DURABILITY AND SUSTAINABILITY OF FIBER

REINFORCED POLYMER (FRP) COMPOSITES

FOR CONSTRUCTION AND REHABILITATION
Proceedings of the Fourth
International Conference (CDCC-11)
Qubec (Qubec) Canada July 20-22, 2011

DURABILIT ET CYCLE DE VIE DES MATRIAUX COMPOSITES

EN POLYMRES RENFORCS DE FIBRES (PRF) EN
CONSTRUCTION ET RHABILITATION
Comptes rendus de la quatrime
Confrence Internationale (CDCC-11)
Qubec (Qubec) Canada 20-22 juillet 2011

Editors diteurs

Brahim Benmokrane

Universit de Sherbrooke
Sherbrooke (Qubec) Canada

Ehab El-Salakawy

University of Manitoba
Winnipeg (Manitoba) Canada

Ehab Ahmed

Universit de Sherbrooke
Sherbrooke (Qubec) Canada

DOCUMENTATION

The organizing committee and the sponsors of the CDCC-11 are not
responsible for the statements or opinions expressed in this publication.
The papers contained herein are published in the exact form submitted by the
authors. Any statements of view expressed in the papers are those of the
authors. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
Le comit dorganisation ainsi que les commanditaires de la CDCC-11
dclinent toute responsabilit quant aux affirmations et points de vue exprims
dans le prsent ouvrage.
Les manuscrits contenus dans ce volume sont prsents dans la forme tablie
par les auteurs. Toutes les opinions avances dans les manuscrits
appartiennent leurs auteurs. Lutilisation des marques commerciales et des
noms de produits ne constitue en aucun cas un endossement ou une
recommandation.

ISBN: 978-2-7622-0196-3

While supplies last, additional copies of the proceedings may be obtained

from / On peut se procurer des exemplaires additionnels des comptes rendus
(Jusqu puisement des stocks) en sadressant :
CDCC-11
Department of Civil Engineering
Faculty of Engineering
Universit de Sherbrooke
Sherbrooke (Qubec)
CANADA J1K 2R1

FOREWORD
Applications of fibre reinforced polymer (FRP) composites in civil structures
have increased significantly in recent years. Since most of these structures are
designed with service lives of more than 50 years, the long term durability of
FRP materials in these applications is a critical issue.
The present conference is the fourth of its kind focusing on the durability,
sustainability and field applications of fibre reinforced polymers (FRP) for
construction and rehabilitation of structures (bridges, buildings, marine
structures, etc.). The first conference was held in Sherbrooke, Qubec, Canada
on August 1998, followed by one in Montral, Qubec, Canada on May 2002.
The third one was held in Qubec City, Qubec, Canada on May 2007. The
objective of this series of internationals conferences is to provide a forum to
academics, researchers, engineers, structure owners, FRP manufacturers, and
delegates from public and industrial institutions to present and exchange views
on present and future research on long-term durability and sustainability of
FRP composite materials in bridges and other infrastructure systems exposed
to harsh environmental conditions.
A total of 75 peer-reviewed papers are included in the proceedings. They have
been authored by various experts in the field from 19 different countries
around the world, including Algeria, Australia, Canada, China, Egypt, France,
Germany, Hong Kong, Italy, Iran, Japan, KSA, Netherland, Portugal,
Singapore, South Korea, Switzerland, United Kingdom and USA. On behalf
of the organizing committee, I am very thankful to the authors for their
valuable contributions to the conference. I am also very grateful to the paper
reviewers for their time and diligence.
We are thankful to the Department of Civil Engineering, Faculty of
Engineering, the Universit de Sherbrooke for their tremendous organizational
support without which this conference would not be possible. Financial and/or
logistic support from the conference sponsors: The ISIS Canada Network
Association, the Natural Science and Engineering Research Council of Canada
(NSERC), the Ministry of Transportation of Quebec (MTQ), the Research
Centre on Applied Polymers and Composites (CREPEC), the Research Centre
on Concrete Infrastructures (CRIB), and the American Composites
Manufacturers Association (ACMA) are gratefully acknowledged. We also
thank our corporate co-sponsors for their active participation.

We have been fortunate to receive help from a host of people in organizing

this conference. Professors Charles E. Bakis, Pennsylvania State University,
Nemkumar Banthia, University of British Columbia and John Myers, Missouri
University of Science and Technology deserve special mention. We are very
grateful to Mrs. Nathalie Valle, the Conference Secretary and Drs. Ahmed
Farghaly, Robert Mathieu, Patrice Cousin, Hamdy Mohamed, Brahim
Tighiouart and Mr. Christian Dulude for their assistance with the various tasks
to ensure the success of this conference.
We extend our heartiest welcome to the conference participants coming from
many near and far parts of the world.
The Editors,
Dr. Brahim Benmokrane, P. Eng.
NSERC & Canada Research Chair
Professor
Department of Civil Engineering
Faculty of Engineering
University of Sherbrooke

Dr. Ehab El-Salakawy, P. Eng.

Canada Research Chair Professor
Department of Civil Engineering
Faculty of Engineering
University of Manitoba

Dr. Ehab Ahmed, P. Eng.

Department of Civil Engineering
Faculty of Engineering
University of Sherbrooke
June 10, 2011

AVANT-PROPOS
Lutilisation des matriaux composites en polymres renforcs de fibres (PRF)
comme matriaux de construction dans le domaine du gnie civil a augment
de faon significative au cours des dernires annes. Compte tenu que les
structures de gnie civil sont conues pour des dures de service de plus de 50
ans, la durabilit long terme des matriaux composites de PRF constitue un
facteur majeur pour ce type dapplications.
La prsente confrence est la quatrime du genre qui traite de thmes se
rapportant principalement la durabilit, au cycle de vie et les applications sur
le terrain des matriaux composites en PRF pour la construction et la
rhabilitation de structures. Les ditions prcdentes ont eu lieu Sherbrooke,
Qubec, Canada en 1998, Montral, Qubec, Canada en 2002 et Qubec,
Qubec, Canada en 2007. Lobjectif de cette srie de confrences
internationales est de procurer un forum aux universitaires, chercheurs,
ingnieurs, propritaires douvrages, manufacturiers de produits de PRF et
membres dinstitutions publiques ou industrielles pour prsenter et changer
leurs points de vue sur les recherches prsentes et futures sur la durabilit
long terme et le cycle de vie des matriaux composites de PRF pour les ponts
routiers et dautres structures exposes des conditions environnementales
svres.
Soixante-quinze (75) articles valus sont inclus dans le compte-rendu de la
confrence. Ces articles ont t rdigs par plusieurs spcialistes dans le
domaine provenant de 19 pays dont lAlgrie, lAustralie, le Canada, la Chine,
lgypte, la France, lAllemagne, la Hong Kong, lItalie, lIran, le Japon,
lArabie Saoudite, la Hollande, le Portugal, Singapore, la Core du Sud, la
Suisse, lAngleterre, et les tats-Unis dAmrique. Au nom du comit
dorganisation, je remercie vivement les auteurs pour leurs contributions trs
importantes cette confrence. Je remercie aussi toutes les personnes qui ont
valu les articles pour leur temps et assiduit.
Nous remercions le Dpartement de gnie civil de la Facult de gnie de
lUniversit de Sherbrooke pour son support exceptionnel sans lequel cette
confrence naurait pas t possible. Les supports financiers et/ou logistiques
des commanditaires de cette confrence : ISIS Canada Network Association, le
Conseil de recherche en sciences naturelles et en gnie du Canada (CRSNG),
le ministre des transports du Qubec (MTQ), le Centre de recherche en
plasturgie et composites (CREPEC), le Centre de recherche sur les

infrastructures en bton (CRIB) et lAmerican Composites Manufacturers

Association (ACMA) sont grandement apprcis. Nous tenons aussi
remercier les co-commanditaires pour leur participation active.
Lorganisation de cette confrence a t rendue possible grce laide et la
participation de plusieurs personnes. Les professeurs Charles E. Bakis, de
Pennsylvania State University, Nemkumar Banthia, de l'University of British
Columbia et John Myers, de Missouri University of Science and Technology
mritent dtre particulirement mentionns. Nous sommes aussi
reconnaissants envers Mme Nathalie Valle, Secrtaire de la confrence et
MM. Ahmed Farghaly, Mathieu Robert, Patrice Cousin, Hamdy Mohamed,
Brahim Tighiouart et Christian Dulude pour leur assiduit dans lexcution de
leurs tches respectives qui garantissent le succs de cette confrence.
Nous souhaitons la bienvenue la plus cordiale toutes les participantes et
tous les participants qui viennent de pays proches et lointains assister cette
confrence.
Les diteurs,
Brahim Benmokrane, P. Eng.
Titulaire de Chaires du CRSNG
et du Canada
Dpartement de gnie civil
Facult de gnie
Universit de Sherbrooke

Ehab El-Salakawy, P. Eng.

Titulaire de Chaire du Canada
Dpartement de gnie civil
Facult de gnie
Universit de Manitoba

Ehab Ahmed, P. Eng.

Dpartement de gnie civil
Facult de gnie
Universit de Sherbrooke
10 juin 2011

ORGANIZING COMMITTEE / COMIT DORGANISATION

Honorary Chair of the Conference / Prsident honorifique de la confrence

Daniel Bouchard, Ministre des Transports du Qubec
Conference Chair / Prsident de la confrence
Brahim Benmokrane, Universit de Sherbrooke
Editorial Chair / Directeur scientifique
Ehab El-Salakawy, University of Manitoba
Co-Chairs / Co-prsidents
Charles E. Bakis, Pennsylvania State University
Nemkumar Banthia, University of British Columbia
John Myers, Missouri University of Science and Technology
Program Co-ordinator / Coordinateur du programme
Ahmed Farghaly, Universit de Sherbrooke
Information System Assistant / Coordonnateur des systmes d'information
Ehab Ahmed, Universit de Sherbrooke
Administrative Assistant / Assistante administrative
Nathalie Valle, Universit de Sherbrooke

NATIONAL ORGANIZING COMMITTEE /

COMIT ORGANISATEUR NATIONAL
Nemkumar Banthia, University of British Columbia
Brahim Benmokrane, University of Sherbrooke
Omar Chaallal, cole de technologie suprieure
Grard Desgagn, Ministry of Transportation of Quebec
Mario Desroches, Sika Canada
Bernard Drouin, Pultrall Inc.
Raafat El-Hacha, University of Calgary
Ehab El-Salakawy, University of Manitoba
Darrell Evans, P.E.I. Department of Transportation
Garth Fallis, Vector Construction Group
Mark Green, Queens University
Pierre Labossire, University of Sherbrooke
David Lai, Ministry of Transportation of Ontario
Aftab Mufti, University of Manitoba
Kenneth Neale, University of Sherbrooke
John Newhook, Dalhousie University
Ghani Razaqpur, McMaster University
Murat Saatcioglu, University of Ottawa
Shamim Sheikh, University of Toronto
Khaled Soudki, University of Waterloo
Allan Wiseman, Public Works Canada (PWGSC)

INTERNATIONAL SCIENTIFIC COMMITTEE /

COMIT SCIENTIFIQUE INTERNATIONAL
R. Al-Mahaidi, Australia

I. Mahfouz, Egypt

L. Bank, USA

U. Meier, Switzerland

N. Banthia, Canada

A. Mosallam, USA

J. Barros, Portugal

G. Monti, Italy

B. Benmokrane, Canada

A. Mufti, Canada

L. Bisby, UK

A. Nanni, USA

O. Chaallal, Canada

K. Neale, Canada

J.F. Chen, UK

D.J. Oehlers, Australia

O. Cosenza, Italy

R. Razaqpur, Canada

C. Dolan, USA

S. Rizkalla, USA

E. El-Salakawy, Canada

R. Sen, USA

M.A. Erki, Canada

C. Shield, USA

A. Fam, Canada

J. Sim, Korea

H. Fukuyama, Japan

S.T. Smith, China

H. GangaRao, USA

L. Taerwe, Belgium

P. Hamelin, France

B. Taljsten, Demark

T. Hamilton, USA

K.H. Tan, Singapore

K. A. Harries, USA

J.G. Teng, China

V.M. Karbhari, USA

R. Tepfers, Sweden

T. Keller, Switzerland

T. Triantafillou, Greece

K. Kobayashi, Japan

T. Uomoto, Japan

J.J. Lesko, USA

Z.S. Wu, Japan

SPONSORS / COMMANDITAlRES

Dpartement de gnie civil, Facult de

gnie, Universit de Sherbrooke
ISIS Canada Network Association

Natural Science and Engineering Research Council

of Canada (NSERC) / le Conseil de recherche en
sciences naturelles et en gnie du Canada (CRSNG)
Ministre des transports du Qubec (MTQ)

The International Institute for FRP in Construction

Research Center on Concrete Infrastructures

Centre de recherche sur les infrastructures en bton

The Canadian Society for Civil Engineering

Socit canadienne de gnie civil

Center for Applied Research on Polymers and Composites

Centre de recherche en plasturgie et composites

American Composites Manufacturers Association

CO-SPONSORS / CO-COMMANDITAlRES
Avensys Inc.
BP composites Ltd
Canadian Association for Composite Structures and Materials
(CACSMA)
Carbonfibreplus
FiReP Rebar Canada Inc.
Hughes Brothers Inc.
Hydro-Qubec
Industrial Materials Institute of NRC (IMI)
ITF Labs
MagmaTech Ltd
Ministre des transports du Qubec (MTQ)
Osmos Canada Inc.
Public Works and Governmental Services of Canada (PWGSC)
Pultrall Inc.
Roctest Ltd.
Schck Canada Inc.
Sika Canada Inc.
Vector Construction Inc.

TABLE OF CONTENTS / TABLE DES MATIRES

Alternative Environmental Knockdown Factors (CE) Derived From

Extensive Experimental Data .....................................................................................

K.A. Harries and B.M. Shahrooz

A Field Study on Durability of Bond Strength between CFRP Wraps and
Concrete.........................................................................................................................

11

S.A. Sheikh, A.D. Caspary and M. Stafford

Bond Behavior of Externally Bonded Fiber Reinforced Polymer
Laminates Subjected to In-Situ Service Loading and Environmental
Conditioning ..................................................................................................................

19

J.J. Myers and N. Muncy

Durability of GFRP Rods in Field Demonstration Projects across Canada ...........

27

A.A. Mufti, J. Newhook, B. Benmokrane, G. Tadros and H.M. Vogel

Effect of Temperature Change on the Behavior of FRP-to-Concrete
Bonded Joints ................................................................................................................

37

W.Y. Gao, J.G. Teng and J.G. Dai

Long-term Bond Behavior of GFRP Rebars in Severe Environments ....................

47

B. Juette, A. Weber and C. Witt

Effect of Various Severe Environmental Exposures on CFRP and SFRP
Confined Circular Concrete Columns ........................................................................

55

M.A. Mashrik and R. El-Hacha

Temperature Effect on Curing Time Effects on the Creep Response of
Fiber Reinforced Polymer Composites Bonded to Concrete ....................................

63

Y. Jeong, A. Jaipuriar, M.M. Lopez and C.E. Bakis

Bond Behaviour of GFRP Bars to Concrete Subjected to Fatigue
Loading and Freeze-Thaw Cycles ...............................................................................

71

J.R. Alves, A.A. El-Ragaby and E. El-Salakawy

Durability Characteristics of New GFRP Dowels for Concrete Pavement .............
M. Montaigu, M. Robert and B. Benmokrane

81

Evaluation of the Durability of CFRP Sheets after Climatic Exposure ..................

91

I. Nishizaki, P. Labossiere, K.W. Neale, M. Demers and T. Tomiyama

Durability of Externally Bonded FRP Composites: Comparison of
Accelerated Test Methods with Real Time Exposure .............................................. 101
H.R. Hamilton, C.W. Dolan and J.E. Tanner
Tensile Capacity of Stressed CFRP Strand Exposed to Extreme
Aggressive Groundwater Environments ................................................................... 109
M. Sentry, R. Al-Mahaidi, A. Bouazza and L. Carrigan
The Effect of Long-Term Loading on the FRP-Concrete Bonding
Strength ........................................................................................................................ 119
C. Mazzotti and M. Savoia
Design Optimization of Composite Structures Considering Bonded Joins
Durability ...................................................................................................................... 129
P. Hamelin, A. Si Larbi and E. Ferrier
Durability of Glass FRP Bars Subjected to Extreme Environmental and
Loading Conditions ...................................................................................................... 139
M. Robert, A. Fam and B. Benmokrane
Fire Testing of FRP Strengthened Reinforced Concrete Columns.......................... 151
N. Bnichou, D. Cree, E.U. Chowdhury, M.F. Green and L.A. Bisby
An Investigation into the Sustainability of FRP Reinforcing Bars .......................... 159
M.C.K. Pearson, T. Donchev and M. Limbachiya
Methods for Evaluating Long-Term Performance of FRP-Structural
Systems Applied to Concrete Bridges ......................................................................... 167
K. Crawford
Creep Behaviour of GFRP Cables, Short and Long Term Test ............................. 175
E. Ferrier, A. Si-Larbi and P. Hamelin
Durability of GFRP bars under Low Temperature Environment and
High Temperature Alkali Exposure .......................................................................... 185
S. A. Sheikh, D.T.C. Johnson and A.D. Caspary
Durability of FRP-Concrete Interface under Fatigue Loading ............................... 195
C. Carloni1, K.V. Subramaniam, M. Savoia and C. Mazzotti

ii

Phenomenological and Experimental Study of Pultruded UD FRP

Composites Static and Deferred Rupture under Combined Flexural,
Compressive and Torsional Loadings ........................................................................ 203
N. Kotelnikova-Weiler and J.-F. Caron
Fatigue Performance CFRP Post-Tensioned Tendons ............................................. 213
A. El Refai, J. West and K. Soudki
Transverse Shear of GFRP Rods: Test Method Development and
Potential for Durability Assessment ........................................................................... 221
T.R. Gentry
Fatigue Life Behavior of Adhesively Bonded Pultruded FRP Joints under
Tensile and Compressive Loading .............................................................................. 229
R. Sarfaraz, A.P. Vassilopoulos and T. Keller
Fatigue Performance of Composite Sandwich Panels with and without
Internal Ribs ................................................................................................................. 237
H. Mathieson and A. Fam
HEMP Fiber Composites for Structural Retrofit ...................................................... 245
D. Asprone, M. Durante and A. Prota
Efficiency of Different FRP-Based Flexural Strengthening Techniques in
Beams Submitted to Fatigue Loading......................................................................... 253
J. Sena-Cruz, J. Barros, M. Coelho and L. Silva
A Reactive Multiphase Multicomponent Approach for Predicting the
Performance of FRP Composites in Concrete ........................................................... 263
M. Boulfiza
Validation of Digital Image Correlation Technique on CFRP and SFRP
wrapped Cylinders Subjected to Freeze-Thaw Environmental Exposure .............. 275
K. Abdelrahman and R. El-Hacha
Experimental Study of the Influence of Matrix System on Pultruded
GFRP Composites' Creep Behaviour under Flexural Loading .............................. 285
S. Barboura, N. Kotelnikova-Weiler, J.-F. Caron and O. Baverel
CFRP Confinement of Non-Seismic Columns under Explosive Loading ............... 299
A. Lloyd, M. Saatcioglu and T.K. Tikka

iii

Performance of Concrete Bridge Barriers Reinforced with Conditioned

GFRP Bars in Alkaline Solution ................................................................................. 309
E.A. Ahmed and B. Benmokrane
Thermal Endurance and Fatigue Performance of Prestressed Concrete
Prism Reinforcement in Civil Infrastructure ............................................................ 319
H.M. Vogel and D. Svecova
FRP Externally Bonded Systems for a Sustainable Strengthening of
Masonry Structures ...................................................................................................... 327
F. Ceroni, B. Ferracuti, M. Pecce and M. Savoia
Bridge Deck-Guardrail Anchorage Detailing for Sustainable
Construction .................................................................................................................. 337
K. Sennah, N. Nikravan, J. Louie, A. Hassan, N. Al-Bayati, M. El-Sayed and
M. Sayed-Ahmed
More than 10 Years Successful Field Applications of FRP bars in
Canada .......................................................................................................................... 345
B. Drouin, G. Latour and H.M. Mohamed
Mechanical Recovery of Epoxy Adhesives Subsequent to Exceeding Glass
Transition Temperature .............................................................................................. 357
O. Moussa, A. Vassilopoulous, J. De Castro and T. Keller
Use of Silane Adhesion Promoter to Enhance FRP-to-Steel Bond
Performance .................................................................................................................. 365
J. Sizemore, J. Aidoo, K. A. Harries and J. Monnell
Structural Health Monitoring of Corroded Steel Reinforcement in FRP
Repaired Beams Using Multiplexed Fibre Bragg Grating Sensors ........................ 375
R. Al-Hammoud, K. Soudki and T. Topper
Short Term and Long Term Properties of Newly Developped Bent GFRP
Reinforcing Bars ........................................................................................................... 383
A. Weber and C. Witt
Sustainability Problems Associated with the Heat and Fire Protection of
Fiber-Reinforced Polymers (FRP) ............................................................................. 391
H. Ibrahim and I. Mahfouz

iv

Effects of Aging on Glass Transition and Creep Behavior of Epoxy used

in FRP Strengthening ................................................................................................... 401
A. Jaipuriar, J. Flood, Y. Jeong, C.E. Bakis and M.M. Lopez
More than Twenty Years of Field Applications of Composites with
Durability Testing and Field Observations ............................................................... 409
E.R. Fyfe, R. Watson and M. McCullagh
Vehicle Crash Testing of a GFRP-Reinforced PL-3 Concrete Bridge
Barrier ........................................................................................................................... 417
K. Sennah, B. Juette, A. Weber and C. Witt
Experimental Setup and Thermal Insulation Assessment of Sandwich
Concrete Walls with GFRP Connectors for Sustainable Construction................... 425
G. Woltman, M. Hanna, D. G. Tomlinson and A. Fam
FRP Composites for Strengthening Bridges in Victoria ........................................... 433
Y.C. Koay
Ageing Effects on the Debonding Behavior of FRP Reinforcements:
Experiments and Damage Modeling ........................................................................... 443
K. Benzarti, F. Freddi, M. Quiertant and S. Chataigner
Numerical Simulation of Thermal Deformations in FRP-Reinforced
Concrete Members in Hot Region............................................................................... 453
A. Zaidi, R. Masmoudi and M. Bouhicha
Effects of the Inter-phase on the Properties of Glass Fibre Reinforced
Polymer (GFRP) Composites ..................................................................................... 461
A.S.M. Kamal, M. Boulfiza and S. Panigrahi
Sustainable Structures Using Hybrid-Composite Beams ........................................ 469
J. Hillman
Considerations for Design of FRP Retrofitted Reinforced Concrete Wall
Panels Subjected to Blast Loading .............................................................................. 477
E. Jacques and M. Saatcioglu
Mechanical Properties and Durability of Sustainable FRP Composites ................ 485
P.V. Vijay, H.V.S. GangaRao and M.P.K. Reddy

Durability Study of Basalt Fiber Reinforced Epoxy Wet Layups Used in

Rehabilitation ................................................................................................................ 493
G. Xian, C. Wang, B. Xiao and H. Li
Comparative Durability Analysis of CFRP Strengthened RC Highway
Bridges .......................................................................................................................... 501
O. Ali, D. Bigaud and E. Ferrier
Axial Behaviour of Concrete Filled FRP Tube Columns after 300-FreezeThaw Cycles .................................................................................................................. 513
H. El-Zefzafy, H.M. Mohamed and R. Masmoudi
CFRP-Strengthened Beams under Repeated Impact Loading ............................... 523
A. Parvin and T.A. Mohammed
Bond Stress-Slip Model for FRP Rebars in Concrete ............................................... 531
S. Quayyum and A. Rteil
A Case Study on the Use of Advanced Fiber Wrap Composites for
Reinforced Concrete Repair of Port Terminal Wharf Caissons .............................. 539
J.C. Percival and T.T. Jimenez
Probabilistic Development of a Life Cycle Inventory (LCI) Dataset for
Pultruded Fiber Reinforced Polymer (FRP) Composites ......................................... 547
S.M. Ali, M.D. Lepech and J.P. Basbagill
Study on Moisture Fickian Diffusion Process of a Pultruded FRP
Composite Material ..................................................................................................... 555
X. Jiang, H. Kolstein, and F.S.K. Bijlaard
Durability of FRP-Strengthened RC Members under Various Loadings ............. 563
A. Parvin and A. Kulikowski
Fatigue Performance of Hybrid SFRP-GFRP-UHPC Beams ................................. 575
D. Chen and R. El-Hacha
SDOF Analysis of Protective Hardening Design for Reinforced Concrete
Columns using a Fiber Reinforced Polymer Wrap ................................................... 585
J. Quek and M.C. Ow
Effect of Mechanical Load on the Thermal Behavior of Concrete Slab
Reinforced with Fibre Reinforced Polymer (FRP) Bars .......................................... 593
H. Bellakehal, R. Masmoudi, A. Zaidi and M. Bouhicha
vi

Proposed Shear Model for Reinforced Concrete Beams Wrapped with

FRP: A Genetic Algorithm Approach ........................................................................ 603
C. Marshall, A. Rteil and M.S. Alam
Long-Term Durability of FRP Reinforcements Damaged by High
Temperature ................................................................................................................ 611
D.Y. Moon, H. Oh and J. Sim
Flexural Behavior of Simple and Fixed-end Beams Strengthened with
FRP bars in NSM Method .......................................................................................... 619
M.K. Sharbatdar, M. Mohamadian and S.M. Jaberi
Fatigue Performance RC Beams Strengthened with Prestressed NSMStrips and Rods ............................................................................................................ 627
F. Oudah and R. El-Hacha
Concrete Demolition of Bridge Barriers and Deck Slabs Reinforced with
GFRP Bars ................................................................................................................... 637
E.A. Ahmed, C. Dulude, S. Goulet and B. Benmokrane
Effect of Freeze-Thaw Cycles on the Compression Strength of Concrete
Filled FRP Tube Cylinders .......................................................................................... 649
H. El-Zefzafy, H.M. Mohamed and R. Masmoudi
Effects of Harsh Laboratory and Field Environmental Conditions on the
Durability of Concrete Wrapped GFRP Bars .......................................................... 659
T.H. Almusallam, Y.A. Al-Salloum, S. Alsayed, S.E. El-Gamal and M. Aqel
Durability of Concrete-Filled FRP Tubes Subjected to Salt Solution ..................... 667
M. Robert, A. Fam and B. Benmokrane
Freeze-Thaw Cycling Effect on the Mechanical Properties of FRPFilament-Wound Tubes .............................................................................................. 677
H. El-Zefzafy, H.M. Mohamed and R. Masmoudi

vii

viii

CDCC-11

DURABILITY AND SUSTAINABILITY OF FIBRE REINFORCED POLYMER

(FRP) COMPOSITES FOR CONSTRUCTION AND REHABILITATION
DURABILIT ET CYCLE DE VIE DES MATRIAUX COMPOSITES
EN POLYMRES RENFORCS DE FIBRES (PRF) EN
CONSTRUCTION ET RHABILITATION

ALTERNATIVE ENVIRONMENTAL KNOCKDOWN FACTORS

(CE) DERIVED FROM EXTENSIVE EXPERIMENTAL DATA
K.A. Harries 1 and B.M. Shahrooz 2
1
2

Associate Professor, Dept of Civil and Env. Engrg, Univ. of Pittsburgh, Pittsburgh PA, USA
Professor, School of Advanced Structures, Univ. of Cincinnati, Cincinnati OH, USA

ABSTRACT
An extensive study investigating the performance of fiber reinforced polymer (FRP)
materials and FRP bond to concrete encompassing 64 permutations of FRP material, test
method, and environmental conditioning was conducted. Based on the findings of this
study, alternative environmental exposure factors - the so-called CE knockdown factors are proposed. Factors associated with material properties remain dependent on the FRP
material with values of 0.90 and 0.80 proposed for CFRP and GFRP, respectively. It is also
proposed that these factors be applied to both FRP strength and modulus values only. Strain
capacity will, therefore, remain unchanged. In addition, factors to be applied to bond
capacity are proposed; these are dependent on expected material quality and are proposed
as 0.90 for preformed CFRP and 0.50 for fabric systems regardless of material.
1. INTRODUCTION
When designing with FRP materials, it is important to distinguish between material
resistance factors, load reduction factors, and environmental reduction factors. This paper
discusses only the latter so-called knockdown factors. These factors are given the notation
CE in ACI 440.2R [1]; this notation is adopted here. The ACI-prescribed values of CE are
given in Table 1. CE factors are not intended to capture effects of sustained such as creep
and fatigue; these effects are typically addressed using FRP strain limits, which are beyond
the scope of the present study.
Conventionally, the CE factor is applied to FRP strength (i.e., CEFu) and strain capacity
(CEu) but not modulus [1]. The results of the present study call into question this
convention, and demonstrate that strength and modulus are affected while strain capacity is
not significantly affected by environmental conditioning. Additionally, the CE factor is
typically only applied to material properties and not when determining bond capacity. This
study shows that environmental exposure has a greater effect on bond behaviour than on
FRP strength or modulus. Finally, this study also indicates a difference in performance
between preformed and hand laid-up FRP with the preformed FRP demonstrating superior
1

durability. This difference between manufactured and hand lay-up materials reflects issues
of quality control and should not be surprising. Nonetheless, existing guidelines [1] do not
acknowledge this difference, effectively penalizing the use of preformed materials.
Table 1. CE values prescribed by ACI 440.2R-08 [1].
Exposure
CFRP Plate CFRP Fabric
Interior Exposure
0.95
Exterior Exposure
0.85
Aggressive Environment Exposure
0.85

GFRP Fabric
0.75
0.65
0.50

The objective of this work is to establish a basis for recommending environmental exposure
factors suitable for design of externally bonded FRP materials. This goal is achieved based
on the results of a database having 64 permutations of FRP material, test method, and
environmental conditioning [2,3] described in the following sections.
2. EXPERIMENTAL PROGRAM
An extensive experimental program investigating the behaviour of three FRP systems
subjected to nine different environmental conditioning protocols was conducted [2,3]. The
effect of environmental conditioning was assessed using four different standard test
methods. Table 2 summarizes the test matrix for this program.
2.1 Environmental Conditioning Considered
Eight different environmental conditioning protocols, in addition to control specimens,
were considered. The ambient control conditions averaged 22oC and 70% RH and provide
the baseline against which results from conditioned tests are compared. The following
environmental conditioning protocols were considered in this study; additional details are
found in references [2] and [3].
Water Exposure: Test specimens were exposed to conditions of 100% humidity at 38oC for
durations of 1000, 3000, and 10,000 hours in accordance with ASTM D2247.
Salt Water Exposure: Test specimens were immersed in substitute ocean water, prepared
according to ASTM D1141, at 22oC for durations of 1000, 3000, and 10,000 hours.
Alkaline Exposure: Test specimens were immersed in a saturated solution of calcium
carbonate (CaCO3) at 22oC for durations of 1000, 3000, and 10,000 hours. During
exposure, the solution was maintained at a pH of 9.5.
Dry Heat Exposure: Test specimens were exposed at 60oC in a forced-draft circulation-air
furnace in accordance with ASTM D3045 for durations of 1000 and 3000 hours.
Diesel Fuel Exposure: Test specimens were immersed in diesel fuel at 22oC for 4 hours, in
accordance with ASTM C581.

Exposure
Baseline
Water
Salt Water
Alkaline
Dry Heat
Diesel
Weathering
Freeze-Heat

Table 2. Durability test program [2,3].

Duration
Description
laboratory
1000, 3000 &
100% RH @ 38oC
10000h
1000, 3000 &
Immersion in salt water @ 22oC
10000h
1000, 3000 &
Immersion in pH 9.5 CaCO3
10000h
solution @ 22oC
1000 & 3000h
60oC
4h
Immersion in diesel fuel @ 220C
1000 cycles
2h of UV @ 63oC
2h @ 100% RH
(4000h)
20 cycles
9h @ -18oC
15h @ 100% RH @ 38oC
(480h)

Baseline
Freeze-Thaw

laboratory
360 cycles
(1583h)

70min @ -18oC
70 min @ 4.5oC and UV

Tests Conducted
3 test methods:
ASTM D3039
tension,
ASTM D2344
SBS & bond test
4 specimens of:
CFRP plate,
CFRP fabric &
GFRP fabric
2 beam flexure
tests of: unretrofit
control, CFRP
plate, CFRP fabric
& GFRP fabric

Weathering Exposure: Specimens were exposed to UV light from a carbon arc source at a
temperature of 63C for 2 hours (ASTM G23) followed by exposure to conditions of 100%
humidity at 22oC for 2 hours. This procedure was repeated for 1000 cycles.
Freeze-Heat Exposure: Test specimens were initially placed in conditions of 100%
humidity at 38oC for 500 hours to allow for moisture absorption prior to freezing. Upon
removal from the humidity chamber, specimens were air dried for 48 hours and then placed
in a freezer at -18oC for 9 hours. Following this step, the specimens were placed back into
the humidity chamber at 100% humidity and 38oC for 15 hours. This freeze-thaw cycling
was repeated a total of 20 cycles.
Freeze-Thaw Exposure: Only reinforced concrete flexural beam specimens having bonded
FRP were subjected to this exposure. Specimens were subject to freeze-thaw conditioning
based on the ASTM C666 protocol. The test protocol extended for 360 cycles, each
consisting of the following four steps: 1) 70 minutes at -18C at 30% RH; 2) 20-minute
ramp up to 4.5C (resulting in 90% RH); 3) 70 minutes at 4.5 C at 50% RH with UV lights
on; and 4) UV lights off and 80-minute ramp down to -18C (resulting in 40% RH).
2.2 Test Methods
Four test methods were used to assess the effects of environmental conditioning. The test
methods and specimen preparation are described as follows; additional details are found in
references [2] and [3].

Tension Test: Tension capacity parallel to the fiber was determined using standard ASTM
D3039 coupons and test method. Baseline and conditioned specimens were cut using a
water-cooled diamond blade into coupons 254 mm long by 25.4 mm wide having 50 mm
long fiberglass gripping tabs at each end. Tension tests on the baseline specimens also
provided fundamental material properties reported in Table 3.
Short Beam Shear Test: The short beam shear (SBS) test provides interlaminar shear
strength (ILSS). This property is not applicable to single ply fabric specimens; therefore,
SBS tests were only conducted on the preformed CFRP plate material. The test was carried
out according to the method of ASTM D2344. Specimens were cut using a water-cooled
diamond blade into coupons 19 mm long by 6.4 mm wide. The specimens were loaded in
three-point flexure over a span of 12.7 mm.
Bond to Concrete: The bond test specimen consisted of two 51 mm concrete cubes spaced
25 mm apart bonded together using 19 mm wide by 127 mm long FRP strips on opposing
faces. The bonded length of FRP on both cubes was 38 mm. The average 56-day
compressive strength for all 192 cubes was 30 MPa. Prior to FRP application, the concrete
was air-dried for 96 hours and a wire brush was used to remove all laitance from the faces
to be bonded. The CFRP plate was bonded using a compatible structural epoxy adhesive.
The fabric systems were bonded during the layup process using the saturating resin.
Custom-made U-shaped collars were used to grip opposing cubes in a universal test
machine and tension was applied in displacement control at a rate of 2.2 mm/min.
Beam Flexure: Sixteen concrete beams 154 mm deep by 203 mm wide by 2440 mm long
reinforced internally with two #3 (9.5 mm) bars, top and bottom, and U shaped W2.9 (4.9
mm) deformed wire stirrups at 152 mm centers were used. The concrete compressive
strength determined at the age of testing was 38 MPa. Beam surfaces to receive FRP were
sandblasted to remove laitance. The CFRP plate was bonded using a compatible structural
epoxy adhesive. The fabric systems were bonded during the layup process using the
saturating resin. A single 2130 mm long layer of FRP was applied to each beam: for the
CFRP plate, a 102 mm wide plate was bonded to the beam soffit; the CFRP fabric repair
was 152 mm wide and the GFRP fabric was 305 mm wide resulting in it extending 51 mm
up each side of the beam. The FRP details are not comparable in terms of strengthening
effect (nor were they intended to be); the objective of this study was only to draw
comparisons between similar beams subjected to different environmental conditioning.
Beams were tested in three-point flexure over a span of 2288 mm.
2.3 FRP Material Systems
Three commercially available FRP systems were included in this study: 1) a preformed
unidirectional CFRP plate; 2) a unidirectional carbon fiber fabric; and 3) a unidirectional
glass fiber fabric. The carbon and glass fabrics were hand laid-up into their CFRP and
GFRP forms using a compatible two-part saturating resin. When bonded to concrete, a
layer of resin was used to prime the concrete surface prior to adding the saturated fabric
sheets. The CFRP plate was bonded to the concrete using a compatible two-part structural
epoxy. Experimentally determined [3] material properties of all three systems are given in
Table 3. Throughout this study, FRP modulus and strength are given in terms of force per
4

unit width of FRP (N/mm). This approach is consistent with that now promulgated by
ASTM D7565 and normalizes for FRP thickness which varies for a hand laid-up fabric
system. In essence, the values reported are E x t and Fu x t, respectively.
Table 3. Material properties of FRP system components.
ASTM
CFRP
CFRP
GFRP
property
units
test
plate
fabric
fabric
material thickness
mm
1.14
varies with layup
D3039 N/mm
tensile strength, Fu x t
2550
650
330
D3039 kN/mm
modulus of elasticity, E x t
142
58
18
D3039
%
1.88
1.06
1.87
ultimate strain, u
o
-6 o
CTE @ 20 C
E831
10 / C
1.48
3.94
13.77
o
glass transition temperature, Tg D4065
C
150
52
45
3. RESULTS OF EXPERIMENTAL PROGRAM
A summary of material properties and performance of all specimens subject to all
environmental conditioning is provided in Table 4. Reported values are normalized by
those obtained from the baseline tests. Detailed results are presented in [3].
Table 4. Average measured properties normalized to those obtained for baseline exposure.

Freeze-Thawe

Freeze-Heat

Weathering

Diesel

Heat 3000h

Heat 1000h

Alkaline 10000h

Alkaline 3000h

Alkaline 1000h

Salt Water 10000h

Salt Water 3000h

Salt Water 1000h

Water 10000h

Water 3000h

Water 1000h

Exposure

GFRP fabric CFRP fabric

CFRP Plate

E x ta 0.96 0.97 1.06 1.01 1.00 1.03 1.02 0.96 1.03 1.01 1.05 1.02 0.99 0.99
a
Fu x t 1.05 1.09 1.15 1.02 1.07 1.12 1.02 1.04 1.13 1.06 1.06 1.03 1.02 1.03
ua
1.02 1.04 1.09 0.84 1.02 1.07 0.97 0.95 1.02 1.01 1.00 0.96 1.02 0.97
FSBSb
1.01 1.02 0.98 1.05 1.03 0.98 1.01 1.00 1.01 1.04 1.07 1.03 0.94 0.97
bondc 1.07 1.14 1.18 1.17 1.30 0.94 1.16 1.18 0.98 0.91 0.98 1.00
1.15
flexd
0.94
E x ta 1.12 0.98 1.05 1.10 0.98 1.14 1.09 1.02 0.90 0.98 1.10 1.00 1.03 1.12
Fu x ta 1.26 0.92 1.08 1.22 0.97 0.97 1.22 0.97 0.97 1.17 1.18 0.92 1.20 1.14
ua
1.14 1.01 1.13 1.22 0.99 0.84 1.17 1.06 1.09 1.06 1.16 1.13 1.30 1.19
bondc 0.91 1.26 1.16 1.29 1.29 1.13 1.13 1.40 1.44 0.61 0.60 1.00
1.28
flexd
1.02
E x ta 1.00 1.00 1.00 1.06 0.94 0.94 1.00 1.00 1.00 1.06 1.00 1.00 0.94 1.00
Fu x ta 0.91 0.82 0.97 1.06 0.88 0.94 0.85 0.85 0.91 0.88 0.94 0.94 0.88 1.00
ua
0.94 0.83 1.01 0.96 0.89 1.03 0.81 0.96 1.07 0.79 0.87 1.01 0.90 0.96
bondc 1.09 0.87 1.01 1.10 1.11 0.87 0.90 0.83 0.52 1.05 1.01 1.06
1.15
flexd
0.97
lowest observed values are presented in bold text; all values are average of four tests except freeze-thaw (2 tests)
a
obtained from D3039 tension test; b obtained from D2343 short beam shear test; c obtained from bond test
d
obtained from flexural beam test; e unretrofit control beams = 0.92

3.1 CFRP Plate

The preformed CFRP plate material was relatively unaffected by the environmental
conditioning conducted. As a performed product, the CFRP plate system has better
consistency and quality and has cured since its manufacture. Additionally, a resin-rich outer
layer protects the underlying carbon tows from exposure. Of note is the consistent
improvement in modulus and strength with increased exposure to water at 38oC. This
observation is likely indicative of the elevated temperature resulting in additional post-cure.
Due to the relatively low CTE of the CFRP plate, few potentially damage-causing residual
stresses are likely to develop due to elevated post-cure temperature; therefore, no
detrimental effects are observed. These results, and the relatively inert behaviour when
exposed to salt water, demonstrate that the resin-rich outer layer of the CFRP plate inhibits
water absorption and is resistant to combined exposure (weathering). In no event does the
modulus or strength of the CFRP plate fall below 95% of the baseline value.
3.2 CFRP Fabric
The hand-laid up CFRP fabric also performed relatively well under the environmental
conditioning conducted. The greater variation compared to the CFRP plate can be partially
attributed to variation introduced in a hand lay-up process. Some increase in system
performance is evident in the conditioning methods that involved elevated temperature; this
increase may be attributed the post-cure effects of elevated temperature. However, there is
also evidence of water absorption that negatively affects the system performance. For
instance, the tension properties of the CFRP fabric are improved following 1000h exposure
to 38oC water (attributed to post-cure); however, the performance degrades at longer
exposures indicating eventual absorption. In this case, the elevated post-cure temperature
may result in some microcracking associated with temperature-induced residual stress
allowing absorption to take place. The additional detrimental effects of salt are also
apparent in the behaviour of the CFRP fabric. Nonetheless, the tensile modulus and
strength do not fall below 90% of the baseline values under any of the exposures
considered.
3.3 GFRP Fabric
The GFRP fabric was less durable than the CFRP products. No post-cure resulting from
elevated temperature was evident, and the detrimental effects of absorption are readily
apparent. As may be expected, the GFRP was particularly affected by exposure to an
alkaline environment. Glass fiber in GFRP applications is known to degrade in an alkaline
environment; for this reason the integrity of the surrounding resin matrix is critical to good
performance. A relatively high CTE can lead to matrix cracking allowing absorption of
concrete pore water with its incumbent alkalinity. The GFRP also indicated deterioration
under both elevated temperature environments. This decline may be an indication that these
environments exposed the specimens to temperatures greater than their glass transition
temperature, Tg. If this is the case, deterioration is expected to be due to the phase change in
the polymer matrix. This effect was not seen in the CFRP fabric specimens that did have a
marginally greater value of Tg.

3.4 Glass Transition Temperature and Water Absorption

Single specimens of all three FRP materials from each conditioning regime were tested to
assess any shift in Tg [3]. In no case was a statistically significant change in Tg recorded.
However, in those specimens where absorption was suspected, distinctive peaks at 0o and
100oC were noted, corresponding to the freezing and boiling points (phase changes) of the
absorbed water. This confirmed the hypothesis that absorption was present in both FRP
fabrics although not in the CFRP plate material.
3.5 Bond to Concrete
Bond to concrete is an interface property and not itself affected by the deterioration of the
FRP material. Certainly, the FRP may deteriorate to a degree where failure occurs prior to
bond strength being developed but this decline does not imply bond failure per se. The
majority of the reported deterioration of bond capacity associated with environmental
conditioning was not bond related but simply reflects the deterioration and failure of the
FRP material. The results of the GFRP bond specimens subjected to alkaline environment,
for instance, exhibit the same degree of deterioration as the comparable tension tests. Some
adhesive failures were noted at the FRP or concrete interfaces of the bond line. These
failures may be attributed to environmental induced degradation of the adhesive layer.
Deteriorated bond capacity of the CFRP plate system subjected to heating conditions
(60oC) may be attributed to the elevated temperature approaching the Tg value of the
adhesive epoxy used. While this value was not determined in the present study, a typical
value would be on the order of 50oC.
3.6 In situ Performance Beam Flexure Tests
The degree of strengthening provided varied considerably. The CFRP plate, CFRP fabric
and GFRP fabric-retrofit beams had capacities 218%, 157% and 149% of their companion
unretrofit beam capacities, respectively. The effect of 360 freeze-thaw cycles on the
unretrofit concrete beam was to reduce its capacity approximately 8%. This reduction is
consistent with results expected from good quality concrete and indicates little damage to
the concrete resulting from the freeze-thaw conditioning. The reductions associated with
freeze-thaw exposure of the CFRP plate and GFRP fabric retrofit beams are only 6% and
4%, respectively. The beams having CFRP fabric subjected to freeze-thaw exposure had no
apparent strength loss. The reduced degree of environmental-induced degradation
associated with the presence of FRP may result from the FRP actually protecting the
concrete from freeze-thaw damage. With the FRP in place, a large region of soffit concrete
is protected from water ingress and, therefore, the effects of freeze-thaw cycles are
diminished.
Primary failure of the retrofit beams was associated with loss of the FRP strengthening
effect. An evaluation of failure modes in this regard can reveal information on the FRP
systems resistance to the combined freeze-thaw and UV exposure provided. The CFRP
plate system exhibited typical intermediate crack-induced (IC) debonding [4] in all cases.
This observation is consistent with the expected mode of flexural failure and was
7

apparently unaffected by the freeze-thaw conditioning. The behaviour of the CFRP fabricreinforced beams changed from one controlled by adhesive failure of the bond between
FRP and concrete to IC debonding for the conditioned beams. This shift is counterintuitive
and may reflect poor quality lay-up or surface preparation of the unconditioned specimens.
Nonetheless, the fact that the conditioned beams performed admirably is an indication that
little freeze-thaw induced deterioration occurred.
The GFRP fabric retrofit beams had such a small effective increase in reinforcement
attributable to the GFRP that the fabric ruptured prior to debonding. In all cases, the FRP
rupture resembled a brooming type of failure typical of tension specimens. Such a failure
is indicative of a good quality FRP having a sound matrix and good fiber continuity [5].
While the failures of the GFRP fabric were similar, the extent of the rupture (brooming)
area was reduced for the conditioned beam specimens indicating that there may be some
degradation of the GFRP system present.
4. RECOMMENDATIONS
Based on the extensive experimental program described, the following environmental
exposure factors for design of externally bonded FRP materials are proposed (Table 5).
Environmental exposure factors associated with material properties are dependent on the
FRP material with values of 0.90 and 0.80 proposed for CFRP and GFRP, respectively. It is
proposed that these factors be applied to both FRP strength (i.e., CEFu) and modulus (CEE)
values. Since the linearity of FRP behaviour is unaffected, strain capacity remains
unchanged since the combinatorial effects of reducing both the stress and modulus equally
result in unchanged strain (i.e., u = CEFu/CEE). This application to strength and modulus,
but not strain, is supported by the present study.
In addition, exposure factors to be applied to bond capacity are proposed (Table 5); these
factors are dependent on expected material quality and are proposed as 0.90 for preformed
CFRP and 0.50 for fabric systems regardless of material. Based on the conditioning
performed in this study, it is deemed that these proposed factors are appropriate for exterior
exposure. The factors can likely be relaxed for interior exposure while environmentspecific factors need to be developed for specified aggressive environments.
Table 5. Reduction factors for environmental exposure.
CFRP
CFRP
GFRP
Plate
Fabric
Fabric
Lowest values observed in this study
0.96
0.90
0.94
Tension Modulus, E x t
1.00
0.92
0.82
Tension Strength, Fu x t
Bond capacity
0.91
0.60
0.52
Recommendations
FRP Material Properties
0.90
0.80
Bond Capacity (to concrete substrate)
0.90
0.50

5. REFERENCES
1.
2.
3.
4.
5.

ACI Committee 440. 2008. Guide for the Design and Construction of Externally
Bonded FRP Systems for Strengthening Concrete Structures. ACI 440.2R-08.
American Concrete Institute, Farmington Hills, MI, 76 p.
Cromwell, J.R., Shahrooz, B.M. and Harries, K.A. 2011. Environmental Durability of
Externally Bonded FRP Materials Intended for Repair of Concrete Structures. ASCE
Journal of Composites for Construction (in review).
Pack, J.R. 2003. Environmental Durability Evaluation of Externally Bonded
Composites. MSCE thesis, Department of Civil and Environmental Engineering,
University of Cincinnati, 289 p.
Teng, J.G. Chen, G.F., Smith, S.T. and Lam, L. 2002. FRP-Strengthened RC
Structures. John Wiley & Sons, New York, 245 p.
Agarwal, B.D. and Broutman, L.J. 1990. Analysis and Performance of Fiber
Composites, 2nd edition. John Wiley & Sons, New York, 449 p.

10

CDCC-11

DURABILITY AND SUSTAINABILITY OF FIBRE REINFORCED POLYMER

(FRP) COMPOSITES FOR CONSTRUCTION AND REHABILITATION
DURABILIT ET CYCLE DE VIE DES MATRIAUX COMPOSITES
EN POLYMRES RENFORCS DE FIBRES (PRF) EN
CONSTRUCTION ET RHABILITATION

A FIELD STUDY ON DURABILITY OF BOND STRENGTH

BETWEEN CFRP WRAPS AND CONCRETE
S.A. Sheikh1, A.D. Caspary1 and M. Stafford2
1
2

Department of Civil Engineering, University of Toronto, Toronto, Canada

R. J. Watson, Inc., East Elmhurst, NY, U.S.A.

ABSTRACT
The use of composite materials, in particular fibre-reinforced polymers (FRP), for
infrastructure strengthening, retrofit and rehabilitation applications has been steadily
increasing in the field. The advantages of FRP as a retrofit material include among others
ease of application, high specific stiffness and strength ratios, and resistance to
electromagnetic corrosion.
The failure modes of externally wrapped FRP include tensile rupture, as well as delamination and/or peeling of FRP material due to shear distortions or cracking. The delamination and peeling modes are directly related to bond properties between the FRP and
concrete.
An investigation to determine the in-situ bond strength between FRP wraps and concrete
exposed to actual field conditions was performed by the University of Toronto. The paper
summarizes the direct tension pull-off tests performed on the two CFRP retrofitted
structures, one a highway bridge repaired in 1996 and the other a 14-storey residential
building retrofitted in 1999. Both structures are located in Toronto. Only the columns in the
bridge were repaired while in the building structures, the beams, columns and walls were
rehabilitated. The repair of both the structures was carried out using external wet-lay CFRP
wraps. The results indicated a variation of bond strength in the range of 2.26 MPa to 3.92
MPa. In a number of field specimens the failure was not achieved due to the limited
capacity of the equipment used. The failure modes included separation of pull-off plates
from the FRP wraps, bond failure at the interface between FRP and concrete, and tensile
failure of concrete.
1. INTRODUCTION
The use of composite materials, in particular fibre-reinforced polymers (FRP), for
infrastructure strengthening, retrofit and rehabilitation applications has been a widely
studied topic in recent years. The advantages of FRP as a retrofit material include high

11

specific stiffness and strength ratios, outstanding fatigue behaviour, resistance to

electromagnetic corrosion and ease of application.
Two of the most common failure modes in concrete structures externally wrapped with
FRP are material tensile rupture and delamination and/or peeling of FRP wrap. The
delamination and peeling modes are directly related to bond characteristics between the
FRP and concrete. Performance of retrofitted structures thus significantly depends on the
bond properties between the FRP sheets and the concrete. As such, the bond strength and
the methods used to determine it are critical to the understanding and evaluation of FRP
retrofitting techniques, as well as the FRP-concrete behaviour and failure mechanism.
While laboratory studies can provide useful information, they can rarely replicate the actual
conditions experienced by a structure in the field. Thus, there exists a critical need for data
from field applications and tests to determine reliable estimates of the bond strength
between concrete and FRP.
Two structures upgraded with FRP in 1990s were used to evaluate the FRP-concrete bond
properties after years of service in the field. The Leslie Street Bridge, built in the sixties, is
located in Toronto, Ontario. The highway has 7 westbound and 6 eastbound lanes at this
location. The columns are about 920 mm to 1010 mm in diameter. The specified concrete
strength was 20.7 MPa (3000 psi). The Amadiyya Abode of Peace, a 14-storey residential
building, is also located in Toronto, Ontario. The exterior west and east faces of the
building each contain 14 beam-column supports. The columns are approximately 830 mm
in diameter, and the beams are approximately 600 mm wide and 1200 mm deep.
2. REPAIR WITH FRP
After decades of overuse and neglect, our aging infrastructure is failing, particularly in
areas that are exposed to years of harsh weathering elements, resulting in corrosion of the
reinforcing steel embedded within concrete structures. Figures 1 and 2, respectively, show
examples of severely corroded bridge columns and exterior building columns before and
after the FRP retrofit (Sheikh et al. 2002, Sheikh 2007). The concrete cover is completely
lost in bridge columns while in the building structure, the damage due to steel corrosion is
still limited with concrete severely cracked. Neither of these conditions is uncommon.
Engineers have been searching for alternative materials to repair and prevent these types of
structural issues and in the past 15 years, fibre reinforced polymers (FRPs) have emerged as
innovative materials that can be useful in many infrastructure rehabilitation/ strengthening
projects.
One highly successful method for infrastructure rehabilitation involves wrapping concrete
columns and beams with FRPs. The confinement provided by FRP significantly increases
the strength and deformation capacity of columns. In the case of beams, the transverse fibre
wrap significantly increases the shear strength. This technique is now widely recognized as
an effective and efficient strengthening, repair, and rehabilitation method. Twenty eight
bents consisting of beam and columns were repaired on the west and east exterior faces of
the apartment building. The current conditions of the repaired columns in Figures 1 and 2,
after exposure to extreme environmental conditions in and around Toronto for up to 15
years, show remarkable performance of FRP repair. The repaired structures showed
12

excellent resistance to freeze-thaw cycles, UV radiation, low temperatures, temperature

cycling, chemicals (alkalis and chlorides), ice, rain, sustained load, impact and fatigue.

Before Repair in 1996

2010

Fig. 1. Severely corroded bridge columns before and after repair

Before Repair

Repair in 1999

2010

Fig. 2. Corroded and damaged building beams and columns and FRP repair
Although, FRP repair has been used for all types of strcutural components, its use in
columns has been uniquely advantageous. Closed FRP wrap provides shear and
confinement reinforcement but the bond between FRP and concrete is not critical in most
such applications. Test data from columns in the field retrofitted with FRP, however,
provide valuable information about the durability and longevity of FRP repair as far as the
bond between concrete and FRP is concerned.
3. PULL-OFF TESTS FOR FRP-CONCRETE BOND
As the first step in the preparation for pull-off tests, the protective paint and barriers are
carefully removed from the surface of FRP which has previously been bonded to the
concrete substrate. The FRP surface is then cleaned in accordance with the requirements of
the manufacturers of the epoxy used for adhering the metal disc to the FRP surface. FRP
areas equal to the size of the disc is cut through to the concrete layer with the help of a core
drill before the discs are attached. After the epoxy has sufficiently cured, a testing apparatus

13

(Figure 3) is used to pull off the discs and measure the value of tensile force required to debond the metal disc and all adhered material from the concrete surface. The disc diameter
of the commercially available system varies between 20 mm and 50 mm. The epoxy is
specified to have a tensile capacity of at least 5.0 MPa. The current test series at the bridge
was part of a program to evaluate FRP-concrete bond at four locations across Canada for
which it was decided to use 50 mm diameter discs for bridge tests (Banthia et al. 2010).
The larger disc was considered to provide a more representative value for bond strength.
For the building, two disc sizes (50 mm and 20 mm) were used to study the effect of disc
size on bond strength. The ideal failure mode is described as cohesive failure within the
concrete at a stress greater than 1.4 MPa, according to available codes and standards (ACI,
2004; ASTM, 1993; CSA, 2004).
Observations from this test include the force required to de-bond the specimen; the location
of the plane of failure and a qualitative assessment of the bond. The location of failure
plane is generally denoted as E (Adhesive epoxy failure premature failure), E/F
(Adhesive epoxy failure combined with cohesive FRP failure within the laminate
premature failure), F (Adhesive failure of FRP concrete interface, or total cohesive failure
within the FRP laminate), F/C % (FRP-Concrete interface failure given as the percentage of
test area retaining concrete/aggregate), and C (Cohesive failure within the concrete). See
Figure 3 for test set-up and location of potential failure planes.

Fig. 3. Pull-Off Test Set-up and Potential Failure Locations

In most applications, load is generally transferred between FRP and concrete through a
combination of shear and tension stresses, particularly in flexural members (Wan et al.,
2002) while the interface between FRP and concrete is subjected to direct tensile stress in
this test. Although, this test may not simulate stress conditions for flexural members, it
provides a useful indication for structural performance. This test is simple and a valuable
tool for investigating qualitative and relative values for different parameters.
3.1 Ahmadiyya Abode of Peace Building
One Beam and one Column (north and south Face) on the south end of the west side of the
structure were selected for tests. For the column tests, the locations of the discs were
selected randomly as the FRP was used to provide confinement. Since the beams were
retrofitted for shear, the discs were specifically placed near the neutral-axis (mid-height),
where shear forces would be maximum. Before the discs were adhered to the members, the
FRP surface was thoroughly cleaned and FRP areas equal to the size of the disc were cut
through to the concrete layer with the help of a core drill. Nine 50mm-disc and eight
20mm-disc tests were installed.

14

3.2 Leslie Street Bridge on Highway 401

Tests were carried out on two columns which were wrapped with CFRP to halt corrosion
due to excessive use of de-icing salts during the winter season. The retrofitted columns
divide the north bound lanes of Leslie Street from the south bound lanes and receive
extensive exposure to ice and salt as a result of heavy traffic (Figure 4). The locations for
discs were selected to represent this exposure. The core marking on the column were
made on two grids; one facing the traffic with a 30 degrees angle to the axis of the
roadandtheotherontheoppositesideofthecolumn. The FRP surface was prepared in
the same manner as described above. A total of 16 tests were completed.Figure 4 shows a
spot at the lower end of the column where FRP has been damaged. This was caused by a
direct hit by a vehicle. Obviously, the test locations were selected at a distance away from
this damaged area. Figure 5 shows two different variations of the commercially available
testing apparatus.

Fig. 4. Leslie Street Bridge, Toronto

Fig. 5. Testing Apparatus and Procedure

15

4. RESULTS
4.1 Observations and Conditions Status
The Leslie Street Bridge pull-off tests were performed in the second week of August in
2009. The temperature ranged between 12oC and 30oC and the relative humidity was 52%
to 99%. FRP surface at most locations on the columns appeared intact. The pull-off tests on
the apartment building beams and columns were carried out in the second week of
November 2009. The temperature during that time in Toronto varied between 2oC and 12oC
and the relative humidity varied between 69% and 82%. In this structrue, the FRP appeared
in excellent condition and much better than the FRP on bridge columns.
4.2 Pull-off Test Values
The maximum pull-off load value recorded during each test was used to calculate the pulloff bond strength by diving it by the disc area of 1964 mm2. Tables 1 and 2 show average
pull-off strength and stadard deviation for all the tests conducted at the Ahmadiyya
Building and Leslie Bridge, respectively. The capacity of the pull-off tester using 50 mm
discs was 3.5 MPa. In several tests, the discs could not be pulled off and the maximum
strength recorded was 3.5 MPa. The minimum pull-off strength observed in bridge columns
was 2.0 MPa. Since only 50 mm discs were used at the bridge, the pull-off strength for the
FRP is thus underestimated due to equipment limitation. Although the same limitation was
experienced for the tests on beams and columns in the building, the average strength from
50 mm and 20 mm discs are very similar. The maximum pull-off strength with 20 mm discs
was 3.92 MPa while the minimum recorded was 1.5 MPa. Figure 6 shows a typical failures
that can be classified as F/C (FRP-Concrete interface failure given as the percentage of test
area retaining concrete/aggregate) and C (Cohesive failure within the concrete). The failure
modes observed at the building include 5 F, 4 C, 4 F/C, and 2 E (premature). Two tests
were terminated due to the capcity limit of the tester. At the bridge the failure modes are as
follows: 4 F, 4 C and 2 F/C. Six tests were terminated due to the capcity limit of the tester.
It should be noted that the FRP repair of the bridge has endured the environmental exposure
for about 15 years under service conditions. The age of repair in the case of the building
structure is 12 years. The average bond strength ranging from 2.6 MPa and 3.1 MPa after
12 to 15 years of service in extreme enviornmental conditions indicates good performance
of FRP repair. Much lower pull-off strength has been reported in some bridges (Banthia
2010) but it was observed that factors such as placement characteristics, surface finish,
infiltration of water behind the FRP layer and other workmanship issues played important
roles in determining the long-term performance of the repair.
Table 1. Average Pull-Off Strength Values at Leslie Street Bridge
Average Pull-Off Strength Standard Deviation (SD)
MPa
MPa
Column 1
3.06
0.63
Column 2
3.14
0.57
All Tests
3.10
0.58

16

Table 2. Average Pull-Off Strength Values at Ahmadiyya Building

Average Pull-Off Strength
Standard Deviation (SD)
MPa
MPa
Column (50mm)
2.78
0.95
Beam (50mm)
2.74
1.26
Column/Beam (50mm)
2.76
1.03
Column (20mm)

2.45

1.00

Beam (20mm)
Column/Beam (20mm)
All Tests

2.72
2.63
2.71

1.27
1.09
1.02

Fig. 6. Typical De-bonded Specimen

5. CONCLUDING REMARKS
Pull-off tests were carried out on an apartment building columns and beams that were
repaired 12 years ago by wrapping them with CFRP and on bridge columns that were
repaired with CFRP 15 years ago. The average bond strength observed in the building
structure was 2.71 MPa for both columns and beams with a standard deviation of 1.02 MPa.
The average bond strength observed in the columns of the bridge was 3.10 MPa and the
standard deviation was 0.58 MPa. A minimum strength of 1.4 MPa for field tests is
considered to be adequate.
After 12 to 15 years of service in the Toronto area which is characterized by typical
Canadian weather and environmental conditions including extreme temperature variations,
freeze-thaw cycles and exposure to deicing salt, the bond between FRP and concrete was
found to be adequate.

17

6. ACKNOWLEDGMENTS
The authors would like to thank the Ministry of Transportation of Ontario for allowing
access to the bridge site and arranging traffic control and the management of Ahmadiyya
Abode of Peace for access to the bulding and providing amenities needed for field work.
Research support from ISIS Canada is also gratefully acknowledged.
7. REFERENCES
1.
2.
3.
4.
5.
6.
7.

ACI Committee 440. 2004. Guide Test Methods for Fiber-Reinforced Polymers (FRPs)
for Reinforcing or Strengthening Concrete Structures. ACI 440.3R-04, Farmington
Hills, MI, 40 p.
ASTM International Committee D-01. 1993. Standard Test Method for Pull-Off
Strength of Coatings Using Portable Adhesion Testers. ASTM D-4541, Pennsylvania.
Banthia, N., Demers, M., Mufti, A. and Sheikh, S. A. 2010. Durability of FRPConcrete Bond in FRP-Strengthened Bridges. ACI Concrete International, 32(8): 4551.
Canadian Standards Association (CSA). 2002. Design and Construction of Building
Components with Fibre-Reinforced Polymers. CSA S806-02, Mississauga, ON, 177 p.
Sheikh, S. 2007. Field and Laboratory Performance of Bridge Columns Repaired with
Wrapped GFRP Sheets. Canadian Journal of Civil Engineering, 34: 403-413.
Sheikh, S.A., DeRose, D. and Murdukhi, J. 2002. Retrofitting of Concrete Structures
with Fibre Reinforced Polymers. ACI Structural Journal, 99(4): 451-459.
Wan, B., Petrou, M.F., Harries, K.A., Sutton, M.A. and Yang, B. 2002. Experimental
Investigation of Bond between FRP and Concrete. The Third International Conference
on Composites in Infrastructure, San Fransisco, CA, June.

18

CDCC-11

DURABILITY AND SUSTAINABILITY OF FIBRE REINFORCED POLYMER

(FRP) COMPOSITES FOR CONSTRUCTION AND REHABILITATION
DURABILIT ET CYCLE DE VIE DES MATRIAUX COMPOSITES
EN POLYMRES RENFORCS DE FIBRES (PRF) EN
CONSTRUCTION ET RHABILITATION

BOND BEHAVIOR OF EXTERNALLY BONDED FIBER

REINFORCED POLYMER LAMINATES SUBJECTED TO IN-SITU
SERVICE LOADING AND ENVIRONMENTAL CONDITIONING
J.J. Myers and N. Muncy
Missouri University of Science and Technology, Rolla, Missouri, USA

ABSTRACT
The use of fiber reinforced polymers (FRP) in repairing and strengthening bridges has been
an active area of research and implementation in recent years. In particular, adhering FRP
to the tension face of reinforced concrete (RC) beams has provided an increase in load
carrying capacity and extended the service life of the structures. The key to the efficient
performance of FRP strengthened concrete members is the bond between FRP sheets and
the concrete to which they are bonded. This study examines the bond durability of FRP
laminates installed on bridge structures in the State of Missouri, USA after several years of
exposure to service conditions including both mechanical and environmental loading.
Since 1998, more than twenty-five (25) bridges have been built and/or repaired using FRP
composite materials in the State of Missouri. One project was selected for this bond
behaviour study investigation presented herein. The paper discusses both the pull-off
strength and failure mode(s) of the bond tests conducted on FRP laminates under field
conditions in both actively stressed and largely unstressed locations. In terms of
environmental behaviour, the investigation examines FRP laminates with and without a
protective ultraviolet (UV) coating to investigate any differential level of degradation due
to UV in locations of direct and intense UV exposure. The study also compares the bond
performance of laminates in locations that are subjected to repeated moisture including wet
and dry cycles/freezing and thawing cycles. Failure modes are reported as well as bond
strength levels.
1. INTRODUCTION
Building and maintaining infrastructure in the United States of America is essential to the
economic and well-being of Americans. In the future, increasing amounts of infrastructure
such as bridges will become deficient and require structural attention [1]. In 2003, the
Missouri Department of Transportation retrofitted the Dallas County Bridge P-0962 with
Fiber Reinforced Polymers (FRP). Carbon FRP laminate sheets were bonded externally to
the concrete surface using a high strength epoxy to help increase the shear and flexural

19

capacity of the bridge [2]. As the FRP strengthening system is exposed to various
weathering effects including ultraviolet light, freeze/thaw, and water exposure, the bond
strength of the FRP-concrete interface over time was of particular interest. A premature
debonding failure occurring between the FRP-concrete interfaces will greatly reduce the
capabilities of the FRP strengthening system.
2. LITERATURE REVIEW
FRP for repair and strengthening of reinforced concrete (RC) systems has been an active
area of research in recent years. FRP has been used to strengthen bridges and provide
several advantages over other strengthening procedures. Some advantages include minimal
traffic disturbance, short construction times, and lightweight materials allowing for easier
construction [2]. In the August 2010 issue of American Concrete Institute (ACI) Concrete
International Magazine, an analysis of FRP performance was investigated on four inservice bridges in Canada. The results from the carbon fiber reinforced polymers indicated
low bond strengths and variability among each test [3]. While the environmental conditions
in Canada may be more severe than those experienced in Missouri, the results from the
Dallas County Bridge will reflect the various environmental effects on FRP experienced in
Missouri.
3. METHODOLOGY
The DYNA Z Pull-Off Tester with digital Manometer (Figure 1) was used to test the
strength of the FRP-concrete interfaces. This was performed by applying discs to the
exposed surface of the FRP using Loctite Professional Heavy Duty Epoxy. The discs were
secured and left for approximately 24 hours to ensure the epoxy reached the proper
strength. Then, a diamond bit was used to core around the disc and into the concrete to a
target depth of 6.4-mm (0.25-in). The pull-off tester was then used to pull the discs off in
direct tension (Figure 1) and a peak value was recorded by the pull-off tester.

Fig.1. DYNA Z Pull-Off Tester being placed over the disc to record the pull-off strength

20

4. RESULTS
Each disc was applied at a specific location to determine the failure mode due to different
environmental conditions. Figure 2 identifies the disc numbers and Figure 3 shows the
failure modes. Table 1 shows the disc locations and the results of each pull-off test. The
concrete temperature for all pull-off tests was approximately 41 F.

Fig. 2. Disc numbers corresponding to the disc location

Fig. 3. Failure modes of the discs

Discs 1, 2, and 3 were applied to panel 1 of the abutment, which did not have any
sandblasting performed (i.e. no surface preparation) on the concrete surface prior to the
application of the FRP. As shown in Figure 3, discs 1 and 2 pulled concrete indicating that
the failure occurred in the concrete rather than at the interface between the FRP and

21

concrete surface. Disc 3 was partially applied to a delamination zone previously induced for
delamination growth studies so the disc 3 results were discarded.
Discs 4, 5, and 6 were applied to panel 3 of the abutment, which had a moderately light
sandblast concrete surface preparation prior to application of the FRP. As shown in Figure
3, all three discs pulled concrete indicating that the failure occurred in the concrete yielding
the most promising failure mode and not between the FRP and concrete surface.
Discs 7, 8, and 9 were applied to panel 6 of the abutment, which had a heavy sandblast
performed on the concrete surface prior to the application of the FRP. As shown in Figure
3, all three discs pulled concrete again indicating that the failure occurred in the concrete
yielding the most promising failure mode and not between the FRP and concrete surface.
Discs 10, 11, and 12 were applied to the lower part of a column where direct exposure to
moisture in the form of water (both liquid and frozen) occurred. As shown in Figure 3, all
three discs pulled concrete but exposed some FRP causing the failure mode to occur
between the FRP-concrete interfaces.
Table 1. Pull-Off Test Results
Standard Deviation
Location
Disc Load (lbs) Stress (psi)
(psi)
Panel 1
1
2683
847
338
Panel 1
2
1170
370
Panel 1
3
1822
575
Panel 3
4
2607
823
141
Panel 3
5
2095
662
Panel 3
6
1717
542
Panel 6
7
943
298
82
Panel 6
8
1356
428
Panel 6
9
1426
450
Column
10
355
112
72
Column
11
792
250
Column
12
687
217
UV Protected
13
920
291
85
UV Protected
14
1187
375
UV Protected
15
646
204
UV Unprotected
16
1892
598
5
UV Unprotected
17
1915
605
UV Unprotected
18
2037
643
Slab
19
472
149
26
Slab
20
587
185
Conversion: 1 psi = 0.00689 MPa

22

Average
(psi)
608

676

392

193

290

601
167

Discs 13, 14, and 15 were applied to the pier bent that had an ultraviolet (UV) protection
applied over the FRP. The discs were pulled and the sub-surface was found to be saturated.
Figure 3 shows that the failure occurred in both the concrete and at the FRP-concrete
interface.
Discs 16, 17, and 18 were applied to the pier bent that did not have UV protection applied
over the FRP. Figure 3 shows that the failure mode was primarily between the FRPconcrete interface. Disc 18 was not cored correctly so the results for disc 18 were
discarded.
Discs 19 and 20 (Not Pictured) were applied underneath the slab of the bridge deck where
the FRP was subjected to the most frequent stress variations. The discs were pulled and the
failure mode occurred in the concrete. The concrete was extremely saturated causing the
discs to be removed under lower loads. This suggests that moisture was trapped between
the outermost layer of concrete and FRP.
It may be noted that two locations fell below the 1.38 MPa (200 psi) minimum bond
strength limit presented in ACI 440 design guideline, namely the column base and
underside of the slab. Both of these locations had evidence of moisture.
The overall bond strength results are generally higher than those recorded on the previous
study undertaken in Canada by Banthia et al. [3]. These variations could be due to the
varied exposure conditions including lower temperatures experienced in Canada and/or
variations in the strengthening system or its application. Figure 4 shows the mean
temperatures between Kansas City, Missouri and Quebec Canada as a point of reference
[4,5]. The average monthly temperature is 2.8 C to 8.3 C (5 F to 15 F) lower in Canada
depending on the month. The colder weather could create longer exposure to freezing
conditions resulting in deterioration or freeze/thaw damage of the FRP.

Temperature(F)

MeanTemperatures
90
80
70
60
50
40
30
20
10
0

KansasCity
Quebec

Jan Feb Mar Apr May Jun Jul Aug Sept Oct Nov Dec
Month

C = (F - 32) * 5/9
Fig. 4. Mean temperatures of Kansas City, Missouri and Quebec, Canada
23

5. CONCLUSION
The results showed that the failure mode occurred within the concrete based on the pull-off
tests for discs 1-9 (excluding disc 3). The results from this work indicated that roughness
did not have a long-term effect on the failure mode, but the roughness did play a role on
FRP-concrete interface bond strength. The highest level of roughness preparation did not
produce the optimal long-term bond strength.
When the FRP was exposed to water either constantly or intermittently such as at the base
of a column (even under low levels of stress), the failure mode occurred between the FRPconcrete interface. Discs 10, 11, and 12 clearly show that some concrete and FRP are
exposed causing the FRP-concrete mode of failure. This location also recorded the lowest
bond strength level. These results suggest that the bond strength of the FRP system is not
nearly as effective over time when exposed to moisture on a regular basis.
When direct daylight was exposed to the UV protected FRP, the failure mode occurred in
the concrete. These discs pulled low compared to the rest of the results due to the concrete
being saturated at the test region. Although the average strength was around 2.00 MPa (290
psi), more pull-off tests should be performed on the UV protected FRP when excessive
moisture is not present in the concrete. When the FRP was not UV protected, the failure
mode occurred between the FRP-concrete interfaces. Although the UV unprotected FRP
could withstand a greater load, the failure mode indicates that the UV protection had a more
desirable failure mode compared to the UV unprotected laminate area.
The discs placed on the underside of the slab recorded the lowest bond strength. The
concrete appeared to be saturated at the time of testing. The failure mode was in the
concrete indicating a desirable failure mode. However, the average bond strength was only
1.15 MPa (167 psi). More pull-off tests should be performed when the concrete is not
saturated to see if the moisture had an impact on the test results.
Based upon the pull-off tests, there is variability between the three tests in each location.
The standard deviation ranged from a high of 2.33 MPa (338 psi) to a low of 0.34 MPa (5
psi) throughout testing highlighting the variability in the test method. Additional bond
testing of more than ten other bridges in Missouri are planned during this calendar year to
increase the data pool on the long-term bond performance of externally bonded FRP
systems.
6. REFERENCES
1.
2.

Missouri S&T. 2010a. Preservation of Missouri Transportation Infrastructure: NonDestructive Testing of FRP Laminates. Missouri S&T, University Transportation
Center Report, http://utc.mst.edu/research/r096.html, December.
Missouri S&T. 2010b. Preservation of Missouri Transportation Infrastructure:
Validation of FRP Composite Technology through Field Testing-In-situ load testing on
Bridges P-962, T-530, X-495, X596, and Y-298. Missouri S&T, University
Transportation Center, http://utc.mst.edu/research/r095.html, December.

24

3.
4.

5.

Banthia, N., Demers, M., Mufti, A. and Sheikh, S. A. 2010. Durability of FRPConcrete Bond in FRP-Strengthened Bridges. ACI Concrete International, 32(8): 4551.
Canada's National Climate Archive. 2011. Monthly Observation Data - Canada's
National Climate Archive - Archives Climatiques Nationales Du Canada |
Meteorological Service of Canada - Service Mtorologique Du Canada.
(http://www.climate.weatheroffice.gc.ca/), January.
National Weather Service Climate. 2011. NOAA's National Weather Service.
http://www.weather.gov/climate/, January.

25

26

CDCC-11

DURABILITY AND SUSTAINABILITY OF FIBRE REINFORCED POLYMER

(FRP) COMPOSITES FOR CONSTRUCTION AND REHABILITATION
DURABILIT ET CYCLE DE VIE DES MATRIAUX COMPOSITES
EN POLYMRES RENFORCS DE FIBRES (PRF) EN
CONSTRUCTION ET RHABILITATION

DURABILITY OF GFRP RODS IN FIELD DEMONSTRATION

PROJECTS ACROSS CANADA
A.A. Mufti1, J.Newhook2, B.Benmokrane3, G.Tadros4 and H.M. Vogel5
1

President, ISIS Canada Research Network, Winnipeg, MB, Canada

Professor, Dalhousie University, Dept. of Civil & Resource Engrg., Halifax, NS, Canada
3
NSERC/Canada Research Chair Prof., Dept. of Civil Eng., Universit de Sherbrooke, QC, Canada
4
Consultant, ISIS Canada Research Network, Winnipeg, MB, Canada
5
Ph.D. Student, University of Manitoba, Dept. Of Civil Engrg., Winnipeg, MB, Canada
2

ABSTRACT
In 2004, the ISIS Canada Research Network initiated a project in which the performance of
Glass Fiber Reinforced Polymer (GFRP) used in demonstration structures across Canada
was studied after 5 to 8 years of exposure to concrete. The study was designed around
apprehensions from the engineering research community to use GFRP as primary
reinforcement in concrete structures due to studies that have indicated that glass fibers may
deteriorate in the relatively high pH environment of concrete. The 2004 study indicated that
the apprehension of the research community was not valid as GFRP was in an excellent
condition. More recently, in 2009, core specimens of GFRP reinforcement were collected
from three of the field demonstration projects studied in the earlier rounds of tests.
Analytical methods including Scanning Electron Microscopy (SEM), Energy Dispersive Xray (EDX), Optical Microscopy (OM) and Differential Scanning Calorimetry (DSC) were
used to assess the condition of GFRP after 10 to 13 years of exposure to concrete in the
selected structures. These analyses were performed at the University of Manitoba by the
ISIS Canada Resource Center and at the University of Sherbrooke with the intent of
providing further information on the durability of GFRP in actual concrete field
environments as opposed to simulated laboratory alkaline environments.
1. INTRODUCTION
The choice of GFRP as a preferred material for concrete reinforcement in North America
relates to its economical competitiveness, its resistance to corrosion, its electromagnetic
permeability, its high strength to weight ratio as well as its excellent fatigue performance
(Memon and Mufti 2004, Sheikh and Homam 2004). However, research performed by
Mufti et al. (2005) emphasizes that numerous and, to some extent, contradictory statements
concerning the durability of GFRP in alkaline concrete environments have been published
in technical literature. Some of these publications suggest that exposure of reinforcement to
simulated concrete pore water and temperatures elevated to 80oC can cause a decrease in

27

the mechanical properties of the material (Porter and Barnes 1998, Bank and Russell 1995,
Sen et al. 2002, Bank et al. 1998). Other experiments have suggested little deterioration of
GFRP reinforcement after 12 months of exposure to alkaline solutions maintained between
20 and 38oC. Accordingly, research performed by Sheard et al. (1997) for the
EUROCONCRETE project indicates that GFRP reinforcement can resist alkaline
environments and that it is suitable for use in a concrete environment.
Results in this paper are intended to provide qualitative and quantitative information on the
performance of GFRP rebar extracted from actual structures across Canada. The samples
represent 10 to 13 years of exposure to concrete environments in the field and are not
bound by the limitations of standard tests performed in laboratory simulated environments.
2. BACKGROUND AND METHODOLOGY
2.1 Field Structures Selected for the Durability Study
From the five structures selected in the initial rounds of tests in 2004, three were
investigated for the durability analysis presented in this paper. The first structure, shown in
Fig. 1(a), is Halls Harbor Wharf. It is located in Nova Scotia on the Bay of Fundy shore
and comprises 45MPa concrete with steel-free precast concrete panels containing ISOROD
GFRP reinforcement and concrete pile cap beams reinforced with a hybrid GFRP-steel bar
system. The structure was 10 years old at the time of extracting the cores. It is typically
subjected to temperatures ranging between -35 and 35oC with wet and dry cycles as well as
splash and tidal salt water. It is also common for the structure to experience daily freezethaw cycles.
The second structure, the Joffre Bridge, is shown in Fig. 1(b). It is located in Sherbrooke,
Qubec, and spans the St-Franois River. The traffic barriers as well as the sidewalks were
designed with 45MPa concrete and GFRP C-bar reinforcement. The structure was 12 years
old at the time of core extraction and is subjected to the same environmental conditions as
Halls Harbour Wharf with the exception of splash and tidal water. It can also be overlaid
by substantial amounts of de-icing salts during the winter months.
The third and final structure is the Crowchild Trail Bridge located in Calgary, Alberta. The
structure is shown in Fig. 1(c) and contains ribbed-deformed GFRP C-bar reinforcement in
the barrier walls and the deck slab. Unlike Halls Harbour Wharf and Joffre Bridge, the
concrete used in the Crowchild Trail Bridge was designed to a compressive strength of
35MPa. The bridge was 13 years old at the time of core extraction and operates under a
thermal range of -15 to 23oC. It is also subjected to freeze-thaw cycles and can be regularly
overlaid with de-icing salt.

28

(a) Halls Harbour Wharf

(b) Joffre Bridge

(c) Crowchild Trail Bridge

Fig. 1. Field Projects Considered for Durability Assessment

2.2 Analytical Methods and Sample Preparation
Experienced contractors were hired to extract at least 5 cores of GFRP reinforced concrete
at various locations in each of the three structures considered. The samples had diameters
ranging from 75mm to 100mm and lengths varying between 75mm and 200mm.
The first analytical method applied to the samples was Optical Microscopy (OM). The
method was used to investigate the condition of the interface between the reinforcement
and surrounding concrete. The cores were carefully cut perpendicular to the axis of three
reinforcing bar segments using a high speed diamond saw. The cutting procedure ensured
that the surface roughness of the segments was minimal to ease sanding/polishing. The
sanding/polishing procedure for OM consisted of three progressively finer sanding steps
and one final polishing step with a 1 alumina powder solution.
The second analytical method considered consisted of performing a Differential Scanning
Calorimetry (DSC) analysis to establish the glass transition temperature Tg of the polymeric
matrix containing the glass fibers in the reinforcing bars. The tests were performed in
Winnipeg, Manitoba, by the Composite Innovation Centre (CIC) using a TA Instrument
Q2000 differential scanning calorimeter having an accuracy of 0.1oC. The values for Tg
will be compared to the values obtained during the first round of testing in order to detect
any changes in the structure and stability of the matrix. It has been documented (Mufti et al.
2005) that the value of Tg can be reduced by the presence of moisture or alkalis in the
polymer, which can cause cracking/hydrolysis of the matrix and fiber-matrix delamination.
The Infrared spectroscopy (FTIR) technique was used as the third method of analysis to
determine the presence of chemical species which are formed during the hydrolysis
reaction. The technique can reveal the presence or not of chemical degradation in the resin
matrix. The chemical durability of the resin matrix is governed to a large extent by the
chemical nature of the structure of the polymeric chain. All resin types have ester bonds in
their structure and this is the weakest link of the polymer. A possible degradation
mechanism of resin is alkali hydrolysis of the ester linkages. Due to the alkaline
environment in concrete, alkali hydrolysis is expected to some extent. During the
hydrolysis reaction, the OH- induces ester linkage attack and the resin chain is broken.
Consequently, the structure of the resin is disrupted and the material properties are
changed. The resin material degrades and eventually will not be able to transfer stresses to

29

the glass fibers and to protect the glass fibers against alkaline attack. Changes in the amount
of hydroxyl groups present in the composite material provide insight into the hydrolysis
reaction. The relative amount of hydroxyl groups in the specimens were measured by
determining the ratio between the maximum of the band corresponding to the hydroxyl
groups at 3430 cm-1 and the band corresponding to the carbon-hydrogen groups at 2900 cm1
in the FTIR spectra. The C-H content is considered constant. Since the vinyl ester resins
naturally contain hydroxyl groups, all the spectra present a strong absorption band in this
region. The FTIR analysis was performed at University of Sherbrooke on GFRP samples
from Joffre Bridge and Halls Harbour Wharf.
Scanning Electron Microscope and Energy Dispersive X-Ray (SEM/EDX) analysis was
also performed on samples of composite reinforcement from each of the three structures
considered in this paper. The SEM portion of the analysis was used to investigate whether
the fibers have been damaged by chemical degradation or other means. Conversely, the
EDX analysis was performed to trace chemicals that are likely to cause a degradation of the
fibers over the service life. The presence of sodium or potassium in the EDX scans is
generally an indication of alkali migration from the concrete pore solution toward the glass
fibers. The samples were cut using a low speed diamond saw to minimize surface damage
and finished with one sanding step and three progressively finer polishing steps using a
0.3 alumina powder solution.
3. ANALYSIS RESULTS AND DISCUSSION
3.1 Optical Microscopy
Selected scans from the OM are shown in Fig. 2 for each of the structures analyzed. None
of the scans taken intermittently along the length of the samples extracted from the cores
revealed degradation of the interface separating the reinforcement from the concrete after
10 to 13 years of exposure to alkalinity, freeze-thaw as well as wet and dry cycles, de-icing
salts, salt water and/or thermal loading. There are no visible gaps between the bars and the
surrounding concrete and cracks, slits or fissures are absent in both materials. The result is
encouraging and indicates that the bond has been preserved between the materials. The
composite reinforcement, the surrounding concrete and the interface are clearly identified
in the figure.
ISOROD
Interface

C-Bar GFRP

C-Bar GFRP
Interface

Interface
Concrete
Concrete

(a) Halls Harbour Wharf

Concrete

(b) Joffre Bridge

(c) Crowchild Trail Bridge

Fig. 2. Interface Scans from the OM Analysis

30

3.2 Differential Scanning Calorimetry

The glass transition temperature values for the GFRP reinforcement removed from the field
structures are presented in Table 1 along with those obtained from the first round of testing
performed by Mufti et al. (2005). The values for the most recent set of tests performed in
the final stages of 2009 represent an average obtained from a total of three samples that
were cut from each bar specimen given to the CIC and University of Sherbrooke for DSC
testing.
A total of two bar sizes were extracted from the concrete cores of Halls Harbour Wharf
and Joffre Bridge, namely 16mm and 9mm diameter. One bar size was extracted from
Crowchild Trail Bridge, namely 16mm diameter. The glass transition temperature values
for the GFRP specimens removed from the demonstration structures are presented in Table
1. It should be noted that the 9mm diameter bars from Halls Harbour Wharf and 16mm
diameter bars from Crowchild Trail Bridge were not considered during the initial round of
tests in 2004. In the absence of these values, a full comparison with the current state of the
reinforcement is not possible. The results presented in Table 1 indicate that the structure of
the polymer matrix was not significantly disrupted by exposure to the environment at the
Halls Harbour Wharf and Joffre Bridge. Furthermore, it can be noted that the Tg value
obtained for GFRP samples (C-Bar 16 mm) from Crowchild Trail Bridge is similar to the
value obtained for the same GFRP bar (C-Bar 16 mm) used in Joffre Bridge.
Table 1. Glass Transition Temperature (Tg)
Structure
GFRP
First round tests
(2004)
Tg (oC)
Halls Harbor Wharf
Isorod (16 mm dia.)
123

Joffre Bridge

Recent 2009
Tests - Tg (oC)
123

Isorod (9 mm dia.)

N/A

118

C-Bar (16 mm dia.)

107

103

C-Bar (9 mm dia.)

127

121

N/A

103

Crowchild Trail Bridge C-Bar (16 mm dia.)

3.3 Infrared Spectroscopy

Typical results of the FTIR analysis for control and in-service GFRP samples from Joffre
Bridge (Joffre Bridge 9 and 15 mm) and Halls Harbour Wharf (Isorod 15 mm) are shown
in Fig. 3. No significant change in the spectra of the specimens from the demonstration
structures was observed. The contents in resin matrix or glass fiber in the small amounts of
powder required in the analysis vary from one sample to another. Such differences affect
the intensity and the shape of the absorption bands.
Table 2 presents the content ratio of OH / CH for the control specimen as well as for those
removed from the Joffre Bridge and Halls Harbour Wharf demonstration structures. The
31

results show very little change in hydroxyl content in the exposed specimens. In effect, the
OH/CH ratio decreases in the sample from the structures (exposed) indicating that no
hydrolysis reaction occurred in the specimens during exposure of the demonstration
structures to natural environmental conditions.

Fig. 3. FTIR spectra for GFRP bars for Joffre Bridge and Halls Harbor Wharf
Table 2. Ratio of OH/CH for control and exposed specimens
Specimen
OH/CH ratio
Reference (C-Bar 9 mm)
0.57
Joffre Bridge (C-Bar 16 mm)
0.52
Halls Harbour Wharf (Isorod 16 mm)
0.50
Joffre Bridge (C-Bar 9 mm)
0.45
3.3 Scanning Electron Microscope and Energy Dispersive X-Ray
A representative micrograph from the SEM analysis performed on the cross-section of
reinforcement used in Halls Harbour Wharf is shown in Fig. (a). A small sand grain from
the surface finish of the reinforcement can be seen at the top of the scan. The scan does not
reveal any evidence of deterioration in the fibers, the polymer or the fiber/matrix interface.
The result confirms that the GFRP reinforcement in the structure has not undergone alkali
attack in the concrete environment. The white interspersed deposits found in the scan are
residues from the polishing process.

32

(a) Halls Harbor Wharf

(b) Joffre Bridge

(c) Crowchild Trail Bridge

Fig. 4. SEM Micrographs of Reinforcement Cross-Section

Fig. 4(b) shows a micrograph obtained from the SEM analysis performed on reinforcement
extracted from Joffre Bridge. Results from this scan as well as the remaining scans from
Joffre Bridge continue to suggest the absence of deterioration in the fibers, the polymer as
well as the interface. Similar conclusions can be made from the scan shown in Fig. 4(c) for
the reinforcement removed from Crowchild Trail Bridge. All structures therefore contain
reinforcement in which glass fibers have not been exposed to alkali attack from the
surrounding concrete.
In order to confirm results from the SEM analysis, a series of EDX analyses were also
performed on the GFRP samples from each structure. Since silica glass is known to
dissolve in strong alkaline solutions (Mufti et al. 2005), the analyses were used to assess
consequent chemical changes in the glass fibers. Results are illustrated in Fig. 5 for each of
the structures considered.

Si

Ca

(a) Halls Harbor Wharf

Si

Si

Ca

Ca
A

(b) Joffre Bridge

(c) Crowchild Trail Bridge

Fig. 5. EDX Scans of Glass Fibers

33

As observed during the initial round of tests, the main elements detected on the surface of
the samples are silica (Si), calcium (Ca) and Aluminum (Al). There is no evidence of
sodium (Na) or potassium (K) in the scans, which suggests that alkali solution from the
concrete pores has not migrated within the reinforcement of the structures considered.
4. CONCLUSIONS
Several conclusions can be drawn from the results of the analyses presented in this paper.
These conclusions pertain to the durability of GFRP reinforcement that has been used in
Halls Harbor Wharf, Joffre Bridge and Crowchild Trail Bridge for a period of concrete
exposure ranging between 10 and 13 years.

Results from OM did not show any sign of gaps or debonding between the GFRP
reinforcing bars and the surrounding concrete in all of the three structures examined.
The results also suggest that there were no cracks, slits or fissures in the concrete,
indicating that the bond was preserved between the materials.

The results from DSC indicate that the structure of the polymer matrix of GFRP
reinforcement was not significantly disrupted by exposure to the environment.
Neither hydrolysis nor significant changes in the glass transition temperature of the
matrix took place after exposure to the combined effects of concrete alkaline
environment and the external natural environmental exposure for 10-13 years.

The FTIR spectroscopy analyses support the results from the DSC analyses and OM
examinations.

Micrographs from the SEM analyses did not reveal any signs of physical damage to
the fibers, polymer or fiber/matrix interface. The result brings confirmation to the fact
that the structures were not exposed to the detrimental effects of alkali attack from the
surrounding concrete.

Accordingly, results from the EDX analysis do not show traces of alkaline solution
within the cross-section of the reinforcement for any of the structures considered.

5. ACKNOWLEDGMENTS
The authors would like to thank Lisa Hibbert and Eugene Rothwell from the CIC for their
help in performing DSC testing on the samples as well as Dr. Ravinder Sidhu from the
Department of Geological Sciences at the University of Manitoba and Dr. Patrice Cousin
from the Department of Civil Engineering at the university of Sherbrooke for their help in
performing the SEM/EDX and FTIR analyses.

34

6. REFERENCES
1.
2.
3.

4.
5.
6.
7.

8.

Memon, A.H. And Mufti, A.A. 2004. Fatigue Behaviour of Second Generation SteelFree Concrete Bridge Deck Slab. Proceedings of the 2nd International Conference on
FRP Composites in Civil Engineering (CICE 2004), Adelaide, Australia, pp. 765-772.
Sheikh, S. And Homam, M. 2004. A Decade of Performance of FRP-Repaired
Concrete Structures. Proceedings of the ISIS-SHM Workshop, Winnipeg, Manitoba,
Canada, pp. 1866-1873.
Porter, M. L. and Barnes, B. A. 1998. Accelerated Durability of FRP Reinforcement
for Concrete Structures. Proceedings of the 1st International Conference on Durability
of Fiber Reinforced Polymer (FRP) Composite for Construction (CDCC), eds.
Benmokrane, B. & El-Salakawy, E., Sherbrooke, Qubec, Canada, pp. 191-201.
Bank, L. C. and G. Russell T. 1995. Accelerated Test Methods to Determine the LongTerm Behaviour of FRP Composite Structures: Environmental Effects. Journal of
Reinforced Plastics and Composites, 14: 559-587.
Sen, R., Mullins, G., and Salem, T. 2002. Durability of E-Glass/Vinylester
Reinforcement in Alkaline Solution. ACI Structural Journal. 99(3): 369-375.
Bank L.C., Gentry, T.R., Barkatt, A., Prian, L., Wang, F. and Mangla, S. R. 1998.
Accelerated Aging of Pultruded Glass/Vinylester Rods. Proceedings of the 2nd
International Conference on Fiber Composites in Infrastructure (ICCI), 2: 423437.
Sheard, P., Clarke, J.L., Dill, M., Hammerslely, G. and Richardson, D. 1997.
EUROCONCRETE - Taking Account of Durability for Design of FRP Reinforced
Concrete Structures. Non-Metallic (FRP) Reinforcement for Concrete Structures,
Proceeding of 3rd International Symposium, Sapporo, Japan, 2: 75-82.
Mufti, A., Onofrei, M., Benmokrane, B., Banthia, N., Boulfiza, M., Newhook, J.,
Bakht, B., Tadros, G., and Brett, P. 2005. Studies of Concrete Reinforced with GFRP
Specimens from Field Projects. ISIS Canada Research Network Technical Report,
Winnipeg, Manitoba, Canada, 65 p.

35

36

CDCC-11

DURABILITY AND SUSTAINABILITY OF FIBRE REINFORCED POLYMER

(FRP) COMPOSITES FOR CONSTRUCTION AND REHABILITATION
DURABILIT ET CYCLE DE VIE DES MATRIAUX COMPOSITES
EN POLYMRES RENFORCS DE FIBRES (PRF) EN
CONSTRUCTION ET RHABILITATION

EFFECT OF TEMPERATURE CHANGE ON THE BEHAVIOR OF

FRP-TO-CONCRETE BONDED JOINTS
W.Y. Gao 1, J.G. Teng 2 and J.G. Dai 3
1

PhD student, Dept. of Civil & Struct. Eng., Hong Kong Polytechnic Univ., China
Chair Professor, Dept. of Civil & Struct. Eng., Hong Kong Polytechnic Univ., China
3
Assistant Professor, Dept. of Civil & Struct. Eng., Hong Kong Polytechnic Univ., China
2

ABSTRACT
Service temperature variations may significantly affect the bond behavior between
externally bonded fiber reinforced polymer (FRP) reinforcement and concrete. This paper
presents a closed-form analytical solution for the full-range debonding process of FRP-toconcrete bonded joints under combined mechanical and thermal loads. The solution is
based on a bilinear bond-slip model and leads to closed-form expressions. Numerical
results from the solution are presented to illustrate the effect of temperature change on the
debonding load (i.e. bond strength). It is shown that provided the material properties are not
affected by a temperature change, a significant but realistic temperature decrease from the
installation temperature can significantly reduce the bond strength due to the development
of significant thermal interfacial stresses. A useful function of the closed-form solution lies
in the interpretation of pull test results: the solution allows the effect of thermal stresses to
be isolated from the effect of property changes of the bondline in obtaining bond-slip
responses from pull tests.
1. INTRODUCTION
In reinforced concrete (RC) structures strengthened with externally bonded fiber reinforced
polymer (FRP) reinforcement, the bond behavior between FRP and concrete often controls
the load-carrying capacity of the strengthened structure [1]. As a result, many studies have
been conducted on the bond behavior of FRP-to-concrete interfaces (e.g. [2-6]). In
particular, the single-lap pull test (Fig. 1) (or a double-lap pull test which can be seen as
two single-lap tests being conducted simultaneously) has been widely used to study the
ultimate load (i.e. bond strength) of FRP-to-concrete bonded joints and the local bond-slip
behavior of the interface (e.g. [2-4, 7]).
FRP-strengthened RC structures in service are likely to experience significant temperature
variations (e.g. ambient temperature changes and exposure to fire) and such variations can
have a significant effect on the bond performance of FRP-to-concrete interfaces. To
understand the bond behavior of FRP-to-concrete interfaces at different temperatures, the
37

pull test has also been used (e.g. [8-12]). Results from such tests reflect directly the
combined effects of a number of factors including temperature-induced interfacial stresses
and temperature-induced property changes in the bondline (the adhesive and the adjacent
parts of the adherends) and the adherends. A key purpose of such pull tests is to determine
the bond-slip curve of the interface at a specific temperature change, for which the effect of
thermal interfacial stresses needs to be isolated from the effect of temperature-induced
bondline property changes, as only the latter should be included in a bond-slip model for
use in a theoretical model for FRP-strengthened RC structures subjected to temperature
variations. This issue has received little attention and indeed in some existing studies, these
thermal stresses were simply ignored in interpreting pull test results (e.g. [9,10]).
This paper presents an analytical solution for the full-range behavior of FRP-to-concrete
bonded joints at a specific temperature variation from the ambient temperature at
installation. The solution is an extension of the existing analytical solution of Yuan et al.
[6] and serves two important purposes: (a) to provide the first theoretical explanation of
bond strength variations observed in tests due to temperature variations; (b) to provide a
rigorous theoretical basis for the experimental determination of the bond-slip curve by
isolating the effect of thermal stresses from the effect of property changes of the bondline.

FRP
Adhesive
Concrete
d

Fig. 1. Schematic diagram of a single-lap pull test.

2. ANALYTICAL SOLUTION
2.1 Assumptions and Notation
Figure 1 shows the theoretical idealization of a single-lap bonded joint where both
adherends are assumed to experience only membrane deformation. Moreover, it is assumed
that the width and thickness of each of the three components (plate, adhesive layer and
concrete prism) are constant in the longitudinal direction. In such a simplified theoretical
model, the adhesive layer (i.e. the interface or the bondline whose deformation represents
the deformation of the actual adhesive layer and that of the adjacent parts of the two
adherends) is subjected to shear deformation, so that mode II interfacial fracture is the
failure mode. This theoretical model is a close approximation of the behavior of a real
bonded joint [6]. In addition, it is assumed that the interface is still within the linear elastic
range of behavior and the properties of the adherends are not affected by the imposition of
the thermal load; these two assumptions means that the degree of temperature change needs
38

to be appropriately limited. The applicability of the solution to bonded joints whose

adherends have experienced property degradations is discussed later in the paper.
In Fig. 1, ,
and are the thickness, width and bonded length of the FRP plate,
respectively, while
and
are the width and thickness of the concrete prism,
respectively. The elastic moduli of the plate and the concrete are
and , respectively.
For convenience of presentation, the left end of the plate ( =0) is referred to as the free end
and the right end ( = ) the loaded end hereafter.
2.2 Governing Equations
Based on the above assumptions, the horizontal equilibrium consideration of the FRP plate
and of the overall joint cross-section give the following equations:
0

(1)
0

(2)

where is the shear stress in the adhesive layer,

is the axial stress in the FRP plate and
is the axial stress in the concrete prism. The constitutive equations for the adhesive layer
and the two adherends are described by
(3)

(4)

(5)

where
and
are the longitudinal displacements of the FRP plate and of the concrete,
and
are the coefficients of thermal expansion of the FRP plate and of
respectively;
the concrete, respectively; and is the service temperature variation (thermal load).
The interfacial slip
that is

is defined as the relative displacement between the two adherends;

(6)

From Eqs. (1)-(6), the following second-order differential equation can be derived:

(7)

where

39

(8)

In Eqs. (7) and (8), is the local bond strength (i.e. the maximum shear stress on the bondis the interfacial fracture energy (the area underneath the interfacial
slip curve) and
bond-slip curve).
Substituting Eq. (6) into Eqs. (2), (4) and (5) yields

(9)

2.3 Bond-Slip Model

A bilinear bond-slip model has been widely used to study the bond behavior of FRP-toconcrete bonded joints (e.g. [6] ); such a bilinear model is also adopted in the present study.
As shown in Fig. 2, the bilinear bond-slip relationship consists of two segments: an initial
linear elastic segment where the bond shear stress increases linearly with the interfacial slip
until it reaches the peak stress (the corresponding slip is denoted by ) and a softening
segment where the shear stress reduces with the interfacial slip until it becomes zero (the
corresponding slip is denoted by ). This bond-slip model (Fig. 2) is described by the
following equation:
when0
when

(10)

0when

Fig. 2 Bi-linear bond-slip model.

2.4 Predictions for the Debonding Process
With the bond-slip model defined above and following the approach of Yuan et al. [6], the
governing equation (Eq. 7) can be solved to find the interfacial shear stress distribution
along the interface and the load-displacement response of the bonded joint under combined
thermal and mechanical loads.

40

For a bond slip model as shown in Fig. 2, the debonding process can be divided into three
stages [6]: (1) the elastic stage (Stage I), during which the load is small and the interfacial
shear stress stays below ; (2) the elastic-softening stage (Stage II), during which the shear
slip at the loaded end ( = ) has exceeded but is smaller than ; (3) the elasticsoftening-debonding stage (Stage III), during which the interfacial slip at the loaded end
and the shear stress there has reduced to zero; during
has exceeded the separation slip
this stage, interfacial debonding initiates at the loaded end and propagates along the
interface. Since the interfacial shear stress and the tensile stress in FRP can be easily
derived from the interfacial slip, only the expressions for the interfacial slip and the loaddisplacement relationship at the loaded end are presented below. Other details of the
analytical solution can be found elsewhere [13].
2.4.1. Elastic stage (Stage I)
During the elastic stage (i.e Stage I), the bond-slip curve is given by the first expression of
Eq. 10. Substituting Eq. 10 into Eq. 7 and making use of the boundary conditions ( = 0 at
= 0;

at = ), the expression for the interfacial slip is obtained:

where

sinh

(11)

The slip at the loaded end (i.e. the value of at = ) is also referred to as the displacement
of the bonded joint and is denoted by . Based on this definition, the load- displacement
relationship is given by:

tanh

sinh

(12)

The initial displacement due to the temperature change can be calculated as (i.e. = 0
and = in Eq. 11):

ctanh

sinh

(13)

At the end of the elastic stage, the interfacial shear stress reaches with a slip at
the loaded end. Substituting = into Eq. 12 leads to the load at the initiation of
interfacial softening (the beginning of the elastic-softening stage):
1

sinh

41

tanh

(14)

2.4.2. Elastic-softening stage (Stage II)

When the loaded end slip first exceeds , the free end slip is still less than , so the right
part of the interface is now in a softening state while the left part of the interface is still in
the linear elastic state. As the load further increases, the length of the softening zone
also increases. In the elastic zone (0
,0
), the interfacial slip has the
same form as Eq. 11:

cosh

In the softening zone (

in the FRP plate are given by

(15)

), the interfacial slip and the axial stress

tanh

sinh

sin

cos

where

(16)

(17)
at the loaded end ( = ) into Eq. 17 yields

tanh

cos

sin

and the displacement

tanh

sin

Substituting the boundary condition

cos

(18)
at the loaded end can be found from Eq. 16 (i.e. = ) to be
tanh

sin

cos
(19)

Eqs. 18-19 can be used to predict the load-displacement relationship for the elasticsoftening stage by varying the value of .
2.4.3. Elastic-softening-debonding stage (Stage III)
At the end of Stage II, the slip at the loaded end reaches , indicating the initiation of
debonding at the loaded end. The corresponding value of is its maximum possible value
(denoted by ) and is determined using the following equation which is obtained from Eq.
19:
42

tanh

sin

cos

(20)
For an infinite bond length, Eq. 20 reduces to
arctan

(21)

As the debonding crack propagates, the location of the peak shear stress
moves away
from the loaded end towards the free end, and as a result, the total length of the intact
interface reduces. Assuming that the length of the debonded interface is , the equations for
the interfacial slip (Eqs. 15 & 16) and the load (Eq. 18) are still valid if is replaced by ( ). Whereas, the displacement can be identified with the consideration of the thermalinduced slip in the debonded length, as shown in following

(22)

3. NUMERICAL RESULTS
Klamer [11, 12] conducted a series of pull tests on double-lap FRP-to-concrete bonded
joints (Fig. 3) at temperatures ranging from -20 oC to 100 oC. The test specimens (including
the installation of strain gauges on the FRP plate) were prepared at 20 oC (the reference
temperature) but the tests were conducted at a different temperature (-20 oC, 20 oC, 40 oC,
50 oC, 70 oC 80 oC or 100 oC). The thicknesses of the FRP pultruded plate and the adhesive
layer were 1.2 mm and 1.5 mm, respectively, while the bond length was 300 mm. Figure 4
presents a comparison between the predicted thermal strains in the CFRP plate and the
experimental values for one of the specimens which was subjected to a temperature
increase of 30 oC (i.e. = 30 oC) before the external load was applied (Klamer [11]
reported the thermal strains of the FRP plate only for this specimen). In making the
predictions, the following parameters were used as provided by Klamer [11]: = 100 mm,
= 75 mm, = 150 mm, =165 000 MPa, =26 800 MPa, = 10.210-6 /oC and =
0.310-6 /oC. From the experimental load-displacement curves of the two specimens tested
at the reference temperature (20 oC), the loads corresponding to the initiation of softening
and debonding were found to be 20 KN and 44.7 KN respectively [13], so the interfacial
parameters were identified as: =0.35 mm, =0.08 mm, =3.07 MPa, and
=0.54
N/mm [6]. The close agreement between the test results and the analytical predictions
demonstrates the validity of the present closed-form solution.
Figure 5 shows a comparison between the debonding loads of FRP-to-concrete bonded
joints obtained from Klamers tests [11] and the analytical solution. In this figure, the
debonding loads are normalized by the corresponding experimental or predicted value for
the reference temperature of 20 oC for a clearer comparison. It is seen that the experimental
debonding load initially increases as the temperature increases (or decreases as the
temperature reduces) but the trend reverses when the temperature is around the glass
transition temperature
of the bonding adhesive which was 62 oC [11]. The analytical
43

debonding load increases monotonically with the temperature. The differences between the
experimental debonding loads and the analytical predictions for temperatures above the
glass transition temperature are due to the omission of the effect of softening of the bonding
adhesive as in the present predictions the bilinear bond-slip law was identified from the test
results for the reference temperature of 20 oC. In addition, any softening of the adherends
was also ignored. The close agreement between the experimental results and the analytical
predictions for temperatures below
further verify the reliability of the closed-form
solution.
Figure 6 presents numerical results from the analytical solution to examine the effect of
temperature change on the normalized debonding load for FRP plates of different
thicknesses (i.e. different tensile stiffnesses), with = 0 oC being for the reference case.
It is clearly seen that for the same temperature rise, the increase in the debonding load is
larger when a thicker FRP plate is used. If a 2.4 mm thick FRP plate is used, a temperature
increase of 50 oC leads to about a 27% increase in the debonding load. This aspect needs to
be considered in engineering practice when an FRP-strengthened structure is subjected to
significant service temperature variations.
300

Analytical predictions
Test data

Strain in FRP (um/m)

250
200
150
100
50
0

Fig. 3. Test specimens of Klamer [11]

Normalized ultimate debonding load (KN)

Normalized ultimate debonding load

1.4
1.2
1
0.8
0.6
0.4
0.2
0

20

100
150
200
Distance from the left end (mm)

250

300

1.3

Analytical predictions
Test data

1.6

0
-20

50

Fig. 4. Thermal strain distribution in the FRP

2
1.8

40

60

80

1.25
1.2
1.15
1.1
1.05
1
-10

100

Temperature ( C)

Fig. 5. Normalized ultimate debonding load

vs. temperature

0.15 mm
0.3 mm
0.6 mm
1.2 mm
2.4 mm

10

20
30
o
Temperature ( C)

40

50

60

Fig. 6. Normalized ultimate debonding load

vs. temperature

44

4. DETERMINATION OF BOND-SLIP CURVES FROM PULL TESTS

It is now clear that the debonding load of an FRP-to-concrete bonded joint can be
significantly affected by a temperature variation from the installation temperature. This has
a significant implication when pull test results are used to derive bond-slip curves. In such a
derivation, the interfacial fracture energy
is directly related to the debonding load of the
bonded joint when temperature variations are not involved (e.g. [4,5 14]). When pull tests
are used to derive bond-slip laws at a temperature change, previous authors have followed
the same approach [9]. In a correct approach, the effect of thermal stresses on the
debonding load needs to be eliminated using the present solution for the establishment of
the bond-slip curve which may include the effect of softening of the bonding adhesive.
Readers are referred to Ref. [13] for detailed discussions.
5. CONCLUSIONS
This paper has presented a closed-form analytical solution for the full-range behavior of
FRP-to-concrete bonded joints under combined mechanical and thermal loads. A bilinear
local bond-slip relationship is employed in the solution, but the general characteristics
observed from the solution are applicable to interfaces with a similar bond-slip model. The
solution provides close-form expressions for the interfacial slip, interfacial shear stress,
stress in the FRP plate as well as the load-displacement relationship for the entire
debonding process. The predictions of the closed-form solution have been compared with
the limited test data available, demonstrating close agreement between them. It has been
shown that the debonding load increases as the temperature increases provided the material
properties of the bonded joint have not been affected by the temperature change; this
increase is more significant when a stiffener FRP plate is used. The present solution can be
used to design against debonding failure in FRP-strengthened steel/concrete structures
subjected to significant service temperature variations. In addition, the solution also
provides a useful tool by which the test results of FRP-to-concrete bonded joints with
temperature variations can be correctly interpreted to obtain bond-slip curves.
6. ACKNOWLEDGEMENTS
The authors are grateful for the financial support received from the Research Grants
Council of the Hong Kong SAR (Project No: PolyU 516509) and The Hong Kong
Polytechnic University provided through a PhD studentship to the first author.
7. REFERENCES
1.
2.
3.

Teng, J.G., Chen, J.F., Smith, S.T., and Lam, L. 2002. FRP-Strengthened RC
Structures. John Wiley & Sons Ltd., Chichester, 266 p.
Chen, J.F., and Teng, J.G. 2001.Anchorage strength models for FRP and steel plates
bonded to concrete. Journal of Structural Engineering, ASCE, 127(7): 784-791.
Yao, J., Teng, J.G., and Chen, J.F. 2005. Experimental study on FRP-to-concrete
bonded joints. Composites Part B: Engineering, 36(2): 99-113.

45

4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.

Dai, J.G., Ueda, T., and Sato, Y. 2005. Development of the nonlinear bond stress-slip
model of fiber reinforced plastics sheet-concrete interfaces with a simple method.
Journal of Composites for Construction, ASCE, 9(1): 52-62.
Dai, J.G., Ueda, T., and Sato, Y. 2006. Unified analytical approaches for determining
shear bond characteristics of FRP-concrete interfaces through pullout tests. Journal of
Advanced Concrete Technology, JCI, 4(1): 133-145.
Yuan, H., Teng, J.G., Seracino, R., Wu, Z.S., and Yao, J. 2004. Full-range behavior of
FRP-to-concrete bonded joints. Engineering Structures, 26(5): 553-565.
De Lorenzis, L., Miller, B., and Nanni, A. 2001. Bond of Fiber-reinforced Polymer
Laminates to Concrete. ACI Material Journal, 98(1): 256-264.
Blontrock, H. 2003. Analysis and Modeling of the Fire Resistance of Concrete
Elements with Externally Bonded FRP Reinforcement. PhD Thesis, Ghent University,
Ghent, Belgium.
Wu, Z.S., Iwashita, K., Yagashiro, S., Ishikawa, T., and Hamaguchi, Y. 2005.
Temperature effect on bonding and debonding behavior between FRP sheets and
concrete. Journal of the Society of Materials Science, 54(5): 474-480.
Leone, M., Matthys, S., and Aiello, M.A. 2009. Effect of elevated service temperature
on bond between FRP EBR systems and concrete. Composites Part B: Engineering,
40(1): 85-93.
Klamer, E. 2006. The Influence of Temperature on Concrete Structures Strengthened
with Externally Bonded CFRP. Research Report, Eindhoven University of Technology,
Eindhoven, Netherlands.
Klamer, E. 2009. Influence of Temperature on Concrete Beams Strengthened in
Flexure with CFRP, PhD Thesis, Eindhoven University of Technology, Eindhoven,
Netherlands.
Gao, W.Y., Teng, J.G., and Dai, J.G. 2011. Effect of temperature variation on the fullrange behavior of FRP-to-concrete bonded joints. (in preparation).
Lu, X.Z., Teng, J.G., Ye, L.P., and Jiang, J.J. 2005. Bond-slip models for FRP
sheets/plates bonded to concrete. Engineering Structures, 27(6): 920-937.

46

CDCC-11

DURABILITY AND SUSTAINABILITY OF FIBRE REINFORCED POLYMER

(FRP) COMPOSITES FOR CONSTRUCTION AND REHABILITATION
DURABILIT ET CYCLE DE VIE DES MATRIAUX COMPOSITES
EN POLYMRES RENFORCS DE FIBRES (PRF) EN
CONSTRUCTION ET RHABILITATION

LONG-TERM BOND BEHAVIOUR OF GFRP REBAR IN SEVER

ENVIRONMENTS
B. Juette1, A. Weber2 and C. Witt3
1

Product Manager ComBAR, Schoeck Bauteile GmbH, Baden-Baden, Germany

Head of R & D, Division of GFR, Schoeck Bauteile GmbH, Baden-Baden, Germany
3
President, Schoeck-Canada Inc., Kitchener, ON
2

ABSTRACT
Increasingly glass fibre reinforced polymer (GFRP) rebars are being installed - instead of
conventional steel reinforcement - in concrete structures located in corrosive, that is moist
and warm, environments. Recent material developments have resulted in much higher
tensile strengths and moduli of elasticity compared to GFRP bars of earlier generations.
These material developments make newer generation GFRP bars suitable for long-term
applications under high sustained stresses. Under these conditions the long-term bond
properties of these bars are critical. For the structural design long-term design values of the
bond strength must be determined.
The bond behaviour and the bond creep behaviour of reinforcing bars in general are
discussed in the paper. The critical issues, especially those applying to GFRP bars, are
defined. The results of a test series on the bond durability and a test series on the bond
creep behaviour of a newest generation GFRP rebar are presented. Finally, a procedure for
the determination of long-term design values of the bond strength using the results of the
bond creep tests is explained in detail.
1. INTRODUCTION
Glass fibre reinforced polymer (GFRP) reinforcing bars were first developed as crack
reinforcement in North America in the 1980s. Massive problems encountered there due to
the corrosion of the steel reinforcement in bridge decks had led to the search for alternative
reinforcing materials to conventional steel rebar. As the prices for stainless steel are
extremely volatile and usually very high, GFRPs continue to be an economic reinforcement
solution for bridge decks and barrier walls.
The first materials available on the market had relatively low tensile strengths and moduli
of elasticity. The approach at the time was, therefore, to install comparatively large
amounts of bars with large diameters (high rebar ratio) in the bridge decks. Thereby the
stresses in the bars were kept relatively small while the widths of the cracks in the decks
47

were limited to acceptable sizes. As the corresponding bond stresses on the bars were
relatively low, bond behaviour was tested only in short-term tests. Little attention was paid
to long-term bond behaviour and bond creep behaviour of these bars. (ACI440; CSA S80602)
In the late 1990s the search for corrosion resistant reinforcing bars began in Europe, where
high prices for stainless steel rebar made its installation increasingly uneconomic. Here the
focus lay on load bearing reinforcement in relatively thin concrete elements, such as
balcony slabs, subjected to high loads.
As the stresses which needed to be permanently sustained by these bars were much higher,
durability and long-term behaviour became a central issue in the qualification tests of these
bars. A testing concept was developed to analyse the bond creep behaviour of these bars
and determine design values of the long-term bond stress which insure acceptable
performance of these bars for service life spans of as much as 100 years.
2. BOND CREEP
2.1 Bond Creep in Reinforced Concrete
Bond creep of steel reinforcement in concrete is defined as the time dependent increase of
the slip of a steel reinforcing bar embedded in concrete under permanent load (Franke
1976).
Bond creep occurs along the embedment length of reinforcing bars as the slip of the bars
increases with time. The stresses in the bar and the surrounding concrete change in that
region, as a result. Bond creep is one of the reasons for the time dependent increase of the
widths of cracks in reinforced concrete elements.
One of the causes of the increase in the relative displacement between the reinforcing bar
and the surrounding concrete lies in the visco-elastic and plastic compression of the
concrete corbels as local compressive stress peaks occur in the immediate vicinity of the
bar ribs. Radial micro cracking of the concrete is most likely more pronounced near the ribs
or the reinforcing bar, also contributing to an increase in the relative displacement between
the concrete and the reinforcing bar.
2.2 Bond Creep in Concrete Reinforced with GFRP Bars
GFRP bars are available from a large number of producers worldwide. A large number of
different materials (fibres and resins) are used in their production. Their geometry and
surface properties and therefore their properties in concrete vary greatly. Whereas bars,
especially those of earlier generations, are often laid, newer generation bars, conceived to
sustain higher permanent tensile stresses, are usually produced in a so called pultrusion
process. In this process the bars are pulled through the machinery at high tension. This
ensures a nearly perfect linear and parallel alignment of the fibres in the bars.

48

Bar surfaces also vary greatly. A number of producers sand coat the finished bars to insure
bond, others apply ribs to the core of the bars during the production process or after
hardening of the bars, others cut ribs into the hardened bars.
The bond between a reinforcing bar and the surrounding concrete is ensured via three
mechanisms: adhesion, shear (force transfer in to the concrete corbels parallel to the bar)
and friction between the bar and the concrete. All three mechanisms are strongly influenced
by surface geometry, texture and consistency. As GFRP bars contain linearly oriented
fibres, they are, unlike steel, not isotropic. As a result, their bond behaviour is additionally
controlled by the inter-laminar shear properties of the material itself (transfer of forces
along the interface between the ribs and the core cross section of the bar).
As a result of all these considerations the bond and bond creep behaviour of the various
GFRP bars available on the world market differ significantly. Detailed and strenuous shortand long-term bond and bond creep testing is required to determine safe design values of
the bond stress for each individual bar.
3. TESTING BOND CREEP BEHAVIOR OF FRP REBARS
3.1 Intent of Testing Series
In conjunction with the applications for general construction authority certifications of a
newest generation pultruded GFRP reinforcing bar in Germany and in The Netherlands
(DIBt 2008, KOMO, KIWA 2009) a testing scheme was developed to determine the
allowable design value of the long-term bond stress sustainable by these bars. The
behaviour of the bars after the concrete has cracked and their bond creep behaviour in
higher grade concretes were to be studied in an extensive testing series using this scheme.
3.2 Test Specimen
Eighteen (18) test specimen were prepared according to the RILEM testing
recommendation RC 6. GFRP rebars with a core / nominal diameter of 8, 16 and 25 mm
were tested. The concrete cubes had side lengths of 150mm and 250mm respectively. The
bars were located in the center of the concrete cubes. The bottom end of the bars extended
beyond the face of the cube to allow for the measurement of the slip at the unloaded bar
end. The embedment length of the bars was four or five times the bar diameter. The rest of
the bar length was covered by plastic pipes to debond the bar (Weber 2009).

Fig. 1. Curing of test specimen in water

49

The concrete recipe was devised so as to obtain a compressive strength of 75MPa (grade
C50/60 according to EC-2). The specimen were cast horizontally and tested vertically. This
configuration corresponds to the concreting direction in real life near-surface installations
of the bars. Each prism was reinforced with two closed carbon steel stirrups (d = 6 mm,
grade 500 steel) located at the beginning and in the middle of the embedment length. The
specimens were cured in water at room temperature. They were at least 28 days old when
tested.
3.3 Testing Concept
The central idea behind the testing scheme is to study the behaviour of the bars in cracked
concrete under extreme yet realistic environmental conditions.
To simulate these conditions alkaline concrete (Portland cement) was used on all specimen.
The specimens were pre-loaded to a slip of 1 mm at the free bar end at room temperature.
For the long-term bond creep tests the specimen were heated to 60C and were kept water
saturated at all times. This environment simulates real life conditions for service life spans
as they are required in real life projects. The load was sustained on the bars for at least 2000
hours without them showing any signs of increasing slip due to bond creep.
Bond Creep Test Sheme
25

Loading over peak

bond stress in MPa

20

15

unloading to 7,5 MPa

10
2000h creep at 60C in wet
alkaline concrete

design value for bond

0
0

1
2
slip at unloaded end of bar in mm

Fig. 2. Concept of bond creep testing

3.4 Pre-Loading of Specimen
In a first step a tensile load was applied on the bars at room temperature until a total slip of
1mm was measured at the end of the bar opposite to the load application. The slip was
recorded as a function of the (bond) force applied on the bars.
The maximum bond stress achieved by the 8mm bars during this pre-loading was between
20 and 26 MPa at a slip between 0.5 and 0.7 mm.

50

30

bond
force in kN
Verbundkraft
in kN

25
20
8 V10
8 V11

15

8 V12
10
5
0
0

0,2

0,4

0,6

0,8

1,2

1,4

slip at unloaded end in mm

1,6

Schlupf am lastabgewandten Ende in mm

Fig. 3. Pre-loading to 1mm slip d = 8 mm ComBAR bars

The maximum bond stress for the 16 mm bars was between 22 and 25 MPa at a slip
between 0.1 and 0.4mm.
120

bond
force in in
kNkN
Verbundkraft

100
80
16 V9
60

16 V10 002

40
20
0
0

0,2

0,4

0,6

0,8

slipamatlastabgewandten
unloaded end inEnde
mmin mm
Schlupf

1,2

Fig. 4. Pre-loading to 1mm slip d = 16 mm ComBAR bars

The maximum bond stress for the 25mm bars was between 16 and 19 MPa at a slip
between 0.3 and 0.4mm.
160

bond
force in kN
Verbundkraft
in kN

140
120
100

V1
V2

80

V3
60
40
20
0
0

0,2

0,4

0,6

0,8

1,2

slipam
at lastabgewandten
unloaded end in Ende
mm in mm
Schlupf

Fig. 5. Pre-loading to 1mm slip d = 25 mm ComBAR bars

51

3.5 Bond Creep Testing

After this pre-loading all specimen were unloaded and prepared for the long-term test. To
simulate real life conditions and to accelerate any aging processes in the bars the specimen
were saturated in water and heated to 60C prior to the application of the load. A constant
tensile load was then applied to the bars. The applied loads and the resulting bond stresses
are shown in Table 1.
bar diameter
[mm]
8mm
8mm
16mm
25mm

Table 1. Applied constant loads, bond stresses ComBAR

specimen
embedment
load [kN]
bond
stress
designation
length [mm]
[MPa]
V4-V6
40
9.60
9.55
V7-V12
40
8.07
8.03
V7-V10
80
30.31
7.54
V1-V6
100
70.47
8.97

The loads were sustained on the bars for at least 2000 hours. The tests were only declared
acceptable if there were no signs of an increase in the bond creep over this period of time.
The slip curves are shown in Figures 6 to 8.
The 8mm bars did not pass the test at a bond stress of 9.5 MPa (see Figure 6). They were,
however, able to sustain a bond stress of 8.0 MPa for up to 5000 hours.
1,0

0,9

displacement
bar end
[mm]
Weg amat
Stabende
[mm]

0,8

0,7

8-V4 9,5 MPa

8-V5 9,5MPa
8-V6 9,5MPa
8-V7 8MPa
8-V8 8MPa
8-V9 8MPa
8-V11 8MPa
8-V12 8MPa

0,6

0,5

not acceptable

0,4

0,3

0,2

0,1

0,0
0

500

1000

1500

2000

2500

Laufzeit
[h]
time
[hours]

3000

3500

4000

4500

5000

Fig. 6. Bond creep curves d = 8mm ComBAR bars at 60C, pre-loaded to 1mm slip
The 16 mm bars were able to sustain a permanent bond stress of 7.5 MPa without showing
signs of growing bond creep.

52

0,50
0,45

displacement
bar end
Weg am at
Stabende
[mm] [mm]

0,40
0,35
0,30

16-V7
16-V8
16-V9

0,25

16-V10

0,20
0,15
0,10
0,05
0,00
0

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

5500

6000

[h]
timeLaufzeit
[hours]

Fig. 7. Bond creep curves 16mm ComBAR, 60C, bond stress 7.5 MPa, pre-loaded to
1 mm slip
The 25 mm bars were able to sustain a permanent bond stress of 9.0 MPa without showing
signs of growing bond creep.
0,50
25-V1
25-V2

displacement
at bar end [mm]
Weg am Stabende [mm]

0,45

25-V3
25-V4
25-V5

0,40

25-V6

0,35

0,30

0,25

0,20

0,15

0,10

0,05

0,00
0

500

1000

1500

2000

2500

3000

3500

4000

4500

5000

[h]
timeLaufzeit
[hours]

Fig. 8. Bond creep curves 25mm ComBAR at 60C, bond stress 9.0 MPa, pre-loaded to
1 mm slip
3.6 Inspection of Specimen after Testing
One concrete cube from each testing series was cut open after the test to inspect the GFRP
bars. Most of the ribs of the 8 and 16 mm bars had been sheared off in the tests. On the
25mm tests the concrete corbels had failed.
4. DERIVATION OF DESIGN VALUES OF BOND STRENGTH
Based on the results of these tests, the design values of the bond strength can be specified
for the specific GFRP reinforcing bar and for individual bar diameters thereof for service
53

lives up to 100 years. The characteristic value of the bond stress is derived by statistical
means from the bond stress sustained by at least five tests over at least 2000 hours without
showing any increase in the bond creep. It applies, of course, only to the concrete grade
tested in the series. In Germany a safety factor of 0.65 is applied to this characteristic value
to obtain the design value of the bond stress for that particular bar and the specific grade
concrete. (Schiel 2008)
In the case of the bar tested in this series the design values of the bond stress in normal
grade concretes were found to be essentially the same as those of steel reinforcement as it is
commonly used in central Europe.
5. CONCLUSIONS
The bond creep properties of GFRP rebars are a critical measure of their durability and
long-term load bearing characteristics. As the materials used in the production of the
various GFRP rebars available on the market differ greatly, and the bars themselves have
different surface profiles, geometries and textures, their bond properties differ greatly.
A testing procedure has been developed to determine the long-term bond behaviour (bond
creep) of GFRP reinforcing bars in pre-cracked concrete sections. In this procedure real life
environmental conditions are simulated by testing the bars in saturated highly alkaline
concrete prisms. To speed up the testing process and to simulate long-term applications of
the bars the tests are conducted at 60C.
In the tests the characteristic or guaranteed value of the bond stress is determined for the
tested concrete for a design service life of up to 100 years. The design value of the bond
stress for the particular grade concrete is derived from the results of these tests by applying
a safety factor of 0.65.
6. REFERENCES
1.
2.
3.
4.
5.

ACI Committee 440. 2004. Guide test methods for fiber reinforced polymers FRPs for
reinforcing or strengthening concrete structures. ACI 440.3R-04, American Concrete
Institute, Farmington Hills, MI, 40 p.
Canadian Standards Association (CSA). 2002. Design and construction of building
components with fibre reinforced polymers. CAN/CSA S806-02, Mississauga, ON,
177 p.
Franke, L. 1976. Einfluss der Belastungsdauer auf das Verbundverhalten von Stahl in
Beton, Verbundkriechen, Heft 268 DAfStb, Beuth Verlag, Berlin, Germany.
Schiel, P., Volkwein, A., Knab, F. 2008. Application for DIBt Certification of GFRP
Reinforcing Bars, ComBAR made by Schck Bauteile GmbH. expert report, Munich,
Germany.
Weber, A. 2009. Bond creep of ComBAR bars with d = 8 mm, d = 16 mm and d = 25
mm, Lab report, Schck Bauteile GmbH, Baden-Baden, Germany.

54

CDCC-11

DURABILITY AND SUSTAINABILITY OF FIBRE REINFORCED POLYMER

(FRP) COMPOSITES FOR CONSTRUCTION AND REHABILITATION
DURABILIT ET CYCLE DE VIE DES MATRIAUX COMPOSITES
EN POLYMRES RENFORCS DE FIBRES (PRF) EN
CONSTRUCTION ET RHABILITATION

EFFECT OF VARIOUS SEVERE ENVIRONMENTAL EXPOSURES

ON CFRP AND SFRP-CONFINED CIRCULAR CONCRETE
COLUMNS
M. A. Mashrik1 and R. El-Hacha2
1
2

MSc Student, Department of Civil Engineering, University of Calgary, Calgary, Canada

Associate Professor, Department of Civil Engineering, University of Calgary, Calgary, Canada

ABSTRACT
This paper evaluates the performance of plain concrete cylinders strengthened with Steel
Fibre Reinforced Polymer (SFRP) and Carbon Fibre Reinforced Polymer (CFRP) sheets
exposed to different aggressive environments. A total of forty-five plain concrete cylinders
(150 mm diameter 300 mm height) were divided into five groups with nine specimens in
each group (three unconfined, three CFRP-confined, and three SFRP-confined). One group
was considered control and kept at room temperature (+22C). The remaining four groups
were subjected to four different environments: (a) prolonged high temperature of +45C for
135 days; (b) prolonged high temperature of +60C along with relative humidity of 96% for
42 days; (c) 90 freeze-thaw cycles; and (d) 471 freeze-thaw cycles. Each freeze-thaw cycle
was 7 hours and 8 minutes between -34C to +34C and a relative humidity of 75% for
temperature above +20C. Monotonic concentric uniaxial compression tests were
conducted on the specimens and the experimental results were correlated with statistical
analysis. Results showed that severe environments had a detrimental effect on the axial
strength of the CFRP wrapped cylinders, but not on that of the SFRP wrapped cylinders.
1. INTRODUCTION
Fibre Reinforced Polymer (FRP) materials have emerged as an economic and efficient
solution for strengthening structural concrete members. FRP sheets can be bonded onto
flexural members to increase flexural and shear capacity, or wrapped around the columns to
improve axial strength and ductility (El-Hacha et al., 1999). Such FRP-confined columns
may be exposed to aggressive environments during their service life. In Canada, exterior
columns (especially in bridges and parking structures) are likely to be subjected to
temperature extremes and freeze-thaw cycles. For the practical acceptance of FRP-confined
columns, it is necessary to evaluate the effects of aggressive environments on the
performance of the FRP confinement. This paper deals with the use of the newly developed
Steel Fibre Reinforced Polymer (SFRP) sheet to wrap plain concrete cylinders and compare

55

with the traditionally used Carbon Fibre Reinforced Polymer (CFRP) sheet; both at room
temperature and under different severe environments.
2. EXPERIMENTAL PROGRAM
2.1 Materials
2.1.1 Concrete
The target 28 day unconfined concrete compressive strength was 40 MPa while the average
of three 100200 mm concrete cylinders tested at 28 days was 43.1 MPa with a standard
deviation of 2.1 MPa.
2.1.2 Steel Fibre Reinforced Polymer (SFRP) Sheet
The SFRP sheet used in this project was type 32-20-12 made of unidirectional brass
coated ultra high strength twisted steel wires with a thickness of 1.23 mm and a net crosssectional area of 0.38 mm2/mm. According to the manufacturer, the ultimate tensile
strength and effective modulus for the SFRP sheet are 986 MPa and 66100 MPa,
respectively, and the ultimate tensile strain at failure is 1.5% (Hardwire, 2010a).
2.1.3 Carbon Fibre Reinforced Polymer (CFRP) Sheet
The CFRP sheet used in this project was SikaWrap Hex 230C made of unidirectional
carbon fibre fabric with a thickness of 0.381 mm. According to the manufacturer, the
ultimate tensile strength and modulus of elasticity are 894 MPa and 65402 MPa,
respectively, and the strain at failure is 1.33% (Sika, 2010b).
2.1.4 Epoxy Adhesive
Sikadur 330 was used to bond the SFRP and CFRP sheets around the cylinders. The
Sikadur330 is a two-component epoxy resin that has a tensile strength of 33.8 MPa,
modulus of elasticity of 3489 MPa and 1.9% elongation at break (Sika, 2010c).
2.2 Strengthening Procedure and Instrumentation
The concrete cylinders and FRP sheets were cleaned off dust before wrapping. The SFRP
and CFRP sheets were cut into desired lengths, allowing an overlap length of 100 mm.
Using the wet-lay up procedure, a thin layer of epoxy adhesive was applied to one side of
the FRP sheets and the sheets were wrapped around the cylinders in a continuous manner.
All cylinders were wrapped with one layer (300mm wide) of CFRP sheet or SFRP sheet
having almost the same axial stiffness (EACFRP= 7.48kN and EASFRP= 7.51kN, where E is
the modulus of elasticity of the FRP sheet and A is the area/300mm width).
The cylinders were instrumented with two strain gauges on two opposite sides. For FRP
wrapped cylinders, both were to measure circumferential strains; for unwrapped cylinders,
one was to measure circumferential strain and the other was to measure axial strain.
56

2.3 Test Set-up and Test Matrix

The cylinders were tested under monotonic concentric uni-axial compression load on a 2
MN compression testing machine at an average rate of 10 kN/sec until failure. Forty-five
cylinders were divided into five groups with nine specimens in each group (three
unconfined, three CFRP-confined, and three SFRP-confined). Table 1 presents the detailed
test matrix of the experimental program.

Specimen
Designation
RT-UW
RT-C
RT-S
HT-UW
HT-C
HT-S
HTRH-UW
HTRH-C
HTRH-S
FT90-UW
FT90-C
FT90-S
FT471-UW
FT471-C
FT471-S

Environmental
Exposure

Table 1. Test Matrix

Duration
Details
(Days)

Room
Temperature

22C

Prolonged High
Temperature

+45C

135

High
Temperature
+ Relative

+60C and
96% RH

42

Freeze-Thaw
Cycles
Freeze-Thaw
Cycles

"-34C" to
"+34C with
75%RH"
"-34C" to
"+34C with
75%RH"

27
(90 cycles)
140
(471 cycles)

Wrapping
Material
Unwrapped
CFRP
SFRP
Unwrapped
CFRP
SFRP
Unwrapped
CFRP
SFRP
Unwrapped
CFRP
SFRP
Unwrapped
CFRP
SFRP

No. of
Specimens
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3

The specimen designation in Table 1 are explained as follows: except for the freeze-thaw
exposure, the first letters represent the environment (RT for room temperature, HT for
high temperature, HTRH for high temperature with relative humidity), followed by
letters representing wrapping material (UW for unwrapped, S for SFRP, C for
CFRP). For the freeze-thaw exposure, the first letters represent the environment condition
(FT for freeze-thaw), followed by number of cycles, and followed by wrapping material.
3. EXPERIMENTAL RESULTS AND DISCUSSION
In this section the results are reported and discussed in terms of ultimate strength, ductility,
stress-strain behaviour, and modes of failure of the concrete cylinders. The results are also
correlated with statistical analysis using the Single Factor Analysis of Variance (ANOVA)
performed to study the significance of different environments on the ultimate strength of
the cylinders. The ANOVA results are presented in Appendix A.
3.1 Ultimate Strength
The average ultimate strengths of the cylinders of different groups are presented in Table 2.
Table 2 shows that the average axial strength of the wrapped cylinders was superior to that
57

of the unwrapped cylinders, irrespective of the environmental conditions, indicating the

efficiency of FRP sheets as strengthening materials. The axial strength of the SFRP
wrapped cylinders was always superior to that of CFRP wrapped cylinders subjected to
same environment indicating that SFRP sheet is more effective than CFRP sheet in strength
enhancement of columns.

Specimen
Designation
RT-UW
RT-C
RT-S
HT-UW
HT-C
HT-S
HTRH-UW
HTRH-C
HTRH-S
FT90-UW
FT90-C
FT90-S
FT471-UW
FT471-C
FT471-S

Table 2. Summary of Axial Strength Results

Avg. Axial
Standard
% Increase or
% Increase
Strength
Deviation
Decrease
w.r.t. UW
(MPa)
(MPa)
w.r.t. RT
58.0
0.81

64.6
1.67
11.4

104.1
2.12
79.5

46.5
4.58

-19.8
57.1
1.66
22.8
-11.6
104.0
0.94
123.7
-0.1
51.2
7.07

-11.7
62.4
1.71
21.9
-3.4
101.8
2.00
98.8
-2.2
49.8
4.94

-14.1
61.2
0.66
22.9
-5.3
96.1
7.56
93.0
-7.7
53.6
2.37

-7.6
62.6
0.90
14.6
-5.0
106.3
0.70
98.3
2.1

% Increase
from CFRP
to SFRP

61.1

82.1

63.1

57.0

69.8

3.1.1 Effect of High Temperature

Table 2 shows that exposure to prolonged high temperature (HT) decreased the strength of
the unwrapped cylinders by 19.8% compared to room temperature exposure. The change in
average strength due to exposure to high temperature is statistically significant (Appendix
A). This is likely due to the micro-cracks developed in the concrete due to high
temperature, and some bond cracks at cement-aggregate interface as the Coefficient of
Thermal Expansion (CTE) are different for cement and aggregate. Prolonged high
temperature caused statistically significant changes in the average strength of the CFRP
wrapped cylinders (Appendix A). The decrease in strength of the CFRP wrapped cylinders
subjected to prolonged high temperature was 11.6% compared to the room temperature
cylinders. The reduction in ultimate strength of CFRP wrapped cylinders may be attributed
to the differences in the CTEs of the fibre and matrix, resulting in fibre slippage and matrix
cracking; and to carbon fibre degradation induced by high temperature. However,
prolonged high temperature did not affect the strength of the SFRP wrapped cylinders.

58

3.1.2 Effect of High Temperature along with Relative Humidity

Table 2 shows that exposure to combined high temperature and relative humidity (HTRH)
decreased the strength of the unwrapped, CFRP wrapped and SFRP wrapped cylinders by
11.7%, 3.4%, and 2.2%, respectively. The ANOVA analysis shows that this exposure did
not cause statistically significant difference in strength. Visual observation before testing
revealed that the steel fibres of the SFRP sheet were corroded at different locations around
the cylinders; however, it did not have any detrimental effect on the strength.
3.1.2 Effect of Freeze-Thaw Cycles
Appendix A shows that exposure to freeze-thaw cycles caused statistically significant
change in the strength of the unwrapped cylinders. Table 2 shows that the strength of the
unwrapped cylinders decreased by 14.1 and 7.6%, respectively, due to freeze-thaw cycles.
When exposed to freezing, the water in capillary pores in concrete freezes and expands
causing a tensile pressure in the concrete. During thawing, water goes back to the capillary
pores again. When, after thawing, another freezing cycle takes place, pressure is created in
the concrete again. In this way, repetitive freeze-thaw cycles causes damage to the concrete.
Exposure to 90 freeze-thaw cycles decreased the strength of the CFRP and SFRP wrapped
cylinders by 5.3% and 7.7%, respectively. Exposure to 471 freeze-thaw cycles decreased
the strength of CFRP wrapped cylinders by 5.0%, but increased the strength of the SFRP
wrapped cylinders by 2.1%. The change in strength was not statistically significant for the
SFRP wrapped cylinders, but was statistically significant for the CFRP wrapped cylinders.
The changes in strength between FT90 and FT471 groups were not also statistically
significant (Appendix A). The two effects of freeze-thaw cycles are thermal incompatibility
and polymer embrittlement. The first effect is related to the different CTEs of matrix and
fibre, as a result of which micro cracking and void generation may occur at matrix-fibre
interface. The second effect is polymer embrittlement at low temperature which increases
the strength and stiffness of the polymer. It may reduce the effectiveness of the matrix to
transfer stresses between fibres, or between the composite and the concrete (Green et al.,
2006). While observing the specimens before test, rust was found on the steel fibres of the
SFRP sheet at different locations; however, it did not affect the strength.
3.2 Ductility
Ductility can be defined as the ability to absorb inelastic energy without losing its load
capacity (Wang et al., 2005). In this study, the ductility index is determined as the area
under the stress-strain curve up to ultimate load. The average ductility indices of the
cylinders of different groups are presented in Table 3. Table 3 shows that the average
energy ductility index of the wrapped cylinders was superior to that of unwrapped
cylinders, irrespective of the environmental exposure. The average ductility of the SFRP
wrapped cylinders was always superior to that of CFRP wrapped cylinders subjected to
same environment, indicating SFRP sheet is more ductile than CFRP sheet. Severe
environments did not cause any statistically significant change in the ductility indices of the
wrapped cylinders, indicating severe environments had no significant impact on the
ductility of the wrapped cylinders.
59

Specimen
Designation
RT-UW
RT-C
RT-S
HT-UW
HT-C
HT-S
HTRH-UW
HTRH-C
HTRH-S
FT90-UW
FT90-C
FT90-S
FT471-UW
FT471-C
FT471-S

Table 3. Summary of Ductility Index Results

Avg. Ductility
Standard
% Increase
% Increase
Index
Deviation
or Decrease
w.r.t. UW
w.r.t. RT
(MPa-)
(MPa-)
35.4

661.4
160.9
1768.4

1233.7
76.7
3385.0

20.4
3.9

-42.4
572.0
92.4
2703.9
-13.5
1236.0
145.1
5958.8
0.2
7.6
1.9

-78.5
645.1
48.7
8388.2
-2.5
1100.3
96.9
14377.6
-10.8
7.5
7.5

-78.8
707.8
707.8
9337.3
7.0
1033.7
1033.7
13682.7
-16.2
20.2
9.2

-42.9
558.5
132.8
2664.9
-15.6
1075.6
180.4
5224.8
-12.8

% Increase
from CFRP
to SFRP

86.5

116.1

70.6

46.0

92.6

3.3 Stress-Strain Response

60

1060

50

884

40

707
RT
HT
HTRH
FT90
FT491

30
20
10
0
0

2000

4000

6000

530
353
177

120

2121

100

1767

80

1414

60

1060
RT
HT
HTRH
FT90
FT471

40
20
0

0
8000 10000 12000 14000

3000

6000

9000

12000

15000

707
353

Axial Compressive Load (kN)

1237
Axial Compressive Load (kN)

70

Axial Compressive Stress (MPa)

Axial Compressive Stress (MPa)

Figure 1 shows the typical stress-strain response of the CFRP and SFRP wrapped cylinders.

0
18000

Circumferential Strain ( )

Circumferential Strain ( )

(a)
(b)
Fig. 1. Typical stress-strain response of the cylinders wrapped with (a) CFRP; (b) SFRP
Figure 1 shows that the stress-strain curves of the wrapped cylinders were not altered by the
severe environments. The curves can be characterized by three zones: (1) linear elastic zone
roughly up to strain level of 500 with slope similar to that of unwrapped cylinders,
indicating the behaviour in this zone is similar to that of unwrapped cylinders. (2) nonlinear transition zone approximately between strain level of 500-6000 (CFRP) and
500-2500 (SFRP). In this zone, after the peak stress, the curve is nearly a plateau for
the CFRP wrapped cylinders, but ascending for the SFRP wrapped cylinders. It indicates
that for the CFRP wrapped cylinders, the rate for the confining sheet to contribute to the
60

load carrying capacity is nearly equal to the rate for the concrete to lose its load carrying
capacity, but for the SFRP wrapped cylinders, the former is higher than the latter. It implies
that SFRP wrapped cylinders show better strain-strain response compared to CFRP
wrapped cylinders. (3) linear ascending zone until failure, indicating concrete is completely
crushed and FRP sheet is fully activated.
3.3 Modes of Failure
Exposure to aggressive environments did not seem to have any significant effect on the
modes of failure of the specimens. All the unwrapped cylinders failed in conventional shear
or cone. The SFRP wrapped cylinders failed either by rupture of the SFRP sheets or by
combination of rupture and debonding at the overlap. All of the CFRP wrapped cylinders
failed by rupture of the CFRP sheets. The failure of the wrapped cylinders was always
brittle and sudden, with no prior warning except for some creeping sound of concrete
crushing. Figure 2 shows the typical failure modes of the wrapped cylinders.

Debonding

Fig. 2. Typical modes of failure of the wrapped cylinders

4. CONCLUSIONS
From the compression tests and statistical analysis of the CFRP and SFRP wrapped
cylinders, the following conclusions were drawn:
FRP confinement was found to be an effective and efficient way to increase the axial
strength and ductility of columns, both at room temperature and under severe
environments. However, for the same axial stiffness of the FRP sheet, one layer of
SFRP sheet was found to be more efficient than one layer of CFRP sheet in strength and
ductility enhancement.
Aggressive environments had a detrimental effect on the axial strength of the CFRP
wrapped cylinders, but not on that of the SFRP wrapped cylinders. Severe environments
had no significant effect on the ductility of the wrapped cylinders.
The axial stress-circumferential strain curves of the CFRP and SFRP wrapped cylinders
were not altered by the presence of aggressive environments. However, cylinders
wrapped with one layer of SFRP sheet exhibit better stress-strain behaviour than the
cylinders wrapped with one layer of CFRP sheet.
Exposure to aggressive environments did not have any significant effect on the failure
mechanism of the cylinders.
The SFRP sheets were susceptible to material degradation in the form of rust under
environments where moisture was involved. However, it did not have any negative
impact on the strength. The CFRP sheets were not susceptible to such damage.

61

5. REFERENCES
1.

2.
3.
4.
5.
6.

El-Hacha, R., Green, M., and Wight, G. 1999. CFRP wrapped concrete cylinders in
severe environmental conditions. Structural Faults and Repair 99, Proceedings of the
8th International Conference and Exhibition on Extending the Life of Bridges, Civil
and Building Structures, London, England, CD-Rom, 13-15 July, 12 p.
Hardwire Composite Armor Systems. 2010a. What is Hardwire?
(http://www.hardwirellc.com/solutions/reinforcments.html).
Sika Inc. 2010b. Sikawrap Hex 230C. Product Technical Data Sheet.
(http://www.sikaconstruction.com/tds-cpd-SikaWrapHex230C-us.pdf).
Sika Inc. 2010c. Sikadur 330. Product Technical Data Sheet, High-modulus.
(http://www.sikaconstruction.com/tds-cpd-Sikadur330-us.pdf).
Green, M.F., Bisby, L.A., Fam, A.Z., and Kodur, V.K.R. 2006. FRP Confined
Concrete Columns: Behaviour under Extreme Conditions. Cement and Concrete
Composites, 28: 928-937.
Wang, H. and Belarbi, A. 2005. Flexural Behavior of Fiber-Reinforced-Concrete
Beams Reinforced with FRP Rebars. Proceedings of the 7th International Symposium
on Fibre-Reinforced Polymer Reinforcement for Concrete Structures (FRPRCS-7),
ACI Special Publications SP-230, Kansas City, USA, 2: 895-914.

APPENDIX A. STATISTICAL ANALYSIS

The null hypothesis of the single factor ANOVA is: means of two or more populations are equal;
the alternate hypothesis is: at least one of the means is different. The ANOVA was conducted on a
95% confidence level. Table 4 shows the results of the ANOVA analysis.

Unwrapped

CFRP
Wrapped

SFRP
Wrapped

Table 4. Summary of the results of ANOVA analysis

Comparison
F-Ratio
F-Critical
P-Value
ALL
2.92
4.06
0.100256
RT
HT
18.30
7.71
0.012867
RT
HTRH
2.71
7.71
0.175314
RT
FT90
8.10
7.71
0.046592
9.35
7.71
0.037762
RT
FT471
1.45
0.294677
FT90 FT471
7.71
ALL
13.41
4.06
0.001735
RT
HT
30.65
7.71
0.005203
RT
HTRH
2.69
7.71
0.176562
RT
FT90
10.72
7.71
0.030652
8.51
0.043337
RT
FT471
7.71
0.09338
0.775163
FT90 FT471
7.71
ALL
2.56
4.06
0.128223
RT
HT
0.0018
7.71
0.968319
RT
HTRH
1.92
7.71
0.238117
RT
FT90
3.12
7.71
0.152056
3.12
0.151841
RT
FT471
7.71
5.44
0.080102
FT90 FT471
7.71

62

Decision
Not Significant
Significant
Not Significant
Significant
Significant
Not Significant
Significant
Significant
Not Significant
Significant
Significant
Not Significant
Not Significant
Not Significant
Not Significant
Not Significant
Not Significant
Not Significant

CDCC-11

DURABILITY AND SUSTAINABILITY OF FIBRE REINFORCED POLYMER

(FRP) COMPOSITES FOR CONSTRUCTION AND REHABILITATION
DURABILIT ET CYCLE DE VIE DES MATRIAUX COMPOSITES
EN POLYMRES RENFORCS DE FIBRES (PRF) EN
CONSTRUCTION ET RHABILITATION

TEMPERATURE AND CURING TIME EFFECTS ON THE CREEP

RESPONSE OF FIBER REINFORCED POLYMER COMPOSITES
BONDED TO CONCRETE
Y. Jeong1, A. Jaipuriar2, M.M. Lopez1 and C.E. Bakis2
1
2

Dept. of Civil & Env. Eng. Pennsylvania State University, University Park, PA, USA
Dept. of Eng. Science and Mechanics, Pennsylvania State University, University Park, PA, USA

ABSTRACT
The objective of this investigation is to experimentally evaluate and numerically model
creep in carbon fiber reinforced polymer (CFRP) bonded to concrete and in neat epoxy.
Two test temperatures and epoxy curing times were considered. Based on the experimental
observations, an epoxy creep model was incorporated into the numerical modeling of the
time-dependent behavior of the pull-off specimens. Results indicate that the effects of
temperature and epoxy aging time on the time-dependent behavior of the FRP compositeconcrete bond are significant. These results provide critical data for characterizing the timedependent behavior of RC structures strengthened with bonded FRP composites.
1. INTRODUCTION
Fiber reinforced polymer (FRP) composites have emerged as a promising method for
rehabilitating existing reinforced concrete (RC) structures on account of their ease of
application, their corrosion resistance, and high strength-to-weight ratio [1]. While many
research activities on bonded FRP systems have been performed to improve predictive
models capable of explaining bond behavior [2], relatively few research efforts have
addressed bond performance under aggressive environmental conditions. In particular, not
much is known about long term FRP/concrete bond behavior under sustained loads. Recent
research [3] demonstrated that FRP-concrete bond behavior is affected by temperature,
sustained load and the presence of flaws at the interface. Specifically, significant creep
occurs in the epoxy when loaded within 7 days of epoxy application. Furthermore, it was
found that elevated temperature combined with sustained load led to significant bond stress
redistribution and a reduction in joint strength [3].
In the present investigation, experimental results from single shear (pull-off) specimens and
neat epoxy tensile coupons under sustained loading are obtained and used to develop finite
element models of FRP strips bonded to concrete. Two primary parameters were
63

considered: test temperature and the epoxy age following installation (curing time).
Experimental results indicate that the effects of temperature and epoxy curing time on the
time-dependent behavior of FRP composites to concrete are significant. The epoxy creep
model used in the numerical simulations captures the general trend of the CFRP timedependent deformations.
2. EXPERIMENTAL PROGRAM
2.1. Test Specimens and Materials
The concrete blocks (127 x 127 x 254 mm) and cylinders (100 mm diameter x 200 mm
long) were cast in one batch. All of the concrete specimens were cured in a moisture
chamber (95% relative humidity at 23C for 28 days). The average 28-day compressive
strength of the concrete was 33.9 MPa. A total of twelve pull-off specimens were
fabricatedthree for each test condition. Carbon FRP (CFRP) strips were attached to the
prepared concrete block according to manufacturers guidelines. A 25.4-mm-wide, single
ply of unidirectional carbon fiber strip was saturated with epoxy and bonded to each
concrete block as shown in Fig. 1. The CFRP composites bonded to concrete blocks were
cured by allowing them to cure in laboratory conditions (35% relative humidity at 20C)
for 7 or 90 days before applying sustained loading. According to the manufacturers
datasheet, full cure should be achieved after 7 days at 20C and 40% relative humidity. The
thickness of the impregnated CFRP sheet was found to be between 0.6 to 1.0 mm. Other
properties of the cured laminate and epoxy were obtained experimentally by Coronado and
Lopez [4]. Room-temperature material properties of CFRP and epoxy used in the finite
element models are shown in Table 1.
Table 1. Material properties used in the finite element models
Property
Unidirectional CFRP
Epoxy
Tensile strength
4.667 MPa
44.5 MPa
Elastic Modulus
208 GPa
2.8 GPa
Ultimate Elongation
2%
Poissons Ratio
0.39

(a)

(b)

Fig. 1. (a) Instrumentation of a pull-off specimen, (b) Fabrication of pull-off specimens.

64

2.2. Sustained Loading and Epoxy Creep Tests

The main parameters examined in the sustained loading of pull-off specimens were test
temperature during sustained loading (RT=20C, HT=30C) and the curing time of the
CFRP prior to sustained loading (7 days, 90 days), as shown in Table 2. Three replicate
specimens were tested for each combination of parameters.
Table 2. Test plan for pull-off specimens.
Specimen
Name
RT-7d
RT-90d
HT-7d
HT-90d

Curing time
(day)
7
90
7
90

Temperature during
sustained loading period
RT (20C)
HT (30C)
RT (20C)
HT (30C)

Sustained loading
period (day)
30
30
30
30

Pull-off specimens were instrumented with small stainless steel discs with machined pinholes for the use of a detachable mechanical (DEMEC) gauge. Seven target points were
fixed on the surface of the CFRP strip and nine reference points were fixed on the concrete.
By measuring the movement of each target point with respect to its corresponding reference
point, the longitudinal displacement of the CFRP strip along its length was calculated. The
accuracy of the DEMC gauge is 2.54e-3 mm. In this investigation, the creep deformation of
the CFRP strip was solely evaluated based on the variation of the longitudinal
displacements of target points D5 and D6 (Fig. 1(a)).
Sustained loading was applied to the prepared specimens as a percentage of a theoretical
debonding failure load, calculated using Cheng and Tengs bond model [5]. A sustained
load equivalent to 27% of the predicted ultimate pull-out load capacity was applied, which
corresponds to a load of approximately 2.67 kN. The sustained loading set-up consisted of a
steel frame in which the pull-off specimens were mounted. The sustained load was applied
by means of hooks and springs. Six loaded specimens were kept in room-temperature
laboratory conditions (20C) as shown in Fig. 2(a) while six others were kept at 30C in an
oven, as shown in Fig. 2(b).

(a)
(b)
(c)
Fig. 2. Creep tests(a) RT pull-off specimens, (b) HT pull-off specimens, and (c) neat
epoxy tensile coupons.
65

Neat epoxy creep tests were carried out in tension at temperatures of 20C and 30C (Fig.
(c)). Prior to testing, the epoxy specimens were aged under conditions similar to those for
pull-off specimens (room-temperature for either 7 or 100 days). Loading was set with dead
weights to provide a normal stress of 10 MPa in the epoxy. Longitudinal strains on the front
and back sides of the specimens were recorded with 25-mm extensometers and averaged to
compensate for bending. Three specimens were tested at each temperature. Further details
of the epoxy creep investigation are given in a companion paper [6].
2.3. Tests Results and Discussion
Results from the sustained loading period are presented in Figs. 3 and 4. Results are
presented as the longitudinal displacement at target points D5 and D6 as functions of time.
It can be observed that at target point D5, HT specimens exhibited larger displacements (of
the order of 0.127 mm) than the RT specimens (which remained very close to 0 mm). This
trend was very clear for all specimens regardless of the epoxy age. It appears that the
presence of a 30C temperature during sustained loading induces larger creep
displacements on the bonded CFRP strip. Target point D6 showed a similar trend for the
90-day cure specimens (with the exception of specimen HT-2). However, the 7-day cured
specimens showed a reversed trendi.e., more deformation occurs in RT specimens
compared to the HT case. Future investigation will determine the cause of the reversed
trend, such as microcracks at the bonded interface of the RT specimens.
Displacement (mm)

0.3
0.2

Displacement (mm)

HT-1
HT-2
HT-3
RT-1
RT-2
RT-3

D5

0.4

0.1
0
-0.1
0

10

20

Time (day)

30

0.3
0.2
0.1

HT-1
HT-2
HT-3
RT-1
RT-2
RT-3

D5

0.3
0.2

0
10

20

Time (day)

30

20

Time (day)

30

40

HT-1
HT-2
HT-3
RT-1
RT-2
RT-3

D6

0.1

-0.1
0

10

(b) Point D6, 7 day

Displacement (mm)

Displacement (mm)

(a) Point D5, 7 day

0.4

D6

0
-0.1
0

40

HT-1
HT-2
HT-3
RT-1
RT-2
RT-3

0.4

0.4
0.3
0.2
0.1
0
-0.1
0

40

(c) Point D5, 90 day

10

20

Time (day)

30

40

(d) Point D6, 90 day

Fig. 3. Effect of temperature on the longitudinal displacement of bonded CFRP strips with
different curing conditions.

66

It was also found that after the instantaneous elastic displacement occurs, the majority of
displacement changes occurred in the first 15 days for the HT specimens. In contrast, the
RT specimens showed that most of the change in time-dependent displacement occurred
during the first 3 days. It can be concluded that the presence of a 30C temperature results
in larger CFRP displacement due to creep, irrespective of the age of the epoxy, and it
stabilizes over a longer period of time than for specimens tested at 20C.
The influence of the epoxy curing time on the creep behavior of bonded CFRP strips can be
observed in Fig. 4. The 7-day cured specimens showed larger creep displacements than 90day cured specimens when the sustained loading was conducted at 20C. Aside from the 7day-cured, 20C-tested outliers (possible microcracking), it can be concluded that
increasing the temperature to 30C during sustained loading results higher creep
displacements and little disparity in the displacements for the 7-day and the 90-day cured
specimens. Therefore, at room temperature it can be expected that larger curing times can
have beneficial effects in reducing creep deformations. Increasing the temperature
decreases the curing time effect, but creep values could be larger in magnitude.

0.1

0.05

Displacement (mm)

Displacement (mm)

0.5

7d-1
7d-2
7d-3
90d-1
90d-2
90d-3

D5

-0.05
0

10

20

Time (day)

30

0.4
0.3

0.1

(a) Point D5, 20C

0.5

0.3
0.2

20

Time (day)

30

40

0.5

7d-1
7d-2
7d-3
90d-1
90d-2
90d-3

D6

Displacement (mm)

Displacement (mm)

0.4

10

(b) Point D6, 20C

7d-1
7d-2
7d-3
90d-1
90d-2
90d-3

D5

D6

0.2

0
0

40

7d-1
7d-2
7d-3
90d-1
90d-2
90d-3

0.1

0.4
0.3
0.2
0.1

0
0

10

20

Time (day)

30

0
0

40

10

20

Time (day)

30

40

(c) Point D5, 30C

(d) Point D6, 30C
Fig. 4. Effect of curing time on the longitudinal displacement of bonded CFRP strips
exposed to different temperatures during sustained loading.

67

3. NUMERICAL PROGRAM
3.1. Creep Constitutive Model
A power-law model was used to describe the creep behavior of the epoxy used in the CFRP
strengthening system. This model is only valid in cases where the stress state remains
essentially constant and low in relation to the yield strength of the polymer, as is assumed
for the sustained loading stage of the pull-off specimens. The time hardening form of the
power-law model can be written as follows:
=Aqntm

(1)

Based on the experimental creep test results, the uniaxial equivalent deviatoric creep strain
rate of the epoxy, (min.-1) was modeled in terms of the uniaxial equivalent deviatoric
stress, q (Pa), and time, t (min.) using Equation 1 [7]. The creep model parameters A and m
were obtained by curve-fitting the model to the experimental creep strain data (Table 3).
The stress exponent n is assumed to be 1 for linear viscoelastic behavior.

A
n
m

Table 3. Creep law parameters used in the finite element analysis.

RT-7d
RT-90d
HT-7d
HT-90d
1.2e-10
1.0e-12
4.3e-10
1.0e-12
1
1
1
1
-0.750
-0.325
-0.950
-0.370

3.2. Finite Element Modeling

Time-dependent nonlinear finite element analyses of the pull-off specimens during
sustained loading were performed using ABAQUS 6.9-2. The finite element mesh used to
model the pull-off specimen is shown in Fig. 5. The specimen was modeled using a 2-D
plane strain analysis. Concrete, epoxy, damage band (concrete-epoxy interface) and CFRP
were modeled using 4-node bilinear elements (CPE4R). Tie contacts were used between
meshes with different element sizes. Normal and shear stresses can be determined along the
entire tie interaction. A plastic damage model was used to predict the constitutive behavior
of the damage band. Details of the formulation can be found elsewhere [2]. The CFRP strip
is assumed to behave linear-elastically in tension up to the failure stress.

Fig. 5. Finite element mesh used for modeling the pull-off specimen.

68

3.3. Numerical Results

The FE model described in this paper was built assuming that creep deformation occurs
only in the epoxy. One of the objectives of the modeling effort is to determine the
suitability of this assumption for the given experimental conditions. Another objective is to
evaluate the capability of the power-law model in simulating creep of bonded CFRP strips.
Figure 6 shows the experimental and numerically simulated longitudinal displacements of
point D6 as functions of time. It can be observed that the FE model captures the general
trend of the time-dependent deformations for all temperatures and epoxy ages, except for
the 7-day cured specimens with sustained loading at 20C (RT-7d). As mentioned earlier,
additional investigation needs to be done on these specimens to determine whether or not
microcracking contributed to the anomalous results. The best prediction was found for the
7-day cured specimen with sustained loading at 30C (HT-7d).
0.5

Exp-1
Exp-2
Exp-3
FEM

0.4
0.3

Displacement (mm)

Displacement (mm)

0.5

0.2
0.1
0
0

10

20

30

Time (day)

40

0.3
0.2
0.1
0
0

50

Exp-1
Exp-2
Exp-3
FEM

0.4

10

20

(a) RT-7d
Exp-1
Exp-2
Exp-3
FEM

0.4
0.3
0.2
0.1
0
0

10

20

30

Time (day)

40

50

(b) RT-90d
0.5

Displacement (mm)

Displacement (mm)

0.5

30

Time (day)

40

0.3
0.2
0.1
0
0

50

(c) HT-7d

Exp-1
Exp-2
Exp-3
FEM

0.4

10

20

30

Time (day)

40

50

(d) HT-90d

Fig 6. Longitudinal displacement of target point D6 as function of time from experimental

(Exp) and numerical (FEM) results.
5. CONCLUSIONS

Results from this investigation indicate significant effects of curing time and temperature
during loading on the time-dependent creep behavior of CFRP composite bonded to
concrete. The higher temperature resulted in larger creep displacements for either curing
time. Longer curing time resulted in less creep deformation only in the specimen loaded at
room-temperature. The epoxy creep model used in the numerical simulations captures the
general trend of the time-dependent CFRP deformations for the temperature and curing
69

times evaluated, with the exception of the 7-day cured specimens with sustained loading at
20C. The possibility of microcracking in the latter specimens needs to be investigated in
future research.
6. ACKNOWLEDGEMENT
This paper is based upon work supported by the National Science Foundation under Grant
No. CMMI-0826461 and an REU supplemental grant. The authors would like to
acknowledge the help provided by Jeff Flood during the experimental work.
7. REFERENCES
1.
2.
3.
4.
5.
6.

7.

ACI Committee 440. 2008. Guide for the Design and Construction of Externally
Bonded FRP Systems for Strengthening Concrete Structures. ACI 440.2R-08.
American Concrete Institute, Farmington Hills, MI, 76 p.
Coronado, C.A., and Lopez, M.M. 2010. Numerical Modeling of Concrete-FRP
Debonding Using a Crack Band Approach. Journal of Composites for Construction,
ASCE, 14(1): 1121.
Gullapalli, A., Lee, J.H., Lopez, M.M., and Bakis, C.E. 2009. Sustained Loading and
Temperature Response of FRP-Reinforced-Polymer-Concrete Bond. Transportation
Research Record, 2131: 155162.
Coronado, C.A., and Lopez, M.M. 2007. Damage Approach for the Prediction of
Debonding Failure on Concrete Elements Strengthened with FRP. Journal Composites
for Construction, 11(4): 391400.
Chen, J.F., and Teng, J.G. 2001. Anchorage Strength Models for FRP and Steel Plates
Bonded to Concrete. Journal of Structural Engineering, 127: 784791.
Jaipuriar, A., Flood, J., Bakis, C.E., and Lopez, M.M. 2011. Effects of Aging on
Behavior of Epoxy used in FRP Structural Strengthening. Proc. 4th Intl. Conf.
Durability and Sustainability of Fibre Reinforced Polymer (FRP) Composites for
Construction and Rehabilitation, CDCC 2011, eds: Benmokrane, B., El-Salakawy, E.
& Ahmed, E., 20-22 Jul. 2011, Quebec City, Quebec, Canada, 8 p.
Shames, I.H., and Cozzarelli, F.A. 1992. Elastic and Inelastic Stress Analysis, PrenticeHall, New Jersey, USA, 722 p.

70

CDCC-11

DURABILITY AND SUSTAINABILITY OF FIBRE REINFORCED POLYMER

(FRP) COMPOSITES FOR CONSTRUCTION AND REHABILITATION
DURABILIT ET CYCLE DE VIE DES MATRIAUX COMPOSITES
EN POLYMRES RENFORCS DE FIBRES (PRF) EN
CONSTRUCTION ET RHABILITATION

BOND BEHAVIOUR OF GFRP BARS TO CONCRETE SUBJECTED

TO FATIGUE LOADING AND FREEZE-THAW CYCLES
J.R. Alves1, A. El-Ragaby2 and E.F. EL-Salakawy3
1

Structural Engineer, Manitoba-Hydro, Winnipeg, MB, Canada

Assistant Professor, Dept. of Civil & Env. Eng., University of Windsor, Windsor, ON, Canada
3
Associate Professor and CRC, Dept. of Civil Eng., University of Manitoba, Winnipeg, MB, Canada
2

ABSTRACT

Due to the mismatching thermal properties between FRP bars, particularly glass, and
concrete, bond of FRPs to concrete is susceptible to more damage as a result of temperature
changes. Whether FRP-reinforced concrete elements can be designed to utilize the full
tensile strength of FRP bars or not depends largely on their bond characteristics to concrete.
Currently, few to none experimental test data are available on the bond behaviour of FRP
bars in concrete elements under different loading and environmental conditions. Therefore,
this research program is designed to investigate experimentally the durability of FRP bond
to concrete elements subjected to different combinations of loading and environments. An
FRP-reinforced concrete specimen was developed to apply axial-tension fatigue or
sustained loads to GFRP bars eccentrically located within the concrete environment. Test
specimens were subjected simultaneously to the dual effects of 250 freeze-thaw cycles (FT)
and sustained load (SUS) followed by 1,000,000 fatigue load cycles. A total of six test
specimens were constructed and tested. The test parameters included bar diameter and
concrete cover. After conditioning, each test specimen was sectioned to two replicates
(halves) for pull-out test. Another series of twelve unconditioned standard pull-out
specimens were constructed and tested as control. It is concluded that freeze-thaw cycles
along with sustained load followed by fatigue loads resulted in significant decrease in the
bond strength compared to the unconditioned specimens.
1. INTRODUCTION
Glass fibre-reinforced polymer (GFRP) bars are more attractive to the construction industry
due to its lower cost compared to other types of FRP materials. In North America, bridge
deck slabs and parking garages are prime examples where GFRP reinforcements are
currently being used. These elements are continuously subjected to wide combinations of
fatigue and/or sustained load from traffic as well as thermal loads due to freeze-thaw cycles
and temperature fluctuations. Such environmental and loading conditions may affect the
mechanical properties of the concrete itself, the FRP reinforcement, and interface or bond
between them. For reinforced concrete structures, the transfer of stresses between the
71

concrete and the reinforcement, both at the serviceability and at the ultimate limit states, is
strongly dependent on the quality of bond. Therefore, bond of FRP reinforcing bars to
concrete elements subjected to thermal and mechanical loads might be a critical design
factor that needs to be investigated.
Research work done to date on the bond characteristics of FRP bars in concrete investigates
mainly the bond strength under static/monotonic load. Currently, few experimental test data
is available regarding the effects of fatigue loading cycles with constant and variable stress
ranges, impact loads as well as environmental effects such as, freeze-thaw cycles, wet-dry
cycles, high and low temperatures. The bond behaviour of FRP bars is influenced by many
factors such as adhesion, bar diameter, bar surface pattern and shape. Weber (2005) and
Esfahani et al. (2005) noticed the increase in bond strength with the increase of concrete
cover up to a value of 2.0db, where db is the bar diameter. Tepfers et al. (1998) and Galati
et al. (2005) found that concentric placement of the bar reduced the bond strength by
approximately 30% to 50% compared to eccentric specimens with 2.0db concrete clear
covers. On the other hand, Benmokrane et al. (1996) and Achillides and Pilakoutas (2004)
reported that larger size bars produce smaller bond stresses and they also lose their adhesive
bond earlier than small size bars (in case of sand-coated GFRP bars). Therefore, in order to
achieve the same bond stress, larger bar diameters need longer embedment lengths.
Furthermore, different loading conditions also were found to affect the bond quality of FRP
bars to concrete. Bakis et al. (1998) investigated the influence of cyclic load on the bond
behaviour of GFRP-reinforced beams. FRP bars with difference surface patterns were
investigated. The authors reported that the cumulative slip increased with increase in the
peak load or with increase in the number of cycles regardless of the surface pattern.
Moreover, Katz (2000) observed 19% to 80% reduction in the bond strength of GFRP in
concrete specimens depending on the surface conditions. The GFRP bars with smooth
surface showed the highest loss in bond strength, while bars containing sand-coating and
helical wraps presented the lowest loss. He and Sun (2006) found that pullout specimens
with GFRP bar containing helically winded glass strands subjected to fatigue loading
showed up to 40% smaller bond resistance and up to 6 times higher peak slip than those
without any loading history.
2. THE EXPERIMENTAL PROGRAM
2.1 Test Specimens
This experimental program is divided into two series: (C) Control Static pullout and (I)
Freeze-thaw cycles along with sustained loads followed by fatigue loading. One of the main
challenges of this research was to develop a test specimen that meets the following
requirements:
Apply axial tension sustained or fatigue loading on FRP bars eccentrically located
within concrete with different concrete cover thickness;
Sustain large number of fatigue load cycles without premature failure of the anchor;
Can be moved easily and placed inside the environmental chamber for the
environmental conditioning;
Allow to perform monotonic pull-out test after conditioning.

72

An FRP-reinforced concrete specimen for Series I was developed to fulfill these

requirements. The specimen consists of three concrete blocks: the test portion (middle
block) and the two end anchor blocks. The three portions were connected together by the
GFRP bar as shown in Fig. 1. The middle block had 200-mm square cross section to meet
the requirement of standard pull-out test specified by the Canadian Standard CSA-S806-02
(CSA-2002). The length of the middle block was 405-mm such that after conditioning, each
specimen was split into two 200-mm cube replicates for pullout test (considering a 5-mm
clearance for the cutting saw). The GFRP bar was always eccentric in the middle block,
while the end anchors were always adjusted to be concentric with respect to the GFRP bar.
Each anchor contained six 16 mm-diameter threaded rods placed perpendicular to the axis
of the GFRP bar to apply the specified loads.
Series C contained 12 unconditioned control specimens to assess the bond characteristics of
the used GFRP bars as well as to provide a datum for the other test specimens using a
standard procedure provided by the CAN/CSA-S806-02 (CSA 2002). The control specimen
consists of a 200-mm concrete cube with a 5db embedment length with eccentric placement
of the bar. Steel pipes were used as anchors, and were cast with expansive grout prior to the
casting of the concrete specimens.

(a) Plan view of test specimen

(b) Photo for the constructed test specimen

Fig. 1. Details of test specimen for Series I

The test parameters included the GFRP bar diameter and the concrete cover. Two bar
diameters, No.16 and No. 19 GFRP bars (15.9 and 18.9 mm in diameter, respectively) were
investigated. Also, three different clear concrete covers of 1.5 db, 2.0 db, and 2.5 db were
studied. These parameters resulted in having a total of 12 specimens in Series I and II; six
specimens each. In each series, the test parameters were bar diameter and concrete cover
where each group was subjected to different conditioning scheme. Table 1 gives the details
of the test program.
73

2.2 Materials Properties

The specimens were constructed using normal weight concrete with an average 28-day
compressive strength of 50 MPa. Since the concrete was cast on several stages, it was
mixed in the McQuade Heavy Structures Laboratory at the University of Manitoba using a
mechanical mixer to obtain consistent concrete properties. At least twelve 100 by 200 mm
and six 150 by 300 mm cylinders were prepared for each concrete batch and cured under
the same condition as their reference specimen to evaluate the compressive and tensile
strength of concrete at the test date.
Table 1. Test matrix and details of the experimental program
Specimens

Name*

Sustained load

Freeze-thaw cycles

Fatigue load

C.16.15
C.16.20
C.16.25
No
No
No
conditioning
conditioning
conditioning
C.19.15
C.19.20
C.19.25
S.FT.F.16.15
S.FT.F.16.20
1,000,000 cycles
S.FT.F.16.25
Series I
30% of GTS
250 Cycles
at 1.5 Hz and 25%
S.FT.F.19.15
of GTS
S.FT.F.19.20
S.FT.F.19.25
Notes: GTS is the guaranteed tensile strength of the GFRP bar.
* C refers to the control specimens; S stands for sustained load, FT freeze-thaw cycles, F
fatigue load, the first two numbers represent the bar diameter (No.16 or No.19), the following
two numbers stands for the concrete clear cover (1.5, 2.0 or 2.5db). So, for example,
S.FT.F.16.15 stands for a specimen subjected to sustained load and freeze-thaw cycles followed
by fatigue loading, considering bar diameter No. 16 mm and concrete clear cover of 1.5db.
Series C

The GFRP bars used in this study (V-RODTM) had a sand-coated surface and were made of
continuous longitudinal fibres impregnated in a thermosetting vinyl-ester resin with a fibre
content of 73% by weight (Pultrall Inc. 2007). Table 2 lists the mechanical properties of
the reinforcing bars as determined by tensile tests on representative sample in accordance
with CAN/CSA-S806-02 (CSA 2002).
2.3 Loading and Environmental Conditioning Schemes
Specimens of series I were subjected to the dual effects of sustained load and freeze-thaw
cycles simultaneously followed by fatigue load cycles. A sustained load of 30% of the
guaranteed tensile strength of the GFRP bar (GTS) was selected (CSA 2002).
Consequently, the applied sustained loads were 42 and 54 kN for No.16 and No. 19 GFRP
bars, respectively. Moreover, the fatigue loading consisted of repeated tensile stress cycles
(tension-tension) with peak loads of 25% of the GTS according to the limits permitted by
the CAN/CSA-S6-06 (CSA 2006). Therefore, the peak tensile load for a fatigue cycle was
74

35 and 45 kN for No.16 and for No. 19, respectively. The lowest tensile load was set to 5
kN for all fatigue cycles. The specimens were conditioned to 1,000,000 cycles in a
sinusoidal wave.
Table 2. Properties of GFRP bars
Bar
Size

Bar
Diameter
(mm)

No.16
No.19

15.9
18.9

Tensile
Modulus of
Elasticity
(GPa)
48.2
47.6

Ultimate
Tensile
Strength
(MPa)
751
728

Ultimate
Tensile
Strain
(%)
1.56
1.53

Coefficient of Thermal
Expansion (at 10-6/oC )
Longitudinal
6.4
6.0

Transverse
29.1
29.0

The freeze-thaw conditioning was designed and applied in accordance with ASTM standard
specification C66603 (ASTM 2003). According to this standard, to simulate the freezethaw effect, the temperature inside the concrete at the level of reinforcement should ramp
between -18 and +4oC allowing for 67% of the cycle duration under freeze conditions.
After several attempts, it was concluded that ramping the environmental chamber
temperature from -25oC to +15oC and maintain the freezing period (at -25oC) for five hours
and the thawing period (at +15oC) for 3 hours satisfied the ASTM C666-03 requirement. A
total of 250 cycles were applied in approximately 4 months at a rate of 2 cycles/day. The
relative humidity was forest to 50% during freezing and 95% during thawing.
2.4 Test Set-up and Instrumentations
The sustained load was applied to the specimens of Series I through a prestressing bed. A
set of steel plates were used to apply the load. Steel angles and two 38-mm diameter thread
rods were used to hold the anchors in place, thus sustaining the load. To avoid relative
movement between concrete and the steel angles, nuts were used in both sides of the
angles. This mechanism allows moving the specimen in and out of the environmental
chamber as well as adjusting the applied load by turning the locking nuts toward or against
each other to maintain constant sustained load. Figure 2 shows a photo for the sustained
load test set-up. A total of twelve 6-mm long electrical strain gauges were attached to the
GFRP bars at the middle of the test block. During prestressing, the load was monitored by a
load cell, while the strains in the GFRP bars were monitored through the strain gauge.
Moreover, thermocouples were attached to the GFRP bars in two specimens: S.FT.F.16.15
and S.FT.F.19.25, which represented the specimens with the smallest and largest concrete
covers. These thermocouples were attached to verify the temperature inside the specimens
during the freeze-thaw cycles.
A universal 1000-kN MTS machine was utilized to apply either the fatigue cyclic loading
or the monotonic pull-out loading. The fatigue load cycles were applied to the test
specimens of Series I in sinusoidal wave at rate of 1.5 Hz under load-controlled conditions.
The monotonic pull-out load was also applied at a load-controlled rate of 10 kN/min in
accordance with the standard test method of CAN/CSA-S806-02 (CSA 2002). A high
accuracy LVDT was used to record the free-end slip of GFRP bars during the pull-out test.
Figure 2 shows details of the fatigue loading and pull-out tests set-up.

75

(a) Sustained Load

(b) Freeze-Thaw Conditioning

(c) Fatigue loading

(d) Pull-out test

Fig. 2. Details of test set-up for different loading and environmental conditioning
TEST RESULTS AND OBSERVATION

In this section, test results regarding bond strength and slip measurements are presented in
terms of bond stress versus free-end slip for different conditioning schemes. The term slip
represents the relative displacement between the concrete block and the GFRP bar. The
bond stress was considered constant along the embedded length. Such assumption is
acceptable since a short embedment length of 5db was used.
2.5 Effect of Conditioning
Figures 3 and 4 show typical bond-slip relationship obtained from the pull-out tests carried
out on the conditioned specimens. For specimens in Series I, all the conditioned test
specimens showed degradation in bond properties between the GFRP bars and concrete.
This degradation was observed in terms of decrease of the bond strength, increase of the
slip, or both. The bond strength decreased by about 20 to 50% for specimens S.FT.F.19.25
and S.FT.F.16.25 with significant increase in slip in the former specimen when compared
to their unconditioned counterparts. It is clear that the fatigue loading largely degraded the
bond resistance of sand-coated GFRP bars.
2.6 Effect of Concrete Cover
Figures 5 shows a comparison of the bond-slip relationship of test specimens of Series I
with different concrete cover thickness, which seems to play an important role. From Fig. 5,
it is clear that the degree of decrease in the bond strength due to combined freeze-thaw,
sustained loading and fatigue conditioning depends on the concrete cover. Increasing the
concrete cover by about 33% (from 1.5db to 2.0db) and 25% (from 2.0db to 2.5db) increased
the residual bond strength (i.e. decrease the loss in bond strength) by about 15 and 25%,
respectively. On the other hand, all control specimens showed constant increase of about
15% in the bond strength when increasing the concrete cover from 1.5db to 2.0db or from
2.0db to 2.5db. This can be explained as follows. Increasing the concrete cover increases the
confinement pressure on the bar which improves the bond strength. Also, cover thickness
plays an important role in decreasing the temperature gradient effects at reinforcement level
which reduces the thermal effects due to temperature fluctuation. However, it can be
concluded that the effect of fatigue loading is more dominant than that of the freeze-thaw

76

conditioning so that the residual bond strength after conditioning is less than that of the
counterpart control specimen.

12

12

Bond Stress (MPa)

15

Bond Stress (MPa)

15

9
6
Series C
(Control

9
6
Series C (Control)
Series I (S.FT.F)

3
0

0
0

0.04

0.08
0.12
0.16
Free-End Slip (mm)

0.2

0.04

0.08
0.12
0.16
Free-End Slip (mm)

0.2

Fig. 4. Effect of different conditioning

schemes on the bond-slip behaviour of GFRP
bar No. 16

Fig. 3. Effect of different conditioning

schemes on the bond-slip behaviour of
GFRP bar No. 19

Figure 6 shows comparisons of the bond strength of GFRP bars to concrete after
conditioning. It can be noticed that all test specimens of Series I experienced decrease of
bond strength after conditioning compared to their counterparts of Series C. Also, it is
noticeable that the larger the bar diameter, the larger the loss in bond strength after
conditioning. This may be due to the effect of thermal stresses on the expansion of the bar
cross section, which is more pronounced for the larger bar diameter (No. 19). Therefore, the
largest cover, 2.5 db, appears to have resisted the thermal loads and was capable of
providing significant increase in bond resistance. For smaller covers (1.5 and 2.0db), the
enlargement in the bar diameter may have caused damage in the surrounding concrete.
From these figures, it can be concluded that using concrete cover of 2db and 2.5db resulted
in the highest bond strengths for GFRP bars No. 16 and No. 19, respectively, either for the
control or the conditioned specimens.
However, the ISIS Design Manual No 3 (ISIS Canada 2007) and CAN/CSA-S6-06 design
code (CSA 2006) recommend the use of the clear covers of 2.5db for external exposure with
a minimum value of 35 10 mm. The CAN/CSA-S806-02 code (CSA 2002) recommends
clear covers of 3.5db with a minimum value of 40 mm for prestressed elements. Compared
to test results, these provisions are not in good agreement with test results for No. 19 mm
GFRP bars subjected to freeze-thaw cycles and fatigue loads. The ACI 440.1R-06 (ACI
Committee 440) only provides a maximum limit of 3.5db for the clear concrete cover.

77

19.15
16.15

12

19.20
16.20

15

19.25
16.25

Bond strength (MPa)

Bond Strength (MPa)

15

Series C (control)
Series I (S.FT.F)

10

9
6
3

5
0

0
0

0.04

0.08
0.12
0.16
Free-End Slip (mm)

0.2

Fig. 5. Effect of concrete cover on the bondslip behaviour of GFRP bars subjected to
Freeze-Thaw cycles, sustained load, and
fatigue

16.15 16.2 16.25 19.15 19.2 19.25

Test Specimens

Fig. 6. Comparison of the bond strength of

GFRP bars to concrete after being
conditioned under different conditioning
schemes

2.7 Mode of Failure

In this section, the mode of failure was analyzed. Five types of failure could be observed in
this experimental program as shown in Fig. 7. Tastani et al. (1984) reported also similar
failure modes for bond failure of FRP bars. The following presents brief description of the
modes of failure obtained for the three test series:
Pure Pullout Failure; where the bar is pulled out of the specimen without any
splitting/cracking in the concrete.
Concrete Cover Failure; where a crack in the smallest cover is suddenly formed with
immediate loss of bond resistance.
V-Shaped Concrete Cover Failure; in this case, it is thought that a cover failure has
already occurred. The final failure is observed with the v-shaped concrete failure,
Diagonal Failure; probably due to the lack of homogeneity of the concrete, a failure
which should follow the V-shape, deviated and cracked the cover in the lateral direction,
Concrete Block Split; where a crack, which started in the smallest cover, propagates
through the whole concrete block.
In Series C, it was observed that by increasing the concrete clear cover, the failure mode
changed. For specimens with No. 16 bars, the clear cover of 2.0db was enough to ensure the
pure pullout failure, while for No. 19 bar the same concrete clear cover (2.0db) was not
enough to prevent cracking. The concrete deterioration due to freeze-thaw conditioning is
visible in both the specimens and cylinders. The concrete compressive and tensile strength
of the cylinders that were kept inside the environmental chamber during the freeze-thaw
conditioning were 15% less than those of the control ones. In this case, the decrease in the
concrete strength also contributed to the splitting mechanism. For Series I, half of the
specimens failed in pure pullout and half failed by cracking. In some specimens, which
failed by concrete block splitting, the surface print of the sand-coating layer on the concrete

78

interface could be clearly seen This observation indicates that significant deterioration
occurred in the sand-coating layer due to the fatigue loading.

(a) Pure pullout

failure

(b) Concrete
cover failure

(c) V-Shaped
concrete failure

(d) Diagonal concret (e) Concreteblock

splitfailure
cover failure

Fig. 7. Different modes of bond failure

5. CONCLUSIONS
Based on the experimental results, the following conclusions can be drawn:
1. In Series C (control specimens), unconditioned specimens presented very stiff
behaviour with no slip up to failure. It can be observed that by increasing the concrete
clear cover, the failure mode changed. For specimens with No. 16 bars, the clear cover
of 2.0db was enough to ensure the pure pullout failure, while for No. 19 bar the same
concrete clear cover (2.0db) was not enough to prevent concrete cracking.
2. Investigating the results of Series I (freeze-thaw cycles and sustained load followed by
fatigue loading), it can be concluded that the effect of fatigue loading is more dominant
than the freeze-thaw cycles, since all specimens in Series II showed significant loss in
bond strength of at least 23%. For specimens with No. 16 bars, it seems that the effect
of fatigue loading counteracts that of the freeze-thaw cycles. While for specimens with
No. 19 bars and covers of 1.5 and 2.0db, the adverse effects of both conditionings are
added together.
3. It seems that a concrete clear cover of 2.0db is an optimum for sand-coated GFRP bars.
However, if subjected to the effect of freeze-thaw cycles, bar No. 19 presented 25%
higher bond resistance than unconditioned specimens for the larger concrete cover of
2.5db.
6. ACKNOWLEDGEMENTS
The authors wish to express their gratitude and sincere appreciation for the financial
support received from the Natural Science and Engineering Research Council of Canada
(NSERC), through Canada Research Chairs program. Also, the equipment provided by a
Canada Foundation for innovation (CFI) grant is greatly appreciated. The help received
from the technical staff of the McQuade Heavy Structural Laboratory in the department of
civil engineering at the University of Manitoba is also acknowledged.

79

7. REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.

Achillides, Z. and Pilakoutas, K. 2004. Bond behaviour of fibre reinforced polymer

bars under direct pullout conditions. ASCE, Journal of Composites for Construction, 8,
No. 2: 173-181.
Bakis, C., Al-Dulaijan, S., Nanni, A., Boothby, T. and Al-Zahrani, M. 1998. Effect of
cyclic loading on bond behaviour of GFRP rods embedded in concrete beams.Journal
of Composites Technology and Research, JCTRER, 20, No. 1: 29-37.
Benmokrane, B., Tighiouart, B. and Chaallal, O. 1996. Bond strength and load
distribution of composite GFRP reinforcing bars in concrete. ACI Materials Journal,
93(3): 246-253.
CSA. 2002. Design and construction of building components with fibre reinforced
polymers, CAN/CSA-S806-02, Canadian Standards Association, Toronto, Canada.
CSA. 2006. Canadian highway bridge design code, CAN/CSA-S6-06, Canadian
Standards Association, Toronto, Canada.
Galati, N., Nanni, A., Dharani, L., Focacci, F. and Aiello, A. 2005. Thermal effects on
bond between FRP rebars and concrete. Elsevier Composites: Part A: 1223-1230.
He, Z. and Sun, Y. 2006. Experimental study on bond behaviour of GFRP rod in
concrete subjected to stress reversal. Federation Internationale du Beton, FIB, Session
10. ID 10-29.
ISIS Canada. 2006. Durability of fibre reinforced polymers in civil infrastructure,
Technical Report, The Canadian Network of Centres of Excellence on Intelligent
Sensing for Innovative Structures, ISIS, Winnipeg, Manitoba, Canada.
ISIS Canada. 2007. Reinforcing concrete structures with fibre reinforced polymers,
ISIS-Manual No.3, version 2, The Canadian Network of Centres of Excellence on
Intelligent Sensing for Innovative Structures, ISIS, Winnipeg, Manitoba, Canada.
Katz, A. 2000. Bond to concrete of FRP rebars after cyclic loading. ASCE, Journal of
Composites for Construction, 4(3): pp. 137-144.
Koller, R., Chang, S. and Xi, Y. 2006. Fibre-reinforced polymer bars under freezethaw cycles and different loading rates. Journal of Composite Materials, 41(1): 5-25.
Mashima, M. and Iwamoto, K. 2004. Bond characteristics of FRP rod and concrete
after freezing and thawing deterioration. ACI Structural Journal, 138: 51-70.
Pulltral Inc. 2007. V-ROD-Product guide specification. Pultral inc, Thetford Mines,
Quebec, Canada.
Shahidi, F., Wegner, L. and Sparling, B. 2006. Investigation of bond between fibrereinforced polymer bars and concrete under sustained loads. Canadian Journal of Civil
Engineering, 33:.1426-1437.
Tastani, S., Pantazopoulou, S. and Karvounis, P. 1984. Local bondslip characteristics
of GFRP bar. ACI Structural Journal, 2(230):1481-1496.
Tepfers, R., Hedlund, G. and Rosinski, B. 1998. Pull-out and tensile reinforcement
splice tests with GFRP bars. Second International Conference on Composites in
Infrastructure, ICCI-98, Tucson, Arizona, 14p.
Won, J. and Park, C. 2006. Effect of environmental exposure on the mechanical and
bonding properties of hybrid reinforcing bars for concrete structures. Journal of
Composite Materials, 40(12): 1063-1074.

80

CDCC-11

DURABILITY AND SUSTAINABILITY OF FIBRE REINFORCED POLYMER

(FRP) COMPOSITES FOR CONSTRUCTION AND REHABILITATION
DURABILIT ET CYCLE DE VIE DES MATRIAUX COMPOSITES
EN POLYMRES RENFORCS DE FIBRES (PRF) EN
CONSTRUCTION ET RHABILITATION

DURABILITY CHARACTRISTICS OF NEW GFRP DOWELS FOR

CONCRETE PAVEMENT
M. Montaigu1, M. Robert2 and B. Benmokrane3
1

Master Student, Department of Civil Eng., University of Sherbrooke, Quebec, Canada

Postdoctoral Fellow, Department of Civil Eng., University of Sherbrooke, QC, Canada
3
Professor, Department of Civil Engineering, University of Sherbrooke, Quebec, Canada
2

ABSTRACT
Dowel bar commonly used for load transfer in concrete pavement slab is traditionally made
from steel. However, once the steel dowel bar corrodes, it may cause faults, such as binding
due to lockout of the dowel bar in concrete pavement, level differences resulting from
spalling or decreased efficiency of load transfer. To solve this problem, many studies on the
performance and feasibility of FRP dowel bars are being undertaken. In particular, glass
fibre-reinforced polymer (GFRP) dowel bar is now considered as a potential solution.
However, GFRP dowels could be damaged by the capillary water of concrete, which
typically has high alkalinity. This study is conducted through a research collaboration
project between the University of Sherbrooke and the Ministry of Transportation of Quebec
(MTQ, Pavement Division). One of the main objectives of this research project was to
evaluate the durability characteristics of new generation of GFRP dowel bars manufactured
using vinylester resin, in simulated environmental conditions experienced by concrete
pavement. An accelerated aging test method was applied to evaluate the long-term
durability of the GFRP dowel bar. The results of this research have contributed to introduce
this type of GFRP dowel bars with enhanced durability characteristics for concrete
pavement applications in the province of Quebec.
1. INTRODUCTION
FRP composites, mainly based on thermoset polymers and glass or carbon fibres, are being
used in infrastructures exposed to harsh conditions involving de-icing salts or marine
environments. A typical dowel bar is more vulnerable to corrosion when it is used to
transfer loads in concrete pavement due to the penetration of water by the joint opening [1].
When corrosion occurs in a dowel bar, freezing could be caused by the expansion, followed
by fractures due to the curling of the concrete pavement slab. Ministry of Transportation of
Quebec (MTQ) has encountered these types of problem with steel reinforced jointed
concrete pavement. More than 350 km of jointed concrete pavement have been rebuilt since
1994 in the province of Quebec. To minimise the corrosion problems, the MTQ is actually
using epoxy-coated steel dowels of 25.1 to 38.1 mm in diameter in the design of new
81

concrete pavement slabs of 225 to 325 mm in thickness. To solve corrosion problems of

metallic dowels (conventional steel, epoxy-coated, etc.), GFRP dowel bars could be
considered now as a feasible solution [2, 3] with high industrial productivity [4, 5].
The MTQ is looking to improve the long-term performance, durability, and safety of
jointed concrete pavement using GFRP dowels. However, long-term durability
performance of GFRP dowels, such as in high alkalinity environment, needs to be
investigated [6].
This study investigates the durability of GFRP dowels using accelerated ageing alkaline
tests at different temperatures and period of time. GFRP specimens were immersed at
different temperatures (up to 60 C.) in alkaline cement Portland solution for a period of 180
days, then tested under transverse direct shear. Based on prediction modelling, long-term
transverse shear strength of the tested GFRP dowels was determined.
2. EXPERIMENTAL PROGRAM
2.1 Material
GFRP dowels made of continuous high strength E-glass fibres impregnated in vinylester
(VE) resin using the pultrusion process were used in this study. Durability tests were
performed on several diameters from 25.4 to 44.5 mm. The initial properties (reference) of
the GFRP dowels were determined by preliminary tests. As there was no effect of the
diameter on the degradation of dowels (less than 2%), the durability study reported in this
paper only presents results for 34.9 and 38.1 mm-diameter dowels. Note that these two
diameters are the most employed diameters in roads infrastructure. The reference
mechanical and physical properties of the tested 34.9 and 38.1 mm-diameter GFRP dowels
are summarized in Table 1.
All GFRP dowels were cut into 300 mm lengths so that the direct shear test can be
performed using ACI 440.3R-04. The bars were divided into two series; 1) the
unconditioned reference samples; 2) conditioned samples immersed in alkaline solution.
2.2 Conditioning of the GFRP Dowels
Accelerated ageing tests were conducted according to the method DBT-2 Recommended
FRP Dowel Bar Test Protocol, proposed by Market Development Alliance of the
Composite Institute [7], in accordance with the ASTM C 581 standard. Cement extract
solution, which is a real representative solution of concrete pores water, was prepared by
mixing commercial Portland cement type 30 (highest rate of CaO) with tap water to get a
pH of 12.60, measured at the beginning of the test and maintained constant throughout the
conditioning period with Ca(OH)2 and cement. Specimens were immersed at three different
temperatures (23, 50 and 60C). Specimens were place in wood containers waterproofed
with a high density polyethylene film and hermetically closed to avoid excessive
evaporation and changes of the pH solution (Figure 1). For each diameter and resin type of
dowel, six specimens were removed from solution and tested under short beam shear tests
after 30, 60 and 180 days at 23, 50 and 60C of conditioning to measure the mechanical
properties of aged dowels.
82

Mechani

Physical

Table 1. Mechanical and physical properties of 34.9 and 38.1 mm diameter GFRP dowels
Property
VE-based GFRP dowels
34.9 mm
38.1 mm
Fibre content (%)
80.7
80.6
Cure ratio (%)
100
100
Tg (C)
124
123
Moisture uptake (%)
0.06
0.07
Relative density
2.12
2.14
-6
-1
LTE long. (x10 C )
7
7.6
-6
-1
LTE transv. (x10 C )
23.5
24
Direct shear strength (MPa)
184
173
Short beam shear strength (MPa)
61.1
53.9
4 points flexural strength (MPa)
1210
1077
Flexural modulus of elasticity (GPa)
50.3
51.6

Fig. 1. GFRP dowels ready for conditioning

2.3 Direct Transverse Shear Tests
Direct shear tests were conducted to characterize the GFRP dowels according to ACI
440.3R-04 guide Test Method B.4. The setup consists in a 230 x 100 x 110 mm steel base
equipped with lower blades spaced of 50 mm face to face, allowing the double direct shear
of the specimen by an upper blade, as shown in Figure 2. Six unconditioned specimens of
300 mm long were tested as reference for each type of dowel at the laboratory conditions
using MTS 810 testing machine equipped with a load cell of 500 kN. The displacement
speed was 1.5 mm/min giving between 30 and 60 MPa/min, until the failure of the
specimen. The direct shear strength is given by the equation 1.
P
(1)
u s
2A
Where u is the direct shear strength (MPa); Ps is the failure load (N); and A is the specimen
section (mm).
83

Fig. 2. Direct shear test setup

2.4 Prediction of Long-Term Behavior
Arrhenius concept [8] was used for the prediction of long-term behaviour of GFRP dowels.
The equation 2 expresses the Arrhenius relation, in terms of the degradation rate [9].
k Ae

Ea
RT

(2)

Where k is the degradation rate; A is a constant relative to the material and degradation
process; Ea is the energy of activation of the reaction; R is the universal gas constant; and T
is the temperature in C. The primary assumption of this model is that only one dominant
degradation mechanism of the material operates during the reaction and that this
mechanism will not change with time and temperature during the exposure.
2.5 Differential Scanning Calorimetry (DSC)
Twelve-milligram to 15-milligram specimens from both unconditioned and aged samples
were sealed in aluminum pans and analyzed in a TA Instruments DSC Q10 calorimeter
equipped with a refrigerated cooling system. Analysis was conducted in modulated DSC
mode. Specimens were heated from 25C to 195C at a rate of 5oC/min. Glass transition
temperature was determined by DSC for each specimen in accordance with ASTM D 1356
standard. Two scans were performed for each specimen. The first scan is useful to
determine the difference of Tg between reference and conditioned specimens. If a decrease
of Tg is observed for conditioned samples, this is an indication of plasticizing effect or
chemical degradation. The second scan gives information about the mechanism of
degradation and if it is irreversible.

84

2.6 Fourier Transform Infrared Spectroscopy (FTIR)

Fourier transform infrared spectroscopy (FTIR) spectra were recorded using a Nicolet
Magna-550 spectrometer equipped with an attenuated total reflectance device. Fifty scans
were routinely acquired with an optical retardation of 0.25 cm to yield a resolution of 4
cm1. 25.4 mm diameter dowels were investigated after 180 days conditioning at 60C.
2.7 Scanning Electronic Microscopy (SEM)
Scanning electron microscopy (SEM) observations and image analysis were performed to
observe the microstructure of vinylester-based dowels specimens before and after aging in
alkaline solution during 180 days at 60oC. All specimens observed in the SEM were first
cut, polished, and coated with a thin layer of gold-palladium by a vapor-deposit process.
After coating the surfaces, microstructural observations were performed on a JEOL JSM840A SEM. These observations were conducted to see the potential degradation of polymer
matrix, glass fibers, or interfaces, if any.
3. EXPERIMENTAL RESULTS
3.1 Direct Shear Strength
Table 2 shows the direct shear strength of vinylester-based dowels after different
conditionings in alkaline solution. Figure 3 shows the retention of the direct shear strength
after 180 days of ageing of the GFRP dowels at various temperatures. As shown in the
Figure 3, vinylester-based dowels were slightly affected by ageing in alkaline solution with
a retention of more than 90% of the initial shear performance after 180 days at 60C in
alkaline solution (pH=12.60). All specimens have kept elastic behaviour until the shear
failure of the fibres (Figure 4). It can be seen that the degradation rate between 23C and
50C is nearly the same that between 50C and 60C, characterizing the exponential effect
of the temperature on mechanisms of degradation.
Table 2. Direct shear properties of aged GFRP dowels
Time of
Temperature
Mean Shear
Standard deviation
o
immersion (days)
( C)
Strength (MPa)
(MPa)
35 mm 38 mm
35 mm
38 mm
0
23
184.3
173.4
1.81
3.19
23
182.1
177.7
2.67
1.24
30
50
180.0
175.7
2.43
2.54
60
178.1
171.0
3.65
1.87
23
180.3
181.9
1.97
1.93
60
50
175.7
177.6
4.17
1.84
60
172.4
168.6
1.85
0.86
23
178.9
176.5
1.87
2.14
180
50
172.7
169.8
3.11
0.73
60
167.5
157.7
3.65
3.27

85

Fig. 3. Direct shear strength retention of GFRP dowels after conditioning in alkaline
solution at 60oC

a)

b)
Fig. 4. Typical mode of failure of 34.9 mm GFRP dowels tested under shear: a)
longitudinal view, b) cross-sectional view
3.2 Effect on Polymer Matrix
Table 3 gives the values of Tg before and after aging in alkaline solution during 180 days at
60oC. No significant effect of aging was observed on vinylester-based dowels. It can also
be seen that no shift of Tg was measured during the second run, leading to the conclusion
that the initial cure ratio of the vinylester resin was very high.
A FTIR analysis of unconditioned dowel and specimens aged in alkaline solution during
180 days at 60oC was conducted (Figure 5). The most interesting region of the FTIR
spectra is located between 3300 cm-1 and 3600 cm-1, which corresponds to the stretching
mode of the hydroxyl groups of the vinylester resin. When hydrolysis reaction occurs, new
hydroxyl groups are formed and the corresponding infrared band increases. Changes in the
peak intensity are quantified by determining the ratio of the OH- peak to the carbonhydrogen stretching peak of the resin, which is not affected by the conditioning. The
experimental ratios of the OH peak to the carbon-hydrogen stretching peak of the core and
86

the surface of vinylester-based dowel immersed in alkaline solution for 180 days at 60oC
were 0.44 and 0.54, respectively, compared to 0.49 for unconditioned samples. The
hydroxyl peak did not show any significant changes. This observation lead to the
conclusion that no chemical degradation of the vinylester resin occurred during the
immersion of the dowels in alkaline solution at 60C for 180 days.
Table 3. Tg of reference and aged dowels
Tg run 1 (oC)
35 mm
38 mm
Reference
123
124
o
180 days in alkaline solution at 60 C
122
123
Conditioning

Tg run 2 (oC)
35 mm
38 mm
124
123
123
122

Fig. 5. FTIR spectra for unconditioned and aged samples

3.3 Microstructural Observations
3.3.1 External Surface
SEM observations of external surface of dowels were performed to investigate the surface
deterioration of the polymer matrix after conditioning in alkaline solution. Figure 6 presents
micrographs of the surface of reference and aged dowels.

a)
b)
Fig. 6. Micrographs of the dowels surface for: a) 34.9 mm VE reference dowel; b) 34.9 mm
VE dowel aged in alkaline solution during 180 days at 60oC
87

It can be observed that no degradation of the vinylester matrix has occurred after 180 days
at 60C aging in alkaline solution. No increase of the number or dimensions of the pores
was observed and the surface remains intact without any cracking or microcracking. No
dimension or weight changes of the dowels have been noticed after the accelerated ageing
in alkaline solution.
3.3.2 Micro-structural Effects
Figure 7 presents micrographs for both reference specimens and specimens aged in alkaline
solution during 180 days at 60oC. The visual and microstructural observations showed no
significant damage on vinylester-based dowel after 180 days of immersion in the alkaline
solution at the highest temperature (60oC). Observations of the fibre/matrix interface and of
the microstructure, in general, demonstrate that the conditionings of vinylester-based dowel
in alkaline solution do not affect the microstructural properties of the GFRP dowel (Figure
7b).

(a)
(b)
Fig. 7. Micrographs at the fibre/matrix interface before mechanical tests for: a)VE reference
dowel, b) 34.9 mm VE dowel aged in alkaline solution during 180 days at 60oC
3.4 Service Life Prediction
Long-term predictions of mechanical properties can be made according to the method
DBT-2, based on shear strengths obtained after 30, 60 and 180 days at three temperatures
of conditioning. Simulations are proposed for the vinylester-based dowels of 34.9 mm
diameter. Following the procedure proposed by Bank et al.[10], the natural logarithm of
time to reach a set of levels of normalized performances versus 1/T, expressed as the
inverse of absolute temperature (1000/K), was used to predict the service life at the Mean
Annual Temperature (6.2oC) in Montral, Qubec, Canada. A coefficient of determination
(R2) value close to 1 is desired. However, the ASTM procedures recommend a minimum
value of 0.80 for acceptability and the obtained R2 values are between 0.96 and 0.99.
Predictions are made for direct shear strength retention as a function of time for an
immersion at 6.2C and the general relation between the PR and the predicted service life at
the average temperature of 6.2oC are drawn (Figure 8). It can be seen from Figure 8 that
vinylester-based dowels present a very high durability in concrete pavement environment.
In fact, the predicted service life of vinylester-based dowels immersed in alkaline solution
88

at an isotherm temperature of 6.2oC to reach a PR less than 90% can be estimated to be

infinite.

Fig. 8. Long-term predictions for 34.9 mm dowels

4. CONCLUSIONS
Based on the results of this study, the following conclusions may be drawn:
- The tested GFRP vinylester-based dowels present a very high durability after 180
days of accelerated ageing in alkaline solution at 60C. Shear strength retention of
90 % was measured after 180 days of immersion in alkaline solution at 60oC.
- The microstructural analysis has shown that no damage has occurred to the internal
microstructure or at the fibre/matrix interface for vinylester-based GFRP dowel.
- According to the long-term predictions, the shear strength retention of vinylesterbased GFRP dowels decreased by less than 10% after 200 years. This result clearly
demonstrates the high stability of vinylester-based GFRP dowels in concrete
pavement environment. For steel reinforced concrete pavement, the ACPA
estimates that steel dowel bars can fail in as little as 7 to 15 years depending on
design and location due to corrosion problems.
5. ACKNOWLEDGEMENTS
The authors would like to express their special thanks and gratitude to the Ministry of
Transportation of Quebec, the Natural Science and Engineering Research Council of
Canada (NSERC), the Fonds quebecois de la recherche sur la nature et les technologies
(FQRNT), Pultrall Inc. (Thetford Mines, Qubec) for donating the GFRP dowels, and the
technical staff of the structural lab of the Department of Civil Engineering at the University
of Sherbrooke.
89

6. REFERENCES
1.

Mauricio, M., Cruz, C.J., Jieying, Z.; Harvey, J.T., Monteiro, P.J.M.; Abdikarim, A.
2005. Laboratory evaluation of corrosion resistance of steel dowels in concrete
pavements. Pavement Research Center, Institute of Transportation Studies, University
of California: Davis, Berkeley.
2. Eddie, D., Shalabi, A., and Rizkalla, S. 2001. Glass-Fiber-Reinforced Polymer Dowels
for Concrete Pavements. ACI Structural Journal, 98(2): 201-206.
3. Porter, M. 2002. Assessment of dowel bar research. Final Report, CTRE, Department
of civil and construction engineering, Iowa State University, Ames, USA, 81 p.
4. Highway Innovative Technology Evaluation Center (HITEC). 1998. HITEC evaluation
plan for Fiber reinforced polymer composite dowel bars and stainless dowel bars. Ohio
Department of Transportation, Ohio, USA.
5. Abo-Qudais, S.A. and Al-Qadi, I.L. 2000. Dowel bars corrosion in concrete pavement.
Canadian Journal of Civil Engineering, 27: 1240-1247.
6. Won, J., Cho, Y. and Jang, C. 2006. The Durability of Glass Fibre-Reinforced Polymer
Dowel after Accelerated Environmental Exposure. Polymers & Polymer Composites,
14(7): 719-730.
7. Litherland, K.L., Okley, D.R., and Proctor, B.A. 1981. The use of accelerated aging
procedures to predict the long term strength of GRC composites. Cement and Concrete
Research, 11: 455-466.
8. Market Development Alliance (MDA). 1998. Recommended FRP Dowel Bar
Durability Test Protocol. Dowel Bar Team 2 of the SPI Composites Institute, Harrison,
USA, 26 p.
9. Nelson, W., 1990. Accelerated testingStatistical models, test plans, and data
analyses. Wiley, New York, 601 p.
10. Bank, L.C, Gentry, T.R, Thompson, B.P, and Russel, J.S. 2003. A Model Specification
for Composites for Civil Engineering Structures. Construction and Building Materials,
17(6-7): 405-437.

90

CDCC-11

DURABILITY AND SUSTAINABILITY OF FIBRE REINFORCED POLYMER

(FRP) COMPOSITES FOR CONSTRUCTION AND REHABILITATION
DURABILIT ET CYCLE DE VIE DES MATRIAUX COMPOSITES
EN POLYMRES RENFORCS DE FIBRES (PRF) EN
CONSTRUCTION ET RHABILITATION

EVALUATION OF THE DURABILITY OF CFRP SHEETS AFTER CLIMATIC

EXPOSURE
I. Nishizaki1, P. Labossiere2, K. W. Neale2, M. Demers3 and T. Tomiyama4
1

Team Leader, Public Works Research Institute, Tsukuba, Japan

Professor, Universite de Sherbrooke, Canada
3
Research Engineer, Universite de Sherbrooke, Canada
4
Senior Researcher, Public Works Research Institute, Tsukuba, Japan
2

ABSTRACT
Structural reinforcement with CFRP sheets is a widely used method for strengthening
concrete structures such as girders, decks and piers. While the efficiency of this method is
well documented, there are still questions about its long term durability. Most of the
original studies on the long-term behaviour of CFRP have been done for aircraft or marine
structures. However, many conditions such as materials for these applications,
environmental factors and expected life time are different than those in structural civil
engineering. The authors carried out a series of exposure tests under natural climates
conditions for CFRP sheets in three characteristic locations in the world, Sherbrooke
(Canada), Tsukuba (mainland Japan) and Oogimi (Okinawa). The natural exposure tests
were continued for ten years and the mechanical properties of the specimens from the three
locations, such as tensile strength and in-plane shear, were evaluated at regular interval.
Although the results showed a slight decrease in mechanical properties over the exposure
period, the tested materials exhibit properties that still allow their use after ten years.
1. INTRODUCTION
The use of composite materials is becoming an increasingly popular method of repairing
and strengthening ageing concrete structures around the world. While the efficiency of this
method is well documented, there are still questions about its long term durability. Most of
the previous studies on the durability of CFRP have been done for aircraft or marine
structures [1-4], and many conditions such as materials, environmental factors and expected
life time are different for structural civil engineering applications. The authors carried out a
series of exposure tests under natural climates conditions for CFRP sheets in three
characteristic locations in the world, Sherbrooke (Canada), Tsukuba (mainland Japan) and
Oogimi (Okinawa). Intermediate results were reported previously [5-7], however the
natural exposure tests were originally planned for ten years, and this report shows some
results from the ten year exposed specimens.

91

2. METHODS
2.1 Specimens
The results from two types of specimens are presented herein as follows:
(1) Laminate 0
Table 1 provides a summary of the specimens. Two types of continuous CF fiber sheet
(products A and B) supplied by different manufacturers were used for the specimens. Both
CF fiber sheets were PAN type. Matrix resins were selected from the recommended
products of each manufacturer (an epoxy resin in both cases). Four-ply CFRP
unidirectional laminates, 250 x 300 mm in size with the fibers in the longitudinal direction,
were prepared with each set of materials. Additional laminates with top coating
(acrylurethane resin) were also prepared to evaluate the protective effect of the latter. The
purpose of Laminate 0 specimens was to evaluate the durability of the fiber itself, hence
after exposure, it was tested in tension in the longitudinal direction.

250

Laminate45-A
Laminate45-B
Laminate45-AC

Direction for 1 st & 4 th layer

Laminate system
4-ply, UD

4-ply, [45/-45]s for the

testing direction

Direction for 2 nd & 3 rd layer

Code
Laminate 0-A
Laminate 0-B
Laminate 0-AC

Table 1. Summary of the specimens

Materials
PAN type CF & Epoxy resin (product A)
PAN type CF & Epoxy resin (product B)
PAN type CF & Epoxy resin (product A)
Top coating (Acrylurethane resin)
PAN type CF & Epoxy resin (product A)
PAN type CF & Epoxy resin (product B)
PAN type CF & Epoxy resin (product A)
Top coating (Acrylurethane resin)

250

250

Fig. 1. Outline of Laminate45

92

Unit: mm

(2) Laminate 45
Figure 1 shows the general configuration of Laminate45 specimens. Tensile tests were to be
performed after exposure in the 45-degree direction, hence this specimen has a [45/-45]s
configuration. The purpose of these specimens was to evaluate how much the shear
properties of the laminates would be affected by the condition between fiber and matrix
resin. Table 1 also provides a summary of Laminate45. The three type of the specimens
prepared as Laminate45 were identical in configuration to the Laminate0 series.
Five identical specimens were prepared for each laminate configuration and for each
exposure site. One set of additional specimens was also tested before exposure, in order to
obtain reference values for each configuration.
In addition to the above specimens, three other types of specimens were also prepared for
the exposure test program. They are configuration (3) to (5), as follow:
(3) CF-sheets: CF sheet without matrix resin
(4) Matrix resin plates: Matrix resin plate used for the laminates
(5) Concrete columns: Concrete columns reinforced with laminate-A or B
The results for these specimens are not reported herein.
2.2 Exposure Test
The exposure tests were initiated in 1997 at the three locations identified above. Table 2
shows the main climatic characteristics of each location.
Table 2. Characteristics of the exposure sites
Exposure locations Latitude Annual mean Annual
mean Climate
temperature
rain fall
Sherbrooke
4537N 4.1C
1084mm
Cold and covered
(Quebec, Canada)
with snow in winter
Tsukuba
3567N 13.6C
1505mm
Moderate climate
(Ibaraki, Japan)
Oogimi
2648N 22.4C
2036mm
Subtropical climate,
(Okinawa, Japan)
close to sea shore
The laminate specimens were installed vertically on steel exposure racks using an
aluminum frame. Five specimens of each type were placed in the each of the three exposure
sites. Exposed specimens were scheduled to be recovered after one, three, five, seven and
ten years of exposure. Fig.2, 3 and 4 show the specimens undergoing exposure under
natural conditions at the three exposure sites.

93

Fig. 2. Specimens under the exposure test (Sherbrooke)

Fig. 3. Specimens under the exposure test (Tsukuba)

Fig. 4. Specimens under the exposure test (Oogimi)

94

2.3 Evaluation Methods

All the specimens were recovered according to the planned schedule. Five coupons were
cut out from each retrieved specimen and mechanical properties were examined with the
coupons. The evaluation methods were as follows.
- For Laminate0: tensile test- ASTM D 3039-76 Standard Test Method for Tensile
Properties of Fiber-Resin Composites
- For Laminate45: in-plane shear test- ASTM D 3518M-91 Standard Practice for In-Plane
Shear Stress-Strain Response of Unidirectional Polymer Matrix Composites
3. RESULTS
3.1 Differences between Materials
Fig.5 shows the average tensile properties measured from Laminate0 specimens while Fig.6
shows the results of shear properties from Laminate45 specimens.
1.2
Tensile modulus/Initial values

Tensile strength/Initial values

1.2
1.0
0.8
0.6
0.4
Laminate0-A

0.2

Laminate0-B
2

10

0.8
0.6
0.4
Laminate0-A

0.2

Laminate0-B

0.0
0

1.0

0.0

12

Exposure years

6
8
Exposure years

10

12

a) Tensile strength
b) Tensile modulus
Fig. 5. Tensile properties of the Laminate0 specimens after exposure variation between
products (Average of the three exposure sites)
In-plane shear modulus/Initial values

In-plane shear strength/Initial values

1.2
1.0
0.8
0.6
0.4
Laminate45-A

0.2

Laminate45-B

0.0
0

6
8
Exposure years

10

1.2
1.0
0.8
0.6
0.4
Laminate45-A

0.2

Laminate45-B

0.0
0

12

6
8
Exposure years

10

12

a) In-plane shear strength

b) In-plane shear modulus
Fig. 6. In-plane shear properties of the Laminate45 specimens after exposure variation
between products (Average of the three exposure sites)
95

The tensile strength of the laminates did not show significant decrease even after ten years
of exposure. On the other hand, the shear strength of the laminates showed a slight
decrease. The decrease of shear strength was almost totally attained after three years of
exposure and remained quite stable afterwards. This suggests that the carbon fiber itself has
a good durability after ten years of exposure. However the interface between matrix resin
and carbon fiber appears to show a slight deterioration. The difference between products A
and B is not clear, however Laminate-B seems slightly more sensitive to deterioration than
Laminate-A. The changes in modulus for each case seem to be more significant than
changes in strength, but the difference was not so remarkable in either cases.
3.2 Effects of the Exposure Conditions
Fig.7 and 8 show the results of tensile and shear properties measured from specimens of the
three exposure sites. The data shown are the average values of uncoated specimens
(Laminate0-A and B).
1.2
Tensile modulus/Initial values

Tensile strength/Initial values

1.2
1.0
0.8
0.6
0.4

Sherbrooke
Tsukuba

0.2

Oogimi
2

6
8
Exposure years

10

0.8
0.6
Sherbrooke

0.4

Tsukuba
0.2

Oogimi

0.0

0.0
0

1.0

12

10

12

Exposure years

a) Tensile strength
b) Tensile modulus
Fig. 7. Tensile properties of the Laminate0 specimens after exposure differences between
exposure sites (Average values of the two products)
In-plane shear modulus/Initial values

In-plane shear strength/Initial values

1.2
1.0
0.8
0.6
Sherbrooke

0.4

Tsukuba
0.2

Oogimi

0.0
0

6
8
Exposure years

10

12

1.4
Sherbrooke

1.2

Tsukuba

1.0

Oogimi

0.8
0.6
0.4
0.2
0.0
0

6
8
Exposure years

10

12

a) In-plane shear strength

b) In-plane shear modulus
Fig. 8. In-plane shear properties of the Laminate45 specimens after exposure differences
between exposure sites (Average values of the two products)
96

The results suggest that the deterioration in Oogimi generally seems to be slightly larger
than in other locations. However the difference may not be significant. Considering the
difference between climatic conditions at the exposure sites, the effect of the exposure
condition seems not to be strong in the tested cases. However the tested cases were carried
out in three locations in the world, the condition was not still comprehensive, hence the
strength of the effect of the environmental condition should be evaluated under wider
variations in the future.
3.3 Protective Effect of the Top Coating
Fig.9 and 10 show the protective effect of the top coating. Each data point is the average
value of three exposure sites for Laminate0-A and Laminate0-AC.
1.2
Tensile modulus/initial values

Tensile strength/initial values

1.2
1.0
0.8
0.6
0.4
Laminate0-A
0.2

Laminate0-AC

1.0
0.8
0.6
0.4
La minate0-A

0.2

La minate0-AC

0.0

0.0
0

6
8
Exposure years

10

12

10

12

Exposure years

1.2

In-plane shear modulus/Initial values

In-plane shear strength/Initial values

a) Tensile strength
b) Tensile modulus
Fig. 9. Tensile properties of the Laminate0 specimens after exposure - Effect of the top
coating (Average values of the three exposure sites)

1.0
0.8
0.6
0.4
Laminate45-A
0.2

Laminate45-AC

0.0
0

10

1.2
1.0
0.8
0.6
0.4
La minate45-A
0.2

La minate45-AC

0.0
0

12

Exposure years

6
8
Exposure years

10

12

a) In-plane shear strength

b) In-plane shear modulus
Fig. 10. In-plane shear properties of the Laminate45 specimens after exposure tests - Effect
of the top coating (Average values of the three exposure sites)
There is a previous study that reports that a top coating has a good protective effect against
deterioration of mechanical properties of pultruded GFRP [8]. However, Fig.9 a) and
Fig.10 a) show that the coating does not provide an effective protection for CFRP laminates
97

and does not prevent strength degradation. Fig.9 b) and Fig.10 b) suggest that a top coating
may prevent a modulus reduction; however, this effect seems to be very small.
4. CONCLUSIONS
Outdoor exposure tests were conducted for ten years in three locations in Canada and Japan
for two types of laminates with continuous CF sheets, and the changes of the tensile and inplane shear properties were evaluated.
The tensile strength of the laminates did not show significant decreases even after ten years
of exposure while, on the other hand, the shear strength of the laminates showed a small
decrease. This result suggests that the carbon fiber itself has a good durability after ten
years of exposure. However, the interface between matrix resin and carbon fiber shows a
slight deterioration. The difference between equivalent products from two manufacturers
was not clear.
Despite the important differences in climatic conditions between the exposure sites, it did
not correlate with differences in strength and modulus between specimens.
The anticipated protective effect of a top coating against the reduction of mechanical
properties was not observed for the CFRP laminates tested, whereas such an effect had
clearly been found for GFRP in a previous study.
Overall, although the results showed a slight decrease in mechanical properties over the
exposure period, the products exhibit properties that still allow their use after ten years.
5. ACKNOWLEDGEMENT
This work was supported in Japan by JSPS KAKENHI (21360209) and, until 2009, by the
ISIS Canada Network of Centres of Excellence.
6. REFERENCES
1.
2.
3.
4.
5.

6.

Backer, D. J. 1984. Flight Service Evaluation of Composite Components on Bell 206L

and Sikorsky S-76 Helicopters. Journal of American Helicopter Society, 29 (2): 3-11.
Dexter, H. B. and Backer, D. J. 1984. Worldwide Flight and Ground-based Exposure
of Composite Materials. NASA Conference Publication, NASA-CP-2321, pp. 17-49.
Yamaguchi, T., Momoshima, Y. and Shirota, T. 1998. Long-term outdoor exposure
test for Carbon Fiber Composites, Reinforced Plastics, 44 (2): 56-62. (in Japanese).
Kudo, R., Ben, G., Hojo, H. and Okubo, H. 1999. Durability and undestructive
prediction of bending strength of CFRP under accelerated exposure condition. Journal
of Japanese Society for Composite Materials, 25 (1): 23-29. (in Japanese).
Labossiere, P., Neale, K. W. and Nishizaki, I. 2003. Effect of Different Long-term
Climatic Conditions on FRP Durability, Proceedings of the Sixth International
Symposium on FRP Reinforcement for Concrete Structures (FRPRCS-6), ed. Tan,
K.H., Singapore, 8 10 July, 2: 779-784.
Nishizaki, I., Labossiere, P., and Sarsaniuc, B. 2005. Durability of CFRP sheet
reinforcement through exposure tests. Proceedings of the 7th International Symposium
98

7.

8.

on Fiber Reinforced Polymer Reinforcement for Reinforced Concrete Structures

(FRPRCS-7), ACI SP-230, Kansa City, MO, USA, 5-9 November, pp. 1419-1428.
Demers, M, Labossiere, P., Nishizaki, I., Sarsaniuc, B. and Neale, K.W. 2007.
Durability of CFRP sheets under natural climatic conditions. Proceedings of the Third
International Conference on Durability and Field Applications of Fibre Reinforced
Polymer (FRP) Composites for Construction (CDCC2007), eds. Benmokrane, B. & ElSalakawy, E., Quebec City, Quebec, Canada, 22-24 May, pp. 151-158.
Nishizaki, I., Kishima, T., and Sasaki, I. 2007. Deterioration of mechanical properties
of pultruded FRP through exposure tests. Proceedings of the Third International
Conference on Durability and Field Applications of Fibre Reinforced Polymer (FRP)
Composites for Construction (CDCC2007), eds. Benmokrane, B. & El-Salakawy, E.,
Quebec City, Quebec, Canada, 22-24 May, pp. 159-166.

99

100

CDCC-11

DURABILITY AND SUSTAINABILITY OF FIBRE REINFORCED POLYMER

(FRP) COMPOSITES FOR CONSTRUCTION AND REHABILITATION
DURABILIT ET CYCLE DE VIE DES MATRIAUX COMPOSITES
EN POLYMRES RENFORCS DE FIBRES (PRF) EN
CONSTRUCTION ET RHABILITATION

DURABILITY OF EXTERNALLY BONDED FRP COMPOSITES:

COMPARISON OF ACCELERATED TEST METHODS WITH REAL
TIME EXPOSURE
H.R. Hamilton1, C.W. Dolan2 and J.E. Tanner3
1

Assoc. Prof., Dept. of Civil & Coastal Eng., Univ. of Florida, USA
Prof., Dept. of Civil and Arch. Eng., Univ. of Wyoming, USA
3
Assoc. Prof., Dept. of Civil and Arch. Eng., Univ. of Wyoming, USA
2

ABSTRACT
The durability of CFRP composite reinforcement externally bonded to concrete is affected
by moisture exposure. This paper provides a comparison of accelerated aging of externally
bonded CFRP composites with that of similar specimens exposed to the environment.
Small beam test specimens (100 x 100 x 350 mm) were constructed with bonded CFRP
reinforcement and exposed to both accelerated conditioning using heated water and to real
time weathering. Two extreme environments were chosen for weathering. One was in a
tidal zone in coastal brackish water in Florida, which resulted in wetting and drying. The
other was in Laramie, WY, which exposed the specimens to extreme temperatures, UV, and
freeze-thaw. The accelerated conditioning protocols were submerged exposure in 60C (140
F) water for 60 days and 100% relative humidity at 60 C (140 F) for 60 days. The
specimens were then tested in three-point bending (flexure) to determine the loss in
capacity due to the exposure.
1. INTRODUCTION
The strength of CFRP composites bonded to concrete deteriorates when exposed to
hygrothermal conditions. The deterioration occurs principally on the bond plane, which is
the interface between the epoxy adhesive and the concrete. The exact mechanism of the
deterioration is not well known; however deterioration is accelerated under elevated
temperature, immersion in water or sustained loads. Tests on a several CFRP systems
indicated that the procedures in this report were capable of differentiating a wide range of
responses of deterioration of CFRP systems subjected to the test procedures.
Environmental exposure conditions and test procedures assure uniform reporting and
evaluation of durability strength reduction factors to field application environments. The
methodology presented in this report allows engineers to obtain standardized comparisons
of alternative CFRP systems, apply the strength reduction factors to design, and to make
decisions on the applicability of a CFRP system for a given environment.
101

2. RESEARCH PROGRAM
The research reported in this paper was part of a larger research project examining the
durability of externally bonded CFRP composites. As part of the research a literature
review was conducted regarding the various potential mechanisms of degradation that
might affect externally bonded composites. It was found that moisture is the most likely
candidate. For bond critical applications, moisture degradation of the bond line can be the
most serious degradation mechanism. Consequently, the bulk of the research program
focused on the development of a simple, convenient bond test and the accelerated
conditioning protocols that can be used to evaluate the susceptibility of a CFRP system to
moisture.
The recommended accelerated conditioning is divided into two primary environments: Wet
and Air. Wet environment calls for test specimens to be submerged in 60 deg. C water for
60 days. Air environment calls for exposure to 100% relative humidity at 60 deg. C for 60
days. (Hamilton et al. 2011)
Along with extensive accelerated conditioning testing, real time outdoor testing was also
conducted. This paper details the results of the real time exposure and compares the results
to those of the accelerated conditioning. Two real time exposure conditions were selected
to compare to the accelerated conditioning. One was in a tidal zone in brackish water in
Florida. The other was on the roof of a building in Wyoming. The Florida exposure
involved wetting and drying of the specimens with little solar exposure while the Wyoming
specimens were exposed to freeze-thaw and solar. Specimens constructed and tested at the
Univ. of Wyoming are denoted as Group 1 and those constructed and tested at the Univ. of
Florida are denoted as Group 2.
3. SPECIMEN CONSTRUCTION
The test specimen is a 100 mm x 100 mm x 356 mm (4 in. x 4 in. x 14 in.) beam, tested in
three-point bending (flexure) on a 305-mm (12-in.) span length (Fig. 1). The specimen is
prepared by providing a full-width half-depth saw cut approximately 2.5 mm (0.1 in.) wide
at midspan. Further details of the development of this specimen configuration can be found
in Gartner et al. (2011). Application of a CFRP system and surface preparation is in
accordance with NCHRP Report 514 (Mirmiran et al. 2004).
To determine if the proposed test procedures could differentiate resistance to the
accelerated aging, two - unidirectional wet-layup CFRP systems (denoted as A and B) and
one unidirectional carbon laminate system (denoted as C) were selected to reinforce the
specimens. Systems A, B, and C were commercially available CFRP composites in which
the fiber and epoxy were supplied jointly. All of these systems consisted of bisphenol A
epoxies with diamine hardeners.
Composite A was a unidirectional carbon fiber fabric and a two-component epoxy resin
with a mixture ratio 2.9:1.0 by mass. The system also included a sealant to provide a
protective coating. To assess the effectiveness of the coating, Group 1 specimens did not
apply the coating and Group 2 specimens used the coating.
102

Composite B consisted of unidirectional carbon fiber fabric, epoxy primer, epoxy putty,
epoxy saturant, and protective top coat.
Composite C consisted of a pre-cured unidirectional carbon fiber polymer laminate and
epoxy putty. Group 1 used a 25 mm (1-in.) wide strip and Group 2 used a 19 mm (-in.)
wide strip to assess the effects on concrete flexure-shear failure of the concrete specimen.

(a)
(b)
Fig. 1. Small beam bond test (a) test setup and (b) externally bonded composite
reinforcement
4. TIDAL CONDITIONING
4.1 Exposure Conditions
With the cooperation of the Florida Department of Transportation (FDOT), 29 specimens
were hung from the northeast fender of the SR 206 Bridge in Crescent Beach, Florida (Fig.
2). The bridge crosses the Matanzas River, which is a brackish water river on the Atlantic
Intracoastal Waterway.

SR206 Bridge

Orlando

Miami

Fig. 2. SR206 bridge in Crescent Beach, FL

Six concrete beams were hung from each fender beam using stainless steel cable and
clamps (Fig. 3). As indicated by the staining and barnacle growth on the fender system, the
beams were held at such elevation as to result in regular wetting and drying during natural

103

tidal fluctuations. Although the FRP composite reinforcement was oriented upward, the
specimens were protected from direct sunlight for most of the daylight hours by the fender
system. This prevented direct UV exposure and large temperature swings from solar
heating. Beam tests were conducted after 12 and 18 months continuous exposure.

Fig. 3. Typical real time beam installation

4.2 Results and Discussion
Twelve months after real time tidal exposure was initiated, 9 beam specimens (3 of
composite A, B, and C) were selected at random to be tested in flexure. All specimens
were retrieved by boat and upon arrival at fender system it was noted that the overall
location and orientation of each beam specimen remained relatively unchanged.
Additionally, it was observed that as a result of being exposed to tidal action all of the
beams had experienced significant barnacle growth (Fig. 4). The highest densities of
barnacle growth were seen on the bottom and side surfaces of each beam specimen. Despite
the large number of barnacles present along the perimeter of each CFRP strip, the strips
themselves experienced very little barnacle growth.

Fig. 4. Typical specimen layout after tidal exposure

After being removed from the bridge supports, the tidal specimens remained dry until
reaching the testing destination. At the testing location all remaining stainless steel support
strands were removed with bolt cutters and the specimens were submerged into a room
temperature, fresh water tank. After 24 hours of fresh water submersion, barnacles were
removed from the concrete substrate with a scraper blade. Care was taken to avoid
damaging the CFRP reinforcing strip. Immediately after removing the excess barnacles,
beam specimens were returned to the room temperature, fresh water tank. After an
104

additional 24 hours of fresh water submersion all nine specimens were tested in flexure to
failure.
After 18 months the remaining 20 beam specimens were removed from the tidal exposure
location to be tested in flexure. At this time there was no noticeable change in the density
of barnacles present on the surface of each concrete specimen compared to the 12 month
tidal exposure specimens. The same cleaning and testing procedures as for the 12-month
tests were used.
Test results after 12 and 18 months of tidal exposure are shown in Fig. 5. The plot presents
the average strength ratio for each parameter. The strength ratio is the ratio of the ultimate
strength of the conditioned specimens to that of the unconditioned specimens made with the
same materials and tested at the same age as the conditioned specimen. System A and B
both showed a decrease in strength after 12 months of 22% and 14%, respectively. System
C, however, actually showed a slight increase in strength after 12 months of exposure.
1.4

Strength Ratio

1.2
1.0
0.8
0.6
0.4

System A

0.2

System B
System C

0.0
0

100

200

300

400

500

600

Exposure Time (days)

Fig. 5. Strength ratios for tidal exposure

The initial reduction in strength of Systems A and B was followed by an increase in
strength over the following 6 months. In fact, System B increased to a strength ratio of
1.04 after 18 months of real-time exposure. After 18 months, the strength ratio of System
C increased further to 1.23. This increase might be due to the buildup of barnacles and
their associated marine adhesive, which protected specimens from deterioration and acted
as additional adhesive or sealant (El-Hawary et al. 2000). Furthermore, the relatively short
exposure time at these mild temperatures was insufficient to result in significant
degradation. The failure mode for all three systems was primarily adhesive with some
mixed-mode failure modes noted in System B.
Modulus of rupture beams were exposed with the FRP system beams for 18 months. The
strength ratio of these concrete beams without CFRP was an average of 1.01, indicating that
the concrete tension strength changed very little as a result of the tidal conditioning.
Consequently, the variations in strength ratios were all a result of changes in the CFRP
composites.

105

5. FREEZE-THAW AND UV CONDITIONING

5.1 Exposure Conditions
Sixty CFRP specimens (20 each of System A, B, and C) were placed on the roof of a six
story building in Laramie, Wyoming (Fig. 6). CFRP direct tension coupons were also
fabricated and exposed along with the beams. The local weather is harsh with high
ultraviolet radiation as well as cold and snow contributing to freeze-thaw and moisture
conditions in the winter. Generally, due to the high elevation, the winters are long with
relatively short spring, summer and fall seasons. No UV protection coating was used on
these specimens to evaluate the effect. Beam and tension tests were conducted after 6, 12,
and 18 months.

Fig. 6. Exposure in Wyoming

5.2 Results and Discussion
Freeze-thaw and UV conditioning was conducted in a real time environment that combined
UV, moisture variation and ambient freeze-thaw cycles present at an elevation of 7200 ft.
The flexural test results of CFRP Composite A, B and C Systems after 18 months are
shown in Fig. 7.
1.2

Strength Ratio

1
0.8
0.6
0.4

CFRP System A

0.2

CFRP System B
CFRP System C

0
0

100

200

300

400

500

Exposure Time (days)

Fig. 7. Strength ratios for solar exposure

106

600

The test results from CFRP System A, B and C solar exposure indicated the following:
System A flexural strength decreases sharply in the first 200 days. The specimens
failed by a mixed mode failure after 6 months and adhesive failure after 12 months.
System B flexural strength decreased sharply in the 200 days. The flexural bond
strength degraded up to 28% and specimens failed by a mixed mode failure after 6
months. At 12 months and beyond, the flexural strength tended to stabilize.
System C flexural test results showed linear strength reduction with time. The
specimens failed by mixed mode failure after 6 months. Adhesive failure mode
occurred after 12 months, and CFRP laminate slip was observed after 18 months.
6. COMPARISON WITH ACCELERATED CONDITIONING
Fig. 8 shows a comparison of the real time exposure results (Roof and Bridge) with results
from accelerated conditioning. The plot gives the average strength ratios for the three
tested CFRP systems. The accelerated conditioning consisted of submerging the small
beam specimens in 60 deg C heated water (60C-S) for the time indicated in the plot
followed by flexural testing to failure. In addition, specimens were subjected to the same
temperature at 100% relative humidity for 60 days (60C-RH). Group 1 indicates specimens
constructed at Univ. of Wyoming and Group 2 indicates specimens constructed at Univ. of
Florida. Details of the accelerated conditioning and testing can be found in Hamilton et al.
2011.
CFRP System

1.4

C
60 days
6 months
12 months
18 months

Strength Ratio

1.2
1
0.8
0.6
0.4
0.2

G2|Bridge

G1|Roof

G1|60C-RH

G2|60C-S

G1|60C-S

G2|Bridge

G1|Roof

G1|60C-RH

G2|60C-S

G1|60C-S

G2|Bridge

G1|Roof

G1|60C-RH

G2|60C-S

G1|60C-S

Conditioning

Fig. 8. Summary of average strength ratios. (G1Group 1, G2Group 2, 60C-S

submerged, 60C-RH100% relative humidity)
The drastic difference in behavior of System C between submerged accelerated
conditioning and wetting and drying is puzzling. The specimens degraded severely after
immersion in water at elevated temperatures. Further, the failure mode in the accelerated
conditions was that of delamination of the composite itself. Yet, relatively cool brackish
water (less than room temperature) did not cause degradation of System C. In fact, the
107

strength of the system increased over the real time exposure period. This appears to
indicate a significant influence of temperature on the precured laminate that was not seen in
the other wet layup systems.
Overall, the freeze-thaw/UV exposure results have consistently lower strength ratios than
those of the tidal exposure specimens. This may be an indication that the combination of
UV degradation along with the cyclic thermal loading due to differences in the coefficient
of thermal expansion (CTE) of the laminate, adhesive, and concrete may be a more
significant mode of deterioration than wetting and drying in brackish water under relatively
mild ambient temperatures. The diurnal temperature changes and the CTE differences
result in a residual stress between CFRP and the concrete that may be sufficient to affect the
bond strength.
The final recommendations of the research project were to submerge the specimens in 60
deg. C water for 60 days to evaluate externally bonded CFRP systems for durability. While
this approach does not necessarily cover all possible environments and conditions,
comparison of the results of the accelerated conditioning with that of the real time exposure
indicates that the recommended protocol will provide conservative results. Longer periods
of real-time exposure, however, are needed to verify that the accelerated conditioning
captures the long-term moisture degradation at ambient temperatures.
7. ACKNOWLEDGEMENTS
The research reported in this paper was sponsored by the American Association of State
Highway and Transportation Officials in cooperation with the Federal Highway
Administration and was conducted by the National Cooperative Highway Research
Program, which is administered by the Transportation Research Board of the National
Academies.
8. REFERENCES
1.
2.

3.

4.

Gartner, A., Douglas, E.P., Dolan, C.W., Hamilton, H.R. 2011. Small Beam Bond Test
Method for CFRP Composites Applied to Concrete. Journal of Composites for
Construction, ASCE, 15(1): 52-61.
Hamilton, H.R., Dolan, C.W., Tanner, J.E., and Douglas, E.P. 2011. Testing Protocol
for Bonded CFRP Durability. Proceedings 10th International Symposium on FiberReinforced Polymer Reinforcement for Concrete Structures, FRPRCS-10, eds: Sen, R.,
Seracino, R., Shield, C. & Gold, W., 2-4 April, Tampa, FL, 20 p.
Mirmiran, A., Shahawy, M., Nanni, A., and Karbhari, V. 2004. Bonded Repair and
Retrofit of Concrete Structures Using CFRP Composites, Recommended Construction
Specifications and Process Control Manual. National Cooperative Highway Research
Program Report 514, Transportation Research Board, Washington, DC.
El-Hawary, M., Al-Khaiat, H., and Fereig, S. 2000. Performance of epoxy-repaired
concrete in a marine environment. Cement and Concrete Research, edited by K.
Scrivener, Elsevier Science, 30: 259-266.

108

CDCC-11

DURABILITY AND SUSTAINABILITY OF FIBRE REINFORCED POLYMER

(FRP) COMPOSITES FOR CONSTRUCTION AND REHABILITATION
DURABILIT ET CYCLE DE VIE DES MATRIAUX COMPOSITES
EN POLYMRES RENFORCS DE FIBRES (PRF) EN
CONSTRUCTION ET RHABILITATION

TENSILE CAPACITY OF STRESSED CFRP STRAND EXPOSED TO

EXTREME AGGRESSIVE GROUNDWATER ENVIRONMENTS
M. Sentry1, R. Al-Mahaidi2, A. Bouazza3and L. Carrigan1
1

Geotech Pty Ltd, Port Melbourne, Australia

Professor, Faculty of Eng. & Industrial Sciences, Swinburne University of Technology, Australia
3
Professor, Dept. of Civil Eng, Monash University, Australia
2

ABSTRACT
Current permanent ground anchor standards require the use of double protection systems
encapsulating the steel strands, to ensure a serviceable design life [1]. FRP reinforcement
has been successfully researched and used as a durable construction alternative to steel [2].
Minimal research has been devoted to developing FRP strands for use in high capacity
ground anchors under aggressive environments.
Finding from experiments investigating the effects of aggressive groundwater solutions on
various CFRP strands under stressed state were carried out on two different CFRP materials
under two varying ground conditions at elevated temperature of 60C under saturated
conditions. Tensile strength performance from this investigation indicated that under
extreme conditions certain CFRP strands are more reliable than others. This paper
summarises findings obtained from an extensive PhD research project.
1.

INTRODUCTION

To ensure adequate design life, conventional ground anchors use HDPE sheaths protecting
the steel strand to overcome environmental degradation [3]. Ground anchors in Australia
are exposed to acidic and alkaline ground conditions depending on grounds mineral
content, soil/rock type and age and groundwater flow path [4].
Minimal research has been conducted into the durability performance of CFPR materials
under conditions common to permanent ground anchor applications. Limited knowledge
on strand protection requirements for CFRP ground anchors is known. Greater
understanding into CFRP limitations when exposed to the aggressive ground environments
(groundwater and mineralogy) will provide sound insight into the protection levels required
when using CFRP material in permanent ground anchors.

109

Works presented investigated the tensile performance of two CFRP stands when exposed in
different aggressive environments under a stressed condition. Two different methods of
analysis were used to assess the material performance under extreme aggressive conditions
including tensile strength and scanning electron microscopy (SEM).
2.

PREVIOUS WORKS

Prior to investigating performance of CFRP strand under stressed conditioned when

exposed to aggressive environments, the authors previously presented findings from
research conducted investigating the effects of CFRP exposure to aggressive ground
environments under unstressed condition [4, 5]. Experiments were carried out under same
exposure conditions as works presented within this paper. Results from the unstressed
exposure samples showed minimal physical and tensile changes when compared to
controlled specimens.
3.

EXPERIMENT

3.1. Aggressive Ground Environment

Aggressive ground conditions generally relate to the aggressive nature of either the ground
(soil/rock/contaminated or hazardous waste fill) or the water within the ground
(groundwater) by means of increased levels of minerals, ions or contaminants.
Groundwater can be defined as the subsurface water located in soil pore spaces and in the
fractures of geological formations [6]. Three simulated ground environments were used to
asses CFRP performance (Table 1).
Table 1. Experimental simulated ground environments
Ground Environment Chemical Composition
pH Level
*
Acidic
NaHSO4 + MgCl2 + NaOH + NaCl
9.5 0.5
Alkaline
NaHCO3 + MgCl2 + NaOH + NaCl* 1.0 0.5
Neutral
H2 O
7.0 0.5
Note: * = added to increase salinity of solution to further accelerate aging of specimen.
3.2. Material Selection
Two separate materials from two different manufactures were used. A 15.2 mm diameter, 7
wire helically wound CFRP (CFRP-01) (this strand replicated the dimensions of a 15.2mm
diameter 7 wire low relaxation steel strand) and a 25 mm x 2 mm CFRP flat strand (CFRP02). Table 2 provides details of each material.

110

Table 2. Technical properties of CFRP test specimens

CFRP-01
CFRP-02
Property
Dimensions
Width (mm)
25
Diameter (mm)
15.2
Nominal Thickness (mm)
2
Effective Cross Sectional Area (mm2)
113.6
32
Linear Weight (g/m)
226
97
Carbon Fibre
Minimum fibre volume ratio
0.62
0.64
Density (g/cm3)
1.5
1.8
Tensile strength (MPa)
4200
4900
Elastic modulus (GPa)
240
230
Resin
Modified epoxy Vinylester
Type
Density (g/cm3)
1.6 - 2.0
1.15
Tensile strength (MPa)
80
>55
Product
Tensile strength (MPa)
2200
2500
Elastic modulus (GPa)
141
140
3.3.

Assessment Methodology

3.3.1. Tensile Assessment

Each sample was prestressed prior to curing in a unique stressing frame (Fig. 1Fig. 1a) to
50%Tult (Tult = ultimate breaking load). Once stressed, each sample was individually
exposed to the respective aggressive solutions. Samples were suspended within curing tank
for the required periods of time. Real time monitoring systems were applied to each
specimen to ensure applied stresses were maintained during curing.
Ultimate tensile capacity assessments were conducted in accordance with ASTM D3039M,
ASTM D3916 and ACI440.3R-04 [7-9]. Method of determining elastic modulus of the
cured specimens was in accordance with ACI440.3R-04. A range between 1000 and 3000
micro strain was used to establish the elastic modulus of each sample post curing as
recommended [9].
The results obtained for sustained loaded samples were compared against unstressed results
[4] to assess and compare tendon performance. Ultimate tensile testing was conducted
using a universal testing machine, calibrated and serviced by Monash University.
3.3.2. Scanning Electron Microscope (SEM) Analysis
SEM analyses micro structural changes to the CFRP tendon post specimen curing.
Assessment on the degeneration of the tendon can assist in understanding what effects this
111

may have on the specimens overall tensile performance. Once cured samples were
mounted in a neutral epoxy resin (Fig. 1Fig. 1b), examined surfaces were carbon coated
and analysed at Monash Universitys Centre for Electron Microscopy (MCEM) using the
JEOL JSM-840A scanning electron microscope. Electron back scatter diffraction was used
to visually assess physical alterations to the CFRP strands. Quantitative analytical
spectroscopy was implemented to assess the depth of penetration of chemicals present in
the aggressive ground solutions each specimen was exposed to during curing.
3.4. Test Program
Specimens were cured at temperatures of 60C to accelerate the curing process and attempt
to induce material deterioration. Temperature remained constant throughout each
experiment. Samples were cured for a period of one, three and six months prior to testing.
Details of the testing regime are summarised in Table 3.

Aggressive

Mounted

solution

Stressing

chamber

specimen

support

(b)

(a)

Fig. 1. (a) CFRP tendon stressing frame; (b) Mounted CFRP sample for SEM analysis
Table 3. Experimental test program
Number of Samples Tested
Exposed Curing
Temp
CFRP-01
CFRP-02
Total
Test
Time
Length
(C)
(mths) Alk Aci Neu Alk Aci Neu
(mm)
1
3
3
3
9
3
3
3
3
Tensile
9
300
60
6
3
3
3
3
3
3
18
Capacity
Total
9
9
9
3
3
3
36
6
1
1
1
1
1
1
6
60
300
SEM
Total
1
1
1
1
1
1
6
4.

RESULTS

4.1. SEM Analysis

Unfortunately due to the curing process, CFRP-02 samples taken for SEM analysis had
been damaged beyond use as a result of gripping issues. This indicates that curing the
CFRP-02 specimens in any solution type at elevated temperatures affects the bonding
112

between the fibres and the adjacent bond interface. Additional assessments needs to be
carried out on CFRP-02 specimens once sustained gripping issues are resolved.
4.1.1. Neutral Solution
CFRP-01 samples showed no signs of fibre fatigue, damage, fibre splitting or de-bonding
with the surrounding matrix after sample exposure to the neutral solution under stressed
conditions (Fig. 2a).
4.1.2. Acidic Groundwater
CFRP-01 cured samples showed no change in state from the controlled results (Fig. 2b).
The inner king wire was fully intact with no signs of deposits, micro-cracking or voids as
a result of the stress-cured curing process. Spectroscopy analysis conducted on the
localised outer fibre boundary layer showed high concentration levels of sodium and
chloride (Fig. 3). No signs of deposit penetration were observed from the visual SEM
analysis.

(a)
(b)
Fig. 2. Stressed CFRP-01 tendon cured for 6 months in (a) neutral solution (b) acidic
aggressive groundwater.

Fig. 3. SEM image and spectrum analysis for acidic groundwater cured CFRP-01 samples
after 6 months.

113

4.1.3. Alkaline Groundwater

CFRP-01 samples showed localised micro-cracks of the outer fibre layer had enabled the
penetration of deposits from the aggressive groundwater solution along a micro-crack
within close proximity to the outer boundary layer (Fig. 4a). Close inspection of the deposit
exposed some minor deterioration of the bond between the epoxy and fibres (Fig. 4b) and
concentrations of Sodium-Chloride crystals (Fig. 4c).
Spectroscopy analysis conducted on the localised outer fibre layer micro-cracks (Fig. 5)
showed high concentration levels of sodium and chloride and minor concentrations of
calcium. Although penetration of the aggressive alkaline solution occurred in the CFRP-01
sample, no sign of fibre deterioration was evident.

(a)

(b)

(c)

Fig. 4. Stressed CFRP-01 samples cured at 60C for 6 months (a) outer fibre layer
penetration; (b) deterioration of inner fibre region; (c) deposit within inner fibre region

Fig. 5. SEM image and Spectrum analysis of CFRP-01 sample after 6 months.
4.2. Tensile Capacity
Due to the flat profile of the CFRP-02 specimen, consistency issues with gripping during
testing resulted in inconsistent findings. Method of specimen anchorage during elevated
temperature curing was established as a reliable method of sustaining load. Further

114

investigations are being carried out to determine an effective gripping system to repeat
these tests.
Visual inspections of CFRP-01 tendon failure locations identified no consistent failure
pattern for all specimens tested over the six months in all curing environments. Failure was
by rupture, located within the gauge length. Failure strains for CFRP-01 specimens cured
in acidic, alkaline groundwater and neutral solutions indicated minimal variation compared
to control results (Fig. 6).
4.2.1. Neutral Solution
Results indicted a decrease in elasticity of 8.8% curing for six months (Fig. 6). Ultimate
strength increases of 14.5%, 15.5% and 16.3% were recorded for neutral stress-cured
specimens. Strength results observed confirms findings from SEM analysis.
4.2.2. Acidic Solution
Elasticity results were consistent, indicating that there is less impact on ductility variations.
Increases in strengths of 10.6%, 17.0% and 14.2% were recorded (Fig. 6). Compared to
unstressed results [4, 5], strength performance of CFRP-01 samples stress-cured out
performed unstressed-cured specimens. Recorded tensile capacity verifies findings from
SEM analysis. Minimal change in physical and engineering properties result from curing
of CFRP-01 samples under stressed conditions.
4.2.3. Alkaline Solution
Elasticity variations of 7.0%, 13.5% and 1.1% were recorded after one, three and six
months respectfully. Strength increases of 10.2%, 14.7% and 13.0% were recorded over
the six month assessment compared to controlled results (Fig. 6). These findings indicate
the stress-cured environment does alter the elasticity properties of the composite structure
in an alkaline solution. Although penetration of the aggressive solution occurred after six
months curing, this did not affect the tensile performance of the CFRP-01 sample.
5.

CONCLUSIONS

Tensile capacity and SEM assessment was carried out under stressed conditions at elevated
temperatures over a six month period on CFRP tendons exposed to aggressive acidic,
alkaline ground environments and neutral environments which can exhibit in ground
environments permanent ground anchors are installed in. CFPR-01 samples did not show
any sign of tensile strength decrease as result of the accelerated curing process even though
SEM analysis indicated penetration of aggressive solution had occurred after six months
curing. Due to gripping issues, no reliable results were obtained for the CFRP-02 samples.
Further research into a reliable gripping system for the flat tendon is required prior to reassessing the performance of this product under accelerated aggressive stressed conditions.

115

2500

2250

2000

1750

Stress(MPa)

1500

1250

1000

750

500

CFRP01 CONTROL

CFRP01 ACI 6mths

CFRP01 NEU 6mths

CFRP01 ALK 6mths

CFRP01 ACI 3mths

CFRP01 NEU 3mths

CFRP01 ALK 3mths

CFRP01 ACI 1mth

CFRP01 NEU 1mth

CFRP01 ALK 1mth

250

0
0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.1

1.2

1.3

1.4

1.5

Strain(%)

Fig. 6. Tensile test results of stressed CFRP-01 samples.

6.

ACKNOWLEDGEMENTS

The authors would like to express their gratitude to Monash University and the civil
engineering laboratory and research sponsor Geotechnical Engineering.
7.

REFERENCES

1.

Littlejohn, G.S. 1987. Ground Anchorages: corrosion performance. Proceedings Institute of Civil Engineers, Part 1 (82): 645-662.
Benmokrane, B., and Chennouf, A. 1997. Tensile and bond properties of AFRP and
CFRP tendons for ground anchor applications. Proceedings of the 3rd Symposium on
non-metallic reinforcements for concrete structures, FRPRCS-3, Sapporo, Japan, 14-16
October, 2: 727-734.
Sentry, M., Bouazza, A., Al-Mahaidi, R., Loidl, D., Bluff, C., and Carrigan, L. 2009.
Carbon Fibre Reinforced Polymer (CFRP): An Alternative Material in Permanent
Ground Anchors. Journal of Australian Geomechanics, 44(3): 47-56.
Sentry, M. 2010. An investigation into the durability and performance effects of CFRP
strand in permanent ground anchors, in Department of Civil Engineering, Monash
University, Clayton. 398 p.
Sentry, M., Bouazza, M., Al-Mahaidi, R., Collins, F., Carrigan, L., Loidl, D., Bluff, C.
2008. Durability of Carbon Fibre Reinforced Polymer (CFRP) strand for use in high
capacity permanent ground anchors. Proceedings of the Australian Structural
Engineering Conference (ASEC), Melbourne, 26-27 June, 8 p.
Sentry, M., Bouazza, M., Al-Mahaidi, R., Collins, F., Carrigan, L., Loidl, D., Bluff, C.
2007. Durability of Carbon Fibre Reinforced Polymer (CFRP) strand for use in high

2.

3.
4.
5.

6.

116

7.
8.
9.

capacity permanent ground anchors. Proceedings of the 14th international conference

on composite structures (ICCS14), Melbourne, Australia, 9 p.
ASTM. 2000. Standard test method for tensile properties of polymer matrix Composite
Materials. (D3039-00), West Conshohocken, Pa.
ASTM. 2002. Standard test method for tensile properties of pultruded glass-fibrereinforced plastic rod. (D3916-02), West Conshohocken, Pa.
ACI Committee 440. 2004. Guide Test Methods for Reinforcing or Strengthening
Concrete Structures. ACI 440.3R-04, American Concrete Institute, Farmington Hills,
MI, 40 p.

117

118

CDCC-11

DURABILITY AND SUSTAINABILITY OF FIBRE REINFORCED POLYMER

(FRP) COMPOSITES FOR CONSTRUCTION AND REHABILITATION
DURABILIT ET CYCLE DE VIE DES MATRIAUX COMPOSITES
EN POLYMRES RENFORCS DE FIBRES (PRF) EN
CONSTRUCTION ET RHABILITATION

THE EFFECT OF LONG-TERM LOADING ON THE CFRPCONCRETE BONDING STRENGTH

C. Mazzotti1 and M. Savoia2
1
2

Associate Professor DICAM - Structural Engineering, University of Bologna, Italy

Professor DICAM - Structural Engineering, University of Bologna, Italy

ABSTRACT:
Results of an experimental campaign concerning the long-term behavior of FRP plates
bonded to concrete are presented. Four CFRP plates with three different bonded lengths
(from 100 400 mm) have been subject to long-term constant traction force. The tests have
been carried out in a climatic room with a constant temperature T = 20 C and humidity RH
= 60%, for about five years. The strain profile evolution with time along the bonded length
has been recorded by using a number of strain gauges along the plates. An appreciable
stress redistribution with time along the plate has been observed. At the end of the longterm test, all specimens have been subjected to failure load in order to verify if the interface
behaviour was affected by the long-term loading application. No remarkable degradation
has been found.
1. INTRODUCTION
Most of the studies carried out, in recent years, regarding the bond between the Fiber
Reinforced Polymer (FRP) and concrete concern the characterization of its performances
under short-term loadings. Nevertheless, it is very important to verify the durability with
time of the retrofit intervention. The effectiveness of the strengthening can in fact be in
principle reduced due to the rheological behaviour of the materials (concrete, adhesive,
FRP) in the case of long-term loadings.
As far as the behaviour of FRP-strengthened elements under sustained load is concerned,
only in the last few years the scientific contributions are closing the gap of knowledge,
starting from the early work of Plevris and Triantafillou [1]. Some contributions can be
found, i.e., experimental investigations [2], [3], [4] and numerical models [5], [6], [7]. A
common result of all research studies is that, under long-term loading, a shear stress
redistribution occurs, during time, along the FRP element due to the viscosity of both
concrete and adhesive. It has not been investigated up to now if the long-term loading could
negatively affect the bond strength of the strengthening system, eventually modifying the
interface law and the transmissible force.
119

In the present paper, the results of an experimental campaign concerning the long-term
behaviour of the bond between FRP plates and concrete prisms are presented. Three
different bonded lengths (100 - 400 mm) have been tested. The time duration of the tests
has been about 2000 days (more than five years of continuous loading). The strain
distribution along the bonded part has been monitored with time.
After the unloading, the specimens have been subjected to instantaneous bond failure test in
order to verify if the previous loading history affected the bond strength. Strains along the
reinforcement during the tests have been recorded and the corresponding shear stress
distributions evaluated, thus allowing for the calibration of shear stress-slip interface laws.
All the data have been compared with analogous information obtained, five years ago, by
tests on reference specimens made of the same constituents. The comparison shows that no
particular damage occurred to the interface behaviour due to the long-term loading.
2. LONG-TERM BOND TESTS
2.1 Specimen Preparation
The long-term behaviour of CFRP plates bonded to concrete has been investigated by
testing three different bonded lengths (B.L. = 100, 200 and 400 mm) and two concrete
specimens (two CFRP plates for each concrete specimen).
The concrete specimen dimensions were 150200600 mm (Fig. 1). Specimens were
fabricated using normal strength concrete (at the age of the long-term loading application,
compression strength was fcm = 52.6 MPa and tensile strength, fctm = 3.81 MPa, from
splitting tests). Further details can be found in [4].

200

CFRP plate
801 2 mm

Adhesive
801 5 mm

150
7 to 11 strain gauges

L9

L2

100 mm
L1

(b)
CONSTANT
LOADS

B.L
.
100
200
400

7 to 11 strain gauges

L1 L2 L3 L4 L5 L6 L7 L8 L9

L1
L11
0

10 10 10 15 15 15 15
10 10 10 20 20 20 30 30 40
10 10 20 20 40 40 50 50 50 50 50

(a)
(c)
Fig. 1: (a) Geometry of specimens and (b) experimental setup for the long term tests,
(c) spacing (mm) between strain gauges along the FRP plates with different B.L.
120

(b)

CFRP plates, 80 mm wide and 1.2 mm thick, have been used; according to technical data
provided by the producer, the mean elastic modulus is Ep = 165000 MPa.
Two opposite surfaces of each concrete block have been grinded with a stone wheel to
remove the top layer of mortar, until the aggregates were visible (approximately 1 mm).
The plates have been bonded to the concrete surfaces by using a 1.5 mm thick layer of a
two components epoxy adhesive, having mean tensile strength of 30.2 MPa and mean
elastic modulus Ea = 12840 MPa. No primer before bonding has been used. The plate
bonded length starts 100 mm far from the front face of the specimens (Fig. 1a). Two CFRP
plates have been bonded to the opposite faces of each concrete specimen, with different
bonded lengths (concrete Block P1: B.L. = 100, 200 mm; Block P2: B.L. = 200, 400 mm),
see the layout reported in Fig. 1a. The curing periods of the anchorages of blocks P1 and
P2 were 15 and 20 days prior to testing, respectively.
2.2 Experimental Setup and Instrumentation
The experimental setup is depicted in Fig. 1a. The concrete block was positioned on a rigid
frame with a steel reaction element at the front side (60 mm height), in order to prevent
longitudinal rigid displacements. The traction force was applied to each couple of CFRP
plates by using a mechanical frame allowing for the magnification of the force applied at
the opposite side of the horizontal arm (leverage system). The force was given as the
weight of a number of steel plates (Figure Error! Reference source not found.Error!
Reference source not found.1b); details on the experimental set-up can be found in [4].
The long-term tests were performed by prescribing a constant force during time.
Along the CFRP plates bonded to concrete, series of 7 11 strain gauges (depending on the
plate length) were placed, along the centerline. For each bonded length, the spacing
between the strain gauges is reported in Fig. 1c.
2.3 Loading Program
Two different loading histories have been prescribed. The FRP plates of specimen P1 have
been subject to a traction force of 12.30 kN (about 50 percent of the maximum
transmissible load, according to debonding tests on analogous specimens, see [8]), constant
in time for 1981 days. The plates bonded to specimen P2 have been subject to the same
sustained load (12.30 kN) for 510 days; after that, a second load step of 3.80 kN has been
applied and maintained for further 1454 days. The longitudinal strains along the plate at
different time intervals have been recorded by an automatic computer system. Further
details can be found in [4].
2.4 Long-term Results
Figures 2a-d show the time evolution of the longitudinal strains along the plates for 100
200 mm (specimen P1) and 200 400 mm (P2) bonded lengths, respectively. For each
strain gauge, the distance from the initial section of the anchorage is reported in
parenthesis. It can be observed that, at medium-low stress levels, plates with different
121

bonded length exhibit a similar behaviour: the rate of strain starts decreasing immediately
after the load application (as in classical creep tests on concrete specimens). The continuous
and uniform reduction of the strain rate for more than five years, as already shown in [4]
using the log-scale of time for a shorter time interval, suggest that after few months of
loading a steady-state creep condition was attained.

Longitudinal strain (

700

E1 (10)

E2 (20)

E3 (30)

E5 (60)

E6 (75)

E7 (90)

800

E4 (45)

700

(10)

Longitudinal strain (

800

600
(20)

500
400

(30)

300
(45)

200
(75)

100

(60)

0
730

1095

1460

E3 (30)

E4 (50)

E7 (120)

E8 (150)

E9 (190)

1825

(20)

400

(30)

300

(50)
(70)

200

(90)

2190

365

730

(a)

2190

(10)

E1 (10)

E2 (20)

E3 (30)

E4 (50)

E5 (70)

E6 (90)

E7 (120)

E8 (150)

E9 (190)

700

Axial strain ()

Longitudinal strain ()

1825

800

600

(20)

500

(30)

400
300

(50)

200

(120)

100

730

1095

1460

nd

Load step
(20)

600

E1 (10)
E5 (100)
E9 (290)

500
400

E2 (20)
E6 (140)
E10 (340)

E3 (40)
E7 (190)
E11 (390)

200

(150)

100

(100)

0
1825

2190

Time (days)

E4 (60)
E8 (240)

(40)

300

(90)

(190)

(60)

(70)

365

1460

900

Load step

700

1095

(b)

(10)
nd

(150)

(190)

(120)

Time (days)

900
2

(10)

500

Time (days)

800

E5 (70)

600

0
365

E2 (20)

E6 (90)

100

(90)

E1 (10)

365

730

1095

Time (days)

1460

1825

(140)
(190)
(240)
(290)
(340)
(390)

2190

(c)
(d)
Fig. 2: Time evolution of strains in FRP plates at different positions along the anchorage:
specimen P1 (a) B.L.=100 mm, (b) B.L.=200 mm; specimen P2 (c) B.L.=200 mm, (d)
B.L.=400 mm. The number in parenthesis indicates the distance (mm) from the initial
section of the anchorage.
As a confirmation, Fig. 4a shows the strain evolution with time of the strain gauges
belonging to the two plates with B.L.=200 mm. Thick lines refer to the specimen P2
subject to the second loading step p2 510 days after the first loading p1 (p2=0.35p1),
whereas thin lines refer to the specimen P1 subject to the initial loading step only (p1).
The evolution with time of the strain curves obtained from specimen P2, after the second
loading step, rapidly attains the same slope of the corresponding curves from specimen P1.
Finally, Figs. 3a-d show the strain profile along the plates at different times, for all the
specimens. The dashed lines join the first experimental strain value recorded at x=10 mm
with the theoretical value of the FRP strain at x=0, evaluated as 0(t)=F/ Ep(t) Ap, with t
being the time after the loading application and Ep(t) the CFRP elastic modulus, taking
into account the material viscosity (very small). As expected, at the beginning of the
bonded length the plates show increasing strain values with time, approaching the limit
value of 0 (no redistribution can occur in the plate free part). The strain profiles are
qualitatively similar, suggesting the increase of the delayed strain with time reduces
122

moving along the bonded length; on the contrary, at a distance greater than 150 180 mm
from the loaded end (a realistic value of transfer length), the delayed strain increase with
time is almost constant along the bonded length, so confirming the experimental results by
Wu and Diab [2]. After the second load increment (Figures 3c,d), the long-term strain
variation is less pronounced, due to the concrete and the interface ageing, and the strain
curves shifting along the bonded length with time is progressively reducing.
The creep strains measured on the CFRP plates suggest that the adhesive and the concrete
cover, constituting the interface where the load is transferred, are subject to creep strains,
thus increasing the compliance of the interface with time. A shear stress redistribution
along the interface then occurs: high shear stresses close to the loaded end decrease and a
longer portion of the bonded length is subject to the shear stress transfer, i.e. the transfer
length increases.
Due to the very high shear stresses at the adhesive and concrete cover level, close to the
loaded end the creep phenomenon may be highly non linear. Hence, the compliance
increase at the beginning of the anchorage can be much higher than far from it. The
evolution with time of the creep deformation at the interface level is then variable along the
anchorage, and may explain the shear stress redistribution along the interface. As a
consequence, the normal stresses in the plate increase far from the loaded section of the
plate, so explaining the large increase of the axial strains with respect to the values recorded
immediately after the load application.
800

800

End of Loading

End of loading
700

30 days

Longitudinal strain ()

Longitudinal strain ()

700

192 days

600

932 days
500

1981 days

400
300
200

30 days
192 days

600

932 days
500

1981 days

400
300
200
100

100

0
0

10

20

30

40

50

60

Distance (mm)

70

80

90

100

20

40

60

(a)

100

120

140

160

180

200

(b)
1000

1000

End of Loading

900

510 days

600

2nd Load step

Axial strain ()

175 days

700

676 days

500

930 days

400

End of Loading

900

30 days

800

Axial strain ()

80

Distance (mm)

1964 days

300

30 days

800

175 days

700

510 days
2nd Load step

600

676 days

500

930 days

400

1964 days

300

200

200

100

100
0

0
0

20

40

60

80

100

120

140

160

180

200

Distance (mm)

40

80

120

160

200

240

Distance (mm)

280

320

360

400

(a)
(b)
Fig. 3: Strain distributions along the bonded length, at different times. Specimen P1: (a)
B.L.=100 mm, (b) B.L.=200 mm and specimen P2: (c) B.L.=200 mm, (d) B.L.=400 mm.
123

3. FAILURE TESTS
After the conclusion of the long-term tests, all the specimens have been subject to failure
tests in order to verify the effect of the long-term loading on the FRP-concrete bond
strength and the type of failure. For all the specimens, failure was caused by shearing of a
layer of concrete 1-2 mm thick. A detailed description of the experimental set-up can be
found in [8]. Similar bond tests were also carried out at the beginning of the long-term
experimental campaign on specimens made with the same materials; a comparison between
results before and after the long-term loading can then be given.
Figure 4b shows the comparison between the bond forces obtained, for the different bond
lengths, before and after the creep tests. The tests clearly show that the long-term loading
did not cause damage of the interface transferring the shear force. A systematic increase of
the bond force can be in fact observed for all the bond length but bl=400 mm; in this case,
during the preparation of the failure test a partial debonding of the front side of the CFRP
plate occurred, thus producing a bond degradation and a reduction of the corresponding
force of about 15%. The force increase can be only partially explained by the strength
increase with time of concrete substrate due to aging; in particular, the compressive
strength measured on cylinders obtained from the concrete specimens is fcm(6 years)=54.2
MPa while at the beginning of the long-term tests was fcm(9 months)= 52.6 MPa. According
to the expression reported in [9], this strength variation should produce a bond force
increase of about 1 kN, much smaller than the measured increase. The increase of bond
strength can be also due to the ageing of the epoxy, for which, nevertheless, there are no
available data.
45

P2-200_E1
P1-200_E1

1000
900

P2-200_E3
P1-200_E3

P2-200_E5
P1-200_E5

P2-200_E7
P1-200_E7

700
600
500
E3

400
300
200
100

Before creep test

After creep test
After creep test

35

E1

Bond force (kN)

Axial strain

800

40

30
25
20
15

E5

10

E7

365

730

1095

Time (days)

1460

1825

2190

100 mm

200 mm

400 mm

Bond length (mm)

Fig. 4: (a) Strain evolution with time recorded by strain gauges E1, E3, E5, E7 (specimens
P1 and P2, B.L.=200 mm) and (b) bond strength measured before and after long-term tests.
From the strain-gauges measures along the FRP bonding, the strain distribution along the
reinforcement for increasing values of the applied force during failure tests have been
obtained; these distributions are reported in Figures 5a-d; points A to C refers to subsequent
steps of the debonding process. The shapes of the curves, both at low level of applied force
and at the onset of debonding or during it, are qualitatively similar to those obtained in
standard conditions (without application of a preceding long-term loading). Only in the
case of bl=400 mm, the debonding occurred before the test produced a strain distribution in
124

the first 40 mm of the reinforcement (yellow area of Figure 5d) almost constant also at the
beginning of the test. This portion of the strain curve has not been then considered in the
subsequent analyses. In order to verify if the strain distribution was affected by the long
term loading, the strain curves along the reinforcement from specimen P2-200 (solid
markers), for increasing values of the applied force, have been compared in Figure 6a with
the analogous results obtained from the reference specimen with the same bond length
(hollow markers). A very good matching can be observed at both low level of applied force
(linear-elastic behaviour) and during debonding.
3000

3000

4 KN

4 KN

12 KN

2500

12 KN

2500

20 KN

20 KN

28 KN

28 KN

Axial strain ()

Axial strain ()

2000

2000

34 kN
35 kN

1500

36 kN
40 kN
point A

1500

point A
point B

point B
point C

1000

1000

point D

500

500

0
0

10

20

30

40

50

60

70

80

90

100

20

40

60

80

100

120

140

160

180

200

x (mm)

x (mm)

(a)

(b)

3000

3000
4 KN

4 KN

12 KN

2500

12 KN

2500

20 KN

20 KN

FRP Debonded part

28 KN

28 KN
2000

Axial strain ()

Axial strain ()

2000

36 kN
37.5 kN
point A

1500

point A
point B

1500

point B
point C

1000

30 KN

point C

1000

point D

500

500

0
0

20

40

60

80

100

120

140

160

180

200

50

100

150

200

x (mm)

x (mm)

(c)

(d)

250

300

350

400

Fig. 5: Strain distribution along the bonded part during the failure tests Specimen P1 (a)
B.L.=100 mm, (b) B.L.=200 mm, Specimen P2 (c) B.L.=200 mm, (d) B.L.=400 mm.
Following a standard procedure [8], the average shear stress between two subsequent strain
gauges and the corresponding CFRP-concrete slip have been evaluated as a function of the
measured strains, thus obtaining a set of couples of average shear stress and slip, used to
calibrate a local interface law [10]. Figure 6b show the interface law calibrated adopting the
data obtained from the reference specimens and from those subject to the creep tests,
together with all the shear stress slip data. The initial branch of the curves has not been
affected by the long-term loading (same linear elastic stiffness); the interface law of the
specimens subjected to the creep test exhibits a higher maximum shear stress and a slightly
steeper softening branch. The fracture energy in the second case is greater, due to the
greater value of the debonding force.

125

Reference test

Axial strain ()

2000

1500

16

4 KN
12 KN
20 KN
28 KN
point A

4 kN-Ref
12 kN-Ref
20 kN-Ref
28 kN-Ref
point A-Ref

point B
point D

point B-Ref
Point C-Ref

Referencebonddata
ReferenceCurve
Bonddataaftercreeptest
Curveaftercreeptest

14
12

Test after
creep load

1000

Shearstress(MPa)

2500

10
8
6
4

500
2
0

0
0

20

40

60

80

100

120

140

160

180

200

0.05

0.1

0.15

x (mm)

Slip(mm)

(a)

(b)

0.2

0.25

0.3

Fig. 6: Reference case (B.L.= 200 mm) and specimen P2 after the long-term test (a)
Strains along the bonded part and (b) shear stress-slip curve.
4. CONCLUSIONS
The results of an experimental campaign concerning the long-term behaviour (about five
years of loading) of bond between concrete prisms and FRP plates are presented.
Bonding with plate with different lengths exhibit a similar behaviour with time,
characterized by a steady-state delayed strains increase reached after about three months of
loading. After more than five years of testing, no remarkable change of quality of the curve
has been observed (steady interface viscosity).
The application of subsequent load steps at different times shows the long-term interface
behaviour is mainly governed by the aging of the concrete (enhancement of the mechanical
properties due to the evolution with time of the hydration level of the cement paste).
Moreover, the post-processing of the experimental results shows a significant redistribution
of the shear stresses along the anchorage due to the creep deformation at the interface level.
The failure tests carried out at the end of the long-term test show that no appreciable
damage occurred due to creep of the interface.
5.

ACKNOWLEDGEMENTS

The financial supports of the (Italian) Dept. of Civil Protection (RELUIS 2009 Grant
Task 3.1: New materials for the seismic retrofit) is gratefully acknowledged. The authors
would like to thank also the Sika Italia S.p.A.
6.

REFERENCE

1.

Plevris, N. and Triantafillou, T.C. 1994. Time-dependent behaviour of RC members

strengthened with FRP laminates. Journal of Structural Engineering, ASCE, 120(3):
1016-1042.
Wu, Z. and Diab, H. 2007. Constitutive model for time-dependent behaviour of FRPconcrete interface. Journal of Composites for Construction, ASCE, 11(5): 477-486.

2.

126

3.

Choi, K.-K., Meshgin, P. and Reda Taha, M.M. 2007. Shear creep of epoxy at the
concrete-FRP interfaces. Composites Part B: Engineering, 38, (5-6): 772-780.
4. Mazzotti C., Savoia M. 2009. Stress redistribution along the interface between concrete
and FRP subject to long-term loading. Advances in Structural Engineering, 12(5): 648658.
5. Savoia, M., Ferracuti B. and Mazzotti, C. 2005. Long-term creep deformation of FRPplated r/c tensile members. Journal of Composites for Construction, ASCE, 9(1): 6372.
6. Diab, H. and Wu, Z. 2007. Nonlinear constitutive model for time-dependent behaviour
of FRP-concrete interface. Composites Science and Technology, 67(11-12): 23232333.
7. Benyoucef, S., Tounsi, A., Adda Bedia, E.A. and Meftah, S.A. 2007. Creep and
shrinkage effect on adhesive stresses in RC beams strengthened with composite
laminates. Journal of Composites Science and Technology, 67(6): 933-942.
8. Mazzotti, C., Savoia, M. and Ferracuti, B. 2008. An experimental study on
delamination of FRP plates bonded to concrete. Construction and Building Materials,
22(7): 1409-1421.
9. Research National Council CNR Committee 2004. Guidelines for Design,
Construction of Externally Bonded FRP Systems for Strengthening Existing Structures.
CNR DT 200/2004, Rome, Italy, 144 p.
10. Ferracuti, B., Savoia, M. and Mazzotti, C. 2007. Interface law for FRPconcrete
delamination. Composite Structures, 80(4): 523-531.

127

128

CDCC-11

DURABILITY AND SUSTAINABILITY OF FIBRE REINFORCED POLYMER

(FRP) COMPOSITES FOR CONSTRUCTION AND REHABILITATION
DURABILIT ET CYCLE DE VIE DES MATRIAUX COMPOSITES
EN POLYMRES RENFORCS DE FIBRES (PRF) EN
CONSTRUCTION ET RHABILITATION

DESIGN OPTIMIZATION OF COMPOSITE STRUCTURES

CONSIDERING BONDED JOINS DURABILITY
P. Hamelin1, A. Si Larbi2 and E. Ferrier3
1

Professor, University of Lyon, Laboratory LGCIE, Lyon, France

Associate Professor, University of Lyon, Laboratory LGCIE, Lyon, France
3
Professor, University of Lyon, Laboratory LGCIE, Lyon, France
2

ABSTRACT
First we introduce the notion of critical length for load transfer between substrates and
adhesive. Then we present viscoelastic measurements which permit to evaluate the
rheological behaviour of the epoxy and to identify its creeping function. For a final
application which concerns mixed steel-concrete structure connected by adhesive join we
are able to design an innovative technological process which consider discontinuous epoxy
blobs with a length defined from durability analysis.
1. INTRODUCTION
In recent years, structural bonding has been much used for the repair of damaged structures
in reinforced concrete [1], [2], considering reinforcement by the addition of composite
pultruded plates, or by the introduction of additional composite rebars (NSM process) [3] in
zones subjected to tension or shear.
The performance and durability of epoxy adhesives has been established and the principal
characteristics considered in the dimensioning rules are the properties in instantaneous and
delayed tensile and shear resistance. In the case of bonding on concrete, the modes of
rupture often depend on the surface treatment of the concrete and of the tensile and shear
resistance properties. We are particularly interested, in this research, in optimizing the
design of bonded joints to ensure load transfer by shear between the concrete and the steel
and also in optimizing the dimensioning of composite concrete-steel beams in order to
achieve a non-fragile ultimate behaviour controlled by the plastic deformation of the steel
and by the absence of any initiation of cracking in the bonded joints following the HSF
methods proposed by P. Hamelin, 2009 [4].

129

2. OPTIMIZATION OF MECHANICAL PERFORMANCE RELATIVE TO

SHEARING BONDED JOINTS: Notion of a critical length in function of glue
nature
2.1. The Theory: Volkersen Expressions
We first consider the standard Volkersen expressions [5]. The expression for the variation
of shear stress as a function of abscissa x of a point on the bonding surface is given by:

x moyen

Where
F = external force
l = length of overlap between the two substrates
h = width of overlap
G = shear modulus of the adhesive
E1 and E2 = longitudinal elasticity
(1)modulus of
substrates 1 and 2
s1 and s2 = thickness if substrates 1 and 2
e = thickness of the layer of adhesive
adh shear resistance of the adhesive

l chx E 1 s1 E 2 s 2 shx

.
l
2 l E1 s1 E 2 s 2
sh
ch

2
2

With: moyen

1/ 2
F
et G E1 s1 E 2 s 2
lh
E1 s1 E 2 s 2 e

(1)

(2)

Fig. 1. Description of the bonded joints [12]

Considering: E2

kE1

, s 2 ps1 ,

1 kp
1 kp

we simplify the expression to: x moy

and

1 kp
kp

(3)

l cosh x
sinh x
with
M.
l
l
2
sinh
2

cosh

GN
E1 s1e

(4)

Since stress is maximal at the extremities, shear failure of the bonded joint takes place for
/ 2 adh we can establish a relationship between adh and moy
Calculation of the force supported by the joint Fr moy .bl
Frd l

adh
A

1
with A
coth Al M tanh Al

GN
(Hamelin, 2009)
4 E1es1

(5)
(6)

2.2. Influence of the Nature of the Adhesive on the Mechanical Behaviour of a SteelConcrete Assembly
With substrate n1 being concrete and substrate n 2 being steel, following the data in
Table 1, we used the expressions above to study, for two types of adhesive (hard epoxy
130

[Sikadur 31] and flexible polyurethane [Sikaflex]), the variation of the load supported for
different lengths of bonded joint (Fig. 2).
Table 1. Data for two different adhesives with two substrates (concrete and steel)
E1 (MPa)

s1 (mm)

E2 (MPa)

s2 (mm)

e (mm)

b (mm)

30 000

50

210 000

5,7

7,000

0,1140

0,1123

2,2531

3,000

55

Load supported by the bonded join

Polyurthane

30,0000

Epoxy

out

25,0000

G (MPa)

29

adh (MPa)

9,5

A(polyurthane)

0,003392

kN )
Load ((N)

20,0000

15,0000

Epoxy

10,0000

Fcritical
5,0000

200

polyurethane

critical

critical

0,0000

400

600

800

1000

Lenght

1200

1400

1600

1800

G (MPa)

4 590

adh (MPa)

19,5

A(epoxy)

0,042678

2000

(mm)

Fig.2. Theoretical force supported by the bonded joint for two types of adhesive
There is a critical, or efficient, bond length above which any increase in length provides no
increase in the force supported by the join (Fig. 2). In fact, the maximum shear stress in the
adhesive is not reached because the failure appears in the concrete with an average
maximum shear stress equal to 2MPa. The critical length which depends on the mechanical
performance of the adhesive can be evaluated considering the same calculation with two
metal substrates (Table 2) which present higher resistance in comparison with concrete.
Table 2. Data for the calculation of the joints bonded load for two metal substrates
E1
(MPa)

s1
(mm)

E2
(MPa)

s2
(mm)

210000

3,25

210000

3,25

1,0000

1,0000

0,0000

2,0000

131

kp

G
(MPa)

e
(mm)

(MPa)

4590

19,5

b
(mm)

0,0739

0,041004

10

Epoxy

Fr (kN)
(N)

criticalsteel

l (mm)

Optimum shear length = 60mm

Fig. 3. Shear resistance of the bonded joint as a function of the length of the joint for two identical
metal substrates and a hard epoxy adhesive (Sikadur 31)
For a width of 10mm, the critical length for epoxy is equal to 60mm (Fig. 3). Consequently,
for a width of 55mm (Table 1) the critical length will be determined by the maximum load
supported by the concrete ( Fcritical concrete * b * criticalsteel 2MPa * 55mm * 60mm ). We can
evaluate on the Fig. 2 the corresponding critical length for load transfer equal to
criticalconcrete 150mm .As the assumption of perfect adhesion is considered the average stress
level for the epoxy is equal to 0.75MPa in this case.
Consequently, if we want to increase the shear load supported by connection, we made an
assembly using discontinuous blobs of adhesive, limiting the length of the join to the
efficient length and increasing the number of blobs and their width to obtain the global and
local shear resistance. The validation of this principle has been demonstrated by Hamelin et
al. [4].
3. DURABILITY PREDICTION OF ADHESIVE JOIN
3.1. Determination of the Creep Behaviour Law of the Adhesive by Mechanical
Spectrometry
As the average shear stress level in the adhesive join is low (1MPa) we will consider a
linear viscoelastic behaviour [7]. We used a METRAVIB type visco-elasticimeter which
allowed the harmonic periodic application of shear stress with a frequency ranging between
1Hz and 100Hz and temperature scanning from -20C to 120C (Fig. 4).

132

Tension test

Shear test

Flexural test

Fig. 4. METRAVIB Viscoelasticimeter

3.2. Results and Interpretation
The behaviour of the polymer becomes extremely visco-elastic as the vitreous transition
temperature is approached. Therefore, as illustrated in Fig. 5 for epoxy resin (Sikadur 31),
the passage of Tg is accompanied by a considerable drop in the mechanical properties. In
our case, the glass transition temperature is 42C.
Epoxy
1,00E+10

1,00E+00

G'
G"
tan delta

tan delta

G', G" (Pa)

1,00E+09

1,00E+08

1,00E-01

1,00E+07

1,00E+06
-20

20

40
60
Temprature (C)

80

100

1,00E-02
120

Fig. 5. Variation of G, G and Tg in function of temperature (epoxy resin)

Presentation of the results (Fig. 6) in the complex plane/ surface (Cole-Cole diagram)
shows that: The curve tends towards an asymptotic point G0 corresponding to the
instantaneous modulus. The curve also tends towards an asymptotic point Ginf
corresponding to the long-term modulus. This representation helps to establish a
rheological model describing the visco-elastic behaviour of the resin over time, frequency
and temperature.

133

3,00E+08
2,50E+08

G' (Pa)

2,00E+08
1,50E+08
1,00E+08
5,00E+07

Ginf

G0

0,00E+00
0,0E+00

5,0E+08

1,0E+09

1,5E+09

2,0E+09

2,5E+09

3,0E+09

G'' (Pa)

Fig. 6. ColeCole diagram (epoxy resin)

To account for the dissymmetrical shape of the curve experimentally obtained in the
complex plane (Fig. 7), the Zener biparabolic model seems to be the most appropriate, and
G 0 G inf
(7)
so the complex modulus is written as G * G
1 i h i k
is a parameter which takes into account the two periods of relaxation.

G0 - Ginf

G0
Ph
Pk

Fig. 7. Zener biparabolic model

The parameters , h, and k of this biparabolic model in parallel are determined by following
the process described by Huet [8] and Yong-Sok [10].
3.3. Experimentally Obtained Parameters
From the previously described tests on the viscoelasticimeter, the following values are
obtained for the epoxy resin, Sikadur 31:

h
0.11

k
0.47

Table 3. Constants of the rheological model

G0
1.3
8.58e7

134

Ginf
2.86e9

After integration, the analytical expression of the biparabolic model then becomes:
G 0 G
As a function of relaxation time of the resin
(8)
G t / G
1 t / h t / k
We can then draw the curve of the shear modulus over time for two constant temperatures:
+20C and +30C (Fig. 8).

Module
de modulus
cisaillement
(MPa)
(MPa)
Shear

3500

Rheological model 20C

3000

Rheological model 30C

2500
2000
1500
1000
500
0
1

10

100

1000

10000
100000
Time (s)
Temps (s)

1000000

1E+07

1E+08

1E+09

Fig. 8. The shear modulus over time for two temperatures for epoxy adhesive (creeping function)
3.4. Synthesis: Notion of Critical Shear Strength and Critical Transfer Length after
Adhesive Creep
Considering the results established above, we can recalculate the load supported by the
adhesive and the critical length for life time prediction of 50 years (Fig. 9). We note that the
maximum shear stress is limited at 6 MPa and the length increases from 60 to 100 mm.
5000
4500

S1: Es=210 000MPa ; Gadh=4 500MPa

4000

Load (kN)

3500
3000
2500

S2: Es=210 000MPa ; Gadh=1 500MPa

2000
1500
1000
500
0
0

50

100

150

200

250

Length (mm)

Fig. 9. Influence of adhesive creeping on load transfer for a life time prediction of 50 years
Now, if we consider the load transfer between concrete and adhesive taking into account
the variation of mechanical properties for the two materials during 50 years, Fig. 10

135

confirms that the maximum stress is limited to 1.0 MPa in the concrete and the critical
length to 250 mm.
45000
40000
C1: Ec=3 000MPa ; Gadh=4 500MPa ; adh=19MPa

35000
30000

C2: Ec=10 000MPa ; Gadh=1 500MPa ; adh=6MPa

Load (kN)

25000
20000
15000
10000

C3: Ec=30 000MPa ; Gadh=4 500MPa ; adh=6MPa

5000
0
0

50

100

150

200

250

Length(mm)

Fig. 10. Load transfer between concrete and steel for different durations
4. CONCLUSION: DESIGN DATA FOLLOWING EUROCODE
RECOMMENDATION
In as far as the failure of the adhesive joint occurs in the concrete or the adhesive, the shear
stress at the ULS of the interface is equal to
f tj

adhu ,d min adh * adh ,e ; with adh 0,8 if TG 50C et adh 0,4 if TG 50C (9)
adh td

Considering durability conditions the shear characteristics of the adhesive at the SLS:

adh, d min adh. f *

adh,e
adh

f ij
with adhf 0,4 at 20C if Tg > 50C
td
adhf 0,1 at 20C if Tg < 50C

(10)
(11)

In the case of load transfer by bonding joins between concrete and steel, we recommend a
discontinuous connection by adhesives blobs with a critical length equal at:
critical 150mm for t=0
and
critical 250mm for t=50years

136

5. CONCLUSION
We have established how it is possible to determine the resistance of an adhesive joint
according to the calculation of its efficient length in function of a design shear stress adu, d .
The evaluation of the weighting coefficient for the stress level of the adhesive and the
variation of the shear modulus have been determined by viscoelastic properties
measurements.
From a technological point of view we recommend gluing by blobs incorporating zones of
hard adhesive (epoxy) and zones of flexible adhesive (polyurethane mastic). Following this
principle of discontinuous assembly, the proposed dimensioning method aims to verify the
equality between the maximum shear flux induced either by the resistance of the concrete
in compression or the steel in tension and the sum of the shear forces supported by each
blob.
6. REFERENCES
1.

Meier, U. 1995. Strengthening of structures with carbon fiber epoxy laminates.

Construction and Building Materials, 9(6): 341-351.
2. Varastehpour H. and Hamelin P. 1996. Experimental study of RC beams strengthened
with FRP plates, Proceeding of ACMBS-2, Canadian Society for Civil Engineering,
Montreal, Quebec, Canada, pp. 527-536.
3. Taljsten, B., Carolin, A., and Nordin, H., 2003. Concrete structures strengthened with
near surface mounted reinforcement of CFRP, Journal of Advanced Structural
Engineering, 6(3): 201-213.
4. Hamelin, P. 2010. La connexion par collage p227-254, Pont mixtes Acier Bton, Un
guide pour des ouvrages innovants, Projet National MIKTI Presse des ponts ISBN
978-2-85978-449-2, 457 p.
5. Volkersen O. 1965. Recherche sur la thorie des assemblages colls, Construction
mtallique (No. 4), pp. 3-13.
6. Schapery, R.A. 1996 Characterization of non linear time dependant polymers and
polymeric composites for durability analysis, Progress in durability analysis of
composite Systems, Edited by Cardon-Fukuda, pp. 21-36.
7. Brinson, L.C., Gates, T.S. 1995. Effects of physical aging on ling term creep behaviour
of polymers and polymers matrix composites, International Journal of Solid and
Structures, 32: 827-846
8. Huet, C. 1963. Etude par une mthode dimpdance du comportement viscolastique
des matriaux hydrocarbures, Thse Doct. Ing. Facult des sciences de lUniversit de
Paris, 210 p.
9. Hamelin, P. 1979. Contribution ltude du comportement rhologique de liants
viscolastiques en vue de lanalyse du fluage des matriaux composites utiliss sous
forme de structures du Gnie Civil, Thse de doctorat es Sciences, Universit Claude
Bernard, 161 p.
10. Yong, S. 1984. Modelisation of viscoelastic behaviour of carbon-epoxy unidirectional
composite. European Mechanics Colloquium (182), Brussel.

137

138

CDCC-11

DURABILITY AND SUSTAINABILITY OF FIBRE REINFORCED POLYMER

(FRP) COMPOSITES FOR CONSTRUCTION AND REHABILITATION
DURABILIT ET CYCLE DE VIE DES MATRIAUX COMPOSITES
EN POLYMRES RENFORCS DE FIBRES (PRF) EN
CONSTRUCTION ET RHABILITATION

DURABILITY STUDY OF GLASS FRP BARS SUBJECTED TO

EXTREME ENVIRONMENTAL AND LOADING CONDITIONS
M. Robert1, A. Fam2 and B. Benmokrane3
1

Postdoctoral Fellow, Department of Civil Eng., University of Sherbrooke, Quebec, Canada

Canada Research Chair Professor, Dept. of Civil Eng., Queens University, Kingston, Canada
3
Canada & NSERC Research Chairs Professor, Department of Civil Engineering, University of
Sherbrooke, Quebec, Canada
2

ABSTRACT
Fiber-reinforced polymer (FRP) materials have emerged as a practical alternative material
for producing reinforcing bars for concrete structures. This is due to their relatively low
cost-to-performance ratio and noncorrosive nature compared to traditionnal steel
reinforcing bars. In addition, FRP materials exhibit properties, such as high tensile strength,
that make them suitable for use as structural reinforcement. However, their durability in an
alkaline environment is still of concern and factors that can affect the long-term behavior of
GFRP materials have to be investigated. Moisture, high pH of surrounding environment,
extreme temperature and the presence of microcracks can affect the long-term properties of
FRP reinforcing bars. This paper summarizes the long-term durability study of glass fibrereinforced polymer (GFRP) reinforcing bars subjected to different conditionnings
simulating the real environment of application. In particular, the behavior of GFRP bars
subjected to moisture, high pH of moist concrete, extreme temperatures, and to tensile
prestress was investigated showing the great durability of GFRP reinforcing bars.
1. INTRODUCTION
Many steel reinforced concrete infrastructures exposed to harsh environments present
durability issues and are likely not to reach the lifetime expected or have already attained
the service limit. Steel corrosion induced by chlorides and/or carbonation are the main
causes of deterioration of steel reinforced concrete structures. The amplitude of the
phenomenon and economic factors have pushed research engineers, industries and public
institutions to develop new technologies to better protect reinforcing steel or providing non
corrosive materials in replacement of steel [1]. In anticipation of offering an alternative
material, research works have been conducted to develop and use fibre reinforced polymer
(FRP) composites in civil engineering applications.
One of the biggest challenges for the acceptance of FRP in civil engineering applications
concerns their durability and more specifically their capacity to maintain their structural
139

performance in severe and changing environmental conditions of service [2]. The

degradation process of GFRP reinforcing bars may be caused by the deterioration of one (or
more) of its phase: fibre, resin matrix or interface [3]. The degradation mechanisms are
more complex compared to those occurring in a monophasic material, such as steel. The
diffusion of moisture and alkalis from the concrete into the polymer matrix can lead to
changes in the mechanical, thermomechanical and physico-chemical properties of GFRP
reinforcing bars [4].
This paper presents an overview of several aspects affecting the long-term durability of
GFRP reinforcing bars subjected to laboratory accelerated aging simulating the genuine
environment of application. The durability of GFRP bars submitted to various
environmental conditions such as wet concrete, extreme temperatures and microcracksinducing pre-loading will be evaluated though the residual mechanical properties.
2. EXPERIMENTAL PROGRAM
2.1 Material
Experimental results presented in this paper were obtained during research work conducted
in the laboratory of durability of FRP composite materials of the department of civil
engineering of the University of Sherbrooke, Sherbrooke, Canada. Sand-coated glass FRP
reinforcing bars manufactured by a Canadian company (Pultrall inc., Thetford Mines,
Quebec, Canada) are used in this study. The bars are made of continuous E-glass fibres
embedded in vinylester matrix using a pultrusion process. The GFRP bars used have a
nominal diameter of 12 mm.
2.2 Test Plan
2.2.1 Aging in Moist Concrete
The aging was performed by immersing the mortar-wrapped GFRP bars in a wood
container specially manufactured for the study. The water level was kept constant
throughout the study to avoid a pH increase which could be due to a water level decrease
and a significant increase of the concentration of the alkaline ions in the solution. The
specimens were completely immersed at three different temperatures (23, 40 and 50oC) and
were removed from the water after four different periods of time (60, 120, 180 and 240
days). After each period, usually six GFRP bars were removed from the water, tested under
tension to compare their tensile strength retention values to those of the reference
specimens. Microstructural and physical analyses were also performed after immersion.
2.2.2 Extreme Temperatures
Properties of GFRP bars submitted to extreme temperatures were measured. All GFRP bar
specimens were tested under extreme temperature conditions using a MTS environmental
chamber cooled by liquid nitrogen. Sub-zero temperatures equal to -100o, -80o, -60o, -40o, 20o and also 0oC were chosen to simulate the effect of Nordic climate on GFRP bars. The
temperatures below -60 oC were used to amplify the effects of temperature on mechanical
140

and microstructural behaviour of the FRP materials. Elevated temperatures equal to 50,
100, 150, 200, 250 and 325oC were chosen to simulate short-term environmental conditions
in the case of fire. All the tests were conducted using a steady-state temperature (heat then
load) regime.
2.2.3 Pre-Loading of GFRP Bars Samples
The effect of cracking and microcracking on the long-term behaviour of GFRP bars was
evaluated. GFRP bars were preloaded at 80 percent of their theoretical ultimate tensile
strength (854 MPa) (UTS) to initiate cracks and microcracks in polymer and glass fibres.
All bars were preloaded under tension according to ASTM D 7205 standard. The rate of
loading ranged between 250 and 500 MPa/min and the maximum load was maintained for
10 minutes, and then was reduced at the rate of 250 to 500 MPa/min. The study of the longterm behaviour of pre-loaded GFRP bars was performed by using accelerated aging in
moist moist concrete, as described above.
2.3 Mechanical Characterization
Tensile tests were performed on bars aged in solution, bars subjected to controlled
temperatures and bars pre-loaded at 80 percent of their UTS. The measured tensile
strengths of the bars before and after exposure were considered as a measure of the
durability performance of the specimens. All bars were tested under tension according to
ASTM D 7205 standard. Each specimen was instrumented with a Linear Variable
Differential Transformer (LVDT) to capture the elongation during testing. The rate of
loading ranged between 250 and 500 MPa/min. The applied load and bar elongation were
recorded during the test using a data acquisition system monitored by a computer. Due to
the brittle nature of GFRP, no yielding occurs and the stress-strain behaviour was linear.
2.4 Physico-chemical Characterization
Scanning Electron Microscopy (SEM) observations and image analysis were performed to
evaluate the microstructure of aged specimens and the integrity of the GFRP material after
different conditionings. These observations were conducted to determine the potential
degradation of the polymer matrix, possible glass fibers and interface.
Glass transition temperature was determined by Differential Scanning Calorimetry (DSC)
in accordance with ASTM E 1356 standard for both the reference specimens and specimens
exposed to conditionings. If decrease in Tg was observed for conditioned samples, it was an
indication of plasticizing effect or chemical degradation. Aged sample maintaining a lower
Tg than for the reference showed an irreversible chemical degradation.
Thermal stability of specimens was investigated by using Thermogravimetric Analysis
(TGA), and the weight loss traces were recorded as function of temperature in the range of
20800 C according to ASTM E 1868 standard. The mass variations recorded could give
indications concerning the different phenomena of degradation occurring in the matrix and
also explain the loss of mechanical properties observed at very high temperatures.

141

Chemical degradation of polymer matrix was characterized by using Fourrier Transform

Infrared Spectroscopy (FTIR). The most interesting region of the FTIR spectra is located
between 3300 cm-1 and 3600 cm-1, which corresponds to the stretching mode of the
hydroxyl groups of the vinylester resin. When a irreversible chemical degradation occurs,
such as polymer hydrolysis reaction, new hydroxyl groups are formed and the
corresponding infrared band increases. Changes in the peak intensity were quantified by
determining the ratio of the OH peak to the carbon-hydrogen stretching peak of the resin,
which is not affected by the conditioning.
2.5 Arrhenius Relation
Equation (1) expresses the Arrhenius relation, in terms of the degradation rate [5]
Ea
k A exp
(1)

RT
where k = degradation rate (1/time); A = constant relative to the material and degradation
process; Ea = activation energy of the reaction; R = universal gas constant; and T =
temperature in Kelvin. The primary assumption of this model is that only one dominant
degradation mechanism of the material operates during the reaction and that this
mechanism will not change with time and temperature during the exposure [6]. Only the
rate of degradation will be accelerated with the temperature increase. Equation (1) can be
transformed into:
1 1
E
exp a
k A
RT
1 E 1
ln a ln A
k R T

(2)
(3)

From Equation (2), the degradation rate k can be expressed as the inverse of time needed
for a material property to reach a given value [6]. From equation (3) one can further
observe that the logarithm of time needed for a material property to reach a given value is a
linear function of 1/T with the slope of Ea /R [6]. Ea and A can be easily calculated by using
the slope of the regression and the point of intersection between the regression and the Y
axis respectively.
3. EXPERIMENTAL RESULTS
3.1 Effect of Immersion
3.1.1 Tensile Strength Retention
Table 1 shows the experimental results obtained during the tensile tests concerning the
ultimate strength of reference and pre-loaded samples aged in moist concrete and tested
after immersion. Figure 1 shows the retention of the ultimate strength of reference and preloaded bars aged in moist concrete according to the duration of immersion of embedded
bars at various temperatures. Table 1 shows a slight decrease of the ultimate tensile strength
142

with immersion duration. The recorded results show that the longer the time of immersion,
the larger the loss of resistance. Furthermore, it is clear that the temperature of immersion
affects the loss of resistance. It can be seen that for duration of immersion of 8 months, the
loss of resistance for aged reference bars is equal to 16, 10 and 9 % at 50o, 40o and 23oC,
respectively. Results concerning aged pre-loaded bars presented in Table 1 show slight
decrease (6 to 11%) of the ultimate tensile strength after 240 days of immersion of preloaded bars embedded in moist concrete. This decrease was similar to the loss of tensile
strength measured on same bar subjected to same conditionings but without preloading [7].
This phenomenon was due to the increasing of diffusion rate of the solution into the sample
due to the presence of cracks and microcracks and to the acceleration of chemical reactions
of degradation with the temperature of immersion, leading to a larger absorption rate of the
solution for the same time of immersion. It was also noted that the loss of elastic modulus
was negligible for both, reference and pre-loaded bars aged in moist concrete.
Table 1. Experimental tensile strength of reference and pre-loaded specimens aged in
moist concrete
Time of immersion
(days)
0
60
120
180
240

Temperature
(oC)
23
23
40
50
23
40
50
23
40
50
23
40
50

Reference Bars
Mean Tensile
COV
Strength (MPa)
(%)
788
6,9
753
8,2
755
3,7
767
3,8
702
2,0
666
7,8
720
3,2
717
2,6
708
4,8
711
2,5
714
3,5
708
7,1
665
9,3

Pre-Loaded Bars
Mean Tensile
COV
Strength (MPa)
(%)
854
2.1
846
5.6
847
6.4
838
4.2
849
2.5
832
8.1
837
3.9
836
3.3
823
5.7
808
3.1
810
4.2
784
7.9
768
9.0

a)
b)
Fig. 1. Tensile strength retention for: a) reference bars aged in moist concrete, and b) preloaded bars aged in moist concrete

143

The tensile test of unconditioned specimens, specimens aged in moist concrete at 50oC
during 240 days and pre-loaded specimens aged in moist concrete at 50oC during 240 days
showed an approximately linear behaviour up to failure. Specimens failed through the
rupture of fibres. Micelli and Nanni [8] also observed similar tensile failure modes of
GFRP bars. The failure was accompanied by the delamination of fibres and resin, as shown
in Figure 2. From this observation, it could be concluded that the aging in moist concrete or
the presence of pre-existing cracks and microcracks due to the pre-loading of bars, have no
significant effects on the failure mode occurring during tensile tests.

Reference

Aged in moist concrete

Fig. 2. Typical failure mode for GFRP bars

3.1.2 Microstructural Effect of Pre-Loading
The micrographs of Figure 3 show the longitudinal bar surface of reference and GFRP bar
pre-loaded at 80 percent of the UTS. Observations of fibres/matrix interfaces and of the
microstructure in general demonstrated that the pre-loading affected the microstructure of
GFRP bar. The only visible damage occurred at the fibre level since the elongation at the
rupture was lower for the fibres compared to the polymer matrix. No damage occurred at
the matrix and fibre/matrix interface.

a)
b)
Fig. 3. Micrograph of longitudinal GFRP bar: a) Reference GFRP bar at low magnification,
and b) GFRP bar loaded at 80% of the UTS at high magnification
3.1.6 Microstructural Effect of Aging in Moist Concrete
The micrographs of Figure 4 show the fibres/matrix interface for reference and pre-loaded
bars aged in moist concrete at 50oC during 240 days. Observation of these interfaces, and of
144

the microstructure, in general, demonstrate that the conditionings of mortar-wrapped bars in

water do not affect the microstructural properties of the GFRP bars, even if the bars show
pre-existing cracks or microcracks.

(a)
(b)
Fig. 4. Micrographs of fibre/matrix interface (X3000): a) Unconditioned bars, b) Preloaded GFRP bar specimen aged in moist concrete at 50oC during 240 days.
3.1.7 Effect on Polymer Matrix
The change in the Tg of GFRP bars embedded in concrete after aging in water was also
investigated. No significant change of the Tg was measured after the aging of reference and
pre-loaded bars in moist concrete during 240 days at 50oC. In fact, the Tg of reference
specimen was 127oC, the Tg of reference specimens aged in moist concrete was 129oC, and
the Tg of pre-loaded specimens aged in moist concrete was 125oC. These results clearly
show that even after 240 days in salt solution at 50oC, the polymer matrix is not affected.
A FTIR analysis of unconditioned bar specimen, and preloaded (80% of the UTS) mortarwrapped bars and aged in moist concrete at 50oC for 240 days was conducted (Figure 5).
The experimental ratio of the OH peak to the carbon-hydrogen stretching peak of the resin
for the 12.7 mm diameter preloaded concrete-wrapped samples and immersed in water for
240 days at 50oC was 0.53 compared to 0.51 for unconditioned samples. So, the hydroxyl
peak did not show any significant changes. This observation led to the conclusion that no
chemical degradation of the polymer occurred during the immersion of preloaded concretewrapped bars.

Fig. 5. FTIR spectra of unconditioned specimens and pre-loaded specimens aged in moist
concrete
145

3.1.8 Long-Term Predictions

Following the procedure proposed by Bank et al. [9], the natural logarithm of time to reach
a set of levels of normalized performances versus 1/T, expressed as the inverse of absolute
temperature (1000/K), was used to predict the service life at the Mean Annual Temperature
(6oC) in Montral, Qubec, Canada. A coefficient of determination (R2) value close to 1 is
desired. However, the ASTM procedures recommend a minimum value of 0.80 for
acceptability and the obtained R2 values are between 0.94 and 0.99. From the Arrhenius
plot, the service life time necessary to reach the established tensile strength retention levels
(PR) can be extrapolated for any temperature. Consequently, predictions are made for
tensile strength retention as a function of time for an immersion at 6C and the general
relation between the PR and the predicted service life at the average temperature of 6oC are
drawn (Figure 6). The predicted time to reach the determined strength property retention
level (PR) for GFRP bars embedded in moist concrete without any preloading at an
isotherm temperature of 6oC is approximately 1 and 210 years for a PR of 90 and 75%,
respectively. Also, the predicted service life of GFRP bars embedded in moist concrete at
an isotherm temperature of 6oC to reach a PR less than 75% can be estimated to be infinite.
These predictions show that the GFRP bars are durable face to the concrete environment,
which is supposed to be well simulated by the immersion of embedded GFRP bars in water.

Fig. 6. General relation between the PR and the predicted service life at a mean annual
temperature of 6oC (Montreal, Quebec)
3.2 Effect of Extreme Temperatures
3.2.1 Effect of Extreme Temperatures on Mechanical Properties
Figure 7 shows the relation between the tensile strength of GFRP bars and the test
temperature respectively. Three different zones were identified in Figure 7: 1) the reference
plateau between -40o and 50oC; 2) the zone where the stiffness increasing at temperatures
lower than -50oC; and 3) the mechanical strength decreasing zone for temperatures near and
above the glass transition temperature (Tg) of the polymer. The mechanical strengths
between -40o and 50oC were stable. In this range of temperatures, the molecular chain
mobility of the polymer did not change since the temperature was below Tg. This result
146

showed that for standard environmental conditions of Nordic countries as Canada and north
of USA (temperature ranging from -40 to 50 oC), the mechanical properties of GFRP bars
were not changed. For temperatures lower than -50oC, the molecular chains mobility of the
polymer decreased, leading to an increase of the mechanical stresses needed to rupture the
material. Below Tg (around 120oC), the matrix was in a glassy state. When increasing the
temperature and reaching the decomposition region, the breaking of molecular bonds
started and the ductility of the material increased, leading to a decrease of mechanical
strengths and stiffness of the material. At very high temperatures (greater than 300oC),
strong degradation of the polymer occurred (combustion, oxidation) and load transfer
provided by the matrix was severely reduced.

Fig. 7. Tensile strength of GFRP bars tested at different temperatures

3.2.2 Effect on Polymer Matrix
Figure 8 shows the mass variation of specimens as a function of temperature measured by
TGA.

Fig. 8. Mass variation of GFRP bar specimen when heated between 20o and 800oC.
147

It was seen that major weight loss occurred between 300 and 450oC. This important drop,
up to 18%, was due to the thermal degradation of the polymer. At these temperatures,
thermal degradation occurred and the molecular chains of the polymer broke, leading to the
formation of microcracks both at the fibre/matrix interface and in the matrix phase (Figure
6). The weight losses measured by TGA showed a major degradation of the resin from
300oC to 450oC. Below this range of temperatures of degradation, the properties were
recovered when the GFRP specimens were heated to the desired temperature and then
tested at the ambient temperature. At this point, the experimental results suggest that
irreversible loss of mechanical properties occurred above 300oC.
The Tg was measured by DSC for reference and specimen heated at 350oC for 2 hours. The
Tg of the reference specimen was equal to 113oC, whereas the specimens heated at 350oC
for 2 hours was equal to 67oC. This showed that the matrix weakened when heated at
elevated temperatures, which explained the causes of the observed large decrease of
mechanical properties.
3.2.3 Microstructural Effect of Extreme Temperatures
The polymer matrix degradation was confirmed by microstructural analysis. The
micrographs presented in Figure 9 show the polymer, the fibre and the interface between
the fibres and matrix for reference specimen and specimen conditioned at 350oC for two
hours. The comparison of these micrographs showed that significant damage occurred for
GFRP bar conditioned at 350oC (Figure 9b) as compared to reference sample (Figure 9a).
The presence of microcracks in specimen tested at 350oC confirmed the degradation of the
polymeric matrix and explained the loss of mechanical properties.

(a)
(b)
Fig. 9. Micrographs of longitudinal fibre/ matrix interface of: a) unconditioned GFRP bar
specimen, b) Conditioned GFRP bar specimen aged in air at 350oC for 2 hours.
4. CONCLUSIONS
Based on the results of this study the following conclusions may be drawn:
1- Even at high temperature (50oC), where the environment is the more aggressive, the
change in tensile strength of GFRP bars embedded in moist concrete is minor.

148

2- No significant microstructural changes were observed after 240 days immersion of

reference and pre-loaded GFRP bars embedded in concrete in tap water at 50oC.
3- The polymer matrix is not affected by moisture absorption and high temperatures: no
changes of the glass transition temperature occur as observed by DSC and no
irreversible degradation occurred to the polymer matrix as observed by FTIR. The
preloading of GFRP bars didnt lead to any further degradation.
4- After 240 days water immersion of pre-loaded bar embedded in mortar, the retention
rate of tensile strength are 95, 96 and 98% at 50, 40, and 23oC, respectively.
5- According the to long-term predictions, the tensile strength retention of GFRP bars
embedded in moist concrete will decreased by 20 and 25% after 100 and 200 years
respectively. It was shown that the service life time allowed to reach tensile strength
retention of less than 50% should be infinite.
6- Tensile properties of GFRP bars increased when the temperature decreased. This
phenomenon was due to the increase of stiffness of the amorphous polymer matrix due
to low temperatures.
7- At severe temperatures experienced in Northern regions such as Canada (temperature
ranging from -40o to 50oC), the tensile strength appeared to be stable, which showed that
the mechanical behavior of GFRP bars was not affected by temperature in this range of
temperature
5. REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.

Nkurunziza, G., Benmokane, B., and Debaiky, A.S., Masmoudi, R. 2005. Effect of
sustained load and environment on long-term tensile properties of glass FRP
reinforcing bars. ACI Structural Journal, 102(4): 615-621.
Robert, M., Cousin, P., Wang, P., Benmokrane, B. 2010. Temperature as an
Accelerating Factor for Long-Term Durability Testing of FRPs: Should there be any
limitations? Journal of Composites for Construction, 14(4): 361-367.
Benmokrane B, Wang P, Ton-That T, Rahman H, Robert J. 2002. Durability of glass
fibre reinforced polymer reinforcing bars in concrete environment. Journal of
Composite for Construction, 6(2): 143153.
Benmokrane, B., Wang, P., Pavate, T., and Robert, M. 2006. Durability of FRP
Composites for Cvil Infrastructure Applications, Chapter 12. Whittles Publishing,
Scotland, pp. 300-343.
Nelson, W. 1990. Accelerated testingStatistical models, test plans, and data analyses.
Wiley, New York, 601 p.
Chen, Y., Davalos, J. F., Ray, I. 2006. Durability Prediction for GFRP Bars Using
Short-Term Data of Accelerated Aging Tests. Journal of Composites for Construction,
10(4): 279-286.
Robert, M., Cousin, P., and Benmokrane, B. 2009. Durability of GFRP Bars Embedded
in Moist Concrete. Journal of Composites for Construction, 13(2): 66-73.
Micelli, F., and Nanni, A. 2004. Durability of FRP rods for concrete structures.
Construction Building Materials, 18(7): 491503.
Bank, L.C, Gentry, T.R, Thompson, B.P, Russel, J.S. 2003. A Model Specification for
Composites for Civil Engineering Structures. Construction and Building Materials,
17(6-7): 405-437.

149

150

CDCC-11

DURABILITY AND SUSTAINABILITY OF FIBRE REINFORCED POLYMER

(FRP) COMPOSITES FOR CONSTRUCTION AND REHABILITATION
DURABILIT ET CYCLE DE VIE DES MATRIAUX COMPOSITES
EN POLYMRES RENFORCS DE FIBRES (PRF) EN
CONSTRUCTION ET RHABILITATION

FIRE TESTING OF FRP STRENGTHENED REINFORCED

CONCRETE COLUMNS
N. Bnichou1, D. Cree2, E.U. Chowdhury2, M.F. Green2 and L.A. Bisby3
1

Institute for Research in Construction, National Research Council of Canada, Ottawa, Canada
Dept. of Civil Engineering, Queens University Kingston, Ontario, Canada
3
Institute for Infrastructure and Environment, University of Edinburgh, Scotland, UK
2

ABSTRACT
This paper presents the results of full-scale fire tests on two concrete columns strengthened
with fibre reinforced polymer sheets and protected with a new insulation system. One
column was circular while the other was square. The columns were subjected to the ULCS101 standard fire. Both columns obtained fire endurance ratings of over 4 hours. At the
end of the fire test, the columns were loaded to failure. The temperatures in the columns
during the fire test and the load capacities of the columns will be presented. Additional
work is required to compare the results against thermal and structural numerical models
developed specifically for circular and square columns.
1. INTRODUCTION
Fibre reinforced polymers (FRPs) are now widely applied for repairing concrete structures
[1, 2]. FRP are known to have excellent strength, stiffness, and corrosion resistance, but
their mechanical properties are significantly reduced at elevated temperatures due to the
properties of the matrix resin. As such, concerns regarding fire resistance have limited FRP
applications in buildings. To address this concern, a new fire insulation material is applied
on the surface of FRP wraps to obtain appropriate fire resistance. Novel materials for
potential use inside a building, including FRPs and insulation materials must pass certain
fire exposure limits as prescribed in North America by CAN/ULC S101 [3] or ASTM E119
[4]. The objective is for the structure protected with the new insulation material to
withstand a fire resistance (fire endurance) test by being exposed to a standard fire for a
specific duration, under strict conditions. Until the fire endurance property of this unique
supplemental insulating fire protection system is known, it will hinder its extensive use as a
reinforcement coating on the interior of buildings.
2. MOTIVATION
To gain insightful information on fire endurance, there has been an on-going collaborative
research program in the development of various insulating materials for protecting FRP
151

wrapped reinforced concrete columns in fire conditions. The initial tests consisted of two
circular columns (Column 1 and Column 2) [5], followed by another three circular columns
(Column 3, Column 4 and Column 5) [6]. The current study reports the work done on one
circular column (Column 5) [6] and one recently tested square column (Column 6). A team
comprising of the National Research Council of Canada (NRC), Queens University,
Canada and industry partner Sika Canada are collaborating on the current research project.
This project involves the full-scale fire test of two concrete columns strengthened with FRP
wrap and insulated with a supplemental fire protective coating. The column fabrication,
FRP strengthening, the insulation fire protection material and fire test results are discussed.
The objectives for these fire tests are to investigate the behaviour of FRP strengthened
reinforced concrete columns protected with an insulation layer and subjected to elevated
temperatures. The strengthened and insulated columns were subjected to the ULC-S101
standard fire, under sustained service load, which simulates the load that the columns
would be expected to experience in an actual fire situation. The results of both experiments
(Columns 5 and 6) are presented in this paper.
3. EXPERIMENTAL PROCEDURE
3.1 Column Test Specimens
The experimental program consisted of two fire tests on one circular column and one
square reinforced concrete column strengthened with the SikaWrap Hex 103C FRP
externally-bonded strengthened system and insulated with Sikacrete-213F, a sprayapplied fire protection mortar. The columns were designated as Column 5 and Column 6,
respectively and photos of the columns before fire testing are shown in Figure 1. Column 5
was 400 mm in diameter while the cross-section of Column 6 was 305 mm by 305 mm.
Both columns had a height of 3810 mm and were designed for 28 MPa concrete strength
using Type 10 Portland cement with crushed carbonate aggregate, maximum aggregate size
of 14 mm, and natural sand as the fine aggregate.
The longitudinal steel reinforcement in Column 5 consisted of eight 20M (19.5 mm
diameter) deformed bars, symmetrically placed with 40 mm clear cover to the spiral
reinforcement. The lateral reinforcement for the columns consisted of 10M (11.3 mm
diameter) deformed steel spiral with a centre-to-centre pitch of 50 mm. The longitudinal
reinforcing bars and the steel spiral had average yield strengths of 456 MPa and 396 MPa,
respectively. The primary longitudinal reinforcement bars for Column 6 were four 25M
deformed steel bars. The lateral reinforcement for the column consisted of 10M deformed
steel bars spaced at 305 mm. The main reinforcing bars and ties all had average yield
strengths of 477 MPa. The steel reinforcement had a clear cover of 40 mm from the exterior
surface of the concrete to the steel ties and a 50 mm concrete cover to the principal
reinforcement. Column 5 was poured vertically into a 400 mm inside-diameter Sonotube,
while Column 6 was cast vertically in a plywood formwork. The columns were
instrumented with Chromel-alumel (Type-K) thermocouples to record the internal
temperatures within the concrete, the reinforcing steel bars, the concrete-FRP interface, the
FRP-insulation interface and the outer surface of the insulation.

152

3.2 FRP Strengthening and Fire Protection

Both columns were strengthened with a SikaWrap Hex 103C unidirectional carbon/epoxy
FRP strengthening system with a Sikadur 300 resin epoxy saturant/adhesive. Column 5 was
strengthened with two layers of FRP wrap while Column 6 was strengthened with three
layers. The carbon FRP has an ultimate tensile strength of 849 MPa and tensile elastic
modulus of 70.5 GPa. To provide fire protection, Sikacrete-213F, a supplemental fire
insulation material, was spray-applied over the surface of the FRP wrapped columns. Prior
to the spray application for Column 5, a 1/8 th inch diameter steel mesh, having 50 mm by
50 mm openings, was attached to the surface of the column to provide reinforcement for the
protection system. Column 5 had an average insulation thickness of 44 mm, while Column
6 had an average insulation thickness on flat surfaces of 40 mm and an average corner
insulation thickness of 51 mm.
3.3 Fire Endurance Test
The fire endurance experiments were conducted by exposing the columns to heat in a
furnace specially built for fire testing loaded columns. The columns were installed in the
furnace by bolting the end plates to the test frame loading head at the top and a hydraulic
jack at the bottom. This resulted in a fixed-fixed end condition for the column, which
simulates the end conditions to be expected in an actual building. Column 5 was loaded in
compression until the applied load on the circular column was 3031 kN, and Column 6 until
the applied load on the square column was 1717 kN. The respective loads were maintained
at a constant value throughout the fire endurance test. Both columns did not fail under their
respective sustained service loads, therefore after 4 hours of fire exposure; the loads were
increased until failure was observed.

(a)
(b)
Fig. 1. Images of the FRP-wrapped and insulated concrete columns (a) circular and (b)
square, prior to fire test.
153

4. RESULTS AND DISCUSSION

4.1 Temperature Behaviour
The fire behaviour of the insulation, FRP wrap, and crack propagation in the insulation in
Column 5 and Column 6 were monitored during the test through small observation
windows (view ports) in the walls of the furnace. Prior to commencing the fire tests, it was
observed on both columns that minor cracks developed on the surface of the dried
insulation, which were due to shrinkage. The initial cracks were approximately less than 5
mm wide but widened to about 10 mm as the test progressed. For both columns, flames
were seen emanating through the small cracks on the surface of the insulation at
approximately 2 hours into the fire test. Cracks on the surface of the insulation, prior and
during the fire endurance test, are depicted in Figure 2. The steel mesh reinforcement
within the insulation of Column 5 may have prevented the spalling of the insulation from
the column surface, however, no spalling of the insulation system was observed from
Column 6 which did not contain a mesh. Prior to the column failures, the fire protection
insulation system remained intact with minor external cracks throughout the length of the
column for the duration of the fire test.
The temperature results for Column 5 and Column 6 at mid-height for horizontally
positioned thermocouples are given in Figure 3 and Figure 4, respectively. The temperature
measurement readings are for the insulation surface, FRP surface, concrete surface and
internal to the concrete. Both Figures 3 and 4 also include the standard fire curve. During
the test on Column 5, a dip in the average furnace temperature was a result of an abrupt
furnace stop at 3 hours into the fire test and required 12 minutes to bring the temperature
inside the chamber back to the prescribed temperature-time curve.

(a)
(b)
Fig. 2. (a) Typical cracks on the surface of the fire insulation system prior to the fire test,
and (b) flame from the cracks on the surface of the fire insulation system during the fire
test.

154

Insulation surface
Concrete surface
Cover 150 mm

1000

Temperature (C)

Temperature (C)

ASTM E-119
FRP surface
Cover 50 mm

ASTM E-119
Concrete surface corner
Concrete surface
Cover 153 mm

800
600
400
200

Insulation surface
FRP surface
Cover 63 mm

1000
800
600
400
200
0

0
-1

2
3
4
Time (hours)

-1

Time (hours)

Fig. 3. Column 5 temperature of thermocouples

at mid-height as a function of time.

Fig. 4. Column 6 temperature of thermocouples

at mid-height as a function of time.

The polymer resin (Sikadur 300) has a glass transition temperature (Tg) of 60C if cured at
ambient temperature. The concrete surface temperature for Column 5 (Figure 3) was shown
to have an average temperature of 60.5C after 29 minutes of fire exposure. This
demonstrated that the insulation system limited the time at which the temperature was
below the Tg of the polymer resin, which is important for the bond performance. The
temperatures recorded at the concrete surface increased at a steady rate until they reached a
plateau at 100C. This plateau is due to moisture evaporation within the insulation. After
approximately 1 h 30 min, the concrete surface temperature began to increase at a higher
rate up to 266C after 4 hours. The temperature at the concrete surface remained below the
manufacturer stated Tg for about 29 minutes, by providing a thickness of approximately 44
mm of insulation. After 4 hours, the highest temperatures recorded at the FRP surface and
concrete surface were 408C and 266C, respectively.
Column 6 had an average concrete surface temperature of 60.1C (Figure 4) after 33 minutes
of fire exposure. Similarly, the corner thermocouples at mid-height detected an average
concrete surface corner temperature of 60.2C (Figure 4) after 22 minutes of fire exposure.
This showed that the corner FRP increased in temperature quicker than the sides. The data
demonstrated that it is possible to maintain the temperature of the concrete surface below
the Tg of the FRP material for about 0.5 hrs by providing an insulation system of
approximately 40 mm thick. After 4 hours, the highest temperatures recorded from the
horizontally positioned thermocouples at the FRP surface and concrete surface were 406C
and 250C, respectively.
For both columns, the test data indicated the insulation system would be unable to maintain
the FRP temperature below the Tg of the polymer resin for prolonged periods of time during
fire. However, the insulation system was able to maintain the temperatures within the
concrete and reinforcing steel below 200C. For example, Figure 3 shows that after 4 hours,
the temperatures in Column 5 on the concrete surface, in the concrete at 50 mm from
surface (at the location of the reinforcement), and in the concrete at 150 mm from surface
are 266C, 148C and 98C, respectively. Similarly, Figure 4 shows that, for Column 6
after 4 hours, the concrete surface, the reinforcement temperatures, the concrete at 63 mm
from the surface, and concrete at 153 mm from the surface are 250C, 193C 164C and
155

136C, respectively. The insulation system was able to maintain the temperatures within
the concrete and reinforcing steel of both columns below 200C. By maintaining this
criteria, no significant deterioration in the mechanical properties of concrete and steel
occurred. Hence, it can be stated that the columns maintained their full unconfined axial
load carrying capacity for greater than 4 hours of exposure to the ASTM E119 standard
fire.
4.2 Structural Behaviour

Column 5-Axial displacement

Column 6-Axial displacement

Column 5-Axial load

Column 6-Axial load

5
0
-5
-10
-15
-20
-25
-30
-35
-40
-45
-50

0
-1000
-2000
-3000
Preload
phase

Fire test
phase

-4000

Applied load (kN)

Axial deformation (mm)

Figure 5 shows the axial deformations and load applied as a function of time for Columns 5
and 6 during the preload and fire test phase. A negative axial deformation or applied load
value indicates compression of the column. The columns experienced an initial load of
approximately 350 kN due to the weight of the hydraulic jack prior to the start of the
preload phase. Column 5 resisted the sustained concentric load of 3054 kN for 4.5 hours.
Therefore, after 4.5 hours, the load was steadily increased until the column reached failure
at a load of 4984 kN. The column failed by apparent crushing of the concrete core.
Similarly, Column 6 resisted the sustained fire load of 1717 kN for 4 hours. Since it did not
fail, the load level was increased to 2641 kN, at which point the column failed in a nonviolent fashion by crushing of the concrete core, which resulted in the insulation
delaminating at the centre of the column, while the insulation remained largely intact at the
extremities of the column. The temperature data obtained during the fire endurance test of
Column 5 and Column 6 show that the FRP wrap had been rendered ineffective even with
an insulation layer due to increased temperatures in excess of its glass transition
temperature. However, due to the superior thermal insulation system, the concrete and
reinforcing steel retained their room-temperature strength.

-5000
-2

-1

2
3
Time (hours)

Fig. 5. Applied load and axial deformation as a function of time for Columns 5 and 6.
The axial deformations of the columns were the result of a combination of load effects and
thermal expansion. During the preload phase of the fire test, Column 5 had an initial 7.3
mm axial deformation due to the applied value of 3054 kN. After the preload phase, the
axial load (3054 kN) was kept constant for 4.5 hours of fire exposure. No major expansion
was observed during heating up to 4.5 hours due to the relatively low temperatures in the
156

concrete. For instance, after 3 hours of fire exposure, the temperature of the rebar for
Column 5 was 105C and the concrete surface temperature was 225C. However, a severe
drop in axial deformation after 4.5 hours occurred from the increasing applied load up to
failure. At failure, Column 5 had an axial deformation of 32 mm.
Similarly, during the preload phase of the fire test, Column 6 had a 6 mm axial deformation
as a result of the applied load of 1717 kN. After the preload phase, the applied axial load
(1717 kN) was kept constant for 4 hours of fire exposure. After 1 hour of fire exposure, the
column began to experience a gradual expansion of approximately 2 mm for up to 3 hours
or until the end of the test. The source of expansion in Column 6 may be from the thermal
expansion of the concrete and a lack of spiral reinforcement as contained in Column 5. For
instance, the temperature of the rebar for Column 6 was 150C and the concrete surface
temperature was 200C after 3 hours. The drop in the deformation curve after 4 hours was
due to the increase in applied load to failure. Column 6 experienced a final axial
deformation of 6.8 mm. In general, the initial degradation for the compressive strength of
concrete is experienced between 200 and 250C, while at 300C strength reduction is in the
range of 15-40%. [7]. In addition, the thermal expansion of concrete is reduced when a
compressive strength of 0.45fc or higher is applied to the specimen during heating [8].
Based on the strength of Column 6 concrete, (36 MPa at 28 days, but tested after one year)
applied load and cross-sectional area, a small expansion from the concrete may have
occurred. The temperatures in the steel reinforcement remained low enough to prevent any
material property degradation. Normally, at 350C or higher, the yield strength of
conventional steel is reduced by 66% of its room temperature yield strength [9]. According
to the temperature of the reinforcement in both columns, the ductility did not increase nor
did the yield strength and ultimate tensile strength reduce.
5. CONCLUSIONS
From results of these full-scale fire endurance tests on one circular and one square FRP
wrapped reinforced concrete column, the following conclusions can be drawn:
1.

The insulation system was effective in protecting the FRP wrapped columns such that
they were able to achieve four hour fire endurance ratings according to ULC S101 and
ASTM E119.

2.

The insulation fire protection material was not able to maintain the temperature of the
FRP below its glass transition temperature for the duration of the fire endurance test.
With an average insulation thickness of 44 mm, the temperature of the concrete surface
for Column 5 reached its glass transition temperature of 60C at about 29 minutes into
the fire test, while Column 6 with 40 mm of insulation material had an FRP
temperature of 60C at 33 minutes into the fire exposure. The difference may be
partially attributed to the mechanism of heat transfer through the insulation in a
circular column versus a rectangular column.

3.

The supplementary insulation fire protection material used in this fire endurance test is
an effective fire protection system. However, the shrinkage cracks in the cured
insulation material should be investigated to further improve the system. Although
157

flames were observed emanating from cracks formed in the insulation of both columns,
the overall insulation remained intact for more than 4 hours of exposure to the ULC
S101 or ASTM E119 standard fire.
6. REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.

Neale, K.W. 2000. FRPs for structural rehabilitation: A survey of recent progress.
Progress in Structural Engineering and Materials, 2: 133-138.
Hollaway L.C. 2010. A review of the present and future utilisation of FRP composites
in the civil infrastructure with reference to their important in-service properties.
Construction and Building Materials, 24: 2419-2445.
ULC. 2007. Standard Methods of Fire Endurance Tests of Building Construction and
Materials. CAN/ULC-S101-07, Underwriters Laboratories of Canada, Scarborough,
ON, 70 p.
ASTM, Test Method E119-10b. 2010. Standard Methods of Fire Test of Building
Construction and Materials. American Society for Testing and Materials, West
Conshohocken, PA, USA, 33 p.
Bisby, L.A., Kodur, V.R., and Green, M.F. 2005. Fire endurance of fiber-reinforced
polymer-confined concrete columns. ACI Structural Journal, 102 (6): 883-891.
Chowdhury, E. 2009. Behaviour of fibre reinforced polymer confined reinforced
concrete column under fire condition. PhD thesis, Department of Civil Engineering
Queens University, Kingston, Ontario, Canada.
Georgali, B., and Tsakiridis, P.E. 2005. Microstructure of fire-damaged concrete-a case
study. Cement & Concrete Composites, 27: 255-259.
Thelandersson, S. 1982. On the multiaxial behaviour of concrete exposed to high
temperature, Nuclear Engineer Design, 75: 271-282.
Chijiiwa, R., Tamehiro, H., Funato, K., Yoshida, Y., Horii, Y., and Uemori, R. 1993.
Development and Practical Application of Fire-Resistant Steel for Buildings, Nippon
Steel Technical Report 58, pp. 47-55.

158

CDCC-11

DURABILITY AND SUSTAINABILITY OF FIBRE REINFORCED POLYMER

(FRP) COMPOSITES FOR CONSTRUCTION AND REHABILITATION
DURABILIT ET CYCLE DE VIE DES MATRIAUX COMPOSITES
EN POLYMRES RENFORCS DE FIBRES (PRF) EN
CONSTRUCTION ET RHABILITATION

AN INVESTIGATION INTO THE SUSTAINABILITY OF FRP

REINFORCEMENT BARS
M.C.K Pearson, T. Donchev and M. Limbachiya
Kingston University, London, UK

ABSTRACT
This paper explores the sustainability of FRP internal reinforcement in the construction
industry and comments on the uses currently employed. Sustainability has many focuses in
industry, but principally it is to ensure that the generations of the future are not negatively
affected by the impact of current developments. FRP has already proven itself to be a
contender in construction against steel with its continually lowering costs of manufacture.
The lower carbon emissions, longer life and correspondingly lower life cycle cost for FRP
reinforced RC elements and the relatively environmentally safe waste are between other
advantages of using those materials. This paper also looks at the manufacture of FRP
materials and how they can be used in industry for FRPs to be a material of the future. This
work expresses not only the idea of resources being renewable or reusable, but also if they
are economically effective in order to give a rounded approach to Sustainability.
1. INTRODUCTION
As we are aware, the challenge facing the world is lowering its carbon emissions. This was
first internationally recognized at the Kyoto Protocol of 1997 which states that countries
must lower their GHG (green house gases) emissions or penalties will be incurred.
The construction industry makes 22.3% of carbon emissions according to the European
Commission for Emissions document, with the energy sector still emitting more (38.2%)
and transport at 23.1% (European Commission, 2010). There are options to make this better
for construction even in times of financial difficulties, but this must be done as a personal
responsibility in order for change to occur globally. America is the largest carbon
contributor; emitting 25.2% of all emissions, China is the second with 15.2% then Russia
with 6.7% and other industrial countries such as Japan, India, Germany and the UK not
producing more than 5% (Nation Master 2010). However, this is justified by their size and
populations as well as their contribution to the global economy.
Steel is widely used in the construction industry and this year (2011) it is predicted that
1306 mmt (million metric tons) will be used: a record level (Recycling Portal 2010). Steel
159

has a high carbon footprint, in particular high yield steel which is preferred in construction,
due to more carbon being used to produce a high yield strength and FRPs carbon emissions
are less than that of steel. Therefore, the continued trend of FRP rebar being used in
preference to steel will lower our carbon emissions over time in construction.
The life cycle costing must be equal if not less than that of steel for FRP to seriously make
a difference to industry. A negative aspect of some FRPs production in terms of materials is
the polymer manufacture, which can be made from fossil fuels. However, in BFRP the
polymer is polysiloxane which is vastly abundant and easily synthesized from siloxanes.
Although this is greatly negative for some FRPs, we must not only look at the long-term
benefits of FRPs with respect to carbon emission.
As well as the direct benefits of FRP, there are indirect benefits with construction, such as
constructing FRPs with concrete. The lightweight rebars and a higher tensile capacity
allows for lighter weight structures. This in turn provides for less concrete and less cover is
necessary to protect the rebars from aggressive environments.
2. ASPECTS OF IMPORTANCE FOR SUSTAINABILITY
Sustainability has evolved in meaning over the years. The original definition was coined as
development that meets the needs of the present without compromising the ability of
future generations to meet their own needs, stated in the Bruntland report in 1987 (World
Commission on Environment and Development (WCED, 1987). In essence, sustainability
must protect the present and the future; therefore the following aspects have been identified
as important in this paper:
2.1 Economy and Life Cycle Costing
Construction accounts for 30.6% of GDP worldwide (CIA, 2010) and therefore has a
massive influence on the world economy. FRP continues to get cheaper and with greater
developments being continually made to FRPs to make them better in construction. That
said it is still more expensive than Steel (Burgoyne and Balafas, 2007). With that in mind,
can FRP be economically more viable that steel? We should not disregard FRPs because
the cost of manufacture is greater than that of steel, as the life cycle costing and other
hidden costs, (such as for repairs) are not included. FRP also gives the opportunity to
design for specialized buildings with longer life spans, and many research papers have
concluded that the life cycle costing for FRP is more beneficial for FRPs as seen in
Nishizakis research. Nishizaki concluded that although the initial costs are high, when
longer life spans are necessary FRP is an efficient choice of material (Nishizaki et al.,
2006).
2.2 Properties
The very properties of FRPs make them sustainable. Polysiloxane, the polymer commonly
used for FRPs, is an inorganic polymer which is already oxidized due to the Si-O bond and
consequently does not suffer from oxidation effects, unlike steel. This property in turn
160

allows for a much greater service life and lower maintenance with fewer necessary repairs.
Moreover, Polysiloxanes have high resistance to UV light and temperature (Mowrer, 2003).
CFRP, GFRP and BFRP all have the same if not greater tensile capacity in comparison to
steel. They are also lightweight (lighter than steel) and have specific stiffness; these
properties add to the repertoire for FRPs being good to construct with. The recent studies
into chemical testing with BFRP have shown great resistance to acid and alkali resistance.
This inertness is a great property to ensure longer life spans of structures. The high
resistance to temperature has also made BFRP a choice in structures for enchanting the fire
resistance of structures as commented by Palmieri et al. (2009).
2.3 Construction Phases
There are definitive phases that a construction material must go through in its life. These
are: manufacture, construction, mid-life and life expansion and end of life as commented by
Burgan and Sansom (2006); these are discussed in relation to FRP usage.
2.3.1 Manufacture
FRP rebar is undoubtedly more difficult to manufacture than steel at present. However, the
carbon emissions and the impact to the environment are less with FRP. Steel contributes
greatly to carbon emissions and temperatures of a little under 1400C are required
constantly, therefore large amounts of energy are used. In order to manufacture Glass,
Carbon and Basalt fibres, temperatures of up to 2400C are used (Lee and Jain, 2009),
which is greater than steel, but currently less FRP is being made and the process of
manufacture is continually being refined. Both FRP and steel rely on fossil fuels to be burnt
(Gerdeen et al., 2006), which are not in infinite supply and this is of concern as Lee and
Jain (2009) point out in their paper outlining the role of FRP. The abundance of Siloxane
(which is any compound containing R2SiO) and Basalt (volcanic rock) for BFRP, make
BFRP a particularly sustainable choice of FRP for the future as well as the products
necessary for CFRP and GFRP.
FRPs contain a greater amount of chemicals in order for the composition to be made as
seen in the Table 1.
Some of these chemicals are rare (Ti for example) and can be hazardous, but for BFRP and
GFRP they are naturally found in the production of the compounds and therefore not added.
Also when controlled they are no greater a threat to the environment than the by-products
of steel (carbon as for CO2 a GHG and unwanted products from the slag).
As for the abundance of materials glass fibres (as SiO2) for FRPs (in particular GFRP) are
in great supply and is easily manufactured using thermosetting plastics such as unsaturated
polyester. Carbon for CFRP too is sustainable to reuse in construction and is greatly
available from graphite. Basalt Fibre is made by the direct extraction of basalt rock
(volcanic) from selected quarries. The rock is then crushed into an extremely fine dust. This
basalt dust is then attached to the polysiloxane polymer, thus reinforcing the polymer. With
161

this argument of materials being in abundance the same can be argued for steel as iron is
the 4th most abundant element of the planet, although the manufacture of steel could be
seen to be more harmful to the environment, as the materials required for FRPs are more
readily available than smelting steel from the extraction of iron ore.
Material Fe
Mild

steel

Table 1. Elemental composition of materials

O
C
Si
Mg Al
Ca
Na K

High
Yield
Steel

GFRP

BFRP

Ti

Br

Cl

2.3.2 Construction
The single greatest disadvantage for FRP has been the lack of design codes. Burgoyne and
Balafas stated in 2007 that this was a main contributor to why FRP is not a financial
success. Uomoto (2007) also agreed that efforts needed to be made for design codes for
FRP and this would mean greater ease and use of FRP in buildings and not just as an
alternative material. Indeed great efforts have now been made and there are now design
codes in many countries such as America (although a Eurocode is now been developed) and
this is due to the continual interest in FRP as seen in all research and specifically the
document ACI 440.1R-06 (Ospina and Gross, 2006).
The lack of sufficiently developed design codes has not stopped engineers building with
FRPs, however, and indeed the properties of FRPs compliment the building industry well.
Their lighter weight makes the construction process safer, boosting the construction
industrys poor health and safety record. This also boosts the construction time of buildings
as seen by (Lee and Jain, 2009). The UK has seen this directly with domes and mosques
been made in cheaper and faster fashion (Kendall, 2007).
2.4 Midlife and life expansion
Life expansion is easy with the rebars as other external FRP can be retrofitted which
mirrors the structural behavior of the internal rebars. The properties of FRPs compliment
strengthening to buildings in need of repair and this has been a market that has greatly
boosted the use of FRP. It is always of benefit to extend the service life of a structure as far
fewer resources are required than to demolish and re-build. This compliments the ideology
of sustainability with reuse of buildings. FRPs also do not require methods like cathode
protection due to their inertness to corrosion and would only require standard monitoring
techniques with sensors and monitoring the deflection of the beam in order to ensure that
the beam is not failing.
162

2.5 End of Life and Waste

With rebars there is very little waste as the design is established well before the build so the
amount used is the amount ordered (waste minimalisation). That said, the waste of FRP
becomes an issue in demolition. Steel offers benefits that FRP in general does not; steel is
readily reusable and recyclable and there are large businesses that specialize in this. Steel is
easily processed from the waste concrete and then recycled, but the reprocessing of steel
also produces CO2 emissions. FRP can also be reused in other aspects such as crushing it
down into aggregate (dispersed reinforcement). Patents have now been passed for the
recycling of waste FRP, as seen by (Kamite et al., 2008) and this is a major step of many to
come for the recyclability of FRPs. There are other methods of using chemical substances
to break down the resin from the polymers, however this is an expensive and hazardous
solution, but with waste of plastics being a growing problem there are many initiatives in to
process FRPs.
The main issue with recycling and therefore reusing FRPs is the way in which they are
manufactured as identified by Bartholomew (2004). FRPs are highly engineered plastics
designed for long lifespan. Therefore they demand high amounts of energy (as they have
high calorific values) making incineration with energy recovery troublesome to the
operators as well (Halliwell, 2009).
FRP waste is in general non-hazardous, however only the dust particles (which are only
produced in cuttings during the manufacture stage) have been studied in laboratory
conditions and have been seen to be harmful, although much less in comparison to other
construction materials (Hesterberg et al. 1998; U.S. Department of Health and Human
Services, Public Health Services, Agency for Toxic Substances and Disease Registry,
2004).
FRP waste according to the European waste list 2000/532/EC, is classified under several
codes (CEFIC Technical Bulletin, 2006) and these must be known in order to organize
waste correctly and legally in Europe. Possibly the most commonly used for construction
purposes at the demolition stage would be CEFIC - 17 02 03. However, this would depend
on the state of the FRP; as stated before it can be classified into Hazardous waste within
the H1 to H14 list of Annex III of EC Hazardous Waste Directive (CEFIC Technical
Bulletin, 2006).
Halliwells paper on Recycling FRP materials, comments on FRPs being difficult to reuse,
as it is difficult to ascertain the loading properties of pre-used FRP beams, and this can be a
lengthy process. FRP waste is also predicted to double, from some 156000t being used in
2000 to a predicted 304000t in 2015 (Harvers, 2002).

163

3. CONCLUSIONS
FRP is a material that has the capacity to allow for greater service life and greater life cycle
costing for construction of structures. FRPs are cost effective in many aspects, have higher
strengths, greater resistance to aggressive environments and are sustainable in our current
global environment. Certainly FRPs have their disadvantages in comparison to steel; they
are not readily reusable and there is no well established design code in the EU. The
finalization of the design code worldwide and standardization of FRP materials will ensure
the future of FRPs. In terms of manufacture, FRP offers great sustainability and great
promise to the world of construction.
4. REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.

10.
11.
12.
13.
14.

Agency for Toxic Substances and Disease Registry (ATSDR). 2004. Toxicological
Profile for Synthetic Vitreous Fibers. Atlanta, GA: U.S. Department of Health and
Human Services, Public Health Service.
Bartholomew, K. 2004. Fibreglass Reinforced Plastics Recycling. Minnesota
Technical Assistance Program.
Burgan, B.A. and Sansom, M.R. 2006. Sustainable Steel Construction. Journal of
Constructional Steel Research, 62(11): 1178-1183.
Burgoyne, C. and Balafas, I. 2007. Why is FRP not a Financial Success? Proceedings
of the 8th International Symposium on Fiber-Reinforced (FRP) Polymer
Reinforcement for Concrete Structures (FRPRCS-8), Patras, Greece, 16-18 July, 10 p.
European Commission. 2010. EU Energy in Figures 2010. Directorate-General for
Energy and Transport (DG TREN).
Gerdeen, J.C.U., Lord, H.W. and Rorrer, R.A.L. 2006. Engineering Design with
Polymers and Composites CRC Press, Taylor and Francis Group.
Halliwell, S. 2009. Recycling of FRP Materials in Construction. ICE Manual of
Construction Materials, pp. 695 706.
Harvers, F. 2002. Green Recycling Label. The Step Towards Closing The
Loop, Proceedings of the 3rd Automotive Seminar, Cambrai/ Lieu Saint Amand,
France.
Hesterberg, T.W., Chase, G., Axten, C., Miller, W.C., Musselman, R.P., Kamstrup,
O., Hadley, J., Morscheidt, C., Bernstein, D.M., and Thevenaz, P. 1998.
Biopersistence of Synthetic Vitreous Fibers and Amosite Asbestos in the Rat Lung
Following Inhalation. Toxicology and Applied Pharmacology, 151(2): 262-275.
Kamite, M. and Kato, M. 2008. Process for Recycling Waste FRP. United States
Patent 7407688.
Kendall, D. 2007. Building the Future with FRP Composites. Reinforced Plastics,
51(5): 26-33.
Lee, L.S. and Jain, R. 2009. The role of FRP composites in a sustainable world. Clean
Technologies and Environmental Policy, 11(3): 247-249.
Mowrer, N.R. 2003. Polysiloxanes. in Ameron International - Performance
Coatings & Finishes.
Nation Master. 2010. Environment Statistics - CO2 Emissions (most recent) by
country.
(http://www.nationmaster.com/red/pie/env_co2_emi-environment-co2164

emissions2010), May.
15. Nishizaki, I., Takeda, N., Ishizuka, Y. and Shimomura, T. 2006. A Case Study of Life
Cycle Cost based on a Real FRP Bridge. Proceedings of the Third International
Conference on FRP Composites in Civil Engineering (CICE2006), eds. Mirmiran, A.
& Nanni, A., Miami, Florida, USA, 13-15 December, pp. 99-102.
16. Ospina, C.E. and Gross, S.P. 2005. Rationale for the ACI 440.1R-06 Indirect
Deflection Control Design Provisions. Proceedings of the 7th International
Symposium on Fiber-Reinforced (FRP) Polymer Reinforcement for Concrete
Structures (FRPRCS-7), eds: Shield, C.K., Walkup, S.L, Busel, J,P. & Gremel, D.D.
Kansas City, USA, pp. 651-670.
17. Palmieri, A., Matthys, S. and Tierens, M. 2009. Basalt Fibres: Mechanical Properties
and Applications for Concrete Structures. Proceedings of the International conference
on Concrete Solutions, Padua, Italy, CRC Press, Balkema, pp. 165-169.
18. Uomoto, T. 2007. Test Methods for FRP Materials. Proceedings of the 8th
International Symposium on Fiber-Reinforced (FRP) Polymer Reinforcement for
Concrete Structures (FRPRCS-8), Patras, Greece, 16-18 July, 9 p.
19. UP resin group of CEFIC. 2006. Classification and Handling of FRP Waste within the
Current EC Legislation.
20. World Commission on Environment and Development (Brundtland Commission).
1987. Our Common Future, 43 p.
21. Zurck. 2011. Worldsteel: Steel Use Will Increase by 10.7 Percent.
(http://www.recyclingportal.eu/artikel/24078.shtml), May.

165

166

CDCC-11

DURABILITY AND SUSTAINABILITY OF FIBRE REINFORCED POLYMER

(FRP) COMPOSITES FOR CONSTRUCTION AND REHABILITATION
DURABILIT ET CYCLE DE VIE DES MATRIAUX COMPOSITES
EN POLYMRES RENFORCS DE FIBRES (PRF) EN
CONSTRUCTION ET RHABILITATION

METHODS FOR EVALUATING LONG-TERM PERFORMANCE OF

FRP-STRUCTURAL SYSTEMS APPLIED TO CONCRETE BRIDGES
K. Crawford
Lead Project Engineer, Directorate of Civil Engineering, 95th Air Base Wing, Edwards AFB, CA,
USA

ABSTRACT
A challenge confronting transportation infrastructure owners using carbon fiber-reinforcedpolymer (CFRP)-structural systems to repair and strengthen reinforced-concrete (RC)
bridges is having necessary protocols to determine and predict the long-term performance
of the CFRP-concrete bridge systems. Adequate procedures to measure CFRP-RC bridge
performance over time and models to predict performance are not currently available to
owners. While a range of environmental and mechanical deterioration factors potentially
impact the CFRP-concrete (CFC) bonding system affecting the long-term durability and
performance of the bridge structural system, the sustained performance of CFRPstrengthened bridges is a function of three factors: (1) design of the CFRP-concrete
structural system, (2) the CFRP material application process to the concrete, and (3)
maintenance of the CFRP-bridge and structural components. Whereas the CFRP design
and application determine the baseline conditions for bridge performance upon construction
completion, the maintenance process sustains bridge performance over time. This paper
considers methods to evaluate the performance of CRP-strengthened concrete bridges by
using non-destructive testing in the CFC bridge maintenance process to examine the
condition (strength) of the CFRP-concrete bonding structure and to monitor changes in the
bonding system. Maintenance procedures that control and mitigate deterioration factors
impacting CFRP-concrete bonding will contribute to long-term performance of the CFCbridge structural system. The 19 CFRP-strengthened bridges on European Corridor 8 in the
Republic of Macedonia are used as a model for the evaluation.
1. INTRODUCTION
The purpose of this paper is to present methods to test the performance characteristics of
large areas of CFRP composite material applied to a number of concrete bridges by
looking at the condition of the bond between the CFRP plates and the concrete structural
member. Because load performance of CFRP-strengthened (CFS) RC bridges is dependent
on the condition and integrity of the CFC bond the objective of this approach is to identify
areas of CFRP material de-bonding from concrete structural members, and thus be able to
167

determine the degree of degradation in structural performance. Ideally the applied CFRP
plates have 100% bonding to the concrete structural members and the designed level of
performance is maintained. However, over time under influence of a range of
environmental and mechanical deterioration factors the bond between the CFRP material
and the concrete will degrade. Therefore to maintain bridge performance de-bonded areas
must be identified and repaired. To examine the integrity of the CFRP-concrete bonding
structure this paper describes a non-destructive testing (NDT) method for testing the bond
structure using a mobile impact-echo (hammer) device. This device generates low
frequency acoustic pulses to detect areas of CFRP plate de-bonded from the concrete
structure through changes in frequency. The NDT concept presented in this paper is
designed to test large areas of CFRP bonded plate on multiple bridges. The device is
currently under development and will undergo field testing in late 2011, in the Republic of
Macedonia, on the highway bridges strengthened in 2001 with CFRP-structural systems.
In RC bridge maintenance programs, a desired objective of bridge owners is to maximize
bridge load performance for its designed service life with a reasonable investment in
maintenance and rehabilitation funds. However, for owners of CFRP-strengthened bridges
a fundamental issue in bridge maintenance protocols concerns how to measure, maintain,
and predict the performance of the CFRP-strengthened bridges at a reasonable cost. Nondestructive testing of the CFC bond with an impact-echo device is an effective method for
testing the bridge CFC bonding structure and evaluate bridge load performance. For CFS
bridge maintenance programs non-destructive testing should meet the following criteria:
(1) NDT testing is easy to apply; (2) the NDT device and testing methods are inexpensive;
(3) the NDT procedure is capable of testing large areas of CFRP plates in reasonable time;
and (4) NDT testing provides sufficient data to quantify the structural condition of the
bond. Testing the condition of the CFC bonding structure is a necessary element in a
comprehensive CFS bridge maintenance program to maintain bridge load performance over
time.
2. BACKGROUND
In 1999, in support of the NATO peacekeeping operations in Kosovo, it became necessary
to move heavy military equipment by road from Bulgaria through the Republic of
Macedonia into the theater of operations, as reported in earlier papers [1,2]. Certain
bridges on the movement route were overloaded above designed safety limits and required
strengthening. The decision was made to use a CFRP-structural system to strengthen 19 of
48 slab and girder bridges (over 90 km) on the highways M2 and M1 (E-870) in
northeastern Macedonia to allow the efficient movement of military heavy equipment
transports (HETS). The 17 bridges on the M2 were strengthened in 2001 and the two M1
bridges strengthened in 2002. A total of 14,600 linear meters of CFRP plates in widths of 5
to 15 cm were applied to bridge girders and decks, a total area of 1,318 m2. 19 bridges,
originally designed with the 1949 M-25 loading scheme, were strengthened with CFRP
using a V600 load scheme, increasing the bending moment at main girder 10-meter midpoint from 75 ton-meter(tm) to 160 tm(bridge B7) and from 50 tm to 110 tm at the main
girder 7-meter midpoint (bridge B18). The application of the FRP structural system
produced a significant increase in the structural member bending moment and bridge
loading capacity. As a consequence the CFC bonding structure becomes a critical element
168

in bridge load performance and must be tested and maintained to sustain the V600 designed
loading. The challenge for the owner becomes how to test and evaluate 14,000 meters of
bonded CFRP plates and how to determine the extent of de-bonding, if any, and its impact on
bridge performance.
Figure 1 shows examples of the numbers of CFRP plates applied to three (of 19) bridges,
with their bonded areas shown in Table I. It has been ten years since the 17 bridges on the
M2 and two bridges on the M1 were strengthened. Testing of CFRP bond on the bridges has
not been performed.
The issue is this: How are the strengthened bridges currently
performing, i.e. is the designed V600 load scheme still sustained by the bonded CFRP plates
as originally applied to the bridges in 2001? Table I shows the CFRP plate area for 10 of
the 19 CFS bridges. The issue for bridge engineers becomes how to evaluate the entire CFC
area on 19 bridges in a timely and cost effective manner. The NDT impact-echo method
proposed in this paper provides an effective solution to evaluate the large bond area on the 19
CFS bridges.

Fig. 1. M1- Bridge B2-N

96 Plates (12cm)

M2 Bridge B36
80 Plates (15 cm)

M2 - Bridge B22
60 Plates (15 cm)

4. FACTORS CONTROLLING CFS BRIDGE PERFORMANCE

Three factors define performance of RC bridges strengthened with CFRP-structural
systems:
(1) Design of CFRP-concrete structural strengthening system
(2) Application of the CFRP material to the concrete structural members
(3) Maintenance of the CFRP-bridge structural components and system
4.1 CFRP Design
When the M2 project was designed in 2000, national codes and ACI-440(in draft) were used.
Per 440 [3] the design for CFRP structural systems required the flexural strength Mn of
the concrete member to exceed the required loading factorial moment Mu such that:
Mn Mu where Mu = [(d a/2) fy As ] i Mi

(1)

where d=depth of section, a=depth of concrete, fy=steel yield strength, and As=cross
section area of flexural steel. is the strength reduction factor and i Mi is the sum of
the factored loading moments on the section. The FRP strengthening limits of the concrete
169

flexural members were controlled by the failure mode and the ductility and serviceability
limits of the structural system.
Table 1. Area of applied CFRP plates for selected bridges
Typical CFRP-Strengthened Bridges
CFRP Plate
(10 of 19 CFS bridges)
Length (m) and Area (m2)
Bridge
No.

Location
On M2

Bridge
Type

Length
(m)

B7
B11
B18
B22
B28
B35
B36
B37
B39
B2-N

14+027
21+876
38+444
41+786
49+631
66+058
67+409
68+452
71+211
M1-Kum

Girder
Slab
Slab
Slab
Girder
Girder
Slab
Girder
Girder
Slab

120
10
36
30
50
52
21
17
85
46

No.
Spa
n
6
1
1
1
3
3
2
1
4
4

Plate
Length

Number
Plates

Plate
Area

1478
198
218
1308
1210
346
1032
415
1778
938

72
26
54
60
60
20
80
36
48
96

116.1
29.6
32.6
196.3
100.6
41.2
82.8
34.4
146.2
112.5

Total CFRP Plate Area for 19 bridges: 1,318 m2

While a strength reduction factor of 0.80 is recommended by ACI 440, because of
uncertainty in the condition of the reinforcing steel a reduction factor of 0.70 was used.
This strength reduction factor improved the reliability of the CFRP-strengthening design
for the M2 bridges.
4.2 CFRP Application
The FRP structural system application to RC bridge members is a process to apply an FRP
composite material with an epoxy compound to a concrete substrate, to form a high-quality
tri-layer FRP-epoxy-concrete structural system. The goal is to achieve 100% bonding of
the FRP material to the concrete substrate. The quality of the bond is critical for achieving
the designed load performance. These conditions are normally specified by the FRP
material manufacturer. Training and certification by the FRP manufacturer of the
contractor, and inspection personnel. The actual application process will require substrate
repair and surface preparation, correction of corrosion related deterioration, and repair and
injection of cracks, normally greater than 0.3 mm. The concrete surface must be prepared
to achieve a specified surface profile. The mixing and application of the bi-component resin
in the correct temperature range is necessary to achieve a long-lasting bond. The CFRP
application process requires controlling key parameters, such as temperature, humidity, and
moisture in the substrate. A good quality control and quality assurance system by the
contractor and the inspectors is necessary to ensure a lasting FRP-concrete bond. In
practice, with no de-bonding factors affecting the bonding structure, the CFRP-epoxyconcrete tri-layer system should last indefinitely and sustain its original designed load
performance. In reality some de-bonding will occur
170

4.3 CFRP-Bridge Maintenance

The third factor necessary for sustaining CFC bridge performance is maintenance of the
bridge, its structural components, and the applied CFRP strengthening system, to include
the composite material and the bonding epoxy. Good maintenance procedures by the owner
are necessary to sustain optimum designed performance of the FRP-strengthened bridges.
The FRP bridge maintenance process consists of three key components: inspection of the
bridge and its FRP system, testing of the FRP-epoxy-concrete bonding structure, and
maintenance of the FRP composite and epoxy bonding material and the bridge concrete and
steel structural components.
NDT evaluation with the described impact-echo device is key to obtaining bond condition
data.
The maintenance program establishes inspection criteria and procedures, sets maintenance
standards and procedures to maintain bridge performance, and sets conditions to maximize
the service life of the bridge structures. Control and mitigation of deteriorating factors and
de-bonding mechanisms is a key component in a bridge maintenance program.
5. METHODS TO EVALUATE CFRP-CONCRETE BOND CONDTION
Of the ten NDT methods described in ACI 228 the best method for detecting flaws in
bonded overlays on concrete slab structures is the impact-echo method [4] .
5.1 Principles of Impact-Echo NDT
The principle of the impact-echo technique generates a transient pulse into the CFRPepoxy-concrete tri-layered structure such that a wave is reflected by the boundaries of the
structure interfaces. The pulse produces P and S waves that propagate into the layered
structure in hemispherical waveforms. The objective of using this NDT method on the
CFRP plate is to accurately identify areas of de-bonding between the plate and concrete.
Using the acoustic signal generated by an impact device applied to a layered bonding
structure provides an effective means to detect and identify areas of voids and delaminated
materials in the bonding system.
5.2 NDT Device and Concept of Operation

Fig. 2. (a) Impact-echo device

(b) Plate impact-signal pattern

171

Figure 2 shows (a) the mobile impact-echo device developed to test CFRP plate bond and
(b) the resulting impact-signal pattern. The four-wheel mechanical impact device consists
of two 15cm cam-actuated levers with hammer pins, aligned 5 cm apart on axes A and B.
Each pin is driven by an adjustable compression spring to obtain an equal impact force on
the CFRP plate to generate optimum acoustic signal intensity. The impact of the alternating
striking pins generates acoustic waves penetrating the CFRP-epoxy-concrete tri-layer
bonding which in turn reflects an acoustic wave to a signal transducer (receiver). The pins
are designed to strike the CFRP plates at four cm intervals alternately on the A and B axis.
A 15cm wide CFRP plate will have 25 impacts per meter, with one impact producing a
bond condition response for 600 mm2 of CFRP plate area. The impact device is designed to
strike the plate with equal force over the length of the CFRP plate generating an acoustic
signal with consistent intensity for a fully bonded plate. If the CFRP plate is bonded for the
length of the concrete structural member the signal frequency will be consistent, i.e.
without variation over the entire length of the plate.
The frequency of the generated pulse is a function of the impact duration. A shorter
duration, < 80 s, generates a higher frequency and is suitable for detecting flaws in
concrete structural members less than one meter in thickness. For the M1 and M2 bridges
the desired depth of flaw detection, from the CFRP plate surface to the closest steel
reinforcing in the concrete is less than 30 cm. The depth of the CFRP plate surface to the
epoxy-concrete interface is less than 1.5 cm and is the focus of interest in the impact-echo
NDT method proposed. Each impact signal received is logged on a data recorder. A
bonded plate will generate a specific frequency, and a de-bonded plate will generate a
lower frequency. The variation in received acoustic frequency will indicate a change in the
CFC bond structure, i.e. potential de-bonding. As testing of the plate bond structure
proceeds and different frequencies are recorded, frequency analysis of the displacement
waveform (reflected pulse) is used to obtain an amplitude spectrum from the time-domain
signal. The frequency of the reflected pulse is a function of wave speed and distance
between the plate test surface and the internal reflecting surface, i.e. a function of the CFC
bonding/de-bonding structure. De-bonded CFRP plate areas are are marked for repair.
6. CFC BRIDGE DETERIORATING FACTORS
There are a number of failure modes and de-bonding mechanisms, e.g. moisture, affecting
bonded CFRP plate. In work by Kotnia[5], two primary modes of bond failure occur in the
plate anchorage areas and the maximum bending moment in the mid-beam region, a focus
for this NDT method. In one theoretical de-bonding load model by Nigro and Savoia [6],
CFRP plate width, length, thickness, and axial stiffness define the de-bonding load for a
structural member. Thus a decrease in plate bonded width and length proportionally
decreases CFS bridge load capacity. Control and mitigation of CFRP plate de-bonding
mechanisms, especially moisture, in a CFC bridge maintenance program will contribute to
extended service life for the bridges.
7. CONCLUSION
This paper is written from the perspective of a transportation infrastructure owner
confronting the issue on how to inspect, test, and maintain a large number of RC bridges
172

strengthened CFRP systems. Methods to efficiently test the bond condition of large areas
of CFRP plates applied to RC bridges are not readily available. The proposed impact-echo
NDT method will identify de-bonded CFC areas on concrete structural members will
provide bridge engineers qualitative data on CFRP-concrete bond condition and the
necessary input for bridge maintenance protocols.

Mobile impact-echo device is a means to rapidly test and evaluate large areas (long
lengths) of CFRP plates on multiple bridges.
Recording the frequency data for the bond establishes a signature for CFRP plate
bond condition at a point in time and becomes a historical record for the bridge.

CFC bond field measurements on the 19 E-870 bridges with the impact-echo device
described in this paper are not yet available. Field testing on a number of the bridges on the
E-870 is scheduled in the autumn of 2011.
8. REFERENCES
1.

2.

3.
4.
5.

6.

Crawford, K. 2007. A Model for the Inspection and Maintenance of FRP-Bridge

Strengthening Systems. Proceedings of the Third International Conference on
Durability and Field Applications of Fibre Reinforced Polymer (FRP) Composites for
Construction (CDCC2007), eds: Benmokrane, B. & El-Salakawy, E., Quebec City,
Quebec, Canada, 22-24 May, pp. 269-276.
Crawford, K., Nikolovski, T. 2006. The Design and Application of CFRP Composite
Material to Strengthen 19 Concrete Highway Bridges in the FY Republic of
Macedonia. Proceedings of the 2nd International fib Congress, Naples, Italy, 5-8 June,
12 p.
ACI Committee 440. 2002. Guide for the Design and Construction of Externally
Bonded FRP Systems for Strengthening Concrete Structures. ACI 440.2R-02,
American Concrete Institute, Michigan, USA, 45 p.
ACI Committee 440. 1998 (Reapproved 2004). Nondestructive Test Methods for
Evaluation of Concrete Structures. ACI 228.2R, American Concrete Institute,
Michigan, USA, 62 p.
Kotynia, R. 2005. De-bonding Failures of RC Beams Strengthened with Externally
Bonded Stripes. Proceedings of International Symposium on Bond behavior of FRP in
Structures (BBFS 2005), eds: Chen and Teng, Hong Kong, China, 7-9 December, pp.
247-252.
Ferracuti B., Martinelli E., Nigro E., Savoia M. 2007. Fracture Energy and Design
Rules against FRP-Concrete Debonding. Proceedings of the 8th International
Symposium on Fiber Reinforced Polymer Reinforcement for Reinforced Concrete
Structures (FRPRCS-8), Patras, Greece, 16-18 July, 10 p.

173

174

CDCC-11

DURABILITY AND SUSTAINABILITY OF FIBRE REINFORCED POLYMER

(FRP) COMPOSITES FOR CONSTRUCTION AND REHABILITATION
DURABILIT ET CYCLE DE VIE DES MATRIAUX COMPOSITES
EN POLYMRES RENFORCS DE FIBRES (PRF) EN
CONSTRUCTION ET RHABILITATION

CREEP BEHAVIOUR OF GFRP CABLES, SHORT AND LONG TERM

TEST
E. Ferrier, A. Si-Larbi and P. Hamelin
Laboratoire de gnie civil et d'ingnierie environnementale, universit claude bernard, lyon i, 82 bd
niels bohr, domaine scientifique de la doua, 69622 villeurbanne cedex, France.

ABSTRACT
This paper deal with the relaxation of GFRP cable used for linear RC unbonded posttensioning method. In 1990, GFRP cables have been used on one span of a prestress
concrete bridge of French Alps Highway. This bridge has been built with a usual steel
cables post-tensionning method. Some GFRP cables have been add for an experimental
purpose. The level of prestress applied on cables was 60 % of the ultimate cable strength.
Some cables have been kept in the same environmental conditions without loadings. After
this period in 2006, on site test are done in order to evaluate the loss in properties of the
GFRP cable, but also the stress cables loss. The result of this experiment allows obtaining
the durability of the cable and the effect of permanent loading on GFRP. Relaxation
functions are obtained in laboratory using time-temperature principle with short time period
relaxation test. Time temperature principles have been applied in order to estimate the loss
in stress of GFRP cables and compare then with the result obtained directly on site after
more than 15 years of relaxation. The paper present the methodology to evaluate the
relaxation of the cable with laboratories testing and compare successfully the relaxation
function to the long term relaxation test obtained on site.
1.

INTRODUCTION

1.1 Bridge Description

The bridge is a 373 m continuous span box girder located in Bardonnex (74, France). The
bridge was made with prestress concrete using bonded conduits post-tensionned method.
Structural reliability is insuring with this steel prestress method but some GFRP cables
have been had for experimental purpose. The use of a composite reinforcing bar for
concrete as an alternative to traditional steel reinforcing bar has many potential advantages.
The material is relatively light and corrosion resistant. However it is important to obtain
representative data on the long-term durability of concrete reinforced with composite rebar
in order to be able to asses the performance of the structures under non-standard conditions
which have to be considered in the design and regulation process.

175

In 1990, an experimental investigation of the reliability of prestress GFPR cables has been
done on this bridge. Few partners (GTM construction, ATMB, LCPC) have decided to
investigate the performance of such prestressing technique. In this purpose GFRP rods have
been prestress using post tensioning method. 5 GFRP bars have been prestress in two
unbonded conduit on each side of the box girder (Fig. 1). At each end of tendon specific
anchorage system have been developed made of steel wire strands placed around GFRP
tendons.

Fig. 1. Cross section on Bridge box girder

(a) Picture of the bridge

(b) Bridge box girder
Fig. 2. Bridge of Bardonnex (74 France)
1.2 Experimental Investigation
In order to investigate the long term behaviour of the GFRP cables and the residual stress in
the tendon after 16 years of loading, several tests are done. The cable stress loss is
essentially due to:
-

An initial anchorage or frictional loss during the prestressed process located to

the tendon anchorage.
Aging of GFRP and Young modulus diminution
Time dependent losses due to relaxation of the cable depending on the
rheological behaviour of GFRP tendon.
176

The total tendon stress loss can be sum-up by relation 1.

f GFRP (t , t 0 ) f GFRP , E (t , D ) f GFRP , R (t , t 0 ) f GFRP , A (t )

(1)

with f GFRP , E (t , D ) : loss due to aging of GFRP (E)

and f GFRP , R (t , t 0 ) : loss due to relaxation of GFRP (R)
and f GFRP , A (t ) : anchorage loss (A)
The loss in stress is measured directly on the bridge tendon thanks to a specific device. The
residual stresses are obtained and compare with the initial one. The next stage is to evaluate
the evolution of mechanical properties of GFRP (tensile strength and Young modulus).
Parts of the free tendon (Fig. 3) are taken on site and tested in lab. This value are compare
to the one obtain 16 years ago on the tendon. In order to study the relaxation behaviour of
the cable, some samples of GFRP tendon are also test using time-temperature principle in
order to obtain the material relaxation function. It is then possible to explain the loss
measured on site thanks to the different laboratory testing.

Fig. 3. Picture of GFRP tendon end anchor

2.

STRESS LOSS OF GFRP CABLES

2.1 Measurement of Stress Loss On-Site

The aim of this part is to evaluate the relaxation of GFRP tendons after 16 years of loading
on the bridge.
In order to obtain this value, a specific device was developed (Fig. 4); the device was made
of a bearing plate placed over the dead end anchorage. A threading steel tube allows to
apply a force on a anchoring cone. The wedge transfers this force to the tendon. It is then
possible thanks to a load cell to measure the load which is necessary to apply to obtain an
additional strain to the tendons. On each GFRP bars, in the mid part, strain gauges were
bonded.
During the loading process, strain gauges and the load cell were connected to a data
acquisition system. Figure 6 illustrate the value of load and strain in function of time. It is

177

clear that when the wedge became free (Fig. 5), the composite strain increases suddenly,
this allows evaluating the value of the residual prestress force (Fig. 6).
Wedge
32

Load cell
20
32

Threading steel tube

40

Bearing plate

32

53

32

20

160
20

20

32

20

32
160

Fig. 4. On site tensile force measurement device

Fig. 5. Photos of wedge before and under maximal loading value

-1000

20

Load (daN)

-1500
-2000

15

-2500
-3000

10

-3500
-4000

GFRP Strain (mm/m)

-500

25

90
12
0
15
0
18
0
21
0
24
0
27
0
30
0
33
0
36
0
39
0
42
0
45
0
48
0
90
0
93
0
96
0
10
00

0
30
60

-4500
-5000

0
Load cell
GFRP strain

Time (s)

Fig. 6. Load and composite strain in function of time

For each cable the value obtains is given by table 1.
Table 4. On site GFRP cable residual prestress force
Cable 1
Cable 2
Cable 3
Cable 4
Force (kN)

40

44

41
178

43

Average
42

An average value of 42 kN is then measure. These values correspond to a 31 % force loss

compare to the initial 61 kN.
2.2 Evolution of Tensile Mechanical Properties
Tensile test are done with a displacement control rate of 1 mm/min. Initial 17 years ago
tensile strength of GFRP was 1330 MPa 150 MPa for a Young modulus of
45900 MPa 2500 MPa where as these value has respectively decrease to
800 MPa 180 MPa and 36500 MPa 2000 MPa. It is then obvious to conclude on a
decrase of 40 % of tensile strength and a 20 % Young modulus decrease (Fig. 7). The loss
in force in GFRP tendon due to the loss in Young modulus properties is given by relation 2.

t 0 , t k f , D E f ,0 (t 0 )

(2)

f GFRP, E t 0 , t A f k f , D E f ,0 (t 0 )
With consideration of a linear damaging process of GFRP
k f ,D 1 a t

(3)

With a: aging factor (0.0117); t : aging time in years; Ef,0 : initial Young modulus; (t0) :
initial prestress strain; Af : composite area
1600
1400

Strength (MPa)

1200
1000
800
600
400

Tensile test after 17 years

Initial tensile test

200
0
0

5000

10000

15000

20000

25000

30000

35000

Strain (m/m)

Fig. 7. Average curve of Tensile test

This result are comparable to the one obtained by Zermeno and all [1] and Abbasi and all
[2].
3.

RELAXATION FUNCTION: SHORT TERM TEST

3.1 Time-Temperature Principle

The construction of the master curve can be done step by step by using the time
temperature superposition WLF principle (William Landel and Ferry [3]). The time
superposition of the short-term creep compliance test data provided the means for the
sequenced, short term data to be collapsed into a single master curve (MC). This principle
is used for long-term creep and fatigue polymer study. It allows a time test reduction.
Brinson and Gates [4] have extended the use of the time-temperature superposition
179

principal to composite. It allows predicting the long-term creep compliance or relaxation

using as input the material properties developed from short-term tests.
It means that a test made with several temperatures on a short period allows assessing the
long-term behavior for a reference temperature by the master curve construction method
using a shift factor at(T). This shift factor is simply defined as horizontal distance required
to shift a relaxation curve to coincide with a reference relaxation curve. The test procedure
is as follows: the test specimens (three for each test) are placed in a thermo-regulated room
at different temperatures ranging from 20 to 80C. The loss is composite force is measured
during a five-hour period, up to 80C. From each single test, we obtain a short-term
relaxation function which is usually referred to as a momentary curve. By repeating this
process at different temperatures, a family of relaxation functions can be obtained. All
those individual relaxation curves can be superposed onto a single master curve. For each
test, we have made 5 short-term relaxation tests (5 hours) with an initial constant strain of
6000 m/m and an initial tensile force of 12.4 kN. In the application of this approach to
obtain data from a short-term relaxation test, it is important that the data should not bias
against any region of the desired curve. Since the application is to be a curve in log time,
the data should approximate to an equal spacing in log time. This has been accomplished
for this work through adjustment of the data sampling rate during the experimental study.
We repeat that the goal of shifting momentary curves to a single reference (master) curve is
to predict response of the GFRP stress.
3.2 Results
Single curve obtain between 20C to 80 C are plot on figure 8 a. Each curve may be shift
in order to obtain a master curve at 20C.
Load cell

Anchoring cone
Relaxation frame

Loading frame

GFRP tendon

1.00 m

Fig. 8. Experimental relaxation frame

14000

14000

12000

12000
10000

Force (N)

20C

8000

40C
60C
70C
80C

6000

Force (N)

10000

8000
6000

4000

4000

2000

2000

20C
40C
60C
70C
80C
Master curve 20C

0
0

5000

10000

15000

20000
Time (s)

100

10000

1000000

(a) Single force-time curve

(b) Master curve construction
Fig. 9. Relaxation result

180

100000000
Time (s)

This master curve is build thanks to the six relaxation curve measured experimentally. The
reliability of this method may be enhance by plotting the logarithm of the shift factor (at) in
function of T-To, with To reference temperature of the master curve (20C) and T
corresponding shift factor temperature. The linear variation of log(at) in function of T-To
generally correspond to a reliability of time-temperature principle.
3.3 Relaxation Function
For considering viscoelasticity, the three-parameter Zener model (or standard linear model)
shown in Fig. 9 can be applied to describe the viscoelastic relaxation of GFRP tendons.
Based on this model, the differential equation of the constitutive relationship is given as
equation 4 which is the relaxation function of three-parameter Zener model shown in Fig.
10. We may investigate the asymptotic behavior of the relaxation by applying Laplace
transform with the limit theorems to Eq. (4):

E2

( E1 E2 )

( E1 E2 )

(4)

The elastic modulus E1 used in the three parameters model is corresponding to the elastic
modulus E obtained from the Eq. (4), which was derived based on elastic theory. Based on
Eqs. (6), the elastic constant E2 can then be determined from the experimental data and the
results are shown in Figs. 7 and 8. It is notable that these elastic constants decrease with the
increasing temperature. The viscosity parameter can be simply calculated from Eq. (13),
combined with the experimental data, and the results are plotted in Fig. 9. Using the master
curve relaxation function obtained, a Zener rheological model is retained to fit with
relaxation GFRP behaviour.

Fig. 10. Zener model

f GFRP , R A f 0 E (t , t0 )

E E
With E (t , t0 ) 1 2
E E2
1

(5)

( E1 E 2 )

1 e
E1 e ( E1 E 2 )

181

(6)

With

1
1
1

K E1 E2

(7)
t

Then E (t , t0 ) K ( E1 K ) e e
With e

(8)

(9)
E1 E2
Viscosity factor (4.109), e : relaxation time, E2 (88 000 MPa) and E1 (36000 MPa) :
Elastic constant
14000
12000

Force (N)

10000
8000
6000
4000
Master curve 20C
Zener Model

2000
0
1

100

10000

1000000

100000000
Time (s)

Fig. 11. Comparison between experimental and modelling

The model allows obtaining, with a good accuracy, the relaxation function and after 16
years of loading the expected residual load obtain with the modelling is 44.5 kN. Finally
the loss due to relaxation is given by:
t

fGFRP, R (t, t0 ) Af E0 K (E1 K) e e 0

4.

(10)

CONCLUSION

This experimental investigation allows to compare real loss of stress obtain on site and
different laboratory testing. The total loss of the GFRP bars can be given by relation 11:
t

f GFRP (t ) Af k f , D E f , 0 (t0 ) Af E0 K ( E1 K ) e e

0 ) f GFRP , A (t )

(11)

After 17 years of loading the residual prestress force measure on site is equal to 42 kN for a
value of 41 kN using the model develop. In this case the loss due to anchorage loss may be
considered to be equal to zero. We have demonstrated how a simple experimental technique
and associated analyses can be applied to characterize the viscoelastic properties of GFRP
bars under aging and relaxation. The well known viscoelastic theory, Zener model, can well
182

describe the time-dependent relaxation behaviors and therefore facilitate the determination
of the parameters, such as elastic moduli and viscosity, from fitting the experimental data.
5.

REFERENCES

1.

Zermeno De Leon, M.E. 1988. Contribution ltude du comportement mcanique de

cbles composites. these INSA LYON 1, France.
Abbassi A., Hogg P. 2005. Temperature and environmental effects on glass fibre rebar:
modulus, strength and interfacial bond strength with concrete. Composites Part B, 36:
394-404.
Williams, M.L., Landel, R.F. and Ferry, J.D. 1955. The Temperature Dependence of
Relaxation Mechanisms in Amorphous Polymers and Other Glass-forming Liquids.
Journal of the American Chemical Society, 77: 3701-3715.
Brinson, L.C. and Gates, T.S. 2003. Viscoelasticity and Aging of Polymer Matrix,
Composites. Comprehensive Composite Materials, Chapter 2.10, pp. 333-368.

2.
3.
4.

183

184

CDCC-11

DURABILITY AND SUSTAINABILITY OF FIBRE REINFORCED POLYMER

(FRP) COMPOSITES FOR CONSTRUCTION AND REHABILITATION
DURABILIT ET CYCLE DE VIE DES MATRIAUX COMPOSITES
EN POLYMRES RENFORCS DE FIBRES (PRF) EN
CONSTRUCTION ET RHABILITATION

DURABILITY OF GFRP BARS UNDER LOW TEMPERATURE

ENVIRONMENT AND HIGH TEMPERATURE ALKALI EXPOSURE
S.A. Sheikh, D.T.C. Johnson and A.D. Caspary
Department of Civil Engineering, University of Toronto, Toronto, Canada

ABSTRACT
An investigation was conducted to evaluate the performance of glass fibre reinforced
polymer (GFRP) reinforcing bars under extreme environmental conditions. These
conditions included 1) low temperature exposure around -40oC and 2) accelerated high
temperature alkali exposure (AAE). The objective of this study is to contribute to the
understanding of the life cycle performance and cost of a typical concrete flexural member
reinforced with internal GFRP bars.
Forty-five tests spread over three bar sizes were conducted to investigate the effects of cold
temperatures on GFRP bars. Results from the extreme cold tests showed minimal decrease
in the material properties. The maximum reduction in tensile strength was of the order of
2%. The modulus of elasticity differed by up to 10% when compared against the reference
samples. The accelerated alkali exposure consisted of 2000 hours of exposure to an alkaline
solution maintained at 60oC, while the bars were subjected to a uniform axial tensile strain
of 3x10-6 (3000 ). The study was conducted using specially constructed frames, as well as
a specifically designed system used to circulate alkali solution from a concentrated
reservoir to individual chambers surrounding the reinforcing bars.
Results from the tests on alkali exposure showed the tested bars consistently retained
tensile strength in excess of 91% of their original strength with only minimal changes to
stiffness. This study shows excellent performance of the tested GFRP bars exposed to
extreme environments indicating the likelihood of high durability during a lifespan of
normal service loads.
1. INTRODUCTION
Specifications for use of FRP bars in concrete structures in Canada (CSA S807-10, ISIS 06)
require, among other details, testing and evaluation of the cold temperature resistance as it
is theorized that extreme cold temperatures can adversely affect the mechanical properties
of FRP resulting in reduced performance of FRP-reinforced structures. Since FRP bars used
as internal reinforcement are subjected to highly alkaline environment, the Standards also
require retention of a certain minimum strength of bars after exposure to a prescribed alkali
185

solution for approximately three months. This requirement includes testing of bars without
load as well as under sustained loads. This paper reports the results of the tests conducted
on GFRP bars for these two conditions and the likely effect of reduced properties on
structural performance of a typical flexural member.
2. COLD TEMPERATURE TESTING
In this test series, bars were preconditioned down to -35oC and then tested in direct tension
in a -40oC environment. Changes in the ultimate strength, modulus of elasticity and
ultimate elongation of the FRP bars were investigated (Johnson, 2009).
2.1. Test Program for Cold Temperature
A total of 45 GFRP bar specimens were tested, the results of which were compared against
a range of control specimens. Three bar sizes were tested (8, 12 and 16 mm diameter bars)
with each size having 15 cold exposed samples. The bars were mounted with metal
couplers used for gripping the samples during testing. Couplers were mounted using a
readily available two part epoxy resin.
Because of the low thermal conductivity of GFRP, preconditioning was done in a low
temperature thermal chamber (Figure 1). The bars remained in the chamber for 96 hours
after which it could be concluded that the core temperature was sufficiently low. The
preconditioning chamber was maintained at -35oC throughout the entire process.

Fig. 1. Preconditioning chamber for cold temperature testing

After preconditioning, the bars were individually transported to a specialized test frame
which used an attached environmental chamber to maintain a -40oC environment
throughout the entire test. Samples were mounted with the couplers outside of the chamber
to keep the epoxy connecting the specimens to the couplers at room temperature during
tensile testing. The frame and chamber along with a mounted specimen are shown in Figure
2. Independent thermocouples were used to verify the internal test conditions of the
chamber throughout the test. The temperature was adjusted and corrected periodically
throughout testing to maintain the -40oC environment from start to finish.

186

Fig. 2. Test chamber and mounted 16mm specimen

Low temperature in the range of -40oC limited the use of any LVDTs, as such the
elongation and modulus of elasticity were determined using surface mounted strain gauges.
Special strain gauges for cryogenic environments and low temperature epoxies were used in
this program.
2.2. Results of Cold Temperature Testing
In general, the results of the cold exposed specimens were similar to those of the control set
regardless of size. The failure stress was the most consistent of the three mechanical
properties compared. The other two (elongation and modulus of elasticity) changed slightly
under the exposure however these changes can be primarily attributed to issues related to
measuring the elongation/strain at low temperatures. The detailed results for each size are
summarized in the sections to follow.
2.2.1 8 mm Test Results
Of the 15 tests conducted at low temperatures, a total of eleven specimens failed by rupture
of the bar, the other four failed by de-bonding of the couplers from the sample. The
mechanical characteristics of exposed specimens are compared against those from a set of
control specimens which were not exposed or tested in a cold environment. The results of
the testing are summarized in Table 1.
Table 1. Summary of results for 8mm tests (Standard Deviation in Parenthesis)

Average Reference
Samples
Average of Cold Exposed
Samples that Ruptured
Percent of Reference
Value
Average of Cold Exposed
Samples that De-bonded

Ultimate Strength
(MPa)

Modulus of
Elasticity (MPa)

Ultimate Elongation
(%)

1374 (30)

59990 (3000)

2.30 (0.17)

1370 (35)

63540 (5800)

2.10 (0.20)

99%

106%

91 %

1141 (32)

60830 (4100)

1.91 (0.18)

187

The results from the 8mm tests show a minimal change in mechanical properties, a small
increase in modulus of elasticity and resulting drop in elongation were expected. For
samples that failed by de-bonding of bars at couplers, the loads were consistently lower
than the value for those specimens that ruptured, however the modulus of elasticity was
similar regardless of failure mode. A typical de-bonding failure and a typical rupture failure
are shown in Figure 3.

Fig. 3. Failure modes for cold temperature samples

2.2.2 12-mm Test Results
For the 15 tensile tests conducted on the 12mm sized bars, a significantly higher frequency
of de-bonding failures was noticed. This was in large part due to the higher tensile force
required to fail the bar. Of the 15 samples tested, 3 failed by rupture and 12 failed by debonding, however for this size the de-bonding loads were similar and in some cases higher
than the rupture loads which indicate that the strength of the coupler for this size is similar
to the strength of the bar itself. The results of the testing are summarized in Table 2.
Table 2. Summarized 12mm Test Results (Standard Deviation in Parenthesis)
Ultimate Strength
(MPa)

Modulus of Elasticity
(MPa)

Ultimate Elongation
(%)

1128 (28)

60190 (3600)

1.88 (0.16)

1164 (26)

55165 (2800)

2.09 (0.06)

Percent of Reference Value

103 %

92%

111%

Average of Cold Exposed

Samples that De-bonded

1109 (42)

56160 (2900)

1.98 (0.10)

Average of Reference
Samples
Average of Cold Exposed
Samples that Ruptured

As was the case for the 8mm bars, again only a small difference was noted between the
control specimens and the exposed ones. The differences can be primarily attributed to
instrumentation rather than the effect of environmental conditions.
2.2.3 16 mm Test Results
The tests on 16mm bars were conducted initially with LVDTs measuring elongation and
then changed to use strain gauge measurements. Of the 15 cold exposed specimens; 8 failed
188

by rupture of the bar and 7 by de-bonding. The 8 specimens that ruptured are further
subdivided into two groups based on their instrumentation method. Specimens which used
LVDT measurements to determine the elongation are labelled as group A of which there
are 6 specimens. The remaining pair of rupture specimens used strain gauges and is labelled
as group B. As was the case with the 12mm samples, the de-bonding failure loads were
similar and in some cases higher than the rupture loads of some of the samples.
Table 3. Summarized 16mm Test Results (Standard Deviation in Parenthesis)
Reference Average
Average of Ruptured Specimens
(Group A)
Average of Ruptured Specimens
(Group B)
Percent of Reference Value
(Group A)
Percent of Reference Value
(Group B)
Average of Cold-Exposed Samples
that De-Bonded

Ultimate Strength
(MPa)
1208 (48)

Modulus of
Elasticity (MPa)
62360 (1800)

Ultimate
Elongation (%)
2.10 (0.05)

1192 (52)

53070 (2500)

2.25 (0.08)

1186

56990

2.09

99%

85%

107%

98%

91%

100%

1221 (60)

55260 (4400)

2.22 (0.22)

For the 16mm group again the ultimate strength remains unchanged and differences in the
modulus of elasticity and elongation were noted and can primarily be attributed to issues
related to instrumentation at cold temperatures.
In general for all sizes strength was not affected by the cold temperature. A variation in
rupture strain and the modulus elasticity measured between 6% and 15% was attributed to
the difficulties in instrumentation under extreme cold temperatures.
3. HIGH TEMPERATURE ALKALI EXPOSURE
CSA-S807 requires that FRP bars in high durability category should retain a minimum
strength of 80% after about 3 months (2000 hours) of exposure to alkali solution without
load and 70% after 3 months of exposure to alkali solution with sustained load causing a
strain of 3000 micro-strain. If the bars can maintain 80% strength retention under load, the
tests without load are not required.
3.1. Test Program for Alkali Exposure under Sustained Load
In this test program, 8 mm diameter bar specimens were subjected to sustained load and
alkali solution at 60oC in especially designed and built test frames as shown in Figure 4
(Caspary 2011). Each frame can accommodate up to 8 specimens under independent
sustained load measured by the built-in load cell. With an anchoring steel coupler and a
hinge at each end of the specimen and the load cell, the total specimen length in the 990
mm high frame was 560 mm. Three frames containing 24 specimens were stacked together
and enclosed in an insulated space as shown in the figure.

189

A small water-tight enclosure around the test zone of each bar specimens housed the alkali
solution at 60oC. The alkaline solution consisted of 118.5g Ca(OH)2, 0.92.4g NaOH and
4.2g KOH per one litre of deionized water providing a pH value of 12 to 13. The solution
was replaced every 24 hours using a pump and a reservoir to maintain its temperature at
60oC. The strain in each bar was measured with the help of strain gauges installed on all the
specimens. The load was adjusted at regular intervals to maintain the tensile strain at 3000
. The specimens remained under sustained load while exposed to 60oC alkali solution for
over 2000 hours when they were removed from the frames and tested for residual
mechanical properties. Nine control specimens were also stored in the laboratory at room
temperature without any exposure to alkali or load to serve as the standard against which
properties of the exposed specimens were to be compared.

Fig. 4. Test frames with specimens subjected to sustained load and alkali exposure
3.2. Test Results of Alkali Exposed specimens under Sustained Loads
All the specimens in this series were tested under tensile load in a Satec High-Capacity
universal testing machine. A total of 21 exposed specimens and nine control specimens
were tested. Additionally, mass change tests were performed on five exposed samples. A
summary of the results is given in Table 4. Exposure to alkali solution resulted in a crystal

190

built-up on the bars due to ion transfer from the alkali solution. Figure 5 shows a
comparison of the physical appearance of the exposed and the original bars.
It is clear from Table 4 that the GFRP bars displayed only a small change in the mechanical
properties tested during this program. The reduction in strength under severe exposure
conditions is about 8% while the allowable limit under sustained load and alkali exposure is
30%. There is practically no change in bar stiffness. A slight increase in the measured value
of the modulus of elasticity of the exposed specimens is attributed to the experimental
scatter and difficulty in measuring deformations. It should be noted that while load
measurements in such tests are reasonably certain, there is a large variation in strain
measurements which in most cases is due to the environmental condition to which the
gauges are exposed. The change in mass as a result of exposure to alkali solution for 2000
hours at high temperature was also found to be minimal.
Table 4. Summary of test results from alkali exposed specimens under sustained loads.
(Standard Deviation in Parenthesis)
Control Sample

Ultimate Strength
(MPa)
1439 (39)

Modulus of Elasticity
(MPa)
57307 (2800)

Alkali-Exposure Average

1324 (89)

57853 (3900)

0.85

Percent of Reference
Value

92.0%

100.9%

100.8%

Mass Change (%)

N/A

Fig. 5. Comparing appearance of exposed samples to control samples

4. EFFECT OF REDUCED GFRP PROPERTIES ON THE BEHAVIOUR OF
FLEXURAL MEMBERS
Various codes and design guidelines limit the adverse effects of environmental exposure by
providing the maximum allowable strength reduction. In the case of alkali exposure under
sustained load that limit is 30%. Figure 6 shows responses of a slab (S1) with tensile

191

strength of GFRP bars at 1200 MPa and 850 MPa which reflects a strength reduction of
29%. The slab is 220 mm thick and originally designed to fail due to

Fig. 6. Moment Curvature relations for sample slabs S1 and S2

concrete crushing with 16 mm bars at 125 mm spacing. Concrete strength is assumed to be
40 MPa. A 29% reduction in the tensile strength of bars still causes the slab to fail at the
same moment due to concrete crushing and results in the same behavior. Behaviour of
another slab (S2) with 16 mm bars at 250 mm spacing is also shown in Figure 6. Due to the
reduction in FRP strength of 29%, the failure mode of the slab changed from concrete
crushing to FRP rupture although the moment capacity is reduced only by about 10%. It
should, however, be noted that the responses shown in Figure 6 do not include any capacity
reduction factors. Inclusion of these factors in design would most likely result in no change
in mode of failure of the structural components because the capacity reduction factor for
GFRP is 0.5 and for concrete 0.75 (CSA S606).
5. CONCLUDING REMARKS
GFRP bars from one manufacturer were subjected to severe environmental treatments as
recommended by various guidelines in North American codes and standards. Results from
the tests on bars subjected to cold temperatures of -40oC and sustained loads combined with
alkali solutions at 60o C are reported here. The sustained loads produced tensile strain of
0.003 in the bars.
Results of the tests show excellent resistance of GFRP bars under cold conditions and alkali
exposure. There was no significant change in the tensile properties of GFRP bars as a result
of exposure to cold temperatures. The alkali exposure under sustained load produced a drop
in tensile strength of about 9% but no changes in bar stiffness were observed. A simple
analysis of slabs reinforced with GFRP bars shows that the changes observed in bar

192

properties as a result of tested exposures will result in no significant changes in the

behavior of the slab.
6. ACKNOWLEDGEMENT
The research reported here is funded by a CRD grant from NSERC and industry partners,
Facca Inc., Schoeck Canada Inc, Pultrall Inc. and Vector Construction Ltd. The authors
would like to thank the technical staff in the Structures Laboratories at the University of
Toronto and at Kinectrics Inc. for their assistance.
7. REFERENCES
1.
2.
3.
4.
5.

Canadian Standards Association (CSA). 2010. Specifications for Fibre-Reinforced

Polymers. CAN/CSA S807-10, Mississauga, Ontario, Canada, 35 p.
ISIS Canada, 2006. Specifications for Product Certification of FRPs as Internal
Reinforcement in Concrete Structures. Winnipeg, Manitoba, Canada, 27 p.
Johnson, D.T.C. 2009. Investigation of Glass Fibre Reinforced Polymer reinforcing
bars as reinforcement for concrete structures. MSc Thesis, University of Toronto. ON,
Canada.
Caspary, A. 2011. Investigation of short-term and long-term behavior of GFRP bars.
MSc Thesis, University of Toronto, ON, Canada. (In Progress)
Canadian Standards Association (CSA). 2006. Canadian Highway Bridge Design
Code. CAN/CSA S6-06, Mississauga, Ontario, Canada, 733 p.

193

194

CDCC-11

DURABILITY AND SUSTAINABILITY OF FIBRE REINFORCED POLYMER

(FRP) COMPOSITES FOR CONSTRUCTION AND REHABILITATION
DURABILIT ET CYCLE DE VIE DES MATRIAUX COMPOSITES
EN POLYMRES RENFORCS DE FIBRES (PRF) EN
CONSTRUCTION ET RHABILITATION

DURABILITY OF FRP-CONCRETE INTERFACE UNDER FATIGUE

LOADING
C. Carloni1, K.V. Subramaniam2, M. Savoia3 and C. Mazzotti3
1

Dept. of Architecture, University of Hartford, West Hartford, USA

Dept. Civil & Env. Enrg., City College of New York (CUNY), New York, USA
3
Dept. DICAM, University of Bologna, Bologna, Italy
2

ABSTRACT
The topic of this research work is the fatigue behavior of externally strengthened concrete
beams with fiber-reinforced polymers. An increase of the fatigue life after strengthening is
generally reported in literature and is attributed to the reduced stress level in the steel
reinforcement due to the presence of FRP. This paper presents an experimental analysis
devoted to investigate the fatigue-induced properties of the FRP-concrete interface. Strain
patterns along the bonded length and the surrounding concrete were measured using digital
image correlation. A decrease of the effective length during cycling is observed, indicating
a different debonding mechanism during fatigue. Post-fatigue response of the FRP-concrete
interface appears to be qualitatively similar to the monotonic one.
1. INTRODUCTION
Modern transportation infrastructures are in serious need of rehabilitation as a result of
material decay, corrosion, impact damage or increase in service load. During the last two
decades, externally bonded uni-directional fiber-reinforced polymers (FRP) composite
sheets have attracted significant attention from the civil engineering community. Most of
the strengthening and retrofit schemes take advantage of the high tensile capacity of the
FRP composite. The strengthening is a result of load-sharing between the concrete structure
and the FRP composite, which is achieved through stress-transfer between the concrete
substrate and the FRP composite. Bond quality directly contributes to the effectiveness of
the stress transfer between FRP and concrete. Debonding of the FRP composite
reinforcement is the most critical concern in this type of applications. Bond characteristics
between FRP sheets and concrete have been extensively studied [1-4]. However, the
majority of these studies deal with monotonic loading or few-cycle loading at very high
force level [5]. Kim and Hefferman [6] noticed that further research is needed to address
limitations in the current knowledge of fatigue response and durability of FRP retrofitting
systems. Most of the recent studies focus their attention on the global response of FRPstrengthened beams subjected to fatigue loading [7-9]. An increase of the fatigue life after
195

strengthening is reported in literature [6] and is attributed to the reduced stress level in the
steel reinforcement due to the presence of FRP. When early debonding of the FRP sheets
occurs, the stress carried by the FRP is redistributed back to the internal reinforcing steel.
Improvement in fatigue performance can be achieved only if adequate bond between FRP
and concrete substrate is assured. In this paper, fatigue direct-shear tests of FRP strips
attached on the surface of concrete blocks are used to investigate the extent of the
interfacial debonding.
2. MATERIAL AND EXPERIMENTAL SET-UP
Direct shear tests were performed under fatigue and monotonic quasi-static loading
conditions adopting the classical pull-push configuration. The dimensions of the concrete
block and the FRP sheet are reported in Fig. 1. The 28-day compressive strength of
concrete was equal to 35 MPa. The FRP composite comprised of continuous uni-directional
carbon fiber sheet in a two-part thermosetting epoxy matrix. The tensile strength and the
Youngs modulus were equal to 3.83 and 230 GPa. The FRP composite sheet was bonded
along the centerline on one side of the concrete block using the wet-layup procedure.
y
FRP

b1

FRP

width of the FRP:

b1=25 mm

Restraint

35

Adhesive

Steel
I-Beam

Steel
I-Beam

Restraint

L=330

152

Concrete
Block

143

L=330

LVDT

nominal thickness of the

FRP: t=0.167 mm

b =125

h=125

Fig. 1. Set-up of the monotonic and fatigue tests.

The surface deformation in the FRP and surrounding concrete during the tests has been
determined from the displacement field measurements using a full-field optical technique
known as digital image correlation (DIC) [10]. A description of DIC is reported in [3,4,10].
3. MONOTONIC QUASI-STATIC TESTS
Three specimens (Test-S1, Test-S2 and Test-S3) were tested under monotonic loading
condition following the test procedure previously described [3,4]. During each test, the
global slip, which is defined as the relative displacement between points on FRP and
concrete initially located at the beginning of the bonded area, was increased at a constant
rate equal to 0.0004 mm/sec, up to failure. Global slip was measured using two LVDTs.
The typical applied load-global slip response (Test-S2) is reported in Fig. 2.

196

P crit = 7.74 kN

Applied Load (kN)

C'
A [0.040 mm, 4.00 kN]
B [0.091 mm, 5.76kN ]
C [0.163 mm, 7.23 kN ]
C' [0.163 mm, 6.64 kN ]
D [0.545 mm, 7.69 kN ]

5
4

3
2
1
0
0.00

0.20

0.40

0.60

0.80

1.00

Global Slip (mm)

Fig. 2. Typical applied load-global slip response for monotonic quasi-static tests (Test-S2).
The initial monotonic linear increase of load is followed by a non-linear relationship
between load and slip. The end of the monotonic load increase with slip is marked by a
sudden load drop (Point C in Fig. 2). The portions of the load response before and after
Point C are referred to as pre- and post-peak responses, respectively. In the post-peak part
of the load response, the load levels off at a nominal constant value equal to 7.74 kN that is
identified as Pcrit. Pcrit was determined as the mean value of the load when the global slip
varied between 0.4 and 0.8 mm. The average Pcrit of the three static tests was 7.69 kN.
The axial strain distribution along the centerline of CFRP composite during the post-peak
response corresponding to point D is shown by the dotted curve in Fig. 3. [1, 3, 4]. The
approximated strain distribution using the function reported in [1] is also shown in Fig. 3.
0.008

yy

LSTZ

0.007
0.006

yy

0.005
0.004
0.003
0.002

Point D (figure 2) - Experiental Data

0.001

Point D (figure 2) - Fitting Curve

0
-0.001
0

20

40

60

80

100

120

140

160

y (mm)

Fig. 3. Axial strain along the FRP composite at point D of the load response in Fig. 2.
During debonding, the strain levels off at a value yy approximately equal to 6700 . The
average value of for the three monotonic quasi-static tests is 6500 . The strain
distribution can be divided into three main regions: (a) the unstressed region; (b) the stress
transfer zone (STZ); and (c) the fully debonded zone. Analysis of the strain data revealed
197

that the stress transfer zone was fully established when the load response attains Pcrit. Once
the STZ was fully established, as the global slip increased, there was a translation of the
STZ further along the length of the FRP sheet while its shape remained constant [1-4].
Within the range of the global slip 0.4mm 0.8mm used to compute Pcrit, the stress transfer
zone was fully established. The average value of the estimated length of STZ for the three
monotonic quasi-static tests is 80 mm.
The cohesive law, which relates the interface shear stress with the relative slip at each point
of the bonded area, can be obtained using the procedure outlined by Taljsten [1-4,11]. From
the measured strain yy along the FRP, the interface shear stress zy was calculated as:

zy Et

d yy

(1)

dy

where E and t are the Young modulus and the thickness of the FRP strip, respectively. The
relative slip, s(y), between FRP and concrete at a given location along the FRP is obtained
by integrating the axial strain in the FRP from the unloaded end up to that point, and
assuming it is zero at the unloaded end. The average values of the maximum shear stress
(max) and the corresponding interfacial slip (so) for the three monotonic quasi-static tests
are found to be equal to 6.43 MPa and 0.043 mm, respectively. The interfacial fracture
energy GF is obtained from the area under the entire zy s curve:
sf

GF zy ds

(2)

The average value of GF for the three monotonic quasi-static tests is 0.80 MPamm.
4. FATIGUE TESTS
Three direct shear tests (Test-F1, Test-F2 and Test-F3) were performed under fatigue
loading. The maximum (Pmax) and minimum (Pmin) applied loads are defined as a
percentage of the ultimate load Pcrit determined in the quasi static tests. If Pmax is less than
50% of the failure load, the test is said to be performed under a low-level cycling
conditions; otherwise the test is said to be performed under high-level cycling conditions.
High-level cycling has been investigated in this paper. The load range was 15%-80%, 15%
-70%, and 15%-60% for Test-F1, Test-F2 and Test-F3, respectively. Fatigue cycling was
performed in load control at 1 Hz. The load-slip response data were collected for ten cycles
at the prescribed intervals of 0.05 mm threshold value in the global slip. Failure of all tested
specimens was associated with progressive debonding of FRP up to a complete separation.
The collected load cycles of Test-F1 specimen are plotted versus the global slip in Fig. 4.
The hysteresis curves are qualitatively similar. The area enclosed by the loop represents the
energy dissipated in the cycle. In Test-F1, this area is nominally constant in the first half of
the test whereas it decreases in the second half while the loading-unloading behavior
becomes approximately parallel. Yun and Wu [12] observed that the area enclosed by the
loops does not change significantly when externally-bonded systems are tested between
198

Load (kN)

10% and 45% of the ultimate load. It should be noted that the total number of cycles at
failure for Test-F1 was 1290, which is significantly smaller than the number of cycles
reported in literature [7-9,12,13]. A smaller number of cycles can be attributed to the range
of fatigue load and the frequency. Bizindavyi et al. [13] tested all specimens at 1Hz.
Gheorghiu et al [8,9] tested at 2Hz or 3Hz depending on the load range and the maximum
number of cycles. Yun and Wu [12] tested externally-bonded systems at 5Hz.
7
6.5
6
5.5
5
4.5
4
3.5
3
2.5
2
1.5
1
0.5
0

Cycle# Cycle# Cycle# Cycle# Cycle# Cycle# Cycle# Cycle# Cycle# Cycle#
210
243
558
956
1055 1246
375 394
676
795

Pmax = 6.00 kN
Pmin = 1.25 kN

0.1

0.2

0.3

0.4

0.5

Global Slip (mm)

0.6

0.7

0.8

0.9

Fig. 4. Applied load vs. global slip response of Test-F1.

The total number of cycles for Test-F2 and Test-F3 was 13192 and 116995, respectively.
The number of cycles dramatically increases if the fatigue-load range decreases. The
enclosed loop area increases in Test-F2 and Test-F3. The ratio of the area of the first block
of cycles over the area of the last block of cycles was calculated to be 1.2 for Test-F2 and
1.38 for Test-F3, indicating that as the load range decreases the dissipated energy increases
within the test as the number of cycles increases.
The slopes of the loops in Test-F1, Test-F2 and Test-F3 are found to be gradually
decreasing with the number of cycles, indicating a progressive loss in the stiffness of the
interfacial bond between CFRP and concrete substrate. Yun and Wu [12] reported that
under low-level fatigue load, the slopes of the hysteresis loops remain unchanged, except
for early loops which show a slightly decrease of the slope. Bizindavyi et al. [13] reported a
decrease of the slope of the hysteresis loops. The rate of change of slope was observed to
increase with increasing load amplitude. Gheorghiu et al. [8,9] reported