxy-impregnated textiles in
342
3. concrete – load bearing capacity and durability. In: Hegger, J., Brameshuber, W., Will, N. (Hrsg.):
Proceedings of the 1st International RILEM Conference, Aachen (2006). RILEM Proceedings PRO
75. RILEM Publications S.A.R.L., ISBN 2-912143-97-7, S. 77-88.
4. Büttner, T., Orlowsky, J., Raupach, M.: Erhöhung der Dauerhaftigkeit textiler Betonbewehrungen
5. durch Epoxidharz-Tränkung. Bautechnik 88 (2011) Heft 5.
6. Hegger, J., Kulas, C., Horstmann, M.: Tailor-made 3D-reinforcements for TRC structures. In: Oeh
7. lers, D. J., Griffith, M. C., Seracino, R.: Proceedings of the 9th International Symposium on Fiber-
Reinforced Polymer Reinforcement for Concrete Structures, Sydney, Australien.
8. Brockmann, T.: Mechanical and Fracture Mechanical Properties of Fine Grained Concrete for Tex
9. tile Reinforced Composites, Aachen, Technische Hochschule, Dissertation. 2006.
10. Deutsches Institut für Bautechnik (DIBT): Allgemeine bauaufsichtliche Zulassung: „betoShell“
11. Platten aus Betonwerkstein mit rückseitig einbetonierten Befestigungselementen zur Verwendung
als hinterlüftete Außenwandbekleidung oder als abgehängte Decke (Z-33.1-577). 31. Juli 2008.
12. Hegger, Josef; Kulas, Christian; Horstmann, Michael: Large-sized facade elements made of textile-
13. reinforced concrete; In: Concrete: 21st Century Superhero: Building a sustainable future ; The 11th
Annual International fib Symposium ; 22nd – 24th June 2009, Business Design Centre, London /
CEB-FIP. - London : emap networks, 2009.
14. Voss, S.: Ingenieurmodelle zum Tragverhalten von textilbewehrtem Beton. Dissertation, Schriften
15. reihe des IMB, RWTH Aachen University, Heft 24, (2008), ISBN 3-939051-03-9.
16. Kulas, C.: Ingenieurmodelle zum Tragverhalten getränkter textiler Bewehrungselemente für Beton
17. bauteile. Dissertation, Schriftenreihe des IMB, RWTH Aachen University, Heft 38, (2013), ISBN 3-
939051-17-9.
343
344
MESH MOULD: ROBOTICALLY FABRICATED METAL MESHES

AS CONCRETE FORMWORK AND REINFORCEMENT
Norman Hack1, Willi Lauer2, Fabio Gramazio1, Matthias Kohler1, 1ETH Zurich, Gramazio and
Kohler Research, Chair of Architecture and Digital Fabrication, 2SEC FCL Singapore, Module II,
Design of Robotic Fabricated High Rises
Abstract: The research project Mesh Mould explores the unification of formwork and rein-
forcement into one single, robotically fabricated construction system. An industrial robot is
used to spatially “weave” a three-dimensional mesh, which acts as porous formwork dur-
ing the process of concrete pouring, and is activated as reinforcement after the concrete has
cured. In a first project phase a three-dimensional polymer extrusion process was devel-
oped allowing quickly assessing a wide range of mesh patterns and subsequently establish-
ing an appropriate relationship of mesh morphology and concrete rheology. Whereas in
the first phase, the loadbearing capacity of the polymer meshes was only of minor concern,
the second phase of the project deliberately focuses on the structural performance of the
meshes. Therefor an automated robotic wire bending and welding tool for steel meshes was
prototypically developed, systematically tested and is currently undergoing further cycles
of development. Concurrently a series of structurally differentiated metal meshes was fab-
ricated semi-automatically and tested in destructive load tests. Particularly well perform-
ing mesh typologies subsequently inform the design of the next iteration of the robotic wire
bending and welding manipulator. As such, the research aims to develop a fully automated
in-situ robotic fabrication process for geometrically complex, loadbearing concrete con-
structions.
INTRODUCTION
The Gramazio and Kohler Research group at the Chair of Architecture and Digital Fabrica-
tion at ETH Zurich, has over the past years explored the impact of robotic fabrication on
architectural design and construction. The chair’s research points into the direction of a
digitally informed and robotically materialized future of architecture. A novel, “digital
materiality” [1], displaying a high complexity and sophistication digitally inscribed into the
material, is rendered possible only through the dexterity of the robot. The research Project
Mesh Mould follows that line of inquiry and investigates how the unification of the two
most labour and cost intensive aspects of concrete structures, the formwork and the rein-
forcement [2], into one single, robotically fabricated construction system allows for a cost
and material efficient fabrication of geometrically complex concrete constructions. The
multitude of interacting parameters which are touching upon the diverse disciplines of ar-
chitecture, robotics, material science, computation, mechanical-, and electrical-engineering
has opted to develop the Mesh Mould technology in iterative cycles. Accordingly the re-
port is structured along two consecutive phases. The first phase of the project focused on
the development a generic robotic fabrication process, allowing to quickly produce and test
345
a wide range of mesh morphologies and geometries. Subsequently the milestones of this
first phase of the project were defined as follows:
- Development of an automated robotic fabrication process for fast prototyping and
subsequent evaluation [3].
- Development of generative computational model as a design tool for topologically
differentiated mesh morphologies.
- Calibration of the relationship between mesh aperture size and concrete rheology
[4].
- Development of an appropriate concrete filling strategy.
A customized toolhead for the spatial extrusion of Acrylnitril-Butadien-Styrol (ABS), a
thermoplastic polymer conventionally used for layerbased 3D printing applications, was
developed. Key features of the extruder were an extrusion thickness of up to 3mm diame-
ter, a temperature differential of 270 ° as well as an additional air-cooling mechanism with
a capacity of 10 liter air per second. The latter allowed the instantaneous hardening of the
thermoplastic, and thus enabled to spatially print complex freeform mesh structures (Fig 1-
3). Despite the relative weakness of the ABS filament with a ultimate tensile strength of
only 28 MPa [5], this generic extrusion process was found to be an effective way for short
evaluation cycles of various mesh patterns upon their performance during the process of
concreting. The experiments generated essential knowledge regarding the rheological be-
havior of concrete within the mesh and in defining an appropriate ratio of mesh aperture
size and concrete viscosity,
Fig. 1: Industrial robot on a mobile platform extruding a 1:1 scale prototype

Fig. 2: Close-up of the spatial extrusion process, ABS hardening in mid air
Fig. 3: Double curved mesh, half-filled with concrete
However the targeted loadbearing capacity of the meshes can only be achieved by improv-
ing the yield strength of the mesh material itself. In this current, second phase of the pro-
ject the research is devoted to developing an automated robotic fabrication process that
builds up on the findings of phase one, but under additional consideration of the factual
structural demands. Thus the goals of the second phase are defined as follows:
- Development of a robotic tool head for processing filament with high yield
strength, such as steel, synthetic-, or natural-fibers.
- The fully automated robotic fabrication of several fully loadbearing mesh speci-
mens for structural testing.
- The exploration of the architectural design implication of such construction system.
346
- The automated fabrication of a small architectural demonstration project, display-

ing the design implications and the structural potential of the system.
FIRST RESEARCH PHASE: INITIAL EXPLORATION INTO DIFFERENTIATED POLYMER

MESHES AND EVALUATION OF CONCRETE FILLING PROCESSES
Preliminary concrete filling experiments showed that a self-compaction concrete with a

slump of 19 cm (tested with a mini-cone with an upper diameter of 7cm, lower diameter 10
cm, height of 6 cm) was clogging inside the mesh, preventing an even distribution of mate-
rial within. This observation was attributed to an unfavorable ratio of exterior and interior
mesh density. Whereas the triangulated interior structure was too dense (side lengths a, b =
15 cm, c = 4 cm) for the concrete to evenly distribute within the mesh, the triangulated
exterior was found to be too open (side length a, b, c = 4 cm) to prevent the concrete from
leaking out. Based on these findings two series of functionally differentiated mesh typolo-
gies were developed aiming to assure sufficient concrete flow inside the mesh while pre-
venting excessive concrete leakage through the outer perimeter. The first series of meshes
(b x h x l = 15 x 50 x 60 cm) was designed for frontal concrete filling. The experiments
were conducted in collaboration with Sika Technology AG using a Sika MonoTop 412 N
repair mortar [6] for wet spray applications. The mortar, with an adhesive strength ≥ 2.0
MPa is particularly well suited for vertical and over-head applications. The mesh specimen
displayed larger triangular apertures on the front than on the back. The open front was de-
signed to allow the Shotcrete to enter the mesh whereas the denser back aimed to prevent
the concrete from shooting through (Fig. 4). One half of each mesh was filled by wet-
spraying the mortar with a 12 mm refurbishment nozzle and an air pressure of 4 bar from a
distance of approximately 40 cm (Fig. 6), the other half was filled without use of the noz-
zle, directly pressing the mortar through the mesh perimeter utilizing the mere pressure of
the concrete pump (Fig. 7).
a. b.
Fig. 4: Two Mesh typologies, a and b, for frontal filling, each with larger apertures on the front (left) and
denser apertures on the back (right). Mesh a consists of a triangulated front with a side length of a, b, c
= 4 cm and triangulated back with a, b = 2 cm and c = 4 cm, Mesh b consists of a triangulated front with a
side length of a, b, c = 6 cm and triangulated back with a, b = 3 cm and c = 6 cm.
A second series of meshes (b x h x l = 20 x 80 x 60 cm) was designed for top-wise filling.

Holcim Singapore supported the testing and developed an application-specific concrete
with a low water-cement ratio and low slump flow; vibrating with a poker vibrator was
required after filling (Table 1, Holcim).
347
Table 1: Two Concrete mixtures used for filling ABS and metal meshes
Holcim mix ETHZ IFB12 mix
Water-cement ratio 0.36 0.34
Ordinary portland cement 450 kg/m³ 500 kg/m³
Undensified silica fume - 43.5 kg/m³
Water 160 kg/m³ 169kg/m³
Aggregates of grain size 0-4 mm 713 kg/m³ 705 kg/m³
Aggregates of grain size 4-8 mm - 1008 kg/m³
Aggregates of grain size 10 mm 1020 kg/m³ -
Modified lignosulphonate based retarder 1.58 kg/m3 -
Polycarboxylate ether superplasticizer 3.65 kg/m3 4.32 kg/m3
The mesh design of the second series followed two main strategies: Firstly the introduction
of filling channels within the mesh (Fig. 5 a-c), and secondly a general decrease of density
of the interior structure (Fig. 5 d-f). For both experimental setups the evaluation parameters
were defined as follows:
1. An even distribution of concrete within the mesh without the occurrence of aggre-
gate nests or voids.
2. A protrusion rate of the concrete from inside through the outer perimeters of the
mesh of approximately 1 cm during vibrating.
3. The mesh’s shape stability under the process of filling, and under the wet load of
concrete.
a. b. c. d. e. f.
Fig. 5: Mesh typologies for top wise filling with apertures ranging from a, b, c ≥ 2cm ≤ 4 cm
12
The recipe was developed by Lex Reiter from Institute for building materials (IFB) at ETH Zurich
348
Findings:
When filling the meshes frontally via the mere pressure of the concrete pump, a high pres-
sure led to considerable deformations of the mesh up to 1.5 cm, particularly noticeable at
the corners and edges. Furthermore the concrete did protrude only approximately half way
from one side of the mesh to the other, suggesting that the mesh should rather be filled
from both sides. Moreover the distribution of the viscous concrete paste was not consistent
and showed the occurrence of voids up to a diameter of 1.5 cm around parts of the horizon-
tal interior mesh structure.
The wet-spayed Shotcrete on the other hand consistently protruded through the front of the
mesh without damaging it. In order to prevent the concrete from shooting through howev-
er, the aperture size on the back side should be further reduced to a size of at most 1.5 cm
for all triangle sides. Furthermore, both frontal filling methods imposed unsymmetrical
loading on the mesh during the process of filling. At a larger scale and without additional
scaffolding this is likely to cause deformations of the overall structure during filling. The
concrete paste allowed a good workability during the manual troweling process and ena-
bled a smooth surface finish (Fig. 8, bottom).
Fig. 6: Concrete application with a wet-spray nozzle.

Fig. 7: Concrete application without nozzle, merely using the pump pressure.
Fig. 8: Cured samples, top: untreated back side, bottom: roughly troweled front side.
With regards to the top wise filling, the partially missing interior structure in regions of the
flow channels (Fig. 5a, b, c) noticeably weakened the overall stability of the specimen. The
wet load of concrete caused buckling and led to deformations up to 3 cm in the channel
area. Furthermore, the concrete did not distribute uniformly from the channels throughout
all regions of the mesh, causing voids of up to 2 cm diameter. A general, even reduction of
the horizontal interior structure on the other hand worked well for the filling and vibrating
process.
Moreover, the introduction of a tetrahedral geometry of the outer perimeter (Fig. 5f) signif-
icantly improved the structural performance during the filling and vibrating process. Fur-
thermore the tetrahedral perimeter increased the flow resistance and showed a positive ef-
fect on the protrusion rate of the fresh concrete (Fig. 9-11).
349
Fig 9: Filling and vibrating process of a mesh with tetrahedral perimeter.

Fig 10: Close-up of concrete protruding through the tetrahedral mesh perimeter.
Fig 11: Troweled and cured concrete sample.
Cutting the filled sample with tetrahedral perimeter in half showed a thoroughly even dis-
tribution of concrete without the occurrence of voids or aggregate nests. In conclusion an
evenly distributed ratio of 1/30 (density interior structure to density outer perimeter) was
found to be most suitable for the purpose of filling from atop and thus acted as a point of
reference for the next meshes fabricated from steel wire.
SECOND RESEARCH CYCLE: IMPROVING THE STRUCTURAL CAPACITY

In order to meet the goal of activating the meshes not only as formwork, but at the same
time as structural reinforcement, two possible lines of inquiry were identified. Firstly, the
co-extrusion of a tension active filament within a polymer matrix, as for example with al-
kali-resistant glass fiber, carbon fiber, aramid or basalt [7], and secondly, the substitution
of the polymer filament with thin, continuous steel wire (d ≤ 3mm). Both options display a
unique set of characteristics with specific advantages and disadvantages. Despite being
prone to corrosion, metal wire displays a multitude of advantages.
The compound of steel and concrete, for example, is a well-established material system
with good mechanical bonding and very similar expansion rates. Steel wire is, compared to
the high-tech fiber compound, a fairly cheap and almost universally available material [8].
Its ability to be monolithically joined through discharge welding makes the connections
between the layers simple and straight forward, even for small diameters [9]. Another ma-
jor advantage can be accounted to the fact that steel can be deformed plastically at fast
rates without having to thermally activate it. Although, for the current research phase the
material question was decided in favor of steel wire, fiber reinforced polymers remain a
promising path for follow-up research.
TOOLING AND PROCESS ENGENEERING

Based on both, the insights won throughout the previous ABS extrusion experiments, and
on the decision to focus the coming research phase on processing metal wire, a concept for
a robotic wire bending and welding manipulator was developed and has evolved over vari-
ous iterations. The initial concept for a wire deformation mechanism was developed (Fig.
12a), and was prototypically realized and tested with commercially available pneumatic
actuators (Fig. 12b). The integration of all required functionalities, including electronic
linear position feedback actuators, retractable components for collision avoidance and a
discharge welding compatible wire bending mechanism resulted in a manipulator size of
350
10 x 8 x 22 cm (Fig. 12 c), appropriately small and agile for the realization of filigree ar-
chitectural geometries.
a. b. c.
Fig. 12: Evolution of wire bending-welding manipulator.
The functional principles of the manipulator in its current state are described in greater
detail below:
1. The upper wire is clamped by two outer brackets (Fig. 13 a).
2. The linear actuator extends to the calculated length (depending on the surface cur-
vature) and bends the wire into a “V” shaped configuration (Fig. 13 b).
3. The upper wire touches the lower wire and an electrical impulse is discharged, spot
welding the two wires together.
4. The outer brackets open up again and a part of the clamping mechanism rotates 90
degree inwards in order to avoid collision with the previously bent part (Fig. 13 c).
5. The tool head moves to the next position and the routine repeats.
a. b. c.
Fig. 13: Bending and welding cycle
351
Tooling: Future challenges

During extensive fabrication tests the current bending-welding manipulator has revealed
several weak spots for further improvement. Among the most challenging future develop-
ments are:
1. The incorporation of a wire pre-straightening mechanism.
2. The design of a robust wire alignment mechanism within the welding electrode in
order to guarantee flawless welding connections.
3. Sufficient dimensioning of the components for processing wire up to 3mm diame-
ter.
4. A wire guidance mechanism for the horizontally oriented wires.
5. Further actuation of collision-prone components.
In order to be able to quantify the structural performance of the metal meshes as rein-
forcement, a robust and reliably repeatable fabrication process is of paramount importance.
As the demanded process stability cannot be guaranteed in the tool’s current iteration, the
bending-welding manipulator presently undergoes further development, specifically ad-
dressing the above mentioned issues. Meanwhile a semi-automated fabrication process was
developed, allowing to reliably fabricate various mesh samples for concrete filling and
structural testing. In this intermediate fabrication setup the robot three-dimensionally bends
single mesh layers, which are subsequently assembled and spot welded manually. The
identification of mesh typologies that are, next to being formwork, additionally complying
with the demands of structural reinforcement, directly feeds back into the refinement and
further specification of the tool.
MESHES FOR INITIAL STRUCTURAL LOADING

On the basis of the successful mesh typologies of the first research phase, a series of rec-
tangular metal meshes (b x h x l = 7 x 9 x 30 cm) with different patterns was fabricated
(Fig. 14). The primary aim of this initial structural test was to identify certain particularly
well performing characteristic of the mesh, which in turn would inform the further devel-
opment of the robotic bending-welding manipulator.
Fig.14: Robotically fabricated meshes for structural loading
The main directions followed in the development of the test specimen were:
1. Increased spatiality through tetrahedral perimeter and spatially diagonal interior
structure (Fig. 15, v1, v2).
2. Increasing number of the horizontally running wires (Fig. 15, v3).
352
3. Vertical build-up as an alternative to the horizontal layout (Fig. 15 c, v4).

4. Varying the density of mesh apertures, from the smallest size producible to the
largest aperture size capable of holding the concrete (Fig. 15, v5 and v6).
5. Perimeter layering of the mesh, double layered perimeter (Fig. 15, v7).
v1 v2 v3
v4
v5 v6 v7
Fig.15: Mesh typologies, versions v1 - v7.
With regards to the concrete, the previous mixture of phase one was slightly modified (Ta-
ble 1, ETHZ IFB). The meshes were filled from atop, using formwork only on the two
open ends and at the bottom of the mesh, leaving the mesh faces with the triangulated pat-
tern exposed. Other than in previous trails a vibrating table was used to consolidate the
samples. The concrete coverage was 1 cm on each side, except on the open mesh ends,
resulting in a sample size of b x h x l = 9 x 11 x 30 cm. The probes cured for 14 days in a
curing chamber with a relative humidity of 95 %. A three point bending test, with two sup-
ports in a distance of 21 cm and a centered load was performed with a movement speed of
0.01 mm/sec (Fig. 16). An additional beam (v8) with no reinforcement served as a refer-
ence.
Fig. 16: Displacement as a function of the sample’s loading force.
353
The mesh volume fraction varied among the different mesh typologies (Fig. 17) and was
relatively low compared to a conventional steel reinforced concrete-, or ferrocement-
structures [10]. For the sake of comparability an equal mesh volume fraction was initially
considered, but finally discarded due to the multifunctional requirements of the mesh. The
structural performance is only one parameter among others, as for example, the concrete
distribution inside the mesh, the concrete protrusion rate through the outer perimeter, the
roughness of the surface structure as a basis for troweling, as well as the degree of com-
plexity for an automated fabrication process. In that regard the load test revealed that a
mesh with a double perimeter (Fig. 17, v7) and a higher volume fraction has a lower struc-
tural performance compared to a single layer mesh with additional horizontals (Fig. 17,
v3), however, with regards to the concrete protrusion rate, for example, the layered buildup
of the latter shows significant advantages during the process of concrete filling.
Fig. 17: Mesh volume fraction in relation to maximum load
The force displacement behavior of meshes v1, v3, v7 (Fig. 16) displayed a significant
yield plateau with an increasing load capacity after the first crack appeared. Additionally
mesh v3 showed the appearance of a second crack, indicating a substantial loadbearing
behavior of the mesh after the appearance of the initial crack. This observation is supported
by a very pronounced plateau of that particular mesh typology. In contrast to the inclining
yield plateaus of v1, v3 and v7, the typologies v2, v5 and v6 displayed a noticeable drop of
load bearing capacity after the appearance of the first crack, indicating that the meshes
were unable to compensate loads beyond that force. The force displacement behavior (Fig.
16) in relationship to the mesh volume fraction (Fig.17) suggests that a minimum volume
fraction of 1.4 % is necessary in order for the meshes to act as reinforcement. Whereas this
observation is valid for meshes with a horizontal build-up, it does not comply with the be-
havior of the vertically buildup mesh v4. Despite a mesh volume fraction of 1.3, the spec-
imen displayed a behavior very similar to the reference beam v8 without any reinforce-
ment. This brittle behavior can be attributed to the weak welding connections, combined
with the complete absence of continuous horizontal wires. The diagonal interior structure
of specimen v2 does not contribute to the structural performance under this specific load
case, could however be activated as helical reinforcement in particularly high stressed ele-
ments.
354
CONCLUSION AND NEXT STEPS

The physical experiments of phase one and two have verified the underlying assumption,
that formwork and reinforcement can be effectively combined into one robotically fabri-
cated construction system. The elements of the research, namely the concrete recipe, the
mesh typology and the robotic fabrication process, are strongly interlinked and can only be
looked at in a mutually depending manner. The challenge of the future research will be to
pinpoint the “sweet spot” at the intersection of the robotic mesh fabrication, the perfor-
mance during filling, the concrete workability during troweling, the final surface quality
and the mesh’s structural performance as reinforcement. With regards to the tool develop-
ments some significant conclusion can already be drawn from the initial load tests: The
minimum mesh volume fraction above 1.5 % implies that certain mesh typologies, as for
example v5 and v6, can only be activated as reinforcement by increasing the diameter of
the wire. This in turn would result in a bulkier bending-welding manipulator and a higher
percentage of steel which does not act as reinforcement (e.g. the horizontal interior struc-
ture). Therefore it seems more appropriate to develop the next version of the bending-
welding manipulator with a particular emphasis on the ability to lay additional horizontally
aligned wires. Besides structural considerations the filling experiments from phase one
identified a slightly structured mesh surface (e.g. mesh v1) as beneficial for the concrete
filling-, and subsequent troweling-process. Following up on the immediate phase of tool
re-development a next series of experiments will direct the focus again towards the archi-
tectural implication of the fabrication system. Thus a wider set of geometries will be tested,
including vertically oriented elements and continuous transitions from vertical to horizon-
tal building elements. For a better workability and an increased stability of the concrete,
especially on such sets of geometry, a modification of the structuration rate in the early age
of the concrete will have to be considered [11].
ACKNOWLEDGEMENTS
We would like to express our gratitude for the support to the members of the SEC Future
Cities Laboratory Singapore, where this research had initially started. Furthermore we
would like to thank our colleagues from Gramazio Kohler Research and the National Cen-
tre of Competence in Research (NCCR Digital Fabrication Agreement # 51NF40-141853)
for their valuable support and inspiration. Sincere thanks go to Heinz Richner and Lex
Reiter for their support with regards to concrete mixtures and testing. We owe particular
gratitude to our industry partner Sika Technology AG, without whom this research would
not have been possible. Also we would like to thank Holcim Singapore for supporting the
concrete experiments in Singapore.
REFERENCES
[1] Gramazio, F. and Kohler, M., “Digital Materiality in Architecture”, (Lars Müller Publishers, Baden,
2008).
[2] Lab, R. H., "Think Formwork - Reduce Costs," in “Structure magazine” , April, 2007. 14-16.
[3] Hack, N., Lauer, W., Gramazio, F. and Kohler, M., "Mesh-Mould: Robotically Fabricated Spatial
Meshes as Reinforced Concrete Formwok," in “Fabricate: Negotiating Design & Making”, Gramazio,
F., Kohler, M. and Langenberg, S., (Gta Verlag, Zuerich, 2014) 224-231.
[4] Hack, N., Lauer, W., Gramazio, F. and Kohler, M., "Mesh-Mould: Differentiation for Enhanced
Performance," in “CAADRIA: Rethinking Comprehensive Design”, (Kyoto Institute of Technology,
Kyoto, 2014) 140 – 148.
355
[5] Tymrak, B., Kreiger, M., and Pearce, J., "Mechanical Properties of Components Fabricated with Open-
Source 3-D Printers under Realistic Environmental Conditions," in “Materials and Design”, 2014, 242-
246.
[6] Sika Technology AG, "MonoTop 412 N," (Online), Available:
https://www.google.ch/webhp?sourceid=chrome-instant&ion=1&espv=2&ie=UTF-
8#q=monotop%20412%20n%20sika. (Accessed 22 04 15).
[7] Gries, T., Roye, A., Offermann, P. and Peled, A., “Textiles”, in "Textile Reinforced Concrete: State-of-
the-Art Report", Brameshuber, W., (RILEM, Bagneux, 2006) 11-27.
[8] Naaman, A. E., "Composites in Construction," in “Thin Cement Composites: Performance Comparison
between Steel and Textile Reinforcements”, Hamelin, P., Bigaud, E. and Jacquelin, E., (University of
Lyon , Lyon, 2005) 1155-1164.
[9] Petrović, Ž., Džiho, E. and Pašić, S., "Capacity Discharge Welding of Small Diameter Wires," in
”Trends in the Development of Machinery and Associated Technology”, 17th International
Research/Expert Conference, Istanbul, 2013.
[10] American Concrete Institute, "State-of-the-Art Report on Ferrocement," (American Concrete Institute,
Detroit, 1994) 17.
[11] Roussel, N., "Thixotropy: From Measurement to Casting Concrete," in “Understanding the Rheology of
Concrete”, Roussel, N., (Woodhead Publishing Limited, Cambridge, 2012) 286-295.
356
RESTORATION OF THE HISTORICAL CONCRETE-GLASS

WINDOWS OF THE TOWN-HALL IN AACHEN
Sergej Rempel, Stephan Geßner, RWTH Aachen University, Institute of Structural Concrete, Mies-
van-der-Rohe Str. 1, 52074 Aachen, Germany
Abstract: Textile-reinforced concrete (TRC), an innovative composite material, creates

new opportunities for the refurbishment of historic buildings. For example, filigree
elements, such as the historic concrete-glass windows of the Aachen City Hall, benefit
from the corrosion-resistant carbon textile reinforcement. The 60 year old windows had to
be repaired because of the corroded steel reinforcement and concrete spalling. Therefore,
the glass stones embedded in the concrete were separated and reinstalled in reconstructed
concrete-glass windows. The steel reinforcement was replaced by carbon reinforcement.
Thus, the historic monuments protection authorities’ terms were respected while avoiding
future corrosion damage.
INTRODUCTION
The gothic Aachen City Hall is adjacent to the Aachen Cathedral, the most distinctive
building in the imperial city of Aachen. It is one of the most important medieval state
buildings in Germany. Erected on the foundations of the King's Hall of Charlemagne from
the 8th century, this palace was built for the coronation ceremonies of German kings. The
palace has been adapted and restored gradually to comply with contemporary requirements
or demands. In 2014, the City of Aachen celebrated the 1200th anniversary of Charle-
magne’s death, for which parts of the old City Hall were refurbished.
The focus of the renovation work was on the Marien tower (fig. 1, left), the main apse of
the former Kingdom Hall. Particularly striking features of this structure are the spire and
the concrete-glass windows, which should reflect the appearance of the quarry stone walls
(fig. 1, right).
As part of the reconstruction of the tower in the 1960s due to its destruction in Second
World War, 19 unique concrete-glass windows were designed and installed by the archi-
tects G. Graubner and U. Fuß.
The concrete-glass windows of Aachen City Hall vary in appearance and size, and consist
of glass stones embedded in a 3,5 cm thick concrete layer. The distance between the rows
of glass stones is about 2,5 cm, reinforced with 5-mm diameter horizontal steel bars. The
windows are surrounded by a steel frame, which is required to anchor the windows in the
wall and served as lost formwork. In the production of the old concrete-glass windows, the
glass stones were positioned on a formwork panel within the steel frame and then concret-
ed. Due to the small concrete cover (around 1,5 cm), concrete spalling occurred and the
corroded reinforcements surfaced. Consequently, the refurbishment of the historic win-
357
dows was essential. The regulatory constraints of the historic monuments protection au-
thority for the refurbishment of the concrete-glass windows were to maintain the appear-
ance of the original windows. Identical glass stones were to be used, reusing those already
used in the old windows. In addition, the position of the glass stones in the windows was
required to remain the same. As a consequence, the thickness of the concrete layer (3,5 cm)
could not be changed. Thus, the thin concrete cover must be maintained. This excludes a
reapplication of the steel reinforcement, which would not be sustainable. The corrosion-
resistant carbon reinforcement used in the textile-reinforced concrete enables the pre-
scribed concrete thickness of 3,5 cm.
Figure 1: View of Aachen town-hall with Marien tower on the right side
WINDOW-SETUP AND MATERIALS

Fig. 2 shows the setup-up of a concrete-glass window. The load transfer in the windows is
uniaxial, parallel to the glass stone rows and concrete rows. At the lateral ends, the loads
are transferred via the perforated plate into the frame. By connecting the frame to the ma-
sonry, the loads are transferred into the masonry.
358
Figure 2: Sectional view of a concrete-glass window
Due to the thin concrete cover in the old concrete-glass windows, the steel reinforcement
corroded and concrete spalling occurred. These damages were to be avoided in the new
windows. The required size of the concrete cover could be realized by using textile rein-
forcement.
As textile reinforcement for the concrete-glass windows of Aachen City Hall, a scrim of
carbon filaments was used. The cross-sectional area of the textile reinforcement mesh is
110 mm²/m, with 4,18 mm² per roving. To increase the ultimate stress of the roving, the
fabric was impregnated with epoxy. By doing so, tensile stress of more than 3.000 N/mm²
can be achieved, corresponding to a tensile force of 12,5 kN per roving. In addition, the
impregnation improves the handling and the workability of the reinforcement. The 90-
degree direction (fill direction) is the main bearing direction of the window. Since the glass
stones enable a continuous installation of the textile reinforcement mesh in only one bear-
ing direction, the mesh was fragmented and the rovings were installed between the glass
stones in the concrete rows. The 0-degree direction (warp direction) is not required for the
load transfer in the concrete-glass windows.
Concrete
The composition of the concrete used for the original windows is not known, and could not
be reproduced. For this reason, a fine-grained composition commonly used for textile-
reinforced concrete with a maximum grain diameter of 5 mm was utilized. The concrete
has a compressive strength of 87 N/mm² and a bending tensile strength of 10,6 N/mm².
The high material strength is particularly suitable for components that have to remain
uncracked in use.
359
Fastening elements
The historical windows were surrounded by steel frames. The frames were used to fix the
fastening elements and ensured that the loads were directed into the masonry. The required
connection of the concrete to the steel frame was secured by hooks. The existing mecha-
nism for the force transmission into the masonry should be maintained. Only the material
of the frame has been replaced by stainless steel to prevent corrosion. In addition, the
hooks were replaced by a perforated plate that was welded to the frame. Thus, a steady
load distribution should be achieved in the frame. The refurbished windows were attached
to the masonry via welded screw sleeves. Therefore, stainless steel screws were used
screwed halfway into the sleeves. The other half of the screws was inserted into a modified
bearing rail, which was anchored in the masonry. After the installation, the bearing rail and
all interstices were filled with color-adjusted mortar. To enable strains, the windows were
installed on an expansion strip.
MANUFACTURING
The manufacturing of the new concrete-glass windows was carried out in accordance to the
production process of the old windows 60 years ago. In the first step, the historical win-
dows were documented in detail, as the appearance of the windows was to remain un-
changed. Therefore, the windows were surveyed and photographed; the glass stones were
labelled. In the next step, the historical windows were carefully disassembled. In this work,
the damage of the historical window became obvious. After the steel frame was disman-
tled, the concrete crumbled after simply tapping on it. This had advantages in that the ma-
jority of the glass stones remained intact and could be reused. This step revealed that the
damage was severe and renovation was urgently needed. Simultaneously, custom-fit an-
choring rails made of stainless steel were manufactured. These replaced the steel frame,
but without completely surrounding the windows, as the steel frame had done. Instead, the
anchoring rails were installed only on the bearing side of the window.
The possibility to realize sharp-edged concrete components and the modified load transfer
enabled the omission of a surrounding frame. Thus, the 2-meter high windows could be
divided into portable segments, and then be mounted standing one above the other on-site
to be installed without lifting equipment. This procedure had advantages in both the manu-
facturing as well as in the assembly. The sharp edges make the stacked segments look like
a monolithic created window. The bigger spacing between some glass rows were already in
existence in the old concrete-glass windows.
360
Figure 3: Concrete-glass window under construction with about 400 glass stones
In the next production step, the glass stones were glued to the formwork with acrylic (up to
400 glass stones per window segment). Here, each glass stone had a specified position,
corresponding to its position in the old window. The conglutination guaranteed the position
of the glass stones during the compaction work, and ensured that no concrete covered the
glass stones. The position of each glass stone was checked with a true to scale plan of the
old windows on a transparency. Before concreting, the carbon reinforcement was built in
and fixed on the perforated plates to ensure central positioning during concreting. No spac-
ers were used. The placing of the concrete was done in the casting process; a table shaker
was used for compacting. The stripping of the formwork had to be done very carefully, as
the acrylic’s adhesion was very strong. Finally, the windows were cleansed of the acrylic
and the concrete surfaces were acidified.
Finally, the window segments were reinstalled in Marien tower. For this purpose, bearing
rails were installed in the masonry, into which the concrete-glass windows were inserted.
In the final step, the bearing rails were plastered with a mortar of a similar color.
361
Figure 4:View of the inside of Marien tower (left) and installed concrete-glass windows seen from outside
(right)
CONCLUSION
The refurbishment of the historic concrete-glass windows of Aachen City Hall was an ex-
traordinary challenge. The use of innovative carbon reinforcement contributed significantly
to the solution of the problem. With the use of epoxy-impregnated carbon textiles and high
performance concrete, a corrosion-resistant solution was found, which also impressed the
critical monument conservators.
ACKNOWLEDGEMENT
The authors thank the DFG and the BMWi for the financial support during SFB 532 period
and the transfer project T08. A special gratitude goes to Hering Bau GmbH & Co. KG and
solidian GmbH for their commitment to support the development of the new composite
material TRC.
362
FERROCEMENT SOLUTIONS FOR AFFORDABLE HOUSING PRO-

TOTYPES IN INDIA
Prof. Dr. Anupama Kundoo, Department of Architecture, Universidad Camilo José Cela, Madrid,
Spain; Prof. Dr. Mike Schlaich, Dr. Arndt Goldack, Chair of Conceptual and Structural Design, TU
Berlin, Germany, Anupama Kundoo Architects; schlaich bergermann und partner
Abstract: In achieving ‘Housing for All’ in the rapidly urbanising Indian context, ferroce-
ment technology promises to be environmentally as well as economically ‘affordable’,
consuming significantly less quantities of building materials with high embodied energy.
Though already used to some degree, ferrocement would make its most significant contri-
bution if used for roofs spanning the whole structure, or speedy wall-systems with built-in
furniture. Form development plays a big part in the design of ferrocement roofs. Shell
structures though efficient, mean complex, expensive formwork, whereas folded-plate
structural systems consist of planar surfaces and often easier to construct. A range of ge-
ometries are already tested in origami (the ancient art of paper-folding) to stiffen thin
sheets of paper, and origami crease patterns serve as inspiration for self-supporting roofs.
The paper discusses prototypes designed by the author, for quick and cheap delivery of
houses, with alternatives for small and larger roofing spans for residences and community
buildings respectively. The prefabrication process, undertaken in backyards of masons’
houses improves social and economic sustainability performance, while the technical de-
sign of units improves environmental performance. Developed through Indo-German col-
laboration, test prototypes were built in Delhi, Chennai, Auroville India; Brisbane and
Sydney in Australia; as also in Bielefeld, Berlin and Steinhagen near Bielefeld in Germa-
ny.
INTRODUCTION
Given the nature of India’s population and unprecedented current rate of urbanization, the
need for environmentally sustainable buildings cannot be overstated, and affordability is a
major concern. While the per capita resource consumption is lower than many countries,
India’s population creates demands that the nation is struggling to meet. Buildings in India
need to be produced with significantly less than globally accepted standards, particularly
for the bulk of the construction needs (housing and small public buildings).
In India, steel and cement are not only considered high-energy materials, but they are also
costly, need expensive infrastructure to produce and transport and are in short supply. Use
of structural ingenuity and engineering is most required therefore to produce lightweight
efficient structures that use significantly less quantities of materials, apart from judiciously
selecting materials based on their low environmental impact. Moreover, increasing the
labour component, paying by including participation of ‘unskilled’ labour will create so-
cio-economic benefits to the local area through a holistic and contextual approach to sus-
tainability.
363
Ferrocement technology promises to be environmentally as well as economically ‘afforda-

ble’, consuming significantly less materials. Though already in use to some degree, the full
potential of ferrocement as a relevant and appropriate material in the housing sector is yet
to unfold. Significantly lighter than reinforced concrete, ferrocement contributes best if
used as the central material for construction of roofs spanning across the whole structure as
also if used to construct wall-systems with built-in furniture, leading to great economy of
time and cost. Prefabricated or cast insitu varies according to the advantages in each for
context. This paper discusses the design of ferrocement building systems developed by the
author in response to the growing concern of affordability of housing in India in economic
as well as environmental terms. These were developed over the last 3 years through exten-
sive research and full-scale prototyping in Delhi, Chennai and Auroville within India;
Brisbane and Sydney in Australia and also in Steinhagen near Bielefeld, Germany. These
investigations were deliberately carried out in contrasting contexts, representing develop-
ing and developed countries; the contexts being either labour-intensive or material-
intensive; and familiar with fine handmade construction, or construction with sophisticated
or hi-tech tools.
FERROCEMENT’S PARTICULAR APPROPRIATENESS TO AFFORDABEL HOUSING

SOLUTIONS IN DEVELOPING COUNTRIES
Given its thinness as compared to reinforced cement concrete, and its resulting lightness,
there is a significant reduction in the consumption of commonly available materials like
sand; cement and wire mesh leading to reduction of costs related to materials used. The
labour component, which is always a variable in the cost of construction, being compara-
tively cheaper in developing countries, however makes this technology particularly suita-
ble and attractive for applications in the context of developing countries. Further, skills for
ferrocement fabrication can be quite easily acquired. Ferrocement’s mass is much less than
RCC, and its dense reinforcement makes it self- supporting even in the beginning of the
curing period, making the construction of formwork unnecessary. This is an inherent ad-
vantage in the efficiency of time and material. A module, which can be prefabricated, will
lend itself to mass-production.
A low-cost, low-tech technology, that has low-maintenance, high durability, high strength
when properly shaped, water-resistance, and has the ability to take on almost any form and
is easily repaired, ferrocement is versatile in many parts of the world. The American Na-
tional Academy of Sciences report of 1973 recommends applications in developing coun-
tries, including low-cost roofing and disaster relief, and recommended an international
committee to coordinate its research and development. Apart from identifying its particular
appropriateness to developing countries, it also identified the need for further research and
experimentation in not only design but also in production techniques that are contextually
appropriate. “Its use, particularly in developing countries, must be preceded by more re-
search and experimentation in design and production techniques suited to construction by
unskilled labor.” (Roumaldi 1973)
TWO PARALLEL AREAS OF RESEARCH AND EXPERIMENTATION

The design and construction of affordable housing prototypes in this paper are classified
into 2 series according to the differing nature of the problems and solutions within the
364
housing sector. Housing units themselves involve fairly small spans and room sizes, while
needing a complex understanding of ergonomic requirements as very diverse activities and
related furniture need to be accommodated within a small space, with a great flexibility of
uses across the day as well as across the different needs of different users. On the other
hand community buildings related to housing require more complex solutions for spanning
larger room sizes, whereas the multipurpose activity spaces contained within these are rela-
tively simple as there is a larger available space and more uniform use within them. The
first series of design research and development for large span community buildings within
housing projects, particularly their roofing system is called ‘Light Matters’, while the se-
cond series focusing on prefabricated elements assembled to form the house itself is called
‘Full Fill Elemental Homes’. Together they contain the basic vocabulary and elements re-
quired in the speedy and affordable construction of housing as a whole, as an integrally
sustainable approach, that is generally appropriate to developing countries but particularly
relevant to the India context.
‘LIGHT MATTERS’: USING ORIGAMI CREASE PATTERNS FOR FORM DEVELOPMENT OF

FERROCEMENT ROOFS
In order for ferrocement to be able to span large roofs, its folded geometric form is a key
aspect of the design. The design process involves the exploration of folding and bending
patterns inspired by origami to lend strength to thin sheets of material, helping it to achieve
rigidity and enabling it to support itself.
Bending and folding, the basis of structural stability and form development
Thinness is an essential characteristic of ferrocement and can make this material unstable
without introducing folding or bending into its form. Abercrombie aptly describes the ma-
terial’s relationship to its architectural form thus: “Ferrocement’s thinness makes it struc-
turally unstable without bending, but it is also its thinness, combined with the great pliabil-
ity of its supporting wire mesh that makes such bending very easy” (Abercrombie 1977).
Form development thus plays a big part in the design and structural stability of ferrocement
roofs. Though efficient, shell structures mean complex, expensive formwork; whereas
folded-plate structural systems consist of planar surfaces and often easier to construct, and
also easier to describe geometrically. The principle of folding to strengthen thin sheets has
been known for a long time. While aerospace and the automotive industry use this princi-
ple to create self-supporting wall, slab elements with high load capacity etc; in contrast, in
the building industry, the principle of folding is less prevalent as perhaps these forms are
less suitable for habitat use despite their structural efficiency and more extensively used for
large span structures such as auditorium, stadia and industrial structures (KUNDOO,
RANGARAO 2013). Due to its properties, ferrocement lends itself to curving or undulat-
ing forms and therefore has great potential to be explored architecturally.
The structural resistance of folded plates relies on their edges. These structures are made of
plates jointed in certain angles by shear resisting connections. The edges of the folded
plates are line-like supports when each plate acts as a slab transferring the load by bending
(Fig. 1). All the loads prefer to be supported by the stiff short span and not by the long
span of plate. The second main aspect of the structural behavior is that these reaction forc-
es at the line like supports are in equilibrium with membrane forces such as normal and
365
tangential forces acting in the plane of the plates. Especially longitudinal prismatic fold-
ings spanning in one direction as in (Fig. 1) behave as girders with a thin walled but deep
cross section made of flat panels. They provide high stiffness and low deflections by very
effective, material saving und economic cross sections.
Figure. 1: Load bearing of folded plates by 1) bending of a slab, 2) membrane forces in the plate and 3) frame
or truss action. (Source: Büttner, O.; Hampe, E. (1984): Bauwerk, Tragwerk, Tragstruktur. Band 2, VEB
Verlag für Bauwesen, Berlin)
THE RELEVANCE OF CREASE PATTERNS FOUND IN ORIGAMI

Origami, the Japanese ancient art of paper-folding transforms flat sheets of paper into
three-dimensional forms through multiple folding, and could serve as inspiration for ex-
ploring forms for self-supporting ferrocement roofs: firstly to take advantage of the range
of possibilities already developed that has continued to evolve and be applied in current
times by mathematicians, scientists and artists, secondly to find a way of building synthetic
surfaces that integrate roofs and walls; thirdly to translate curved shells into facetted sur-
faces that can be easier to describe geometrically and easier to build including formwork;
fourthly to arrive at forms that contribute to minimum wastage. Origami forms are devel-
oped out of a sheet, as also chicken mesh or panels used in formwork, remaining continu-
ous surfaces even though the polygons join other polygons in a different plane (KUNDOO,
RANGARAO 2013). Origami can be a very befitting entry point, an intuitive and sponta-
neous approach to explorations of suitable forms for ferrocement surfaces integrating
roofs and walls, where the ‘art of folding’ itself unfolds the structural behaviour.
Selected origami crease pattern and chosen form for prototyping as roof surface
Paul Jackson’s book, ‘From sheet to Form’ formed the basis of the quest for appropriate
form to be taken up for development as a synthetic roof surface prototype. After construct-
ing all relevant forms in paper, observing their stability and investigating their suitability to
habitation related requirements, the form selected was as shown below in (Fig. 2 to 8). The
flattened paper consisted of vertical creases in one direction and diagonal creases in two
directions such that the intersections of the thus formed rhombus shapes intersected with
the edges of the rectangular sheet of paper. The form was ideal due to its regular
geometry and stability but also as the geometry of the open form contained 2 parallel
edges in the front elevation which would be ideal to locate a door, while when completely
folded it formed a regular hexagon. The form was complete and regular when flattened
before creasing, when completely folded, as well as when positioned as a shell.
366
Figure 2, 3, 4: Initial form development in paper in a workshop at The National Association of Students of
Architecture, Sonepat, India, 2013. (Source: Anupama Kundoo)
Figure 5, 6, 7, 8: Selected form has the special characteristic of forming a regular hexagon when deployed.
Brisbane 2012. (Source: Anupama Kundoo)
Force mechanisms in the proposed structure

These folded plate structures are loaded by self weight and wind loads, in some areas snow
and earthquake are also an issue. In many cases the self weight governs the design of this
folded plates stressed by membrane forces acting in the plane of the plates. The local bend-
ing in the triangular plates is of minor importance.
Formwork. And a version without formwork

The fact that this form can be deployed to obtain a smaller more compacted object than the
unfolded surface opened up a whole new area of research into deployable formwork that is
made of recycled cartons (Fig. 9, 10, 11). This meant that form work which is often a hur-
dle being time consuming but also expensive, could now be very easy and affordable. Giv-
en that the form was stiff for paper, the enlarged full-scale form was likely to be stiff
enough if produced in recycled carton. As the form work and steel mesh can be collapsed
and transported, it may be ideal also for disaster relief and speedy constructions. This form
has the promise of enabling many further areas of economy in time and cost.
367
Figure 9, 10, 11: Form work tested in Brisbane, Australia and Auroville, India 2013. (Source: Anupama Kun-
doo )
First prototype tested in Auroville, Tamil Nadu
The first prototype, later left behind as childrens’ pavilion, was constructed in a residential
community called Citadines, in Auroville’s city centre, first in reused corrugated carton.
The dimensions and sizing were based on habitation requirements as well actual materials
to avoid wastage. A range of alternatives were tried for easy fabrication of steel mesh as
well as workmanship issues in plastering the surface from two sides. Finally due to as yet
unresolved issues at that stage, the first prototype was made without formwork, breaking
down the identical rhomboid elements into separate frames and erecting them in situ to be
plastered from both sides. The basic module was a rhombus folded across the longer diag-
onal. Modules were made in the steel workshop using 8mm dia bars of tor steel with two
layers of GI (Galvanized Iron) chicken mesh (22 standard wire gauge, 0.8 mm dia) on ei-
ther side (Fig.12). Frame panels were arranged in position easily as the side elements were
absolutely vertical and the 2 further sloping panels automatically met at the centre where
they leaned towards each other. Wooden logs supported these in position during welding.
Masons applied cement plaster 1:2, on either sides of the self supporting mesh surface
without formwork and cured for 10 days. This proved to be fairly quick and easy and was
completed in 4 days (Fig. 13).
Second prototype tested in Sonepat near Delhi, India

A second prototype was built in Sonepat near Delhi at Gateway College campus that host-
ed the student convention, NASA, and built at the entrance area of the architecture building
(Fig.14). The key improvement in the second prototype was to improve the quality of plas-
tering in terms of achieving a thinner section not exceeding 25 mm. The surface was also
raised on 450 mm high walls to achieve a higher space.
Figure 12, 13, 14: First and second prototypes at Citadines, Auroville and at Gateway College campus,
Sonepat near Delhi, India. (Source: Anupama Kundoo)
368
Further structures constructed for exhibitions in Australia and Germany

A 1:1 scale model of the above was further developed at UQ in Brisbane after solving as a
residential unit. Under the exhibition ‘CUSP: Designing into the next decade’ this research
was showcased as part of 12 designers whose works were considered relevant for the fu-
ture, and toured 7 Australian cities. In the meantime, a further full-scale structure ‘Light
House’ was exhibited as part of ‘The new modesty: an architecture of restraint’ at the Kun-
stverein Bielefeld and DAZ Berlin. showcasing the work of 5 international architects (Fig.
15, 16). Constructed with styrofoam of the same thickness it represented spatial authentici-
ty but was accompanied by a sample of the actual ferrocement. After dismantling the struc-
ture was taken to Steinhagen (Fig.17) at the request of a public school that wished to ex-
periment further by using the styrofoam as a formwork to cast a ferrocement structure and
serve as a pavilion for the school students. The research was conducted in different coun-
tries representing the two extreme ends of current trends in habitat and urban development
globally. To produce prototypes simultaneously in a developed country context and anoth-
er developing country context addresses its relevance in each context given the huge dif-
ferences in socio-economic scenarios as well as in population densities and natural re-
sources.
Figure 15, 16, 17: Exhibitions of ongoing investigation at Sydney, Australia, Bielefeld and Berlin, Germany
2013, Prototype at School in Steinhagen. (Source: Anupama Kundoo)
FULL FILL ELEMENTAL HOMES

‘Full Fill Elemental Homes’ are envisioned as speedy and affordable solutions to con-
structing housing and toilets that have low environmental impact, using an appropriate
combination of sophisticated and low-tech. Consisting of specially designed modules of
prefabricated ferrocement hollow block units, it can be assembled on the site in less than 6
days including foundation.
Figure 18, 19, 20, 21: Ferrocement production in Masons houses, prototype ferrocement elements for hous-
ing and finished housing prototype. (Source: Anupama Kundoo)
369
Since folding is involved in giving strength to thin ferrocement elements, the concept is
based on ergonomically designing folds that form open box units that can be utilised to
fulfil and accommodate a range of ancillary storage needs that are fundamental to the effi-
cient performance of small dwelling spaces. The voids created inside the blocks are de-
signed to efficiently accommodate all storage needs of the user, from clothes to books to
kitchen utensils, even the kitchen sink itself, and other belongings so that all furniture be-
comes redundant. The void of the spaces can remain empty of furniture or sanitary fittings,
and therefore achieve more space while saving the additional cost and time involved in
furnishing homes. Small spaces are often burdened by the way furniture occupies them but
these units make small spaces smart using the thick container walls to their full capacity, to
be filled by the residents. The 25 mm thick ferrocement elements involve chicken mesh,
welded mesh and small diameter steel reinforcement, significantly reducing the quantities
of high-embodied-energy materials used in the construction. Blocks are produced in the
backyards of masons homes to provide them with additional income, rather than in facto-
ries, thereby reducing costs while helping local economy. A range of window, doors, roof
elements and other necessary building components are all produced in ferrocement exclud-
ing any complex hardware in their design. The voids in the walls that can be colonised by
the users, are emphasised, rather than the walls themselves, through the use of the happy
colours. Roofs for the housing and toilet units consist of moulded prefabricated interlock-
ing U shaped profiles according to the size and weight restrictions involved in easy trans-
portation and assembly.
The prefabrication process, undertaken in backyards of masons’ houses improves social

and economic sustainability performance, while the technical design of units improve envi-
ronmental performance. Affordability is due to efficiency and inclusivity. Rather than pro-
ducing the elements in a factory, the elements are produced in a decentralised arrangement
in the home backyards of masons who are assisted by their helpers. This idea has already
found enthusiasm in the villages around Auroville, South India where local masons have
ferrocement experience. Masons can conduct training for other masons in the villages who
want to participate. The pieces are checked for quality through a local engineer, and direct-
ly delivered to the site for assembly creating further economy. The design is suitable for
all, not only the poor. The structures can be used as immediate shelters necessary in farm
house plots in remote areas, disaster relief homes, youth hostels, student housing, as well
as guest houses in environmentally sensitive locations. The houses can be dismantled
equally simply in a day.
India has a huge housing deficit and an acute shortage of toilets not only in urban areas.
Most available solutions take too long to construct, have cost-escalations by the time they
are delivered, and are not affordable to the target group. India has a huge demand for pre-
fabricated solutions to housing, toilets and other buildings that are affordable not only in
money terms but in terms of all other natural resources. There is a demand for decentral-
ised solutions that are environmentally low impact and for an appropriate balance between
high-tech and low tech approaches given that many areas have no access to stable electrici-
ty, and given the level of skills and knowledge of a large section of the population. Full Fill
elements are designed as a quick response to the above demand. Given that the units are
prefabricated and involve an assembly in less than 6 days in most cases, there will not be
any cost escalation reassuring the target group who are investing their precious funds to
meet their housing and sanitation needs. As the elements are built using materials and
skills that are universally available, the same system designed and investigated for Tamil
370
Nadu state is replicable across the country, regardless of urban or rural areas, or climatic
zones as these can be easily produced in both areas and continue to be advantageous eve-
rywhere.
Since the proposed construction system will be significantly cheaper and quicker to con-
struct, the main impact will be a quick reduction in the urgent demand for housing and
sanitation solutions across the rapidly developing country. This reduction will not come
with a negative environmental impact but rather provide solutions and set new consump-
tion standards for building construction in the future. The system requires no initial in-
vestments, no factory space for production and no storage space, the economic benefits are
significant. Geometry and engineering know-how delivers the maximum built space for the
least use of materials as the form of the product contributes to its strength. The material
reduction involved in using ferrocement technology makes it extremely reasonable com-
pared to other prevalent building technologies. Steel reinforcement bars are replaced by
steel mesh; 25 mm thick wall elements are used instead of 125; and the economic benefit is
proportional.
Socially there are many benefits too. The skill can be easily acquired. Masons and helpers
can produce these elements in the various backyards of their houses rather than in a facto-
ry. They can work according to their need and in flexible hours or weekends there by hav-
ing an additional income source for days when they are not occupied on a daily wage la-
bour. They are part of a larger craft-network managed by a local sustainable construction
company, and their pieces are quality controlled by visiting engineers and supervisors.
Since steel mesh replaces reinforcement bars, all materials can be easily sourced locally
and can be bought in small quantities according to the need. Expert masons can also
choose to work as trainers for others desiring to produce ferrocement elements at their
home yards.
Environmentally, the system offers a low impact solution as it uses far less resources than
regular standard construction by replacing steel bars with steel mesh, but also by using
very little material at all given the thin construction elements. As these are complete ele-
ments there is no need for any additional finishes like plaster and paint. The units come
with storage space built in, so this makes the budget for furniture redundant. The units are
durable, but can also be dismantled in a day if necessary and the site can be left as before
with no negative impact. The dismantled units can be reassembled elsewhere. Transporta-
tion energy and use of electricity are minimal, cow-carts suffice where these are available.
Production is decentralised and develops and uses local skills.An integral development
solution, such an enterprise generates local economy, spending a small percentage of the
budget on materials and a large part of it on the labour, thereby benefitting livelihood
while providing quick and affordable solutions to toilets, houses and other spaces. As the
elements are individually hand-made and not standardised factory mass-produced ones,
there is a great advantage in that tailor made solutions or colour options do not come at an
extra cost.
FURTHER WORK
Abercrombie wrote in 1977: “Compared to other building techniques, ferrocement has re-
mained a puzzling freak. Its design criteria are based mostly on experience, not on scien-
tific experiment, and without accepted data to explain its unusual properties, those proper-
371
ties have gone regrettably underutilized.” Through a combination of theoretical research

and continued hands-on full-scale experimentation on site, further research and investiga-
tion of the above proposals will continue in Madrid and Berlin, with the assistance of engi-
neers from TU Berlin, and the team of craftsmen in India supervised by a local engineer
and architect. These prototypes which have been preliminarily executed in India are being
further developed for introduction to the main-stream after thorough testing with European
standards. The next step is design optimisation, to streamline the product and arrive at fur-
ther economy and efficiency before introducing this as a versatile housing solution in the
rapidly urbanising world.
CONCLUSIONS
The above explorations use Ferrocement as the central material in construction unlike cur-
rent uses as secondary elements. The wall and roof systems allow quick construction and
examine shelter or the building envelope as a synthetic solution of a single surface, without
categorizing this as wall or roof, leading to further economy and efficiency, particularly in
the case of larger spans. These proposals are holistic aesthetic solutions to affordable con-
struction not only in money terms but also in environmental terms. A low-cost, low-tech
technology, that has low-maintenance, high durability, high strength when properly
shaped, water-resistance, the ability to take on almost any form and is easily repaired, fer-
rocement can prove to be a versatile material to address affordable housing and could be
relevant beyond India, in many parts of the world.
ACKNOWLEDGEMENTS
Structural design and onsite supervision in India: Dr. Venkata Rangarao, Hyderabad, In-
dia,Greg Killen
Research assistants: Alba Balmaseda, Sonali Phadnis, Yashoda Joshi.
Local Construction Contractor: Sekar Sokkalingam, Aryan Constructions.
REFERENCES
[1] Abercrombie, S., ‘Ferrocement’. ( Schocken Books, New York,1977).
[2] Jackson, P., ‘Folding Techniques for Designers From Shhet to Form’, (Laurence King Publishing, Lon
don, 2011).
[3] Kundoo A. and Rangarao V., ‘Form Development and Ferrocement: A Synthesis of Art and Science’,
in ‘Go Green with Ferrocement’, 2nd National Convention of Ferrocement, Pune, July, 2013 (Ferroce-
ment Society, Pune) 84-90.
[4] Kundoo, A., ‘ New Building Approaches rather than New Building Materials’, JA Journal of
Architecture, Volume 1, Number 02, June 2012. Issue Focus: The sustainability in Architecture and
infrastructure development, Chandigarh College of Architecture, Chandigarh.
[5] Kundoo A., The Plasticity of Ferrocement: It’s Potential for Architectural Application and Influence on
Architectural Form’, 10th International Symposium on Ferrocement and Thin Reinforced Cement Com-
posites, La Habana, October 2012 (International Ferrocement Society) 451-456.
[6] Romualdi, J. P., et. ‘ Ferrocement: Applications in Developing Countries’, (National Academy of
Sciences, Washington,1973).
372
THE MORPHOGENETIC DESIGN AND INFLUENTIAL

DEMIURGIC QUALITIES OF A HYBRID FERROCEMENT
BUILDING SYSTEM
Owen Waldschlagel, Intact Structures Ferrocement Building Systems, New York USA; Stephen A.
Bohlen, A.D.G. Engineering P.C., Newark New Jersey; Joseph V. Lieber P.E., Aamer Islam P.E.
Abstract: Intact Structures building components and methods of construction are capable
of creating and producing municipal and commercial buildings as well as apartment com-
plexes and private homes. The fabrication of the components and the methods of con-
structing the buildings have unique inherent qualities. These are that they can be fabricat-
ed with hand tools at a remote site, and then constructed utilizing the built in point to point
methods to erect the building with “willing” unskilled labor. The macro of this system is
that the completed armatures of the municipal, commercial and residential buildings can be
fabricated in mass production at a factory to specific designs and be delivered as prefabri-
cated village or a commercial complex. This type of construction would arrive at the build-
ing site at a much lower cost to volume equation than conventional block or post and beam
building systems. The following paper will contain a detailed documentation of the Intact
Structures ferrocement building design as constructed in Oaxaca, Mexico. It will include a
review of the development process via photos, mechanical drawings and illustrations de-
picting the multiple designs and capabilities of the buildings armatures that are currently
constructed at the Intact Structure training facility. The architectonic and engineering spe-
cifics of the Intact Structure’s ferrocement building system will be addressed and reviewed
as well as their anthropocentricity.
INTRODUCTION
After reading “Concrete” A Seven Thousand Year History by Reese Palley, it occurred to
me that the trade of ferrocement could very well create and be the zeitgeist for housing and
building systems in the 21st century. I believe it could and that it should be introduced and
implemented by the Civil Engineers to help integrate it and protect the integrity of the
methods of construction by the “Trade Secret Laws.”
Intact Structures designs and methods are capable of implementing a Ferrocement Building
System that modulates through 4 components to adapt to various, topographical, socio-
economical and educational circumstances. “As has been noted, all things will appear and
disappear through causes and conditions.”
Intact Structures is capable of meeting minimum and maximum building codes without
diminishing the designs attributes or the structures integrity.
373
Intact Structures in Oaxaca, Mexcio
Design Features
Intact Structures Ferrocement Building system can be altered to upgrade or expand, adapt-
ing to the changes of circumstance. Some people need a house/ shelter as soon as possible.
Intact Structures building materials are available and delivered expeditiously with low site
impact addressing the immediacy of the short and long term disaster site requirements
MISSION STATEMENT
The primary focus of the prototype is:
To develop a vocational training program that produces a housing system that is sustaina-
ble and constructed by mentored local people. Providing long-term skilled job creation.
Where an earthquake or typhoon has occurred this trade of ferrocement construction will
be needed to survive the aftermath. Rebuilding for a productive future by constructing
− A Medical Center
− Cisterns and Retaining walls.
− Homes - Schools and Municipal Buildings.
The photograph below is a scene from the construction of the prototype in Mexico.
374
ECONOMIC ADVANTAGES
The material in ferrocement construction is approximately one fifth the volume and the
cost of the conventional materials for the same square feet of construction. The method of
construction utilizes temples and point to point construction to fabricate the buildings
frame and wire work. This creates vocation and saves time and allows the cement mortar
to be applied without form work. The product features conservation and economy in both
material and the way the end product functions.
The tools used to construct the armature of Intact Structures by hand
Wire tie tool | Steel pipe | Rebar bender | Rebar cutter | Tie wire
The materials used to construct the armature of Intact Structures

− Rebar
− 6x6 wire
− ½ x ½ wire mesh
− Sand and Portland cement
− Mason Tools- Mixing hoes-Hawks-Trowels
− Water Tube level- Story Pole
THE TEMPLATE FABRICATION TABLE
This table can be constructed on the job site: Or arrive at the job site on a well-equipped
factory truck stocked with the materials, a mixer and applicator. Or: These components can
be mass-produced in a factory to be delivered to the building site in a container, as a pre-
fabricated well designed multiple use building system.
375
The photograph below is of the template fabrication table with the Intact Structures com-
ponents in position.
This photo is the portable template cloth that is capable of producing the fabrication table
on site.
A PROPITOUS GOAL
Establish an accredited certified training program for Ferrocement Technicians bonded by

the Civil Engineering Professionals and “Trade Secret Laws.” This program could encom-
pass a building project from the site planning stage, to utilizing the use requirements build-
ing codes with the end results developing into human resources and vocational training
programs providing certification for different grades of Intact Structures Ferrocement
Technicians, that could and would construct hopefully self sustaining villages.
376
INTACT STRUCTURES DESIGNS ARE MORPHE-GENETIC
They are flexible in terms of use, meaning that they are capable of multiple design configu-
ration utilizing the same components.
The components are illustrated below:
THE E-DESIGN ARMATURES FOLLOW THE FOOTPRINT OF THE SLAB LAYOUT
The drawing below illustrates, the designs footprint in an on-site slab. It is possible to sal-
vage and utilize the existing slabs in areas where the original houses were destroyed.
The drawing below illustrates, the armature as constructed from the components.
377
This system can be placed on the 2nd FL. slab as well because the total constructed weight
is lighter and stronger than conventional.
THE BASIC DESIGN
The photo below depicts the water cistern being constructed on site.
The photo below depicts the constructed armature (The Columns in this photo are not al-
ways required)
The walls have a 6” depth for insulation, wiring and plumbing. Flexible tubes could follow
the rebar paths to circulate water-cooling the building outer surface while heating the water
for domestic use. One method of insulating the walls to fill the void with earth. The pho-
tograph below depicts the front elevation of the protypes armature during construction,
featuring the completed cistern below the slab.
378
The photograph below depicts the side elevation of the completed prototype. At this site
the grade change of 4’ was utilized for cistern placement.
This cisterns outer wall are also the perimeter foundation It has a capacity of 3000 gal of
water which is used for drip Irrigation and domestic needs.
THE BASIC ELLIPTICAL DESIGN WITH RENDERING
The elliptical design building during the application of the cement.
379
A rendering of the design utilizing roof tiles and cultured stones vertical surface
Delicious herbs and veggies can be cultivated in the window boxes. The much needed
fresh food required after a disaster could be grown in a controlled environment above the
pollution and debris of the inundated site.
FLEXIBILITY OF DESIGN
The drawing below depicts the alteration of the original design and footprint by utilizing
the modular aspects of the components. These structures can be altered at a later time to
meet the changing requirements of the circumstances.
The drawing below depicts the modular aspects of Intact Structure
The photo below shows the back of the prototype building
380
ILLUSTRATIONS OF THE DESIGNS AND APPLICATIONS OF INTACT STRUCTURES

FERROCEMENT BUILDING
Below is an illustration of a complex of garden apartments and the potential for municipal
buildings, schools or commercial units
Below is an illustration depicting the people rebuilding their homes.
381
The retaining wall creates terraced farming and stop erosive flooding while providing a
gravity fed water source to living and working building complexes.
ILLUSTRATION OF THE CANTILEVERED RETAINING WALL
With irrigation system
This system can follow the contours of a mountain road or an eroding hillside to sustain the
topography eliminating washout and collapse. A geotextile fabric installed on the substrate
slab below grade will capillary the above ground drainage into the perforated drainpipe and
direct the run off water into the ferrocement water tank. To be stored for future uses.
Environmental impact
The Intact Structure Retaining Wall increases cultivatable land area by extending the hori-
zontal plane of the hillside as in step farming. It controls the ground water that would oth-
erwise continue down the slope eroding, flooding and lost forever. It could be used for
gravity feed, spray or drip irrigation farming, conventional irrigation or other needs of the
home or industry. It could be directed immediately to catch basins or depleted aquifers’ to
replenish the water table where needed possibly miles away. Because of the looming lack
of adequate water crisis in many areas of the world this is a very Intriguing aspect of the
design.
Social impact
Intact Structures Designs will produce cost effective durable Ferrocement Housing Sys-
tems with multiple design configurations. In tandem with the accreditation of a vocational
training program designed and implemented by the Civil Engineering Professionals to help
integrate and protect the integrity of the building system. Intact Structures goal is to utilize
water collection and water management to create the circumstance of self sustaining hous-
ing and village complexes by enabling the people to collect and store rainwater and runoff.
The cantilevered ferrocement retaining wall will foster the development of and cultivate
the food and water sustainability of the people’s independent healthy life style.
Economic advantages and sociological advantages
382
The Material in ferrocement construction is approximately one fifth the volume and cost of
the conventional materials used to construct the same square feet of wall and roof areas.
The method of construction utilizes templates and point to point construction to fabricate
the buildings frame and wire work. This type of vocational training is very expeditious and
adds intrinsic value to the lives of those who acquire it by granting them a certification in
the skilled trade of Intact Structures Ferrocement Building Systems. Intact Structure de-
signs features conservation and economy in both material and the way the end product
functions. It appears that because of the low site impact of this construction process it is
possible to revitalize devastated areas incrementally, maintaining cultural ties and human
resources by offering an alternative to forced relocation. The flexibility of the material can
help maintain indigenous details on the buildings. As civilization once again hopefully
begins in the “home.”
383
384
PROPOSED JOINTS FOR A FERROCEMENT BUILDING RESI-

DENTIAL SYSTEM
Msc. Julia R. Alvarez, Civil Engineer Department. Construction and Tropical Architecture Center
(CECAT in Spanish) of Technical University of Havana (CUJAE). Avenue 114 No. 11901
between Ciclovia y Rotonda, Mariano - CP 19390, Havana, Cuba; Dr. Hugo Wainshtok Rivas;
Eng. Debora Acosta
Abstract: The existing housing deficit in our country demands the search for new structural
solutions that respond to our climatic conditions in fast, economic and resistant ways. The
Ferrocement Residential Building System (SERF in Spanish), formed by prefabricated panels of U
cross section form, allows varying their dimensions easily, with small modifications to the molds
and it facilitates their use in walls, floors and roofs. This technology permits to obtain a great
variety of solutions, for isolated housing or residential buildings up to 5 stories high, with
significant savings of materials, equipment and time of execution. Taking into account the great
importance of joints in precast structures, some variations of joints between panels used in this
system are analyzed in this paper. The proposed joint permits to obtain a flat surface. For the
design of the joints different aspects were taken into account like the constructive easiness,
assembly and mainly their structural behavior.
INTRODUCTION
The ACI Committee 549 defines ferrocement as a type of reinforced concrete, built in form
of a thin sheet of hydraulic cement mortar and reinforced with layers of mesh not very sep-
arate and reinforced by continuous wires of relatively small diameter [1]. The mesh can be
of metallic wire or another material that it is appropriate. The mortar and their composition
should be compatible with the mesh [1, 2]. Some authors define ferrocement based in spe-
cific surface concept (relationship of the superficial area of the steel and the volume of the
composite) or the steel dispersion in the mortar mass.
Naaman, gives minimum values of specific surface in two direction of 0,8 cm-1, although
he recommends to double those values in the case of being used for deposits of liquids, he
defines also 1,8 % as minimum value for reinforcement factor (volume of the reinforce-
ment related with the volume of the mortar). The group of San Carlos has worked with
lower values, arriving at a total specific surface of 0.35 cm-1 with quantities of steel in
order of 200-300 kg/m3 mortar, calling this material reinforced mortar, which they consid-
er a synonym for ferrocement. Those limits have been fixed arbitrarily by all authors[1].
The steel dispersion can be up ten times greater than in conventional concrete. With results
in a large resistance to cracking of the mortar. This distribution of the steel contributes to a
higher resistance and elasticity, allowing to decrease the dimensions of the traverse section
elements. Therefore the weight and the volume of the materials can decrease in more than
50% and the reinforcement in 35% compared to similars structures of reinforced concrete
[3].
385
This material is introduced in Cuba in the decade of the 60 ties of the last century, based on
the experience in the construction of boats in Canada and New Zealand. The Cuban fishing
fleet was in crisis, and it demanded a radical change, and was decided to use ferrocement
as an alternative, making use of its multiple advantages. After several intents it was possi-
ble to build the FC-IV, boat which fulfilled the structural and naval requirements. Starting
from this experience there was a peak in the constructions with this material. Under the
guidance of Dr. Hugo Wainshtock new models of ships were built for fishing, recreation,
passenger transport and sport. The next step was to build pool, water deposits, houses and
sculptures of big dimensions [3].
FERROCEMENT HOUSING
The housing deficit is one of the main problems in Cuba, and it is a necessity to search for
structural solutions that respond to our climatic conditions and give an answer in a quick,
economic and resistant way to the natural phenomena that occur in our geographical area.
In the year 1986, the first housing elements with ferrocemento were built in Santiago de
Cuba. The panels had transversal section in U form, with or without partition in the ends.
They were formed on a concrete floor with wooden molds, allowing variations in size us-
ing the same mold type [3]. This experience extended to other cities of Cuba, like Guantá-
namo, Camagüey, Pinar del Rio and Havana with little modifications in the molds that
were used.
The U section panels were selected because it allows varying its dimensions easily, with
small modifications to the molds, it facilitates its use in walls and slabs. Also an appropri-
ate resistance is achieved with little reinforcement. In 1988 the first two-storied house was
built, in San José de Las Lajas, Mayabeque. The technology has evolved since and nowa-
days it is a group of machines and devices recognized for its simplicity, mobility, economy
and productivity (Fig.1). Now appropriate tools exist for the dissemination of the Ferroce-
ment Residential Buildings System (SERF).
The creation of the necessary machinery for the panel’s construction was realized by Dr.
Sergio Marrero Osorio of the Mechanical Engineering School, and structural design was
directed by Dr. Hugo Wainshtok of the Centre for Construction and Tropical Architecture
(CECAT in Spanish) of the Technical University of Havana CUJAE. This technology al-
lows achieving a great variety of solutions for isolated houses or residential buildings up to
a height of 5 stories, with significant savings of materials, equipment and time of execution
[1].
The system is composed of a small quantity of different prefabricated elements. They can
be combined with a high grade of flexibility, allowing to build houses and social or admin-
istrative buildings. The weight of the elements is lower than 140 kg, what allows the trans-
portation and mounting up to a first level in a manual way, with 4 men. The assembly of
higher levels needs accessories and equipment with manual control. The production don’t
need highly qualified worker, nor expensive equipment. The technology consists of a ma-
chine with suspended immersion vibrator and externally activated, a group of slipping in-
terchangeable molds in dependence of the traverse section of the element, a couple of light
rails, two manipulation pincers, a conventional wheelbarrow and a pulley system for the
movement inside the plant and for the assembly [3]. It is made for quick, easy and econom-
386
ic transfer and assembly at the construction site. This equipment has been used in several
countries of America with satisfactory results.
Figure 1: Technology used in the SERF system
JOINTS
The connection among the elements is one of the problems of more difficult solution in the
design and construction of precast buildings, they assure the structural continuity of the
unit, and achieve that the components work as a system. In its conception it is necessary to
keep in mind technological aspects, such as production easiness and assembly of the ele-
ments, to facilitate the execution and to achieve all this guaranteeing the structural behav-
ior of the system, according to the project conception. In the ïn situ¨ constructions, the
continuity is achieved with salient bars of the element. This advantage cannot be reached
totally in the prefabricated structures.
Two types of joints are used to transmit shear efforts, the plane and keyed joint. Tests car-
ried out and both types of connections demonstrated that the specimens have an elastic
lineal behavior until the appearance of the first cracks, in the filler material or in the inter-
face between this material and the face of the prefabricated element. It is the failure for
adherence in the interface that determines the beginning of a new stage. In plane joints the
loss of adherence means the failure of the connection, in the keyed joints after the loss of
adherence begins the shear mechanical transfer, this advance from the border of the ele-
ment towards the filler material [4].
In the keyed joints, the first crack appears in the corner of the key, in the filler material, in
this place there are a concentration of tensions and the crack spreads slantingly when in-
creasing the load. The failure can occur by local shear in the key. The geometry of the bor-
der of the element influences in the resistant capacity. The size, the slenderness and the
density of the key (λ), are aspects that influence in the behavior of the connection. It has
been observed that the resistance of the joint is increased when the depth of the key is in-
creased. The longitude also influences in the resistance and in the way it fails. For keys
with a small width, the failure is caused for shear in the base of the key, while for big val-
ues failure occurs in the corners of the key. An increment of the λ values, produces an in-
crease of the resistant capacity, but this increment has a limit, values higher than 0.5 can
cause the failure in the key of the precast element [4]. The interlock aggregate is other as-
pect that influence in the structural behavior of the joint. It is produced when part of aggre-
387
gate particle in one side of the crack stand up. The wedge effect between the aggregate and
the fissures of the concrete causes successively relative displacements of traverse and tan-
gential components. This effect induces tensions that cause compressions in the concrete,
which are transmitted through the cracks.
For the design and conception of a joint it is necessary to consider aspects that affect its
behavior, for example the technological factors. The first condition is a correct filling and
vibration of the concrete or mortar filler. The presence of steel bars contributes resistance
to the joint through mechanisms like dowel action, tensile force in the steel, but these bars
can hinder the correct filling of the joint, factor that it is necessary to keep in mind. On the
other hand the grade of adherence is affected by factors like the cleaning of the surfaces,
and the humidity. The joints have to achieve the impermeability to water and air, acoustic
and thermal isolation; conserving an aesthetically appropriate aspect and guaranteeing the
durability of the material of the filler. In this article, variations of connections are analyzed,
that can be used for walls, floor and covers.
CHARACTERIZATION OF THE JOINT IN THE SYSTEM SERF
The joint used in the SERF system is the result of a parametric design of the panels. This
design was carried out by Dr. Sergio Marrero. The panels use the keyed joint to transmit
shear effort, this article analyzes some variations of joint, permitting plane surface. For the
design technological and structural aspects were considered.
Figure 2: Joint representation
The joints have presented some difficulties, the system considered slope roofs to evacuate
water rapidly, but in reality there is filtering and it´s necessary to add impermeabilization
system. Also the technology used in Cuba to produce the panels doesn't present sufficient
accuracy to achieve a total linearity, which influences the non accuracy of the assembly.
PROPOSED JOINTS
Thus, new joints are proposed. The panels have the U form, based in the whole construc-
tive system. In the new analysis a possible collocation of steel bar in wall joints was con-
sidered to guarantee the continuity between the foundation and the beam. In the design of
roof panels it was considered to place of a cord of asphaltic emulsion to guarantee the im-
permeability of the joint.
388
Walls joints
Case 1 Case 2 Case 3
Figure 3: Wall joints
The proposal 1 and 3 require a panel thickness of 3 cm, and the second one requires 2.5
cm. Those variations were analyzed, proposal number two was selected when a thermal
and acoustic isolation is applied, proposal three will be used without thermal and acoustic
isolation since it´s thickness will cover those necessities. The proposal number one is dis-
carded because the requirements of the system can be solved with a smaller panel thick-
ness.
Joints for Floor or Slabs
Case 1 Case 2 Case 3 Case 4
Case 5 Case 6 Case 7 Case 8
Figure 4: Joints for floor and roof slabs
In a first selection process the first three variations are discarded for constructive aspects,
the asphalt is the material that first is filled in the joint, which provokes it to slip through
389
openings and could cause discoloring difficult to solve. Also it prevents the mortar from
adherence to the surface of the joint and therefore it doesn't complete its structural func-
tion. In a second analysis the proposals seven and eight were discarded because in thin
joints filled with mortar appeared longitudinal fissures caused by the retraction of the con-
crete with too much frequency. It shows to be necessary to provide a mechanical anchor-
age for the joint to work structurally, this is obtained variations the vertical surface of the
panel, missing in those proposals. In another elimination number six is discarded due to the
constructive complexity to fill the joint with mortar possibly leaving empty spaces, which
can affect their structural capacity. Therefore this leaves alone the variants four and five to
analyze.
ANALYSIS OF PROPOSED JOINTS

Strength
For the calculation of the shear resistance of the joint the next expression is used [5]:
(1)
: Concrete resistance.
Dimension of traverse section.
This is the simplified expression which allows to adopt a more conservative value, it
doesn´t consider the influence of the keyed effect [5]. In the proposed joint for the system
there will be bolts employed for earthquake load, but this formula is used since the bolts
only work when the mortar has failed, in this case it will be considered as shear strength
that the manufacturer indicate according to the type of bolt used [6]. The relations of shear
acting, shear resistance, load and factor of security for each case is expressed in the next
table.
Table 1: Relation of shear acting, shear resistance, load and security factor for wind load.
Case (kN/m) (kN) (kN) F.S
Wall 4.04 5.09 46.98 9.23
Case 4 6.77 8.53 58.59 6.87
Case5 6.75 8.50 83.43 9.82
Table 2: Relation of shear acting, shear resistance, load and security factor for earthquake load.
Case Fasteners (kN) (kN) F.S
Wall M-10 29.40 90 3.06
Case 4 M-14 17.06 200 11.72
Case5 M-20 17.00 240 14.12
390
OTHER INDICATORS
Other indicators are analyzed to select the joint.
1. Joint mass: For their influence in the weight, more mass augments the weight of the el-
ements.
2. Joint volume: It is the factor that determines the quantity of necessary materials, it is an
economic indicator.
3. Contact area: On this indicator depends the permeability of the union between the faces
of the joint and the panel, the larger the contact area a water drop has to travel bigger the
possibility that it evaporates and does not filter through the cover.
4. Moments of inertia of the joint respect the axis X and Y: It offers an idea of the rigidity
of the joint.
5. Construction facility: The speed with which the joint can be executed it is an important
factor that determines the times of execution and it facilitates the process of construction
continuity.
Table 3: Countable indicators
INDICATORS CASO 4 CASO 5
Shear strength (kN) 58.59 83.43
Joint mass: (kg) 25.66 22.96
Contact area: ( ) 185.79 225.79
Joint volume (litros) 10.97 9.81
X :217480236.59 X :221198863.52
Moments of inertia
Y :317780949.81 Y :281716334.63
CONCLUSIONS
SERF is a very economical prefabricated system and of simple manipulation, its develop-
ment in America has generalized in the last years. The proposal number two for wall pan-
els will be used when a material of thermal and acoustic isolation is applied or in countries
where it doesn't require it. For slabs it is the variant 5 that possess bigger shear strength,
bigger contact area and smaller material volume. In other analyzed aspects significant dif-
ferences don't exist among the proposed solutions. The joints resist the permanent, use and
wind loads with appropriate factors of security. In the case of seismic loads the use of bolts
guarantees the stability of the panels in case of failure of the mortar joint.
391
REFERENCES
[1] Wainshtok, Hugo. "Ferrocemento Diseño y Construcción”. Editorial Fábrica Comunicación,
Ríobamba, Ecuador, 2010, pp. 35 – 63.
[2] Naaman, Antoine E. “Evolution of continuous reinforcement for Ferrocement and thin cement com-
posites”. Ninth International Symposium on Ferrocement and thin reinforced cement composites,
Baly , Indonesia, may 2009.
[3] Wainshtok, Hugo; Lizazo, Yen-Liu. “Sistema de edificios residenciales sismoresistentes de
ferrocemento (SERF): Una opción para la construcción de viviendas de interés social”. En actas de
X Simposio Internacional del ferrocemento y compuestos delgados de cemento reforzado, La
Habana, Cuba, 2012 pp 375-392.
[4] Navarro Campos, N. “Applicability of analytical models of shear joints in precast structures.” Doc-
toral thesis, Warsaw University of Technology, 1986.
[5] Hernández Caneiro, J. A; Hernández Santana, J. J. “Hormigón estructural. Diseño por estados
límites.” Parte I. La Habana: Formato Digital, 2010, pp. 327 – 387.
[6] Acosta Álvarez, Debora. “Diseño de alternativas para juntas entre paneles del sistema SERF.”
Trabajo de diploma, ISPJAE, 2014.
392
32 YEARS WITH FERROCEMENT

FROM SMALL PREFABRICATED BRIDGES TO PONTOONS FOR
FLOATING HOUSES
Carlos A. Gschwind, Carlos Gschwind Construcciones
Abstract: We began to work with ferrocement in 1983. The first development was a small
prefabricated bridge system. Culverts of 2.00, 3.00 and 4.00 m of span made with parabol-
ic arched plates. The intention was to replace the reinforced concrete tubes and the ARM-
CO steel culverts. The result was successful because the plate was more resistant, more
durable and cheaper than other products in the market. Then, we built walls for houses,
with two thin walls of ferrocement stuffed with granulated volcanic stone. In 1993 we be-
gan to use polypropylene fiber as reinforcement for ferrocement structures. Under the ne-
cessity to replace imported gabions and mattresses filled with stone that has a very expen-
sive transport cost, we developed hollow parts stuffed with sand, ground, mud or any local
material. In 1994 we began to build swimming pools of different kinds and sizes. So far,
we made more than 200 pools. The greatest is of 20 x 43 m and 1.000.000 liters. An Ar-
gentinean company, whose plan was to build 30 floating houses near Buenos Aires city,
entrusted us to build pontoons for them. We also developed pontoons with a bottom that
has undulations and a curve beam on the top all around it to provide greater resistance for
the entire hull.
CULVERTS
Argentinian Patent N° 234827 and 252629
I began to work with ferrocement in 1983 when I was living in the south of Argentina,
where roads needed many culverts. Steel culverts and concrete tubes were been used, so I
decided to develop a new system of culverts for replacing them. The concrete tubes of 1,00
m of diameter were unpractical and the Armco steel was very expensive. With the small
ferrocement bridges of 2,00 m; 3,00 m and 4,00 m span there was an appropriate solution,
cheap and practical.(Figures 1, 2 and 3). This system was a set of undulated and parabolic
plate of 1,20 m wide and only 50 mm thick. One plate of 4,00 m span weighs less than one
tube and has an hydraulic section that replaces 7 tubes but with the cost of only 3 tubes. In
1985 I sent a paper about Ferrocement Folded Arched Culvert to the 2nd International
Symposium in Bangkok, Thailand. It was published in the corresponding Proceedings.
After that, other paper was published at JOURNAL OF FERROCENT, Volume 17, Num-
ber 1, January of 1987, after attending the World Exhibition of Achievements of Young
Inventors that was organized in Plovdiv, Bulgaria, in November 1985. The exhibition was
organized by the World Intellectual Property Organization (WIPO) From Genève, Switzer-
land. Some years later, in 1993, after several investigations, I began to use polypropylene
fiber to improve ferrocement impermeability.
393
6’6’’ span plate 10’ span plate 13’ span plate
Figure 1: Plates for Culverts
Figure 2: Transporting and Assembling plates
Figure 3: Finished Bridges
394
WATER TANKS AND SEPTIC TANKS
Following market trends, I developed some new products: water tanks (Figure 4) and sep-
tic tanks. In very windy areas, the high own weight was an advantage. Based on that fact
and using old fibercement as mold, I developed ferrocement water tanks and septic tanks.
The wall was 15 to 20 mm thick.
Figure 4: Typical examples of water tanks
SUPPORT WALL
By using abundant volcanic material (pozzolana stone) found in south Latin America re-
gion, I developed a wall for prefabricated houses. The type module was 90 x 260 cm and
20 cm width. The internal and external faces were a thin plate of ferrocement (20 mm),
joined to horizontal and vertical ribs. These ribs formed a sort of pigeonholes which were
filled with pozzolana. This structure, confined into holes, provided enough strength to sup-
port a three floor building. Besides, this wall filled with pozzolana provides an excellent
thermal isolation, ideal for the Andes Mountains where in winter temperatures could reach
– 20° C but in summer temperature rises up to 40° C.
PLATE FOR INDUSTRIAL WALL AND ROOF.

Other innovative application for this material, was an undulated plate of 1,20 m wide and
3,00 m, 4,00 m or 5,00 meter length. At that time, the steel was still very expensive, so this
development was a real solution. Where climate is windy, roofs and walls with good
weight are more convenient than steel pieces.
LARGE DIAMETER PIPES WITH RIBBED WALLS (ARGENTINIAN PATENT N°

236487)
To make culverts installation easier without special machinery, I developed a pipe with
two plates joined by screws and nuts. For example: half tubes of 1,50 m diameter,
weighted only 250 kg/m. There are many of them installed in north of Santa Fe province,
Argentina since 1987.
395
Some years later, I built large diameter pipes of 5,00 m length,(Figures 5, 6 and 7) without
using mold. Thanks to these ferrocement pipes, that were not expensive and they were
easy to build, I was able to win public biddings. A good example is the repairing of a very
old pluvial drainpipe in Santa Fe city, Argentina.
Figure 5: Transporting tubes Figure 6 :Placing the tubes
Figure 7: Placing the tubes of 1,75 m diameter and 5,00 m long
PRECAST RECESSES FOR CEMETERIES (INDUSTRIAL MODEL N° 53144)

Another demand in the market were precast recesses for cemeteries (Fig 11) According to
a social traditions, in my country, each town has a cemetery with a battery of tombs. With
this system, there were built with only two walls and a bottom. The objective was to reduce
costs and to make the structure lighter.
396
Figure 11: Assembling and finished precast recesses for cemeteries
GABIONS – HOLLOW PARTS FOR COASTLINE DEFENSE (ARGENTINIAN

PATENT N° 318149) (FIGURES 12 AND 16)
These products were designed specifically for places of difficult accessibility or with avail-
able rocks to fill gabions and/or where mattresses are distant or very expensive. They are
hollow parts (Figure 13) of different shapes. Thanks to this, they can be transported in tan-
dem floating to their final destination. Gabions are filled with sand, land and even mud
(Figure 14) and they are later sealed (Figure 15). They could be manufactured in different
sizes and placed, linked and locked to one another. They can be placed over geotextiles
that prevent the passage of soil particles.
Figure 12: Design and patent
397
Figure 13: Hollow parts Figure 14: Filling with sand and land
Figure 15: Covering and sealing gabions Figure 16: Coast protection after 6 years installation
Tetrahedron of Ferrocement: Four-faced hollow pyramidal block.

Due to its lightness when it is empty, is not necessary to use special equipment and it is
possible to be handled only by two workers. It can be filled with any material found near
the final location. It was designed and built specifically to protect coasts, bridge pillars and
culverts from water erosion.
SWIMMING POOLS
Twenty years ago, seeing great manufacturing faults in swimming pools built with brick
wall, reinforced concrete or pre-molded plates, I developed a new building system for
swimming pools. Taking advantage of ferrocement and my construction experience, I built
a monolithic structure with perimeter columns (ribs of 50 x 5 cm each 2,00; 2,50 or 3,00
m) and a base of 0,20 x 0,50 x 1,20 to 1,50 m. Half of this base is under the bottom part of
the piscine. Another very important procedure is to treat the soil under the piscine. In a
zone where the soil changes the volume due to its humidity tenor, it is important to let the
soil works, without transmitting weight to the piscine structure. The secret is to make lon-
gitudinal and transversal furrows and put straw or paper cement bags. Then, cover all the
surface with polypropylene film and above to put the steel meshes, etc. Another secret is a
beam over and around the piscine joining all the ribs (columns) (Figure 17). It gives the
final of the monolithic structure. This procedure gives good results, offering 10 years of
398
guarantee, I had never have to repair any swimming pool. Over the past 20 years, I built
more of 200 piscines from 10.000 to 1.000.000 liters. Across the years, I realized that there
were many people with disabilities that did not enjoy pools. So I developed a lift for
swimming pools, which is not made of ferrocement but of stainless steel. (Figures 23 and
24).
Figure 17: Steel meshes Figure 18: Wet plaster
Figure 19: 100.000 liters piscine Figure 20: Familiar piscine of 200.000 liters
Figure 21: 80.000 liters piscine Figure 22: 1.000.000 liters piscine.
399
Figure 23: Stainless steel swimming pool lift Figure 24: Stainless steel and fiber swimming pool
lift.
PONTOONS FOR FLOATING HOUSES.
An Argentine Company, planned to build 30 floating houses near Buenos Aires city, and
started looking for pontoons in USA and Canada. There are many floating homes there and
their pontoons are made of steel. It was a great problem for Argentina due to the periodical
maintenance work. They reject this option. Then they travelled to Holland where they
found pontoons made of reinforced concrete, which is a very heavy and expensive materi-
al. Finally, the solution was the ferrocement. I began with the idea of building a hull for a
Floating House that could serve to sail also. Double hull was the ideal solution. Two hulls
which walls of only 50 mm thick are separated by 150 mm but joined by longitudinal and
transversal ribs of 50 mm wide with space between two hulls filled with polyurethane. Wa-
tertight compartment were completely watertight and practically impossible to sink. The
emplacement of the neighborhood allowed building hulls but not for sailing, so there was a
need of building a simple hull.(Figures 25, 26 and 27) The wall had only 50/60 mm thick.
The bottom was undulated with transversal reinforced ribs. The lateral walls finished in a
superior belt-beam that unified the effort of all the structure. Thus the hull is converted in a
monolithic structure. Example: pontoon of 5,00 x 7,00 x 1,45 meter weigh 12.000 kg and
to sink empty only 0,35 m.
400
Figure 25: Metallic framework pontoon. Figure 26: Plastering pontoon.
Figure 27: First pontoon of 5,00 x 7,00 x 1,45 m Figure 28: Wood house finished upon ferrocement
newly launched. pontoon.
Figure 29 :Wood house under construction upon Figure 30: Finished wood house upon a ferrocement
ferrocement pontoon of 6,00 x 9,00 m. pontoon of 5,00 x7,00 m.
ANNEXE
There is a wide list of other products that I developed and produced.
FERROCEMENT PRECAST HEXAGONAL SILOS.
It is well known, that Argentina is a great grains producer. Because of that, I had devel-
oped a silos system to be installed in field. I designed a “Z”-shape plate to build a hexago-
nal silo. Each plate is 1,00 m high and 2 or 3 parts of 2,00 m long to assemble a set of dif-
ferent height and capacity.
PREFABRICATED BEAMS FOR BUILDING SLABS.
With similar shape, section and price, I was able to achieve 40% more strength than similar
pre-tense beams. Using unfolded steel mesh and a steel bar of 10 mm, and using a mortar
with fiber, very well vibrated into a mold, it is possible to obtain a beam that works at trac-
tion and compression in both senses.
401
MIXED SLABS.
Steel beams with steel meshes and concrete formed a ferrocement mixed slabs. It is a
quicker and cheaper system in comparison to reinforced concrete slabs. Using steel beams
and steel meshes according to structural calculus and unfolded steel mesh (know as Rib
Lath) it can be made slabs without mold. Then, plastering a fiber mortar of 25 or 30 mm is
possible to obtain a slab for roof.
FERROCEMENT FLEX BEAM FOR PROTECTION ROADS.

It is a practical and cheaper system for replace the steel flex-beam in urban street.With
similar shape to steel flex-beam, can support 2,00 or 3,00 m of construction. It is possible
to obtain a good side protection for streets and roads.
SPA (Jacuzzi) (Figure 31)

Since ferrocement swimming pool do not need to use mold, it is possible to give different
tailored shapes and sizes.
Figure 31 - Small pool for SPA (Jacuzzi)
CONCLUSIONS.
It is important to stand out the great cost and time reduction with Ferrocement products as
regards to Reinforced Concrete products. For example: a pontoon of 5,00 x 7,00 x 1,50 m
needs only 5,00 m3 of mortar. A reinforced concrete pontoon for the same performance
need more of 20,00 m3 of concrete (measurement 5,00 x 7,00 x 4,00 m) With only a 25%
of mortar it is possible to obtain the same result. Ferrocement is waterproof and the rein-
forced concrete needs an additional treatment to be waterproof. Ferrocement allows sav-
ings in time and money.
For a structure made of Ferrocement is necessary to use only a third part of the cement
needed for a similar structure made with reinforced concrete. Due to the 60% lower con-
tent of cement in its preparation, it is less contaminant for the atmosphere.Ferrocement has
many different applications, leveraged on a very low cost and the possibility of getting
402
tailored shapes. In site construction and fast montage, are also advantages of this versatile
material.
REFERENCES – MANUSCRIPT
Carlos A. Gschwind is Building Technician – National University of Litoral – Arg.
Degree in Business Management – National University of Litoral – Argentina.
Inventor of 20 products.
Address: Urquiza 3747 - 6° B – (3000) Santa Fe – Argentina
e-mail: carlosgschwind@gmail.com, web: www.sitioferrocemento.com, www.pool-lift.net
403
404
FERRO-PAST AND FERRO-FUTURE
Michael Pemberton, Trafalgar Marine Technology Ltd. Master Mariner. AMRINA. Member of the
Concrete Society and the Concrete Bridge Development Group. Martin Pullan. EngTech
TMIET.AMICT, AMRINA. Member Engineering Council. Member of the Institution of
Engineering and Technology, Associate Member of the Institute of Concrete Technology. ACI 549
Committee Member. Antony Tucker. Naval Architect, Tucker Designs. Fellow Member YDSA.
Abstract: This paper comprises a summary of the Author’s Ferrocement activity since at-
tending the Nervi Symposium of 1981, prior to taking a licence for the UK and Europe to
use the spray and laminating technology invented by Iorns and Watson of the Fibersteel
Company of California. The Nervi Symposium was the first and largest International Fer-
rocement Symposia. Over 200 people attended, only two of the delegates, the author, Mi-
chael Pemberton from Leeds, UK. and Martin Iorns from California, were entrepreneurs.
The remaining delegates were ferrocement academics from around the world. This paper
asks the question - where in our construction education agenda is ferrocement today? This
paper follows the establishment of Ferrocement Laminates UK Ltd. in 1982 and the devel-
opment work to build pontoons and then roof leaves of the deMenil Museum, Houston.
These roof leaves are still claimed to be the world’s most accurate and technically ad-
vanced ferrocement / concrete castings. The paper continues with reference to work in UK
developing sewer lining systems, FCL, (Ferro-Monk) and Russell. (Alphacrete). The paper
describes the building of pontoons for marinas and to support houses. Some extraordinary
ferrocement work is discussed and illustrated. The paper describes the benefits of the spray
and laminate system of manufacture and discusses the potential for future development of
high tensile laminated ferrocement plates to be used as permanent shuttering. Historical
ship recovery and replication is discussed. The paper concludes with a reference to the
solar roof of the Stavros Niarchos Cultural Centre in Athens. Which is without doubt, the
most remarkable application of ferrocement construction to date. “If your put steel into the
ocean – better to wrap it in quality concrete – or the costs of maintaining that steel (paint-
ing) will dissipate the revenue earned from the structure.” This applies to ships and all
steel structures in the ocean!
FERRO-PAST AND FERRO-FUTURE

REF. 1.
This is a summary of the author’s attendances at Ferrocement Symposia from 1981 to date.
It is not a definitive record of all the Ferrocement Symposia held every four years but those
he attended and presented. It was these symposia which most influenced the author and the
paper provides some interesting anecdotes about the material and its application over the
past thirty five years. The author asks the questions:
“Why has Ferrocement not been more widely used commercially?”
405
“Why is the material not readily acknowledged as the first example of reinforced concrete
– the world’s most important construction material?”
“Why does the teaching and history of ferrocement not feature more in the engineering
curriculums today?
“Where are today’s construction innovators to gather their knowledge of Ferrocement and
its history now that the IFIC in Bangkok is no more?
From these symposia and with help from the International Ferrocement Information Cen-
tre, Bangkok the author gained his knowledge of the material and subsequently endeav-
oured to market laminated ferrocement materials.
FERRO 1. THE NERVI SYMPOSIUM, BERGAMO, ITALY, JULY 22-24 1981 REF. 2.
The author attended and met many people with interests in Ferrocement. He particularly
remembers meeting many young professors who subsequently progressed through their
careers to end as heads of their departments. Notable amongst them are Professor’s R N
Swamy (Sheffield UK), Antoine Naaman, (Ann Arbour, Michigan) and Surendra Shah
(North-western University USA. and P. Paramasivan (Singapore) and Lilia Austriaco from
IFIC.
Despite the knowledge and enthusiasm of our leading professors with their understanding
of ferrocement, the world has not made sufficient commercial use of the material. The
question to consider and debate is - why not? The International Ferrocement Information
Centre and Lilia Austriaco, provided an excellent service, for researchers and enthusiasts,
by disseminating knowledge and publishing / distributing the Journal of Ferrocement.
Conventional ferrocement became widely used in the developing countries of Africa and
India for numerous applications. The American Concrete Institute produced a State of the
Art Report which provided guidance to good practice. Unfortunately the development and
adoption of ferrocement manufacturing did not really establish within the concrete industry
and except for the extraordinary use on special projects the material became difficult to
market and has become very much forgotten from the academic agendas. The very basic
principles and understanding of why ferrocement works seems to have been superseded by
the desire to experiment with many new materials to form exotic composites.
Ferrocement Laminates UK Ltd

After securing a licence to use the Iorns/Watson/Fibresteel laminating technology the au-
thor established Ferrocement Laminates UK limited in 1982 with an international board of
experts in the ferrocement field. The company initially made laminated ferrocement pon-
toons for floating piers and a floating fuel station for a marina. The company carried out
development work for the DeMenil Museum Project and then secured a very large order to
make the roof leaf units for the museum and export them from UK to Texas, USA.
406
Fig. 1. The DeMenil Project Team Leeds Yorkshire 1983. The author is pictured with white safety helmet
and Renzo Piano stood to his left.4th From Left, is John Salmons (Crown Foundry), Prof.Gurdev.Singh Leeds
University) Dr.Salam Alwash (FCL) and Martin Iorns are picture to the left side.
Fig 2. Floating Pontoon. Letterfinlay Pier, Scotland, installed 30 years ago.
407
Fig 3. The deMenil Museum Roof Leaves – 1985.
Sewer Renovation Systems
In 1986 the market appeared for renovating brick lined sewer systems in the UK. Subse-
quently FCL worked with (Monk) to form Ferro Monk linings. This company continued to
carry out your sewer lining works for 20 years FCL (Ferromonk) and Russell (Alphacrete)
and Kennedy were the companies working in the sewer lining market at the time. Some
sprayed mortar and some injected mortar. Invert and crown pieces were made in precast
Ferrocement. Extensive ferrocement testing was carried out by Paul Nedwell at Manches-
ter UMIST for the Water Research Council to verify the qualities of Ferrocement for appli-
cation in sewer systems. These tests comprised running ferrocement samples in a circula-
tion tank with hot /dry and chlorinated water and cycled for several months. The epoxy
paint coated steel clamping frame used to hold the ferrocement samples under test in the
middle of the test tank, was found to have completed disintegrated to a pile of rust at the
bottom of the tank, while the ferrocement samples were as good as new. Recent tests on
ferrocement samples cast 25 yrs. ago and exposed to the weather on Manchester University
roof has shown that there is no penetration into ferrocement samples made with a 2/1 ce-
ment mortar? Carbonation <1 mm in 25 mm samples! When correctly made, this is a re-
markable material!
FERRO 5. MANCHESTER UMIST 6-9TH SEPT.1994.

REF. 3 AND 4.
At this Ferrocement symposia the author particularly remembers the conference dinner
where Professor Rowland Morgan, from Bristol University, the authority on Concrete
408
Shipbuilding and author of several very interesting ferrocement papers, stood up amongst
the dinners to state. “We - with our interests in Ferrocement - are a secret society”. That
was in 1994 and here we are, twenty years on, and we ferrocement enthusiasts are still a
secret society!
FERRO 6. THE LAMBO SYMPOSIUM UNIVERSITY OF MICHIGAN 1998.

REF. 5
At this symposium the author met and discussed with Larry Mahan the building of his rep-
lica ship ‘Larinda’. Larry Mahan had made such a good job of his ship that he took it to
display at a traditional boat and yacht builders show on America’s East Coast and he won
the first prize for his lovely ship. But some one noticed the hull was made of Ferrocement
and so they took away his prize. That was not the only bad luck to befall Larinda. Some
years later she was sheltering in Halifax, Nova Scotia from a tropical revolving storm.
When the storm came ashore Larinda was crushed between a quay wall and a large historic
steel warship and sunk – right under the main sewage outlet for the town of Halifax. Poo!
Eventually the ship was recovered, cleaned and rebuilt and still operates from that area. A
sad story with a very sad ending, worthy on more detailed account one day! Around 2000
the most interesting ferrocement work was the architecturally spectacular Yambu Cement
Company Ltd in Saudi Arabia, constructed by Derek Russell of Alphacrete Construction
Linings and Engineer Patrick Jennings. The Architects were Franco Audrito and
D.Angelotti. Studio 65, Turin. Italy Fig 4 and 5. The ferrocement panels/armatures for this
construction were ‘injected’ with mortar Panels made of a polystyrene core with an arma-
ture of meshes on top and bottom surfaces. The panels were then hand finished.
Fig 4. Sculptures for the gate of Yanbu’al Bahr Cement Plant, Saudi Arabia. (courtesy of l’industria italiana
del Cemento – Anno LXX N.6 Guigno 2000)
Pontoons for floating houses

From 2005 to 2010, the authors of this paper were engaged building ferrocement pontoons
to support houses on the River Thames.
409
Fig 5. Pictured here is one of four pontoons which support an 80 tonne cedar house.
The Stavros Niarchos Cultural Centre Athens. 2013 -2015

The most remarkable ferrocement work in Europe today is the construction of the solar
roof for the Stavros Niarchos Foundation cultural centre in Athens. This most unusual roof
100m x 100m is constructed with ferrocement skins top and bottom with a steel frame in
between. It might well be described as the modern equivalent of the ancient Parthenon
Temple on the Acropolis.
SO WHY DO WE LAMINATE FERROCEMENT?
There are several ways of constructing ferrocement and over the years methods have been
found to increase productivity. The original and traditional method of making an armature
of wire and hand plastering has been used since Lambot’s invention in 1848 and Patent of
1855 but is too laborious to be commercial. The application of positive displacement mix-
ers and ‘mono’ type pumps to convey the mortar permits spraying of mortar onto a mould.
The embedding of reinforcements into a plastic mortar is the essential methodology of the
laminating techniques invented in California in the 1960s.
This was proved to be the best method to manufacture the roof leaves for the DeMenil pro-
ject in 1985 and subsequently for the manufacture of sewer lining systems. So why does
the spray and laminating technology produce the best in quality ferrocement?
1. Firstly, the mixing of mortar in a positive displacement mixing machine and then
passing that mortar through a mono type pump is almost self-controlling.
2. A ferrocement mortar mix of 2 to 1 sand/cement ratio of plastering sand and cement
with the appropriate plasticisers, produces a mortar with <.40 water cement ratio
which can be pumped through a pipe and sprayed.
3. Using flat sheet reinforcements of expanded metal and flat sheets of reinforcing
meshes is the most efficient way of handling the reinforcing materials and placing
them into a plastic (wet) mortar.
410
4. There is no need to roll out and flatten rolls of small mesh materials and there is no
need to wire together the small meshes to an armature.
5. There is no need to work by hand plastering to force mortar into an armature and
worry about the correct penetration and total encapsulation of the wires.
6. The spray and laminating technology can be used for both vertical walls and hori-
zontal casting.
7. The spraying of mortar produces excellent compaction against the mould surface.
Thus replicating the features of that mould surface when demoulded.
8. The spraying process drives out any excess moisture and therefore delivers an ex-
cellent low water-cement ratio with good compaction at the mould surface. This al-
so provides the impermeable ‘gel coat’.
9. Surface cracking is controlled by two layers of expanded metal which is set next to
the gel coating. These three layers - of gel coat plus expanded metals (1) and ex-
panded metals (2) provides an impermeable barrier with crack control and yet
might also be regarded as a ‘sacrificial layer’ protecting the subsequent layers of re-
inforcements.
10. IMPORTANT – Expanded Metal meshes dissipate surface cracking whereas round
wire meshes of a square welded mesh concentrate cracks along the wires.
11. IMPORTANT - These three layers of super quality ferrocement - forming the first
8 -10mm of the section are the essential barrier to chloride penetration and crack
control.
IF…repeat…IF…. the surface of ‘gel coat’ is damaged and corrosion starts to
form on the bare metal where expanded metal is sheared in manufacturing, the
oxidation is minimal and is restricted to the cut ends of the sheared metal.
12. The sheet metal used to manufacture the expanded metal is galvanised in the first
place!
13. “The process of oxidation (rusting) of the fine wire ends of expanded metal sheet-
ing is never stronger than the surrounding matrix of quality mortar so ‘spalling’ as
known in normal poor concrete work, does not occur in this type of quality ferro-
cement. In fact, the slight expansion of fine wire oxidation actually ‘tightens’ the
matrix in the area and further oxidation is prevented. Thus the ‘gellcoat’ and ex-
panded metal crack control form a perfect protective layer.
14. IMPORTANT - The placing of ‘steel reinforcement’ closer to the outside of any
element cross section and away from the ‘neutral’ axis provides a more efficient
use of steel.
15. IMPORTANT - We have found that 2 to 3 mm of cover of sprayed mortar is ade-
quate to prevent moisture penetration and two layers of expanded metal can be re-
garded as a ‘sacrificial layer’ protecting the more important reinforcements subse-
quently added into the section.
The concrete industry appears not to recognise the very simple ferrocement truths de-
scribed here as it creates rules and regulations requiring cover of concrete to reinforcing
bars. Perhaps the industry has ignored the benefits of ferrocement in their self-interest to
sell the maximum amount of sand cement and aggregate when in fact a more careful use of
a better quality material (ferrocement) has been ignored, or possibly forgotten for the
411
commercial interests of the sand, cement aggregate industry with their desire to sell more
volume / tonnage of material? It is hard for conventional concrete workers and engineers to
understand and realise that a cover of 2 – 3 mm of quality sprayed concrete as opposed to a
more usual 20 – 30 mm – of poured and vibrated concrete is impermeable in the marine
environment.
FUTURE FERRO
In the Author’s opinion, there is a need to re-educate our concrete engineers in the history
of ferrocement. This wonderful material which, in recent years, has slipped from the engi-
neering curriculums. The future use of sprayed and laminated ferrocement now requires
the application of new innovations to take advantage of recent improvements in cement
technologies and combine the new knowledge to improve the old. Mechanically mixing
and spraying is the most efficient method to produce, deliver and apply mortar. Spraying of
mortar to vertical and horizontal surfaces creates the 2 - 3 mm of high quality ‘gel coating’
which produces an impervious / protective layer. The close proximity of two layers of ex-
panded metal mesh provides a superior ‘cracking control ‘close to that surface. Horizontal
casting is the most economical way to handle mesh materials and embed them into the
mortar. It is also the most cost effective way to cast ferrocement plates.In some jointing
applications, it is preferable to inject mortar into a joint or into a mould containing the re-
quired amount of reinforcements. Use of rolled up wire meshes and the multiple fastening
of meshes to form an armature is time consuming, laborious and therefore an expensive
operation which is avoided by horizontal laminated construction.
Horizontally casted panels can be joined by ‘concrete welding’ to form cell structures.
Stressing cables can be passed through the cell structures with walls of ferrocement plates
which act as permanent shuttering. With the development of self-compacting mortar, super
strength grouts and with the application of some innovative fibrous materials together with
appropriate plasticisers and new stressing methods, it means that today we can produce a
‘super ferrocement’ more similar to a steel plate than ever before, but with the ’ low / no ‘
qualities of concrete. Methods of joining panels of horizontally cast ferrocement and the
concrete welding of joints to produce cellular structures and the use of these panels as
permanent shuttering are the future for ferrocement applications. The ultimate concrete
construction material. The laborious tasks of constructing an armature into which a hand
plastered mortar is applied may still be practical in developing countries with inexpensive
labour but is not commercially practical in developed countries. Trafalgar Marine Tech-
nology Ltd is developing systems and Intellectual Property in new, innovative and disrup-
tive construction systems of laminated and stressed ferrocement plates for permanent shut-
tering. Particular applications are for marine renewable energy structures - Wind Turbine
Gravity Structures and Turbine Towers.
Trafalgar Marine Technology Ltd has patents pending relating to recovery and replication
of historic ships. These systems can also be applied to modern concrete ship building. Be-
tween 2000 and 2014 the author proposed the recovery of the first composite clipper ship
’’City of Adelaide’ which was a ‘hulk’ at the Scottish Maritime Museum in Irvine, Scot-
land. It was proposed to return the ship to Sunderland UK, where she was original built
(1864). Where she would become the focus of a maritime museum. It was proposed to
sheath the ship in a ferrocement cocoon to make the ship waterproof and strong enough to
float and be towed away. After delivery to a dry dock, the ferrocement cocoon could be
412
removed sequentially and re-assembled to make a mould from which an exact replica ship
in Ferrocement could have been constructed. Unfortunately the project failed and the old-
est clipper ship in the world was transported from Scotland to Adelaide, Australia.
REFERENCES
1. Naaman, Antoine E. Ferrocement and Laminated Cementatious Composites. 1st Ed. ISBN: 0-
9674939-0-0. (2000).
2. Iorns. M. E. “Shotcrete Laminating – An Improved Ferrocement Construction Technique”, Ferro 1,
Nervi International Symposium on Ferrocement. July 22-24, 1981.
3. A C I Committee 549 “State-of-the-Art Report on ferrocement” 549-R97.
4. Pemberton J. M. “DeMenil Museum Project and Potential Applications of Ferrocement /concrete
laminates.”, Ferro 5 – International Symposium on Ferrocement, Manchester, 6-9 September 1994.
5. Pemberton J. M. “Ferrocement - an insight and a review – So what is new”, Sixth International
Symposium on Ferrocement, University of Michigan, Ann Arbor, June 1998.
See www.snfcc.org for more information on the SNFCC Athens Solar Roof, www.traf-mar-tech.co.uk, pul-
lan-engineering.uk
413
414
CHAPTER 9
PRODUCTION
TECHNOLOGIES
415
416
LOCALLY ADAPTED BIAXIAL WARP KNITTED TEXTILES AS

REINFORCEMENT OF FOLDED CONCRETE ELEMENTS
A. Koch1, C. Kerschl2, T. Gries1 and W. Brameshuber2, 1Institut für Textiltechnik (ITA) of

RWTH Aachen University (Germany), 2Institut für Bauforschung (ibac) of RWTH Aachen
University (Germany)
Abstract: The use of textile reinforcements for building applications has been on the rise
since some years now (e.g. as plaster or floor reinforcement). One of the most sustainable
developments is textile reinforced concrete (TRC). TRC can save up to 80 % of concrete as
compared to a steel reinforced structure for similar applications [1]. A biaxial warp knitted
structure, made with AR-glass or carbon fibres, increases the mechanical properties of
concrete components significantly. On the basis of its non-corrosive property, building
components with TRC can be realized with very thin cross-sections and with a high bear-
ing capacity. A significant advantage of TRC is the formability of the textile reinforce-
ment. Curved TRC-elements of high quality finish are possible due to the inherent drapa-
bility of the textile in comparison to steel reinforcement. The minimum radius of the curve
decides the required level of drapability of the textile. The drapability of a biaxial warp
knitted textile is mainly depending on the stitching pattern (for e.g. pillar, tricot or plain).
Textiles which use the “pillar” stitching pattern have a higher stiffness compared to a
“plain” knitted textile. In the course of a public funded project we want to realize locally
adapted textiles which combine two different levels of drapability in one single fabric. The
overarching aim is generating the possibility to manufacture adapted textiles for special
folded elements. A major challenge is to classify the possible combinations of stitching
patterns to determine their influence on the drapability of the textiles on basis of a cantile-
ver test. The public funded project “ConcreteFold” started in June, 2014 and has a 2 years
research period. This paper will present the project scope as well as the first results.
INTRODUCTION
Textile reinforced concrete is a growing research topic since the last 20 years in Germany
(especially in Aachen and Dresden). The basic research was done during two collaborative
research projects funded by the DFG (German Research Foundation) from 1999 until 2011
at the RWTH Aachen University and TU Dresden. The technical University of Dresden
had the leading subject “Strengthening steel reinforced concrete with textile reinforced
concrete” and the RWTH Aachen University conducted research on analyzing the behavior
of new precast TRC-elements (especially for façade applications). TRC has many ad-
vantages in terms of ecology, economy, durability and new possibilities of design com-
pared to steel reinforced concrete. The most important advantages of TRC are [2, 3]:
Corrosion free (especially the textile reinforcement)
High drapability of the textile reinforcement compared to steel reinforcement (new
scope for design)
417
Realization of very thin and structurally strong elements (5-50 mm)

High surface quality and high durability due to using a fine grain concrete matrix
Since 1999 many building applications are realized by using TRC components. These de-
monstrator buildings allows realization of using TRC as a building material and provide
the opportunity to get a long-term evaluation. RWTH University emphasized this philoso-
phy by constructing their own buildings (e.g. facades of institute buildings and auditorium
buildings) as the first TRC-demonstrator buildings. The following two pictures show the
TRC façade of the ITA machine hall (Fig. 1, left) and a free-formed TRC deck chair (Fig.
1, right).
Fig. 1: Façade of the ITA machine hall (left) and a TRC deck chair (right)
The drapability of biaxial warp knitted textiles is one of the advantages and enables a new
freedom of concrete designs with an integrated reinforcement. Therefore, the textile design
is the key between the textile structure and its draping behavior. The textile design can be
adapted by changing the following factors:
Yarn material (warp and weft direction, e.g. carbon or AR-glass)
Yarn titer (warp and weft direction, e.g. 800 - 3300 tex)
Mesh size (warp and weft direction, e.g. 5 - 30 mm)
Stitching pattern (pillar, tricot, plain)
In this case the drapability is mainly influenced by the stitching pattern. The standard
stitching pattern regarding biaxial warp knitted textiles are pillar, tricot and plain. The fol-
lowing overview shows stitching pattern in a technical sketch and a photograph (Fig. 2,
left) and shows the connection to the mechanical properties of the textile (Fig. 2, right).
418
Pillar Tricot Plain
Stitching pattern Pillar Tricot Plain

Cross-section geometry
of the warp rovings
Bond properties with
respect to the concrete low medium good
matrix
Bending stiffness high medium low
Resistance to
high medium low
1 cm 1 cm 1 cm
displacement
Fig. 2: Standard stitching pattern and its properties
The high influence of the stitching patterns is the result of the compaction of rovings under
different stitching patterns. The pillar stitching pattern extremely compacts the warp yarn
and generates a circular cross-section geometry of the yarn. In comparison, the plain stitch-
ing pattern only keeps the yarn together and doesn`t compact it. The degree of compacting
is directly proportional to the bending stiffness of the textile.
The textile design enables a desired textile drapability behavior on destined textile parts
due to the influence of the above mentioned factors. This knowledge led to start a research
project in June 2014 to study the possibility of locally adapted textiles for folded concrete
elements. The focus is on the manufacturing of folded concrete elements with the help of a
new folding tool to fold TRC plates whilst the concrete is in a green stage.
PROJECT SCOPE
The research project “ConcreteFold” funded by the federal ministry of economics and en-
ergy (BMWE) started in June 2014 in cooperation with RWTH Aachen University (ITA,
ibac, LfP), Florack Bauunternehmung GmbH, Germany, ingema t+h ingenieurgesellschaft
mbH, Germany and W+S bau-instandsetzung GmbH, Germany. Different partners with
different expertise work on a manufacturing process for precast folded TRC-elements. The
difficulty lies in creating proper edges of the folded TRC element. The textile reinforce-
ment should stay in a defined position in the inside of the edges and not shift to the con-
crete surface.
Two solutions are proposed by the project consortium to answer this research question.
One solution works with an overlap of the textile reinforcement to enable a displacement
during the folding process. The second and more difficult solution is to develop a locally
adapted textile which enables a defined deformation of the textile during the folding pro-
cess. The following two pictures show the mentioned solutions as schematic sketch.
419
Concrete matrix
Betonmatrix Concrete matrix
Betonmatrix
Standard reinforcement
Normalbewehrung Standard reinforcement
Normalbewehrung
Locally adapted
Konfektionierte reinforcement
Bewehrung Locally adapted
Konfektionierte reinforcement
Bewehrung
Fig. 3: Two different project approaches
PRODUCTION OF LOCALLY ADAPTED BIAXIAL WARP KNITTED TEXTILE
Biaxial textiles are based on two fibre systems (warp and weft). The warp yarns will be
placed on top of the weft yarns and bonded by using a thin knitting yarn (usually PES 167
dtex at ITA). The knitting yarn can be inserted in different ways (stitching pattern) by us-
ing the appropriate steel disk guidance mechanism or an electronic control system. These
components control the guide bar motion of the machine, which are used for the knitting
fibre placement. ITA uses a Karl Mayer Malimo machine to produce the biaxial warp knit-
ted textiles with a working width of up to 50 inches.
During the preliminary tests, ITA stopped the textiles production of a pillar bond textile
and changed the stitching pattern to a plain stitching pattern. On the basis of a much higher
knitting yarn consumption, the knitting yarn supply was also configured. After a textile
production length of approx. 4-5 cm, the stitching pattern was changed again (from plain to
pillar). The following picture (Fig. 4) shows the first textile demonstrator and illustrates the
main project idea.
Weft direction
Section A
(stiff)
Warp direction
Section B
(flexible)
Section A
(stiff)
Fig. 4: First demonstrator of a locally adapted biaxial warp knitted textile
A first manual handling test easily showed the difference of the textile bending stiffness
only by changing the stitching pattern on a local section. Fig. 5 shows the textile structure
with the pillar stitching pattern (standard, left) and the locally adapted textile (right). The
bending stiffness is very low on the adapted local section.
420
60 mm 60 mm
Fig. 5: First manual handling test (Comparison of the two textile structures)
In a second step, ITA placed a locally adapted textile on top of a wooden folded form to
perceive the bending behavior of the textile. The developed textile nicely adapts to the
shape of the wooden form. Fig. 6 shows the innovative locally adapted textile in an inner
and outer formed edge.
100 mm 100 mm
Fig. 6: The locally adapted textile in an inner and outer form
During the research project ITA will measure this change of the bending stiffness and test
further options for local adaption. The bending stiffness can be measured by doing a canti-
lever test with the standard DIN 53362 “Testing of plastics films and textile fabrics (ex-
cluding nonwovens), coated or not coated fabrics - Determination of stiffness in bending -
Method according to Cantilever”.
The results of the cantilever test can lead to an overview of locally adapted textiles and
their bending properties. This overview can be a formal tool to choose the best textile for a
set geometry of a folded TRC element.
FOLDING PROCESS OF THE TRC-ELEMENTS
After the development of a locally adapted textile, a flexible formwork is required to man-
ufacture folded TRC elements in a green stage of the concrete matrix. Complex shaped
concrete elements can be realized by using three dimensional formworks and with the use
of a short fibre reinforced concrete matrix. But the manufacturing process of these design
elements is complex and extremely time-consuming. Therefore, the project consortium will
develop a flexible folding tool, which allows folding a flat concrete plate in a green stage.
Fig. 7 shows a drawing of the first construction concept (left) and a three dimensional pic-
ture (right).
421
formwork
abgestufte Formgrößen
TRC-element
formwork
connection rod
durchgesteckte Stangen
joint perforated plate

Lochblech
Fig. 7: Flexible formwork for different folded TRC elements
The size of the folding tool is set on 0.5 to 1.5 m. The numbers of edges and the maximum
angle have to be defined for the first prototype of the formwork. The engineering office
ingema t+h ingenieurgesellschaft mbH will do the first construction details in December
2014.
RESULTS OF THE FOLDING TESTS
In order to find a suitable concrete mixture for the folding process the fresh concrete be-
havior was studied at the Institute of Building Materials Research (ibac) using a folding
tool and different mixture designs. The mixtures should have a high green strength to be
able to remove the first formwork after a few minutes and a high ductility to bear the
movement of the flexible formwork. CEM II/A-LL 42.5 R was used with metakaolin as
binder with a water to binder ratio varying between 0.40 and 0.42. The maximum grain
size was 1 mm. As an additive a PCE based super-plasticiser was used.
A simple folding tool allowing an angle between 0 and ± 90° (Fig. 8) was built to test the
different mixtures in regard to their tendency of cracking during the folding process.
Fig. 8: Simple folding tool for angles between +0 ° and +90°
The concrete was cast in two layers into a formwork with dimensions of 250 x 1000 x 15
mm³. The locally adapted textile was placed between the two layers during the casting pro-
cess. After a certain setting time the specimen was moved with the help of the Polyeth-
ylene foil that had been placed in the formwork prior to casting into the folding tool. One
side of the folding tool is tilted up (0 to +90°) or down (0 to -90°). Afterwards the speci-
men hardens on the folding tool for 24 hours.
422
The experiments have shown that angles between -15° and +90° can be achieved without
the development of cracks in the stretched side of the specimen. Angles between -15° and -
90° cannot be achieved without cracks. It has been shown that with the application of an
additional Polyethylene foil on the surface the crack formation can be avoided. The adhe-
sion forces between Polyethylene foil and fresh concrete apparently bridge the cracks suc-
cessfully. A further advantage of the Polyethylene foil is improved exposed concrete quali-
ty and reduced evaporation.
In order to check the position of the textile the specimen was cut perpendicular to the fold-
ing edge. In Fig. 9 the position of the locally adapted textile is shown in a specimen folded
with an angle of +90°.
Fig. 9: Position of the locally adapted textile in the concrete specimen
So far, only single folded edges have been investigated. In this case it can be seen that the
folding process did not lead to a displacement of the textile. In order to prove this fact for
specimens with more than one edge a new folding tool will be built in association with
ingema t+h ingenieurgesellschaft mbH.
CONCLUSIONS AND FURTHER OUTLOOK
The research project “ConcreteFold” offers the possibility to develop a new and simple
manufacturing process for folded TRC elements. Besides the development of a flexible
formwork to enable this folding process, the development of a locally adapted textile plays
a major role of the project. A new type of textiles can be produced with a local adaption
regarding their bending stiffness. This adaption allows the use of these textiles for folded
TRC elements, because of their adaptability on set folded edges. The project will do the
folding tests with the help of a new developed folding tool in 2015. The first demonstrators
of the folded TRC elements will be analyzed regarding the defined textile placement. The
difference between a standard textile and the locally adapted textiles should demonstrate a
better textile placement in the inside of a folded TRC element. At the end of the project a
wall will be built to demonstrator a combination of many folded TRC elements.
The research plans to take the first steps towards realizing the true design potential of TRC
as a material. However, the material and formwork research is only the first step towards
institutionalizing the use of TRC in complex folded component manufacturing. The per-
formance limits of such folded elements require to be tested according to the norms. A
quality testing procedure to ensure proper placement of textiles within the advanced form-
423
work ensures higher level of replicability. In addition, non-destructive testing (NDT) to see
the exact positioning of textiles within the hardened concrete component will help illustrate
the efficiency of the newly developed textile material as well as the formwork.
ACKNOWLEDGEMENT
The authors would like to acknowledge the financial support from

the “ConcreteFold”-project funded by the BMWE for carrying out
the project. The authors would like to thank Florack Bauun-
ternehmung GmbH, Germany, ingema t+h ingenieurgesellschaft
mbH, Germay, W+S bau-instandsetzung GmbH, Germany and
Lehrstuhl für Plastik (LfP) der RWTH Aachen University for the
good cooperation.
PROJECT CONSORTIUM
REFERENCES
[1] Gries, T., Tomoscheit, S., EU Life06 Insu-Shell: Environmentally Friendly Facade Elements made
of thermal insulated Textile reinforced concrete, Bund Deutscher Architekten, Ortsverein Aachen,
Aachen 30.10.2009.
[2] Janetzko, S., Methodik zur Gestaltung von Bewehrungssystemen für textilbewehrten Beton, Ph. D.
Thesis, Institut für Textiltechnik (ITA) der RWTH Aachen University, Germany, 2013.
[3] Arbeits- und Ergebnisbericht / Sonderforschungsbereich 532 "Textilbewehrter Beton - Grundlagen
für die Entwicklung einer neuartigen Technologie", Aachen, 2002.
[4] DIN 53362, Bestimmung der Biegesteifigkeit - Verfahren nach Cantilever, Oktober 2003
424
EXTRUSION OF TEXTILE REINFORCED CONCRETE
Wolfgang Brameshuber, Institute of Building Materials Research, Aachen University
Abstract: Textile reinforced concrete has been developed in the last 20 years to a modern
and quite broadly applicable composite material. One of the major issues of new materials
is the technique to produce building members in a professional manner. Therefore different
production technologies have been developed in the past, e. g. normal casting, laminating
technique, spraying and laminating, spinning, and, as a good example for a continuous
process, the extrusion. There are different issues which have to be solved if one wants to
extrude members of textile reinforced concrete. Firstly the fresh concrete properties must
meet this process. This includes special mixing processes and mix compositions. The flow
of concrete has to be studied depending on the geometry of the cross-section of the mem-
ber. And last but not least a system has to be implemented which allows the guidance of
the textile into the flow of concrete without disturbing the inner structure and damaging the
structure of the textile. In this paper the development of mixture design, mixing procedure
and the mouth of the extrusion machine will be described and some experimental results in
tensile tests are shown.
INTRODUCTION
A lot of publications dealing with the bond, tensile and shear properties, the dynamic be-
haviour and the load-carrying capacity under sustained loads of textile reinforced concrete,
the modelling of the mechanical behaviour, the durability have been presented in the last
years, e. g. /Bra03, Bra06, Cur13a, Cur13b, Rem15/. This shows that textile reinforced
concrete has been investigated in a very scientific way. Applications of textile reinforced
concrete still are in the status of single prototypes, e. g. /Cur08, Heg10, Sch13/. This might
be caused by the fact that production methods for members of textile reinforced concrete
still are in the status of highly manual works, which increases the cost of this building ma-
terial quite much. Therefore different methods have to be developed to economically pro-
duce members of textile reinforced concrete /Bra08/. One system is the laminating tech-
nique using high flowable matrices and a layer-wise casting of concrete and fabrics. In this
case three-dimensional cross-sections are not producible, and therefore this method often is
used for façade elements /Heg12a, Heg12b/. For the production of three-dimensional
building members the usual casting method with high flowable concretes is at that time the
most common option. This means the reinforcement has to be made with stiff elements, e.
g. epoxy-resin impregnated textiles, e. g. /Rem13/. For maintenance and strengthening the
spraying method in combination with laminating is the most suitable working process, e. g.
/Mot07, Mot08/. Basement sealings against ground water as well as strips for strengthening
existing structures are examples for the spraying and laminating method. For pipes or poles
the spinning method has been approved. In this case it has been shown that this production
method in connection with simple spacer systems is very well applicable to circle-shaped
cross-sections /Har10, Bra08/. However the thin TRC-elements might be produced eco-
nomically by a continuous method, e. g. pultrusion /Pel08/ or extrusion /Kra14/. In this
425
contribution the extrusion method will be described, and first experiences for the produc-
tion of TRC with extrusion are shown. For two-dimensional elements the equipment of a
common extruder developed for model productions in the laboratory has been modified for
guiding fabrics into the flow of the matrix.
DESCRIPTION OF EXTRUDER
The extruder used is a special equipped laboratory machine for optimising mixture designs
and mixing processes. In Figure 1 the machine is shown. It consists of a pre-worm gear and
a main worm gear, taking over the function of transport and mixing the concrete. The main
worm gear ends in a nozzle, which is installed behind the pressure packer.
pre-worm gear
funnel
main worm gear

nozzle
pressure head
pressure unit
Figure 1: Extruder, opened (left) and completely mounted (right)
The laboratory extruder allows to measure the velocity of extrusion, temperature in the
pressure unit, the absolute value of vacuum, if applied, and the press power at the pressure
head, the motor current and the revolution speed of the main worm gear and the pre-worm
gear. These parameters have been documented for each mixture depending on the mixing
regime and the adjustment of the extruder. Usually extrusion and mixing process cannot be
separated from each other, because the transport in the chambers of the worm gears addi-
tionally is a mixing of the concrete and leads to a change of the fresh concrete properties.
The common mixing consists of a pre-mixing and a immediately following dispersing pro-
cess, which is continued by the dispersing procedure in the chambers of the worm gears.
Some aspects of the mixing process are described in the following.
MIXING OF CONCRETE AND EXTRUSION
Instead of mixing and dispersing another option has been chosen to prepare the concrete
mix for extrusion: the high-speed intensive mixing equipment. An Eirich intensive mixer
with variable revolution speed and the possibility of measuring the main parameters of the
426
mixing process, means revolution speed, current uptake in dependence of time and temper-
ature during the mixing process has been chosen. Two basic mixes have been designed for
the first steps of investigating the rheological properties for concretes suitable for the ex-
trusion process. They are given in Table 1.
Table 1: Mix design of basic mixtures
base material unit mixture type 1 mixture type 2
1 2 3 4
CEM I 52.5 or CEM I 42.5 (c) 700 781
fly ash HKV (f1) - 167
fly ash SWF (f2) 210 -

kg/m³
silica fume (s) 70 167
quarzitic aggregate 0–0.6 mm 970 671
water (w) 256 314
methyl cellulose % by mass of c 1.5 1.5
w/c 0.37 0.40

-
w/ceq = w / (c+0.4·f+s) 0.30 0.31
Mixture 1 is a consistent development of mixture 2, which is the basic mixture for textile
reinforced concretes developed at Aachen University /Bra06/. Both mixes contain methyl
cellulose to improve the possibility of extrusion, but for normal cast concretes superplasti-
cizers have to be added.
In Figure 2 different mixing regimes have been tested , B1 and B5 with a nearly constant
revolution time of 500/minute for dry mixing and wet mixing, and B4 and B8 with a revo-
lution time of 500/minute within the first 1.5 minutes (dry mixing) and 800/minute after a
short interruption for adding water, respectively. The difference between both mixtures B1
und B5 and mixtures B4 and B8 is the type of cement, for B1 und B4 CEM I 42.5 and for
B5 and B8 CEM I 52.5. All mixes are based on mixture type 1, see Table 1.
427
Figure 2: Influence of cement type and mixing regime on the flow parameter of concretes for extrusion
In this example the revolution of the mixing device clearly dominates the torque as well as
the current uptake. Indirectly these parameters are an indication of the flow parameters of
the concrete. Due to the high consistency of concretes for extrusion no direct method to
measure the viscosity and the flow limit exists. Some methods like penetration needles or
slab viscosity measurement devices had been tested but found not to be applicable. Abso-
lutely the torque is higher for a revolution of 800/minute instead of 500/minute, but the
maximum is considerably earlier. An increase of torque is similar to a decrease of the con-
crete viscosity. A constant revolution means that the dispersion is finished. A lot of mix-
ture designs with varying contents of methyl cellulose have been tested for finding the op-
timum mix in connection with the dispersing procedure in the extrusion process but cannot
be described within this contribution. Actuallly, the mix design suitable for the extrusion
process has to be found for the equipment of mixing and extruding in particular cases. Fig-
428
ure 3 shows examples of extrusion results choosing mix designs which do not fulfill the
rheological properties of the fresh concrete.
Figure 3: Results of extrusion with unsuitable concretes, no vaccum applied (left), vacuum applied (right)
In Figure 4 the current uptake by the motor of the main worm gear is given in dependence
of the current uptake in the intensive mixer unit. It is obvious that more parameters for
approving the extrusion process are needed. It shows however that all mixtures, which
don´t exceed a certain limit of current uptake in the main unit in principle fulfill the condi-
tions for extrusion.
Figure 4: Current uptake of the main extrusion worm gear vs. that of the mixing unit
429
EXTRUDED TEXTILE REINFORCED CONCRETE
The major task for embedding textiles during the extrusion process is to control the con-
crete flow. Figure 5 shows the extension of the pressure head with the textile feed channel.
In comparison to pure concrete or fiber reinforced matrices the flow of the concrete ap-
proaching this feed element has to be changed by narrowing elements. Otherwise the con-
crete which has to flow through the meshes of the textiles is too slow. This leads to unsuit-
able results of extruded specimen, see as an example Figure 6. At the moment coated tex-
tiles compared to non-coated textiles have given the best results. In this case the coating
consisted of styrol-butadien-acrylate. Tests using epoxy-resin-coated textiles are scheduled
for the next step. The chosen feeding device for the textiles does not allow the use of un-
coated types.
Figure 5: Extrusion with feeding device for textiles
Figure 6: Lateral adjustment of textiles when concrete flow is not adapted
430
TENSILE TESTS
Some preliminary tensile tests on strips of extruded textile reinforced concrete have been
carried out to check the influence of the extrusion process on the tensile behaviour of the
composite. Therefore specimens have been casted in a traditional way using the laminating
method. The textile used was identical, but the concrete had to be modified in that way that
the methyl cellulose has been replaced by a superplasticizer. This makes a direct compari-
son somewhat difficult due to the different changed rheological properties of the concretes.
These might mainly influence the bond, however there is no other possibility to compare.
The compressive strength of the matrix with methyl cellulose is resembling to that with
superplasticizer, reaching about 100 MPa after 28 days.
The tensile tests had been carried out according to the RILEM recommendation for tensile
tests on TRC /Rec14/. In Figure 6 a typical result of the tensile tests comparing laminated,
casted specimen and extruded specimen is shown. From these preliminary tests it can be
concluded, that the load and deformation behaviour of extruded textile reinforced concrete
is comparable to the common one. The scatter of results – the dashed lines show the ex-
treme curves for the extruded TRC – is larger for the extruded concrete concerning the
deformation. The tensile strength of the composite is somewhat higher for the extruded
TRC. The mechanical properties after 2 days differ slightly between the mixture for lami-
nating and extrusion (see Fig. 6 compressive strength and fluxural strength).
Figure 6: Comparison of tensile tests with laminated and extruded textile reinforced concrete
SUMMARY AND OUTLOOK
Mixing procedure and dispersion due to the transport by the worm gear in the extruder
strongly influence the design concept of the concrete mixture suitable to extrude. The ex-
trusion process is quite sensitive regarding the rheological properties of the highly viscous
cementitious paste, which at the moment contains a certain amount of methyl cellulose.
431
The major task for feeding textiles into the concrete flow is to influence this by varying the
cross-section in the pressure head depending on the consistency of the fresh concrete. Sim-
ulating the flow of the concrete using numerical tools will be carried out as a next step to
avoid too much empirical experimental work. However the numerical simulation of highly
filled materials including tribology effects due to the walls hasn´t been solved satisfactorily
until today. On the experimental side a new nozzle has been designed for three-
dimensional cross-sections.
ACKNOWLEDGEMENT
The author greatly acknowledges the financial support by the German research foundation
within the transfer project T04 “extrusion of TRC”.
LITERATURE
/Bra03/ Brameshuber, W. ; Koster, M. ; Hegger, J. ; Voss, S. ; Gries, T. ; Barle, M. ; Reihardt, H.-W. ;
Krüger, M.: Textile Reinforced Concrete (TRC) for Integrated Formworks. Farmington Hill :
American Concrete Institute, ACI SP-224, 2004. - In: Thin Reinforced Cement-Based Products
and Construction Systems, Vancouver, Spring 2003, (Dubey, A. (Ed.)), S. 45-53
/Bra06/ Brameshuber, W.; RILEM; RILEM TC 201-TRC: Textile Reinforced Concrete. State-of-the-Art
Report of RILEM Technical Committee 201-TRC. Bagneux, RILEM, 2006. - Report 36 ISBN 2-
912143-99-3
/Cur08/ Weiland, S. ; Ortlepp, R. ; Hauptenbuchner, B. ; Curbach, M.: Textile Reinforced Concrete Flex-
ural Strengthening jof RC-Structures - Part 2: Application on a Concrete Shell. Farmington Hill :
American Concrete Institute, ACI SP-251, 2008. - In: Design and Applications of Textile Rein-
forced Concrete, ACI Fall Convention 2007, Fajardo, Puerto Rico, 14.-18.10.2007, (Aldea, C.-M.
(Ed.)), 18 Seiten
/Cur13/ Curbach, M.: Concrete light - Possibilities and vision. Structural Concrete 14 (2013) 2, S. 87-88
/Heg12a/ Hegger, Josef; Kulas, Christian; Horstmann, Michael: Spatial textile reinforcement structures for
ventilated and sandwich facade elements. In: Advances in Structural Engineering - An Internation-
al Journal, Vol. 15, Number 4, April 2012
/Heg12b/ Hegger, Josef; Kulas, Christian; Horstmann, Michael: Spatial textile reinforcement structures for
ventilated and sandwich facade elements. In: Advances in Structural Engineering - An Internation-
al Journal, Vol. 15, Number 4, April 2012
/Pel08/ Peled, A. ; Mobasher, B.: Effect of Processing on Mechanical Properties of Textile-Reinforced
Concrete. Farmington Hills, Mich. : American Concrete Institute, ACI SP-250, 2008. - In: Textile-
Reinforced Concrete. ACI Fall 2005 Convention, Kansas City, (Dubey, A. (Ed.)), 14 Seiten
/Rem15/ Rempel, S.; Will, N.; Hegger, J.; Beul, P.: Filigrane Bauwerke aus Textilbeton. Leistungsfähigkeit
und Anwendungspotential des innovativen Verbundwerkstoffs. In: Beton- und Stahlbetonbau 110
(2015), Special, S. 83-93
/Rec15/ RILEM Recommendation “Uniaxial tensile test –Test method to determine the load bearing be-
havior of tensile specimens made of textile reinforced concrete”, to be published in Materials &
Structures, 2015
/Sch13/ Schneider, Maike; Scholzen, Alexander; Chudoba, Rostislav: Filigrane Schalenkonstruktion aus
Textilbeton In: [Textilbeton - in Theorie und Praxis : 5. Anwendertagung Textilbeton ; 24./25.
September 2013, Friedrichshafen]. – TUDALIT. Leichter bauen - Zukunft formen. - Dresden :
TUDALIT. - 5 (2013) 9, S./Art.: 8-8
432
CHAPTER 10
STRENGTHENING
433
434
TEXTILE-REINFORCED MORTAR (TRM) AS RETROFITTING

MATERIAL OF MASONRY-INFILLED RC FRAMES
Lampros N. Koutas1, Stathis N. Bousias2 and Thanasis C. Triantafillou3 , 1 Research Associate,

Dept. of Civil Engrg., Univ. of Nottingham, University Park, Nottingham, NG7 2RD, UK, mail:
Lampros.Koutas@nottingham.ac.uk, 2 Associate Professor, Dept. of Civil Engrg., Univ. of Patras,
Patras, GR-26500, Greece, mail: sbousias@upatras.gr, 3 Professor, Dept. of Civil Engrg., Univ. of
Patras, Patras, GR-26500, Greece, mail: ttriant@upatras.gr
Abstract: The contribution of masonry infills to the seismic resistance of existing rein-
forced concrete structures is significant, both before separation of the infill from the sur-
rounding frame occurs and during large cycles of imposed deformations near collapse.
Strengthening of this type of structures usually aims at increasing the resistance and de-
formation capacity of the frame itself, setting aside the contribution of infills as a source of
strength reserve. An alternative route at improving the performance of existing building
structures, while avoiding the drawbacks of the approach above, is to convert masonry
infilling to a more reliable source of resistance by guaranteeing its contribution over the
whole spectrum of structural response. This is sought and examined experimentally in the
present study, by employing the very promising technique of textile-reinforced mortars
(TRM). The application of TRM is examined in this study on nearly full-scale as-built and
retrofitted 3-storey masonry infilled frames, subjected to cyclic loading. The results of the
tests are presented and discussed along with the presentation of an analytical model ac-
counting for the TRM contribution to the global response. Finally, the results of the numer-
ical analyses, which were carried out after the implementation of the model to the open-
source software OpenSees, are being compared to the experimental results.
INTRODUCTION
Textile-reinforced mortar (TRM) is an advanced composite material comprising multi-

directional textiles (typically two-directional) made of high-strength fibers in combination
with inorganic matrices, such as cement-based mortars. This material, which can also be
found in the literature as textile-reinforced concrete (TRC), has attracted the interest of
research community since last decade. It can be used either for fabrication of precast ele-
ments (e.g. [1],[2]) or for structural retrofitting of reinforced concrete (RC) (e.g. [3]-[7])
and masonry elements (e.g. [8]-[10]). In most of the cases of structural retrofitting the use
of TRM as externally bonded reinforcement has been proved to be effective and promising
for future applications. In this study, in the framework of a broad experimental campaign
which is combined with an analytical study, TRM is used for the first time as retrofitting
material in the case of masonry-infilled RC frames. For this purpose two nearly full-scale
as-built and retrofitted 3-storey masonry infilled frames were tested under cyclic loading,
whereas an analytical model for the behavior of the TRM layers was developed and veri-
fied via numerical simulations of the tested infilled frames. The main objective of this
study is the evaluation of the strengthening technique as a means of increasing the lateral
strength and the deformation capacity of the infilled frame.
435
EXPERIMENTAL PROGRAMME
A 2:3 scale model of a 3-storey frame representing a full-height internal bay of an existing
non-ductile building built in Southern Europe in the ‘60s, was selected. Two identical
frames, i.e. SP#1 and SP#2, were tested with the first serving as control specimen (unretro-
fitted frame) and the second being the TRM-retrofitted. As illustrated in Fig. 1a, each sto-
rey of the frame was 2.0 m high by 2.5 m wide (column centerline distance). Columns
were of rectangular cross-section 170x230 mm (the long side parallel to the plane of the
frame), whereas beams were of T-section. The column longitudinal reinforcement com-
prised six 12 mm-diameter deformed bars which were lap-spliced only at the base of the
first story over a 60-bar-diameter splicing length. The transverse reinforcement both for
columns and beams consisted of plain bars with 90-deg end hooks.
(a) (b)
Figure 1: (a) Geometry of the bare frame and (b) test setup.
The infill walls consisted of two individual wythes with an internal gap of 60 mm between
them. Each wall was built by using perforated, fired clay bricks laid with the perforations
running parallel to the unit’s length. Special care was taken to ensure that each wall was
well supported and confined by the surrounding frame. The total wall thickness (including
the internal gap) was equal to the width of the columns (170 mm). The mean compressive
strength of the bricks perpendicular to the perforations was 11.3 MPa while the thickness
of the bed and the head mortar joints was approximately 10 mm. The flexural and com-
pressive strength of the mortar was (average of nine specimens) 2.59 MPa (SP#1)/2.63
MPa (SP#2) and 12.62 MPa (SP#1)/13.31 MPa (SP#2), respectively.
The strengthening scheme for the retrofitted specimen SP#2 was based on the observations
on SP#1 and on results obtained from tests on small-scale specimens conducted by the au-
thors [11]. It comprised strengthening of the infill walls via two-sided application of TRM
layers externally bonded on the faces of the infills and column strengthening to suppress
the eventual shear/lap-splice failure. The infill at first storey received two TRM layers,
whereas the rest infills received one TRM layer. An illustrative sketch of the strengthening
scheme is given in Fig. 2a.
436
The TRM layers were anchored to the surrounding RC members. The textiles were simply
extended to the columns and bonded over the full height, whereas specially designed an-
chors were placed at the bottom slab of the 1st and 2nd storey to provide proper connection
(Fig. 2b). Details on the development of the anchors and their verification can be found in
[11]. At the top of the infills the textiles were extended to the beam surfaces up to the point
where the web of the beam meets the flange. At the 1st and 2nd stories an extra textile patch
was applied at the area along the wall-beam interface, as a means of improving the connec-
tion between them (Fig. 2b). In the case of the front side of the 1st storey, instead of apply-
ing an extra textile patch, textile-based anchors were used. More details can be found in
[12].
For the closed jackets at the columns ends a commercial carbon-fiber textile (with a mesh
size of 10x10 mm, weighting 348 g/m2 and 50-50% weight distribution in the two direc-
tions) was used. The textile used in TRM layers which were applied on the masonry sur-
faces consisted of a polymer-coated E-glass textile (of 25x25 mm mesh size, surface
weight 405 g/m2 and 50-50% weight distribution in the two directions). The mortar used as
binding material was a commercial fiber-reinforced cement-based mortar mixed with re-
dispersible polymers.
Figure 2: (a) Strengthening scheme and (b) various strengthening steps.
Both specimens (control and retrofitted) were subjected to a sequence of quasi-static cycles
with increasing amplitude under an inverted-triangular pattern of forces, thus simulating
typical distribution of the horizontal seismic forces imposed to a multi-storey structure.
The displacement/forces were applied to the specimens at each floor level via three servo-
hydraulic actuators, one per storey (Fig. 1b). The RC base of the frame was fixed to the
laboratory strong floor via prestressing rods while two transverse steel trusses were in-
stalled to prevent out-of-plane displacements of the specimen. A different value of axial
load was applied at each storey so as to better simulate the confining effect the frames have
on the infills. This was achieved by using different sets of prestressing bars per storey. Fi-
nally, a dense instrumentation network (including potentiometers, strain gages, etc.) was
employed.
437
EXPERIMENTAL RESULTS
Table 1 summarizes the main results of the cyclic tests, whereas Figure 3 illustrates the
response of both specimens (SP#1-unretrofitted and SP#2-retrofitted) in terms of base
shear force versus: (a) top drift ratio (Fig. 3a) and (b) 1st storey drift ratio (Fig. 3b). The
strengthening scheme resulted in significant increase of the lateral strength of the infilled
frame as well as in enhanced deformation capacity. In particular, the maximum base shear
was increased by approximately 55% in both directions of loading, while the top drift ratio
at instant of maximum base shear was increased by approximately 56%.
Table 1: Summary of main test results
Max. Base Shear Vmax (kN) IDR(%) at the top IDR(%) at the top
Specimen
+ - @ Vmax @ 0.8Vmax
Unretrofitted (Sp.#1) 264 252 0.43 0.64

Retrofitted (Sp. #2) 407 395 0.67 1.14
Percentage increase 54.2% 56.7% 55.8% 78.1%
As illustrated in Figure 3, the behavior of both specimens was almost linear elastic up to
the first diagonal cracking on the body of the 1st storey’s infill wall. The greater elastic
stiffness of SP#2 compared to SP#1 is attributed to the presence of TRM layers and to the
favorable masonry-concrete connection details. The development of several diagonal
cracks on the body of 1st storey’s infill as well as of few minor cracks on the body of 2nd
storey’s infill, resulted in gradual decrease of the lateral stiffness in both specimens. The
activation of the textile fibers in tension though, in the case of SP#2, enhanced the overall
lateral strength of the infilled frame.
Failure of SP#1 was attributed to the shear failure of one 1st storey’s column just below the
beam-column joint, in combination with the progressive crushing of bricks after the maxi-
mum load was reached. Damage of the 1st storey after testing SP#1 is shown in Fig. 4a. In
the case of the retrofitted infilled frame (SP#2), debonding of TRM from the beam surface
on the back side of the 1st storey and local crushing of the 1st storey’s infill at the two upper
ends neighboring the columns, led to the gradual lateral strength decrease. The connection
details provided in SP#2 were proved to be effective in anchoring the TRM layers to the
surrounding RC members. Only at very large deformation levels of the 1st storey
(IDR>2%) local debonding of the TRM layers was observed as a result of either rupture of
anchors (at the top slab) or insufficient bond length of the textiles bonded to the column
surfaces. It is also important to note that column pre-emptive shear failure of the type ob-
served in SP#1 (control) specimen was suppressed in the retrofitted SP#2. Damage of the
1st storey after the test of SP#2 is illustrated in Fig. 4b.
438
Figure 3: Base shear of control and retrofitted specimens versus: (a) top drift ratio; (b) 1st storey drift ratio.
An interesting aspect of the behavior of TRM was the capacity of the material to sustain
high load even in large shear deformations without experiencing fibers rupture. It was ob-
served that the cementitious mortar (matrix material of TRM) could not sustain the large
drift demands following the post-peak strength cycles, and thus failed in shear and disinte-
grated. The local failure of the mortar enabled the textile to deform independently from the
masonry substrate and carry tensile forces by its ability to distort in shear without experi-
encing rupture of fibers. This constitutes a further advantageous characteristic of the TRM
system: the structural integrity of the textile is maintained, rendering it capable of contain-
ing the masonry infill and reducing the risk of out-of-plane collapse or of becoming dan-
gerous to users. A comprehensive description of the specimens’ behavior and more de-
tailed test results can be found in [12].
Figure 4: Damage of 1st storey after testing: (a) SP#1 and (b) SP#2.
ANALYTICAL MODELING
An analytical model for TRM-strengthened masonry-infills RC frames is introduced,

which is based on the use of a pair of strut and tie elements per infill diagonal. The pre-
sented model was implemented in OpenSees [13] (an open-source software for non-linear
439
structural analysis) and was employed to simulate the response the experimentally tested 3-
strorey frames. The values of the physical parameters characterizing the response of the
infills were derived from standard tests on masonry sub-assemblies, whereas the properties
of the composite material used for strengthening were obtained through tensile tests on
TRM coupons. Numerical analyses were carried out to validate the developed analytical
model. The simplified approach of diagonal struts/ties representing the response of retrofit-
ted masonry infilled RC frames is based on the use of a pair of alternatively activated ele-
ments (a compression-only strut and a tension-only tie) which are placed along each diag-
onal of each portal (Fig. 5a). During a time increment of the dynamic response of the struc-
ture the strut element mimics the behavior of the diagonal that is under compressive stress
state, whereas the tie – which is usually employed only in retrofitted infilled frames - ac-
counts for the behavior along the opposite (tensioned) diagonal, relying on the externally
bonded material to carry the developing tensile forces.
For the strut element the model of Fardis and Panagiotakos (1997) has been adopted. It
consists of a multilinear backbone curve with simple linear hysteresis rules (Fig. 5b). More
details about this model can be found in [14]. To account for the contribution of the exter-
nally bonded TRM layers to the response of the masonry infill under lateral cyclic loading,
an equivalent-tie element model was developed for the panel diagonal under tension.
Based on the macroscopic behavior of the TRM layers in the 3-storey frame test, the axial
force (tensile) versus elongation response of the tie is modelled as bi-linear with simple
linear hysteresis rules (Fig. 5c). The concept of this model is that the fibers of the textile
layers, crossing a given crack pattern in the direction parallel to the strut, develop tensile
forces which are projected in the tie-direction. As a result of the TRM behavior during the
test, the ultimate force carried by the fibers is limited to a value (the so-called effective
force) smaller than the one corresponding to fibers rupture. One of the objectives of this
study was to calibrate the model in terms of the effective force and in turn of the effective
strain value, with the experimental results. More details on the tie-model parameters and
how they are determined can be found in [15].
(a) (b) (c)

Figure 5: (a) Infill modeling scheme, (b) strut-constitutive law and (c) tie-constitutive law.
Several numerical analyses were carried out in OpenSees in order to validate and calibrate
the proposed tie-model. Some of the model parameters (i.e. the modulus of elasticity of the
tie) were experimentally determined, whereas the rest were calibrated with the test results
along with a sensitivity analysis. As a result, it was found that the crucial parameter - and
440
the most sensitive among the others – was the effective strain of the TRM which is linked
with the effective force. The best agreement with the test results was achieved for a value
of effective strain equal to 0.8% for 1 TRM layer, whereas a change in the effective strain
value of ±25% resulted in a ±7% variation of the base shear strength. The sensitivity of the
model to the rest of the parameters was found not to be considerable. Even in the case of
the parameters that control the hysteresis loops shape, the selection of extreme set of val-
ues did not result in a significant change of the results. The latter can easily be perceived
by the curves in Fig. 6. Besides, for a specific set of these values the agreement achieved
was almost excellent (Fig. 6b).
Figure 6: Comparison between analytical and experimental results for SP#2.
CONCLUSIONS
This paper presents a study on the seismic retrofitting of 3-storey masonry infilled frames
employing textile-reinforced mortar (TRM) as externally bonded reinforcement in combi-
nation with special anchorage details. The two, 2:3-scaled, test structures (as-built and ret-
rofitted) were subjected to cycles of predefined force pattern to evaluate the efficiency of
the proposed technique. An analytical model was developed to simulate the macroscopic
behavior of TRM layers, and was validated with the test results. The main conclusions
drawn are summarized as follows:
• The retrofitting scheme results in an enhanced global response of the infilled frame both
in terms of lateral strength and deformation capacity.
• Application of TRM over the entire surface of infills should be supplemented with an
adequate infill-frame connection: this can be achieved by using custom-fabricated tex-
tile-based anchors which considerably delay debonding of TRM.
• Textile-reinforced mortar layers can accommodate large shear deformations by the abil-
ity of the textile to distort in shear sustaining at the same time its structural integrity.
• A modeling scheme that makes use of single strut and tie elements to model the behav-
ior of a retrofitted infill panel can adequately reproduce the response of the tested
frames.
• The so-called effective force is the crucial parameter of the analytical model developed
to account for the TRM contribution. The effective force is analytically calculated based
on the TRM effective strain value. Calibration of this parameter to more test results that
will be available in the future is essential for the general use of the model.
441
ACKNOWLEDGEMENTS
This research has been co-financed by the European Union (European Social Fund – ESF)
and Greek national funds through the Operational Program "Education and Lifelong Learn-
ing" of the National Strategic Reference Framework (NSRF) - Research Funding Programs
HERACLEITUS II (full support) and THALES (partial support) - Investing in knowledge
society through the European Social Fund.
REFERENCES
[1] Brameschuber, W., Koster, M., Hegger, J. and Voss, S., ‘Integrated formwork elements made of tex-
tile-reinforced concrete’, ACI SP-250 Textile-Reinforced Concrete, A. Dubey Ed. (2008)
[2] Papanicolaou, C.G. and Papantoniou, I.C., ‘Mechanical behavior of textile reinforced concrete
(TRC)/concrete composite elements’, J. Adv. Conc. Tech. 8(1) (2010) 35-47.
[3] Triantafillou, T.C., Papanicolaou, C.G., ‘Shear strengthening of reinforced concrete members with
textile reinforced mortar (TRM) jackets’, Mater. Struct. 39(1) (2006) 93-103.
[4] Bournas, D.A., Lontou, P.V., Papanicolaou, C.G. and Triantafillou, T.C., ‘Textile-reinforced mortar
versus fiber-reinforced polymer confinement in reinforced concrete columns’, ACI Struct. J. 104(6)
(2007) 740-748.
[5] Brückner, A., Ortlepp, R. and Curbach, M., ‘Anchoring of shear strengthening for T-beams made of
textile reinforced concrete (TRC)’, Mater. Struct. 41(2) (2008) 407-418.
[6] Bournas, D.A., Triantafillou, T.C., Zygouris, K. and Stavropoulos, F., ‘Textile-reinforced mortar
versus FRP Jacketing in seismic retrofitting of RC columns with continuous or Lap-spliced deformed
bars’, J. Comp. Constr. 13(5) (2009) 360-371.
[7] Elsanadedy, H.M., Almusallam, T.H., Alsayed, S.H. and Al-Salloum, Y.A., ‘Flexural strengthening of
RC beams using textile reinforced mortar–Experimental and numerical study’, J. Comp. Struct. 97
(2013) 40-45.
[8] Papanicolaou, C.G., Triantafillou, T.C., Karlos, K. and Papathanasiou, M., ‘Textile-reinforced mortar
(TRM) versus FRP as strengthening material of URM walls: in-plane cyclic loading’, Mater. Struc.t
40(10) (2007) 1081-1097.
[9] Harajli, M., ElKhatib, H. and San-Jose, J.T., ‘Static and cyclic out-of-plane response of masonry walls
strengthened using textile-mortar system’, J. Mat. Civ. Eng. 22(11) (2010) 1171-1180.
[10] Babaeidarabad, S., De Caso, F. and Nanni, A., ‘URM walls Strengthened with fabric-reinforced ce-
mentitious matrix composite subjected to diagonal compression’, J. Compos. Constr. 18(2) (2014),
04013045.
[11] Koutas, L., Pitytzogia, A., Triantafillou, T.C. and Bousias, S.N., ‘Strengthening of infilled reinforced
concrete frames with TRM: Study on the development and testing of textile-based anchors’, J. Comp.
Constr. 18(3) (2014), A4013015.
[12] Koutas, L., Bousias, S.N. and Triantafillou, T.C., ‘Seismic strengthening of masonry-infilled RC
frames with TRM: Experimental study’, J. Comp. Constr. (2014) 04014048, doi:
10.1061/(ASCE)CC.1943-5614.0000507.
[13] McKenna, F., Fenves, G.L., Scott, M.H. and Jeremic, B. ‘Open System for Earthquake Engineering
Simulation (OpenSees)’, Pacific Earthquake Engineering Research Center, University of California,
Berkeley, CA (2000).
[14] Fardis, M.N. and Panagiotakos, T.B. ‘Seismic design and response of bare and masonry-infilled rein-
forced concrete buildings Part II: infilled structures.’ J. Earthquake Eng. 1(3) (1997) 475-503.
[15] Koutas, L., Triantafillou, T., and Bousias, S. ‘Analytical Modeling of Masonry-Infilled RC Frames
Retrofitted with Textile-Reinforced Mortar’, J. Compos. Constr. (2014) 04014082, doi:
10.1061/(ASCE)CC.1943-5614.0000553.
442
CONFRONTATION OF TEXTILE REINFORCED CONCRETE (TRC)

WITH CARBON AND STAINLESS STEEL GRIDS FOR
STRENGTHENING OF MASONRY
R. Contamine (1), J.P. Plassiard (2) and P. Perrotin (2), (1) Clermont Université, Université Blaise
Pascal, Institut Pascal, BP 20206, F-63000 Clermont-Ferrand, France, (2) Université de Savoie,
LOCIE, CNRS, Polytech Annecy-Chambéry, Savoie Technolac, F-73000 Chambéry, France
Abstract: A significant need for strengthening techniques of existing masonry structures is

emerging. In response, the bonding of textile reinforced concrete (TRC) has been devel-
oped during recent years. To investigate the influence of the nature of the textile rein-
forcement, pull-out test between the TRC and the masonry has been developed. Tow na-
tures of textile reinforcement grids were tested: carbon and stainless steel. For all the fif-
teen specimens tested the displacement fields are measured by digital image correlation.
This experiment showed that strengthening of masonry structures by contact moulding of
TRC with stainless steel grids directly on the masonry is technologically feasible. The re-
sults show a different post-peak behaviour of the force – displacement curves: ductile for
stainless steel TRC and brittle for carbon TRC. This different behaviour is not only related
to the ductile behaviour of the stainless steel but also to a difference in the failure mode
between the stainless steel TRC and the carbon TRC. The stainless steel TRC showed a
failure mode by progressive delamination. It has the major benefits to be detectable with
the naked eye (allowing a simplified assessment of damages in the event of accidental
stress) and to allow high energy dissipation.
INTRODUCTION
The unreinforced and under-reinforced masonry structures represent a significant propor-

tion of the structures in France and around the world. These structures are particularly
prone to the spreading of macro-cracks due to the differential settlement of foundations,
wind, seismic loading, etc… Those loads can led to the failure of the structure. In response,
the reinforcement and repair of masonry structures by contact moulding of textile rein-
forced concrete (TRC) has grown over the last ten years [1-5]. When they are used as rein-
forcement for reinforced concrete or masonry structures, TRCs are primarily loaded under
tensile, even for the shear reinforcement [6]. However, direct tensile tests are insufficient
to characterize TRCs used as strengthening for reinforced concrete or masonry. Indeed,
previous studies shows that between the loading zone and the free tensile zone, a pull-out
effect is influencing the loading-displacement behaviour [7], [8]. Moreover, the bond be-
haviour between the TRC plate and its substrate is also influencing the loading-
displacement behaviour of the TRC [9], [10]. Therefore, a “two in one test”, tensile test
and single-lap shear test, has been developed in this study to take into account all of the
effects influencing the loading behaviour between the TRC and the masonry. To allow the
study of the global and local behaviour, the full-field displacement have been measured on
the surface of the 15 specimens according to the load thanks to digital image correlation.
443
However, this paper focuses on the study of the global behaviour. Although this study was
performed at room temperature, this is a preliminary study for the development of TRCs
having effective mechanical behavior under unusual temperature condition: above the glass
transition temperature of usual pre-impregnation polymers. Therefore, the studied TRCs
are reinforced with textiles which are non pre-impregnated with polymer products. Good
experience feedbacks obtained through TRCs reinforced with carbon textiles encourage us
to be innovative and aim for strengthening of structures by TRCs having a ductile behav-
iour. In order to meet this objective, TRCs reinforced with ductile stainless steel weaves
(steel TRCs) were tested and compared to TRCs reinforced with carbon textiles (carbon
TRCs). Indeed, in the literature, there are already studies of TRCs reinforced with steel
grouts, but using brittle multifilament steel grouts [3-5], and showing a brittle behavior.
MATERIALS
The brickwork specimens manufactured for the experimental tests are made by hollow clay
bricks (POROTHERM GF R20®). The supplier’s data give a compressive class strength
RC 80 (160 tons/ml) and a minimal compressive strength “fb” (for a design with the Euro-
codes) of 10 MPa. The bricks are 500 mm long, 299 mm wide and 200 mm thick, and they
were assembled with a special mortar 1 mm thick. The matrix is a cementitious mortar,
previously developed for the repair and reinforcement of reinforced concrete structures
with TRC. This thixotropic matrix allows for moulding the TRC on surfaces of any shape
and orientation. The mortar is consisting of a hydraulic binder (CEM I 52.5 N HTS), an
ultrafine quicklime, a silica fume, a siliceous sand 0/400 μm, a sand 400/800 μm and a
redispersible polymer powder (3 % of the mass of binder + aggregates). The TRC plates
are made of the previously described matrix associated with a textile reinforcement. Two
types of textile reinforcement were used, a carbon grid and a stainless steel grid (Figure 1).
The carbon grid with a mesh size of 10 mm is made of multifilament yarns of carbon
(number of filaments per yarn: 12000, diameter of filaments: 7 µm, tensile modulus: 230
GPa, ultimate strain: 2,1 %, strength: 4900 MPa, strength of the grid: 226 kN/m) stuck
with thermo bonded thermoyanrs. The steel grid is made of monofilaments (1 mm diame-
ter) weaved with a mesh size of 5 mm (strength of the grid: 123 kN/m, figure 2). The initial
stiffness of the steel grid is exactly similar to the stiffness of the carbon grid.
Figure 1: Illustration of the two types of textile Figure 2: Load-strain curves of the carbon textile grid
reinforcement and the steel textile grid (suppliers’ data). The load is
given for a textile width of 10 cm
444
TESTING METHODS
Tensile single-lap shear test

The geometry and boundary conditions of the test are shown in Figure 3. The test specimen
is made of two hollow bricks on which a TRC plate is contact moulded. Moreover, in order
to study the tensile behaviour of the TRC, the TRC plate is extended of 500 mm beyond
the brickwork. Aluminium end plates were then bonded with epoxy resin 50 mm after the
free end of the TRC plate. It should be noted that TRC are implemented corresponding to
the real site conditions: without mould ensuring a constant thickness of the TRC, by con-
tact moulding utilizing a trowel. Therefore, the thickness of the TRCs depends on the im-
plementation and, on the number of textile layers. The loading was applied by a displace-
ment rate of 1 mm/min. The displacement field is measured by an image correlation de-
vice, which used an image shooting frequency of 2 Hz. The 7D software was used to per-
form the image analysis [11].
Figure 3: Illustration of the tensile sin- Figure 4: Location of the references measured with digital image
gle-lap shear test correlation
Series of experiments
For both type of textile reinforcement, a parametric study was conducted concerning the
textile reinforcement ratio of the TRC. TablesA1 defines the layout of the tested compo-
sites and the average thickness of these TRCs.
445
Table 1: Layouts of the tested TRC

Group Composite Number Fibre volume ratio
Specimen
specimen thickness Textile of textile along the longitudinal
name
name (mm) layers direction (‰)
Carbon-1-i 4 Carbon 1 11,5
Carbon- Carbon-1-ii 4 Carbon 1 11,5
TRC Carbon-2-i 8 Carbon 2 11,5
Carbon-2-ii 5 Carbon 2 18,5
Steel-1-i 4 Steel 1 39,3
Steel-1-ii 4 Steel 1 39,3
Steel-TRC
Steel-2-ii 6 Steel 2 52,4
Steel-3-ii 7,5 Steel 3 62,8
Use of the displacement field (DIC)

Digital image correlation (DIC) allows measuring the full-field displacement measurement
on the surface of the specimens according to the load. The global tensile behaviour of the
TRCs is characterised with load/number of textile layers-mean strain curves. The mean
strain was calculated by dividing the mean displacement between Reference 1 and 2 (Fig-
ure 4) by the length of the measurement zone (200 mm).
The global behaviour of the bond between the TRC and its substrate is characterised with
the load-displacement curve. The displacement was calculated between references 2 and 3
(Figure 4).
The tensile behaviour

The load-strain curves of all the tested specimens are illustrated on the Figure 5. For the
same number of textile layers and a close thickness between those TRCs (Table 1), the
steel-TRCs are stiffer than the carbon-TRCs. Unlike to the carbon-TRCs, the steel-TRCs
show an important ductility.
Figure 5: Load-strain curves of all the tested specimens
446
The carbon-TRCs show a global stiffness (the ratio between load and strain) lower than the
carbon grid stiffness (except for one specimen of carbon-2, Figure 6). Actually, for a strain
higher than respectively 0,9 ‰ and 1,8 ‰ the global stiffness of respectively carbon-1
specimens and one specimen of carbon-2 is lower than the textile stiffness. Therefore car-
bon-TRCs show a limited contribution of the mechanical performances provided by the
matrix. The strength efficiency ratio (Erσ) of the TRC was calculated as the ratio between
the ultimate load per number of textile layers of the TRC and the ultimate load of a textile
layer. It appears that the strength efficiency ratios of the carbon-TRCs are very low (be-
tween 20% and 28%).
The confrontation of Figure 6 and Figure 7 shows that unlike the carbon-grid, the steel grid
allows a significant contribution of the mechanical performances provided by the matrix.
Indeed, for strains lower than 10 ‰, the global stiffness of the steel-TRCs are significantly
higher than the textile global stiffness. Moreover, the junction between the textile load-
strain curve and the TRC load/number of textile layers-strain curves is located at a strain
included between 15 ‰ and 28 ‰ according to the steel-TRCs layouts (compared to 0,9 ‰
and 1,8 ‰ for carbon-TRCs). This significant difference of behaviour between carbon-
TRCs and steel-TRCs is due to a stronger bond between the steel-textile and the matrix
than between the carbon-textile and the matrix. This is mainly due to a difference of bonds
at different scales: the material, the roving, and the textile. The strength efficiency ratio
(Erσ) of the steel-TRC is independent from the number of textile layers. Its value of 60±3
% is more than twice higher than the one of the carbon-TRCs, but there is still room for
improvement.
Figure 6: Load/number of textile layers-strain curves Figure 7: Load/number of textile layers -strain curves
of the carbon-TRCs of the steel-TRCs
The failure modes
All of the carbon-TRCs showed a failure mode on a predominant crack localized at the end
of the brickwork (Figure 8). Thanks to the study of the local behaviour with the digital
image correlation results, it has been shown that this is a pull-out failure mode. This failure
mode is already well detailed in the scientific literature [12], [7], [8]. On the other end, all
of the steel-TRCs showed a failure mode by delamination of the anchor part of the TRC,
from the beginning to the end of the anchor (Figure 9). The failure occurs when the length
of the working part of the TRC becomes too short to transmit the load from the TRC to the
brick work.
447
Figure 8: Illustration of the failure mode of car- Figure 9: Illustration of the spreading of the delamina-
bon-1-i which is representative of all of the car- tion along the TRC anchor of Steel-2-ii which is repre-
bon-TRCs. sentative of all of the steel-TRCs.
The anchor behaviour

The load-displacement curves of all the tested specimens are illustrated on the Figure 10.
Like for the tensile behaviour, for the same number of textile layers and a close thickness
between this TRC (Table 1), the steel-TRC are stiffer than the carbon-TRC (Figure 11).
Unlike to the carbon-TRCs, the steel-TRCs show an important ductility of the anchor.
The initial qualitative behaviour of the load/number of textiles layers-displacement curves

(anchor curves, Figure 11) is close to the one of the load/number of textiles layers-strain
curves (tensile curves, Figure 6). However, for the carbon-TRCs, unlike the tensile curves,
the last part of the anchor curves is non-linear. This non linearity in the last part of the
curves is due to the failure mode by pull-out of the carbon-TRCs [7]. For the steel-TRCs,
unlike the tensile curves, the last part of the anchor curves shows a constant load according
to the increase of the displacement. This is due to the failure mode by delamination of the
steel-TRCs. It cannot be observed on the tensile curves because the delamination of the
anchor occurs at a constant load and strain of the part of the specimen where the tensile
behaviour is measured.
Figure 11: Load/number of textile layers

Figure 10: Load-displacement curves of the anchor tests displacement curves of the anchor tests
(focus on the first 2 mm of displacement)
448
CONCLUSIONS
This experiment showed that strengthening of masonry structures by contact moulding of
TRC with stainless steel grids directly on the masonry is technologically feasible. Alt-
hough, the initial stiffness of the steel grid is exactly similar to the stiffness of the carbon
grid, the steel-TRCs are stiffer than the carbon-TRCs. Moreover, the strength efficiency
ratio of the steel-TRCs is more than twice higher than the one of the carbon-TRCs. The
results show a different post-peak behaviour of the force – displacement curves: ductile for
stainless steel TRC and brittle for carbon TRC. This different behaviour is not only related
to the ductile behaviour of the stainless steel but especially to a difference in the failure
mode between the stainless steel TRCs and the carbon TRCs. The Carbon TRCs showed a
pull-out failure mode which is brittle, localized and hardly detectable before failure. On the
other end, the steel TRC showed a failure mode by progressive delamination. It has the
major benefits to be detectable with the naked eye (allowing a simplified assessment of
damages in the event of accidental stress) and to allow high energy dissipation. For the
carbon-TRCs and the steel-TRCs, due to their respective failure mode, the last part of the
tensile curves differs from the last part of the anchor curves. Therefore, in the field of
structural reinforcement, for the two types of tested TRCs subject to a pull-out loading
(due to a crack bridging), there is a threshold of load above which the tensile characteriza-
tion of the TRC should be no more appropriate in the design rules. Finally, the steel textile
could be greatly improved because the strength efficiency ratio (Erσ) of the steel-TRCs is
low (60±3 %) and the ultimate strain of the steel TRCs (between 3.4% and 5.3%) is really
lower than the ultimate strain of the steel grid (36%).
REFERENCES
[1] Prota, A., Marcari, G., Fabbrocino, G., Manfredi, G. and Aldea, C. ‘Experimental in-plane
behavior of tuff masonry strengthened with cementitious matrix grid composites’, J. Compos. Con-
str. 10(3) (2006) 223–33.
[2] Harajli, M., Eikhatib, H. and San-Jose J.T. ‘Static and cyclic out-of-plane response of masonry walls
strengthened using textile-mortar system’, J. Mater. Civ. Eng. 22(11) (2010) 1171–80.
[3] De Felice, G., De Santis, S., Garmendia, L., Ghiassi, B., Larrinaga, P., Lourenço, P.B., Oliveira,
D.V., Paolacci, F. and Papanicolaou, C.G., ‘Mortar-based systems for externally bonded strengthen-
ing of masonry’, DOI 10.1617/s11527-014-0360-1, Mater. Struct. (2014).
[4] Razavizadeh, A., Ghiassi, B. and Oliveira, D.V., ‘Bond behavior of SRG-strengthened masonry
units: Testing and numerical modeling’, Constr. Build. Mater. 64 (2014) 387–397.
[5] Malena, M. and De Felice, G., ‘Debonding of composites on a curved masonry substrate: Experi-
mental results and analytical formulation’, Compos. Struct. 112 (2014) 194–206.
[6] Contamine, R., Si Larbi, A. and Hamelin, P., ‘Identifying the contributing mechanisms of textile
reinforced concrete (TRC) in the case of shear repairing damaged and reinforced concrete beams’,
Eng. Struct. 46 (2013) 447–458.
[7] Contamine, R., Si Larbi, A. and Hamelin, P., ‘Contribution to direct tensile testing of textile rein-
forced concrete (TRC) composites’, Mater. Sci. Eng. 2011;528: 8589–8598.
[8] Contamine, R., Junes, A. and Si Larbi, A., ‘Tensile and in-plane shear behaviour of textile rein-
forced concrete: Analysis of a new multiscale reinforcement’, Constr. Build. Mater.51 (2014) 405–
413.
[9] Ortlepp, R., Hampel, U. and Curbach, M., ‘A new approach for evaluating bond capacity of TRC
strengthening’, Cement Concrete Comp., 28 (2006) 589–597.
[10] D’Antino, T., Carloni, C., Sneed, L.H., Pellegrino, C., ‘Matrix–fiber bond behavior in PBO FRCM
composites: A fracture mechanics approach’, Eng. Fract. Mech. 117 (2014) 94–111.
[11] Vacher P, Dumoulin S, Morestin F, Mguil-Touchal S. Bidimensional strain measurement using
digital images. Proc Inst Mech Eng 1999;213:811–7.
[12] Hegger, J., Will, N., Bruckermann, O. and Voss, S., ‘Load–bearing behaviour and simulation of
textile reinforced concrete’, Mater. Struct. 39 (2006) 765–776.
449
450
SHEAR BEHAVIOUR OF MASONRY WALLS STRENGTHENED BY

EXTERNAL BONDED FRP AND TRC
L. Bui1, A. Si Larbi2, N. Reboul1, E. Ferrier 1, 1LGCIE, Université Claude Bernard LYON 1, 82

bd Niels Bohr, 69622 Villeurbanne, FRANCE, 2, Université de Lyon, Ecole Nationale
d’Ingénieurs de Saint-Etienne (ENISE), Laboratoire de Tribologie et de Dynamique des Systèmes
(LTDS), UMR 5513, 58 rue Jean Parot, 42023 Saint-Etienne Cedex 2, France
Abstract: This experimental study focuses on the behaviour of hollow concrete brick ma-
sonry walls, especially walls reinforced with composite materials under in-plane loading
conditions. This work is a step towards defining reliable seismic strengthening solutions.
Indeed, in France, more stringent seismic design requirements for building structures have
been considered with the replacement of old design codes. Thus, an experimental program
has been performed at the laboratory scale. Six walls have been submitted for shear-
compression tests - five walls are reinforced by (1) – fibre-reinforced polymer (FRP) strips
using E-glass and carbon fabrics and/or (2) a textile-reinforced concrete (TRC), and the
last wall acts as a reference. It is noted that the composite strips are mechanically anchored
into the foundations of the walls to improve their efficiency. All of the walls share the
same boundary and compressive loading conditions, which are representative of a seismic
solicitation. Nevertheless, masonry wall performances and anchor efficiency are only eval-
uated under monotonic lateral loadings. A comparative study on global behaviour and on
local mechanisms is performed and, in particular, highlights that the mechanical anchor
systems play an important role in improving the behaviour of reinforced walls (by FRP and
TRC) and that the solutions for strengthening by TRC permit the upgrade of the walls’
ductility with a lower strength compared with the solutions with FRP.
INTRODUCTION
Masonry has a long history as a building technique. In brief, due to seismic actions, walls
in a building can be subjected to shear forces both in the in-plane and out-of-plane direc-
tions. Although the out-of-plane failures should not be overlooked, practitioners (in a broad
sense, including the scientific community) tend to make the in-plane seismic response of
shear walls their first priority; they indeed appear as key vertical components to bear seis-
mic loading.
Solutions for repairing or strengthening masonry structures are many and are varied. Nev-
ertheless, externally bonded fibre-reinforced polymer (FRP) composites are often preferen-
tially chosen by prime contractors [2], mostly because of their lightweight and their ease of
use. However, the reinforcing efficiency of FRP is rarely fully valued when they are only
externally bonded to structural elements. FRP mechanical properties are limited because of
the debonding of the composite sheets. To address this issue, an adequate mechanical an-
chorage system needs to be set up to enhance the bond (between a masonry structure and
its foundation) performance. The benefits of this solution – in terms of the FRP efficiency
451
and lateral load resistance of a masonry wall – have now been widely acknowledged in the
case of out-of-plane actions [3].
In addition, in the context of sustainable development and health and safety conditions for
workers, consideration should be given to an alternative material to FRP, which is often
manufactured with highly toxic epoxy resins. The idea is to substitute these resins with
cementitious materials while preserving or even improving the dissipative capacity of rein-
forced structures. From this perspective, textile-reinforced concrete (TRC) composites,
which combine a suitable fine-grained mortar with the latest generation of textile fabrics,
would benefit from promotion.The efficiency of TRC for strengthening masonry structures
has recently been investigated ([4], [5], [6]-[9]). Compared with FRP, TRC composites
show a nonlinear tensile behaviour with multiple matrices cracking, giving them a greater
deformation capacity, a priori more suitable for seismic reinforcement [6].
Although instructive, these studies lack diversity for studied materials, reinforcement con-
figurations, applied normal loads and slenderness ratio of walls. This work therefore, on
the one hand, is aimed at further developing the existing experimental database, with spe-
cial emphasis on identifying the performances of anchorage devices, particularly in the
framework of a comparative study between FRP and TRC composites. On the other hand,
this paper tries to help identify and clarify damage dissipative mechanisms and their im-
pact on the failure modes of the masonry walls. To attain the aforementioned objectives, an
experimental campaign has been performed, based on static monotonic shear tests, which
are a simplified way to simulate stress states resulting from earthquakes. Certainly, inertial
effects and the inherent cyclical nature of seismic actions are not addressed in the present
study. However, this work can be regarded as a first step towards the definition of efficient
reinforcement solutions. The approach is to test some strengthening configurations to have
relevant and valuable information and to offer prospects that would be appropriate to as-
sess with more realistic loadings in terms of earthquake hazards.
Masonry walls
A series of six walls has been built. The hollow concrete block units, whose dimensions
are 500 mm long, 200 mm high and 75 mm thick, belong to Group 2 according to Euro-
code 6, with a strength class B40 (characteristic compression strength of 4 MPa). However,
these blocks have been halved lengthwise before being assembled to make walls dimen-
sions compatible with the limited means of the laboratory in terms of space and actuator
capacity (Block work size at reduced scale : 250 x 200 x 75 mm3). These blocks are as-
sembled with a mortar composed of Portland cement (CEM I 52.5) and sand in the propor-
tion 1:3 with a water/cement ratio equal to 0.5. These walls are built on a reinforced con-
crete foundation and are superposed by a concrete loading beam (Figure 3)
452
Reinforcement
a. Strengthening materials
Two types of composites have been used: the first composite is a fibre-reinforced polymer
(FRP) while the latter composite is a textile-reinforced cementitious composite (TRC).
¾ FRP composite
The fibre-reinforced composite materials consist of a two-component epoxy matrix and bi-
directional fabrics made of either carbon (CFRP) or glass (GFRP). Their mechanical char-
acteristics have been measured on six specimens according to ISO 527-1. The obtained
results are listed in Table 1.
Table 1: Mechanical characteristics of composites
Composite strengthening Young modulus Tensile strength Ultimate strain
Nominal thick-
ness (mm)
system (GPa) (Mpa) (μm/m)
CFRP 0,48 105 1700 16000
GFRP 1,7 7,2 100 13800
¾ TRC composite
The components of selected composite, referenced to works performed by Contamine et al.
[10], are listed in Table 2, that results from a compromise between all of the aforemen-
tioned parameters, including workability and thixotropy (to apply strengthening materials
more easily). The composite material contains the 4.36 % AR-glass fibre volume fraction,
that is, 2.18% in each principal direction. The textile used for reinforcement is a bidirec-
tional warp-knitted grid fabric.
Table 2: Mix design for TRC composite
Micro-mortar**
Nature of fibres Glass-AR Grain size <2 mm*
TEX 1200* Silica-fume yes*
Fibre diameter 19 μm* Thixotropy yes*
Number of filament/yarn 1600* Shrinkage ∼ 0*
Knitted grid size 5 mm x 5 mm* Tensile strength 5 MPa**
Tensile strength (yarn) 1102 MPa* Compressive strength 40 MPa*
Young modulus 74000 MPa* Young modulus 1700 MPa*
* Provided properties; ** laboratory characterization
453
Figure 1: Tensile behaviour of TRC materials during uniaxial tensile tests
Direct quasi-static tensile tests, whose protocol has been validated [11], have been per-
formed to mechanically characterize the TRC reinforcements. Stress-strain curves are giv-
en in Figure 1. Two different behaviour laws appear in Figure 1. Indeed, without the im-
pregnating resin (latex), TRC exhibits a bilinear behaviour whereas by using latex, the evo-
lution law is nearly linear because of a more homogeneous yarn contribution.
b. Anchorages
To reduce or ideally remove the overturning effects due to lateral loads on walls and to
best maximize the potential of each reinforcement solution by a priori improving its effi-
ciency, a connector, in the form of an anchorage, has been introduced between walls (on
their lower part, on both sides) and their foundation footings. Given the existing solutions
and their well-established performances (easy application and high strength), only connec-
tion solutions based on carbon fibres and an epoxy resin have been considered. The an-
chorage system (MAPEI) is an anchor made from monodirectional carbon fibres with at
least 36 yarns, each including 12,000 fibres. The anchorage strength given by the manufac-
turer is 30 kN at the ultimate limit state.
Figure 2: Description of a MAPEI anchor
c. Strengthening configurations
On the objective in reinforcing a structure is to improve its strength capacity, to enhance its
ductility, or both, different strengthening configurations with TRC and FRP composites
have been proposed. The reinforcing material is always applied symmetrically on both wall
surfaces.
454
The experimental program consists of testing six masonry walls to failure, including an
unreinforced wall (as the reference specimen) and five TRC or FRP-reinforced specimens
(see Figure 3). With FRP composites, three strengthening patterns have been proposed.
The first wall, referenced CGRW, has been reinforced with both carbon fibre-reinforced
polymer (CFRP) and glass fibre-reinforced polymer (GFRP) to significantly improve
strength capacity. The two remaining walls, for their part, have been reinforced by either
CFRP sheets (CRW wall) or GFRP sheets (GRW wall) to combine strength capacity and
ductility. Concerning TRC solutions, it is clear that without overlooking the loadbearing
capacity, it is desirable to increase the ductility of reinforced walls, which is the main mo-
tivation of our choices. In the first pattern (TRCW1 wall), only one TRC layer, is applied
on each side of both faces. However, in the second selected pattern, the main objective is
to improve the dissipation capacities, so a vertical strip is applied in the middle of the wall.
Each strip is made of three TRC-layers.
« Concrete loading beam »
URW TRCRW1 TRCRW2

(One layer without latex) (3 layers with latex)
CGRW CRW GRW
Figure 3: Unreinforced wall and FRP/TRC-reinforced walls
Testing procedure
As mentioned above, monotonic in-plane combined compression-shear tests have been
selected. The vertical axial load (N) is centred and applied by means of a hydraulic actua-
tor with 200 kN capacity. This actuator is combined with a force sensor that is located on a
stiff steel plate and is placed on the concrete loading beam. These devices are held by a
steel profile connected to the strong floor through steel rods, which allows us to control the
applied vertical load (Figure 4).
455
Spreader beam
Load cell
Hydraulic jack Load cell
Vertical load actuator
Post-tension rod
Wall
Beams
Reaction wall
Threaded rod
Figure 4: Test set-up
The vertical load is should represent the weight of the upper floors. In literature, this value
varies greatly and depends on the masonry compression strength. For a given wall geome-
try, the axial compression load magnitude influences the overall performances, in particu-
lar, the efficiency of strengthening solutions. Papanicolaou et al. [7] have noted that rein-
forcement efficiency is reduced as vertical compression load rises. Because this study is
motivated by assessing the shear contribution of the reinforcements, a relatively low value
of approximately 6% of the masonry compression strength has been adopted. It corre-
sponds to a mean vertical stress of 0.2 MPa (15 kN). In practical terms, the vertical load is
very slowly applied up to the target value equal to 15 kN. It is kept constant during the test
because of a force control of the vertical actuator. At this time, horizontal load is imposed
under quasi-static monotonic conditions. This loading step is performed under displace-
ment control at a rate of 0.015 mm/s to better capture post-pic behaviour. Lateral loading is
stopped when masonry walls have obviously failed, that is, when lateral force drops signif-
icantly. According to Tomazevic [12], failure is ascertained when horizontal load falls by
20%. This criterion will be adopted herein.
Experimental results
a. Global behaviour
All walls show a nonlinear behaviour after an initial linear elastic branch. A significant
deviation is experienced in the length of the linear zone although stiffnesses are close. A
dramatic increase in the ultimate strength and in the second phase stiffness is clearly ob-
served, even if it is conditioned by reinforcing materials and strengthening configurations.
Similarly, dissipation capacities increase overall in varying degrees that must be assessed
and quantified. It must be underlined that only the wall CGRW behaves as brittle (as a re-
sult of a sudden and premature failure) (Figure 5a). To evaluate the performances of these
walls in terms of appropriately chosen and unbiased indicators, the experimental load-
displacement curves will be idealized according to the trilinear model proposed by Toma-
zevic [12] (Figure 5b).
456
(a) (b)
Figure 5: (a) - Curves of load versus horizontal displacement at the top of the wall; and (b) - Idealized tri-
linear diagram
The three phases of the conventional trilinear diagram are bounded by three characteristic
points, which facilitate discussion and comparison on wall behaviours. The first zone,
called elastic, ends at lateral load Vcr and displacement dcr. They mark the formation of the
first significant cracks in the wall, which entail a change in stiffness [1]. As masonry walls
exhibit highly nonlinear behaviours, a conventional Vcr value is generally adopted. Accord-
ing to Tomazevic’s study [12], Vcr is equal to 70% of the maximum resistance Vmax. The
second zone extends to the maximum lateral load Vmax and displacement dVmax.
At last, the ultimate zone is characterized by the ultimate load Vdu, corresponding to 80%
of the maximum load and the maximum displacement du on the softening branch. Kel, μu
and Ediss are the initial stiffness (the slope of the first phase, considered as elastic), ductility
coefficient and dissipation energy, respectively. This last parameter provides information
on resistance and deformation capability. Indeed, it is determined as the area below the
idealized load-displacement diagram and is given by the following equation:
(1)
The ductility coefficient, obtained by dividing the ultimate horizontal displacement (du) by
the elastic displacement (dcr), reflects the deformation capacity of shear walls in the post-
elastic zone. This parameter is a decisive criterion for paraseismic construction.
457
Table 3: Summary of experimental results
Ediss(kN.mm)
Vcr = (70%Vmax)
Vu =(80%Vmax)
(mm2)
(MPa)
μ = du/dcr
Kel(kN/mm)
dVmax (mm)
Vmax (kN)
dcr (mm)
du (mm)
×
(kN)
(kN)
Walls Failure modes
=
×
.
URW - - - 7.88 0.88 11.25 4.59 9.00 12.98 8.97 14.78 123.90 Flexural
CFRP 0.48x60 Flexural+
CGRW 3.82 82020 35.79 3.66 51.13 8.39 40.90 8.70 9.78 2.38 285.30
GFRP 1.7x400 shear
Flexural+
CRW CFRP 0.48x60 0.3 31316 22.40 2.50 32.00 9.94 25.60 13.20 8.96 5.28 324.26
shear
Flexural+
GRW GFRP 1.7x400 3.52 50702 35.53 3.06 50.75 8.89 40.60 17.70 11.61 5.78 708.24
shear
Flexural+
TRCRW1 TRC 3x200 3.31 4854 8.61 0.75 12.30 5.85 9.84 12.80 11.48 17.07 133.49
shear
Flexural+
TRCRW2 TRC 9x200 13.98 22135 18.92 1.07 27.03 7.32 21.62 17.60 17.68 16.45 403.80
shear
-thickness of composite band; - width of composite band; - total cross section area of strengthening
(= ; l and t- width and thickness of masonry wall; - vertical reinforcement ratio;
¾ Strength capacity
From these results, it appears that reinforced walls achieve substantially higher ultimate
loads than the reference unreinforced wall, regardless of reinforcement type. Shear strength
increases from 110% (TRCW1) to 450% (CGRW). TRC reinforcements are slightly less
efficient (175% on average) than FRP reinforcements in terms of strength capacity. It must
be noted that the TRCW1 wall appears as an exception because its reinforcement has a
slight effect on the ultimate load. It can be explained by the excessively low stiffness of
this reinforcement, only comprising one TRC layer.
¾ Dissipation energy
In addition to the ductility factor, the dissipation energy (Ediss) will help us position, at least
in order of magnitude, the potentials of FRP and TRC as masonry reinforcements. It is
clear that for the adopted configurations, both with TRC (with the exception of the
TRCRW1 wall with a very low reinforcement ratio) and FRP, results are conclusive be-
cause increases in the range of 200 to 500% can be expected. To better understand, it is
important to correlate the above performance indicators with observed failure modes, dam-
age mechanisms and kinematics and with data at the local scale. This is the purpose of the
rest of this paper.
(a) (b) (c)

Figure 6: Comparative diagrams of the different indicators (a-strength capacity, b-ductility coefficient, c-
dissipation energy)
458
b. Failure modes
This section is devoted to providing information regarding failure modes based on rein-
forcement types and patterns, so that the efficiency of strengthening materials in the global
lateral behaviour and their impact on damage mechanisms can be precise.
¾ Unreinforced wall:
An unreinforced wall exhibits a flexural failure mode (Figure 7) characterized by horizon-
tal cracks on the left part of the wall (on the side where lateral load is applied) due to ten-
sile stresses in bed mortar joints and by toe crushing (on the right part) at the compressed
corner. It is worth noting that observation of this failure mode has given valuable infor-
mation for designing strengthening configurations in the present study.
¾ FRP-reinforced walls:
Failure modes of FRP-reinforced walls are presented in Figure 8. First, it must be noted
that they are different and depend, among other things, on adopted reinforcement configu-
rations.
The sudden failure of the CGRW wall is singular and results from the premature failure of
anchorages, which are subject to tensile forces in the left part (side where lateral load is
applied). This wall has been reinforced over a large surface area on which both glass and
carbon textiles have been laid up, resulting in a significant increase in the wall stiffness
(compared with an unreinforced wall but also to other strengthened specimens). Under
loading, the wall begins a rigid body motion that induces high shear stresses at the bottom
of the wall which could explain, even partially, the observed failure mode. As a conse-
quence, the use of anchorage devices certainly improves the lateral strength of reinforced
walls. However, to avoid sudden ultimate failures of walls and to ensure sufficient energy
dissipation capabilities, an anchorage design cannot be decoupled from the stiffness or
reinforced walls (taking into account both reinforced surface area and strengthening mate-
rials).
The other two FRP-reinforced walls are not affected by anchor failure and show similar
damage mechanisms. These walls exhibit coupled shear-flexure failure modes. In both
specimens, shear cracks initiate in the middle of the wall (unreinforced zone) and propa-
gate along the compressed diagonal for the CRW wall (which is a typical shear failure pat-
tern), even along the edge of the unreinforced zone (in particular for GRW wall), thus re-
flecting that reinforcements can bridge cracks and also influence their propagation. Never-
theless, the collapse of these walls (or very marked deterioration) occurs by the crushing of
the lower right corner (highly subject to compression) and ends and by the splitting of the
extreme unit block. With the failure of this block, the wall tends to turn about the right toe,
as found for an unreinforced wall (Figure 8 (b) and (c)).
459
(b1) (b2)
Figure 7: (a)- Flexural failure for unreinforced wall (URW) and Failure modes and damage mechanisms for
TRC-reinforced walls (b1)- TRCRW1 wall and (b2)- TRCRW2 wall
(a) (b) (c)
Figure 8: Failure modes and damage mechanisms for FRP-reinforced walls: (a)- sudden failure of
CGRW wall; (b) and (c)- coupled shear-flexure failure of CRW and GRW walls
¾ TRC-reinforced walls:
As indicated above for FRP-reinforced walls, the addition of TRC reinforcements changes
the failure mode. TRC-reinforced walls have failed by combined shear-flexure (Figure 7b).
Both reinforcement patterns lead to cracks at the left bottom part (side where load is ap-
plied) and in strengthening strips (unlike FRP) over horizontal joints. In the case of the
TRCRW1 wall (low reinforcement ratio and low reinforced surface area), cracks develop
along a horizontal joint and go through a block unit in the centre with an inclined crack
that suggests a reaction to a shear solicitation. At the ultimate limit state, a macro crack has
been observed that corresponds to the failure of mortar used in TRC without any textile
degradation. In scientific literature, this failure mode is commonly referred to as the « peel-
ing-off » failure [13] (Figure 7 b1). In the case of the TRCRW2 wall (reinforced by three
TRC layers on a larger area), damage and failure mechanisms are different. Cracks grow
horizontally and spread over the height of the strengthening strips. Finally, multi-crack
initiation contributes, beyond what is observed on the wall, to the global damage of the
specimen. It must be noted that until failure, which ultimately occurs by crushing the right
bottom corner, crack opening displacement (without having been measured) remains lim-
ited (Figure 7 b2).
460
In light of the obtained results, it is important to emphasize that: - anchors can be consid-
ered useful and efficient systems, but their use must be correlated at this time with the
global stiffness of the wall (reinforcements’ nature and surface area) to avoid their sudden
and premature failure; - dissipation mechanisms are improved and concentrated for all re-
inforced walls, regardless of the adopted reinforcements’ type and pattern. However, the
dissipation process is controlled because it occurs on the reinforcement itself for TRC ma-
terials, whereas it is uncontrolled with FRP-debonding.
CONCLUSIONS
The present experimental study has focused on masonry walls reinforced by TRC and FRP
composites that are subject to monotonic in-plane combined shear-compression tests. The
main findings are as follows: - Reinforcements, regardless of their nature and the adopted
layout diagram (provided that reinforcement ratio is sufficient), enable us to extend the
structural integrity field of masonry walls; - TRC reinforcements lead to lower perfor-
mance levels than FRP reinforcements in terms of lateral strength capacity, but they signif-
icantly increase their ductility capacity;- TRC and GFRP seem to be more appropriate than
CFRP in terms of ultimate displacement capacity; - A low TRC reinforcement ratio only
marginally modifies global masonry wall performances; - Dissipation mechanisms, which
differ between FRP and TRC-reinforced walls, have been clarified; in particular, they dif-
fuse dissipation processes (micro-cracks) in TRC composites; - Anchorage systems are
appropriate (and technologically possible) to improve the in-plane performances of rein-
forced masonry walls ;- Reinforcement design is limited by the compressive strength of
concrete hollow blocks; - The rigid body motion of a strengthened wall depends on the
reinforcement pattern (reinforcement ratio, axial rigidity and reinforced surface area) and is
likely to substantially limit the contributions of anchorage systems.
REFERENCES
[1] Capozucca R. “ Experimental analysis of historic masonry walls reinforced by CFRP under in-plane
cyclic loading”; Composite Structures 2011;94:277-289
[2] Lanivschi C. E. “State of the art for strengthening masonry with Fibre Reinforced Polymers”; Buletinul
Institutului Politehnic Din IASI Publicat de Universitatea Tehnică, Gheorghe Asachi” din Iaşi Tomul
LVIII (LXII), Fasc. 2, 2012 Secţia ConstrucTII. Arhitechtura.
[3] Dillon S. Lunn, Rizkalla S. H., Maeda S., Ueda T. “FRP Anchorage Systems for Infill Masonry Struc-
tures”; APFIS 2012 Hokkaido University Japan, 2-4 february 2012.
[4] Bernat E., Gil L., Roca P., Escrig C. “Experimental and analytical study of TRM strengthened brickwork
walls under eccentric compressive loading”; Construction and Building Materials 2013;44:35-47
[5] Carozzi F. G., Milani G., Poggi C. “Mechanical properties and numerical modeling of Fabric Reinforced
Cementitious Matrix (FRCM) systems for strengthening of masonry structures"; Composite Structures
2014;107:711-725.
[6] Papanicolaou C.G., Triantafillou T.C., Karlos K., Papathanasiou M. “Textile-reinforced mortar (TRM)
versus FRP as strengthening material of URM walls: in-plane cyclic loading”; Materials and Structures;
2007;40:1081-1097.
[7] Papanicolaou C. G., Triantafillou T., Lekka M.: “Externally bonded grids as strengthening and seismic
retrofitting materials of masonry panels”; Construction and Building Materials 2011; 25:504–514.
[8] Franca da Porto, Flavio Mosele, Claudio Modena “In-plane cyclic behaviour of a new reinforced mason-
ry system: Experimental results”; Engineering Structures 2011; 33:2584–2596
[9] Prota A.; Marcari G.; Fabbrocino G.; Manfredi G.; and Aldea C. “Experimental In-Plane Behavior of
Tuff Masonry Strengthened with Cementitious Matrix–Grid Composites”; Journal of composites for
construction ASCE, 2006;10(3):223-233.
461
[10] Contamine R., Si Larbi A., Hamelin P. “Evaluation of the TRC (Textile Reinforced Concrete) solutions
in the case of RC beams shear strengthening”; Proceedings of 2nd international RILEM conference on
strain hardening cementitious composites, Rio de Janeiro, Brazil; p. 89–95, 2011.
[11] Contamine R., Si Larbi A., Hamelin P. “Identifying the contributing mechanisms of textile reinforced
concrete (TRC) in the case of shear repairing damaged and reinforced concrete beams”; Eng
Struct;2013;46:447–58.
[12] Tomazevic M., Lutman M. and Petkovic L. “Seismic Behavior of Masonry Walls: Experimental Simu-
lation”; J. of Struct. Engg., ASCE, 1996;122(9):1040-1047.
[13] Magenes G. and Calvi G.M. “In-plane seismic response of brick masonry walls”, earthquake engineering
and structural synamics”;1997;26:1091-1112.
[14] Banholzer B. “Bond behaviour of a multi-filament yarn embedded in a cementious matrix”; Philosophy
doctoral thesis, Schriftenreihe Aachener Beiträge zur Bauforschung, Institut für Bauforschung der
RWTH Aachen, Nr. 12; 2004.
462
SHEAR STRENGTHENING OF PRESTRESSED CONCRETE BEAMS

WITH TEXTILE REINFORCED CONCRETE
Martin Herbrand, RWTH Aachen University, Institute of Structural Concrete, Mies-van-der-Rohe

Str. 1, 52074 Aachen, Germany, mherbrand@imb.rwth-aachen.de
Abstract: A large part of existing highway bridges in Germany exhibits calculative shear
capacity deficits under static and cyclic loading. More structures are therefore expected to
demand refurbishment and strengthening within the next years. Besides common strength-
ening methods, the use of textile reinforced concrete (TRC) offers an innovative alternative
for strengthening measures by combining the advantages of lightweight glued CFRP-
stripes and additional concrete layers. Two full scale tests on I-shaped prestressed concrete
beams under cyclic shear loading were therefore carried out at the Institute of Structural
Concrete at RWTH Aachen University. Previous tests on identical non-strengthened beams
served as a reference for the tests on beams strengthened with TRC. These preliminary
tests have shown that a considerable increase of the shear fatigue and static shear strength
can be obtained by strengthening the web with TRC. By increasing the shear strength with
TRC, the service life of deficient concrete bridges can thus be extended efficiently.
INTRODUCTION
Many bridges in Germany and other European countries were built in the 1960s and ´70s,
making assessment, maintenance and refurbishment of the existing network increasingly
important 1. Furthermore, the German design rules have changed reducing the calculative
shear capacity of concrete members without shear reinforcement and increasing the re-
quired minimal shear reinforcement ratios for members with shear reinforcement. The
shear design of many existing bridges was performed using the principal tensile strength
criterion according to former design codes so that no shear reinforcement was necessary.
Current design codes 2 require a minimum shear reinforcement that exceeds the existing
shear reinforcement of many existing bridges. Hence, a large number of existing concrete
bridges in Germany is deficient according to current design codes.
As mentioned before, the shear design for the bridge superstructures in longitudinal direc-
tion was based on the principal tensile strength criterion according to the German code for
prestressed concrete 3,4. In contrast, the shear check according to the current design code
for concrete bridges 2 is based on the so-called “strut-and-tie model with crack friction” 5.
Since the load of the load model for design of bridges has been increased several times in
the past 6,7 because of rising traffic loads (Fig. 1) and due to the more conservative new
shear check, more shear reinforcement is now required in the web.
463
700 Highway Concrete;

600 Railway
Stone; 19,6%
Traffic Volume [10 tkm]
Prestressed Concrete;
500 Ship 0,6% 66,4%
9
485
400 Composite;
425 5,6%
300 346 380
236 Steel;
200
78 125 170 7,7%
115
100 72 62 73 76 86 94 Timber;
65
50 51 55 62 67 62 65 90 0,1%
0
a) b)
Fig. 1: a) Heavy Goods Traffic in Germany 8; b) Materials of highway bridges 9
Therefore, more structures are expected to be strengthened within the next years. Since a
replacement of these bridges is not possible or reasonable in many cases, the analysis,
evaluation and development of efficient strengthening methods becomes more important.
Many strengthening methods have proven to be suitable for the shear strengthening of
bridges, e.g. additional external prestressing, additional concrete layers, additional steel
reinforcement in slots or glued CFRP-stripes (Carbon Fiber Reinforced Polymer). Howev-
er, the applicability and effectiveness of these methods are also influenced by some disad-
vantages. Besides these common strengthening methods, the use of textile reinforced con-
crete (TRC) offers an innovative alternative for strengthening measures by combining the
advantages of lightweight glued CFRP-stripes and additional concrete layers, which pos-
sess better bond characteristics and lower temperature sensitivity. As the textile reinforce-
ment does not require protection against corrosion, thin layers are possible. For the above
reasons, full scale tests on I-shaped prestressed concrete (PC) beams (h = 0,7 m, l = 6,5 m)
under cyclic shear loading were carried out at the Institute of Structural Concrete at RWTH
Aachen University. Previous tests on non-strengthened beams served as a reference for the
tests on members strengthened with textile reinforced concrete.
EXPERIMENTAL INVESTIGATIONS
Test setup
Two full scale shear tests under cyclic loading were performed on I-shaped prestressed
concrete beams strengthened with Textile Reinforced Concrete (TRC). The experimental
results of the strengthened test beams were compared to similar test beams from previous
research projects without TRC which served as a reference 10,11. The test setup is shown
in Fig 2. The beams had a total length of 6,5 m and cross-section height of 0,7 m. The
point loads were located in the third points of the beam so that the shear slenderness of the
specimens amounted to a/d = 3,3.
464
Fig 2: Test setup and position of the point loads

The cross-section of the test beams had a total width of 0,6 m and a web width of 0,1 m.
The cross-section and the reinforcement layout of the test specimens are shown in Fig 3.
The first test was performed on a member without shear reinforcement strengthened by
several layers of TRC (I-O-5 TRC). The second test was performed on a member with a
low amount of shear reinforcement (ρw = 0,22%) which was also strengthened by TRC
(M-22-7 TRC). The test beams were subjected to 1,2 to 3,1 million load cycles using dif-
ferent highest and lowest loads.
600 600
6Ø10
110
6Ø10
110
StØ8/12
5
2Ø8 BüØ8/125
2Ø8
100
400
100 StØ6/25
400
TRC Layer TRC Layer (ρw=0,22%)

(25 mm) (25 mm)
15
2Ø8 2Ø8
95
95
StØ8/25 StØ8/25
190
190
95
95
6Ø10 6Ø10
a) 400 b) 400 [mm]
Fig. 3: a) cross-section of the test specimen without shear reinforcement (I-O-5 TRC) b) cross-section of
the test specimen with shear reinforcement (M-22-7 TRC)
Material properties
The steel reinforcement in each beam consisted of normal strength steel bars (fyk =
500 MPa). The beams were prestressed using two tendons, each consisting of three 0,6”
(15,2 mm) strands of prestressing steel St1570/1770 with a cross-sectional area of 3 × 140
mm². An additional layer of 25 mm of TRC was applied on each side of the web after pre-
stressing. Normal strength concrete with a compressive strength of fcm = 43 MPa similar to
many existing bridges was used. The concrete was planned to be similar to the test beams
without TRC (I-O-5 and M-22-7). However, the concrete strength of the members with
TRC turned out to be higher than of the previous beams. The compressive cylinder
strength fcm,cyl and the modulus of elasticity Ecm were measured on concrete cylinders
465
(300 mm height, 150 mm diameter) and the centric tensile strength fctm on core samples
(91 mm height, 45 mm diameter) of a small sample beam on the day testing started
(Tab. 1). Additionally, the yield strength f0,2, the tensile strength ft and the modulus of elas-
ticity Es of the Ø6 stirrups are also given in Tab. 1.
Tab. 1: Material properties

Concrete Steel
fcm,cyl fctm Ecm f0,2 ft Es ρw Pmt σcp,mt Vcrack
Specimen
[MPa] [MPa] [MPa] [MPa] [MPa] [MPa] [%] [kN] [MPa] [kN]
I-O-5 29,4 2,85 22700 - - - 0 320 1,78 176
I-O-5 TRC 42,5 2,90 26800 - - - 0 327 1,82 188
M-22-7 32,0 2,55 23900 587 626 200777 0,22 314 1,75 145
M-22-7 TRC 43,0 3,10 25100 595 633 203800 0,22 329 1,83 185
The shear reinforcement ratios of the test beams ρw, the prestressing force Pmt on the day
testing began and the initial shear crack force Vcrack are summarized in Tab. 1 for each
specimen. For the members without shear reinforcement Vcrack is a calculated value, since
the fatigue of the concrete was investigated. For the members with shear reinforcement
Vcrack is a measured value, since shear cracks occurred during the initial loading.
Textile reinforcement material

In order to ensure an adequate combination of matrix and textile reinforcement for the
strengthening of the beams, tensile tests were conducted on specimens with dimensions of
100 mm width, 880 mm length and a thickness of approximately 25 to 30 mm (Fig 4, b)).
Combinations of two different matrices (shotcrete with a maximum aggregate size of 4 mm
and polymer-modified dry-spray mortar (SPCC) with a maximum aggregate size of 2 mm)
and one to four layers of textile grid (alkali-resistant glass /carbon impregnated with epoxy
resin / styrene-butadiene / unimpregnated carbon grid) were investigated. In total, 32 ten-
sile tests were performed. The stress-strain curve of the two specimens V12-1 and V12-2
with SPCC and four layers of unimpregnated carbon grid (Fig 4, a)) is shown in Fig 4, c).
These test specimens exhibited the highest initial cracking stresses of all specimens (σc =
2,6 MPa). Moreover, the unimpregnated carbon grid was easy to apply during the pretests.
Therefore, unimpregnated carbon grid with an area of at = 55 mm²/m, which was also pre-
viously used in an innovative structural project 12, was used as textile reinforcement mate-
rial. Theoretically, single carbon fibres have a tensile strength of up to 3950 MPa with a
modulus of elasticity of 238.000 MPa. However, the actual tensile strength of unimpreg-
nated carbon grid in combination with concrete can be much low er. The mean tensile
strength of the carbon grid in these tests was σt = 1136 MPa.
466
Fig. 4: a) unimpregnated carbon grid; b) test setup for tensile tests; c) stress-strain curve of TRC with two
layers of unimpregnated carbon grid
Fig. 5: a) Web of specimen after sandblasting ; b) determination of depth of roughness; c) application of

shotcrete; d) application of textile reinforcement
Strengthening of the test beams

The strengthening of both beams was performed 2-3 weeks after prestressing. In the first
step both sides of the web were roughened by sandblasting (Fig 5a) after which the depth
of roughness was measured by applying gypsum to an area of 20 × 20 cm on the web (Fig
5b). The depth of roughness was calculated from the amount of plaster that was necessary
to create a planar surface on the web. The values had a range of 1,1 to 2,4 mm. In the se-
467
cond step, shotcrete and the textile reinforcement were applied alternately. Two layers of
textile reinforcement and three layers of shotcrete were applied on each side of the web of
both specimens (Fig 5c and d).
The used shotcrete was a polymer-modified dry-spray mortar (SPCC) with a maximum
grain size of 2 mm. The thickness of each layer of shotcrete averaged about 8 mm so that
the total thickness of the reinforcement amounted to about 25 mm. The mechanical proper-
ties of the shotcrete were determined by prism testing with a compressive strength fcm,prism
= 49,9 MPa and a flexural tensile strength of fctm = 5,6 MPa. After the strengthening proce-
dure the surface of the TRC layer was moisturised for at least three days in order to ensure
proper hydration.
Test program
In the first step, both specimens were subjected to an initial static loading until the highest
load Vmax was reached. The specimens were then subjected to cyclic loading with a fre-
quency of 1,0 to 2,5 Hz depending on the maximum deflection. The initial highest load
Vmax of the member without shear reinforcement (I-O-5 TRC) was defined as 75 % of the
theoretical shear crack load Vcrack according to the principal tensile stress criterion in order
to induce fatigue failure in the concrete. After 106 load cycles the maximum load was in-
creased to 90 % of Vcrack (Tab. 1). The test was then aborted after an additional 0,18·106
load cycles.
The specimen with shear reinforcement

Tab. 2: Cyclic loading on the specimens
(M-22-7 TRC) was subjected to a maxi-
Load cycles Vmax Vmin ΔV mum load of 110 % of the actually
Specimen ·103 [kN] [kN] measured shear crack load Vcrack (Tab. 1)
[kN]
in order to induce fatigue failure in the
Ni Ni
stirrups. The number of load cycles Ni
1000 1000 103 47 56
I-O-5 and the corresponding maximum and
1011 2011 102 35 67
minimum loads of each specimen are
1000 1000 141 79 62 summarized in Tab. 2. After the cyclic
I-O-5 TRC
180 1180 171 109 62 testing the remaining shear capacity of
M-22-7 1853 1853 160 103 57 both specimens was determined.
2000 2000 204 147 60
M-22-7 TRC
1100 3100 204 118 86
The applied load was determined by measuring the hydraulic pressure in the jack. Two
displacement transducers were placed in the third points to measure the deflections of the
beams. Photogrammetry was used for measuring crack widths in the web in a rectangle of
300x225 mm (Fig 6a). Four sets of three displacements trancducers were used on the web
to measure crack widths and the angle of the principal stresses. In addition, the concrete
strains on the surface of the TRC were measured by strain gauges (Fig 6b).
468
Fig. 6: a) Photogrammetry for determining shear crack widths; b) displacement transducers and strain
gauges on the web
The deflection of the beams during the cyclic loading is illustrated in Fig 7. It shows that
the beam I-O-5 TRC exhibits larger deflections than the beam I-O-5 due to the highest
load, which was increased by 40 %. However, the beam strengthened with TRC did not
show any signs of fatigue failure. In the following step the loading was increased further so
that the maximum load was almost equal to the initial shear crack load of I-O-5. At this
level the original beam without TRC would have failed immediately, whereas the beam
with TRC sustained another 180.000 load cycles. Although additional load cycles would
have been possible, the testing was aborted at this point due to large deflections. The re-
maining capacity of the specimen I-O-5 TRC was determined to Vu = 233 kN, whereas the
original specimen without TRC had a remaining capacity of Vu = 158 kN.
20,0 25,0
I-O-5 TRC Modification of
test setup Increase of
Deflection [mm]
I-O-5 20,0
Deflection [mm]
15,0 loading
Vmax/min = 204/147 kN
Increase of 15,0
Vmax/min = 141/79 kN loading Vmax/min = 204/118 kN
10,0
10,0
Modification of
Vmax/min = 160/103 kN
5,0 test setup M-22-7 TRC
Vmax/min = 103/47 kN 5,0 M-22-7
0,0 0,0
0 500.000 1.000.000 1.500.000 0 1.000.000 2.000.000 3.000.000
a) Load cycles b) Load cycles
Fig. 7: Deflection at maximum load subject to the number of load cycles for a) beams without shear rein-
forcement; b) beams with shear reinforcement
The highest load of the specimen M-22-7 TRC was increased by about 30 % compared to
the previous specimen M-22-7. In the previous experiment various stirrups failed during
the first 106 load cycles which can be seen from the progression of the load deflection
curve (Fig 7b). In contrast, the beam strengthened with TRC did not exhibit any damage on
the stirrups after 2⋅106 load cycles, after which the amplitude was increased further. After
the increase of the amplitude some stirrups failed and the deflection grew moderately.
Even then, the beam was able to sustain another 1,1⋅106 load cycles after which the test
469
was aborted. The remaining capacity of the beam M-22-7 TRC amounted to Vu = 350 kN.
The remaining of the original beam M-22-7 was not determined due to its considerable
damage in the stirrups. However, another previous test beam M-22-3 with the same pre-
stressing but subjected to smaller highest loads had a remaining capacity of Vu = 264 kN. It
can therefore be concluded that the TRC strengthening had a considerable effect on the
remaining shear capacity for the beams with shear reinforcement as well.
CONCLUSIONS AND FUTURE WORK

The use of textile reinforced concrete (TRC) offers an innovative alternative for strength-
ening measures by combining the advantages of light glued CFRP-stripes and the better
bond characteristics of an additional concrete layer. A possible field of application has
been investigated and described in the paper: Firstly, a considerable increase of the shear
fatigue strength can be obtained by strengthening the web with TRC. Secondly, the static
shear capacity also increases considerably due to the TRC strengthening. In addition to this
preliminary comparison of the tests, a more detailed investigation is required regarding the
actual fatigue design check of beams strengthened with TRC. Here, further experimental
investigations can show to what extent the models for the fatigue design of prestressed
concrete beams with and without shear reinforcement can be applied if textile reinforced
concrete is used.
REFERENCES
1. Naumann, J. (2010), “Brücken und Schwerverkehr - Eine Bestandsaufnahme“ (“Bridges and Heavy
Goods Traffic – an Inventory“), Bauingenieur Vol. 85, No.1, pp.1-9.
2. DIN EN 1992-2 (2013), “Eurocode 2: Design of concrete structures - Part 2: Concrete bridges - De-
sign and detailing rules; German version, April 2013.
3. DIN 4227 (1953), German Design Code “Prestressed concrete - Guidelines for design and construc-
tion“.
4. BMV (1969), German Design Guideline: “Additional provisions to DIN 4227 for prestressed con-
crete bridges“, Federal Ministry of Transport, Building and Urban Development.
5. Reineck, K.-H. (2001), “Hintergründe zur Querkraftbemessung in DIN 1045-1 für Bauteile aus
Konstruktionsbeton mit Querkraftbewehrung“, Bauingenieur 76, Iss. 4, pp. 168-179.
6. DIN Fachbericht 101 (2009), German Design Code “Actions on Bridges”.
7. DIN EN 1991-2 (2012), “Eurocode 1: Actions on structures - Part 2: Traffic loads on bridges; Ger-
man version, August 2013.
8. Ickert, L. (2007), “Abschätzung der langfristigen Entwicklung des Güterverkehrs in Deutschland bis
2050“, Final Report for the Federal Ministry of Transport, Building and Urban Development of
Germany, Basel.
9. Kaschner, R.: Auswirkungen des Schwerlastverkehrs auf die Brücken der Bundesfernstraßen. Heft
B68, Bundesanstalt für Straßenwesen (BASt), 2009.
10. Teworte, F.; Hegger, J (2013a), “Fatigue of prestressed beams without web reinforcement under cy-
clic shear“.Beton- und Stahlbetonbau 108, Iss. 1, pp. 34 - 46.
11. Teworte, F.; Hegger, J (2013b), “Fatigue of prestressed beams with web reinforcement under cyclic
shear“.Beton- und Stahlbetonbau 108, Iss. 7, pp. 475 - 486.
12. Scholzen, A. et al.: Thin-walled shell structure made of textile reinforced concrete: design, dimen-
sioning and realization”, Beton- und Stahlbetonbau 107, Iss. 11, pp. 767–776.
470
EVALUATION OF THE FIB BULLETIN 14 DESIGN GUIDELINE

FOR EXTERNAL TRC BENDING REINFORCEMENT
Svetlana Verbruggen1, Jan Wastiels1, Tine Tysmans1, Sven De Sutter1, Maciej Woźniak1, 1
Department of Mechanics of Materials and Constructions, Vrije Universiteit Brussel, Belgium
Abstract: Due to their advantages related to fire safety and their relatively low cost, Textile
Reinforced Cement (TRC) composites are emerging as external reinforcing materials for
concrete structures. The current design guidelines for external reinforcement are mostly
elaborated for the Carbon Fibre Reinforced Polymers (CFRP) material solutions. Consider-
ing the major differences between both the CFRP and TRC materials, the validity of these
CFRP design guidelines for the TRC external reinforcement should be verified and eventu-
ally the design rules should be adapted. This paper contributes to the evaluation of the Eu-
ropean FIB bulletin 14 design guideline for TRC by comparing 2.3-meter-span externally
reinforced concrete beams designed according to the FIB bulletin 14 with experimental
results from four point bending tests. The results show that the design guidelines underes-
timate the ultimate load by almost 20 % due to a difference in occurring failure mode (ex-
perimental: failure in composite action; predicted: failure by loss of composite action).
These observations indicate that the FIB design guideline cannot be used to accurately pre-
dict the bending behaviour of a TRC-reinforced beam and thus that adapted calculation
techniques are needed.
INTRODUCTION
Recent developments in Textile Reinforced Cements (TRC) have led to the possibility to
use these materials in structural applications such as the external reinforcement of concrete
structures. The use of a continuous fibre reinforcement allows a controlled fibre distribu-
tion and the possibility to align the fibres along the principal stress directions. All papers
studying the use of these TRCs as external shear [1-6] and bending [7-11] reinforcement
indicate the feasibility of the concept and the high potential of this technique. In compari-
son with the existing external reinforcing techniques like Carbon Fibre Reinforced Poly-
mers (CFRP), the use of TRCs can present some advantages like the material fire re-
sistance, and the lower material cost. However, most of the studies on TRCs are not able to
reach the same loadbearing performance as the current CFRP solution. This is mostly the
consequence of the use of open grid textile structures, resulting in a relatively low fibre
volume fraction (typically 5 %). These open grids are used for manufacturing reasons, as
the mortar can easily penetrate the fibre structure through the grid openings.
At the Vrije Universiteit Brussel (VUB) Inorganic Phosphate Cement (IPC) [12] (commer-
cially available under the name Vubonite [13]) was developed. This is a cementitious ma-
terial which is suitable to be used as a matrix material for standard E-glass fibre compo-
sites due to its neutral pH after hardening. IPC consists of a powder-liquid combination
471
which hardens at room temperature and becomes a strong, durable, heat resisting and fire
safe material. Its relatively small grain size (between 10 and 100 µm) enables the impreg-
nation of dense fibre textiles up to high fibre volume fractions (up to 25 % [14]). As a re-
sult, a cementitious composite with high tensile (up to 60 MPa for IPC reinforced with
randomly oriented glass fibre textiles) and compressive (80 MPa) capacities and which is
heat- and fire resistant (highest European class A1), is created, what makes it appropriate
for structural applications and thus also for external reinforcement of concrete structures.
In previous studies the authors of this paper already indicated the possibilities of this high
fibre volume fraction cementitious composite as an external bending reinforcement for
concrete beams [15-17]. However, no appropriate design guidelines exist for external rein-
forcement made of IPC TRC, or any type of high fibre volume fraction TRC in general.
Therefore this paper verifies the validity of the CFRP design guidelines for TRC external
reinforcement. The applicability of the European FIB bulletin 14 design guideline [18] will
be verified for IPC TRC as an external bending reinforcement for concrete beams.
DESIGN PRINCIPALS
In the recent decades several regional design codes and guidelines, such as the FIB bulletin
14 (Europe), ACI-guide (USA), JSCE-recommendation (Japan) and CSA-code (Canada)
are published concerning the calculation of externally reinforced concrete structures. These
design guidelines are mostly elaborated for the CFRP strengthening and repair technique.
Generally the design of externally reinforced concrete structures is similar to the design of
internally steel reinforced alternatives (Eurocode 2 [19]). However, special attention needs
to be paid to the loss of composite action (and thus loss of connexion) between the external
reinforcement and the concrete substrate, mainly due to the low tensile strength of the con-
crete material. Depending on different phenomena causing shear en tensile forces in the
bond zone, the debonding can initiate at different places along the beam’s length: the end
anchorage, flexural or shear cracks or by end shear failure (Figure 1). The FIB bulletin 14
presents different design guidelines depending on the debonding location. Considering the
complexity of these debonding phenomena and the wide variety of available calculation
methods, the FIB bulletin 14 offers the user the choice between three significantly different
calculation approaches for the debonding at the end anchorage and at flexural cracks.
Figure 1: Depending on different phenomena causing shear en tensile forces in the bond zone, the debonding
can initiate at different places along the beam’s length.
472
The fact that the current design for debonding is based on experiments performed on the
CFRP material and that there exist severe variances in the constituents of IPC TRC and
CFRP composite materials, could lead to new failure types and mechanisms for the IPC
TRC external reinforcement. Additionally, mayor differences exist in the materials tensile
stress-strain behaviour as illustrated in Figure 2. Apart from a considerable difference in
achievable tensile strength, a deviating overall evolution is obtained: the CFRP composite
shows a linear stress-strain behaviour, while the IPC TRC exhibits three different stages, as
illustrated in the upper left corner of Figure 2.
Figure 2: Considerable differences exist in the tensile stress-strain behaviour of CFRP and IPC TRC.
Test set-up
Four point bending tests with third point load are performed on reinforced concrete beams
with a span of 2.3 m (total length of 2.5 m), a height of 0.3 m and a width of 0.2 m. The
loading is displacement controlled using a servo-hydraulic actuator with an initial dis-
placement rate of 0.2 mm/min. After the cracking moment is clearly surpassed (load of 60
kN), the displacement rate is increased to 2 mm/min. Figure 3 illustrates the test set-up.
Three different beam types are tested. The first beam type is a reference reinforced con-
crete beam without any external reinforcement. The second type is a steel-reinforced con-
crete beam that is externally strengthened by gluing a strip made of IPC reinforced with 16
glass fibre mat layers (resulting in a nominal thickness of 8 mm). To simulate actual load-
ing conditions a third beam type is tested where a steel-reinforced concrete is precracked
before it is repaired with the same amount of externally bonded IPC TRC as for the second
beam type. In both cases the external reinforcement is applied over the entire tensioned
lower surface of the beam and thus continues over the supports. The IPC TRC is glued
onto the concrete using a two-component epoxy glue (PC 5800/BL [23]). Before attaching
the reinforcement to the beams, the concrete surface is pretreated to remove the laitance
layer. All beams have an internal steel reinforcement that consists of two longitudinal bars
with a diameter of 16 mm and stirrups with a diameter of 6 mm placed every 100 mm in
473
the shear zones of the beams. This shear reinforcement is overdimensioned in order to
cause a bending failure in the zone of constant moment. The reinforcement bars are welded
to steel plates at the end of the beams; this eases the positioning and placing of the rebars
and eliminates potential problems with the anchoring length of the reinforcement bars. The
stirrups are welded to the longitudinal bars as well. Apart from Linear Variable Differential
Transformers (LVDT), monitoring the vertical displacement of the beams, the Digital Im-
age Correlation (DIC) measuring technique is applied. Two cameras follow the side edge
of the beam and are able to measure a 0.4 m wide zone within the constant moment area at
the centre of the beam. The analysis is done using the VIC3D-2009 software package from
Correlated Solutions where a subset size of 21 pixels, a step size of 5 pixels and a strain
window size of 11 are used.
Figure 3: Three beam types are tested in a four point bending test with third point loading.
Materials
Concrete
The concrete has been mixed with the following mass proportions: 375 kg Portland cement
CEM I, 32.5R, 210 kg water, 690 kg sand (0/2), 1125 kg gravel (6/14). This results in a
compressive strength of 35.0 MPa, a Young modulus of 34.0 GPa and a modulus of rup-
ture of 5.3 MPa. The values are the mean values of 8 specimens for the compressive
strength and 3 for the modulus of rupture. These material characteristics are experimentally
determined after 48 days. The compressive strength is the cylindrical strength calculated
from tests performed on cubic shaped specimens, with sides of 150 mm. The young modu-
lus can be calculated from the compressive strength. All calculations are based on [19, 24].
474
IPC TRC
The matrix material IPC is obtained by mixing Vubonite liquid component and high per-
formance Vubonite powder in the mass proportion of 1/0.82. The IPC matrix is reinforced
with 16 randomly in-plane oriented fibre textiles, being chopped strand mats Vetrotex M5,
with a surface density of 300 g/m², resulting in a fibre volume fraction of 21 %. As such a
tensile strength of 58.4 MPa, an ultimate strain of 1.2%, a Young modulus for the stage I
of 12.5 GPa and a Young modulus for the stage III of 4.8 GPa are obtained. These charac-
teristics are the mean values of 10 specimens.
Steel reinforcement
The traditional steel reinforcement consists of ribbed bars made of S500 steel, of which the
material characteristics are not experimentally derived, so the standard values of 500 MPa
for the yield stress and 200 GPa for the young modulus are assumed.

Comparison analytical and experimental load-deflection behaviour
The load-deflection curves (Figure 4) show that both externally reinforced beams reach a
37.6 % higher ultimate load compared to the reference beam (209 kN versus 152 kN), in-
dicating the additional capacity and thus the potential of the IPC TRC external bending
reinforcement. The effectiveness of an IPC TRC patch as a repair reinforcement method is
proven by the equal loadbearing capacity of the pre-cracked and non-pre-cracked beam.
The presence of pre-cracks only influences the stiffness of the initial linear uncracked zone
of the curve and thus shows the independency of the history of the concrete structure. The
pre-cracked beam exhibits a lower stiffness in this initial stage due to the cracks which are
already present. The influence of pre-cracking a concrete beam prior to the application of
IPC TRC external reinforcement is already discussed in detail in [15]. Just as in the previ-
ous studies using IPC TRC as an external reinforcement on small scale beams [15-17], the
load at which the externally reinforced beams lose their high initial stiffness (50 kN) is 43
% higher than for the reference beam (35 kN). This stiffness retention results in an up-
wards shift of the second part of the curve, causing a lower deflection for the same applied
load.
475
Figure 4: Comparison of the experimental load-deflection behaviour indicates the effectiveness of a TRC
repair patch and confirms the retention of initial high stiffness.
The dotted lines in Figure 4 represent the analytical predictions of the load-deflection be-
haviour, based on the integration of the curvature [18, 25]. Contrarily to the previous ob-
servations [15-17], the load at which the high initial stiffness is lost for the externally rein-
forced beams is quite close to the calculated value of 47.7 kN. Considering the fact that the
reference beam is not capable of reaching its predicted cracking load of 46.7 kN, this devi-
ating observation is rather attributed to an overestimation of the tensile strength of the con-
crete. Besides the experimental cracking moment, also the post-cracking stiffness of the
externally reinforced beams is higher than the one of the non-externally-reinforced refer-
ence beam, which is caused by the presence of a greater total amount of reinforcement and
thus stiffness hereof. The stiffnesses and their respective increase are well predicted by the
analytical calculations.
Comparison analytical and experimental failure behaviour
As predicted by the analytical model, the reference beam fails by steel yielding and con-
crete crushing at a load of 152 kN (analytical: 140 kN). The application of IPC TRC rein-
forcement over the entire bottom surface results in a failure load of 209 kN. At this ulti-
mate load the externally reinforced beams exhibit several failure modes: steel yielding and
concrete crushing, debonding at flexural cracks and debonding at the end anchorage. All
these failure modes are indicated in Figure 5.
476
Figure 5: At the ultimate load several failure Figure 6: The difference in vertical displacement sud-
modes are present for the IPC TRC-reinforced denly increases at 165 kN, but debonding already starts
RC beams. at 80 kN.
As long as the IPC TRC is bonded to the concrete, the vertical displacement of the con-
crete substrate and the external reinforcement are identical, so the difference between both
is close to zero. From the moment debonding starts the IPC TRC will have a larger down-
ward displacement and thus the mutual difference will increase. DIC measurements on
these beams (Figure 6) indicate that the first difference in vertical displacement between
the concrete and the IPC TRC is noticed at a load of about 80 kN, and thus that debonding
already starts at this load. However this difference remains limited up to a load of 165 kN,
introducing the final failure.
The analytical calculations following the FIB bulletin 14 predict a composite failure at the
load of 208.9 kN determined by tensile failure of the IPC TRC. Approaches 1 and 2 of the
FIB bulletin 14 predict debonding at flexural cracks at a load of 179.3 kN and 167.9 kN
respectively. No debonding at the end anchorage is predicted. The debonding at shear
cracks, predicted at a load of 197.6 kN, is not taken into account, considering the overdi-
mensioned shear reinforcement. The observed failure load of 208.6 kN corresponds really
close to the predicted composite failure load (208.9 kN), but exceeds the analytical pre-
dicted debonding load by 20 %. The predicted debonding value corresponds rather to the
sudden differential increase in vertical displacement between the concrete and the IPC
TRC, although it does not lead to beam failure.
One of the causes of this underestimation of the analytical force might be the continuation
of the external reinforcement over the supports in the experiment, whereas the FIB bulletin
14 is rather intended for the strengthening and repair scenario, where the external rein-
forcement stops at a certain distance from the supports. Additionally, the FIB bulletin 14
corresponds to a safe and conservative design, where an underestimation is strongly pre-
ferred over an overestimation. However, these results indicate the significant difference
between the design for CFRP strengthening and external bending reinforcement made of
IPC TRC, or any other high fibre volume fraction TRC in general. This demonstrates the
need for adapted design guidelines for external TRC reinforcement.
477
CONCLUSIONS
Four point bending tests on 2.3-meter-span externally TRC reinforced concrete beams in-
dicate an increase in ultimate load and in load at which the initial high stiffness is lost
compared to the reference non-externally reinforced concrete beam. Additionally, the in-
dependency of this external reinforcing technique on the beam history is shown, as the pre-
cracked and non-pre-cracked alternatives exhibit a similar load-deflection behaviour.
Comparison of the experimental results with their analytical predictions according to the
European FIB bulletin 14 design guideline, yields a correct prediction of the beam stiffness
and the stiffness increase with respect to the reference beam. Nevertheless, the FIB bulletin
14 underestimates of the ultimate load by almost 20 %, partially due to a continuation of
the reinforcement over the supports and a conservative design philosophy. Still, this under-
estimation indicates the need for adapted calculation techniques.
ACKNOWLEDGMENTS
Research partially funded by a scholarship of the Institute for the Promotion of Innovation
through Science and Technology in Flanders (IWT-Vlaanderen) (first author), by the Brus-
sels Capital Region through the Innoviris Strategic Platform Brussels Retrofit XL (first and
fourth author) and by Fonds Wetenschappelijk Onderzoek-Vlaanderen (FWO) (fifth au-
thor). The authors gratefully acknowledge the cooperation with the company TRADECC,
trough the delivery of the epoxy glue.
REFERENCES
[1] Triantafillou, T.C. and Papanicolaou, C.G., “Shear strengthening of reinforced concrete members
with textile reinforced mortar (TRM) jackets”, Mater Struct, Vol.39; p.93-103; 2006.
[2] Brückner, A., Ortlepp, R. and Curbach, M.; “Textile reinforced concrete for strengthening in bend-
ing and shear”; Mater struct; Vol.39; p.741-748; 2006.
[3] Brückner, A., Ortlepp, R. and Curbach, M.; “Anchoring of shear strengthening for T-beams made of
textile reinforced concrete (TRC)”; Mater struct; Vol. 41; p.407-418; 2008.
[4] Blanksvärd, T., Täljsten, B. and Carolin, A.; “Shear strengthening of concrete structures with the use
of mineral-based composites”; J compos constr; Vol.12, No.1; p.25–34; 2009.
[5] Contamine, R., Si Larbi, A. and Hamelin, P.; “Evaluation of the TRC solutions in the case of RC
beams shear strengthening”; Proceedings of SHCC2; Rio de Janeiro, Brazil; p.89-95; 2011.
[6] Contamine, R., Si Larbi, A. and Hamelin, P., “Identifying the contributing mechanisms of textile
reinforced concrete (TRC) in the case of shear repairing damaged and reinforced concrete beams”,
Eng struct, Vol.46, p.447-458, 2013.
[7] Täljsten, B. and Blanksvärd, T.; “Mineral-based bonding of carbon FRP to strengthen concrete
structures”; J compos constr; Vol.11, No.2; p.120-128; 2007.
[8] Bisby, L.A., Roy, E.C., Ward, M. and Stratford, T.J.; “Fibre reinforced cementitious matrix systems
for fire-safe flexural strengthening of concrete: Pilot testing at ambient temperature” Proceedings of
ACIC, Network Group for Composites in Construction; Chesterfield, U.K.; 2009.
[9] Ombres, L.; “Flexural analysis of reinforced concrete beams strengthened with a cement based high
strength composite material”; Compos struct; Vol.94; p.143-155; 2011.
[10] D’Ambrisi, A. and Focacci, F.; “Flexural strengthening of RC beams with cement-based compo-
sites”; J compos constr; Vol.15, No.5; p.707-720; 2011.
[11] Ombres, L.; “Debonding analysis of reinforced concrete beams strengthened with fibre reinforced
cementitious mortar”; Eng fract mech; Vol.81; p.94-109; 2012.
[12] European Patent Office “EP 0 861 216 B1, Inorganic Resin Compositions. Their Preparation And
Use Thereof”. May 2000.
[13] www.vubonite.com
478
[14] Remy, O. and Wastiels, J.; “Development of impregnation technique for glass fibre mats to process
textile reinforced cementitious composites”, Plast Rubber Compos, Vol.39, No 3-4-5, p.195-199;
2010.
[15] Verbruggen, S., Tysmans, T. and Wastiels, J.; “TRC or CFRP strengthening for reinforced concrete
beams: An experimental study of the cracking behaviour”; Eng Struct, Vol.77, p.49-56; 2014.
[16] Verbruggen, S., Wastiels, J., Tysmans, T., Remy, O. and Michez, S.; “The influence of externally
bonded longitudinal TRC reinforcement on the crack pattern of a concrete beam”; Proceedings of
ICCRRR, Cape Town, South Africa, p.1259-1265, 2012.
[17] Verbruggen, S., Wastiels, J., Tysmans, T. and Puystiens, S., “Comparison between TRC and CFRP
as external reinforcement for plain concrete beams”; Proceedings of ICCM19; Montreal, Canada;
ISBN: 978-0-9696797-1-4; p.2252-2260; 2013.
[18] CEB-FIP; “fib bulletin 14 Externally bonded FRP reinforcement for RC structures”; Lausanne,
Switzerland; ISBN 2-88394-054-1; 2001.
[19] CEN (Comité Européen de Normalisation), “Eurocode 2: Design of concrete structures – Part 1-1:
General rules and rules for buildings”; EN 1992-1-1; 2005.
[23] TRADECC, PC 5800/BL, http://norborn.no/wp-content/uploads/2011/12/PC-5800-BL-TD.pdf
[24] Belgisch instituut voor de normalisatie (BIN). Proeven op beton; Drukproef; addendum 1”, Belgi-
sche Norm, Brussels, August 1973.
[25] Matthys S.; “Structural behaviour and design of concrete members strengthened with externally
bonded FRP reinforcement”; Doctoral thesis, Universiteit Gent, Faculty of Engineering; Gent, Bel-
gium; 2000.
479
480
UTILISATION OF LINEAR TRACTION SEPARATION BOND LAW

TO IMPROVE THE NUMERICAL SIMULATION OF THE BENDING
BEHAVIOUR OF HYBRID TRC-CONCRETE BEAMS
Maciej M. Wozniak (1), Tine Tysmans (1), John Vantomme (1,2), Svetlana Verbruggen (1) and
Sven de Sutter (1), (1) Department of Mechanics of Materials and Constructions, Faculty of
Engineering, Vrije Universiteit Brussel, Brussels, Belgium, (2) Department of Civil and Materials
Engineering, Royal Military Academy, Brussels, Belgium
Abstract: The use of Textile Reinforced Concrete/Cement (TRC) materials in the building
industry has recently been widely investigated. Among all of the utilisations of TRC, the
most researched ones are strengthening and retrofitting of existing concrete structures as
well as stay-in-place formwork for new concrete construction. To model the mechanical
behaviour in such applications, the bond between TRC and concrete is a crucial parameter.
Therefore this paper proposes to numerically model the bond by the linear traction separa-
tion law, and investigates the influence of the essential parameters (bond strength, maxi-
mum slip, initial stiffness) on the global structural response of a hybrid TRC-concrete
structure. Subject of both simulation and physical experiment are concrete beams longitu-
dinally reinforced with TRC and subjected to four-point bending. The initial bond parame-
ters are identified from shear tests performed on small-scale specimens. As comparison
between numerical model and experiments shows, the inclusion of the bond behaviour
with the linear traction separation law improves the simulation of the mechanical behav-
iour of the hybrid beams. The results of the numerical simulation prove moreover a great
importance of the bond parameters on the overall stiffness of the beams. In the future, it
would be interesting to examine if – and up to what level - the non-linear traction-
separation (bond-slip) model can improve the simulation even more.
INTRODUCTION
Textile Reinforced Concrete/Cement (TRC) is recently widely investigated as an alterna-

tive solution to Fibre Reinforced Polymers (FRP) for externally strengthening reinforced
concrete structures. The main advantage of the TRC system compared to the FRPs is its
better fire behaviour – higher fire resistance, better thermal compatibility and lack of emit-
ting toxic fumes during fire [1]. This results in the high research interest for this material
for various building applications:t, flexural strengthening [2][3][4], shear strengthening
[5][6][7][8], column confinement [9][1], as a stay-in-place (SIP) formwork [9][10][11] for
concrete structures and as well as masonry reinforcement [12][13]. All these publications
indicate that the bond between TRC and concrete is an important failure mechanism. Con-
sequently, this research is dedicated to the investigation of the bond behaviour.
481
The state-of-the-art on the bond between TRC and concrete consist of analysis of direct
shear tests [14][15][16] as well as full beam tests [17][18]. These results obtained by dif-
ferent research groups are difficult to compare as the mechanical bond behaviour greatly
depends on the composition of the applied TRC. However, the debonding mechanism
clearly differs from the one of FRP: TRC delaminates in the first layer of the reinforcing
fibre grid or directly at the interface between TRC and concrete, while the most common
failure for FRP-reinforced concrete structures is due to tensile failure in the concrete sub-
strate [19]. Even though models of bond between TRC and concrete are developed analyti-
cally [16][20], there is a lack of numerical simulations for this bond behaviour.
In this paper, the authors tackle the problem of the bond between TRC and concrete using
a linear traction-separation law that can be found in commercially available software such
as ABAQUS. The necessary parameters for the modelling of the materials and the bond are
obtained by laboratory tests done by the same authors in previous research [21]. Finally,
the model is validated by small-scale beam tests. The paper is divided into two parts: the
first part focuses on the experimental set-up of four-point bending on plain concrete beams
(65cm length) externally reinforced with TRC. The second part presents the numerical
simulation of these tests and provides a detailed description of the model parameters and
their identification.
FOUR-POINT BENDING ON SMALL SCALE BEAMS
Motivation
Despite the recent research on the bond between TRC and concrete [14-20], the interaction
between both materials remains uncertain. Additionally, the TRC used in this previous
research had a limited fibre volume fraction (<2%). Recent developments in TRCs have
led to the possibility to impregnate denser fibre structures and thus achieve a significantly
higher fibre volume fraction (>20%). In order to understand this bond behaviour and to be
able to focus on higher fibre volume fraction TRCs, our research group has performed
bending tests on plain concrete beams externally reinforced by gluing 4mm thick TRC
plates on the tensed bottom side. Such configuration eliminates additional influences of the
internal (steel) reinforcement, leading to the pure contact between TRC and concrete.
Used materials
Concrete
In order to simulate the conditions of existing old structures, composed of concrete with
poor mechanical properties, a low strength concrete mix was prepared. The mean charac-
teristics were experimentally derived: compressive strength 16.8MPa, modulus of rupture
2.5MPa and Young modulus 28.2GPa.
TRC
The cementitious matrix was made of Inorganic Phosphate Cement (IPC) and reinforced
with 2D random chopped strand mats composed of E-glass fibres. The IPC was developed
at the Vrije Universiteit Brussel and is based on a two-component system, consisting of a
482
calcium silicate powder and a phosphate acid based solution of metaloxides [22]. This IPC
matrix is a fine grained mortar, enabling the impregnation of dense fibre textiles, resulting
in high fibre volume fractions (>20%). The choice of the randomly oriented strand mats is
motivated by the possibility of using additional bolts for improving connection properties.
The compressive behaviour of this TRC is linear up to failure (around 80MPa). The tensile
behaviour on the other hand is significantly different as shown in the Figure 1. The initial
stiffness of the TRC is high (15.4GPa) as a result of the composite action between matrix
and fibres. At the value of about 7 MPa the matrix cracks and the behaviour of the TRC
depends only on the fibre reinforcement, resulting in a much lower post-cracking stiffness
(4.4GPa) than the initial one.
Figure 1 Nonlinear tensile behaviour of the modelled TRC
Specimen preparation
Seven beam types were tested, varying in concrete surface preparation and the presence of
additional bolt connections. At least two specimens per type were examined. The concrete
surface pre-treatments, prior to the application of the TRC, were:
- None – no additional concrete pre-treatment was done
- Diamond saw – removing of the bottom concrete laitance layer by a diamond saw
(less than 10mm)
- Chemical treatment – additional to the diamond saw pre-treatment, application of
12% HCl-solution for 15 minutes, proceeded by water cleaning and neutralising by
a 1% ammoniac solution
- Extra roughening - additional to the diamond saw pre-treatment, the surface was hit
by pick
For each pre-treatment type (not for the non-pre-treated one) an alternative beam type with
and without additional bolt connections was investigated. The bolts were located 110mm
from the ends of the beam and in the middle of the width. 45mm deep holes were drilled,
in which a thread rod (Φ 12) was fastened using a chemical anchor. For all beam types the
external TRC was glued to the concrete substrate using a commercial two-component
epoxy glue (PC 5800/BL; produced by TRADECC).
483
Previous research indicated the independency of the mechanical behaviour on the pre-
treatment type [21][23]. Therefore only the diamond saw pre-treated beams will be repre-
sented in this paper, in order to maintain the overview.
Experimental set-up
The beams were tested in four point bending. Their cross-section was 90 x 80mm, while
their length equalled 650mm, with a distance between the roller supports of 600mm with a
load application every 200mm. The displacement was measured by an LVDT placed in the
middle of the beams span.
NUMERICAL SIMULATION
Geometry and boundary conditions
The specimens were simulated in the finite element software ABAQUS with correspond-
ing dimensions and material properties as in the laboratory experiment. The symmetry
boundary conditions were applied to save computing time, so only a quarter of the speci-
men was considered. The supports were modelled as rigid half-cylinders at the correspond-
ing locations. The interaction between the specimen and these supports was considered as a
default ‘hard’ contact. Concrete was modelled with solid elements by using the Concrete
Damaged Plasticity model, with the direct tensile and compressive strength obtained from
experiments (as indicated in “Used materials”). The TRC plates were simulated by shell
elements also with the Concrete Damaged Plasticity material law and with the parameters
described above in “Used materials” and identified by the authors in earlier research [24].
The interaction (and thus bond) between TRC and concrete is described in the following
section.
Modelling of the bond

The bond between TRC and concrete, experimentally achieved with an epoxy glue, was
modelled with a surface-based cohesive behaviour, available within the finite element
software ABAQUS. This model assumes a linear traction-separation law, which is
common simplification to the real-life behaviour of an epoxy glue. Such a model enables
fast calculations with a limited amount of parameters. An elastic relation associates the
normal (tn) and shear stresses (ts, tt) with respectively the normal (δn) and shear (δs, δt)
separations (relative displacements between both surfaces) across the bonded surface (see
Equation 1) by the bond stiffness. The assumed stiffness (Knn=150; Kss= Ktt=36; Kns= Knt=
Kst=0 MPa/mm) was found by the authors in previous experiments [25].
[26] (1)
The model enables the simulation of bond failure by defining a maximal shear stress (tmax)
for all stress components and a maximal slip between the bonded surfaces (δmax). In this
case the values of tmax = 0.58MPa and δmax = 0.7mm were adapted [25].
484
Load application
The loading was applied as displacements, located as in the laboratory experiment. The
values of the applied forces were retrieved as reactions within the supports. This method-
ology allows the use of forces within the displacement-force diagrams not as prescribed
values, but as an output of the model. This enables the prediction of the ultimate force, as
all of the materials’ models include a post-failure, softening behaviour.

The results are divided into two parts: the first, dedicated to the specimens without bolted
connections, assuming bond properties as mentioned in the previous section; and the se-
cond, describing specimens with bolted connections, where the bond was assumed to be
perfect. The comparison between the experiments and the numerical simulation is done by
matching the diagrams of the applied force to the deflection in the middle of the beams
span.
Specimens without bolt connections

Four specimen types (ten specimens) without bolts were tested. All of the specimens failed
by debonding at a similar ultimate load (average 10.9kN). The figure below shows a repre-
sentative experimental load-deflection curve (diamond saw pre-treatment) in comparison
with the numerical simulation.
Figure 2 Applied force to midspan deflection - good correlation between the numerical simulation including
bond properties and laboratory experiment of the specimens without bolt connections
The numerical model presents a good simulation of the experimental mechanical behav-
iour: the initial stiffness as well as the softening of the composite beam due to concrete
cracking is predicted correctly. Moreover, the ultimate load is similar for both the simula-
tion (11.4kN) and the experiment (10.9kN). The major drawback of the simulation is the
lack of ductility, which is achieved by the experimental beam.
485
Specimens with bolt connections

Three specimen types (six specimens) with additional bolt connections were tested. All of
the specimens failed by concrete crushing and tensile failure of the TRC reinforcing plate.
Additionally, debonding between the reinforcement and the concrete substrate occurred,
but was stopped by the bolts and thus led to a substantially higher ultimate load (average
15.6kN). That is why the bond in the simulation was modelled as perfect. The figure below
shows a representative experimental load-deflection curve (diamond saw pre-treatment) in
comparison with the numerical simulation.
Figure 3 Applied force to midspan deflection - good correlation between the numerical simulation assuming
perfect bond and laboratory experiment of the specimens with bolt connections
Similar to the previous simulation, the model is able to predict the experimental mechani-
cal behaviour. Both the initial stiffness and the softening due to concrete cracking are well
grasped by the model. The experimental ultimate force (15.6kN) is reasonably close to the
simulated value of 17.8kN. However, the ultimate deflection is somewhat (19%) underes-
timated by the numerical model, which can be caused by the overstiffening due to the as-
sumption of a perfect bond.
CONCLUSIONS
This paper presents the comparison between the numerical model and the laboratory exper-
iments on the plain concrete beams, externally reinforced with TRC plates. The main con-
clusions can be summarized:
- Bolts improve the mechanical behaviour of the beam, as they prevent the further de-
velopment of debonding leading to final beam failure
- In the case of the non-bolted specimens, debonding is the major failure mechanism
and the bond properties have to be taken into consideration for a correct numerical
model
- The identified bond parameters can be assumed to be correct for predicting the fail-
ure load and can be used in future work.
486
In the future, the matter of bond has to be further researched in depth. This will allow a
better prediction of the overall stiffness of the composite beams and will contribute to
the design codes for the utilisation of TRC in structural elements.
REFERENCES
[1] Peled A. Confinement of damaged and nondamaged structural concrete with FRP and TRC sleeves. J
Compos Constr 2007:514–23.
[2] D’Ambrisi A, Focacci F. Flexural strengthening of RC beams with cement-based composites. J
Compos Constr 2011.
[3] Ombres L. Flexural analysis of reinforced concrete beams strengthened with a cement based high
strength composite material. Compos Struct 2011;94:143–55.
[4] Pellegrino C, D’Antino T. Failure due to delamination in concrete elements strengthened with
cementitious composites. VIII Int Conf Fract Mech Concr Concr Struct Fram 2013:1–6.
[5] Brückner a., Ortlepp R, Curbach M. Anchoring of shear strengthening for T-beams made of textile
reinforced concrete (TRC). Mater Struct 2007;41:407–18.
[6] Triantafillou T, Papanicolaou C. Shear strengthening of reinforced concrete members with textile
reinforced mortar (TRM) jackets. Mater Struct 2006;39:93–103.
[7] Si Larbi A, Contamine R, Ferrier E, Hamelin P. Shear strengthening of RC beams with textile
reinforced concrete (TRC) plate. Constr Build Mater 2010;24:1928–36.
[8] Contamine R, Si Larbi a., Hamelin P. Identifying the contributing mechanisms of textile reinforced
concrete (TRC) in the case of shear repairing damaged and reinforced concrete beams. Eng Struct
2013;46:447–58.
[9] Papanicolaou CG, Papantoniou IC. Mechanical Behavior of Textile Reinforced Concrete (TRC) /
Concrete Composite Elements. J Adv Concr Technol 2010;8:35–47.
[10] Remy O, Verbruggen S. Cement composite stay-in-place formwork: a concept for future building
systems. Proc 18th … 2011.
[11] Verbruggen S, Remy O, Wastiels J, Tysmans T. Stay-in-Place Formwork of TRC Designed as Shear
Reinforcement for Concrete Beams. Adv Mater Sci Eng 2013;2013:1–9.
[12] D’Ambrisi A, Feo L, Focacci F. Experimental and analytical investigation on bond between Carbon-
FRCM materials and masonry. Compos Part B Eng 2013;46:15–20.
[13] Valluzzi MR, Oliveira D V., Caratelli A, Castori G, Corradi M, Felice G, et al. Round Robin Test for
composite-to-brick shear bond characterization. Mater Struct 2012;45:1761–91.
[14] D’Ambrisi A, Feo L, Focacci F. Experimental analysis on bond between PBO-FRCM strengthening
materials and concrete. Compos Part B Eng 2013;44:524–32.
[15] Carloni C, H. SL, T D. Interfacial Bond Characteristics of Fiber Reinforced Cementitious Matrix for
External Strengthening of Reinforced Concrete Members. In: J.G.M. van M, G. R, C. A, R.C. Y,
X.X. Z, editors. VII Int. Conf. Fract. Mech. Concr. Concr. Struct., Toledo, Spain: International
Center for Numerical Methods in Engineering (CIMNE); 2013, p. 129–37.
[16] Ortlepp R, Hampel U, Curbach M. A new approach for evaluating bond capacity of TRC
strengthening. Cem Concr Compos 2006;28:589–97.
[17] Ombres L. Debonding analysis of reinforced concrete beams strengthened with fibre reinforced
cementitious mortar. Eng Fract Mech 2012;81:94–109.
[18] Pellegrino C, Tinazzi D, Modena C. Experimental study on bond behavior between concrete and FRP
reinforcement. J Compos Constr 2008:180–90.
[19] Lu XZ, Teng JG, Ye LP, Jiang JJ. Bond–slip models for FRP sheets/plates bonded to concrete. Eng
Struct 2005;27:920–37.
[20] D’Ambrisi A, Feo L, Focacci F. Bond-slip relations for PBO-FRCM materials externally bonded to
concrete. Compos Part B Eng 2012;43:2938–49.
[21] Verbruggen S. Reinforcement of concrete beams in bending with externally bonded textile reinforced
cementitious composites. PhD Thesis, Vrije Universiteit Brussel, Brussels, BELGIUM: 2014.
[22] Cuypers H. Analysis and design of sandwich panels with brittle matrix composite faces for building
applications. PhD thesis. Vrije Universiteit Brussel, Brussels, BELGIUM: 2001.
487
[23] Verbruggen S, Wastiels J, Tysmans T, Remy O, Michez S. The influence of externally bonded
longitudinal TRC reinforcement on the crack pattern of a concrete beam. Proc. 3rd Int. Conf. Concr.
Repair, Rehabil. Retrofit., Cape Town: Taylor & Francis Group; 2012, p. 1259–65.
[24] Wozniak M, Tysmans T, Vantomme J. Finite Element Modelling of Glass Fibre Reinforced
Composites with Inorganic Phosphate Cement Matrix : Comparison of Inbuilt Abaqus Concrete
Models. In: Hoa S Van, Hubert P, editors. 19th Int. Conf. Compos. Mater. ICCM19, Montreal,
Canada: BytePress.com; 2013.
[25] Silghem E, Vandenplas P. Design of externally bonded cement composite reinforcement of concrete
beams. Master Thesis, Vrije Universiteit Brussel, Brussels, BELGIUM: 2013.
[26] Abaqus Theory Manual. Abaqus User’s Man. Version 6.8, n.d.
488
SHEAR STRENGTHENING OF RC BEAMS USING TEXTILE

REINFORCEMENT IN CEMENT OR EPOXY BASED MATRICES
Zoi C. Tetta1, Lampros N. Koutas2, Dionysios A. Bournas3, Brwa H. Salihi4 ,1 PhD Candidate,
Dep. of Civil Engineering, Univ. of Nottingham, Nottingham, UK, zoi.tetta@nottingham.ac.uk , 2
Research Associate, Dep. of Civil Engineering, Univ. of Nottingham, Nottingham, UK,
koutasciv@gmail.com, 3 Assistant Professor, Dep. of Civil Engineering, Univ. of Nottingham,
Nottingham, UK, dionysios.bournas@nottingham.ac.uk, 4 Post-graduate student, Dep. of Civil
Engineering, Univ. of Nottingham, Nottingham, UK, brwace@gmail.com
Abstract: This paper presents an experimental study on shear strengthening of rectangular

reinforced concrete (RC) beams with advanced composite materials. Key parameters of
this study include: (a) the strengthening system, namely textile-reinforced mortar (TRM)
jacketing and fiber-reinforced polymer (FRP) jacketing, (b) the strengthening configura-
tion, namely side-bonding, U-wrapping and full-wrapping, and (c) the number of the
strengthening layers. In total, 14 RC beams were constructed and tested under bending
loading. One of the beams did not receive any strengthening and served as control beam,
eight received TRM jacketing, whereas the rest five received FRP jacketing. It is conclud-
ed that the TRM is generally less effective than FRP in increasing the shear capacity of
concrete however the effectiveness depends on both the strengthening configuration and
the number of layers. U-wrapping strengthening configuration is much more effective than
side-bonding in case of TRM jackets and the effectiveness of TRM jackets increases con-
siderably with increasing the number of layers.
INTRODUCTION AND BACKGROUND
Structural retrofitting of existing reinforced concrete (RC) structures is a constantly grow-

ing need due to their deterioration. One of the most common structural deficiencies is the
poor shear capacity of RC beams or bridge girders. The use of fiber reinforced polymers
(FRP) as externally bonded (EB) reinforcement in shear strengthening of RC members has
become very popular over the last two decades. Following the studies of Triantafillou 1998
[1] and Khalifa et al. [2] a big effort was made by researchers worldwide to further investi-
gate this technique [i.e. 3-4]; with all the results showing the high effectiveness of using
EB FRP in shear strengthening of RC beams. However, the FRP strengthening technique
has a few drawbacks mainly associated with the use of epoxy resins (i.e. high cost, poor
performance in high temperatures, inability to apply on wet surfaces). In an attempt to al-
leviate the problems arising from the use of epoxies, researchers have introduced a novel
composite material, namely textile-reinforced mortar (TRM), which combines advanced
fibers in form of textiles (with open-mesh configuration) with inorganic matrices, such as
cement-based mortars. Over the last decade it has been reported in the literature that TRM
is a very promising alternative to the FRP retrofitting solution. TRM has been used for the
strengthening of RC members [i.e. 5-10] and masonry elements [i.e. 11-13], as well as for
the seismic retrofitting of masonry-infilled RC frames [14].
489
Shear strengthening of RC beams with TRM has been investigated by few researchers [15-
19]. In these studies various parameters were investigated including the number of layers
[15, 17, 19], the strengthening configuration [18] and the mechanical anchorage of the
jackets [16, 19]. The effectiveness of TRM versus FRP in shear strengthening of RC
beams, has only been investigated on a limited number of specimens [15, 19]. In particular,
in [15] it was concluded that TRM jackets are 45% less effective than their FRP counter-
parts, based on the results of two specimens retrofitted with closed jackets. Moreover,
Tzoura and Triantafillou [19] on the basis of four specimens retrofitted with U-jackets
concluded that TRM jackets are nearly 50% less effective than their counterparts in case of
non-anchored jackets, whereas in case of mechanically anchored jackets the TRM system
is marginally inferior to the FRP system. This paper presents the first systematic study on
the effectiveness of TRM versus FRP jackets in shear strengthening of RC beams.
EXPERIMENTAL PROGRAMME
Test Specimens and Investigated Parameters

The main objective of this study was to compare the effectiveness between TRM and FRP
jacketing in shear strengthening of RC beams. A total of 14 rectangular medium-scale RC
beams were constructed and tested as simply supported in three-point bending as shown in
Fig. 1a. The total length of the beams was equal to 1677 mm, whereas the effective flexur-
al span was equal to 1077 mm. To emulate old detailing practices, the beams were de-
signed to be deficient in shear in one of the two shear spans. To achieve this, the critical
shorter shear span of 460 mm length did not include any transverse reinforcement, whereas
the larger shear span included 8-mm diameter stirrups at a spacing of 75 mm. Details of
reinforcement is given in Fig. 1b. Strengthening was applied only at the critical shear span.
The key investigated parameters of this study comprise: (a) the strengthening system
(TRM or FRP), (b) the strengthening configuration, and (c) the number of layers. One
beam was tested as-built without receiving strengthening and served as control specimen
(CON). The rest 13 beams were divided in two main groups. The first group comprised 8
beams strengthened with TRM jackets, whereas the second group comprised 5 beams
strengthened with FRP jackets. Three different strengthening configurations were applied
on each group’s specimens, namely Side-Bonded jackets (SB), U-Wrapped jackets (UW)
and Fully-Wrapped jackets (FW). For the SB and the UW configurations the specimens of
the first group received from 1 to 3 TRM layers, whereas the specimens of the second
group received 1 and 2 FRP layers. For the FW configuration the first group specimens
received 1 and 2 TRM layers, while only one specimen of the second group received 1
FRP layer. The notation of specimens is X_YN, where X refers to the strengthening con-
figuration (SB, UW and FW), Y denotes the type of the binding material (M for Mortar or
R for Resin) and N denotes the number of layers (1, 2 or 3).
490
Figure 1: (a) Test set-up; (b) cross section; (c) geometry of the carbon textile (dimensions in mm); (d) im-
pregnation of the fibers with epoxy resin; (e) application of a layer of mortar on the top of the final textile
layer.
Materials, Strengthening Procedure and Experimental Set-up

The average concrete compressive strength of all the specimens was 22.4MPa, whereas the
average tensile strength was 2.64MPa. The steel used as internal reinforcement had a yield
stress of 549 MPa. The same reinforcement was used in both strengthening systems; the
only difference between the two systems was the binding material (epoxy resin in case of
FRP and cementitious mortar in case of TRM). This reinforcement comprised a commer-
cial textile with equal quantity of high-strength carbon fibers in two orthogonal directions
(Fig. 1c). The weight of the textile was 348 g/m2, whereas its nominal thickness was 0.095
mm. The tensile strength and the modulus of elasticity of the carbon fibers were 3800 ΜPa
and 225 GPa, respectively. For the specimens receiving mortar as binding material an in-
organic dry binder was used with water-binder ratio in the mortar equal to 0.23∶1 by
weight. The average flexural strength and compressive strength of mortar on the day of
testing was 9.4MPa, and 28.7MPa, respectively. For the specimens receiving epoxy adhe-
sive as binding material, a commercial adhesive was used with an elastic modulus of 3.8
GPa and a tensile strength of 30 MPa.
Prior to the application of the bonding agent the concrete surface was brushed from the
dust. For FRP-jacketed specimens the first textile layer was applied on the top of the first
resin layer and was then impregnated in-situ with resin using a plastic roll (Fig. 1d). For
additional textile layers the same process was repeated. For TRM-jacketed specimens the
mortar was applied with a smooth metal trowel. After application of the first mortar layer
on the (dampened) concrete surface, the textile was applied and pressed slightly into the
mortar, which protruded through all the perforations between the fiber rovings. The next
mortar layer covered the textile completely, and the operation was repeated until all textile
layers were applied and covered by the mortar (Fig. 1e).
The beams were subjected to monotonic loading at a displacement rate of 0.02 mm/s. The
vertical displacement was measured at the position of load application using an external
LVDT. Moreover, measurements from the potentiometers placed at the critical shear span
in one side of the beam, were utilized to monitor the average shear strain of the span. Addi-
tionally, the Digital Image Correlation (DIC) technique was employed to monitor relative
displacements within the critical shear span.
491
Results
The response of all specimens tested is given in Fig. 2 in the form of load - displacement
curves. Key results are also presented in Table 1. It includes: (1) The peak load. (2) The
failure mode. (3) The shear resistance of the critical shear span, VR. (4) The contribution of
the jacket to the total shear resistance, Vf. (5) The degree of strengthening (VR,str/VR,con). (6)
The average shear strain of the critical shear span at peak load, γPmax. (7) The degree of the
deformation capacity enhancement of the critical span (γPmax,str/γPmax,con). The average shear
strain at the critical shear span, γ, was obtained from readings of the two potentiometers
placed in X configuration. The control beam (CON) failed in shear at an ultimate load of
51.8 kN after the formation of a large shear crack in the critical span as shown in Fig. 3a.
Figure 2: Load versus vertical displacement curves for all tested specimens.
All beams strengthened with SB or UW FRP jackets failed in shear at an ultimate load sub-
stantially higher than that of the control beam. In all these specimens, failure occurred due
to FRP debonding (Fig. 3b, Fig. 3c); the excellent bond between the resin and the concrete
substrate resulted in peeling off of the FRP jackets with part of the concrete. Specimen
FW_R1 reached its ultimate moment capacity and failed in flexure. With only the excep-
tion of specimen FW_M2 which failed in flexure, all the TRM-strengthened specimens
failed in shear and displayed considerably higher shear resistance compared to the control
specimen. The failure of specimens with 1 TRM layer was associated with damage on the
TRM jacket (Fig. 3d). The load-drop in these specimens was attributed to the following
local phenomena: (a) slippage of the vertical fiber rovings through the mortar and (b) par-
tial rupture of the fibers crossing the shear crack. The nature of these local phenomena did
not allow for a brittle failure mode.
492
Table 1: Summary of test results
Specimen (1) Peak (2) Failure (3) VR (4) Vf (5) VR,str/ (6) γPmax (7) γPmax,str/
load (kN) mode (kN) (kN) VR,con (‰) γPmax,con
CON 51.8 shear a 29.7 - - 1.51 -
SB_M1 56.6 shear b 32.4 2.7 1.09 2.00 1.32
UW_M1 78.2 shear b 44.8 15.1 1.51 4.00 2.65
FW_M1 111.2 shear b 63.7 34.0 2.15 3.78 2.50
SB_M2 88.7 shear c 50.8 21.1 1.71 3.63 2.40
UW_M2 120.2 shear c 68.8 39.2 2.32 4.34 2.87
FW_M2 152.8 flexural 87.5 57.8 d 2.95 - -
SB_M3 108.9 shear c 62.4 32.7 2.10 5.10 3.38
UW_M3 131.1 shear c 75.1 45.4 2.53 3.07 2.03
SB_R1 105.0 shear c 60.2 30.5 2.03 4.82 3.19
UW_R1 113.4 shear c 64.9 35.3 2.19 3.33 2.21
FW_R1 150.3 flexural 86.1 56.4 d 2.90 - -
SB_R2 124.5 shear c 71.3 41.6 2.40 3.26 2.16
UW_R2 126.2 shear c 72.3 42.6 2.43 2.54 1.68
a
Tensile diagonal cracking; b slippage of the vertical fiber rovings through the mortar and partial fibers rupture; c
debonding of the jacket; d lower limit of Vf due to the flexural failure.
Figure 3: (a) Shear crack in the control beam; (b), (c) SB_R1 and UW_R1 FRP debonding; (d) local damage
of the jacket in SB_M1; (e) debonding of the jacket in SB_M2; (f) abrupt debonding of jacket in UW_M3;
(g), (h) effect of strengthening configuration and number of layers on the shear capacity enhancement.
Specimens SB_M2 and UW_M2 attained higher load with respect to specimens SB_M1
and UW_M1, while their failure was attributed to debonding of the TRM jacket at a large
part (approximately 2/3) of the shear span which was accompanied by peeling off of the
concrete cover (Fig. 3e). This type of failure, although it was brittle, it was not as explosive
as in the case of FRP-strengthened beams. Specimen FW_M2, after the formation of a
shear crack reached its ultimate moment capacity and (identically to FW_R1) failed in
flexure. Specimens SB_M3 and UW_M3 failed in shear at even higher loads when com-
pared to specimens SB_M2 and UW_M2. Specimen SB_M3 failed in a similar way with
specimen SB_M2 (Fig. 3e), whereas the failure mode of specimen UW_M3 was unique
among all TRM-strengthened specimens, as debonding of the U-jacket occurred at the full-
length of the shear span (Fig. 3f) and was explosive as in case of all FRP-strengthened
specimens.
493
Strengthening the beams with TRM or FRP jackets resulted in an increase in the average
shear strain over the critical shear span, which at peak load varied from 1.32 to 3.38 times,
compared to the control specimen. This increase is mainly attributed to redistribution of
shear stresses in the shear span, which ultimately led to a more dense crack pattern and
hence to an increased shear deformation capacity.
Strengthening Configuration and Number of Layers

As illustrated in Fig. 3g the benefit of applying UW instead of SB jackets is more pro-
nounced in TRM than FRP system, especially as the number of layers increase, while FW
jacketing was the most effective configuration for both strengthening systems. The effect
of the number of layers on the shear capacity enhancement for SB and UW strengthening
configurations is illustrated in Fig. 3h. Doubling the amount of reinforcement (two instead
of one layers) resulted in dramatic increase of the TRM jackets effectiveness. In particular,
this increase was equal to 7.8 and 2.6 times for the SB and UW jackets, respectively. The
corresponding increase when resin was used as binder was 1.35 and 1.2 times. Applying a
third TRM layer resulted in strength increase of 1.55 and 1.15 times for SB and UW jack-
ets, respectively.
A possible explanation for the difference between the two strengthening systems, regarding
the effect of increasing the number of layers from one to two, could be found in the ob-
served failure modes. All specimens retrofitted with FRP jackets exhibited the same fail-
ure; failure of the concrete substrate with no damage in the composite jackets. However, in
the case of TRM specimens a change in the failure mode was witnessed when the number
of layers was increased from one to two or three. In specific, when one layer was applied
the failure was attributed to local damage of the TRM jacket (partial rupture of vertical
fiber rovings and slippage of them through the mortar). The increase in the number of lay-
ers in that case prevented these local phenomena and the damage was shifted to the con-
crete substrate. The authors believe that the significantly improvement of TRM jacket be-
haviour when a second layer of textile was provided is linked to the mechanism of transfer-
ring forces from the textile reinforcement to the matrix. It seems that by providing just a
second layer of textile, the mechanical interlock, which is the main mechanism of transfer-
ring forces from the reinforcement to the matrix in TRM systems, is drastically improved.
This improvement might be attributed to the fact that two (at least) overlapping textile lay-
ers create a denser mesh-pattern than just one due to the possible offset between the two
layers. Provided that mortar will not fail in shear; the denser mesh-pattern in turn creates
conditions for improved mechanical interlocking characteristics, which ultimately results
in altering the failure mode.
Deformation Aspects of Jackets Based on DIC

Each curve of Fig.4 illustrates the relative displacement of each point along the beam
height with respect to the bottom of the beam (at the middle cross-section of the critical
shear span) at the instant of peak load.
494
Figure 4: Relative vertical displacements along the beam height.
The control beam exhibited concentration of the vertical deformations at a specific level,
which is related to the development of a single shear crack at that level. The specimens
with one TRM layer exhibited identical behaviour, with concentration of the deformations
at a single level. Better distribution of deformations was observed in the rest specimens,
indicating that the jackets were activated over a broader area due to better redistribution of
stresses. It is evident that in TRM jackets the force transferring mechanism is being modi-
fied with additional layers, thus resulting in a performance of the TRM jackets similar to
the performance of the FRP jackets.
TRM vs FRP Effectiveness Factor
For calculating the FRP contribution to the shear capacity of RC members most of the de-
sign models use the effective stress of the FRP (σeff). The shear force carried by the FRP,
Vf, can be calculated using the Eq. (1) [1].
V f = ρ f σ eff bw 0.9d (1)
Where ρf is the geometrical reinforcement ratio of the composite material, bw is the width
of the beam, d is the effective depth of the section, and tf is the total thickness of the com-
posite material.
Table 2: Effective strains and effectiveness factor, k for SB and UW jackets.
σeff (MPa) εeff (‰) k (σeff,TRM/σeff,FRP)

TRM (SB_M1) 90 0.4
1 layer 0.09
FRP (SB_R1) 990 4.40
Side–Bonded jackets
TRM (SB_M2) 344 1.53
2 layers 0.51
FRP (SB_R2) 675 3.00
TRM (UW_M1) 491 2.18
1 layer 0.42
FRP (UW_R1) 1168 5.19
U–Wrapped jackets
TRM (UW_M2) 637 2.83
2 layers 0.92
FRP (UW_R2) 693 3.08
The format of Eq. (1) can also be used for the calculation of the shear force carried by
TRM jackets [15]. Table 2 includes the values of σeff, εeff (εeff is the effective strain, and is
495
calculated by dividing σeff by the modulus of elasticity of the fibers, Ef) and the effective-
ness factor k, which is defined as the ratio of the TRM to FRP jackets effective stress. It
should be noted that a value of k = 0.55 obtained in [15] from tests on two rectangular RC
beams retrofitted with TRM and FRP closed jackets. In the present study k varies from
0.09, which corresponds to one layer of SB jacket, to 0.92 which corresponds to two layers
of UW jacket. By increasing the number of layers from one to two the effectiveness factor
increases substantially, whereas the same happens when UW jackets are applied instead of
SB jackets.
CONCLUSIONS
This paper presents an experimental investigation on the effectiveness of TRM and FRP
jackets in shear strengthening of rectangular RC beams. The main conclusions drawn from
this study are summarized as follows:
• The effectiveness of TRM jackets depends on both the strengthening configuration and
the number of layers. The TRM versus FRP effectiveness factor varies from 0.09, which
correspond to one layer of side-bonded jacket, to 0.92, which correspond to two layers
of U-wrapped jacket. TRM jackets are more effective in increasing the beams defor-
mation than FRP jackets.
• U-wrapping (UW) strengthening configuration is much more effective than side-
bonding (SB) in case of TRM jackets. On the contrary, in case of FRP jackets the UW
configuration was found only slightly more effective than the SB configuration. Full-
wrapping (FW) is the most effective strengthening configuration for both strengthening
systems.
• A major difference between TRM and FRP strengthening systems is observed by in-
creasing the number of layers from 1 to 2. The considerably higher effectiveness of
TRM jackets when two instead of one textile layers are applied is linked to the better
mechanical interlock conditions created by the overlapping of at least two textile layers.
REFERENCES
[1] Triantafillou, T.C., ‘Shear strengthening of reinforced concrete beams using epoxy-bonded FRP com-
posites’, ACI Struct. J. 95(2) (1998) 107-115.
[2] Khalifa, A., Gold, W.J., Nanni, A., MI, A.A., ‘Contribution of externally bonded FRP to shear capaci-
ty of RC flexural members’, J. Comp. Constr. 2(4) (1998) 195-202.
[3] Galal, K. and Mofidi, A., ‘Shear strengthening of RC T-beams using mechanically anchored unbond-
ed dry carbon fiber sheets’, J. Perform. Constr. Facil. 24(1) (2010) 31-39.
[4] Koutas, L. and Triantafillou, T.C., ‘Use of anchors in shear strengthening of reinforced concrete T-
beams with FRP’, J. Comp. Constr. 17(1) (2013) 101-107.
[5] Bournas, D.A., Lontou, P.V., Papanicolaou, C.G., Triantafillou, T.C., ‘Textile-reinforced mortar ver-
sus fiber-reinforced polymer confinement in reinforced concrete columns’, ACI Struct. J. 104(6)
(2007) 740-748.
[6] Bournas, D.A., Triantafillou, T.C., Zygouris, K., Stavropoulos, F., ‘Textile-reinforced mortar versus
FRP Jacketing in seismic retrofitting of RC columns with continuous or Lap-spliced deformed bars’,
J. Comp. Constr. 13(5) (2009) 360-371.
[7] Bournas, D.A., Triantafillou, T.C., ‘Bond strength of lap-spliced bars in concrete confined with com-
posite jackets’, J. Comp. Constr. 15(2) (2011) 156-167.
[8] Bournas, D.A., Triantafillou, T.C., ‘Bar buckling in RC columns confined with composite materials’,
J. Comp. Constr. 15(3) (2011) 393-403.
496
[9] Bournas, D.A., Pavese, A., Tizani, W., ‘Tensile capacity of FRP anchors in connecting FRP and TRM
sheets to concrete’, Engin. Struct. 82(1) (2015) 72-81.
[10] Elsanadedy, H.M., Almusallam, T.H., Alsayed, S.H., Al-Salloum, Y.A., ‘Flexural strengthening of
RC beams using textile reinforced mortar–Experimental and numerical study’, J. Comp. Struct. 97
(2013) 40-45.
[11] Papanicolaou, C.G., Triantafillou, T.C., Karlos, K., Papathanasiou, M., ‘Textile-reinforced mortar
(TRM) versus FRP as strengthening material of URM walls: in-plane cyclic loading’, Mater. Struc.t
40(10) (2007) 1081-1097.
[12] Harajli, M., ElKhatib, H., San-Jose, J.T., ‘Static and cyclic out-of-plane response of masonry walls
strengthened using textile-mortar system’, J. Mat. Civ. Eng. 22(11) (2010) 1171-1180.
[13] Koutas, L., Pitytzogia, A., Triantafillou, T.C., Bousias, S.N., ‘Strengthening of infilled reinforced
concrete frames with TRM: Study on the development and testing of textile-based anchors’, J. Comp.
Constr. 18(3) (2014), A4013015.doi: 10.1061/(ASCE)CC.1943-5614.0000390.
[14] Koutas, L., Bousias, S.N., Triantafillou, T.C., ‘Seismic strengthening of masonry-infilled RC frames
with TRM: Experimental study’, J. Comp. Constr. (2014), 04014048. doi: 10.1061/(ASCE)CC.1943-
5614.0000507.
[15] Triantafillou, T.C., Papanicolaou, C.G., ‘Shear strengthening of reinforced concrete members with
textile reinforced mortar (TRM) jackets’, Mater. Struct. 39(1) (2006) 93-103.
[16] Brückner, A., Ortlepp, R., Curbach, M., ‘Anchoring of shear strengthening for T-beams made of tex-
tile reinforced concrete (TRC)’, Mater. Struct. 41(2) (2008) 407-418.
[17] Al-Salloum, Y.A., Elsanadedy, H.M., Alsayed, S.H., Iqbal, R.A., ‘Experimental and numerical study
for the shear strengthening of reinforced concrete beams using textile-reinforced mortar’, J. Comp.
Constr. 16(1) (2012) 74-90.
[18] Azam, R., Soudki, K., ‘FRCM strengthening of shear-critical RC beams’, J. Comp. Constr. 18(5)
(2014), 04014012. doi: 10.1061/(ASCE)CC.1943-5614.0000464.
[19] Tzoura, E., Triantafillou, T.C., ‘Shear strengthening of reinforced concrete T-beams under cyclic
loading with TRM or FRP jackets’, Mater. Struct. (2014), doi: 10.1617/s11527-014-0470-9.
497
498
BEHAVIOUR OF HYBRID COMPOSITE-CONCRETE BEAMS

UNDER STATIC FLEXURAL LOADING: A COMPARATIVE
EXPERIMENTAL ANALYSIS
Sven De Sutter (1), Tine Tysmans (1), Svetlana Verbruggen (1), Maciej Wozniak (1) and Matthias
De Munck (1), (1) Department of Mechanics of Materials and Constructions (MeMC), Faculty of
Engineering Sciences, Vrije Universiteit Brussel, Brussels, Belgium
Abstract: The use of composites such as TRCs (Textile Reinforced Cements) and FRPs
(Fibre Reinforced Polymers) in civil structures is nowadays mostly limited to strengthen-
ing and repair applications and stay-in-place formworks. The authors propose to integrate
these materials into a new floor system - suited for renovation purposes - in order to facili-
tate the manual installation by lightweight elements. Lightweight hybrid beams and inter-
mediate placed sandwich panels constitute the formwork for the finishing concrete layer.
The hybrid beams themselves are composed of a TRC box with FRP reinforcement inside
and concrete on top. This paper focuses on the behaviour of these hybrid beams under stat-
ic flexural loading. Therefore, four hybrid beams with a different amount of Carbon Fibre
Reinforced Polymers (CFRPs) and concrete are tested until failure under four point load-
ing. 3-m-span hybrid beams with a different amount of reinforcement are compared to the
Eurocode requirements in SLS (Serviceability Limit State) and ULS (Ultimate Limit
State). Further on, this paper proves the ability to simulate the experimental behaviour by
an analytical model and elaborates on the composite action of the hybrid beams regarding
strains, neutral axis and bending stiffness. This paper proves and quantifies the weight- and
structural advantages of hybrid composite-concrete beams and provides a first insight in
the possible future application of composites for the design of a lightweight floor concept.
INTRODUCTION
Textile Reinforced Cements (TRC) are investigated in various application domains, where
strengthening of concrete structural elements such as columns [1], beams [2] or slabs [3] is
one of the most elaborated topics. Reinforced concrete (RC) structures with TRC are inves-
tigated as a possible substitute for the today’s used Carbon Fibre Reinforced Polymers
(CFRP) which lack the necessary fire resistance. TRC is also used for stay-in-place form-
work elements for floors and beams [4, 5, 6]. The application of TRCs in structural bearing
elements requires high performant materials. Therefore, researchers at the Vrije Universi-
teit Brussel (VUB) developed Inorganic Phosphate Cement (IPC) [7], a fine grained ce-
mentitious matrix, which is pH neutral after hardening. Hereby IPC can be combined with
low-cost E-glass fibres and dense fibre textiles, resulting in a high performant TRC
(GFR.IPC) with fibre volume fractions of up to 25%. The stress-strain relation is obtained
by performing 10 tensile tests with rectangular specimens (Figure 1, [8]). Because of the
matrix cracking (Point A, Figure 1), GFR.IPC exhibits a non-linear stress-strain relation in
tension. The research related to this GFR.IPC currently covers several domains such as
499
thin shells [9, 10], external reinforcement [2], stay-in-place formworks [8, 6] and floor
systems [11].
Its high mechanical properties allow the use of

this material in applications where both struc-
tural performance and fire safety are required,
e.g. a floor system within residential buildings.
In response to the demands of the housing ren-
ovation sector for lightweight and rapidly in-
stallable structural solutions, the authors pro-
pose a new hybrid floor system that is com-
posed of TRC and FRP composites to reduce
the installation weight without losing structural
capacity.
Figure 1: the non-linear stress-strain diagram of

GFR.IPC [8]
Different authors [12 - 16] have studied the structural behaviour of hybrid FRP-concrete
beams. Structural hybrid systems enable new opportunities in the building industry by
combining the good properties of different dissimilar materials such as the tailorable char-
acteristics of FRP and the compressive behaviour of concrete.
The excellent mechanical properties of FRPs are complementary with the improved struc-
tural properties of fire safe TRCs such as GFR.IPC (highest European class A1). There-
fore, the authors combine TRC, FRP and concrete in a hybrid floor system - composed of
beams, sandwich panels and a compression layer - that optimally benefits from all materi-
als (Figure 2). The hybrid beams are composed of a hollow GFR.IPC box at the bottom
with a CFRP strip inside. Concrete is cast on top of this box to provide the necessary stiff-
ness. Sandwich panels are placed in between and form together with the beams the Stay-in-
Place (SiP) formwork for the final concrete layer (casting stage). After hardening of the
concrete compression layer, all components act as one monolithical floor (hardened stage).
This paper focusses on the structural behaviour of the hybrid beams, related to their limits
in casting stage.
Figure 2: The composition of the hybrid floor (left); Hybrid beams are composed of a GFR.IPC box, with a
CFRP strip inside and concrete on top (right). All beam dimensions can be found in Figure 4.
500
MATERIALS AND METHODS
Four hybrid beams with a span of 3000 mm have been tested. A displacement controlled
(1 mm/min) hydraulic actuator of 500 kN is used to load the hybrid beams by four point
bending (500 mm between the point loads). Four strain gauges (concrete: PL-120-11;
GFR.IPC: FLA-30-11) are attached to the middle section (top, middle concrete, middle
GFR.IPC box, bottom).
Figure 3: The experimental set-up
Two different types of hybrid beams (Figure 4 &Table 1) can

be distinguished by varying two dimensional parameters: the
thickness of CFRP (hc: 0 and 3 mm; [17]) and the height of
concrete (hcon: 25 and 50 mm). Two identical beams of every
beam type have been tested, resulting in a total of four beams.
The width of the cross-section is 100 mm, the height and
thickness of the GFR.IPC box is 75 mm and 5 mm respective-
ly. As these first tests focus on the structural behaviour of the
hybrid beams, the concrete layer is cast over the whole width
Figure 4 : Dimensions of the (Figure 4) of the box, in contrast to the future application.
hybrid beams.
The top box surface is roughened with small gravel (4-7 mm) to improve the bond at the
interface between GFR.IPC and concrete. Detailed material properties are given in Table 2.
The concrete compressive strength is the experimental mean of 6 specimens, tested after 28
days. All compressive tests are performed on cylindrical specimens according to [18]. The
young modulus is also calculated from this compression strength [18]. For GFR.IPC and
CFRP, all properties (except the compressive strength) are the experimental mean values
of 3 specimens, determined by tensile tests. GFR.IPC has an initial stiffness of 10.1 GPa
(uncracked matrix) and a final stiffness of 3.8 GPa (cracked matrix).
501
Table 1: Dimensions of the four tested hybrid beams
Hybrid beam nr. 1&2 3&4
Height carbon (hc) [mm] 0 3.0
Height concrete (hcon) [mm] 25 50
Table 2: Characteristic properties of the considered materials
Material Tensile Compressive Young Specific Den-

strength strength Modulus sity
[MPa] [MPa] [GPa] [g/cm³]
Concrete n/a 53.2 37.8 2.5
GFR.IPC 50.5 80 [19] 10.1 – 3.8 2.05
CFRP 2375 n/a 137 1.6
During casting stage, the installation elements (sandwich panels + beams) are put in place
and the top concrete compression layer is poured on top of this formwork. Consequently,
the hybrid beams have to bear their own self-weight (15 kg/m), the weight of the interme-
diate sandwich panels (12 kg/m²) and the liquid concrete (2500 kg/m³). For a span of 3 m,
this results in a design moment of minimum 1.5 kNm (Ultimate Limit State, ULS) and a
deflection limit of L/500 (= 6 mm; most severe restriction) at a Serviceability Limit State
(SLS) moment of 1.1 kNm. Load safety factors are applied according to [20]. Plotting the-
se limit states against the experimental moment-deflection behaviour (Figure 5) gives a
first idea about the structural feasibility of these concepts regarding the Eurocode require-
ments. In design phase however, material safety factors must be applied. Figure 5 clearly
shows the spare capacity of these beams (except beams 1 & 2 for SLS). Beams 3 and 4
largely exceed these limits, revealing the fact that these beams are rather dimensioned for
the hardened stage and thus over-dimensioned for casting stage. The self-weight of the
beams is not taken into account.
The absence of carbon reinforcement in beams 1 and 2 has a clear influence on the stiff-
ness in comparison to the carbon strip reinforced beams. Beams 1 and 2 failed due to ten-
sile failure at the bottom of the GFR.IPC box. These beams ultimately exceed the ULS
load limit (at a deflection above 40 mm), but fail to limit their deflection within the SLS
area. A minimum of carbon reinforcement is thus also needed in casting stage. A further
increase of the amount of CFRP and concrete corresponds to a further increase in stiffness
(beams 3 & 4). Both hybrid beams fulfil both limit state requirements, but failed due to
unexpected interlaminar shear failure between the layers of the GFR.IPC box (more elabo-
502
rated in par. 3.2). The spare stiffness of beams 3 and 4 is mainly the result of the used
amount of carbon and concrete. The amount of CFRP strips cannot be reduced as they are
embedded within the TRC box and necessary in the hardened stage, but the amount of con-
crete can be reduced to a minimum. This reduction of concrete has a significant positive
effect on the installation weight. The additional amount of concrete necessary for the hard-
ened stage can be added during casting of the compression layer. The hybrid beams, as
tested in this paper, weigh on average 10% less than prestressed beams (>16 kg/m) that are
used nowadays to construct beam-and-block floors. However, further reduction of concrete
can increase this weight reduction up to 50% [21].
Figure 5 : All CFRP reinforced hybrid beams experimentally fulfil both limit states.
Analytical simulation
The main goal is to validate the simulation of the structural behaviour of the beams and
fully understand the material behaviour and its influence on the experimental results.
Therefore, no safety factors are implemented in the analytical model. In a first step (3.2.1),
beams 1 & 2 are discussed because they only consist of concrete and GFR.IPC. Secondly,
beams 3 & 4 (3.2.2) are considered to validate the influence of a significant amount of
CFRP on the simulation. In both paragraphs, special attention is paid to the interaction be-
tween the material properties and the reflection of this within the analytical simulation. A
theoretical prediction of the moment-deflection behaviour of the hybrid beams is made by
an implementation of the analytical design formulas based on Eurocode 2 [18] in the math-
ematical software environment Matlab. This model discretizes the hybrid cross-section and
applies different strain distributions to find an equilibrium between all discretized forces.
For every balanced strain distribution, the corresponding loadbearing capacity (Eq. 1) and
deflection (Eq. 2) are calculated until the point of failure (Ultimate Limit State, ULS) is
reached. A 95 % confidence interval - until the failure of one tested beam - indicates the
probability that the true value lies within this range [22].
(1)
503
(2)
With: d: the lever arm between the compression and tension resultant force; Fcomp: com-
pression resultant force; Fapp: the applied experimental force; a: distance between point
load and support (1250 mm); EI: cross-sectional bending stiffness of the considered beam;
L: length of the beam (3000 mm); q: self-weight of the beam, expressed as line load.
Beams 1 & 2 – without carbon strip reinforcement
Beams without carbon strip reinforcement clearly lack the necessary stiffness to be used
within the final floor system, but they allow the validation of the analytical model regard-
ing the concrete and GFR.IPC, without the additional influence of CFRP. A great similari-
ty is obtained between the analytical and experimental stiffness (Figure 6).
Figure 6 : A good agreement is achieved between the analytical simulation and the experimental data for
beams 1 and 2 (without carbon reinforcement).
The only significant deviation between the analytical simulation and the experimental data
is visible near the experimental point of failure. This divergence is due to the gradual fail-
ing of the materials toward the end of the experiment. By the absence of carbon strip rein-
forcement, the GFR.IPC at the bottom of the box failed in tension. Besides a good predic-
tion of the stiffness evolution, an accurate prognosis of the failure load is obtained. The
analytical model simulates bending failure at 3.19 kNm, which is comparable to the aver-
age experimental failure moment of 3.17 kNm.
The strain evolution is measured during testing at four different locations along the height
of the beam in the middle of the span (Figure 7). The strain at the bottom of the beam is
based on the experimental measurements of the other three strain gauges, as due to a tech-
nical error, the strain at this location is not recorded correctly. A linear relation between the
504
data is noticed from the beginning to the end of the test, so we can state that plane sections
remain plane and thus that the assumption of Bernoulli is valid. If we compare the analyti-
cal prediction of the strain at the bottom of the beam (1.1 %, stated as the maximum allow-
able strain) with the measured values (0.8 %) at the point of failure, we see a small differ-
ence, probably due to small deviations in the positioning of the strain gauges. The neutral
axis slightly shifts between the original uncracked section and the final cracked section. A
small shift of 2 mm (from 19 mm to 17 mm) is necessary to restore the force equilibrium
during loading. However, this shift is not fully due to the cracked – and thus not contrib-
uting – concrete (as is the case for a standard RC beam), but also partially to the decrease
in stiffness of the GFR.IPC at higher loads, as was already indicated in Figure 1. The ab-
sence of much concrete below the neutral axis, a deliberate choice within the geometry,
reduces the influence of the cracked concrete. The bigger influence of the decreasing stiff-
ness of GFR.IPC on the other hand can be proven by the analytical calculation of the bend-
ing stiffness EI (Figure 7 right).
Figure 7: The maximum strain distribution of beam 1 (left) varies from 8 ‰ in tension to 2.5 ‰. The bending
stiffness decreases (right) due to the stiffness variation of the GFR.IPC. A similar behaviour is noticed for
beam 2.
In traditional RC deflection calculations, two bending stiffnesses are calculated: one with
initial material properties and one when the tensile concrete is cracked. Consequently, RC
deflections are determined by a weighted average of both deflections in cracked and
uncracked state. The difference between the cracked and uncracked EI for beam 1 rises
from 1.2 % in the beginning to 5.0 % near the end of the experiment. In a standard RC
beam however, this difference can easily go up to 30 % due to the amount of concrete be-
low the neutral axis. On the other hand, this difference remains constant for a RC beam
because of the constant material stiffness assumptions. This hybrid beam 1 is partly made
of GFR.IPC, which loses its stiffness with an increasing load (Figure 1). This mechanism
seems to have a major influence on the deflection calculation as EI drops by 38 % between
the beginning and the end of the experiment. The drop of both EI’s starts at an applied
moment of 0.5 kNm, exactly the moment where the experimental deflection curve (Figure
6) start to deviate from the first linear part.
505
Beams 3 & 4 – with 3 mm carbon strip reinforcement

Beams 3 & 4 differ from beams 1 & 2 in: (i) carbon strips – with a total thickness of 3 mm
– are added at the inside bottom of the GFR.IPC box; (ii) the concrete height is increased
to 50 mm in order to keep the neutral axis around the separation ‘box-concrete’.
Figure 8 : A good agreement is achieved between the analytical simulation and the experimental data for
beams 3 and 4 (with 3 mm carbon strip reinforcement) until the experimental failure due to delamination
within the GFR.IPC.
Similar to the not-carbon strip reinforced beams (1 & 2), a good agreement between the
analytical simulation and the experimental data is achieved for beams 3 and 4, especially
within the SLS zone (Figure 8). The major difference is situated at the failure state: the
analytical failure moment of 17.3 kNm – due to concrete crushing – is not reached as the
beams experimentally failed at an average moment of 10.1 kNm. This difference is a con-
sequence of the experimental failure type: while the analytical model focusses on failure in
bending, a delamination between the layers of the GFR.IPC box has led to the premature
failing of beams 3 & 4 (Figure 9, left). This delamination has gradually been built up, re-
sulting in a slowly decreasing stiffness near the end of the experiment. Compared with the
SLS and ULS limits however, these beams have a sufficient stiffness and loadbearing ca-
pacity in casting stage. In hardened stage, a higher loadbearing capacity will be reached by
the addition of a concrete compression layer. At failure, the maximum compressive (1.3
‰) and tensile (3 ‰) strains are situated far below their limits (3.5 ‰ and 11 ‰ respec-
tively). This small concrete strain reaffirms the possibility to lower the amount of concrete
(Par. 3.1) for the hybrid beams in casting stage. A maximal exploitation of the material
characteristics will further reduce the installation weight (a reduction of 50% in compari-
son with prestressed concrete beams can be achieved) without affecting the structural ca-
pacities. Beams 3 & 4 have a similar upwards shift of the neutral axis, varying from 46 to
40 mm. However, the percentage of the gradually decreasing stiffness of GFR.IPC is sig-
nificantly smaller in comparison to beams 1 & 2. Carbon strips significantly increased the
stiffness of the cross-section, reducing the share of GFR.IPC to the overall stiffness: the
bending stiffness drops by only 8% (Figure 9, right). This small difference has a positive
influence for deflections at loads beyond the cracking moment.
506
Figure 9 : A cross-section of beam 3 (left); The drop in bending stiffness decreases (right) due to the in-
creased influence of the carbon strips. A similar trend and failure mechanism is noticed for beam 4.
CONCLUSIONS
As a first step towards the build-up of a hybrid floor, 2 types of hybrid composite-concrete
beams are tested under static flexural loading. Only the CFRP reinforced hybrid beams
showed a sufficient stiffness and loadbearing capacity to fulfil all SLS and ULS require-
ments. The analytical model is capable of precisely simulating the moment-deflection be-
haviour. Changes in stiffness or experimental failure of the beams could be explained by
the material behaviour and its influence on the bending stiffness. The fulfilment of the lim-
it states, the accurate simulation and the in-depth comprehension of the material behaviour
of these hybrid beams allows us to extend the research to the level of the entire hybrid
floor.
ACKNOWLEDGMENTS
The authors gratefully acknowledge the funding of the Brussels Capital Region through the
Innoviris Strategic Platform Brussels Retrofit XL for funding the research of the first and
the third author and Fonds Wetenschappelijk Onderzoek-Vlaanderen (FWO) for funding
the research of the fourth author through a PhD scholarship, as well as the co-operation
with TRADECC and Rector.
REFERENCES
[1] Triantafillou T.C. and Papanicolaou C.G.; “Textile Reinforced Mortars (TRM) versus Fibre Rein-
forced Polymers (FRP) as strengthening materials of concrete structures”; ACI-SP-230-6; 99-118,
(2005).
[2] Verbruggen S.; “Reinforcement of concrete beams in bending with externally bonded textile rein-
forced cementitious composites”; Doctoral thesis; Vrije Universiteit Brussel, Faculty of Engineer-
ing, Brussels, Belgium; (2014).
[3] Brückner A., Ortlepp R. and Curbach M.; “Textile reinforced concrete for strengthening in bending
and shear”; Materials and structures; Vol. 39; 741-748; (2006).
[4] Papantoniou I.C. and Papanicolaou C.G.; “Textile Reinforced Concrete (TRC) for precast Stay-in-
Place formwork elements”; Tailor made concrete structures; 475-481; (2008); ISBN 978-0-415-
47535-8.
507
[5] Papanicolaou C.G. and Papantoniou I.C.; “Mechanical Behavior of Textile Reinforced Concrete
(TRC)/Concrete Composite Elements”; Journal of Advanced Concrete Technology; Vol. 8 (1); 35-
47; (2010).
[6] Verbruggen S., Remy O., Wastiels J. and Tysmans T.; “Stay-in-place formwork of TRC designed as
shear reinforcement for concrete beams”; Advances in Materials Science and Engineering, Model-
ing, Characterization, and Processing of Advanced Composites; Vol. 2013; (2013).
[7] Wu X. and Gu J.; inventors; “Inorganic Resin Compositions, Their Preparation And Use Thereof”;
European patent EP 0 861 216 B1; (1997).
[8] Remy O. “Lightweight Stay-in-Place formwork:“a concept for future building applications””. Doc-
toral thesis; Vrije Universiteit Brussel, Faculty of Engineering; Brussels, Belgium; (2012).
[9] Tysmans T., Adriaenssens S. and Wastiels J.; “Form finding methodology for force-modelled anti-
clastic shells in glass fibre textile reinforced cement composites”; Engineering structures; Vol. 33;
2603–2611; (2011).
[10] Cauberg N., Tysmans T., Adriaenssens S., Wastiels J., Mollaert M. and Belkassem B.; “Shell Ele-
ments of Textile Reinforced Concrete using Fabric Formwork: A Case Study”; Advances in Struc-
tural Engineering; Vol. 15 (4); (2012).
[11] De Sutter S., Tysmans T. and Wozniak M; “Analytical modelling and experimental testing of hybrid
composite-concrete beams in a lightweight floor system” Proceeding of the 7th International Con-
ference on FRP Composites in Civil Engineering (CICE2014); Vancouver, Canada; (2014); R. El-
Hatcha, editor; ISBN 978-1-77136-308-2.
[12] Correia J.R., Branco F.A., Ferreira J.G.; “GFRP-concrete hybrid cross-sections for floors of build-
ings”; Engineering Structures; Vol. 31; 1331-1343; (2009).
[13] Gai X., Darby A., Ibell T., Evernden M.; “Experimental investigation into a ductile FRP stay-in-
place formwork system for concrete slabs”; Construction and Building Materials; (2012);
http://dx.doi.org/10.1016/j.conbuildmat.2012.08.050.
[14] Hai N.D., Mutsuyoshi H., Asamoto S. and Matsui T.; “Structural behavior of hybrid FRP composite
I-beam”; Construction and building materials; Vol. 24; 956-969; (2010).
[15] El-Hacha R. and Chen D.; “Behavior of hybrid FRP-UHPC beams subjected to static flexural load-
ing”; Composites: Part B; Vol. 43; 582-593; (2012).
[16] Aref A.J., Kitane Y. and Lee G.C. “Analysis of hybrid FRP-concrete multi-cell bridge superstruc-
ture”; Composite structures; Vol. 69; 346-359; (2005).
[17] TRADECC “PC® CARBOCOMP”, http://www.frp.co.il/uploadimages/12.pdf, July 2007.
[18] CEN (Comité Européen de Normalisation); “Eurocode 2: Design of concrete structures – Part 1-1:
General rules and rules for buildings”; EN 1992-1-1; (2004).
[19] Cuypers H.; “Analysis and Design of Sandwich Panels with Brittle Matrix Composite Faces for
Building Applications”; Doctoral thesis; Vrije Universiteit Brussel, Faculty of Engineering, Brus-
sels, Belgium; (2002).
[20] CEN (Comité Européen de Normalisation); “Eurocode 1: Basis of design and actions on structures –
Part 1: Basis of design”; EN 1991-1; 1994.
[21] De Sutter S., Tysmans T., Verbruggen S. and Wozniak M.; “Shape and size optimization of hybrid
concrete –composite beams in a lightweight floor system”; Proceedings of the 3rd International
RILEM Conference on Strain Hardening Cementitious Composites (SHCC3); Dordrecht, Nether-
lands; (2014); E. Schlangen, M.G. Sierra Beltran, M. Lukovic, G. Ye, editors; ISBN 978-2-35158-
150-6; 349-356.
[22] Brown M.A.; “A step-by-step guide to non-linear regression analysis of experimental data using a
Microsoft Excel spreadsheet”; Computer Methods and Programs in Biomedicine; 191-200
508
REINFORCED MASONRY USING TEXTILES
Dorothea Saenger; Wolfgang Brameshuber, Institute of Building Materials Research and Chair of
Building Materials, RWTH Aachen University
Abstract: The idea of strengthening masonry constructions with textiles emanates from the
same principle as textile reinforced concrete. The low tensile strength of masonry is to be
increased by the use of textiles taking advantage of the benefits of this relatively new rein-
forcement material compared to steel. Textile reinforcement can be bonded externally in
mortar-based systems of new or existing masonry building components. Particularly for
strengthening new masonry constructions textiles can be also embedded even in thin-layer
mortar bed joints. Masonry strengthening using textiles can only be ensured if the bond
performance between textile and mortar as well as between mortar and masonry unit is
high enough to transmit the tension forces acting on masonry. This can be achieved by
selecting compatible textiles to mortar and masonry unit, which – when applied in/on ma-
sonry –lead to an improvement of their mechanical properties. Therefore, the first mile-
stone of a recent research project is to find suitable textiles for use in masonry for both
application methods. With the help of an adapted uniaxial tensile test method originally
developed for textile reinforced concrete, the increase of the tensile strength of textile rein-
forced mortar specimens can be examined. This paper presents the test method and first
results of ongoing tests.
INTRODUCTION
Masonry is a heterogeneous building material consisting of masonry units bonded with

mortar in the horizontal joints and occasionally also in the vertical joints. The unit material
can be homogeneous or heterogeneous depending on the type of masonry unit used. Be-
sides the compressive and shear strength, one of the essential mechanical properties for the
design of masonry building components particularly subjected to flexural loads is the flex-
ural tensile strength. Examples of building components subjected to flexural loads are
basement walls under earth pressure or outer leaf of cavity walls under wind load. Depend-
ing on the leading flexural load case – parallel or perpendicular to the bed joint – as well as
the geometry and boundary conditions of the considered masonry building component a
differentiation of the flexural tensile strength is made. As concrete, the tensile strength and
thereby the flexural tensile strength of masonry is quite low compared to the resistance to
pressure. Surpassing the masonry flexural tensile strength a unit and/or a joint failure can
occur in both load cases namely parallel or perpendicular to the bed joint [1].
In seismic vulnerable zones, earthquake-induced horizontal accelerations generate high

horizontal stresses on masonry constructions. The resistance to the in-plane shear stresses
is however low, so that the fail-free recordable horizontal load is limited. To prevent fail-
ure and increase the load-bearing capacity of masonry building components subjected to
509
out-of-plane or in-plane loads, reinforcement can be added to masonry. One of the tradi-
tional methods is to embed reinforcing bars horizontally in the bed joint or collocate rein-
forcing bars surrounded from mortar in cavities of masonry units. Another more modern
and effective method is to use textiles for masonry strengthening taking into account the
same advantages of reinforcing concrete with textiles. Textiles can be embedded in thin-
layer mortar not needing such a thick mortar cover like rebars. This enables the use of thin-
layer mortar, which is the leading mortar in the German market due to its significantly
larger compressive strength and its high bond performance to masonry units compared to
general purpose mortar [2]. Textiles can be also applied in mortar-based systems like ren-
dering or plaster bonded to the tensile strained side of masonry surfaces enabling the
strengthening of masonry externally. For this reason it is one of the most suitable methods
for strengthening particularly existing masonry constructions like heritage buildings and
therefore nowadays gaining upswing in Mediterranean countries subjected to earthquake.
A summary of methods for mortar-based strengthening of masonry structures is given in
[3].
The aim to reinforce the bed joints with textiles is to improve the shear strength and the
flexural tension strength parallel to the bed joints (see Fig. 1). Apart from this, it is at-
tempted to improve the flexural tension strength parallel and perpendicular to the bed joint
(see Fig. 2) as well as the shear strength of masonry by applying textile reinforced render-
ing on the masonry surface. This reinforcing method may be less effective than the other
one, because of the externally application. Combining the two types of textile reinforce-
ment a much better three-dimensional strengthening of new masonry buildings could be
achieved.
Fig. 1 Textile-reinforcement embedded in the bed joint for strengthening of new masonry constructions
510
Fig. 2 Textile reinforced plaster or rendering for externally bonded strengthening of masonry
To ensure strengthening of masonry using textiles a high bond performance between the
composite materials is required. On one hand the transfer of tensile stresses acting on ma-
sonry building components is determined by the bond properties between reinforced mor-
tar and masonry unit surface. On the other hand once the mortar has cracked, the capacity
to transfer and absorb higher acting tensile stresses than the mortar tensile strength depends
on the bond performance between textile and mortar. This demonstrates the importance of
material compatibility and how their mechanical properties can be improved if used in con-
junction. Therefore, the first milestone of a recent research project investigating textile
reinforced masonry especially for new constructions is to find suitable textiles for the use
in bed joints and in rendering/plaster. The bonding behaviour under tensile load of textile
reinforced mortar should be examined with the help of uniaxial tensile tests [4]. In this
process different geometries and material properties of textiles, mortar as well as rendering
or plaster should be taken into account.
TENSILE TESTS
General
With the aim to characterize the stress-strain behaviour of textile reinforced mortar for
uniaxial tension loading, uniaxial tensile tests are being conducted according to the rec-
ommendations of RILEM TC 232-TDT [4]. These recommendations include information
about the specimen shape, the kind of load introduction and the anchorage length of the
specimens. The testing method derives from the round robin tensile test for textile rein-
forced concrete [5] carried out by seven participants from all over the world.
As mortar is a different material with a significantly lower compressive strength than con-
crete, the uniaxial tensile test method may need to be adapted to the specific material prop-
511
erties of mortar. The first results of the ongoing tensile tests on textile reinforced mortar
specimens are presented below. Besides varying the material combination (2 mortars in
combination with 2 textiles), the load introduction type (glued or screwed steel plates) and
the test direction of the textiles (warp and weft direction) were varied to optimize/adapt the
test method for textile reinforced mortar.
Materials
The process of choosing testing materials will be iterative in order to find the most suitable
textile and mortar or rendering for the selected application. Up to now, two AR-glass tex-
tiles were used (see Fig. 3). These textiles were chosen primarily to calibrate the test meth-
od, without regarding their properties and aptness for the intended specific use in masonry
at all. One of the textiles is uncoated (T1) and has been often used as reference for many
investigations on textile reinforced concrete in the past. The other textile is coated with
polyvinyl acetate (PVAC) (T2) and usually finds application in renderings and thermal
insulation composite systems. The characteristics of both textiles are shown in Table 1.
With respect to the mortar two representatives of mortar compression classes as well as
application thicknesses were chosen so far. The thin-layer mortar (M1) is suitable for the
bed joint of thin-layer masonry, while the general purpose mortar (M2) is applied in thick-
bed masonry. The material properties of both mortars are given in table 2.
Fig. 3 Structure of the AR-glass textiles T1 (left) and T2 (right)
Table 1: Characterization of the AR-glass textiles
weave construction coating yarn fineness mesh width

textile
0° 90°
type
- - [tex] mm
T1 core - 2400 8.4 8.3
T2 turner PVAC 320 6 5
512
Table 2: Characterization of the mortars according to DIN EN 1015
grain oven-dry compression flexural tension

size density strength strength
mortar type class
[mm] [kg/dm³] [N/mm²] [N/mm²]
M1 thin-layer mortar M10 0-1 0.84 10.80 2.50
M2 general purpose mortar M5 0-4 1.78 5.40 2.10
Test Specimens
The mortar specimens tested were strips with rectangular shape (l = 500 mm, w = 60 mm
and t = 10 mm) and a width ratio of 5:1. These parameters were chosen according to the
recommendation in [4]. The specimens constituted of two layers of mortar and one layer of
textile located in between. To cast the specimens, a hand-lamination process was used.
Within a steel formwork, first a thin mortar layer was applied, then the textile was put over
this mortar layer and covered with a second thin mortar layer. Fig. 4 shows the production
of the specimens. The specimens were stored in a humid environment, demoulded after 2
days and stored in water for another 12 days. Finally, they were stored in laboratory cli-
mate (20 °C/65 % rel. humidity) 1 day before testing. A total of 8 series with each 6 spec-
imens were produced.
Fig. 4 Production process of the rectangular specimens
Test Method
The test procedure used to analyse the tensile stress-strain behaviour of textile reinforced
mortar specimens is shown in Fig. 5. During the experiments particularly attention was
paid in gripping the specimen. On one hand a slipping between clamping and specimen
513
should be avoided by gripping the specimen with sufficient and homogeneous distributed
compression. On the other hand serious damages on the mortar in the clamping area should
be avoided by not exceeding the low mortar compression strength. In the majority of cases,
the load introduction was carried out by steel plates which were screwed to the mortar
specimen according to [4] (Fig. 5, left). The load was transferred to the specimen by means
of friction. To avoid local load concentrations during the test, a rubber interlayer was col-
located between the steel plates and the specimen. To exert a sufficient and controlled
tightening pressure, the screws were tightened with a specific moment of torque using a
torque wrench. From the set moment of torque, the pre-compression on the gripping area
can be calculated approximately.
Alternatively to the gripping method with screws, steel plates were glued to the specimen
one day before testing. The advantage of this method compared to the other one is, that no
mortar damage in the clamping area may occur, because no pressure must be exerted on
the specimen for clamping the steel plates. In Fig. 5, right, this testing method is shown.
The clamping length of both testing methods was 115 mm. The uniaxial tensile tests were
carried out in a displacement controlled way with a rate 0.5 mm/min. The load was meas-
ured with standard load cells. The deformations were recorded by compact strain transduc-
ers (DD1) attached on both surfaces in the middle of the specimen. The base length ana-
lysed was 250 mm for both testing methods.
Fig. 5 Test setup; screwed steel plates (left) and glued steel plates (right)
514
Results
During the tests the following problems were observed:
• partial or full pull-out of fibres (Fig. 6)
• partial compressive failure of mortar in the clamping area
• failure outside the measuring range
• no rupture of textile
Similar problems were also reported within the scope of optimizing the test method for
textile reinforced concrete [5]. Some recommended solutions for textile reinforced con-
crete are however material dependent and must be adapted for textile reinforced mortar.
Since pull-out of fibres may occur under relatively low gripping pressure, the applied pre-
compression on the clamping area was increased from originally 2.00 N/mm² to 4.25
N/mm² for new test series (Fig 10, left). Despite tightening the specimens of these new test
series with a higher pre-compression, partial pull-out failure of the yarns could not be
avoided at all. For this reason and due to the less inner bond between the filaments no
complete rupture of the uncoated textile occurred apart from the outer filaments of the
yarns as shown in Fig. 6. In addition to the partial pull-out problem, in some tests partial
compressive failure of the mortar in the clamping area led to pull-out failure outside of the
measuring area.
The stress-strain curves were determined for each tensile test. The measured force was
related to the cross-section of the specimen to obtain the tensile stress. To determine the
strain of the tensile specimen, the measured deformation of the strain transducers were
averaged and then related to the measuring length. In addition, the first crack stress and the
type of failure as well as the crack pattern were documented for each single test. The re-
sults were evaluated with respect to the main influencing parameters tested: testing meth-
od, material combination, applied gripping force and textile test direction (see Fig. 7, 8 and
10).
In the tests only two of the three stages of the concrete multi-layer stress-strain-behaviour
[5] were achieved, namely uncracked and multiple cracking. The multiple cracking was
defined as diffuse. The third stage, namely complete cracking [5], when the tensile strength
of the textile is reached, was not recorded. The comparison between both testing methods
for the same material combination shows that the tensile stresses obtained with the glued
steel plates were lower than with the screwed steel plates (Fig. 7). This demonstrates that
without applying a pre-compression force in the clamping area the anchorage length of the
specimen with the glued plates should be increased to improve the bond between textile
and mortar and thus to obtain better results.
515
tensile stress σ [N/mm²] M1-T1

4
1
glued steel plates
screwed steel plates (2.00 MPa)
0
0 2 4 6 8
longitudinal strain ε [mm/m]
Fig. 6 Tensile failure of external yarns and pull- Fig. 7 Stress-strain curves of specimens
out of internal yarns under uniaxial tension; weft direction;
comparison of specimen gripping
Fig. 8 presents a comparison of the stress-strain curves obtained by testing different mate-
rial combinations. The comparison of both textiles applied in combination with mortar M1
generally demonstrates a poor bond of textile T2 with the mortar in opposition to textile T1
(Fig. 8, left). One of the reasons for less bond performance between textile T2 and mortar
M1 could be the small mesh size and the smooth surface of this textile due to its coating.
The results of the tensile tests on specimens reinforced with textile T1 using two types of
mortar (Fig. 8, right) show the influence of the mortar on the curve progression.
Due to the low compressive strength of mortar M2 (5 N/mm²) the first crack stress reached
was much smaller than of mortar M1 (10 N/mm²). Fig. 9 shows the failure case (pull-out of
fibres) and the crack pattern of these specimens. Both material comparisons (textile and
mortar) generally give evidence of the importance of selecting matching materials to im-
prove the properties of the composite material.
516
tensile stress σ [N/mm²] M1 tensile stress σ [N/mm²] T1

4 4
T2 M2
3 T1 3 M1
2 2
1 1
0 0
0 2 4 6 8 10 0 2 4 6 8
longitudinal strain ε [mm/m] longitudinal strain ε [mm/m]
Fig. 8 Stress-strain curves of specimens under uniaxial tension; weft direction; comparison of both textiles
(left) and comparison of both mortars (right)
Fig. 9 Crack pattern of reinforced mortar specimens using M1 (left) and M2 (right), both reinforced with
the AR-glass T1
Comparing the two diagrams in Fig. 10, the influence of the pre-compression for specimen
gripping (2.0 N/mm² or 4.25 N/mm²) and the influence of the testing direction (weft or
warp) can be analysed. All specimens shown in these diagrams are made with the same
material combination (M1-T1) and reached, due to the mortar used, approximately the
same first crack stress. Both the improvement of the anchoring as well as the change of
textile testing direction from weft to warp reached a higher maximal stress before failure.
517
tensile stress σ [N/mm²] M1-T1 tensile stress σ [N/mm²] M1-T1

8 8
weft direction (4.25 MPa) warp direction (4.25 MPa)
weft direction (2.00 MPa)
6 6
4 4
2 2
0 0
0 2 4 6 8 0 2 4 6 8
longitudinal strain ε [mm/m] longitudinal strain ε [mm/m]
Fig. 10 Stress-strain curves of specimens under uniaxial tension for both testing directions
Concluding, textile reinforced mortar shows a similar bonding behaviour under tensile load
as textile reinforced concrete. Despite similarities, extensive investigations on textile rein-
forced mortar for reinforcing masonry are required, due to differences concerning the spe-
cific mortar properties – influenced additionally by contact with the masonry unit – and the
planned specific application.
OUTLOOK
In further uniaxial tensile tests the anchorage length of longer specimens should be extend-
ed to avoid pull-out failure as far as possible and to determine the minimum anchorage
textile length. Apart from the pre-compression and the length of gripping, pull-out of fibres
may occur with textiles providing a poor bond/interlocking with the mortar matrix. For this
reason besides improving the test method, suitable textiles should be selected based on the
gained results and findings for both reinforcing methods – in the bed joint or externally
bonded on the masonry surface. In addition to the bond performance between textile and
mortar, the bond performance of reinforced mortar/rendering to the masonry unit/surface
should be investigated. The overriding goal is to investigate and analyse the load-bearing
and deformation behaviour of textile reinforced masonry components for their dimension-
ing.
REFERENCES
[1] Schmidt, U.: Modelling of masonry under horizontal loads. Bagneux: RILEM, 2010. In: Proceedings of
the International RILEM Conference on Materials Science (MatSci), Vol. II: HetMat Modelling of Het-
erogeneous Materials, Aachen, September 6-8, 2010, (Brameshuber, W. (Ed.)). pp. 125-134 ISBN 978-
2-35158-108-7
[2] Brameshuber, W.; Saenger, D.; Winkels, B.: Recent developments in masonry construction. In: Mau-
erwerk 18 (2014), Nr. 3/4, pp. 151-163 ISSN 1432-3427
518
[3] Papanicolaou, C.; Triantafillou, T.; Roca Fabregat, P.; Increase of load-carrying capacity of masonry
with textile reinforced rendering, Mauerwerk (Masonry), Vol. 19, No 1, 02.2015, pp. 40-51, ISSN
1432-3427
[4] RILEM TC 232-TDT: Recommendation of RILEM Úniaxial tensile test - Test method to determine the
load bearing behaviour of tensile specimens made of textile reinforced concrete´, Materials and Struc-
tures, 2015, to be published
[5] Brameshuber, W. et al.: Round Robin Tensile Test for Textile Reinforced Concrete, Materials and
Structures, 2015, to be published
519
520
APPLICATIONS OF FERROCEMENT IN STRENGTHENING OF

HEAT DAMAGED REINFORCED CONCRETE CIRCULAR
COLUMNS
Danie Roy A.B.*, Umesh Kumar Sharma** and Pradeep Bhargava***, * Research Scholar, Civil
Engineering Department, Indian Institute of Technology Roorkee, danieroy2002@yahoo.com,
**Associate Professor, Civil Engineering Department, Indian Institute of Technology Roorkee, ***
Professor Civil Engineering Department, Indian Institute of Technology Roorkee
Abstract : The main purpose of the present investigation is to examine the capability of
commonly used Ferrocement techniques in restoring the structural capacity of heat dam-
aged reinforced concrete columns. A series of 33 reinforced concrete circular short column
of size 150 x 450 mm were cast using normal strength concrete. After 150 days of ageing,
concrete specimens were exposed to elevated temperatures ranging from room temperature
to 900 ºC in an electric furnace. The overall response of strengthened specimens was inves-
tigated vis-à-vis un-strengthened specimens in terms of axial compression, axial displace-
ment and axial stress strain behaviour. It was observed that the columns exposed to differ-
ent temperatures experienced a reduction in ultimate load carrying capacity. The study
shows that Ferrocement jacketing techniques though indicated only nominal improvement
in the strength capacities, however they were quite capable in restoring the secant stiffness
capacity of heat damaged RC columns.
INTRODUCTION
Since columns are the most important structural components of any building or structure,
investigating damage caused in columns due to fire or heat and their subsequent restoration
becomes important. As columns are the most commonly found in buildings, this study
deals with the influence of elevated temperature and resulting restoration of short circular
columns. Previous studies have shown that Ferro cement jacketing (FC) can also provide
an effective confinement strengthening in such concrete columns (1-7). However, only
limited research has been reported on the repairing of fire damaged concrete columns (8 &
9). The present study explores the potential of use of above mentioned strengthening
schemes in restoring the compressive mechanical performance of RC circular columns that
have been damaged by heat by subjecting them to certain elevated temperatures. The effec-
tiveness of FC jackets on heated damaged reinforced concrete columns in terms of ductili-
ty, stiffness, strength gain, energy dissipation, failure mode were studied under this study.
EXPERIMENTAL PROGRAMS
An experimental programe was designed to examine the efficiency of above mentioned

techniques to restore the structural performance of heat damaged circular columns. A se-
521
ries of 33 reinforced concrete circular columns were constructed using normal strength
concrete. The details of the specimens are illustrated in Table 1 and Fig. 1. The first two
letters (CC) in the abbreviation denote circular column, the numeral (3, 6 and 9) indicates
the temperature i.e. 300°C, 600°C & 900°C, CY indicates cyclic heating, NA refers to
Natural Cooling and WQ for Water quenching specimens.
Temperature of Monotonic / Cy-

Designation Type of Cooling
exposure (°C) clic Heating
CCA Ambient ---- ----
CC3
300 Monotonic Normal
CC3 FC
CC3 CY
300 CY Cyclic Normal
CC3CY FC
CC6
600 Monotonic Normal
CC6 FC
CC9 NA
900 NA Monotonic Normal
CC9 NA FC
CC9 WQ
900 WQ Monotonic Water Quenching
CC9 WQ FC
Table 1 Specimen Details
Materials properties
Concrete was prepared with ordinary Portland cement, Natural River sand, and crushed
limestone aggregate of maximum size 12.5 mm with tap water. Cylinder compressive
strength of 37.19 MPa was obtained after 28 days. The concrete mix consisted of 450
kg/m3 Portland cement, 658 kg/m3 washed sand, 1034 kg/m3crushed limestone, and 202.5
kg/m3 tap water. The steel used for longitudinal reinforcement was 10 mm diameter with
520 MPa yield strength and while that used for lateral ties 6 mm diameter bars of 590 MPa.
The weld mesh used in the FC jacketing had square openings of (13 x 13 mm) and wire
diameter of 0.96 mm with of yield strength 385 MPa. The slurry used in preparing FC
jacket included ordinary Portland cement, natural sand which was less than 600 micron and
522
silica fume. The slurry mix proportions were 1:0.6:0.15:0.35:0.01 by weight of cement,
sand, silica fume, water and super plasticizer respectively. Prepared slurry mix had a com-
pressive strength of 68.06 MPa and had high flow-ability (32 seconds) as measured by a
standard ASTM C939 flow cone test.
Figure 1. Size of specimens, reinforcement details and position of thermocouples
Casting, curing and thermal testing
A concrete cover of 12.5 mm was provided at side in all the cylinders. A cover of 15 mm
was also provided between the ends of the longitudinal bars at the top and bottom surfaces
of the specimens to prevent direct loading. The water curing period lasted for 28 days after
which the specimens were kept in the laboratory at ambient temperature and humidity con-
ditions for another 122 days.
A programmable electrical furnace intended for a maximum temperature of 1200 °C was

used for heating the specimens. The temperature inside the furnace was measured and rec-
orded with specially installed thermocouples. After 150 days, the specimens are heated in
the furnace to different target temperatures ranging from ambient temperature to 900 °C.
Rate of heating was set at 10 °C/min and each target temperature was maintained for 3 hrs
to achieve a thermal steady state (Fig 2). After exposing the specimens to target tempera-
tures for the desired time duration, the specimens were allowed in the furnace for natural
cooling till room temperature.
523
Figure 2. Typical time-temperature curves
Observation during heating
Assessment of fire damaged concrete usually starts with visual observation of colour
change, cracking and spalling of concrete surface. The colour of concrete cylinder changed
to pink when the specimens were exposed to 300n°C temperature and became light greyish
at 600 °C. However, the specimen’s changed to ash white colour when exposed to 900 °C.
There was no visible crack on the surface of the specimens heated up to 300 °C. However
while insignificant hairline cracks were observed at 600 °C. The number of crack became
relatively pronounced at 900 °C. No spalling was observed in 300 °C and 600 °C tempera-
tures. On exposure to temperature of 900 °C no instant spalling was recorded in any of the
specimens and the number of cracks became relatively pronounced at 900 °C (Fig 3). The
structure of the cement mortar after temperature exposure was observed to have become
loose because of the pore expansion owing to the vaporization of the absorbed water. No
spalling was noted in any of the heated and naturally cooled specimens. However in case
of water quenched specimens spalling was observed (Fig 3).
Figure 3. Test specimens exposed to different temperature
524
Repairing and strengthening of heat damaged columns
After cooling, the heat damaged columns which were heated to 900 °C were cracked and
spalled columns were first repaired and their section was restored before taking up
strengthening. The weak concrete was removed by means of steel wire brush, chisels and
hammer. The surfaces of specimen were cleaned thoroughly to ensure no dust remains on
surface. On the spalled area a bonding agent was applied on the surface of column to ac-
complish good bonding between the old concrete and new repair material i.e micro con-
crete (Fig 4). After repairing the heat damaged specimens, wherever required, all the spec-
imens were strengthened with FC jacketing. In all the specimens a primer epoxy bonding
was applied after the surface was cleaned in order to provide good bonding between the
substrata and the new strengthening material. A gap of 20 mm in un-strengthen concrete
was left at the ends of concrete specimen to prevent jacket from direct loading while test-
ing. The FC jacketing was carried out with two layers of welded wire mesh as shown in
(Fig 4). An overlap of 100 mm was provided in the lateral direction of the wire mesh.
Wooden mallets were used to keep the wire mesh close to the surface of prisms. Slurry
with high compressive strength of 68.06 MPa was poured into the mould to form 25mm
thick FC jacket around the heated specimen. After 24 hrs, the strengthened specimens were
demoulded and cured using damped gunny bags for another 28 days.
Figure 4. Application of bonding agent and application of micro concrete
Instrumentation and test setup
The specimens were tested after complete cycle of heating, cooling and strengthening.
Wooden ply piece of 3 mm were placed at the ends of the specimens before testing to en-
sure leveled surfaces and to distribute the load uniformly. The specimens were tested under
monotonic concentric compression at a constant displacement rate of rate of 0.1mm/min to
capture the complete post peak behavior using 2500 KN capacity UTM. The axial defor-
mation of the specimens at the central region (gauge length 200 mm) was monitored with
the help of two linear variable displacement transducers (LVDT). Two horizontal LVDTs
525
were also fixed on opposite faces of the column. Loads were recorded through in built load
cell in the universal testing machine (UTM). The recorded data from the LVDTs and load
cell were fed into a data acquisition system and stored on a computer.
ANALYSIS AND DISCUSSION OF RESULTS
The results are presented graphically in terms of load-deformation and stress-strain re-
sponse based on the test data. The axial compressive behaviour of strengthened columns in
terms of their stiffness, ductility, ultimate strain and the ultimate strength is evaluated and
compared to those of the control specimens and heated damaged specimens. Load strain
curves were plotted and strength, stiffness and deformability were computed.
Test observations and failure modes

Crushing failure was observed in all control and heat damaged columns, at the ends due to
the effect of stress concentration in these regions. The failure was particularly brittle, sud-
den and explosive, in nature in case of control specimens. But in case of heat damaged
specimens the failure was gradual and exhibited ductile behavior with the increase in load
(Yaqub et al. 2013 and Roy et al. 2014). In case of specimens heated to 900 °C, as the lat-
eral ties got exposed to heat due to spalling of cover concrete, a crushing failure and open-
ing of ties was noted. Generally in FC jacketed specimens also, the initial cracks formed at
the ends of the specimen. With gradual increase in loading numerous vertical cracks were
observed throughout the surface of the specimen. Beyond peak load the wire mesh started
to bulge with breaking noise. This condition was observed in all the FC specimens. The
steel wires of the welded mesh in the vertical direction bulged, whereas those in horizontal
direction got broken in full length of the specimens, resulting in crack propagation
throughout the length of the specimen in vertical direction. In this case the failure and sep-
aration between the core and FC laminates was observed. The FC jacketed specimen’s
failure mode is shown in Fig. 5.
Figure 5. Failure pattern of control, heat damaged and strengthened specimens
Influence of FC strengthening on axial strength

It can be seen that a reduction in strength in the range of 4 % to 81 % was observed after
heating the specimens. The Strength of heat damaged FC jacketed specimens the axial
526
compressive strength was higher than the unstrengthen heat damaged specimens, but re-
mained less than the control specimens except for 300 °C heat damaged specimens, Fig. 6.
1000
1000
CCA CCA
CC3
800
800
CC6
CC3 FC
CC6 FC
Axial Load (kN)

CC3 CY
Axial Load (kN)
600
600 CC3 CYFC
400
400
200 200
0 0
0 1 2 3 4 0 1 2 3 4
Axial Displacement (mm) Axial Displacement (mm)
1000
800
CCA
Axial Load (kN)
CC9
600
CC9 FC
CC9 WQ
400 CC9 WQFC
200
0
0 2 4 6 8 10 12 14
Axial Displacement (mm)
Figure 6. Comparison of control, heat damaged specimens and heat damaged strengthened columns
Effect of strengthening on stiffness

In this study, strain of the specimens was measured as change in length normalized with
original length and the secant stiffness of all the specimens was calculated from the meas-
ured ultimate axial load divided by the measured ultimate displacement at the mid-height
of the specimens (Roy et al. 2014). After heating to different temperatures, it was noted
that the stiffness of the heat damaged columns reduced considerably when compared to
control columns, This is attributed to micro cracking and softening of the concrete due to
the evaporation of water after heating. The stiffness of FC jacketed specimens was higher
than both the control and heat damaged specimens up to 600 °C. However in specimen
heated to 900 °C, it was much greater compared to the heat damaged specimens but less
than the of control specimens. This may be due to an increase of cross-sectional size, im-
provement of dimensional stability and integrity of the material.
527
CONCLUSIONS
The following points highlight the principal findings and conclusions derived from the test
results:
1. Subjecting the RC columns to elevated temperatures for a period of 3 hrs reduced their
axial strength by about 4 %, 8 %, 30 %, 79 % and 81 %.
2. The use of circumferential jacketing with FC composites alone was insufficient to re-
store the nominal axial strength heat damaged columns.
3. The repair schemes investigated in this study failed in restoring the original axial
Strength of the undamaged columns.
4. The reduction in axial stiffness of the heat damaged columns was more prominent than
the reduction in compressive strength
REFERENCES
1. Mansur, M.A., and Paramasivam, P., ‘Ferrocement under Combined Bending and Axial Loads’, Inter-
national Journal of Cement Composites and Lightweight Concrete. 7 (3) (1985) 151-158.
2. Mansur, M.A., and Paramasivam, P., ‘Ferrocement Short Columns under Axial and Eccentric Compres-
sion’, ACI Structural Journal. 87 (5) (1990) 523-529.
3. Nedwell P, Ramesht M, Rafei-Tanhanaki S., ‘Investigation into the Repair of Short Square Columns
Using Ferrocement’, In : Proceedings of the Fifth International Symposium on Ferrocement. (1990)
277–85.
4. Takiguchi, K. and Abdullah, ‘Experimental Investigations on Ferrocement as an Alternative Material to
Strengthen Reinforced Concrete Columns’, Journal of Ferrocement. 30 (2) (2000) 177-190.
5. Rathish, K.P, Rao, C.B.K., ‘Constitutive behavior of high-performance ferrocement under axial com-
pression’. Magazine Concrete Research, 58 (10), (2006). 647–56.
6. Kondraivendhan, B., and Pradhan, B., ‘Effect of ferrocement confinement on behavior of concrete’,
Construction and Building Materials, 23 (3), (2009), 1218-1222.
7. Xiong, G.J., Wu, X.Y., Li, F.F., and Yan, Z., ‘Load carrying capacity and ductility of circular concrete
columns confined by ferrocement including steel bars’, Construction and Building Materials. 25 (2011)
2263–2268.
8. Yaqub, M., Bailey, C.G., Nedwell, P., Khan, Q.U.Z., and Javed, I., ‘Strength and stiffness of post-
heated columns repaired with ferrocement and fibre reinforced polymer jackets’, Composites: Part B 44
(2013) 200–211.
9. Danie Roy, A.B., Sharma, U.K., and Bhargava, P., ‘Strengthening of Heat Damaged Reinforced Con-
crete Short Columns’, Journal Structural Fire Engineering 5 (4) , (2014), 381-398.
528
REPAIR CORRODED REINFORCED CONCRETE COLUMNS

USING ALKALI-ACTIVATED SLAG FERROCEMENT
Shuai Fang, Eddie Siu-Shu Lam, and Wing-Ying Wong, Dept. of Civil & Environmental
Engineering, The Hong Kong Polytechnic University, Hong Kong
Abstract: A new method for repairing corroded reinforced concrete columns using alkali-
activated slag (“AAS”) ferrocement is proposed. Ten full-scale column specimens were
prepared and seven of them suffered artificial accelerated corrosion at a mass loss of main
reinforcement of 8.9 %. Three corroded specimens were subsequently strengthened. This is
achieved by replacing concrete cover by AAS ferrocement jackets with two layers of stain-
less steel wire meshes (“SSWM”). The specimens were tested under axial, small and large
eccentric compression, respectively. Experimental results have shown the corroded col-
umns suffer loading capacity losses from 21 to 30 % as compared with the control speci-
mens. AAS ferrocement jacket can enhance the loading capacity of corroded columns by
18 to 38 %. Confinement effect of AAS ferrocement is observed on specimens under axial
and small eccentric compression.
INTRODUCTION
Old reinforced concrete structures in coastal areas are found to exhibit different extents of
defects caused by reinforcement corrosion, such as spalling of concrete and deteriorative
loading capacity, which have posed a serious threat to structural safety. Kashani et al. [1]
indicated that corroded reinforcements with mass loss of 15.5 % experienced a reduction of
78 % in elongation compared with control ones. Corrosion also weakened the buckling
capacity of reinforcements. In addition, expansion stress due to corrosion of reinforcement
can result in cracks and spalling of concrete which exposes reinforcements to aggressive
environment and aggravates their corrosion [2, 3]. Ferrocement has been successfully used
to strengthen reinforced concrete structures because of its outstanding mechanical perfor-
mance, fire resistance and low cost [4-6]. Ganesan and Anil [6] investigated the strength of
RC columns confined by ferrocement under axial compression. Increases of 8 %, 29 %,
and 42 % in strength were achieved by confinement of ferrocement with 2, 3, and 4 layers
of meshes.Compared with Portland cement, slag is regard as a more environmentally
friendly material which reduces energy consumption by 90 % in production [7]. It can be
activated in alkaline environment and has higher strength than cement [7-9]. Moreover,
AAS presents strong resistance to chemical attack because of dense microstructure and a
low calcium silicon ratio in hydration products ensuring chemical stability in corrosive
environment [7]. In this paper, a new type of ferrocement, comprising of AAS mortar, is
proposed to strengthen corroded reinforced concrete columns.
529
EXPERIMENT
Ten specimens were prepared. One specimen was used to test corrosion degree of rein-
forcements, and the others were divided into three groups according to loading type as fol-
low: 1. axial compressive test, 2. small eccentric compressive test with an eccentricity of
30 mm, and 3. large eccentric compressive test with an eccentricity of 130 mm. Each group
includes a control specimen, a corroded specimen and a strengthened specimen for test
(Table 1).
Table 1 Details of specimens
ID of specimens Details Loading types
A0 Corroded -
A1/A2/A3 Axial compression
B1/B2/B3 Control / Corroded / Strengthened Small eccentricity
C1/C2/C3 Large eccentricity
Material properties
One mix proportion for C30 concrete with a water ratio of 0.7 was considered. The materi-
als used in preparing the concrete include ordinary Portland cement, crushed granitic
coarse aggregate and local river sand. Sodium chloride (1 % by mass of cement) was
mixed into concrete. Mix proportion of AAS mortar was optimized from a preliminary
study. It exhibited 28-day flexural and compressive strength of 4.8 MPa and 110.9 MPa,
respectively. Activators used in this study are liquid sodium silicate with a concentration of
47.6 % and sodium hydroxide with a purity of 99 %. Mix proportions of concrete and AAS
mortar are shown in Table 2.
Rebars of 12 mm diameter and plain bars of 6 mm diameter were used as the main rein-
forcement and stirrup, with yield strength of 548 MPa and 444 MPa, respectively. Ferro-
cement used to strengthen columns consists of two layers of SSWM encapsulated in AAS
mortar. The SSWM has grids of 9.8 mm × 9.8 mm and wire diameter of 0.9 mm. Direct
tensile test was performed on single wire. Yield strength of stainless steel wire, defined as
the stress with a residual strain of 0.2 %, is 317 MPa.
530
Table 2 Mix proportions of concrete and AAS mortar (kg/m3)
Concrete AAS mortar
Water 210 Water 173
Cement 280 Slag 710
Coarse aggregate (10 mm) 395 Liquid sodium silicate 224
Coarse aggregate (20 mm) 790 Sodium hydroxide 41
Sand 705 Sand 1065
Sodium chloride 2.8 Barium nitrate 7.1
Total 2382.8 Total 2220.1
Preparation of specimens
The specimens are 900 mm height with a cross section of 200 mm × 200 mm (Fig.1). Con-
crete cover for main reinforcement is 38 mm referring to the design standard in Hong
Kong in 1950s. 90 ° hooks are used for all the stirrups. In addition, the stirrups and main
reinforcements at two ends of corroded and strengthened specimens were protected by
epoxy to avoid specimen failing in this part due to reinforcement corrosion. The specimens
were removed from the formwork 24 hours after casting, and were wet-cured for 28 days.
Averaged 28 days compressive strengths of the concrete, estimated using 100 mm cubes, is
32.5 MPa.
Fig. 1 Schematic view of specimens (Unit: mm) Fig. 2 Strengthening steps
531
After curing, specimens A0, A2, A3, B2, B3, C2, and C3 were subjected to direct current
to accelerate corrosion of reinforcement artificially. The specimens were immersed in wa-
ter and electric current was kept at 0.68 A for 42 days. To evaluate the effect of artificial
accelerated corrosion, main reinforcements and stirrups at the mid-height of Specimen A0
were cut off and cleaned in hydrochloric acid solution. Average mass loss of main rein-
forcements and stirrups are 8.91 % and 20.98 %, respectively. Mass loss of stirrups is more
than double that of main reinforcements. This is because the thinner concrete cover results
in earlier corrosion on stirrups. Corroded stirrups and undamaged main reinforcements
formed activated galvanic-couple cell which aggravated corrosion of stirrups acting as an-
ode [10]. Strengthening work was applied to specimens A3, B3, and C3. Strengthening
method consisted of three steps as follows: 1. Remove concrete cover to expose the rein-
forcements; 2. Wrap the concrete core with two layers of SSWM and overlap the SSWM at
a length of 120 mm; and 3. Damp the concrete core and cast AAS mortar over the concrete
core using wooden formwork (Fig. 2).
Experimental setup
Load was applied through a universal testing machine by displacement control. Loading
rate is 0.1 mm/min in the pre-peak portion and 0.05 mm/min in the post-peak portion.
Compressive test was terminated when the loading capacity reduced to 80% of the maxi-
mum.
Failure modes
Fig. 3(a)-(c) show failure modes of axial compressive specimens. Failure of specimen A1
was initiated by longitudinal cracks in concrete, and followed by main reinforcement buck-
ling and concrete failure. Compared with the control specimen, the corroded specimen A2
possessed a brittle failure. With increasing load, width of concrete cracks increased rapidly
to the maximum of 5 mm. When the load reached its peak, concrete cover at mid-height
spalled off and in succession, main reinforcements buckled. Loading capacity experienced
a sudden drop. This phenomenon can be explained by evidences observed after test. A cor-
roded stirrup at mid-height was found fractured, which lost its confinement to concrete
core and resulted in a brittle failure. The strengthened specimen A3 lost its loading capaci-
ty on account of compressive failure of ferrocement jacket. A corner of ferrocement jacket
cracked at the top due to stress concentration and led to spalling of mortar. Mortar inside
SSWM was found to be crushed. However, specimen A3 exhibited a ductile response in
post-peak stage because of confinement action provided by SSWM on concrete core.
Fig. 3(d)-(f) show the failure modes of specimens under small eccentricity. These three
specimens lost their loading capacity due to failure of main reinforcements and concrete at
compression side. Longitudinal cracks developed at compression side and transverse
cracks developed at the other side. Ferrocement jacket of specimen B3 remained intact
when the load reached its maximum. The crooked rod in the black ellipse in Fig. 3(f) indi-
cates that concrete core has been crushed and lost its bonding with ferrocement jacket.
532
a. Specimen A1 b. Specimen A2 c. Specimen A3 d. Specimen B1 e. Specimen B2
f. Specimen B3 g. Specimen C1 h. Specimen C2 i. Specimen C3
Fig. 3 Failure modes of specimens
Fig. 3(g)-(i) show the failure modes of specimens under large eccentricity. Both the control
specimen C1 and corroded specimen C2 responded by yielding of main reinforcements
under tension and failure occured by compressive failure of concrete. More flexural cracks
at smaller spacing were observed on the specimen C1 as compared with specimen C2. For
the strengthened specimen C3, the corbel lost its load transferring ability before it reached
its capacity leading to failure.
533
Load-deformation relationships
Load-axial deformation relationships of specimens A1-A3 and mid-height moment-lateral
deflection relationships of specimens B1-B3 and C1-C3 are given in Fig. 4-6. Mid-height
moment was obtained by multiplying the load by measured eccentricity. The corroded
specimens experienced severe reduction in both loading capacity and stiffness. Compared
with the corroded specimens, loading capacities are improved for the strengthened speci-
mens and are close to the control specimens. They exhibited better ductility in post-peak
stage. The strengthening method also enhanced the stiffness and reduced the peak defor-
mation of specimens.
Fig. 4 Load-axial deformation relationships of speci- Fig. 5 Mid-height moment-lateral deflection rela-
mens under axial compression tionships of specimens under small eccentricity
Fig. 6 Mid-height moment-lateral deflection relationships of specimens under large eccentricity
Loading capacity and ductility

Yield load, peak load, yield deformation, maximum deformation and ductility of all speci-
mens and ratios of peak load of the specimens P to that of control specimen Pcon (P/Pcon)
are given in Table 3. Maximum deformation is assumed when the load is reduced to 85 %
of its maximum according to Park method [11]. Deformation ductility is obtained by max-
imum deformation divided by yield deformation. As shown in Table 3, loading capacities
of corroded specimens decreased by 28 %, 30 %, and 21 % as compared with control spec-
imens under axial, small eccentric, and large eccentric compression, respectively. The
strengthened specimens achieved increase in loading capacities by 37 %, 38 %, and 18 %
as compared with corroded specimens under axial, small eccentric, and large eccentric
534
compression, respectively. Strengthening effect on specimen under large eccentric com-

pression was less than that of specimens under axial and small eccentric compression. This
is due to the fact that corbel of the strengthened specimen failed prematurely.
Table 3 Loading capacity, peak deformation and ductility
Maximum
Yield Peak load Yield defor-
Specimen P/Pcon deformation Ductility
load (kN) (kN) mation (mm)
(mm)
A1 960.6 1117.3 1.00 0.49 1.39 2.84
A2 736.3 804.0 0.72 0.79 1.25 1.58
A3 1062.5 1104.5 0.99 0.67 1.64 2.45
B1 588.2 706.6 1.00 0.61 2.48 4.07
B2 429.0 497.0 0.70 0.73 2.66 3.64
B3 597.0 684.2 0.97 0.63 3.30 5.24
C1 207.9 235.7 1.00 1.36 3.89 2.86
C2 159.1 187.0 0.79 1.12 3.16 2.82
C3 200.8 220.1 0.93 0.97 2.19 2.25
CONCLUSIONS
In this paper, ten full-scale column specimens were prepared and nine of them were sub-
jected to three different types of compressive loads. Effect of corrosion of reinforcement
on performance of specimens was examined. Based on the experimental results, the fol-
lowing conclusions can be drawn:
1. Corrosion of reinforcement can seriously reduce the loading capacity, stiffness, and
ductility of reinforced concrete columns. A corrosion ratio of 8.9 % on main reinforce-
ments leads to a loading capacity loss up to 30 %.
2. Corroded columns strengthened by AAS ferrocement obtained loading capacity reha-

bilitations by 37 %, 38 %, and 18 % under axial, small eccentric, and large eccentric
compression, respectively. The strengthening method recovered the loading capacity of
corroded specimens to 93 % or above of original level.
3. The proposed strengthening method is effective and suitable for columns under axial
compression and small eccentricity than those under large eccentricity.
535
REFERENCES
[1] Kashani, M.M., Crewe, A.J., and Alexander, N.A., ‘Nonlinear stress–strain behaviour of corrosion-
damaged reinforcing bars including inelastic buckling’, Eng. Struct. 48 (2013) 417-429.
[2] Zhang, R., Castel, A., and Francois, R., ‘Concrete cover cracking with reinforcement corrosion of
RC beam during chloride-induced corrosion process’, Cem. Concr. Res. 40 (3) (2010) 415-425.
[3] Mohammed, T., Otsuki, N., Hisada, M., and Shibata, T., ‘Effect of crack width and bar types on
corrosion of steel in concrete’, J. Mater. Civ. Eng. 13 (3) (2001) 194-201.
[4] Shah, S.P., Lub, K.B., and Ronzoni, E., ‘A summary of ferrocement construction and a survey of its
durability’, Mater. Struct. 19 (4) (1986) 297-321.
[5] Kaushik, S.K., Singh, K.K., Kumar, R., and Sharma, T.P., ‘Performance of ferrocement encased
concrete under fire exposure’, J. Ferrocem. 26 (3) (1996) 191-196.
[6] Ganesan, N., and Anil, J., ‘Strength and behavior of reinforced concrete columns confined by ferro-
cement’, J. Ferrocem. 23 (2) (1993) 99-108.
[7] Shi, C., and Qian, J., ‘High performance cementing materials from industrial slags – a review’, Re-
sources, Conserv. Recycl. 29 (3) (2000) 195-207.
[8] Bilim, C., and Atis, C.D., ‘Alkali activation of mortars containing different replacement levels of
ground granulated blast furnace slag’, Constr. Build. Mater. 28 (1) (2012) 708-712.
[9] Fernández-Jiménez, A., Palomob, J.G., and Puertas, F., ‘Alkali-activated slag mortars: Mechanical
strength behavior’, Cem. Concr. Res. 29 (8) (1999) 1313-1321.
[10] Otsuki, N., Miyazato, S., Diola, N.B., and Suzuki, H., ‘Influences of bending crack and water-
cement ratio on chloride-induced corrosion of main reinforcing bars and stirrups’, ACI Mater. J. 97
(4) (2000) 454-464.
[11] Paulay, T., and Priestley, M.J.N., ‘Seismic design of reinforced concrete and masonry buildings’,
(Wiley Intersci. Publ., 1992).
536
537
538

Ferro Final 12052015

Uploaded by

Document Information

Original Title

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

Ferro Final 12052015

Uploaded by

Copyright:

Available Formats

FERRO-11

11th International Symposium on Ferrocement

Edited by Wolfgang Brameshuber

RILEM Publications S.A.R.L.

RILEM, The International Union of Laboratories and Experts in Construction Materials,

RILEM PROCEEDINGS (PRO)

PRO 77: International Conference on ‘Material Science - AdIPoC - Additions Improving

RILEM REPORTS (REP)

International Technical Committee

Local Organizing Committee

Textile reinforced Concrete – from the idea to a high performance mate-

Classification of approaches to dimensioning and assessment of light-

2 - Load Carrying Behavior

Textile reinforced refractory concrete at high temperature: mechanical

Bearing behavior of impregnated textile reinforcement page 71

Flexural behavior of mesh-and-fiber reinforced cementitious composites page 79

Hysteresis parameter identification and reliability assessment of ferro-

Behavior of precast ferrocement walls under cyclic loading page 103

Flexural behavior of mesh-and-fiber reinforced cementitious composites page 111

Strength assessment of sandwich-type ferrocement structural walls under

Analysis of tensile response of TRC using digital image correlation tech-

Implementation of experimental data in analyses of textile reinforced

Ultimate limit state assessment of TRC structures with combined normal

Analysis and geometric design of sealing junctions for precast ferroce-

Elastic-Plastic Analysis of I- Shaped Normal Strength Ferroce-

Study of earthquake resistant behavior in prefabricated ferrocement of

Analysis on pull-out behaviour of continuous multi-filament glass yarns

The Mechanics of Natural Jute Textile Reinforced Concrete page 253

Corrosion in Ferrocement page 265

6 - Maintenance and Sustainanbility

Combined Formwork and textile reinforcement System for a mineral

Structural concrete elements reinforced with textiles – sustainability per-

Electrical resistivity measurements on TRC tensile coupons page 317

Integrated Monitoring of TRC using carbon fibres page 325

Mesh Mould: Robotically Fabricated metal Meshes as Concrete Form-

Restoration of the historical concrete-glass windows of the town-hall in

Ferrocement solutions for affordable housing prototypes in India page 363

The morphogenetic design and influential demiurgic qualities of a hybrid

Proposed Joints for a Ferrocement Building Resi-dential System page 385

32 years with ferrocement - from small prefabricated bridges to pontoons

Ferro Past and Ferro Future page 405

Extrusion of textile reinforced concrete page 425

Shear behaviour of masonry walls strengthened by external bonded FRP

Shear strengthening of prestressed concrete beams with textile reinforced

Utilisation of linear traction separation bond law to improve the numeri-

Shear strengthening of RC beams using textile reinforcement in cement

Behaviour of hybrid composite-concrete beams under static flexural load-

Reinforced masonry using textiles page 509

Applications of ferrocement in strengtheining of heat damaged reinforced

Repair corroded reinforced concrete columns using alkali-activated slag

FERROCEMENT: PROGRESS REVIEW AND CRITICAL NEED

Antoine E. Naaman, Professor Emeritus, Department of Civil and Environmental Engineering,

SUMMARY-EVOLUTION OVER THE LAST FIVE DECADES

3. Evolution in Reinforcement. Here progress has been quite impressive on several

4. Evolution in Mechanical Performance. Bending resistance in ferrocement plates

5. Evolution in Equipment and Construction Methods. Here the evolution has

FERROCEMENT MODEL CODE AND CRITICAL NEEDS AND MISSION OF

“..The American Concrete Institute is a leading authority and resource worldwide

TEXTILE REINFORCED CONCRETE – FROM THE IDEA TO A

ALTERNATIVE REINFORCING MATERIALS

THE BEGINNING OF RESEARCH ON AND DEVELOPMENT OF TRC

Figure 3: Early textile reinforcement element, taken from DD 210102

Research in the 1990s in Germany