frcm composites for strengthening corrosion-damaged
TRANSCRIPT
FRCM Composites for Strengthening Corrosion-damaged
Structures: Experimental and Numerical Investigations
Thèse
Mohammed Elghazy
Doctorat en génie civil
Philosophiae doctor (Ph.D.)
Québec, Canada
© Mohammed Elghazy, 2018
Les matrices cimentaires renforcées de fibres (MCRF) pour
renforcer les structures en béton endommagées par la
corrosion: investigations expérimentales et numériques
Thèse
Mohammed Elghazy
Sous la direction de :
Ahmed El Refai, directeur de recherche
iii
Résumé
La corrosion des armatures en acier est l'un des mécanismes les plus destructifs pour les structures
en béton armé. La corrosion nuit non seulement à l'intégrité structurale et à l’aptitude au service
de la structure endommagée, mais peut aussi entraîner des défaillances inattendues ou des ruptures
fragiles. Malgré les dispositions rigoureuses de la plupart des codes de pratique pour éviter la
corrosion, des signes de dommages dus à la corrosion sont toujours signalés.
Récemment, des systèmes à matrice cimentaire renforcée de fibre (MCRF) ont été proposés
comme une technique innovante de renforcement/réparation pour les structures en béton afin de
surmonter les inconvénients associés à l'utilisation des systèmes de polymères renforcés de fibres
(PRF). Bien que l'utilisation de composites MCRF pour renforcer les éléments en béton non
endommagés ait prouvé son efficacité, très peu est connu sur la viabilité de leur utilisation pour
renforcer les éléments en béton endommagés à divers niveaux dus à la corrosion. De plus, les
comportements de post-réparation et la durabilité à long-terme des éléments corrodés et renforcés
par les systèmes MCRF et qui seront probablement exposés aux mêmes conditions
environnementales qui prévalaient avant leur réparation, n'ont pas retenu l'attention des chercheurs
dans la littérature. De plus, la plupart de nos infrastructures, telles que les ponts et garages de
stationnement, sont susceptibles d'être endommagées par la corrosion tout en étant soumises à des
charges oscillatoires qui provoquent de la fatigue. À ce jour, aucune information n'est disponible
sur l'effet de la combinaison de la charge de fatigue et de la corrosion dans les structures renforcées
par les systèmes MCRF.
Dans ce travail, les comportements monotones et de fatigue en flexion des poutres en béton
endommagées par la corrosion et renforcées par des systèmes MCRF ont été étudiés en plus de
leur performance à long-terme, c'est-à-dire après une exposition à un environnement corrosif après
leur renforcement. Le travail comprend des investigations expérimentales et numériques. Les
prédictions analytiques et les formulations théoriques actuellement disponibles dans les codes de
conception ont été aussi vérifiées par rapport aux résultats expérimentaux. Le programme
expérimental consistait à tester trente (30) poutres en béton à grande échelle de 150 × 250 × 2800
mm. Les poutres ont été construites et testées en configuration de charge à quatre points. Un
processus accéléré de corrosion a été utilisé pour corroder les armatures d'acier en traction dans le
tiers central des poutres. Les paramètres d'essai comprenaient le niveau de corrosion (représenté
iv
par 10, 20 et 30% de perte de masse dans l'acier de traction), le type de système de renforcement
utilisé (Polyparaphénylène benzobisoxazole (PBO-MCRF), MCRF de carbone et PRF), la quantité
de composites MCRF (1, 2, 3 et 4 couches), le schéma de renforcement MCRF (couches ancrées
aux extrémités par rapport aux couches continues sous forme U) et le régime de chargement
(monotone et fatigue).
Les résultats des tests ont montré que l'utilisation de composites MCRF améliorait
significativement le comportement en flexion des poutres corrodées. Les composites MCRF ont
contrôlé le mode de défaillance des poutres renforcées plutôt que le niveau de corrosion des barres
d'acier. Les poutres renforcées par la MCRF ont montré une augmentation de leurs résistances
ultimes variant entre 7 et 65% de celles des poutres vierges (poutres ni corrodées ni renforcées) en
fonction du type, de la quantité et du schéma de la MCRF utilisée. L'exposition des poutres
réparées par la MCRF à d’autres cycles de corrosion a entraîné une réduction de 23% de la perte
de masse de l'acier. Le schéma en U était plus efficace que le schéma d'ancrage aux extrémités à
retarder le délaminage des couches de MCRF dans les poutres renforcées et testées à court terme.
Il a également atténué l'effet des fissures de corrosion longitudinales et, par conséquent, a
augmenté l'efficacité du renforcement MCRF. Les essais de fatigue ont montré que la corrosion
des barres d'acier diminuait considérablement la résistance à la fatigue des poutres non renforcées.
Le renforcement avec des composites MCRF a augmenté la durée de vie en fatigue des poutres
endommagées par la corrosion de 38 à 377% de celle des poutres corrodées non-renforcées.
Cependant, le renforcement par MCRF n'a pas restauré la durée de vie en fatigue des poutres
vierges.
Dans l'étude numérique réalisée dans ce travail, des modèles d'éléments finis (ÉF)
tridimensionnels (3D) ont été développés pour simuler le comportement non linéaire des poutres
corrodées et renforcées par des composites MCRF et PRF à l'aide du progiciel ATENA-3D. Les
résultats de l'analyse numérique étaient en bon accord avec ceux obtenus expérimentalement en
termes de modes de défaillance, de déformations, de capacités de charge et de flèches. Les modèles
ÉF développés ont été capables de capturer le comportement non-linéaire des poutres testées avec
une bonne précision. Une étude paramétrique a ensuite été menée pour étudier l'effet de la
résistance à la compression du béton et de l'épaisseur de recouvrement des armatures sur l'efficacité
de renforcement des systèmes composites. Il a été observé que la rupture des poutres renforcées
v
par des FRCM était indépendante de la résistance à la compression du béton ou de l'épaisseur de
de recouvrement et était régie uniquement par le glissement du tissu dans la matrice.
Sur le plan analytique, les équations de conception de l’ACI-549.4R-13 (ACI 2013) ont été
évaluées à l'aide des données expérimentales obtenues à partir des tests. Il a été conclu que les
formulations théoriques de l’ACI-549.4R-13 peuvent raisonnablement prédire les résistances
ultimes des poutres renforcées ancrées à l'extrémité mais sous-estimer celles des poutres ancrées
en continu sous forme U. Un facteur de schéma de 1,1 a ensuite été proposé pour calculer la
résistance nominale des poutres renforcées par MCRF sous forme U.
Le résultat de ce travail a été publié (ou soumis pour publication) dans cinq articles de revues et
cinq conférences, comme détaillé tout au long de la thèse.
vi
Abstract
Corrosion of steel reinforcement is one of the most destructive mechanisms for reinforced
concrete (RC) structures. Corrosion not only impairs the structural integrity and the serviceability
of the damaged structure, but it may also lead to unexpected and brittle failures. Despite the
rigorous provisions of most codes of practice to avoid corrosion, evidences of corrosion damage
are still being reported.
Recently, fabric-reinforced cementitious matrix (FRCM) systems were proposed as an innovative
strengthening/repair technique for RC structures to overcome the drawbacks associated with the
use of the well-documented fiber-reinforced polymer (FRP) systems. While the use of FRCM
composites to strengthen un-damaged RC members has proven its efficiency, very little is known
about the viability of their use to retrofit RC members with various levels of corrosion damage. In
addition, the post-repair performance and the long-term durability of the FRCM-strengthened
corroded members, which most likely will be exposed to the same environmental conditions that
have prevailed prior their repair, have not received attention in the literature. Moreover, most of
our infrastructures such as bridges and parking garages are susceptible to corrosion damage while
continuously being subjected to oscillatory loads that cause fatigue. To date, no information is
available about the effect of combining fatigue loading with corrosion in FRCM-strengthened
structures.
In this work, the monotonic and fatigue flexural behaviors of corrosion-damaged RC beams
strengthened with FRCM systems were investigated in addition to their long-term performance,
i.e. after further exposure to corrosive environment following their strengthening. The work
includes experimental and numerical investigations. The analytical predictions and theoretical
formulations that are currently available in the design codes have been verified against the
experimental results. The experimental program consisted of testing thirty (30) large-scale RC
beams of 150×250×2800 mm. The beams were constructed and tested under four-point load
configuration. An accelerated corrosion process was utilized to corrode the bottom steel
reinforcement in the middle third of the test specimens. The test parameters included the level of
corrosion damage (represented by 10, 20, and 30% mass loss in the tensile steel), the type of the
strengthening system used (Polyparaphenylene benzobisoxazole (PBO-FRCM), C-FRCM, and
FRP), the amount of FRCM composites (1, 2, 3, and 4 layers), the FRCM strengthening Scheme
vii
(end-anchored versus continuously wrapped layers), and the loading regime (monotonic and
fatigue).
The test results showed that the use of FRCM composites significantly enhanced the flexural
behavior of the corroded beams. FRCM governed the failure mode of the strengthened beams
rather than the level of corrosion damage of the steel bars. FRCM-strengthened beams showed an
increase in their ultimate strengths that ranged between 7 and 65% of that of the virgin (neither
corroded nor strengthened) beam based on the type, amount, and Scheme of the FRCM used.
Exposing the repaired beams to post-repair corrosion resulted in 23% reduction in the steel mass
loss. The U-wrapped scheme was more efficient than the end-anchoring scheme in delaying the
delamination of the FRCM plies in the short-term repaired beams. It also mitigated the effect of
the longitudinal corrosion cracks and consequently increased the post-repair strengthening
effectiveness of FRCM systems. Fatigue tests showed that corrosion of steel bars dramatically
decreased the fatigue life of the unstrengthened-beams. Strengthening with FRCM composites
increased the fatigue life of the corrosion-damaged beams by 38 to 377% of that of the corroded-
unstrengthened beams. However, FRCM strengthening did not restore the fatigue life of the virgin
beams.
In the numerical study carried out in this work, three-dimensional finite element (FE) models
were developed to simulate the nonlinear behavior of the corroded beams strengthened with FRCM
and FRP composites using the software package ATENA-3D. The results of the numerical analysis
were in good agreement with those obtained experimentally in terms of failure modes, strains,
load-carrying capacities, and deflections. The developed FE models were able to capture the non-
linear behavior of the tested beams with good accuracy. A parametric study was then conducted
to investigate the effect of concrete compressive strength and thickness of concrete cover on the
strengthening effectiveness of the composite systems. It was observed that failure of FRCM-
strengthened beams was independent of the compressive strength of concrete or the thickness of
the concrete cover and was governed only by fabric slippage within the matrix.
Analytically, the design equations of ACI-549.4R-13 (ACI 2013) were assessed using the
experimental data obtained from the tests. It was concluded that the theoretical formulations of
ACI-549.4R-13 can reasonably predict the ultimate strengths of the end-anchored strengthened
beams but underestimated those of continuously-anchored beams. A scheme factor of 1.1 was then
viii
proposed to calculate the nominal strength of beams strengthened with continuously-wrapped
shape of FRCM.
The outcome of this work has been published (or submitted for publication) in five journal
articles and five conferences, as detailed throughout the thesis.
ix
Table of Contents
Résumé ……………………………………………………………......………………….…….. iii
Abstract ………………………………..………..……..……………………………..………… vi
Table of Contents …………………….………..……..…...……….………………...…..……… ix
List of Figures ……….…………….…..………..……..……………….………………....…… xiv
Acknowledgements……….………….…...……..……..…………………..……………....…… xx
Chapter 1 ....................................................................................................................................... 1
1.1 Background and Problem Definition................................................................................ 1
1.2 Thesis Structure ................................................................................................................ 3
Chapter 2 ....................................................................................................................................... 6
2.1 General ............................................................................................................................. 6
2.2 Corrosion of Steel Reinforcement .................................................................................... 6
2.2.1 Corrosion Mechanism of Steel Bars in Concrete ...................................................... 6
2.2.2 Accelerated Corrosion Process ................................................................................. 9
2.2.3 Effect of Steel Corrosion on Concrete Structures ................................................... 11
2.2.4 Behavior of Corroded-RC Beam under Monotonic Loads ..................................... 12
2.2.5 Behavior of Corroded-RC Beam under Fatigue Loads........................................... 15
2.3 FRCM Composites ......................................................................................................... 16
2.3.1 FRCM Acceptance Criteria and Design ................................................................. 18
2.3.2 FRCM Tensile Characterization ............................................................................. 18
2.3.3 Bond Behavior of FRCM Composites .................................................................... 20
2.4 FRCM-Strengthened Beams Under Monotonic Load .................................................... 21
2.5 Fatigue and Durability of FRCM-strengthened Beams.................................................. 25
2.6 Findings of Literature Review........................................................................................ 26
Chapter 3 ..................................................................................................................................... 27
3.1 Research Significance .................................................................................................... 27
3.2 Research Objectives ....................................................................................................... 27
3.3 Methodology .................................................................................................................. 28
3.3.1 Experimental Work Program .................................................................................. 28
3.3.2 Numerical Analysis ................................................................................................. 31
3.3.3 Analytical Investigation .......................................................................................... 32
Chapter 4 ..................................................................................................................................... 33
Résumé ...................................................................................................................................... 34
x
4.1 Abstract .......................................................................................................................... 34
4.2 Introduction and Background ......................................................................................... 34
4.3 Experimental Program.................................................................................................... 36
4.3.1 Test Specimen ......................................................................................................... 37
4.3.2 Accelerated Corrosion Aging ................................................................................. 39
4.3.3 Materials ................................................................................................................. 40
4.3.4 FRCM Equivalent Axial Stiffness .......................................................................... 42
4.3.5 FRCM Repair Schemes........................................................................................... 43
4.3.6 Repair Technique .................................................................................................... 44
4.3.7 Test Setup and Instrumentation .............................................................................. 45
4.4 Test Observations ........................................................................................................... 46
4.4.1 Corrosion Cracks and Mass Loss ............................................................................ 46
4.4.2 Modes of Failure ..................................................................................................... 46
4.4.3 Load-deflection Response ....................................................................................... 48
4.4.4 Strength Analysis .................................................................................................... 49
4.4.5 Ductility Performance ............................................................................................. 53
4.4.6 Strain Response ....................................................................................................... 54
4.5 Theoretical Predictions ................................................................................................... 58
4.5.1 Design Provision ..................................................................................................... 61
4.6 Conclusions .................................................................................................................... 61
4.7 Notation .......................................................................................................................... 63
Chapter 5 ..................................................................................................................................... 65
Résumé ...................................................................................................................................... 66
5.1 Abstract .......................................................................................................................... 66
5.2 Introduction and Background ......................................................................................... 66
5.3 Experimental Program.................................................................................................... 68
5.3.1 Test Specimen and Materials .................................................................................. 69
5.3.2 Accelerated Corrosion Process ............................................................................... 71
5.3.3 FRCM Systems ....................................................................................................... 72
5.3.4 Strengthening Schemes ........................................................................................... 75
5.3.5 FRCM Installation .................................................................................................. 77
xi
5.3.6 Instrumentation and Test Setup .............................................................................. 77
5.4 Test Results .................................................................................................................... 78
5.4.1 Corrosion Observations .......................................................................................... 78
5.4.2 Failure Mechanisms ................................................................................................ 80
5.4.3 Strength Response ................................................................................................... 81
5.4.4 Fabric Strains .......................................................................................................... 86
5.4.5 Ductility and Energy Absorption ............................................................................ 87
5.5 Discussion ...................................................................................................................... 89
5.6 Predicted Strength Results ............................................................................................. 92
5.7 Conclusions .................................................................................................................... 95
Chapter 6 ..................................................................................................................................... 97
Résumé ...................................................................................................................................... 98
6.1 Abstract .......................................................................................................................... 98
6.2 Introduction and Background ......................................................................................... 98
6.3 Experimental Program.................................................................................................. 100
6.3.1 Test Specimen and Materials ................................................................................ 101
6.3.2 Accelerated Corrosion Process ............................................................................. 103
6.3.3 FRCM Composites................................................................................................ 104
6.3.4 FRCM Schemes .................................................................................................... 106
6.3.5 Repair Methodology ............................................................................................. 108
6.3.6 Test Setup and Instrumentation ............................................................................ 108
6.4 Test Results and Discussion ......................................................................................... 108
6.4.1 Corrosion Crack Pattern ........................................................................................ 108
6.4.2 Steel Mass Loss..................................................................................................... 109
6.4.3 Flexural Cracks Pattern and Failure modes .......................................................... 111
6.4.4 Flexural Response ................................................................................................. 114
6.4.5 Load-carrying Capacities ...................................................................................... 116
6.4.6 Ductility and Energy Absorption .......................................................................... 118
6.4.7 Strain Response ..................................................................................................... 120
6.5 Predicted Strengths ....................................................................................................... 123
6.6 Conclusions .................................................................................................................. 124
Chapter 7 ................................................................................................................................... 126
xii
Résumé .................................................................................................................................... 127
7.1 Abstract ........................................................................................................................ 127
7.2 Introduction and Background ....................................................................................... 127
7.3 Experimental Program.................................................................................................. 129
7.3.1 Test Specimen ....................................................................................................... 130
7.3.2 Accelerated Corrosion Technique......................................................................... 132
7.3.3 FRCM Composites................................................................................................ 133
7.3.4 FRCM Strengthening Configuration..................................................................... 135
7.3.5 Strengthening Procedure ....................................................................................... 137
7.3.6 Test Setup and Instrumentation ............................................................................ 138
7.4 Test Results and Discussion ......................................................................................... 139
7.4.1 Corrosion Crack Patterns and Actual Steel Mass Loss ......................................... 139
7.4.2 Monotonic Test Results ........................................................................................ 140
7.4.3 Fatigue Test Results .............................................................................................. 143
7.5 Conclusions .................................................................................................................. 151
Chapter 8 ................................................................................................................................... 154
Résumé .................................................................................................................................... 155
8.1 Abstract ........................................................................................................................ 155
8.2 Introduction and Background ....................................................................................... 155
8.3 Experimental Investigation .......................................................................................... 158
8.3.1 Test Matrix ............................................................................................................ 158
8.3.2 Test Specimen ....................................................................................................... 158
8.3.3 Externally-bonded Composite Systems ................................................................ 160
8.3.4 Strengthening Procedure and Configuration ......................................................... 161
8.3.5 Test Setup and Instrumentation ............................................................................ 162
8.4 Numerical Simulation .................................................................................................. 162
8.4.1 Constitutive Laws ................................................................................................. 163
8.5 Results and Discussion ................................................................................................. 170
8.5.1 Crack Pattern at Failure ........................................................................................ 170
8.5.2 Load-deflection Response ..................................................................................... 172
8.5.3 Load-carrying Capacity ........................................................................................ 177
8.5.4 Strain Response ..................................................................................................... 179
xiii
8.5.5 Ductility ................................................................................................................ 182
8.6 Parametric Studies ........................................................................................................ 183
8.6.1 Effect of Concrete Compressive Strength (fc′)..................................................... 183
8.6.2 Effect of Concrete Cover ...................................................................................... 186
8.7 Conclusions .................................................................................................................. 188
Chapter 9 ................................................................................................................................... 191
8.8 Summary ...................................................................................................................... 191
8.9 Conclusions .................................................................................................................. 191
8.9.1 Effect of Corrosion on RC Beams ........................................................................ 191
8.9.2 Short-term Performance of Corroded Beams Strengthened with FRCM ............. 192
8.9.3 Long-term Performance of Corroded Beams Strengthened with FRCM ............. 194
8.9.4 Validation of ACI-549.4R-13 Design Equations .................................................. 195
8.9.5 Fatigue Performance of Corroded Beams Strengthened with FRCM .................. 196
8.9.6 Numerical simulation ............................................................................................ 197
8.10 Recommendation for Future Work .............................................................................. 198
8.11 Impact of Current Research.......................................................................................... 198
References .................................................................................................................................. 199
Bibliography .............................................................................................................................. 212
xiv
List of Figures
Figure 2.1: Corrosion process of steel bars inside concrete [5] ...................................................... 8
Figure 2.2: A schematic of microcell and macrocell types of corrosion [5] ................................... 8
Figure 2.3: FRCM composite system ........................................................................................... 17
Figure 2.4: Different fabric configurations [64] .......................................................................... 17
Figure 2.5: Actual and idealized stress-strain curve of FRCM coupon in tension [67]................ 19
Figure 2.6: Failure mechanism of a fiber bundle embedded in cementitious matrix [74] ............ 21
Figure 4.1: Typical dimensions and reinforcement details of the test specimen .......................... 38
Figure 4.2: Specimens connected in series inside the corrosion chamber .................................... 40
Figure 4.3: Strengthening materials: a) unbalanced PBO fabric, b) unidirectional carbon fabric,
and c) unidirectional carbon fabric ............................................................................................... 41
Figure 4.4: Idealized tensile stress-strain curve of FRCM coupon specimen [54] ....................... 43
Figure 4.5: Repair schemes: (a) Scheme I and (b) Scheme II ...................................................... 44
Figure 4.6: Repair procedure: a) removing the deteriorated concrete, b) patch repair, c)
roughening the concrete surface with sandblasting, and d) FRCM application ........................... 45
Figure 4.7: Positions of the electrical strain gauges along the outer fabric .................................. 45
Figure 4.8: Corrosion cracks pattern for specimen CU ................................................................ 46
Figure 4.9: Typical modes of failure: (a) SY-CC in beam CR-1P-I, (b) FD in beam CR-2P-I, (c)
FS in beam CR-2P-II, (d) MC-SFM in beam CR-3C-II, and (e) LR in beam CR-1FRP-I ........... 48
Figure 4.10: Effect of number of PBO-FRCM plies on the load-deflection curves ..................... 50
Figure 4.11: Effect of the repair scheme on the load-deflection curves ....................................... 50
Figure 4.12: Effect of FRCM materials on the load-deflection curves......................................... 51
Figure 4.13: Normalized ultimate load versus the equivalent stiffness ........................................ 53
Figure 4.14: Load-strain curves for specimens with repair Scheme I .......................................... 55
Figure 4.15: Load-strain curves for specimens with repair Scheme II ......................................... 56
Figure 4.16: Strain profile in the PBO fabric for specimen CR-4P-I ........................................... 57
Figure 4.17: Strain profile in the PBO fabric for specimen CR-4P-II .......................................... 58
Figure 4.18: Strain profile in the carbon fabric for specimen CR-3C-II ...................................... 58
Figure 4.19: Stress and strain distribution at ultimate stage ......................................................... 59
Figure 5.1: Test specimen geometry and reinforcement details. (All dimensions in mm) ........... 71
Figure 5.2: Specimens inside the environmental chamber during a dry cycle ............................. 72
xv
Figure 5.3: FRCM systems: a) PBO-FRCM (Unbalanced PBO fabric) and b) C-FRCM
(Unidirectional carbon fabric)....................................................................................................... 73
Figure 5.4: Stress-strain relationships for FRCM-tensile coupons [97] ....................................... 74
Figure 5.5: Strengthening schemes: a) Scheme I and b) Scheme II ............................................. 76
Figure 5.6: FRCM installation procedure: a) removing the deteriorated concrete, b) patch
repairing and sandblasting, and c) installation of PBO-FRCM composite ................................... 77
Figure 5.7: Profile of steel bars: a) uncorroded bar, b) corroded bar extracted from CSA-4P-I
(12.6% mass loss), c) corroded bar extracted from CSB-3C-II (18.6% mass loss), and d)
corroded bar extracted from CUC (22.5% mass loss) .................................................................. 79
Figure 5.8: Actual and theoretical mass loss versus the duration of corrosion process ................ 79
Figure 5.9: Modes of failure: a) FRCM delamination, b) fabric slippage with partial fabric
debonding within the matrix, and c) matrix cracking with extensive fabric slippage .................. 81
Figure 5.10: Load-deflection relationships for corroded-unstrengthened beams ......................... 82
Figure 5.11: Load-deflection relationships for corrosion-damaged FRCM-strengthened beams: a)
beams of Group A, b) beams of Group B, and c) beams of Group C ........................................... 83
Figure 5.12: Normalized strength versus the FRCM equivalent axial stiffness, Kf ...................... 85
Figure 5.13: Load versus outer fabric strain for beams strengthened in a) Scheme I and b)
Scheme II ...................................................................................................................................... 86
Figure 5.14: Normalized ductility index versus stiffness factor 𝛽𝑓 % ......................................... 89
Figure 5.15: Normalized energy absorption index versus stiffness factor 𝛽𝑓 % .......................... 89
Figure 5.16: Effect of corrosion damage on the ultimate strength of strengthened beams .......... 90
Figure 5.17: Normalized strength versus stiffness factor 𝛽𝑓 % .................................................... 91
Figure 5.18: Predicted versus experimental flexural response for beams strengthened with a) two
layers of PBO-FRCM in Scheme I, b) four layers of PBO-FRCM in Scheme II, and c) three
layers of C-FRCM in Scheme II ................................................................................................... 94
Figure 6.1: Schematic of the testing procedure of the short- and long-term beams ................... 101
Figure 6.2: Typical dimensions and reinforcement details of the test beam (all dimensions in
mm) ............................................................................................................................................. 102
Figure 6.3: A schematic of the electrical connection .................................................................. 104
Figure 6.4:FRCM systems; a) PBO-FRCM (unbalanced PBO fabric) and (b) C-FRCM
(unidirectional carbon fabric) - all dimensions in mm ................................................................ 105
xvi
Figure 6.5: Idealized tensile stress-strain curve of FRCM coupon specimen ACI 549.4R-13 [54]
..................................................................................................................................................... 106
Figure 6.6: FRCM repair schemes; a) Scheme I and b) Scheme II ............................................ 107
Figure 6.7: Corrosion cracks patterns; a) typical corrosion cracks pattern for short-term
specimens (beam CU); b) beam CRL-4P-I; c) beam CRL-4P-II; and d) beam CRL-3C-II ....... 110
Figure 6.8: Failure mode of a) beam CU due to steel yielding and concrete crushing; b) beam
CRS-4P-I due to FRCM delamination; c) beam CRL-4P-I due to premature FRCM delamination;
d) beam CRL-4P-II due to PBO-fabric debonding from matrix; and e) beam CRL-3C-II due to
fabric slippage ............................................................................................................................. 112
Figure 6.9: Load-deflection relationships of a) short-term beams; b) long-term beams; c) beams
repaired with PBO-FRCM (short-term and long-term); and d) beams repaired with C-FRCM
(short-term and long-term) .......................................................................................................... 116
Figure 6.10: Load-strains relationships for a) beams repaired with PBO-FRCM and b) beams
repaired with C-FRCM ............................................................................................................... 122
Figure 7.1: Geometry and reinforcement details of the test specimen (all dimensions in mm) . 132
Figure 7.2: Specimens connected in series inside the corrosion chamber .................................. 133
Figure 7.3: a) Unbalanced PBO fabric and b) Unidirectional carbon fabric. ............................. 134
Figure 7.4: FRCM strengthening schemes: a) Scheme I and b) Scheme II ................................ 136
Figure 7.5: FRCM strengthening procedure: a) removing the deteriorated concrete after
corrosion, b) patch repair, c) PBO-FRCM application, and d) C-FRCM application ................ 137
Figure 7.6: Test setup .................................................................................................................. 138
Figure 7.7: Corrosion crack pattern for specimen FCU .............................................................. 139
Figure 7.8: Profile of the steel bars: a) Uncorroded bar, b) corroded steel bar extracted from
beam MCS-4P-II, and c) fatigue rupture of a corroded steel bar extracted from FCS-3C-II ..... 140
Figure 7.9: Failure modes of the monotonically tested beams: a) Steel yielding followed by
concrete crushing; b) FRCM delamination; c) Fabric slippage and partial debonding; and d)
Matrix cracking followed by fabric slippage .............................................................................. 141
Figure 7.10: Load-deflection relationships of the monotonically tested beams ......................... 142
Figure 7.11: Variation of fatigue life of the strengthened beams with the stiffness factor 𝛽𝑓 .... 144
Figure 7.12: Load-deflection Hysteresis for a) beam FCU; b) beam FCS-4P-I; c) beam FCS-4P-
II; and d) beam FCS-3C-II .......................................................................................................... 147
xvii
Figure 7.13: Fatigue cracks at midspan of the strengthened beams (Side view) ........................ 148
Figure 7.14: Fatigue rupture of steel bars in a) beam FCU and b) beam FCS-4P-I ................... 148
Figure 7.15: Effect of fatigue cycles on the stiffness of the tested beams .................................. 149
Figure 7.16: Effect of fatigue cycles on the concrete and fabric strains ..................................... 150
Figure 8.1: Geometry and details of steel, P-FRCM, and C-FRP reinforcement of the tested
beams (all dimensions in mm)..................................................................................................... 159
Figure 8.2: a) PBO fabric used in the FRCM composite system and b) Carbon sheets used in the
FRP composite system ................................................................................................................ 160
Figure 8.3: Strengthening procedure of corroded beams: a) removing the deteriorated concrete,
b) patch repairing and sandblasting, c) installation of P-FRCM composite, and d) installation of
C-FRP sheets ............................................................................................................................... 162
Figure 8.4: Constitutive laws of concrete and cementitious matrix: a) compressive hardening law,
b) compressive softening law, and c) tensile softening law ........................................................ 164
Figure 8.5: C-FRP/concrete interfacial bond stress-slip model according to Lu et al. [117]
adopted for various concrete mixes ............................................................................................ 167
Figure 8.6: PBO-fabric /matrix interfacial bond stress-slip model according to D’Ambrisi et al.
[120] ............................................................................................................................................ 168
Figure 8.7: a) Meshing of P-FRCM strengthened beams, b) reinforcement layout for beams
strengthened with 4 layers of P-FRCM, and c) reinforcement layout for beams strengthened with
C-FRP sheet ................................................................................................................................ 169
Figure 8.8: Numerical and experimental crack patterns at failure for a) beam UUa, b) beam CS-
A-1C, and c) beam CS-A-4P ...................................................................................................... 172
Figure 8.9: Numerical and experimental load-deflection responses for beams of group A ....... 173
Figure 8.10: Numerical and experimental load-deflection responses for beams of group B ..... 176
Figure 8.11: Gain and decline in % in the ultimate loads, Pu, with respect to that of the control
beam (UU) .................................................................................................................................. 177
Figure 8.12: Fiber and concrete strain response ......................................................................... 180
Figure 8.13: Tensile steel strain response for beams CS-A-1C and CS-A-4P............................ 181
Figure 8.14: Strain profile of the internal and external reinforcement at ultimate ..................... 181
Figure 8.15: Effect of the concrete compressive strength on the load-deflection response for
beams a) CS-A-1C, b) CS-A-2P, and c) CS-A-4P...................................................................... 185
xviii
Figure 8.16: Effect of the concrete cover on the load-deflection response for beams a) CS-A-1C,
b) CS-A-2P, and c) CS-A-4P ...................................................................................................... 188
xix
To the researchers and engineers who appreciate the value of science and
knowledge
To the ones who were
and will always be
in my mind….
in my heart….
.…and in my Soul
Mohammed Elghazy
xx
Acknowledgements
Every journey has to have an ambitious and a defined end. My journey as a PhD student started
in December 2014. My main objective was to acquire my PhD at Laval. During the last three years,
I have developed another goal for myself: to explore my own feasibility as a researcher and how
to strengthen myself as an individual. The outcomes were greatly dependent on people, financial
status, and the surrounding environment.
Therefore, I would like to express my sincere appreciation and gratitude to my thesis supervisor
Prof. A. El Refai for his guidance and continuous encouragement during the course of this work.
Thanks for being a friend and a firm supervisor. I would like also to express my sincere gratitude
to our collaborators Prof. A. Nanni from University of Miami and Prof. U. Ebead from Qatar
University, for their encouragement and support during my research.
I would also like to acknowledge and thank all the technicians of the structural laboratory at
Laval University, especially, Mr. R. Malo for his precious assistance through the experimental
tests. The help provided by Mr. K. Attia, Mr. A. Abbadi, and Mr. N. Allam during the tests is
greatly appreciated.
I would like to acknowledge Laval University and Qatar Foundation for the financial support.
The donation of composite materials provided by Ruredil, Italy and Simpson Strong-Tie, USA
represented by Mr. Brad Erickson is greatly appreciated.
I would like to express my sincere thanks and appreciation to my mother. Without her
unconditional love, support, and encouragement, the achievement of this work would not be
possible. I cannot end my acknowledgement without expressing my deep appreciation to my sister
from the bottom of my heart for her encouragement and endless love
1
1. Chapter 1
Introduction
1.1 Background and Problem Definition
Repair/strengthening of reinforced concrete (RC) structures is motivated by several factors
including aging, change in use, increased loads, code compliance, and environmental damages
such as corrosion. Corrosion of steel reinforcement is one of the major durability concerns for
concrete structures especially in coastal areas and in cold regions where de-icing salts are heavily
used. Pitting corrosion reduces the cross-sectional area of the steel bars and may lead to significant
loss of their ductility [1,2]. The expansion of corrosion products causes concrete cracking and
impair the composite action between steel and concrete. As a result, the load-carrying capacity and
the service life of the corroded member are considerably jeopardized [3–5].
Corrosion of steel reinforcement combined with fatigue stresses significantly reduces the fatigue
life of RC structures and may lead to unexpected failures especially in bridges [6,7]. According to
the U.S. Federal Highway Administration report published in 2002 [8], approximately 15% of the
highway bridges in USA are considered structurally deficient. Their maintenance and repair costs
exceed 8.3$ billion dollars annually. In Canada, the total direct costs of corrosion were 23.6$
billion dollars in 2003 [9]. The statistics for Europe, Asian Pacific countries, and Australia are not
by any means better than those in North America. Government agencies and industrial firms are
looking for more durable and less costly materials and techniques to maintain and repair our
infrastructures. Hence, structural engineers are in continuous search for new construction materials
and innovative rehabilitation techniques for such deficient structures.
In the last decades, externally-bonded strengthening technologies based on organic matrices,
referred to as fiber-reinforced polymers (FRP), have proven success in restoring the serviceability
and strength of RC structures [10–12]. However, several problems associated with the use of FRP
have been documented. FRP materials are flammable and prone to deterioration and the loss of
their mechanical and bond properties at high temperatures [13]. Their epoxy-based agents lose
their stiffness and strength and change from a stiff material to a viscous material with poor
properties when exposed to elevated temperatures, which affect their bond strength [14,15].
2
Structures located in hot climates or those at risk of fire can easily be vulnerable when exposed to
elevated temperatures. The toxicity nature of epoxy and its poor thermal compatibility to the
concrete substrate add another dimension to the drawbacks of FRP systems [16].
In order to overcome the drawbacks associated with the use of FRP composites, the need to
replace the organic binder (epoxy) by an inorganic binder has been raised. The use of fabric-
reinforced cementitious matrix (FRCM) systems has been introduced as an alternative promising
strengthening technique to FRPs. The FRCM system is a composite material consisting of fabric
meshes made of long dry-woven intermittent in two orthogonal directions and embedded in
cement-based matrix that serves as a binder. The embedded grid is shielded between the mortar
layers thus minimizing its fire vulnerability. In addition, the compatibility between the mortar used
and the concrete substrate is inherited since both materials have cement as a common “base”. The
cement-based mortars used in FRCM also act as barriers against chloride ions penetration thus
protecting the main reinforcing bars from corrosion attack. Its lightweight, high tensile strength,
and ease of application makes the system very appealing to engineers. With their innovative
features, FRCM systems can ensure the endurance of the rehabilitation process and consequently
the sustainability of the strengthened structure.
Recently, significant efforts have been made to introduce FRCM composites in the construction
industry and the use of FRCM composites to strengthen RC structures was initiated in several field
applications. A significant amount of research has been devoted to study the flexural and shear
behaviors of undamaged RC members strengthened with FRCM composites. Nonetheless, the
feasibility of using FRCM composites to strengthen corrosion-damaged RC structures has received
little attention. The challenge in using FRCM to repair/strengthen corrosion-damaged concrete
members rises from the fact that the technique uses cementitious mortars rather than epoxies,
which necessitates proper surface preparation to ensure adequate mechanical and chemical
adhesion to the concrete substrate. Factors such as the absorption properties of the substrate, its
degree of carbonation, its moisture condition, and its cleanness, in addition to the existence of
micro-cracks and/or contaminants on the repaired surface may affect the desired performance.
Unfortunately, surfaces of corroded concrete structures are characterized by their random texture,
unpredictable crack distribution, and concrete fragmentation. Volume expansion resulting from
corrosion of steel bars in addition to the existence of corrosion products cause the weakness of the
3
surrounding concrete and the loss of its integrity. In fact, these conditions might consist a serious
obstacle to the use of FRCM in strengthening corroded structures. To the authors’ knowledge, very
few of the previous studies in which FRCM technique was adopted has addressed this problem,
not to mention the long-term performance of the strengthened members being exposed to the same
conditions that have caused their deterioration. Therefore, the work presented herein aimed at
tackling these problems and filling the gap in our knowledge on the use of FRCM composites in
strengthening corrosion-damaged RC elements.
1.2 Thesis Structure
The purpose of this work is to investigate the short- and long-term structural performance of
corrosion-damaged RC beams strengthened with FRCM composites. This thesis is organized in
nine chapters as follows:
• Chapter (1) provides background on the subject as well as the definition of the research
problem.
• Chapter (2) provides a comprehensive literature review on topics that are related to the work
of this study. This includes a review on the corrosion mechanism in RC structures, a review on
the flexural behavior of corrosion-damaged RC members under both monotonic and fatigue
loads, a full historical review on previous research that was carried out to determine the
mechanical properties of FRCM composites, and an inclusive review on the flexural behavior
of RC members strengthened with FRCM composites.
• Chapter (3) highlights the research significance along with the research objectives of this work.
A detailed description of the utilized methodology to achieve these objectives is also included.
• Chapter (4) presents the first journal paper submitted to the American Society of Civil
Engineers (ASCE) Journal of Composites for Construction. The paper is titled “Corrosion-
Damaged Reinforced Concrete Beams Repaired with Fabric-Reinforced Cementitious Matrix
(FRCM)”. The paper reports on the feasibility of using FRCM to strengthen/repair RC beams
with moderate level of corrosion damage. The test results of eleven large-scale RC beams were
presented including those of nine beams that were subjected to accelerated corrosion process
for 70 days to obtain an average mass loss of 12.6% in the tensile steel reinforcing bars and
4
those of two beams that were tested as controls. The effect of various levels of corrosion damage
on the flexural response of the FRCM-strengthened beams was investigated throughout Chapter
5. In this paper, the test parameters included the number of fabric plies (1, 2, 3, and 4), the
FRCM repair scheme (end-anchored and continuous U-wrapped strips), and FRCM materials
(carbon and PBO). The analysis and discussion of the test results were presented in terms of
mode of failure, load-deflection response, load carrying capacity, ductility performance, and
strain responses. In addition, the experimental results were compared to the current design
guidelines of ACI-549.4R-13.
• Chapter (5) presents the second paper that was published in the Journal of Composite
Structures. The paper is titled “Effect of Corrosion Damage on the Flexural Performance of RC
Beams Strengthened with FRCM Composites.” The paper reports on the effect of the level of
corrosion damage on the flexural response of RC beams strengthened with FRCM composites.
Three theoretical tensile steel mass losses were considered in this study, namely 10, 20, and 30
%, which represented moderate, severe, and very severe degrees of corrosion damage. The test
parameters also included the fabric type (PBO and carbon), the number of FRCM layers (two,
three, and four), and the strengthening scheme (end-anchored and continuously wrapped). The
test results were analyzed and discussed to highlight the effect of various test variables on the
failure mechanism, the flexural performance, the fiber, steel, and concrete strain responses, the
ductility, and the flexural strengths of the tested beams. In addition, the experimental results
were compared to the current design guidelines of ACI-549.4R-13.
• Chapter (6) presents the third journal paper that was submitted to the Journal of Construction
and Building Materials and was titled “Post-repair Flexural Performance of Corroded Beams
Rehabilitated with Fabric-Reinforced Cementitious Matrix (FRCM) under Corrosive
Environment”. The test results of nine RC beams were reported in this paper including those of
one beam that were neither corroded nor repaired, another beam that was corroded and not
repaired, and six beams that were corroded and then repaired in two phases. Beams of Phase I
(short-term) were subjected to an accelerated corrosion process for 210 days before being
strengthened with FRCM whereas beams of Phase II (long-term) were initially subjected to
accelerated corrosion for 70 days, then repaired with FRCM and exposed to further corrosion
5
for 140 days. The test parameters also included the type of FRCM materials (PBO and Carbon)
and the FRCM repair scheme (end-anchored and continuous wrapping).
• Chapter (7) presents the fourth journal paper that was submitted to the ASCE’s Journal of
Composites for Construction. The paper is titled “Fatigue and Monotonic Behavior of
Corrosion-damaged Reinforced Concrete Beams Strengthened with FRCM Composites.” This
paper aimed at investigating the potential of using FRCM composites to restore the fatigue life
of RC beams that were severely damaged due to corrosion (20% tensile steel mass loss). The
paper provides better understanding on the flexural behavior of FRCM-strengthened beams
damaged due to corrosion. The test results of twelve beams were reported. The test parameters
included the fabric material (PBO and Carbon), the number of FRCM plies, the strengthening
configuration, and the type of loading (monotonic and fatigue). The results were discussed and
presented in terms of fatigue life, fatigue behavior, modes of failure, strain response, and
stiffness degradation.
• Chapter (8) presents the fifth paper that was submitted to the Journal of Engineering Structures
and titled “Finite Element Modeling and Experimental Results of Corroded Concrete Beams
Strengthened with Externally-bonded Composites”. This paper presents the numerical
simulation of corroded RC beams strengthened with carbon-FRP (CFRP) and PBO-FRCM
composites under flexural loading using ATENA software package. The CFRP composite was
modeled as discrete reinforcement bonded directly to the concrete substrate without binder
while the PBO-FRCM was modeled using a more detailed approach that involved modeling the
fabric and the matrix layers. Interfacial bond stress-slip models were adopted at the
CFRP/concrete and the PBO-fabric/matrix interfaces to simulate the failure mechanism
observed during the test. The load-carrying capacities, load-deflection responses, and load-
strains responses were evaluated and compared with the experimental results to validate the
accuracy of the model. The validated models were used in a parametric study to investigate the
effect of varying the concrete compressive strengths on the flexural behavior of the FRCM-
strengthened beams after corrosion.
• Chapter (9) includes a summary of this study and the overall conclusions based on the
experimental, analytical, and numerical results obtained. Recommendations for future work are
also suggested in this chapter.
6
2. Chapter 2
Literature Review
2.1 General
The corrosion of the steel reinforcement in concrete structures is a major challenge facing
structural engineers while maintaining ageing infrastructures. Corrosion usually causes
engineering and economic problems, which has led to a considerable amount of research work
devoted to study the performance of corroded-RC flexural elements under both monotonic and
fatigue loads. In recent years, advanced composite materials in the form of externally-bonded
fabric-reinforced cementitious matrix (FRCM) have been introduced as innovative strengthening
techniques for RC structures. Several publications have been published in recent years to document
the effectiveness of FRCM composites to strengthen undamaged concrete structures. Nevertheless,
their application to strengthen/repair corrosion-damaged concrete elements is still lacking.
This chapter provides a comprehensive review on the corrosion of steel reinforcement in concrete
structures as well as the use of externally-bonded FRCM composites to strengthen RC flexural
elements. A review of the available literature concerning the research work on corroded FRCM-
strengthened flexural elements is presented along with the factors that affect their structural
behavior.
2.2 Corrosion of Steel Reinforcement
2.2.1 Corrosion Mechanism of Steel Bars in Concrete
The high alkaline environment of concrete normally protects the steel reinforcement against
corrosion by creating a passive film of iron oxides at the steel/concrete interface [5]. In addition,
the concrete cover works as a barrier to protect the steel bars against harsh environmental
conditions [5]. Exposure to sea salt spray in marine environment or de-icing salts in cold regions
are the most common causes of corrosion in concrete worldwide. Steel corrosion in concrete can
be triggered by chloride attacks and concrete carbonation. Chloride ions diffuse through the
concrete pores and cracks until they reach the surface of the steel bars. Chlorides work as catalysts
to initiate the corrosion reactions in the weak spots on the steel surface. In case of concrete
7
carbonation, the carbon dioxide gas (CO2) in the atmosphere dissolves in water (H2O) to form the
carbonic acid (H2CO3). The acid then penetrates into the concrete and drops its pH level to less
than 8.5. Consequently, the passive layer previously formed on the bar surface decades and
corrosion reactions are then initiated [5,17].
Corrosion of steel in concrete is an electrochemical process. It involves the transfer of charge
(electrons) from one location to another [17]. The corrosion reaction is similar to a galvanic cell
that consists of an anode and a cathode as shown in Figure 2.1. At the anode, the iron (steel bar)
dissolves or oxides and the ferrous electrons are released as illustrated in Equation (2.1). The
released electrons react with water and oxygen at any other location of the steel bar that acts as a
cathode to form dissolved hydroxide ions (Equation (2.2)).
Fe Fe+2 + 2e− (Anode reaction) Eq. (2.1)
2e−+H2O +1
2O2 2OH− (Cathode reaction) Eq. (2.2)
The ferrous ions (Fe+2) at the anode react with the hydroxide ions (OH−) that are released from
the cathode to form ferrous hydroxide [Fe (OH)2] as per Equation (2.3). In the presence of
sufficient amount of water and oxygen, the ferrous hydroxide reacts to produce ferric hydroxide
[Fe (OH)3] as shown in Equation (2.4). Finally, the ferric hydroxide [Fe(OH)3] decays to hydrated
ferric oxide [Fe2O3. H2O] or rust [5]. This reaction is shown in Equation (2.5). The volume of the
corrosion products is much greater than the volume of steel (volume of rust is about 6 times of that
of iron). Therefore, longitudinal cracks parallel to the reinforcing bars generally appear in the
concrete cover as indication of corrosion. The cracks allow to more quantities of moisture and
oxygen to reach the steel bars, which boosts the rate of corrosion reactions.
Fe+2+2OH− Fe (OH)2 Eq. (2.3)
4Fe (OH)2+ O2 + 2H2O 4Fe (OH)3 Eq. (2.4)
2Fe (OH)3 Fe2O3. H2O + 2H2O Eq. (2.5)
8
Figure 2.1: Corrosion process of steel bars inside concrete [5]
Corrosion of steel in concrete can occur as micro-cells (uniform iron removal) or as macro-cells
(local iron removal). Figures 2.2a and 2.2b show schematics of micro-cell and macro-cell types of
corrosion, respectively. At the beginning, corrosion pits start to take place before they expand and
propagate and uniform corrosion can be seen along the steel bars. The pits are usually initiated by
the chloride attacks at the weak spots on the bar’s surface, which creates an electrochemical
potential difference between the anodes and cathodes. In micro-cells, both the cathodes and anodes
are very close as shown in Figure 2.2a. Macro-cell reactions can occur over large distances on the
surface of the steel bar or even between two different bars. In macro-cells, the typical pits are
separated by large passive area of steel that acts as small concentered anodes (Eq. (2.1)) that are
surrounded by large cathodes as shown in Figure 2.2b (Eq. (2.2)).
Figure 2.2: A schematic of microcell and macrocell types of corrosion [5]
a) Microcell b) Macrocell
9
2.2.2 Accelerated Corrosion Process
A significant amount of experimental work has been carried out to investigate the influence of
corrosion of steel reinforcement on the structural performance and durability of concrete structures
[3,18,19]. Most researchers used accelerated corrosion techniques to achieve the desired level of
corrosion damage within reasonable amount of time. Table 2.1 summarizes some of previous
accelerated corrosion tests that have been conducted.
Accelerated corrosion can be used to simulate natural corrosion without remarkable change in
the structural response that would be encountered due to natural corrosion [20–22]. The corrosion
process is accelerated by electrical polarization of the steel reinforcement. Chloride salt is usually
used to activate the corrosion process. Some researchers added chloride salts to the concrete mix
[10,20,23] while others partially submerged the test specimens in a salted solution [18,22,24,25].
During accelerated corrosion, the steel bar is connected to the positive terminal of a power supply
to enforce the bar to act as anode and consequently Fe2+ ions start to dissipate from the surface of
the bar. Embedded stainless steel bars or tubes or external plates made of copper or stainless steel
are then connected to the negative terminal of the power supply to act as cathode. Both methods
are presented in Table 2.1.
The level of corrosion-damage of steel bars can be quantified as the percentage of the mass of
the steel bars lost to rust. The steel mass loss due to corrosion can be measured experimentally or
predicted theoretically. Measuring the mass of lost metal, after the completion of the tests, is the
most widely adopted measure of corrosion levels in laboratory tests. The actual mass loss can be
determined according to ASTM G1-90 [26] standard, which provides mechanical, chemical, and
electrolytic techniques to remove corrosion from the steel bars.
On the other side, the mass loss of steel bars can be predicted theoretically using Faraday’s law
[18,22,24,27]. This law relates the mass loss to the corrosion current and to the time of exposure
to corrosion as follows:
m =I t a
n F Eq. (2.6)
where m = mass loss (in grams); I = intensity of the impressed current (in Ampere); t = the time
of corrosion (in seconds); a = the atomic mass of iron (55.847gm); n = the number of electrons
10
transferred during the corrosion reaction (n = 2 in case of iron); and F = Faraday’s constant, which
equals to 96 500 C/equivalent.
For practical application, the current density, i, instead of the current intensity, I, as follows:
m =(S×𝑖 ) t a
n F Eq. (2.7)
where S is the surface area of the corroded steel and i is the impressed current density.
Table 2.1: Summary of some previous accelerated corrosion tests
Study Specimen type Current density
(μA/cm2) Cathode type Corrosion environment
Almusallam et al.
(1996) [25] Bond pull-out 10400 External stainless-
steel plate
Constant immersion in
3% NaCl solution
Alonso et al. (1998)
[28] Bond pull-out 3, 10, or 100 External stainless-
steel plate
Concrete cast with 3%
NaCl by weight of
cement
Mangat and Elgarf
(1999) [24] Beams 1000, 2000, 3000,
or 4000
External copper
plate
Constant immersion
3.5% NaCl solution
Masoud et al.
(2001) [10] Beams 150
Internal stainless-
steel bar
Concrete cast with 3%
NaCl by weight of
cement
El-Maaddawy and
Soudki (2003) [20] Concrete prisms
100, 200, 350, or
500
Internal stainless-
steel bar
Concrete cast with 5%
NaCl by weight of
cement
Malumbela et al.
(2010) [21] Beams 189 External stainless-
steel bars
4 days immersion in 5%
NaCl solution followed
by 2 days drying
Mancini et al.
(2014) [23] Concrete prisms 200 External stainless-
steel plate
Concrete cast with 5%
NaCl by weight of
cement
Ou and Nguyen
(2016) [22] Beams 600 External copper
plate
Constant immersion
5% NaCl solution
In natural corrosion, the measured corrosion current density is in the range of 1 to 100 μA/cm2
[28–31] while the impressed current density used in accelerated corrosion ranges between 3 and
10400 μA/cm2 [25,30]. The density of the impressed current should be kept constant during the
accelerated corrosion to ensure uniform corrosion mechanism [20]. Much research has been
carried out to investigate the optimal current density that should be used in accelerated corrosion
techniques to simulate natural corrosion within reasonable amount of time [24,27,28]. It has been
reported that the use of a current density higher than 250 μA/cm2 in accelerated corrosion increases
11
the corrosion cracks width [32], decreases the bond strength [18,27], and may affect the
mechanical properties of the steel bars [19,22] in comparison to those measured during natural
corrosion at similar mass losses. These observations were attributed to the change of the
composition of the corrosion products due to the use of high current densities [27]. El-Maaddawy
and Soudki (2003) [20] concluded that varying the current density level between 100 and 200
μA/cm2 had no impact on the concrete strain response. However, increasing the current density
above 200 μA/cm2 caused a significant increase in the concrete strain response and the widths of
corrosion cracks due to higher concentrations of corrosion products around the steel bars. In
addition, it was concluded that the use of current densities less than 200 μA/cm2 would be small
enough to obtain corrosion-damage similar to that reported in the field.
2.2.3 Effect of Steel Corrosion on Concrete Structures
Corrosion of steel reinforcement in concrete structures damages the reinforcement itself, the
surrounding concrete, and ruins the composite action between the steel bars and concrete.
Corrosion significantly affects the tensile behavior of the steel bars. The external surfaces of the
corroded bars significantly change due to the formation of irregular pits. Corrosion alters the shape
of the bar cross-section, which varies randomly along the corroded length. It also reduces the cross-
sectional area of the bar and consequently its load-carrying capacity. The non-uniform distribution
of corrosion pits cause stress concentration, which reduces the residual strength and ductility of
the corroded bars compared to that of the uncorroded ones having the same cross section [33,34].
The level of corrosion-damage (i.e. the mass loss of the bars) and the distribution of pits along the
length of the corroded bars significantly influence their stress-strain response in both tension and
compression [33]. For the same corrosion mass loss, corrosion has a more significant effect on the
tensile behavior of plain bars than that of deformed bars and has a more pronounced impact on the
smaller bars than on the larger bars [2].
In an interesting study by Ou et al [22], the tensile behavior of naturally-corroded and artificially-
corroded steel bars were compared. The current density used was 600 μA/cm2. The test results
indicated that the tensile strength and ductility of both naturally- and artificially-corroded bars
were decreased by increasing the level of corrosion damage. The yield and ultimate tensile
strengths of artificially-corroded bars were similar to those of naturally-corroded bars. However,
the ultimate strains of the naturally-corroded bars were smaller than those of the artificially-
12
corroded ones. These observations indicated that corrosion pits were more uniformly distributed
in artificially-corroded bars than in naturally-corroded ones.
Corrosion damage has a more pronounced impact on the fatigue behavior of the corroded steel
bars rather than their tensile behavior [35,36]. Corrosion of steel bars significantly decrease their
fatigue life under tensile and compressive cyclic loads [33]. In addition, corroded bars fail in a
very brittle manner by sudden fatigue rupture. The fatigue life of corroded bars decreases
dramatically with the increase of corrosion level and the fatigue stress range [19].
On the other side, the expansion of the corrosion products results in expansive stresses in the
concrete surrounding the corroded bars. Once these stresses exceed the concrete tensile strength,
cracking and spalling of concrete cover occur. Moreover, corrosion products work as an isolation
layer at the steel/concrete interface, which deteriorates bond and loosens the composite action
between steel and concrete [37]. When deformed bars corrode, the height of their lugs decreases,
and rust fills the inner cracks of the deteriorated concrete. Therefore, their bond performance
become similar to that of plain bars, where just friction is the dominant parameter [38,39]. In
advanced stages of corrosion, the cracked concrete can peel off, and bond between the two
materials can be totally lost. Many researchers have investigated the effect of corrosion on the
bond characteristics at concrete/steel interface. Most studies have found that bond increases
slightly as the steel bar corrodes and decreases rapidly as the corrosion cracks take place [40]. It is
important to note that the reduction in bond are dependent of the thickness of the concrete cover,
the bar diameter, and the presence of stirrups (confinement) [41,42].
2.2.4 Behavior of Corroded-RC Beam under Monotonic Loads
Much research work has been devoted to study the structural behavior of corroded RC beams
under monotonic loads. In the following, a summary of some of the previous work is presented.
In an early study by Almusallam et al. (1996) [43], the structural behavior of RC elements with
different levels of corrosion damage was investigated. The test specimen was 63.5×305×711 mm
and reinforced by 5 steel bars of 6 mm diameter. The whole length of the steel reinforcement was
subjected to an accelerated corrosion process. A sharp reduction in the ultimate strength was
observed as the degree of corrosion increased. For instance, 10 and 28% mass loss resulted in a
reduction in the ultimate strengths of the specimens by 20 and 65%, respectively. The corroded
13
specimens failed by bond-shear failure due to the loss of bond between steel and concrete. Thus,
the ductility of the corroded specimens was significantly reduced.
Vidal et al. (2007) [44] conducted a long-term experimental program to study the structural
behavior of 36 beams of 150×280×3000 mm. The beams were reinforced with two steel bars of 12
or 16 mm diameter. The beams were stored in a chloride environment under service loads for 14
and 17 years to corrode naturally. The test results showed that the steel mass loss was 15 and 30%
after 14 and 17 years respectively. In addition, the flexural cracks during the corrosion exposure
had no significant impact on the rate of corrosion. The inspection of the corroded bars indicated
that neither the distribution of the corrosion pits nor their intensity was uniform along the steel
bars despite the fact that all specimens were subjected to uniform environmental conditions. The
flexural strength and stiffness of the corroded beams were significantly influenced by the level of
corrosion damage.
Torres-Acosta et al. (2007) [45] investigated the behavior of twelve simply-supported RC beams
(100×150×500 mm) with several levels of corrosion damage. The whole length of the steel bars
was subjected to an accelerating corrosion process. Longitudinal cracks parallel to the steel bars
were observed at the end of the corrosion process. Their width increased as the corrosion level
increased. All of the corroded beams showed a more brittle response as the degree of corrosion
damage became higher. The residual flexural strengths also decreased at higher corrosion levels
and were significantly influenced by the depths of the pits on the surface of steel bars.
Gu et al. (2010) [46] investigated the flexural behavior of naturally- and artificially-corroded RC
beams. The naturally-corroded beams with steel mass loss up to 15% failed due to flexural while
increasing the mass loss to 20% changed their mode of failure to rupture of steel bars. On the other
hand, the beams subjected to an accelerated corrosion process up to 25% mass loss failed due to
flexural whereas those subjected to 40% mass loss failed due to rupture of steel bars. This
observation was attributed to the non-uniform characteristics of the corrosion damage and the
significant change in the mechanical properties of naturally-corroded steel bars. The load-carrying
capacities of the corroded beams were reduced due to the reduction in cross-sectional area of the
steel bars and the degradation in their mechanical properties due to corrosion while the decline in
the beams’ stiffness was mainly attributed to the loss of bond at the steel/concrete interface.
14
Malumbela et al. (2010) [47] conducted an experimental study to investigate the influence of
varying the steel mass loss on the ultimate flexural capacity of corroded RC beams. The test
specimen had dimensions of 153×254×3000 mm and were reinforced with three deformed steel
bars of 12 mm diameter. The middle 700 mm of the steel bars were subjected to an accelerated
corrosion process. The results showed that the maximum mass loss was obtained in the middle of
the corroded zone with less mass loss reported at the two edges of the corrosion zone. The use of
a sustained load during the corrosion stage had no significant effect on the rate of corrosion. The
ultimate flexural capacity of the tested beams decreased linearly with the measured steel mass loss.
The ultimate capacity of the beams decreased at a rate of 0.70% per each 1% steel mass loss.
Xia et al. (2012) [48] conducted an experimental study on the effect of corrosion of steel
reinforcement on the stiffness and flexural strength of RC beams. The beams were tested under
four-point load configuration. The tensile reinforcement was corroded within the maximum
moment zone only to provide sufficient anchored length to avoid debonding failure. Nine levels
of corrosion damage (4 to 12% steel mass loss) were achieved using an accelerated corrosion
process. At the end of the corrosion process, longitudinal cracks parallel to the steel bars were
observed. The average and maximum crack widths were in the range of 0.32 to 2.38 mm, and 0.68
to 4.5 mm, respectively. The average and the maximum crack widths increased as the degree of
corrosion increased. The test results showed that the corroded and uncorroded beams had the same
stiffness until 60% of the ultimate load after which the stiffness of the corroded beams significantly
decreased. The residual flexural strength of the corroded specimens decreased as the level of
corrosion increased. The corrosion level changed the failure mode from under-reinforced ductile
failure to brittle collapse due to the rupture of the steel bars at the highest degree of corrosion
damage (12% tensile steel mass loss).
Dang and François (2014) [49] studied the structural behavior of RC beams (280×150x3000) mm
naturally-corroded under harsh environmental conditions over several years. The test observations
revealed that the non-uniform distribution of the corrosion intensity due to the random formation
of pits on the surface of the corroded bars not only affected the flexural strength of the beams but
also changed their ductile failure (steel yielding followed by concrete crushing) to brittle rupture
of the steel bars. The reduction in ultimate capacities corresponding to a given steel mass loss was
lower than the reduction in ultimate deflections. Therefore, it was concluded that the service life
15
of the corroded structure may be limited by the reduction of ductility rather than the reduction of
the load-bearing capacity. This observation was related to the significant change in the mechanical
properties of the corroded steel bars.
2.2.5 Behavior of Corroded-RC Beam under Fatigue Loads
The fatigue life of concrete structures is usually governed by the endurance of the steel
reinforcing bars under repetitive loads and is rarely controlled by concrete [50–52]. Corrosion
damage significantly decreases the fatigue endurance of the steel bars, which significantly reduces
the fatigue life of RC beams. Many researchers investigated the effect of corrosion on the fatigue
performance of RC beams. A summary of some of these investigations is presented as follow:
Yi et al. (2010) [53] carried out a laboratory study to investigate the performance of RC beams
with corroded reinforcement under fatigue loading. The test specimens (150×300×3600 mm) were
reinforced with two deformed steel bars of 20 mm diameter each. The beams were subjected to
accelerated corrosion process to achieve eight degrees of corrosion damage in the range of 3.25 to
11.6 % steel mass loss. All of the beams were subjected to the same level of fatigue load oscillating
between 11 and 52% of the load-carrying capacity of the control (virgin) beam. The test results
revealed that corrosion of the steel reinforcement significantly reduced the fatigue life of the
beams. The fatigue life decreased as the level of corrosion damage increased. For example, the
non-corroded beams safely sustained 2 million fatigue cycles while beams with 3.25 and 11.6%
steel mass loss failed after 626,000 and 89,000 cycles, respectively. In addition, all of the corroded
beams failed due to the sudden rupture of one of the steel bars. The rupture occurred at locations
of corrosion pits where maximum mass losses were encountered.
The performance of corroded RC beams under repeated loads were investigated by Sun et al.
(2015) [6]. The test results showed that corrosion damage of the steel reinforcement had a
significant adverse effect on the fatigue life of the beams. All of the corroded beams failed by
brittle fatigue fracture of one of the steel bars. The fatigue performance of the corroded RC beams
was described by two stages. At the first stage, flexural cracks occurred accompanied by rapid
increase in the mid-span deflection of the beams and reduction in their flexural stiffness. The
second stage represented the largest part of the fatigue life of the beams. In this stage, deflections
and stiffness of the beams became more stable until sudden rupture of the corroded bars occurred.
16
Recently, Zhang et al. (2017) [7] investigated the effect of corrosion on the fatigue behavior of
seven RC beams (120×200×1500 mm). Six out of the seven beams were subjected to an
accelerated corrosion process to achieve different levels of corrosion damage in their tensile steel
reinforcement (2.8 to 15.4 % steel mass loss). The corroded beams were subjected to fatigue with
lower and upper limits of 10 and 60% of the ultimate flexural capacity of the virgin beam,
respectively. The test results showed that the fatigue life of the corroded beams decreased rapidly
with the increase of the level of corrosion damage. Strains in both concrete and steel increased
with the increase of the fatigue life due to the accumulated fatigue damage. The midspan
deflections of the beams increased rapidly at approximately 5% of the fatigue life followed by a
stable stage until sudden failure of the beams occurred. The beams failed by sudden fatigue rupture
of the steel bars. The microscopic scan of the fracture surface of the bars showed that fracture
consisted of three zones namely, the crack initiation region, the crack growth region, and the abrupt
rupture zone.
2.3 FRCM Composites
Fabric-reinforced cementitious matrix (FRCM) system is a composite material consisting of
fabric meshes made of dry fibers embedded in cement-based matrix (serving as binder) as shown
in Figure 2.3. When adhered to concrete or masonry, the FRCM system acts as an externally-
bonded reinforcement. FRCM is considered as the natural evolution of ferrocement and FRP
composites. FRCM have recently emerged as a viable technology for repairing and strengthening
RC and masonry structures [54]. FRCM has been mentioned in the literature review with many
acronyms such as textile-reinforced mortar (TRM) [55–57], textile-reinforced concrete (TRC)
[58], and mineral-based composites (MBC) [59].
17
Figure 2.3: FRCM composite system
The fabric is a manufactured planar textile structure made of fibers, yarns, or both, which is
assembled by various means such as weaving, knitting, tufting, felting, braiding, or bonding of
webs to give the structure sufficient strength and other properties required for its intended use [54].
Figure 2.4 illustrates different fabric configurations. The fabric can be made of high performance
fiber such as carbon, Polyparaphenylene benzobisoxazole (PBO), alkali-resistant glass (AR-glass),
and basalt. Many researchers reported the significant influence of fabric material, geometry, and
configuration on the mechanical properties of the FRCM composite [54,60–63]. The penetration
of the cementitious matrix to the fabric is a fundamental factor for bond between the fabric and
the matrix to develop and consequently for the mechanical properties of FRCM composites.
Therefore, the fabrics consist of meshes with total coverage area less than 2/3 of the total area of
the fabric to provide sufficient matrix penetration [54,64].
Figure 2.4: Different fabric configurations [64]
18
2.3.1 FRCM Acceptance Criteria and Design
FRCM composites should not be produced randomly by selecting and mixing available
commercial materials. Therefore, the developed FRCM systems must go through various tests and
meet certain criteria to evaluate the mechanical properties, bond strength, durability, and fire
resistance of the composite to be accepted for structural applications. The ICC Evaluation Services
(ICC-ES) proposed guidelines for the test methods to evaluate the properties of FRCM composites
as well as their acceptance criteria [65] . The International Building Code (IBC) (Section 104.11.1)
[66] requires a product report that evaluates the characteristics of the FRCM system according to
AC434 guidelines [65] to be approved for strengthening and repair applications.
The American Concrete Institute (ACI) has developed design guidelines for the structural
applications of FRCM composites (ACI-549.4R, 2013) [54]. This document provides the
necessary tools for the design and use of FRCM systems. It covers background information, FRCM
material properties, and design guidelines for axial, flexural, and shear strengthening applications.
It is important to note that the design methodology provided ACI-549.4R was examined and
validated through the analytical investigation of this thesis.
2.3.2 FRCM Tensile Characterization
FRCM composites are typically applied to the tension face of RC members to work as
supplemental tensile reinforcement. Thus, characterizing their tensile behavior is a fundamental
parameter for design. Much research was carried out to investigate the behavior of FRCM coupons
under tensile loading [67–70]. The results of some of these investigations can be summarised as
follows:
The behavior of FRCM coupons under tensile load included three distinct stages namely,
uncracked stage (State I), cracking stage (State IIa), and crack growth stage (State IIb) [67,70] as
shown in Figure 2.5. The actual and idealized stress-strain response of an FRCM coupon under
tensile load is also shown Figure 2.5. The first stage represents the uncracked elastic behavior of
FRCM in which the stiffness of the system is approximately equal to that of the cementitious
matrix. In the second stage, the initiation and the formation of cracks within the matrix result in a
significant loss of stiffness. The length of this part of stress-strain response is mainly dependent
on the bond quality at the fabric/matrix interface and the mount of fibers used. In the third stage
19
(State IIb), the matrix cracks continue to grow and widen and the applied load is fully carried by
the fabric until failure takes place due to their rupture or slippage within the matrix. It is important
to note that the tensile behavior of FRCM owns the advantage of pseudo-ductility due to the
combination between micro-cracking and cracking mechanisms [68,69].
Figure 2.5: Actual and idealized stress-strain curve of FRCM coupon in tension [67]
It can be depicted from the actual stress-strain curve (Figure 2.5) that the slope of the cracking
stage (State IIa) is similar to that of the crack growth stage (State IIb). The transition from the
uncracked stage to the cracking stage is called the bend-over-point (BOP) or the transition point
[67,70]. Therefore, the actual tensile stress-strain curve of FRCM coupon could be idealized to a
simple bi-linear curve with a bend-over point [67,70] as shown in Figure 2.5. The tensile behavior
of FRCM systems can be characterized by the following parameters: the tensile modulus of
elasticity of the uncracked specimen, 𝐸𝑓∗; the tensile modulus of elasticity of the cracked specimen,
𝐸𝑓; the ultimate tensile strain, ԑ𝑓𝑢; the tensile strain corresponding to the transition point, ԑ𝑓𝑡; the
ultimate tensile strength, 𝑓𝑓𝑢, and the tensile stress corresponding to the transition point, 𝑓𝑓𝑡 [54].
Table 2.2 gives the characteristics of PBO-FRCM and C-FRCM coupons that were used in this
study.
20
Table 2.2: Tensile characteristics of PBO-FRCM and C-FRCM coupons [65]
FRCM property Symbol
PBO-FRCM Carbon-FRCM
Mean STD Mean STD
Uncracked tensile modulus of
elasticity, GPa
𝐸𝑓∗ 216 65 74 19
Cracked tensile modulus of elasticity,
GPa
𝐸𝑓 18 2 12 3
Tensile stress corresponding to the
transition point, MPa 𝑓𝑓𝑡 54 12 66 7
Tensile strain corresponding to the
transition point, %
ԑ𝑓𝑡
0.0172 0.0044 0.1020 0.0449
Ultimate tensile stress, MPa 𝑓𝑓𝑢
241 11 150 8
Ultimate tensile strain, % ԑ𝑓𝑢
1.7565 0.1338 1.000 0.1405
Note: Coupon tested with 6 in (150 mm) long tabs.
STD: Standard Deviation
2.3.3 Bond Behavior of FRCM Composites
Understanding the stress transfer mechanisms in FRCM/concrete joints is essential to maximize
their potential as strengthening materials. The fact that FRCM composites consists of dry-fiber
fabric embedded in a mortar prevents the matrix from full impregnating and bonding to each fiber.
Therefore, the weakness of FRCM/concrete joints usually occurs at the fabric/matrix interface
rather than at the matrix concrete interface [71,72] as is typically observed in FRP composites [73].
Large slips between the fabric and the matrix are usually observed during the debonding between
the FRCM layers and the concrete substrate. The debonding mechanism is complicated because of
the so-called telescopic failure that occurs due to the differential slip between the external fibers
and the inner fibers, which makes the composite vulnerable to slipping and/or debonding [69].
The failure mechanisms at the fabric/matrix interface can be modeled as a fiber bundle containing
sleeve and core filaments as shown in Figure 2.6 [74] The bond between the external fibers (the
sleeve fibers) and the matrix as well as the frictional bond between the internal fibers (the core
fibers) govern the bond behavior of FRCM composites and constantly control their strengthening
contribution. The failure at the fabric/matrix interface occurs in three different mechanisms: a)
21
debonding between the external fibers and the matrix, b) slippage between the internal core fibers,
or c) a combination of the two mechanisms [70,74].
Figure 2.6: Failure mechanism of a fiber bundle embedded in cementitious matrix [74]
There are some other factors that affect the bond behavior of FRCM composites. Ombers [75]
reported that increasing the number of FRCM layers changed their mode of failure from fiber
slippage to delamination at the fabric/matrix interface. It was reported that the fabric strain
decreased when debonding occurred. These observations were also consistent with the findings of
D’Amberisi et al. [76]. The bond capacity of FRCM composites increased with an increase in the
bond length. The effective bond length was reported to be in the range of 150-330 mm [76–78].
Nevertheless, varying the width of FRCM composites have no impact on their bond capacity [77].
This was attributed to the independent behavior of the longitudinal fiber bundles as mentioned
earlier. Arboleda et al. (2014) [79] investigated the bond and tensile behavior of carbon and PBO
FRCM composites after exposure to freeze/thaw cycles, high temperature, water vapor, and
immersion in sea water. The test results indicated that the exposure to these hash environments
had no significant impact on the bond and the tensile behaviors of FRCM composites.
2.4 FRCM-Strengthened Beams Under Monotonic Load
Many studies have been conducted to investigate the flexural behavior of undamaged RC beams
strengthened with different FRCM composite systems. Nevertheless, only one study investigated
the feasibly of using FRCM to strengthen corrosion-damaged RC beams. Some of the previous
studies are summarised in the following sections:
Brückner et al. (2006) [58] conducted an experimental work to investigate the feasibility of using
FRCM as externally bonded strengthening technique. RC slabs of 100 mm depth and 1600 mm
22
span were tested under four-point load configuration. Different numbers of FRCM layers were
applied. Improvements in the load carrying capacity, shear loading, and ductility were observed
by increasing the number of FRCM layers used.
Täljsten and Blanksvärd (2007) [59] investigated the behavior of RC slabs strengthened with C-
FRCM and C-FRP systems. Six specimens of 100 mm depth, 1000 mm width, and 4000 mm length
were tested until failure under four-point load configuration. The test results showed the
effectiveness of using FRCM composites as a strengthening technique not only to improve the
ultimate strength of the tested slabs but also to enhance their ductility in comparison to those
strengthened with FRP systems. In addition, small RC beams were tested in bending to investigate
the effect of using different types of cementitious matrices on the bond and ultimate strengths of
the specimens. It was reported that using polymer-modified cementitious mortars enhanced the
bond between the fabric and the matrix as well as the bond at the FRCM/concrete interface, which
caused an increase in the load-carrying capacities of the beams.
Ombres (2011) [80] studied the structural behavior of RC beams strengthened with PBO-FRCM.
Twelve RC beams of 150×250×3000 mm were tested under four-point load configuration up to
failure. The test results showed the efficiency of the PBO-FRCM in increasing both the yielding
and ultimate strengths of the beams. The flexural capacities of the tested beams increased between
10% and 44% in comparison to those of the control beams depending on the amount of FRCM
used. The number of applied plies of FRCM significantly changed the mode of failure from
concrete crushing in case of 1 ply of FRCM to delamination at the concrete/FRCM interface when
2 and 3 plies of FRCM were used. The assumption of strain compatibility between the fabric and
the mortar was verified until 70 to 80% of the failure load, while significant fabric slip in the
mortar was observed at failure. The ductility of the strengthened specimens with 1 ply of FRCM
increased (ductile failure was reported) while it decreased when large number of plies were used
(brittle failure was reported).
D’Ambrisi and Focacci (2011) tested RC beams of 250 mm depth and 400 mm width with
different span lengths (2200 mm and 1600 mm) [81]. The short beams were tested under three-
point bending configuration while the long beams were tested under four-point load configuration.
The effects of using different fiber materials (PBO-FRCM, C-FRCM and CFRP), two different
cement-based matrices (M50 and M75), and three different strengthening configurations
23
(continuous U-shaped strip, U-shaped strip only at the beam ends, and no wrapping) on the
structural performance of the beams were investigated. The experimental results showed that:
a) Using 2 plies of PBO-FRCM with axial stiffness, Af Efu, of about 83% of that of 1 ply of
C-FRP improved the ultimate capacities of the strengthened beams by 30% of that of the
unstrengthened beam (all beams had the same strengthening scheme).
b) Using transverse strengthening changed the mode of failure from sudden detachment of
the strengthening material from concrete to debonding of fibers from the matrix.
Consequently, enhancements in the ultimate capacity and ductility were observed.
Specimens with continuous U-shaped strips showed 18% higher load-carrying capacities
than those with no wrapping.
c) The M50 matrix made of pozzolanic cement, selected silica aggregates, polycarboxylic
water-reducing admixtures, and organic adhesion promoter showed less bond strength
between the fiber and the matrix and between the matrix and concrete than that of the M750
matrix, which was made of composite high-fineness cement binder, adhesion promoter,
inorganic nanoparticles, micro-aggregates, and a new-generation of high-effectiveness
polycarboxylic water-reducing admixtures.
Schladitz et al. (2012) [55] reported on the structural behavior of large scale RC slabs of 230 mm
depth and 6750 mm span strengthened with different number of layers of C-FRCM systems. All
specimens were subjected to four-point load configuration until failure. The test results showed a
significant increase in the ultimate loads as the number of layers increased. It was reported that
using 1 layer and 4 layers of FRCM increased the ultimate loads by 67% and 245%, respectively.
Fabric rupture was observed at failure in all of the strengthened specimens even when high amount
of FRCM layers was used. Significant decrease in deflection was reported when high amount of
FRCM was used.
Si Larbi et al. (2012) [82] conducted an experimental and analytical study to investigate the
effectiveness of alkali-resistant AR glass-FRCM as a strengthening system for RC structures. Five
beams were strengthened and tested under four-point load configuration. The results showed that
FRCM increased the ductility of the strengthened specimens in comparison to the FRP-
24
strengthened ones. On the other hand, using AR glass-FRCM and C-FRP increased the ultimate
capacity by 30% and 80%, respectively.
Elsanadedy et al. (2013) [83] conducted an experimental and numerical investigation to study
the structural behavior of RC beams strengthened with basalt-FRCM systems. Six specimens of
150×200×200 mm were tested under four-point load configuration up to failure. The specimens
were strengthened with continuous U-shaped strips of FRCM. Cementitious and polymer-
modified cementitious mortars were used. The test results were compared with specimens
strengthened with FRP composites. The experimental results approved the effectiveness of using
basalt-FRCM in increasing the flexural strengths of the tested specimens slightly less than FRPs.
Using polymer-modified cementitious matrix provided better bond between the fabric and the
matrix and between the FRCM and the concrete substrate. In addition, 3D finite elements models
were developed to predict the flexural behavior of the tested beams. The basalt-FRCM and FRP
composites were modeled by 4-node shell elements. Bond between composites and concrete was
modeled through the tiebreak surface-to-surface contact definition of LS-DYNA to account for
both normal and shear forces at the interface. A bond-stiffness coefficient, defined as the ratio of
the composite material stiffness to its tensile bond strength, was introduced and recommended not
to be less than 225 for basalt-FRCM to avoid premature debonding failure. The predicted results
were in a good agreement with the experimental results.
Babaeidarabad et al. (2014) tested 18 beams of 1829 mm length, 260 mm depth, and 152 mm
width under three-point load configuration [84]. The test parameters included the concrete strength
(low and high strength), and the number of PBO-FRCM plies used (1 and 4 layers). The specimens
strengthened with 1-ply of fabric showed an increase in the ultimate flexural strength by 32% and
13% for low and high strength concrete, respectively. The specimens failed by fabric slippage
within the matrix while FRCM delamination from the concrete substrate governed the failure of
all specimens strengthened with 4-plies. The flexural strength of specimens with 4-plies of FRCM
was enhanced by 92% and 73% for low and high strength concrete, respectively. A decrease in the
pseudo-ductility of the beams was observed when higher amount of FRCM was used. This work
confirmed the findings of Loreto et al. (2013) [85].
Only one study on the effectiveness of using FRCM composites in repairing corrosion-damaged
RC beams was found in the literature. El-Maaddawy and El Refai (2016) carried out an
25
experimental study to investigate the structural behavior of severely corroded (22% steel mass
loss) T-beams strengthened with either carbon or basalt FRCM systems [86]. The corrosion
damage was restricted to the middle third of the beam span. The FRCM systems were internally-
embedded within the corroded-repaired zone or/and externally bonded along the beam span. The
test results indicated that basalt-FRCM system could not restore the original flexural capacity of
the beam whereas the C-FRCM system restored 109% of the capacity. The use of basalt-FRCM
did not change the brittle failure of the corroded beams (steel rupture) whereas the beams
strengthened with C-FRCM failed by steel yielding followed by FRCM delamination. It was also
concluded that the use of a combination of internally-embedded and externally-bonded C-FRCM
layers was more effective in improving the strength and ductility of the beams than the use of the
same amount of FRCM layers internally embedded within the corroded-repaired region.
2.5 Fatigue and Durability of FRCM-strengthened Beams
Little attention has been paid to investigate the fatigue and durability performance of the RC
beams strengthened with FRCM. The fatigue and long-term performance (after exposure to
environmental conditions) of undamaged RC beams strengthened with PBO-FRCM was
investigated by Aljazaeri and Myers (2016) [87]. Eight beams (203×305×2133 mm) were
strengthened with 1 or 4 layers of PBO-FRCM. The beams were divided into two groups. The
beams of the first group were tested immediately after strengthening while those of the second
group were exposed to varying cycles of freezing and thawing, elevated temperatures, and high
relative humidity before being tested. All of the beams were subjected to fatigue load of 2 million
cycles. The applied fatigue load ranged between 35 and 65% of the specimen ultimate flexural
strength. The results indicated that PBO-FRCM composites enhanced the fatigue performance of
RC beams as well as the residual flexural strength after experiencing fatigue loading. The
enhancement in the fatigue and post-fatigue responses was dependent on the number of layers of
FRCM used. All strengthened beams successfully survived two million cycles without any
evidence of FRCM debonding. Moreover, exposing the PBO-FRCM strengthened beams to high
temperature and humidity did not affect their fatigue performance.
Recently, Pino et al (2017) carried out an experimental study to investigate the fatigue
performance of RC beams strengthened with PBO-FRCM [88]. Ten beams (152×305×1829 mm)
26
were strengthened with 1, 3, or 4 layers of PBO-FRCM. All beams were subjected to 2 million
cycles before testing. The minimum fatigue load was 20% of the specimen theoretical yield load
while the maximum fatigue load varied between 75 and 90% of the specimen theoretical yield
load. The results showed that FRCM improved the fatigue performance of the RC beams. The
fatigue life of the tested specimens decreased with the increase in the maximum fatigue load above
75% of the yield strength. The failure mechanism of the strengthened beams was governed by the
fatigue rupture of the steel bars rather than the failure of FRCM.
2.6 Findings of Literature Review
The literature survey presented in this chapter highlighted in detail the different aspects may
affect the behavior of corroded and FRCM strengthened beams under monotonic and fatigue loads
as well as the long-term performance of FRCM strengthened beams. The findings of the conducted
literature survey can be summarized as follow:
• Accelerated corrosion with impressed electrical current density less than 200 μA/cm2 would
be small enough to obtain corrosion-damage similar to that reported in the field.
• Little emphasis had been given to study the viability of using externally bonded FRCM
composites to strengthened/repair corroded beams at various levels of corrosion damage. In
additions, the effect of FRCM type, amount, and configuration on the flexural response of the
strengthened beams are needed to be quantified.
• To date, no date is available in the literature concerning the durability (post-strengthening
performance) of FRCM strengthened beams after exposure to corrosive environmental
conditions. The effect of FRCM composites on the rate of the corrosion activities was not
investigated.
• The behavior corroded RC beam strengthened with FRCM under fatigue load and the potential
of FRCM composites to restore the fatigue life of corroded beams were not investigated.
• Little effort has been devoted to develop numerical models able to predicted the non-linear
flexural response of FRCM strengthened RC beams.
27
3. Chapter 3
Research Objectives and Methodology
3.1 Research Significance
The conducted literature review presented in Chapter 2 revealed that little research has been
devoted to study the viability of using FRCM to strengthen corrosion-damaged RC elements.
Therefore, the structural behavior of such elements after being exposed to corrosion needs to be
investigated before recommending their use in the field. It is also required to quantify the
strengthening effect of FRCM systems when used in such applications. Corrosion-damaged
structures are often vulnerable to the same deterioration mechanism after repair, which may require
further repair during their service life. To date, no data is available in the literature concerning the
post-repair performance of such structures. In addition, no research has investigated the fatigue
behavior of corrosion-damaged RC beams strengthened with FRCM composites. The lack of
understanding these aspects represents significant obstacles to broad applications of the FRCM
composites. The scope of this study consisted of experimental, numerical, and analytical
investigations. The experimental program was designed to provide a thorough understanding of
the behavior of corrosion-damaged RC beams strengthened with FRCM composites through
design, construction, instrumentation, and testing of 30 large-scale beams as will be detailed in the
following sections. It is thought that the information provided in this thesis is of great value to
designers who wish to use the FRCM composites and also for the development of code provisions
and recommendations.
3.2 Research Objectives
The present study aims at providing technical information about the structural performance of
corrosion-damaged RC beams strengthened with FRCM composites under monotonic and fatigue
loading. The research includes experimental investigations, numerical modeling, and analytical
investigation of the current design code provisions. The objectives of this work can be summarized
as follows:
28
1. To investigate experimentally the viability of using different FRCM systems to strengthen
corrosion-damaged RC beams subjected to various degrees of corrosion.
2. To investigate the effect of FRCM material, number of plies, and strengthening schemes on
the behavior of corrosion-damaged FRCM-strengthened beams.
3. To investigate experimentally the post-strengthening (long-term) behavior of corrosion-
damaged FRCM-strengthened beams under monotonic loads.
4. To investigate experimentally the behavior of corrosion-damaged FRCM-strengthened beams
under fatigue loads.
5. To develop numerical models that are able to predict the non-linear flexural response of
corrosion-damaged FRCM-strengthened beams.
6. To examine the validity of the current design provisions used to predict the flexural strength
of RC members strengthened with FRCM composites on corrosion-damaged FRCM-
strengthened members.
3.3 Methodology
3.3.1 Experimental Work Program
The experimental work included casting and testing thirty large-scale reinforced concrete (RC)
beams of 150 × 250 mm cross section and 2800 mm long. Twenty-seven beams were subjected to
an accelerated corrosion process before being repaired, strengthened, and tested. The main test
parameters included the type of FRCM composites (PBO-FRCM and Carbon FRCM), the level of
corrosion damage (10, 20, and 30% tensile steel mass loss), the amount of FRCM used (1, 2, 3,
and 4 plies of FRCM), the FRCM scheme (end-anchored and continuously wrapped), the short-
and long-term exposure to corrosion, and the type of loading (monotonic and fatigue).
The test matrix of the experimental work program is shown in Table 3.1. The test specimens were
divided into four main groups [A], [B], [C], and [D]. Each group was designed to achieve one or
more of the objectives of the experimental program. The beams were labeled following the X-Y-
Z-N format. ‘X’ represents the beam condition (UU, CU, and CS referring to Uncorroded-
Unstrengthened, Corroded-Unstrengthened, and Corroded-Strengthened, respectively) while ‘M’
29
and ‘F’ refer to Monotonic and Fatigue loading, respectively. ‘Y’ refers to the anticipated
percentage of the tensile steel mass loss due to corrosion. ‘Z’ describes the utilized composite
system (PBO, C, and C-FRP referring to PBO-FRCM, C-FRCM, and Carbon FRP, respectively).
Finally, ‘N’ describes the number of layers and the strengthening scheme of the applied composite
system (I and II refer to the end-anchored and the continuous wrapping schemes, respectively).
3.3.1.1 Group [A]: Control Beams under Monotonic Loading
Group [A] consisted of five control beams that served as benchmarks. Two beams were tested in
flexure up to failure to determine the ultimate capacity of the uncorroded-unstrengthened (virgin)
beam. Three other beams were subjected to accelerated corrosion process for 70, 140, 210 days to
obtain medium (10% mass loss), severe (20% mass loss), and very severe (30% mass loss) levels
of corrosion damage. The beams were then tested in flexure without being strengthened.
3.3.1.2 Group [B]: Short-term Monotonic Behavior of FRCM-strengthened Beams
Group [B] consisted of sixteen beams that were subjected to accelerated corrosion process to
obtain 10, 20, and 30% mass loss in the reinforcing steel bars. The beams were then tested
immediately after being strengthened with different FRCM systems and configurations (short-
term). In addition, one beam was strengthened with C-FRP sheets for comparison. Table 1
illustrates the details of the corrosion level, the strengthening system used, the number of plies,
and the strengthening scheme of each beam. The objective of testing the beams of Group [B] was
to investigate the effect of corrosion-damage level on the flexural performance of the FRCM-
strengthened beams. It also aimed at examining the effectiveness of different FRCM composite
systems in strengthening the corrosion-damaged beams. The test results of Group [B] dictated the
selection of the parameters investigated in the following groups. They also provided guidelines for
the optimal design of FRCM-strengthened beams.
30
Table 3.1: Test matrix of the experimental program
No. Beam Exposure time in days
(Expected mass loss %)
Composite
system (No. of
plies)
Strengthening
scheme
Group [A]: Control beams
1 UUMa 0 (0%) - -
2 UUMb 0 (0%) - -
3 CUM-10% 70 (10%) - -
4 CUM-20% 140 (20%) - -
5 CUM-30% 210 (30%) - -
Group [B]: Short-term beams
6 CSM-10%-PBO-1I 70 (10%) PBO-FRCM (1) I
7 CSM-10%-PBO-2I 70 (10%) PBO-FRCM (2) I
8 CSM-10%-PBO-4I 70 (10%) PBO-FRCM (4) I
9 CSM-10%-PBO-2II 70 (10%) PBO-FRCM (2) II
10 CSM-10%-PBO-4II 70 (10%) PBO-FRCM (4) II
11 CSM-10%-C-2II 70 (10%) C-FRCM (2) II
12 CSM-10%-C-3II 70 (10%) C-FRCM (3) II
13 CSM-10%-CFRP-1I 70 (10%) C-FRP (1) I
14 CSM-20% -PBO-2I 140 (20%) PBO-FRCM (2) I
15 CSM-20% -PBO-4I 140 (20%) PBO-FRCM (4) I
16 CSM-20%-PBO-4II 140 (20%) PBO-FRCM (4) II
17 CSM-20%-C-3II 140 (20%) C-FRCM (3) II
18 CSM-30%-PBO-2I 210 (30%) PBO-FRCM (2) I
19 CSM-30%-PBO-4I 210 (30%) PBO-FRCM (4) I
20 CSM-30%-PBO-4II 210 (30%) PBO-FRCM (4) II
21 CSM-30%-C-3II 210 (30%) C-FRCM (3) II
Group [C]: Long-term beams
22 CSM-10%:30%-PBO-4I 70 (10%):140 (30%) PBO-FRCM (4) I
23 CSM-10%:30%-PBO-4II 70 (10%):140 (30%) PBO-FRCM (4) II
24 CSM-10%:30%-C-3II 70 (10%):140 (30%) C-FRCM (3) II
Group [D]: Beams under fatigue
25 UUF 0 (0%) - -
26 CUF-20% 140 (20%) - -
27 CSF-20%-PBO-2I 140 (20%) PBO-FRCM (2) I
28 CSF-20%-PBO-4I 140 (20%) PBO-FRCM (4) I
29 CSF-20%-PBO-4II 140 (20%) PBO-FRCM (4) II
30 CSF-20%-C-3II 140 (20%) C-FRCM (3) II
31
3.3.1.3 Group [C]: Long-term Behavior of FRCM-strengthened Beams
Group [C] consisted of three specimens that were subjected to accelerated corrosion process for
70 days (10% mass loss in the steel reinforcing bars). The beams were then strengthened with
PBO- or C-FRCM systems using different strengthening schemes as shown in Table 1. After being
repaired, the beams were exposed to further corrosion for 140 days prior to testing (corresponding
to 30% of steel mass loss). The objective of testing this group of beams was to investigate the long-
term behavior of the FRCM-strengthened beams by simulating a strengthened beam in-service.
The effect of wrapping the beams with FRCM on the corrosion rate and the effect of the harsh
environmental conditions on the integrity of the FRCM systems were also investigated. The test
results of this group of beams were compared to those of the corresponding beams of Group B (the
short-term group).
3.3.1.4 Group [D]: Fatigue Behavior of FRCM-strengthened Beams
Group [D] consisted of six specimens. One beam was neither corroded nor repaired and acted as
a control beam. Five beams were subjected to accelerated corrosion process corresponding to 20%
mass loss in the tensile reinforcement. One of the five corroded beams were tested without being
strengthened while the other four corroded beams were strengthened with PBO- or C-FRCM
systems based on the results obtained from the previous tests. The test parameters included the
number of plies and the strengthening scheme as shown in Table 1. All beams of this group were
tested under fatigue loading until failure occurred. The aim of testing this group of beams was to
examine the effectiveness of FRCM systems in restoring the fatigue life of the damaged-then-
strengthened beams.
3.3.2 Numerical Analysis
A nonlinear finite element (FE) analysis was carried out on the investigated beams. The
numerical analysis included the development of three-dimensional FE models to simulate the
flexural behavior of corrosion-damaged RC beams strengthened with C-FRP and PBO-FRCM
composites. The corrosion damage of the steel bars was represented by a reduction in the cross-
section area of the bars based on the actual mass loss obtained during the tests. The C-FRP
composite system was modeled as discrete reinforcement bonded directly to the beam soffit
without binder while the PBO-FRCM composites were modeled using a more detailed approach
32
that involved modeling the fabric and the matrix layers separately. Interfacial bond stress-slip
models were adopted at the CFRP/concrete and the PBO-fabric/matrix interfaces to simulate the
failure mechanisms observed during the tests. The accuracy of the FE models was validated against
the experimental test results. The validated models provided a valuable supplementary to the
laboratory tests and can be used as a numerical platform to predict the performance of RC flexural
members strengthened with FRP and FRCM composites. In addition, the models were used in a
parametric study to investigate the effect of varying the concrete compressive strengths on the
flexural behavior of the FRCM-strengthened beams.
3.3.3 Analytical Investigation
The analytical investigation included the analysis of the test results using the available design
provisions pertinent to RC members strengthened with FRCM composites. The test results of each
beam were compared to the values predicted using the guidelines of ACI 549.4R-13. Comparing
the analytical with the experimental results aimed at enhancing the current design guidelines by
considering new parameters such as the effect of continuous wrapping on the ultimate flexural
strength of the FRCM-strengthened beams.
33
4. Chapter 4
Corrosion-damaged Reinforced Concrete Beams Repaired
with Fabric-Reinforced Cementitious Matrix (FRCM)
Mohammed Elghazy, Ahmed El Refai, Usama Ebead, and Antonio Nanni
Journal of Composites for Construction, ASCE. Submitted in the revised form: Fabraury 10, 2018
Status: Under review
Résumé
La performance structurelle des poutres en béton armé endommagées par la corrosion et réparées
avec la matrice cimentaire renforcée de fibres (MCRF) a été étudiée. Onze poutres à grande échelle
ont été construites et testées en flexion en configuration de charge à quatre points. Neuf poutres
ont été soumises à un processus de corrosion accélérée pendant 70 jours afin d’obtenir une perte
de masse moyenne de 12,6% dans les barres d'armature en acier, tandis que deux autres poutres
ont été testées comme poutres témoins. Une autre poutre corrodée a été réparé avec les polymères
renforcés de fibres de carbone (PRFC) avant d'être testée. Les paramètres d'essai comprenaient le
nombre de couches de fibres (1, 2, 3 et 4), le schéma de réparation de MCRF (couches ancrées aux
extrémités et couches continues sous forme U) et les matériaux MCRF [carbone et
polyparaphénylène benzobisoxazole (PBO)]. Les résultats des tests ont montré que la corrosion
réduisait légèrement les résistances de plastification et ultime des poutres. L'utilisation de MCRF
a augmenté la capacité ultime des poutres corrodées entre 5% et 52% et leur résistance de
plastification entre 6% et 22% de celles de la poutre vierge non-corrodée. Les poutres réparées
avec des bandes de MCRF en U ont montré une capacité et une ductilité plus élevées que celles
réparées avec des couches ancrées aux extrémités ayant un nombre similaire de couches. Un gain
élevé de la capacité de flexion et un faible indice de ductilité ont été rapportés pour les poutres
avec une plus grande quantité de couches MCRF. Un nouveau facteur a été incorporé dans les
équations de conception de l'ACI-549.4R-13 pour tenir compte du schéma MCRF.
Mots-clés des auteurs : Corrosion; Matrice cimentaire renforcée de fibres; Flexion; Béton armé;
Réparation; Renforcement.
34
4.1 Abstract
The structural performance of corrosion-damaged reinforced concrete (RC) beams repaired with
fabric-reinforced cementitious matrix (FRCM) was investigated. Eleven large-scale RC beams
were constructed and tested in flexure under four-point load configuration. Nine beams were
subjected to an accelerated corrosion process for 70 days to obtain an average mass loss of 12.6%
in the tensile steel reinforcing bars while two other beams were tested as controls. One corroded
beam was repaired with carbon fiber-reinforced polymer (C-FRP) before testing for comparison.
The test parameters included the number of fabric plies (1, 2, 3, and 4), the FRCM repair scheme
(end-anchored and continuous U-wrapped strips), and FRCM materials [carbon and
polyparaphenylene benzobisoxazole (PBO)]. Test results showed that corrosion slightly reduced
the yield and ultimate strengths of the beams. The use of FRCM increased the ultimate capacity of
corroded beams between 5% and 52% and their yield strength between 6% and 22% of those of
the uncorroded virgin beam. Beams repaired with U-wrapped FRCM strips showed higher capacity
and higher ductility than those repaired with the end-anchored bottom strips having similar number
of layers. A high gain in the flexural capacity and a low ductility index were reported for specimens
with high amount of FRCM layers. A new factor was incorporated in the design equations of the
ACI-549.4R-13 to account for the FRCM scheme.
Authors’ keywords: Corrosion; Fabric-reinforced cementitious mortars; Flexure; Reinforced
concrete; Repair; Strengthening.
4.2 Introduction and Background
Corrosion of steel reinforcement is one of the main causes of the deterioration of reinforced
concrete (RC) structures. Corroded structures suffer from loss of cross section of the bars, bond
deterioration, and concrete spalling, which can jeopardize the structure’s safety [44,45,48]. Several
techniques had been adopted to repair and strengthen corroded structures, with the use of
externally-bonded steel plates and more recently the epoxy-bonded fiber-reinforced polymers
(FRP) being the most common techniques.
Numerous studies have documented the advantages of using FRP as repair materials for
corrosion-damaged structures [89–91]. However, concerns about the poor fire resistance of epoxy
35
[16], the incompatibility with the concrete substrate [92], and the loss of ductility of the
strengthened structures [93] have been widely reported. In a desire to overcome these drawbacks,
the fabric-reinforced cementitious matrix (FRCM) systems, with their cement-based adhesives,
have been introduced as a promising alternative to the FRP systems [55–59]. FRCM systems are
characterized by their lightweight, high tensile strength, corrosion resistance, and ease of
application. More importantly, the mortars in the FRCM composites act as barriers against chloride
ions penetration, which may protect the reinforcing bars from corrosion. Their mechanical
properties are strongly influenced by the fabric’s material and geometry and the ability of the
cementitious matrix to impregnate the fabric. The bond strength at the fabric/matrix interface and
at the FRCM composite/concrete interface greatly affect the performance of the strengthened
element [74].
Several studies have reported on the use of FRCM in strengthening RC flexural
[16,56,58,92,94,95]. Most of these studies have related the improvement in the performance of the
strengthened elements to the fabric type and the number of layers used. Parameters such as the
FRCM strengthening scheme and the axial stiffness of the FRCM system used were rarely
reported. D’Ambrisi and Focacci (2011) [81] reported 6 to 46% gain in the load-carrying capacity
of RC beams strengthened with carbon and PBO-FRCM systems. Beams strengthened with PBO-
FRCM performed better than those strengthened with C-FRCM due to the superior bond
characteristics of the former system at the fabric/matrix interface. The use of polymer-modified
cementitious matrix in the PBO-FRCM improved the bond of the fabric to the matrix and
consequently increased the ultimate capacity of the strengthened beams.
In another study, Loreto et al. (2013) [85] reported an increase between 35 and 112% in the
flexural capacity of RC slabs strengthened with PBO-FRCM depending on the volume fraction of
the fabric used. The authors reported that increasing the number of FRCM layers reduced the
ductility of the strengthened slabs. Slabs strengthened with one ply failed due to slippage of the
fabric within the matrix whereas those repaired with four plies failed by fabric delamination at the
fabric/matrix interface. These results were also confirmed by Babaeidarabad et al. (2014) [84].
In a comparison between the performance of FRCM- and FRP- strengthened beams, Elsanadedy
et al. (2013) [83] reported that basalt-FRCM systems were less effective than carbon-FRP systems
36
in enhancing the flexural strength of the beams, yet the FRCM-strengthened beams showed more
ductility at ultimate.
On the other hand, the feasibility of using FRCM systems to strengthen corrosion-damaged
concrete structures have received little attention. The challenges in repairing corroded RC elements
are two-fold, namely the absence of a sound concrete substrate due to corrosion and the durability
of the repair system should corrosion resumes. To the authors’ knowledge, only two studies [86,96]
have documented the effectiveness of using FRCM systems to restore the ultimate capacity and
serviceability of corroded beams. El-Maaddawy and El Refai (2016) [86] reported on the flexural
response of T-beams repaired with carbon and basalt FRCM systems after a mass loss of 22% in
their tensile reinforcement due to corrosion. It was concluded that the basalt-FRCM system could
not restore the original flexural capacity of the beam whereas the C-FRCM system restored 109%
of the capacity. The authors reported that the use of a combination of internally-embedded and
externally-bonded C-FRCM layers was more effective in improving the strength and ductility of
the beams than the use of the same amount of FRCM layers internally embedded within the
corroded-repaired region.
This paper reports the results of the flexural tests conducted on corrosion-damaged RC beams
repaired with FRCM systems. The test program included the type of the FRCM used (carbon and
PBO), the FRCM reinforcement ratios (represented by the number of fabric plies bonded to the
concrete substrate, namely 1, 2, 3, and 4 plies), and the FRCM repair scheme (end-anchored and
U-wrapped strips). The paper also reports on the failure modes, the load-carrying capacities, the
ductility, and the straining actions at different stages of loading of the tested beams. Theoretical
formulations are also presented to predict the flexural response of the beams.
4.3 Experimental Program
Eleven large-scale RC beams were constructed and tested as follows: two specimens were neither
corroded nor repaired (UU), one specimen was corroded but not repaired (CU), seven specimens
were corroded then repaired with different FRCM systems, and one specimen was corroded and
repaired with carbon-FRP (CFRP) sheets. The test matrix is shown in Table 4.1.
The beams were labeled following the X-Y-Z format. ‘X’ represents the beam condition (UU,
CU, and CR referring to Uncorroded-Unrepaired, Corroded-Unrepaired, and Corroded- repaired,
37
respectively). ‘Y’ denotes the number and type of the FRCM layers applied (1P, 2P, 4P, 2C, 3C
and 1FRP referring to one layers of PBO-FRCM, two layers of PBO-FRCM, four layers of PBO-
FRCM, two layers of C-FRCM, three layers of C-FRCM, and one laminate of Carbon FRP
respectively). Finally, ‘Z’ describes the FRCM strengthening schemes (I and II) as will be detailed
in the following sections.
4.3.1 Test Specimen
The test specimen was 2800 mm long with a 150×250 mm rectangular cross section. All beams
were reinforced with 2-15M deformed bars at the bottom (As = 400 mm2) and 2-8M deformed
bars at the top (As' = 100 mm2). The tensile reinforcement ratio was kept at 1.067% that represents
a typical ratio associated with under-reinforced RC section. All of the specimens had a constant
moment span of 800 mm and two shear spans of 880 mm. The shear spans were reinforced with
10M deformed stirrups spaced at 100 mm to avoid a premature shear failure. A hollow stainless-
steel tube with external and internal diameters of 9.5 mm and 7 mm, respectively, was placed at
80 mm from the specimen tension face to act as cathode during the accelerated corrosion process.
Typical dimensions and reinforcement details of the test specimen are shown in Figure 4.1.
38
Table 4.1: Test matrix
*SY-CC = Steel Yielding followed by Concrete Crushing; FD = Fabric Delamination; FS
= Fabric Slippage; MC-SFM = Matrix Cracking with Separation of Fabric within the
Matrix; LR = C-FRP Laminate Rupture.
Figure 4.1: Typical dimensions and reinforcement details of the test specimen (all dimensions in
mm)
Specimen Average mass loss
(%)
𝜌𝑓
(%)
𝐾𝑓 = 𝜌𝑓𝐸𝑓
(MPa)
Mode of Failure*
UUa, UUb - - - SY-CC
CU 12.9 - - SY-CC
CR-1P-I 13.7 0.02 24.1 SY-CC
CR-2P-I 13 0.04 48 FD
CR-4P-I 12.6 0.08 95.3 FD
CR-2P-II 13 0.04 48 FS
CR-4P-II 12.3 0.08 95.3 FS
CR-2C-II 12.5 0.125 93.5 MC-SFM
CR-3C-II 12.1 0.185 139.1 MC-SFM
CR-1FRP-I 11.6 0.153 99.7 LR
39
4.3.2 Accelerated Corrosion Aging
Salt (NaCl) measured as 5% of the cement weight was added to the concrete mix used to cast the
middle-bottom of the corroded specimens with a height of 100 mm (Figure 4.1). Corrosion of the
main reinforcement was localized in the middle 1200 mm of the beam’s span. The accelerated
corrosion process was induced by using a power supply to impress a constant electrical current of
380 mA on the tensile steel bars. The applied current corresponded to a current density of 180
µA/cm2. The invention of this process was produce the desired corrosion damaged within
reasonable amount of time but without remarkable change in the structural response that would be
encountered due to natural corrosion [21]. As there is no laboratory testing standards for the
accelerated corrosion of RC specimens, the current density was chosen based on a previous study
by El Maaddawy and Soudki (2003) [20] that limited the maximum impressed current intensity to
200 µA/cm2 to represent well the natural corrosion in terms of its resulting products. It is also
important to note that this current density is larger than the current densities in real structures which
normally ranged between 0.1 and 100 µA/cm2 [29]. Therefore, this area still needs extensive
research and standardization. During the accelerated corrosion process, the bottom reinforcement
acted as anode whereas the stainless-steel tube acted as cathode and the salted concrete acted as
electrolyte.
The test specimens were electrically connected in series to ensure that the induced current was
uniform in all specimens (Figure 4.2). With applying the anodic current, all specimens were
subjected to wet-dry cycles that consisted of 3 days wet followed by 3 days dry in a large
environmental chamber. The wet-dry cycles provided water and oxygen necessary for the
corrosion process. In this study, a 10% mass loss in the reinforcing bars was anticipated to
represent moderate corrosion damage which commonly encountered in real structures. According
to Faraday’s law that relates the mass loss to the electrical current and the exposure duration, the
duration of corrosion exposure required to achieve this mass loss was 70 days.
40
Figure 4.2: Specimens connected in series inside the corrosion chamber
4.3.3 Materials
Two types of ready mix concrete namely, normal and salted with the same water/cement ratio
were used to cast the beams. Standard concrete cylinders (150×300 mm) were prepared from each
concrete batch and were tested in compression on the day of testing at least after 28 days of casting.
Table 4.2 lists the compressive strengths of both mixes. Prior to FRCM application, the corroded
beams were repaired using local commercial cementitious repair mortar (Sikacrete-08SCC) having
a compressive strength of 55.4 MPa (standard deviation of 5 MPa) and flexural strength of 3.4
MPa (standard deviation of 0.3 MPa) as determined by the authors. The yield strengths of the
longitudinal reinforcing steel bars of diameter 15 and 8 mm were 466 MPa (with a standard
deviation of 4.2 MPa) and 573 MPa (with a standard deviation of 17.7 MPa), respectively, as tested
by the authors.
Two commercial FRCM systems (PBO and carbon) in addition to carbon-FRP composites were
used to strengthen the corroded specimens (Figure 4.3). The fabric properties in the primary
direction as reported in the manufacturers’ data sheet are shown in Table 4.3. The PBO fabric
consists of an unbalanced net of spaced fiber rovings organized along two orthogonal directions
as shown in Figure 4.3a. The associated inorganic cementitious matrix had a compressive strength
of 43.9 MPa (standard deviation of 0.4 MPa) and a flexural strength of 3 MPa (standard deviation
of 0.3 MPa) after 28 days as determined by the authors. On the other hand, the C-FRCM composite
consists of unidirectional net made of carbon-fiber strands oriented in one direction (Figure 4.3b)
41
and impregnated in an inorganic cementitious matrix of compressive strength of 42.1 MPa
(standard deviation of 4.3 MPa) and flexural strength of 3.2 MPa (standard deviation of 0.3 MPa)
after 28 days as determined by the authors. Finally, the carbon-FRP composite consists of
unidirectional carbon fiber sheet (Figure 4.3c) and an epoxy resin. According to the manufacturer’s
data sheet, the composite has a tensile strength of 0.89 GPa, a tensile modulus of 65.4 GPa, and an
ultimate elongation of 1.33%. Table 4.4 lists the properties of the FRCM composite systems as
reported by Ebead et al. (2016) [97].
Table 4.2: Concrete compressive strengths
Compressive strength
(MPa)
Standard deviation
(MPa)
Coefficient of variation
(%)
28-day Normal concrete 37.9 0.8 2
Salted concrete 33.5 1.1 3.2
Testing day Normal concrete 41.8 4.8 11.4
Salted concrete 41.2 0.6 1.6
Figure 4.3: Strengthening materials: a) unbalanced PBO fabric, b) unidirectional carbon fabric,
and c) unidirectional carbon fabric
42
Table 4.3: Fabric properties in the primary direction as given in the manufactures’ data sheet
Fabric Area per unit width
(𝐴𝑓) (mm2/m) Tensile strength
(GPa)
Elastic modulus
(GPa)
Ultimate strain
(%)
PBO 50 5.8 270 2.15
Carbon 157 4.3 240 1.75
CFRP 381 3.45 230 1.5
Table 4.4: Mechanical properties of FRCM systems [97]
FRCM system Cracked tensile modulus
of elasticity, Ef (GPa)
Ultimate tensile
strength, ffu (GPa)
Ultimate strain,
εfu (%)
PBO-FRCM 121 1.55 1.4
C-FRCM 75 0.97 1.25
4.3.4 FRCM Equivalent Axial Stiffness
According to the ACI 549.4R (2013) provisions [54], the tensile stress-strain curve of the FRCM
coupon can be represented by a simple bilinear curve as shown in Figure 4.4 The first linear
segment represents the behavior of the FRCM composite prior to cracking and is characterized by
the uncracked modulus of elasticity, 𝐸𝑓∗. The second linear segment represents the cracked
behavior with a reduced cracked modulus of elasticity, 𝐸𝑓. An equivalent axial stiffness, Kf, was
utilized to compare between the two FRCM systems used in this study based on their cracked
elastic modulus and the cross-sectional area of the fabric as given by Equation (4.1):
𝐾𝑓 = 𝜌𝑓𝐸𝑓 = [(𝑁𝐴𝑓)/𝑑𝑓]𝐸𝑓 Eq. (4.1)
where
𝜌𝑓 =𝑁𝐴𝑓
𝑑𝑓
𝜌𝑓, Af, and 𝐸𝑓 are listed in Table 4.1, Table 4.3, and Table 4.4, respectively. The equivalent axial
stiffness, Kf, of each repaired specimen is shown in Table 4.1. It is important to note that for beams
strengthened with the continuously wrapped PBO-FRCM layer (scheme II), the fibers located on
the lateral sides of the beams were neglected in estimating 𝜌𝑓 (and consequently 𝐾𝑓) due to their
minor contribution to the flexural strength of the beams.
43
Figure 4.4: Idealized tensile stress-strain curve of FRCM coupon specimen [54]
4.3.5 FRCM Repair Schemes
Two FRCM repair schemes were utilized in this study, as shown in Figure 4.5. Scheme I
consisted of one or more FRCM flexure plies having 150 mm width (equal to the width of the
beam) and applied to the soffit of the beam over a length of 2400 mm. The fabrics were oriented
so that their primary direction was parallel to the longitudinal axis of the beam. The plies were
anchored at each end using one U-shaped transverse strip of 300 mm width and 200 mm height as
shown in Figure 4.5a. Scheme II consisted of bottom flexural strips similar to those of Scheme I
but wrapped with an additional U-shaped continuous ply along the beam’s span (Figure 5b). The
primary direction of the U-wrapped PBO ply was oriented parallel to the longitudinal axis of the
beams. For instance, the beam CR-4P-II consisted of 3 bottom flexural strips plus one U-shaped
layer, with the primary fibers of all 4 layers running parallel to the longitudinal axis of the beam.
Therefore, the 4 layers of the PBO-fabric contributed to the flexural performance of the beam. On
the other hand, the carbon fabric is a unidirectional fabric. Therefore, the bottom strips of the C-
FRCM were oriented parallel to the longitudinal axis of the beams whereas the U-shaped layer
was oriented in the transverse direction and therefore did not contribute to the flexural behavior of
the beam (Figure 4.5b). For example, specimen CR-3C-II was repaired with 3 flexural strips in the
longitudinal direction plus one U-shaped layer in the transverse direction. Only 3 layers of the C-
FRCM were considered later in the analysis of this beam.
44
Figure 4.5: Repair schemes: (a) Scheme I and (b) Scheme II
4.3.6 Repair Technique
Corroded specimens were repaired before applying the FRCM repair system. Figure 4.6 depicts
the repair procedure. The deteriorated concrete was first removed using a hydraulic hammer. The
corroded steel bars were then brushed, and the beams were repaired using Sikacrete-08SCC
mortar. After 7 days of curing in ambient temperature, sandblasting was used to roughen the
concrete substrate. The beam’s substrate was damped in water for 2 hours before applying the first
layer of the cementitious matrix with a thickness of 3 to 4 mm. Then, the fabric was installed and
coated with a second layer of matrix of similar thickness. The procedure was then repeated until
the specified number of layers was attained.
45
Figure 4.6: Repair procedure: a) removing the deteriorated concrete, b) patch repair, c)
roughening the concrete surface with sandblasting, and d) FRCM application
4.3.7 Test Setup and Instrumentation
All beams were instrumented at mid-span with a 60 mm long concrete strain gauge bonded to
the top surface of the beams and 5 mm steel strain gauges bonded to the tensile reinforcing bars.
The repaired specimens were instrumented with 5 mm strain gauges installed directly on the outer
fabric of the FRCM composite and distributed along the beam span as shown in Figure 4.7. The
beams were tested under four-point loading configuration as shown in Figure 4.1. The load was
applied in displacement control at a rate of 2 mm per minute using a MTS actuator. Beam
deflections were measured by means of three linear variable differential transducers (LVDTs)
located at mid-span and under the point loads. A data acquisition captured the readings of strain
gauges and LVDTs at all stages of loading.
Figure 4.7: Positions of the electrical strain gauges along the outer fabric
46
4.4 Test Observations
4.4.1 Corrosion Cracks and Mass Loss
Due to corrosion, continuous longitudinal cracks parallel to the reinforcing bars were observed
as shown in Figure 4.8 for specimen CU. No concrete spalling was observed. All of the corroded
specimens did not meet the ACI 318-14 [98] service requirements that limits the maximum crack
width in service to 0.40 mm ACI 318-14 [98]. The average and maximum measured crack widths
after corrosion were determined as 0.7 and 1 mm, respectively.
Figure 4.8: Corrosion cracks pattern for specimen CU
Visual inspection of the corroded beams revealed the existence of several corrosion pits randomly
dispersed along the surface of the bars. Six steel coupons, 200 mm long each, were extracted from
each corroded bar after testing. The actual mass losses of the examined bars were determined
according to the ASTM G1-03 standards [26]. The average tensile steel mass loss for each
specimen are listed in Table 4.1. The average, minimum, and maximum steel mass loss determined
for all specimens were 12.6, 11.5, and 13.7%, respectively.
4.4.2 Modes of Failure
The modes of failure of the tested specimens are summarized in Table 4.1 and shown in Figure
4.9 for the tested beams. Beams UU (control) and CU (corroded unrepaired) failed by yielding of
the steel bars followed by concrete crushing (SY-CC). A similar mode of failure was observed in
specimen CR-1P-I as shown in Figure 4.9a. No loss of bond was observed between the PBO-
FRCM and the concrete substrate while loading. The PBO fabric remained intact with its matrix
until crushing of concrete occurred at ultimate. For the other repaired specimens, four different
modes of failure were observed:
47
a) FRCM delamination (FD): this type of failure occurred at the fabric/matrix interface with
complete delamination between the fabric and the first layer of the matrix adjacent to the concrete
substrate (Figure 4.9b). The delamination was caused by the propagation of flexural cracks to this
thin layer of the matrix and the relative deformation between the fabric and the matrix. This mode
of failure was reported for specimens CR-2P-I, and CR-4P-I.
b) Fabric slippage (FS): slippage occurred between the PBO U-shaped fabric and its
cementitious matrix (Figure 4.9c). Cracks were first observed in the matrix of the U-shaped FRCM
layer followed by the gradual slippage of the fabric until the FRCM strengthening action was lost.
This mode of failure was observed in specimens CR-2P-II and CR-4P-II. It should be noticed that
the continuous PBO-U-shaped ply mitigated the FRCM delamination. Therefore, specimens that
failed in this category showed a more ductile behaviour compared to that observed in specimens
that failed due to FRCM delamination.
c) Matrix cracking and fabric separation from the matrix [MC-SFM)]: this type of failure was
reported for specimens with C-FRCM namely, CR-2C-II and CR-3C-II, as shown in Figure 4.9d.
As the applied load increased, progressive cracking in the cementitious matrix associated with the
separation of the carbon fabric from the matrix was observed. Matrix cracking took a web pattern
as shown in Figure 4.9d for the bottom of specimen CR-3C-II. This mode of failure was more
brittle than that observed in the PBO-repaired specimens, which can be attributed to the superior
characteristics of the cementitious matrix of the PBO-FRCM compared to those of the C-FRCM
counterparts.
d) C-FRP laminate rupture (LR): this mode of failure was reported for specimen CR-1FRP-I
(Figure 4.9e). A longitudinal crack initiated at mid span at the laminate/concrete interface followed
by the sudden rupture of the laminate. This mode of failure was consistent with the high strains
recorded in the laminate at ultimate.
48
Figure 4.9: Typical modes of failure: (a) SY-CC in beam CR-1P-I, (b) FD in beam CR-2P-I, (c)
FS in beam CR-2P-II, (d) MC-SFM in beam CR-3C-II, and (e) LR in beam CR-1FRP-I
4.4.3 Load-deflection Response
Load-deflection relationships of the tested beams are shown in Figure 4.10 to Figure 4.12. The
flexural response of the virgin beam (UU), the corroded-unrepaired beam (CU), and the FRP-
repaired beam (CR-1FRP-I) are also shown for reference. The load-deflection curve of specimen
CU indicated that corrosion slightly reduced the load-carrying capacity and stiffness of the beam.
The load-deflection curve of the repaired beams consisted of three segments with two turning
Concrete
crushing
FRCM delamination
Fabric slippage
Matrix crushing
Laminate rupture
49
points indicating the concrete cracking and the yielding of the tensile steel. The flexural response
of the repaired beams was highly dependent on the FRCM repair scheme, its type, and the number
of FRCM layers used.
Figure 4.10 shows the load-deflection relationships of the beams repaired with PBO-FRCM using
Scheme I. All of the beams showed similar stiffness prior to yielding of steel reinforcing bars
indicating the slight influence of the FRCM composite on the flexural response at this stage.
Increasing the number of the PBO plies increased the post-yielding stiffness of the repaired
specimens in comparison to the control ones. Specimen CR-1FRP-I (repaired with one layer of C-
FRP fabric) showed higher post-yielding stiffness than that of specimen CR-4P-I (repaired with
four layers of PBO fabric). However, the later specimen showed slightly higher load carrying
capacity with more ductile mode of failure than the former one.
Figure 4.11 shows the effect of the FRCM scheme on the load-deflection response of the PBO-
repaired beams. Specimens repaired with two and four PBO plies in Scheme II showed a slight
enhancement in the pre-yielding and post-yielding stiffness, which can be attributed to the
enlargement of the beam width and the effect of the continuous U-wrapped strips in delaying the
delamination of the FRCM.
Figure 4.12 compares the load-deflection responses of the Carbon- and PBO-FRCM repaired
beams using scheme II. It can be noticed that specimens repaired with C-FRCM showed higher
post-yielding stiffness than that of their PBO-repaired counterparts. The former specimens
exhibited a sudden drop after reaching the ultimate load whereas specimens repaired with PBO-
FRCM showed a gradual decreasing trend after reaching the ultimate. This can be related to the
brittle mode of failure reported for specimens repaired with C-FRCM.
4.4.4 Strength Analysis
Table 4.5 gives the strength results of the tested beams. The experimental yield, 𝑃𝑦𝑒𝑥𝑝
, and
ultimate, 𝑃𝑢𝑒𝑥𝑝
, loads of all specimens were normalized to those of the virgin specimen. It can be
noticed that corrosion of the main reinforcement reduced the yield and ultimate loads by 8% and
5%, respectively. The reduction in the load-carrying capacity due to corrosion was smaller than
the steel mass loss due to the good anchorage of the bars within the shear zone, which allowed a
tied-arch action to be developed when the specimen approached failure [10].
50
Figure 4.10: Effect of number of PBO-FRCM plies on the load-deflection curves
Figure 4.11: Effect of the repair scheme on the load-deflection curves
51
Figure 4.12: Effect of FRCM materials on the load-deflection curves
Table 4.5: Strength results
Specimen 𝑃𝑦
𝑒𝑥𝑝
(KN)
𝑃𝑢𝑒𝑥𝑝
(KN)
Normalized loads** 𝑃𝑢𝑡ℎ
(KN)
𝑃𝑢𝑒𝑥𝑝
𝑃𝑢𝑡ℎ
ϕ𝑚𝑃𝑢𝑡ℎ
(KN)
𝑃𝑢𝑒𝑥𝑝
ϕ𝑚𝑃𝑢𝑡ℎ
𝑃𝑦𝑒𝑥𝑝
𝑃𝑢𝑒𝑥𝑝
UUa, UUb* 75.1 79.7 1 1 78.5 1.02 70.65 1.13
CU 69.5 76.1 0.92 0.95 70.6 1.08 63.54 1.1
CR-1P-I 71.1 82.9 0.95 1.05 74.9 1.10 67.4 1.23
CR-2P-I 79.5 86.4 1.06 1.08 81.1 1.06 73 1.18
CR-4P-I 83.3 99.6 1.11 1.25 93.3 1.06 84 1.19
CR-2P-II 85.4 102.2 1.14 1.28 81.1 1.26 73 1.4
CR-4P-II 91.3 114.4 1.22 1.44 93.3 1.23 84 1.36
CR-2C-II 79.8 104 1.06 1.30 93.4 1.11 84 1.25
CR-3C-II 90 120.6 1.16 1.52 106 1.14 95.3 1.27
CR-1FRP-I 77.9 96.5 1.04 1.21 - - - -
* Average values reported **Normalized with respect to the yield and ultimate loads of the virgin beam
4.4.4.1 Effect of Number of FRCM Plies on Strength
The use of a single PBO-FRCM layer in specimen CR-1P-I restored 95 and 105% of the yield
and ultimate loads of the virgin beam, respectively. Increasing the number of PBO-FRCM layers
52
further increased the yield and ultimate loads (specimen CR-2P-I restored 106 and 108% and
specimen CR-4P-I restored 111 and 125% of the yield and ultimate loads, respectively). However,
the strength enhancement was not linearly proportional to the added number of FRCM layers.
A similar trend was encountered in specimens repaired with Scheme II. Increasing the number
of FRCM layers enhanced the yield and ultimate strengths of the repaired beams. Specimen CR-
4P-II showed an increase of 22 and 44% of the yield and ultimate loads, respectively, compared
to 14 and 28% for specimen CR-2P-II. Similarly, the use of two layers of C-FRCM in specimen
CR-2C-II increased the yield and ultimate strengths by 6 and 30%, respectively, compared to 16
and 52% for specimen CR-3C-II.
4.4.4.2 Effect of FRCM Repair Scheme on Strength
Scheme II was more effective than Scheme I in restoring the yield and load-carrying capacity of
the repaired beams. This was depicted from the results of the beams repaired with two and four
PBO-FRCM layers. The enhancement in yield load was 6 and 14% for specimens CR-2P-I and
CR-2P-II, respectively. Their corresponding ultimate strengths increased by 8 and 28%,
respectively. The use of four layers of PBO-FRCM with Scheme II in specimen CR-4P-II
increased the yield and ultimate loads by 22 and 44%, respectively, in comparison to 11 and 25%
for specimen CR-4P-I having the same number of PBO-fabric layers.
4.4.4.3 Effect of Axial Stiffness on Strength
Figure 4.13 shows the effect of changing the axial stiffness of the strengthening system, Kf, on
the normalized ultimate load of the tested specimens. Specimens with similar axial stiffness of
their repair system didn’t show similar ultimate capacities. This can be depicted from the results
of specimens CR-2P-I and CR-2P-II having the same axial stiffness of their FRCM system but
with different repair schemes. The former specimen showed a load-carrying capacity of 86.4 KN
versus 102.2 KN for the later one. Similarly, specimens CR-4P-I and CR-4P-II, also having the
same axial stiffness, showed 99.6 KN and 114.4 KN, respectively. This finding was also
demonstrated in specimens CR-4P-II and CR-2C-II having almost similar axial stiffness but
repaired with two different FRCM systems. Specimen CR-4P-II showed a load carrying capacity
of 114.4 KN whereas the specimen CR-2C-II showed a load carrying capacity of 104 KN. This
finding indicates that the axial stiffness of repair system, Kf, should not be used solely to compare
53
the strengthening actions of different FRCM systems without taking into account the material
properties, the fabric architecture, and the repair scheme used. The equivalent axial stiffness of
FRCM systems needs to be further calibrated with more experimental tests to implement other
parameters which is out of scope of this study.
Figure 4.13: Normalized ultimate load versus the equivalent stiffness
4.4.5 Ductility Performance
The ductility index, ΔI, defined as the ratio of the midspan deflection of the beam at ultimate, δu,
to its midspan deflection at yielding, δy, was used to quantify the ductility of the tested specimens.
In general, a higher ductility index means a higher ability of the beam to redistribute moment and
to exhibit large overall deformation and energy dissipation [84]. Table 4.6 lists the deflections at
yielding and ultimate and the ductility indices for all of the tested beams normalized to that of the
virgin beam.
It can be noticed that corrosion of the steel bars increased the ductility index of the corroded
beam by 15%. All beams repaired with PBO in Scheme I restored the ductility of the virgin beam
except beam CR-4P-I that showed a ductility index 13% less than that of the virgin beam. For this
set of beams, increasing the number of PBO plies decreased the ductility of the repaired beam. The
ductility indices of beams CR-4P-I, CR-2P-I, and CR-1P-I were 2.4, 2.8, and 3.0, respectively. On
the other hand, the CFRP-repaired specimen (CR-1FRP-I) did not restore the ductility of the virgin
54
beam and had a similar ductility index of its FRCM-repaired counterpart (CR-4P-I) having similar
axial stiffness.
The set of beams repaired with PBO in Scheme II restored the ductility of the virgin beam. In
fact, these beams showed 2 to 13% increase in their ductility indices as compared to their
counterparts repaired with scheme I. However, increasing the number of the PBO plies in Scheme
II had a less pronounced effect on the ductility index than in Scheme I. Beams CR-4P-II and CR-
2P-II had ductility indices of 2.8 and 2.9, respectively, which indicates that doubling the number
of plies in Scheme II resulted in only 3.5% reduction in the ductility index of the beam.
The ductility indices of the beams repaired with C-FRCM (CR-3C-II and CR-2C-II) were lower
than that of the beams repaired with PBO-FRCM having same repair scheme. Both beams couldn’t
restore the ductility of the virgin beam. Their ductility index was 14 and 22% less than that of the
virgin beam, respectively. This reduction in ductility was attributed to their brittle mode of failure
that was due to the rapid loss of the strengthening action at the fabric/matrix interface. It is
important to note that increasing the number of carbon plies in this set of beams increased the
ductility index of the beam, which is contrary to what has been noticed in the PBO-repaired beams.
This increase was attributed to the increase in the ultimate load of the C-FRCM repaired beams
with similar yielding deflections in comparison to their PBO-counterparts.
4.4.6 Strain Response
Table 4.6 lists the strains measured at midspan in both concrete and the outer fabric at ultimate.
Figure 4.14 and Figure 4.15 show the load-strain curves for specimens repaired with Scheme I and
Scheme II, respectively. Similar to the load-deflection responses of the repaired beams, the load-
strain curves consisted of three segments with two turning points that indicated the concrete
cracking and the yielding of the tensile steel.
55
Table 4.6: Ductility indices and strains at ultimate
Specimen
Midspan
deflection (mm) Ductility index Concrete strains
at ultimate (µ𝜖)
Fiber strains at
ultimate (µ𝜖) δy δu ΔI ΔInorm
**
UUa, UUb* 11.7 32.9 2.8 1.0 -3311 -
CU 10.9 35.4 3.2 1.15 -2992 -
CR-1P-I 11.6 35.2 3.0 1.08 -2711 14921
CR-2P-I 11.8 33.0 2.8 1.0 -2342 8670
CR-4P-I 12.9 31.5 2.4 0.87 -2421 9541
CR-2P-II 11.2 32.0 2.9 1.02 -3491 11261
CR-4P-II 12.7 35.5 2.8 1.0 -2761 9598
CR-2C-II 12.6 27.6 2.2 0.78 -2370 5753
CR-3C-II 12.4 30.1 2.4 0.86 -2262 5991
CR-1FRP-I 12.3 30.4 2.5 0.88 -2351 13772
* Average values reported **Normalized with respect to the yield and ultimate loads of the virgin beam
Figure 4.14: Load-strain curves for specimens with repair Scheme I
56
Figure 4.15: Load-strain curves for specimens with repair Scheme II
Figure 4.14 shows that, prior to yielding, all repaired specimens showed a similar increase in
concrete strains as the applied load increased. This increase in concrete strains continued after
yielding but at different rates depending on the repair system used. Specimen CR-1FRP-I showed
the highest rate of increase in concrete strains when compared to the PBO-repaired ones. On the
other hand, specimen CR-1P-I recorded the maximum tensile strains in the outer fabric of FRCM
(14921 μɛ) as no fabric delamination was observed for this specimen until failure. Specimens CR-
2P-I and CR-4P-I, repaired with two and four plies, failed by FRCM delamination and therefore
recorded low tensile strains in the PBO fabric (8670 μɛ and 9541 μɛ, respectively).
As shown in Figure 4.15, the concrete strains measured in the PBO-repaired specimens were
higher than those recorded in their C-FRCM counterparts. For instance, specimens CR-2P-II and
CR-2C-II recorded concrete strains of 3491 and 2370 μɛ, respectively. It was observed that
concrete strains of the C-FRCM specimens increased at higher rate than that of strains of the PBO-
FRCM specimens. This can be depicted from the strains recorded for specimens CR-2C-II and
CR-3C-II in Figure 4.15. On the other hand, the tensile strains in the C-FRCM at failure was lower
than those in PBO-FRCM. Specimens CR-2C-II and CR-3C-II recorded tensile strains in the outer
fabric at failure of 5753 μɛ and 5991 μɛ, respectively, whereas their counterparts CR-2P-II and
CR-4P-II recorded tensile strains of 11262 μɛ and 9598 μɛ, respectively. These findings were
consistent with the mode of failure of the C-FRCM repaired specimens where premature matrix
57
cracking and fabric separation were encountered. They were also consistent with the measured
ductility indices for these beams.
The distribution of the outer fabric tensile strains along the beam axis are plotted in Figure 4.16
to Figure 4.18 for specimens CR-4P-I, CR-4P-II, and CR-3C-II, respectively, at a service load
equal to 60% of ultimate (0.6 Pu), at the yielding load (Py), and at two post-yielding loads equal
to 0.9 Pu, and Pu. It can be noticed that the strains in the fabric increased with the increase of the
applied load until yielding occurred. Post yielding, a significant increase in fabric strains were
observed, with the maximum increase occurring in the constant moment zone. This finding
indicates that the FRCM system became more effective in resisting the applied loads after yielding
of the steel bars. The repair scheme had marginal effect on the fabric strain profiles as can be
depicted from Figure 4.16 and Figure 4.17.
It is important to note that relying solely on the effective strains in the fabric may be misleading
in predicted computations based on prefect bond between the FRCM and the concrete substrate.
In fact, the slip of the fabric gives origin to a pseudo-strain that can capture the effectiveness of
FRCM strengthening in design. The assumption of prefect bond suggested by the ACI 549.4R-13
committee [54] is a simplification that appears justifiable and easy to implement by engineers.
Figure 4.16: Strain profile in the PBO fabric for specimen CR-4P-I
58
Figure 4.17: Strain profile in the PBO fabric for specimen CR-4P-II
Figure 4.18: Strain profile in the carbon fabric for specimen CR-3C-II
4.5 Theoretical Predictions
The flexural behavior of the tested beams were predicted according to the provisions of the ACI
318-14 building code [98] and the ACI 549.4R-13 committee [54]. Perfect bond was assumed
between the fabric and the cementitious matrix and between the FRCM and the concrete substrate.
A bilinear-elastic behavior of the FRCM repair system was presumed up to failure. The cracked
tensile modulus of elasticity, Ef, of the FRCM system was used after cracking,
59
The FRCM effective tensile strain at failure, 휀𝑓𝑒 , was limited to the FRCM design tensile
strain, 휀𝑓𝑑, as given in Equation (4.2a) (ACI 549.4R [54]). The effective tensile stress in the FRCM
at failure, 𝑓𝑓𝑒 , was calculated in accordance with Equation (4.2b)
휀𝑓𝑒 = 휀𝑓𝑑 ≤ 0.012 Eq. (4.2a)
𝑓𝑓𝑒 = 𝐸𝑓휀𝑓𝑒 Eq. (4.2b)
Strains in concrete, steel reinforcing bars, and FRCM systems were computed in accordance with
Equation (4.3) using the strain compatibility principle as shown in Figure 4.19.
𝜀𝑓𝑒
𝑑𝑓−𝑐𝑢=
𝜀𝑠
𝑑−𝑐𝑢 =
𝜀𝑠′
𝑐𝑢−𝑑′=
𝜀𝑐
𝑐𝑢 Eq. (4.3)
Figure 4.19: Stress and strain distribution at ultimate stage
The nominal flexural strength, Mn, was calculated in accordance with Equations (4.4) as follows:
𝑀𝑛 = 𝑀𝑠 + 𝑀𝑓 + 𝑀𝑠′ Eq. (4.4)
where,
𝑀𝑠 = 𝑇𝑠 ( d − 𝛽1 𝐶𝑢
2 ) Eq. (4.4a)
𝑀𝑓 = 𝑇𝑓 ( d − 𝛽1 𝐶𝑢
2 ) Eq. (4.4b)
𝑀𝑠′ = 𝐶𝑠′ (𝛽1 𝐶𝑢
2− 𝑑′) Eq. (4.4b)
60
𝑀s, 𝑀f, and 𝑀s were calculated with respect to the centroid of the equivalent rectangular stress
block as shown in Figure 4.19. The concrete stress block factors, 𝛽1 and 𝛼1, and the modulus of
elasticity of concrete, Ec, were calculated as follows (ACI 318-14 [98]):
β1 = (4εc
′ −εc(Cu)
6εc′ −2εc(Cu)
) Eq. (4.5)
α1 = (3εc
′ εc(Cu)−[εc(Cu)]2
3β1(Cu)εc′2 ) Eq. (4.6)
𝐸𝑐 = 4700√𝑓c′ Eq. (4.7)
휀𝑐′ = 1.7𝑓𝑐
′/𝐸𝑐 Eq. (4.8)
The force equilibrium was satisfied in accordance with Equations (4.9) and as shown Figure 4.19:
𝑇𝑠 + 𝑇𝑓 = 𝐶 + 𝐶𝑠′ Eq. (4.9a)
Where,
𝑇𝑠 = 𝑅𝑐𝑜𝑟 𝐴𝑠𝑓𝑦 Eq. (4.9b)
𝑇𝑓 = 𝑁𝐴𝑓𝑏𝑓𝑓𝑒 Eq. (4.9c)
𝐶 = 𝛼1𝑓𝑐′𝛽1𝑐𝑢𝑏 Eq. (4.9d)
𝐶𝑠′ = 𝐴𝑠′ 𝐸𝑠휀𝑠
′ Eq. (4.9e)
Where, 𝑅𝑐𝑜𝑟= 1 – average tensile steel mass loss.
Table 4.5 lists the theoretical ultimate loads, 𝑃𝑢𝑡ℎ, for all of the tested specimens. Good agreement
between the experimental and theoretical values was obtained especially for specimens repaired
in Scheme I. However, the capacities of specimens repaired with Scheme II were under-estimated.
The theoretical formulations adopted do not account for the effect of the U-shaped FRCM layers
on the flexural response of the beams. The obtained results suggested the increase of the nominal
capacity, Mn, of FRCM-repaired beams with U-wrapped layers by 10% to account for the scheme
of the FRCM used.
61
4.5.1 Design Provision
According to the provisions of the ACI 549.4R [54], the design flexural strength, 𝑀𝐷, is
calculated in accordance with Equation (4.10). The strength reduction factor, 𝜙𝑚, is given by
Equation (4.11). In addition, the ACI-549 committee limits the increase in the nominal flexural
strength provided by the FRCM reinforcement by 50% of the flexural capacity of the structure
prior to repair. Table 4.5 lists the theoretical design load ϕ𝑚𝑃𝑢𝑡ℎ and the ratio 𝑃𝑢
𝑒𝑥𝑝/ϕ𝑚𝑃𝑢𝑡ℎ. It can
be noticed that applying both the flexural strength reduction factor and the 50% increase limitation
makes the gap between the experimental and design values lager, especially for the specimens
repaired with Scheme II.
𝑀𝐷 = 𝜙𝑚𝑀𝑛 Eq. (4.10)
𝜙𝑚 = {
0.90 for ɛ𝑡 ≥ 0.005
0.65 +0.25(ɛ𝑡−ɛ𝑠𝑦)
0.005−ɛ𝑡−ɛ𝑠𝑦
0.65 for ɛ𝑡 ≤ ɛ𝑠𝑦
for ɛ𝑠𝑦 < ɛ𝑡 < 0.005 Eq. (4.11)
4.6 Conclusions
This study investigated experimentally and analytically the structural performance of corrosion-
damaged RC beams repaired with PBO- and C-FRCM systems. The following conclusions can be
drawn:
• An average mass loss of 12.9% in the tensile steel reduced the yield and the ultimate loads of
the beam by 8% and 5%, respectively. The corroded-unrepaired specimens failed to meet the
provisions of the ACI-318 standards for crack width criteria.
• PBO-FRCM repaired specimen showed slightly higher ultimate load carrying capacities with
more ductile mode of failure than those of C-FRP repaired specimen with similar axial
stiffness and repair scheme.
• Repairing corrosion-damaged RC beams with PBO- and C-FRCM restored 105 to 144% and
130 to 152%, respectively, of the original load-carrying capacity of the virgin uncorroded
beam. The gain in capacity was highly dependent on the number of fabric layers, their
material, and the scheme used.
62
• Beams repaired with PBO-FRCM systems failed in a ductile mode due to either fabric
delamination (repair Scheme I) or fabric slippage within the matrix (repair Scheme II),
whereas beams repaired with U-wrapped C-FRCM systems showed a more brittle failure due
to matrix cracking and complete separation of the fabric.
• Beams repaired with C-FRCM showed higher post-yielding stiffness than that of their PBO-
repaired counterparts. The former beams exhibited a sudden drop after reaching the ultimate
load whereas the later beams showed a gradual decrease after reaching the ultimate.
• Increasing the number of FRCM layers increased the yielding and ultimate loads of the
repaired beams. However, specimens with similar axial stiffness didn’t show similar ultimate
capacities. More tests are required to calibrate the axial stiffness, Kf, to implement parameters
such as the material properties, the fabric architecture, and the repair scheme used.
• U-wrapped FRCM scheme was more efficient than the bottom end-anchored scheme in
increasing the ultimate capacity of the repaired beams. The PBO-repaired beams with scheme
II showed ultimate strengths 15 to 18% more than those repaired with scheme I.
• Beams repaired with PBO-FRCM systems showed a more ductile behavior than their
counterparts repaired with C-FRCM. Most of the PBO-repaired beams restored the original
ductility whereas the C-FRCM repaired beams showed lower ductility than that of the virgin
beam.
• Strain values recorded during the tests indicated that the assumption of prefect bond suggested
by the ACI-549.4R-13 committee [54] is a justifiable simplification for easy implementation
by practicing engineers.
• The theoretical formulations of the ACI-549.4R-13 committee reasonably predicted the
ultimate strengths of the FRCM-repaired beams with Scheme I but underestimated those
repaired with Scheme II. A scheme factor of 1.1 is then proposed while calculating the
nominal strength of beams repaired with U-shaped FRCM.
63
4.7 Notation
The following symbols are used in this paper:
Af = equivalent area of fabric per unit width (mm2/mm)
As = cross-sectional area of tension steel reinforcement (mm2)
𝐴𝑠′ = cross-sectional area of compression steel reinforcement (mm2)
b = width of the beam (mm)
C = compression force provided by concrete (kN)
Cs’ = compression force provided by the compression reinforcement (kN)
cu = distance from extreme compression fiber to neutral axis (mm)
d = distance from top of the beam to the centroid of tension steel (mm)
d’ = distance from top of the beam to the centroid of compression steel (mm)
df = distance from top of the beam to the centroid of fabric reinforcement (mm)
Ef = cracked elastic modulus of the FRCM composite (GPa)
Es = elastic modulus of steel reinforcement (GPa)
Ec = elastic modulus of concrete (MPa)
𝑓𝑐′ = compressive strength of concrete (MPa)
𝑓𝑓𝑒= effective tensile stress in FRCM composite at failure (MPa)
ffu = ultimate tensile strength of FRCM composite (MPa)
fy = yield strength of steel reinforcement (MPa)
MD = design flexural strength (kN-m)
Mf = moment contribution of FRCM reinforcement to flexural strength (kN-m)
Ms = moment contribution of the tensile steel reinforcement to flexural strength (kN-m)
Ms’ = moment contribution of the compression steel reinforcement to flexural strength (kN-m)
Mn = nominal flexural strength (kN-m)
N = number of fabric layers
Rcor = corrosion reduction factor
Ts = tension force in steel reinforcement (kN)
Tf = tension force in FRCM reinforcement (kN)
휀𝑐 = compression strain in concrete (mm/mm)
휀𝑐′ = compression strain of unconfined concrete corresponding to 𝑓𝑐
′ (mm/mm)
64
εcu = concrete strain at ultimate (mm/mm)
휀𝑠′ = tensile strain in compression steel reinforcement (mm/mm)
휀𝑠𝑦 = tensile yield strain of steel reinforcement (mm/mm)
휀𝑡 = the net tensile strain in extreme tensile steel reinforcement at the nominal strength (mm/mm)
휀𝑓𝑑 = FRCM design tensile strain (mm/mm)
휀𝑓𝑒 = effective tensile strain in FRCM composite at failure (mm/mm)
εfu = ultimate tensile strain of FRCM composite (mm/mm)
𝜌𝑓 = fabric reinforcement ratio
𝜅𝑓 = equivalent axial stiffness (MPa)
∆𝐼 = ductility index
β1 = ratio of depth of equivalent rectangular stress block to depth to neutral axis
α1 = multiplier of 𝑓𝑐′ to determine intensity of the equivalent block stress for concrete
𝜙𝑚= strength reduction factor
65
5. Chapter 5
Effect of Corrosion-damage on the Flexural Performance
of RC Beams Strengthened with FRCM Composites
Mohammed Elghazy, Ahmed El Refai, Usama Ebead, and Antonio Nanni
Journal of Composites Structures. Date of acceptance: August 16,2017
( https://doi.org/10.1016/j.compstruct.2017.08.069 )
Résumé
Cet article rend compte du comportement en flexion des poutres en béton armé endommagées
par la corrosion et renforcées par différents systèmes de matrice cimentaire renforcée de fibres
(MCRF). Trois groupes de poutres ont été soumis à une corrosion accélérée pendant 70, 140 et
210 jours pour obtenir une perte de masse théorique dans les barres d'acier de traction de 10%,
20% et 30%, respectivement. Les paramètres d'essai comprenaient le type de fibres (PBO et
carbone), le nombre de couches de MCRF (deux, trois et quatre), et le schéma de renforcement
(couches ancrées aux extrémités et couches continues sous forme U). Les résultats des tests ont
montré que les composites MCRF gouvernaient la défaillance des poutres renforcées plutôt que le
niveau de dommage auquel les poutres étaient soumises en raison de la corrosion. Les résultats des
tests sur les poutres endommagées par la corrosion ont confirmé que la contribution des composites
MCRF compensait de manière significative l'impact de la corrosion sur la résistance. Les poutres
renforcées par MCRF présentaient une augmentation de la résistance qui variait entre 7 et 55% de
celle de la poutre vierge selon le type, la rigidité axiale et le schéma de la MCRF utilisé. Les
poutres renforcées ont montré des indices d'absorption d'énergie qui se situaient entre 111 et 153%
de celui de la poutre vierge. Les formulations théoriques de l'ACI-549.4R-13 ont raisonnablement
prédit les résistances ultimes des poutres renforcées ancrées à l'extrémité, mais ont sous-estimé
celles des poutres renforcées par des couches continues sous forme U.
Mots-clés des auteurs : Corrosion; Matrice cimentaire renforcée de fibres; Flexion; Béton armé;
Réparation; Renforcement.
66
5.1 Abstract
This paper reports on the flexural behavior of corrosion-damaged reinforced concrete (RC)
beams strengthened with different fabric-reinforced cementitious matrix (FRCM) composites.
Three groups of beams were subjected to accelerated corrosion for 70, 140, and 210 days to obtain
theoretical mass loss in their tensile steel bars of 10%, 20%, and 30%, respectively. The test
parameters included the fabric type (PBO and carbon), the number of FRCM layers (two, three,
and four), and the strengthening scheme (end-anchored and continuously wrapped). Test results
showed that FRCM composites governed the failure of the strengthened beams rather than the
damage level to which the beam was subjected due to corrosion. The reported load-carrying
capacities of the corrosion-damaged beams confirmed that the contribution of FRCM composites
significantly offset the impact of corrosion damage on strength. FRCM-strengthened beams
exhibited an increase in strength that ranged between 7 and 55% of that of the virgin beam based
on the type, the axial stiffness, and the scheme of the FRCM used. The strengthened beams showed
energy absorption indices that ranged between 111 and 153% of that of the virgin beam. The
theoretical formulations of ACI-549.4R-13 reasonably predicted the ultimate strengths of the end-
anchored strengthened beams but underestimated those continuously anchored beams.
Authors’ keywords: Corrosion; Fabric-reinforced cementitious matrix; Flexure; Reinforced
concrete; Repair; Strengthening.
5.2 Introduction and Background
Corrosion of steel reinforcing bars is inevitable. Despite the stringent provisions specified by
most of existing building codes, corrosion is still being reported in reinforced concrete (RC)
structures due to the continuous exposure to harsh environments, proximity to sea-shores, and the
use of de-icing salt. The transfer of water, oxygen, and aggressive agents such as carbon dioxide
and chloride into concrete leads to corrosion and consequently to concrete cracking, spalling, and
deterioration. Steel corrosion results in reduction in the cross-sectional areas of the bars and the
loss of bond at the steel/concrete interface. These phenomena significantly reduce the structural
integrity of the RC member and may lead to its premature collapse [44,45,48,53]. As a result,
engineers face a big challenge not only to assess the corrosion-damaged structure but also to select
the appropriate strengthening technique.
67
The superior properties of epoxy-bonded fiber-reinforced polymer (FRP) products and their non-
corrosive characteristics have inspired engineers to adopt them in strengthening applications.
Several studies have reported on the effectiveness of using FRP composites in strengthening
corrosion-damaged RC structures [10,90,91]. Although the epoxy bonding agents used with FRPs
are commonly durable and resistant to the harsh environmental conditions, various problems
associated with their performance at high temperature have been reported [16]. In addition, epoxies
have low compatibility with the concrete substrate [92] and can’t be applied on wet surfaces or at
low temperatures. In order to overcome such drawbacks, fabric-reinforced cementitious matrix
(FRCM) systems were introduced as promising alternatives.
FRCM composites consist of one or more fabric mesh made of long dry-woven embedded in a
cement-based matrix that serves as a binder. The fabric may be made of carbon (C), glass (G), or
Polyparaphenylene benzobisoxazole (PBO) while the matrix is an inorganic hydraulic or non-
hydraulic cementitious mortar that holds in place the reinforcing meshes. Many studies have
demonstrated the effectiveness of FRCM composites in enhancing the flexural and shear
performances of damage-free RC beams [57,87,99]. Babaeidarabad et al. (2014) [84] investigated
the flexural performance of RC beams strengthened with PBO-FRCM. The results showed that the
strength gain ranged from 13 to 93% of that of the unstrengthened beams depending on the amount
of FRCM layers used. Fabric slippage within the matrix and FRCM delamination were reported
as two distinct modes of failure. Similar results were reported by Loreto et al. (2013) [85].
Schladitz et al. (2012) [55] also reported that increasing the volume fraction of the fabric used in
strengthening RC slabs increased their load-carrying capacities. However, this gain in strength was
accompanied by a decrease in the ductility of the strengthened slabs. D’Ambrisi and Focacci
(2011) [81] demonstrated that the strengthening effectiveness of FRCM composites was highly
dependent on the type of fabric and the bond between the matrix and the fabric.
The use of FRCM in strengthening corrosion-damaged RC members have been rarely reported.
The challenge in strengthening such members arises from the potential loss of integrity of the
member following the corrosion of its reinforcement. It also arises from the potential loss of bond
between the deteriorated concrete and the new mortar applied. The study conducted by El-
Maaddawy and El Refai (2016) [86] evidenced the feasibility of using basalt and carbon FRCM
systems to strengthen severely corrosion-damaged RC beams that suffered 22% mass loss of their
68
steel bars. It was reported that basalt-FRCM system could not restore the original flexural capacity
of the beams whereas the C-FRCM system restored 109% of the capacity. El-Maaddawy and El
Refai (2016) [86] also reported that the beams strengthened with a combination of internally-
embedded and externally-bonded C-FRCM layers have restored both their strength and ductility.
No results were reported on the effect of the FRCM strengthening scheme or the degree of
corrosion damage on the flexural behavior of the strengthened beams.
The present work is part of a large experimental program that aims at filling the gap in knowledge
on the effectiveness of FRCM composites in strengthening corrosion-damaged RC structures. The
tested beams presented herein were subjected to three levels of corrosion damage prior to
strengthening. FRCM composites having different amounts of fibers, different mechanical
properties, and different schemes were used to strengthen the corrosion-damaged beams. The test
results report on the gain in the yield strengths, the ultimate strengths, and the ductility of the
strengthened beams. The influence of the damage degrees, in addition to other parameters, on the
flexural performance of the strengthened beams is presented and discussed.
5.3 Experimental Program
The experimental program is summarized in Table 5.1. Data of some of the tested beams were
previously reported in Elghazy et al. [100] and are presented herein for comparison purposes. The
test parameters included the corrosion level, the strengthening scheme, and the number of FRCM
layers used. The beams were subdivided into three groups (A, B, and C) and were subjected to
accelerated corrosion process for 70, 140, and 210 days, respectively. At the end of the corrosion
process, one beam in each group was not strengthened and was used as a benchmark while other
beams were strengthened with the designated FRCM systems and configurations. In addition, two
virgin beams (i.e. not corroded nor strengthened) were used as controls.
The beams were labeled following the X-Y-Z format. ‘X’ represents the beam condition (UU,
CU, and CS referring to Uncorroded-Unstrengthened, Corroded-Unstrengthened, and Corroded-
Strengthened, respectively) and the beam’s group (A, B, and C). ‘Y’ denotes the number and type
of the FRCM layers applied (2P, 4P, and 3C referring to two layers of PBO-FRCM, four layers of
PBO-FRCM, and three layers of C-FRCM, respectively). Finally, ‘Z’ describes the FRCM
strengthening schemes (I and II) as will be detailed in the following sections.
69
5.3.1 Test Specimen and Materials
Figure 5.1 shows the beam geometry and the reinforcement details. All beams were designed
according to ACI 318-14 provisions [98]. All beams were 2.8 m long with 150×250 mm
rectangular cross section. The bottom and top reinforcement consisted of two 15M (diameter 15
mm) and 8M (diameter 8 mm) deformed bars, respectively. The shear spans were reinforced with
10M (diameter 10 mm) deformed stirrups spaced at 100 mm to avoid shear failure. To accelerate
the corrosion process of the bottom steel bars, a hollow stainless-steel tube with an external
diameter of 9.5 mm and a wall thickness of 2.5 mm was placed at 80 mm from the beam soffit to
act as cathode. Salt (NaCl) weighed as 5% of the cement weight was added to the concrete mix
used to cast the middle-bottom of the corroded beams with a height of 100 mm (Figure 5.1). Details
about the accelerated corrosion process are presented in the following section.
Six standard concrete cylinders (150×300 mm) were prepared from the normal and salted
concrete. Table 5.2 lists the compressive strengths of both mixes after 28 days and on the day of
testing. Prior to FRCM application, the corroded beams were repaired using a commercial
cementitious repair mortar (Sikacrete-08SCC) having a compressive strength of 55.4 MPa
(standard deviation of 5 MPa) and a flexural strength of 3.4 MPa (standard deviation of 0.3 MPa)
as tested by the authors. The yield strengths of the longitudinal reinforcing steel bars of diameter
15 and 8 mm were 466 MPa (with a standard deviation of 4.2 MPa) and 573 MPa (with a standard
deviation of 17.7 MPa), respectively, as tested by the authors.
70
Table 5.1: Summary of the test results
*Average values reported **
SY-CC = Steel Yielding followed by Concrete Crushing; FD = FRCM Delamination; FS-PED = Fabric Slippage followed by Partial Debonding; MC-FS =
Matrix Cracking followed by Fabric Slippage
Specimen wavg
(mm)
Avg. Mass
loss (%)
𝜌𝑆
(%)
𝜌𝑓
(%)
𝐾𝑓
(MPa)
𝛽𝑓
(%)
𝑃𝑦
(kN)
𝑃𝑢
(kN) 𝑃𝑦 𝑁𝑜𝑟𝑚 𝑃𝑢 𝑁𝑜𝑟𝑚
𝑃𝑢𝑡ℎ
(kN)
𝑃𝑢
𝑃𝑢𝑡ℎ
εfu
(µϵ)
Mode of
failure**
Virgin beams
UUa, UUb* - - 1.07 - - - 75.1 79.7 1 1 81.9 0.97 - SY-CC
Group A: Theoretical mass loss of 10%
CUA 0.65 12.9 0.93 - - - 69.5 76.1 0.93 0.95 72.31 1.05 - SY-CC
CSA-2P-I 0.72 12.6 0.93 0.04 48 2.58 79.5 86.4 1.06 1.08 83.64 1.03 7743 FD
CSA-4P-I 0.75 12.6 0.93 0.08 95.3 5.11 83.3 99.6 1.11 1.25 94.56 1.05 9442 FD
CSA-4P-II 0.68 12.3 0.94 0.08 95.3 5.1 91.3 114.4 1.22 1.44 100.1 1.14 9526 FS-PFD
CSA-3C-II 0.58 12.1 0.94 0.185 139.1 7.41 87 120.6 1.16 1.51 104.78 1.15 6000 MC-FS
Group B: Theoretical mass loss of 20%
CUB 1 18 0.88 - - 64.5 74.2 0.86 0.93 68.27 1.09 - SY-CC
CSB-2P-I 1.1 19.6 0.86 0.04 48 2.8 71.8 85.6 0.96 1.07 78.2 1.09 8180 FD
CSB-4P-I 1.15 19.4 0.86 0.08 95.3 5.54 79.6 102.6 1.06 1.29 89.38 1.15 10659 FD
CSB-4P-II 0.95 19.5 0.86 0.08 95.3 5.55 80.7 102.9 1.07 1.29 94.6 1.09 8253 FS-PFD
CSB-3C-II 1.15 18.6 0.87 0.185 139.1 8.01 78.8 123.3 1.05 1.55 99.95 1.23 5530 MC-FS
Group C: Theoretical mass loss of 30%
CUC 1.65 22.5 0.83 - - - 64.1 72.2 0.85 0.91 64.68 1.12 - SY-CC
CSC-4P-I 2 22.7 0.83 0.08 95.3 5.78 82.5 102.8 1.1 1.29 86.85 1.18 7339 FD
CSC-4P-II 1.6 21.1 0.84 0.08 95.3 5.66 80.4 111.1 1.1 1.39 93.1 1.19 9653 FS-PFD
CSC-3C-II 1.5 21.5 0.84 0.185 139.1 8.31 75.2 109.3 1 1.37 97.77 1.12 4885 MC-FS
71
Figure 5.1: Test specimen geometry and reinforcement details. (All dimensions in mm)
Table 5.2: Concrete compressive strengths
Compressive strength
(MPa)
Standard deviation
(MPa)
Coefficient of variation
(%)
28-day Normal concrete 37.9 0.8 2
Salted concrete 33.5 1.1 3.2
Testing day Normal concrete 41.8 4.8 11.4
Salted concrete 41.2 0.6 1.6
5.3.2 Accelerated Corrosion Process
A DC galvanostatic power supply was used to impress a constant electrical current of 380
milliamps (mA) with an approximate density of 180 µA/cm2. The density level was chosen less
than 200 µA/cm2 to represent natural corrosion encountered in the field and to avoid the bond loss
at the steel/concrete interface as recommended in [20]. All beams were connected in series in a
large environmental chamber as shown in Figure 5.2. Therefore, the reinforcing bars acted as
anodes, the stainless-steel bars acted as cathodes, and the salted concrete acted as electrolyte.
During the corrosion process, the beams were subjected to consecutive wet-dry cycles that
consisted of 3 days wet followed by 3 days dry each.
72
Figure 5.2: Specimens inside the environmental chamber during a dry cycle
The theoretical mass loss of the steel bars due to corrosion was calculated using Faraday’s law
that relates the mass loss to the electrical current and the exposure duration as follows:
𝑚 =𝐼𝑡𝑎
𝑛𝐹 Eq. (5.1)
where m = mass loss (in grams), I = impressed current (in Ampere), t = corrosion duration (in
seconds), a = the atomic mass of iron (55.847grams), n = the number of electrons transferred
during the corrosion reaction (n = 2 for iron), and F = Faraday’s constant (96.500 C/equivalent).
As previously mentioned, three mass losses were considered in this study namely, 10% for beams
of group A, 20% for beams of group B, and 30% for beams of group C, which represented
moderate, severe, and very severe degrees of corrosion damage, respectively.
5.3.3 FRCM Systems
In this study, two commercially available FRCM systems (PBO-FRCM and C-FRCM) were used
to strengthen the corrosion-damaged beams. The PBO fabric consisted of an unbalanced net of
spaced fiber rovings running along two orthogonal directions as shown in Figure 5.3a. The
associated inorganic cementitious matrix had a compressive strength of 43.9 MPa (standard
deviation of 0.4 MPa) and a flexural strength of 3 MPa (standard deviation of 0.3 MPa) as
determined by the authors. The C-FRCM composite consisted of unidirectional net made of
carbon-fiber strands that were oriented in one direction as shown in Figure 5.3b. The associated
matrix had a compressive strength of 42.1 MPa (standard deviation of 4.3 MPa) and a flexural
73
strength of 3.2 MPa (standard deviation of 0.3 MPa). As noted in Table 5.3, the area per unit width
of the carbon fabric was a little more than three times that of the PBO fabric.
Figure 5.3: FRCM systems: a) PBO-FRCM (Unbalanced PBO fabric) and b) C-FRCM
(Unidirectional carbon fabric)
In a previous study conducted by Ebead et al. [97], direct tensile tests were performed on PBO-
and C-FRCM coupons having one layer of embedded fabric in order to characterize their
mechanical properties. The tests were conducted according to AC-434 provisions [65]. Test results
indicated that the stress-strain relationship of FRCM consisted of three distinct stages as shown in
Figure 5.4. In the uncracked stage, the PBO-FRCM coupons showed higher initial stiffness than
their C-FRCM counterparts. After cracking (second stage), the fabric transferred the load back to
the matrix until the tensile strength of the fabric was reached at ultimate (third stage). According
to ACI-549.4R-13 provisions [54], the tensile behavior of each FRCM system could be
characterized by its cracked tensile modulus of elasticity, Ef, its tensile strength, ffu, and its ultimate
strain, εfu. Table 5.4 lists the results obtained from the direct tensile tests for both FRCM systems
that were used in this study [97].
74
Table 5.3: Fabric properties as given in the manufacturers’ data sheets
Fabric Area per unit
width
(𝐴𝑓) (mm2/m)
Tensile
strength (GPa)
Elastic
modulus
(GPa)
Ultimate
strain
(%)
PBO (primary direction) 50 5.8 270 2.15
PBO (secondary direction) 15 5.8 270 2.15
Carbon 157 4.3 240 1.75
Figure 5.4: Stress-strain relationships for FRCM-tensile coupons [97]
Table 5.4 : Mechanical properties of FRCM systems [97]
FRCM system Cracked tensile modulus
of elasticity, Ef (GPa)
Ultimate tensile
strength, ffu (GPa)
Ultimate strain,
εfu (%)
PBO-FRCM 121 1.55 1.4
C-FRCM 75 0.97 1.25
75
5.3.4 Strengthening Schemes
The two strengthening schemes used in this study are illustrated in the schematics shown in
Figure 5.5a and Figure 5.5b. Scheme I consisted of one or more FRCM flexure plies having 150
mm width (equal to the beam width) and applied to the soffit of the beam over the middle 2.4 m.
The fabric was oriented so that its primary direction was parallel to the longitudinal axis of the
beam. The flexural plies were anchored at each end using one U-shaped transverse strip of 300
mm width as shown in Figure 5.5a. Scheme II consisted of one or more flexural plies similar to
those used in Scheme I but enclosed with one U-shaped continuous ply running along the clear
span of the beam (Figure 5.5b). For the PBO-FRCM composite used in Scheme II, the primary
direction of the U-wrapped ply was oriented parallel to the longitudinal axis of the beams and was
counted as an additional flexural ply. However, while the bottom flexural plies of the C-FRCM in
Scheme II were oriented parallel to the longitudinal axis of the beams, the U-shaped continuous
ply was oriented in the transverse direction and therefore did not contribute to the flexural
resistance (Figure 5.5b).
For comparison purposes, the equivalent axial stiffness coefficient, 𝐾𝑓, for each FRCM system
was determined based on their cracked modulus and the cross-sectional area of the fabric
embedded within the FRCM composite as follows:
𝐾𝑓= 𝜌𝑓𝐸𝑓 Eq. (5.2)
𝜌𝑓 = 𝑁 𝐴𝑓
𝑑𝑓 Eq. (5.3)
where N is number of the fabric layers in the matrix (shown in each beam’s label in Table 5.1),
b is beam width (150 mm), Af is the fabric area per unit width (Table 5.3), df is the effective depth
of the fabric, Ef is the cracked modulus of the FRCM composite in N/mm2 (Table 5.4), and 𝜌𝑓 is
the fiber reinforcement ratio. Values of 𝜌𝑓 and 𝐾𝑓 are listed in Table 5.1 for all the tested beams.
It is important to note that for beams strengthened with the continuously wrapped PBO-FRCM
layer (Scheme II), the fibers located on the lateral sides of the beams were neglected in estimating
𝜌𝑓 (and consequently 𝐾𝑓). However, their contribution to the flexural strengths of the beams was
considered in the analysis as will be detailed later.
76
The ratio of contribution of FRCM composites to that of steel reinforcement was expressed by
the stiffness factor, 𝛽𝑓, as given in Equation (5.4).
𝛽𝑓 =𝜌𝑓𝐸𝑓
𝜌𝑠𝐸𝑠=
𝐾𝑓
𝐾𝑠 Eq. (5.3)
Where, 𝜌𝑠 is the tensile steel reinforcement ratio and Es is the steel elastic modulus. The steel
reinforcement ratio, 𝜌𝑠, of the corrosion-damaged beams was determined after considering the
actual average steel mass loss due to corrosion. The computed values of 𝜌𝑠 and 𝛽𝑓 are listed in
Table 5.1 for each of the tested beams.
Figure 5.5: Strengthening schemes: a) Scheme I and b) Scheme II
77
5.3.5 FRCM Installation
Prior to FRCM application, all layers of unsound deteriorated concrete were removed as shown
in Figure 5.6a. The corroded steel bars were then cleaned, and a commercial repair mortar was
used to repair the damaged zone. After curing, the beam surface was sandblasted as shown in
Figure 5.6b. The hand lay-up method recommended by ACI-549.4R-13 provisions [54] and the
FRCM manufacturers was adopted during installation. The beam substrate was first soaked in
water for 2 hours before applying the first layer of the cementitious matrix with a thickness of 3 to
5 mm. The fabric was then installed in place and was gently impregnated into the cementitious
matrix. A second layer of matrix having the same thickness was immediately applied as shown in
Figure 5.6c. The procedure was then repeated until the specified number of layers was achieved.
All strengthened beams were left for curing for 28 days in laboratory conditions before being
tested.
Figure 5.6: FRCM installation procedure: a) removing the deteriorated concrete, b) patch
repairing and sandblasting, and c) installation of PBO-FRCM composite
5.3.6 Instrumentation and Test Setup
All beams were instrumented at mid-span with 60 mm long strain gauges on the top concrete
surface and with 5 mm strain gauges bonded to the tensile steel bars. The FRCM-strengthened
78
beams were instrumented with 5 mm strain gauges installed directly on the outer fabric layer of
the FRCM composite at mid-span and at the loading points.
The beams were tested to failure in a four-point bending configuration as shown in Figure 5.1.
All tests were conducted under displacement control at a rate of 2 mm/minute using a MTS
actuator. Deflections at midspan and under the loading points were monitored using three linear
variable differential transducers (LVDTs). All gauges and LVDTs were connected to a 20-channel
data acquisition system that captured the readings at a rate of 5 readings/sec.
5.4 Test Results
The following sections report on the corrosion observations and the test results of both
unstrengthened and strengthened beams in terms of their modes of failure, strength response,
strains in FRCM, and ductility. This is followed by a discussion on the effect of corrosion damage
on the flexural performance of the tested beams.
5.4.1 Corrosion Observations
At the end of the corrosion process, rust stains and longitudinal cracks running parallel to the
corroded steel bars were observed. Corrosion-induced cracks were also observed at the bottom of
the beams and/or on their sides at the level of the tensile steel bars. Table 5.1 lists the average
corrosion crack width, wavg, for each beam. The maximum crack widths were 1.5, 2.8, and 3.5 mm
for groups A, B, and C, respectively. Wider cracks indicated higher steel mass loss due to
corrosion. All the corrosion-damaged beams did not meet the service requirements of ACI 318-14
[98] that limits the maximum crack width to 0.40 mm.
After testing the beams, the corroded steel bars were carefully removed and cut into coupons of
200 mm length each. Visual inspection of the corroded bars revealed the existence of several
corrosion pits that were randomly dispersed on the surface of the bars within the corrosion zone
(Figure 5.7). The mass loss of the corroded bars were determined for each beam according to
ASTM G1-03 provisions [26]. Table 5.1 lists the actual steel mass loss for each beam. Average
mass losses of 12.5, 19, and 22% were determined for groups A, B, and C, respectively,
corresponding to theoretical mass losses of 10, 20, and 30% as predicted from Equation (5.1). The
discrepancy between the theoretical and actual mass losses in the steel bars is shown in Figure 5.8.
79
As reported in [101], this discrepancy was attributed to the reduction in the rate of rust production
as corrosion progresses.
Figure 5.7: Profile of steel bars: a) uncorroded bar, b) corroded bar extracted from CSA-4P-I
(12.6% mass loss), c) corroded bar extracted from CSB-3C-II (18.6% mass loss), and d)
corroded bar extracted from CUC (22.5% mass loss)
Figure 5.8: Actual and theoretical mass loss versus the duration of corrosion process
a)
b)
c)
d)
80
5.4.2 Failure Mechanisms
The failure mechanisms of all the tested beams are summarized in Table 5.1. The virgin beams
(UU) and the corroded unstrengthened beams (CUA, CUB, and CUC) showed classical modes of
failure of under-reinforced beams in which large flexural cracks appeared within the constant
moment zone after steel yielding followed by concrete crushing. On the other hand, three distinct
failure mechanisms were observed in the strengthened beams as shown in Figure 5.9 and were
described as follows:
1) FRCM delamination (FD): this failure mode was encountered in all beams strengthened with
PBO-FRCM systems in Scheme I. The delamination occurred at the fabric/matrix interface
adjacent to the concrete substrate due to the propagation of flexural cracks. This mode of failure
is shown in Figure 5.9a for beam CSA-4P-I.
2) Fabric slippage with partial fabric debonding within the matrix (FS-PFD): this failure mode
was observed in all beams strengthened with PBO-FRCM systems in Scheme II. Vertical flexural
cracks were first observed in the FRCM matrix in the moment zone while inclined cracks extended
in the shear spans. As the applied load increased, gradual fabric slippage within the matrix was
observed until failure occurred. At failure, partial debonding was noticed in the U-shaped ply at
locations where several cracks formed as shown in Figure 5.9b for beam CSC-4P-II. As will be
detailed later, this mode of failure resulted in a more ductile behavior than that observed when the
first mode of failure occurred. This was attributed to the wrapping effect of the continuous U-
shaped FRCM layer that delayed the delamination of the bottom flexural plies.
3) Matrix cracking with extensive fabric slippage (MC-FS): this mode of failure was observed in
all beams strengthened with C-FRCM in Scheme II. Vertical flexural cracks formed in the
cementitious matrix of the U-shaped FRCM layer after steel yielding. As the load increased,
progressive cracking accompanied with large fabric slippage was observed at the beam’s soffit
(Figure 5.9c for beam CSA-3C-II). It should be noticed that this mode of failure was more brittle
than that reported in the PBO-strengthened beams.
81
Figure 5.9: Modes of failure: a) FRCM delamination, b) fabric slippage with partial fabric
debonding within the matrix, and c) matrix cracking with extensive fabric slippage
5.4.3 Strength Response
The load-deflection relationships for the corroded-unstrengthened (CU) beams are shown in
Figure 5.10. Corrosion of steel bars slightly affected the flexure response of the beams with slightly
notable impact on the beams’ stiffness. Due to corrosion, the yield and ultimate loads of beams of
groups A, B, and C were reduced by 7 and 5%, 14 and 7%, and 15 and 9%, respectively, of the
corresponding values of the virgin beams. The yield and ultimate strengths of the corrosion-
damaged beams decreased at average rates of 0.66 and 0.40%, respectively, per 1% of mass loss.
The insignificant effect of corrosion of the tensile steel bars on the beam’s strength could be
attributed to a transition in the beam behavior from a pure flexural action to a tied-arch action due
to bond deterioration within the flexure (corroded) zone while remaining anchored in the shear
(uncorroded) zones. It might also be attributed to the fact that corroded bars lost their lugs, which
increased the measured mass loss without affecting their effective cross section area as shown in
Figure 5.7.
82
Figure 5.10: Load-deflection relationships for corroded-unstrengthened beams
Figure 5.11 shows the load versus the midspan deflection curves for the FRCM-strengthened
beams. All the strengthened beams of groups A, B, and C showed similar flexural response. The
load-deflection curves consisted of three stages with two turning points indicating the concrete
cracking in tension and the yielding of the steel bars. FRCM systems slightly enhanced the stiffness
of the strengthened beams prior to steel yielding, which indicated their marginal contribution in
the beam’s behavior. However, the post-yielding response was significantly dependent upon the
number and the type of the FRCM plies used.
(a)
83
(b)
(c)
Figure 5.11: Load-deflection relationships for corrosion-damaged FRCM-strengthened beams: a)
beams of Group A, b) beams of Group B, and c) beams of Group C
The yield load, 𝑃𝑦, ultimate load, 𝑃𝑢, and normalized yield and ultimate loads to those of the
control specimen, 𝑃𝑦 𝑁𝑜𝑟 and 𝑃𝑢 𝑁𝑜𝑟, respectively, are presented in Table 5.1. It can be noticed that
all the FRCM-strengthened beams fully restored the yield and ultimate strengths of the virgin beam
except beam CSB-2P-I that restored only 96% of the yield capacity. The gain in strengths was
highly dependent upon the amount of fabric, the FRCM strengthening scheme, and the type of the
84
FRCM used. The use of two layers of PBO-FRCM in beams CSA-2P-I and CSB-2P-I increased
their ultimate strengths by 8 and 7%, respectively, of that of the virgin beam while using four PBO-
FRCM plies in beams CSA-4P-I, CSB-4P-I, and CSC-4P-I increased their ultimate strengths by
25, 29, and 29%, respectively.
The use of a U-shaped fabric layer in Scheme II was more efficient in increasing the load-carrying
capacities of the strengthened beams than the use of end anchors in Scheme I. For instance, the
use of four plies of PBO-FRCM with Scheme II in beams CSA-4P-II, CSB-4P-II, and CSC-4P-II
increased their ultimate strengths by 44, 29, and 39% compared to 25, 29, and 29% increase in
beams CSA-4P-I, CSB-4P-I, and CSC-4P-I, respectively. This increase in strength was associated
with the contribution of the longitudinal PBO fibers attached to the lateral surfaces of the beams.
Moreover, wrapping the beam with a continuous U-shaped PBO layer delayed the premature
delamination of the fabric and consequently increased the strengthening effectiveness of the
FRCM system.
The use of three layers of C-FRCM in Scheme II increased the load-carrying capacities of beams
CSA-3C-II, CSB-3C-II, and CSC-3C-II by 51, 55, and 37%, respectively. It was obvious from the
test results that, despite the various levels of corrosion, the difference in gain in strengths observed
in most of the tested beams strengthened with similar FRCM systems was marginal. This finding
will be discussed in the following sections.
The equivalent axial stiffness of the FRCM composites, Kf, was utilized to compare between the
load-carrying capacities of the strengthened beams (normalized with respect to that of the virgin
beam UU). These results are shown in Figure 5.12. It can be noticed that almost doubling the
equivalent axial stiffness of PBO-FRCM systems from 48 MPa (2 plies in beam CSA-2P-I) to 95.3
MPa (4 plies in beam CSA-4P-I) increased the gain in the load-carrying capacities from 10.3 kN
(for beam CSA-2P-I) to 23.5 kN (for beam CSA-4P-I). A similar trend was observed for specimens
CSB-2P-I and CSB-4P-I. This finding revealed that the gain in ultimate strengths was
approximately proportional to the axial stiffness of the PBO-FRCM systems represented by the
coefficient Kf. This finding was also confirmed by the similar ultimate fiber strains recorded for
specimens CSA-2P-I, CSA-4P-I, CSB-2P-I, and CSB-4P-I (Figure 5.13a).
85
Figure 5.12: Normalized strength versus the FRCM equivalent axial stiffness, Kf
On the other hand, the beams strengthened with three layers of C-FRCM in Scheme II (beams
CSA-3C-II, CSB-3C-II, and CSC-3C-II) showed average load-carrying capacities only 7.5%
higher than those of their counterparts strengthened with four layers of PBO-FRCM in the same
scheme whereas the equivalent axial stiffness of three layers of C-FRCM was 146% higher than
that of four layers of PBO-FRCM. This finding was attributed to the premature cracking of the
matrix associated with the C-FRCM system, which limited the strengthening effectiveness of the
system. This was also confirmed by the strain results shown in Figure 5.13b. The ultimate strains
measured in the fibers of beams strengthened with C-FRCM were obviously lower than those
measured in the fibers of beams strengthened with PBO-FRCM.
FRCM configuration had a notable effect on the flexural response of the beams. Beams
strengthened with four plies of PBO-FRCM systems using Scheme II (beams CSA-4P-II, CSB-
4P-II, and CSC-4P-II) showed an increase in their ultimate strengths that ranged between 29 and
44% with an average gain of about 37%. Their counterpart beams CSA-4P-I, CSB-4P-I, and CSC-
4P-I strengthened with the same amount of FRCM layers (similar FRCM equivalent stiffness) but
in Scheme I showed an average increase in their strengths of about 28%. As previously noted, this
increase in strength was associated with the contribution of the longitudinal fibers attached to the
lateral surfaces of the beams and with the wrapping effect of the U-shaped fabric on delaying the
delamination of the PBO-FRCM fabrics.
86
5.4.4 Fabric Strains
Table 5.1 lists the strains recorded in the outer fabric layer at ultimate load, εfu, for all the
strengthened beams. Figure 5.13a and 5.13b shows the fabric strains in the beams strengthened in
Scheme I and II, respectively.
(a)
(b)
Figure 5.13: Load versus outer fabric strain for beams strengthened in a) Scheme I and b)
Scheme II
87
It is important to note that the ultimate fiber strains measured during the tests were lower than
those obtained from the direct tensile tests conducted on FRCM coupons listed in Table 5.4.
Measured strains ranged between 4885 μɛ for beam CSC-3C-II and 10659 μɛ for beam CSB-4P-I
while the tested coupons showed strains of 12500 μɛ for C-FRCM and 14000 μɛ for PBO-FRCM
[97]. This discrepancy in the strain results between the flexural and tensile specimens was
attributed to the different failure mechanisms observed in both the strengthened beams and the
coupons. While three distinct modes of failure were encountered in the strengthened beams, all the
tested FRCM coupons failed due to fabric slippage in tension.
5.4.5 Ductility and Energy Absorption
The ductility index, ΔI, was determined for all beams as the ratio of deflections at ultimate, 𝛿𝑢,
to the deflection at yielding, 𝛿𝑦. The energy absorption index, ψ, represented by the area under the
load-deflection curve up to the ultimate point was also determined. Both indices were utilized to
evaluate the ductility of the tested beams. Table 5.5 lists the deflections, the ductility indices, and
the energy absorption indices for each beam. It was noticed that the corrosion-damaged beams
showed higher ductility indices than that of the virgin beams. The ductility index increased with
the increase of the corrosion level. The unstrengthened beams of groups A, B, and C (average mass
loss of 12.9, 18, and 22.5%, respectively) showed ductility indices of 3.2, 3.3, and 3.4 representing
114, 118, and 121%, respectively, of those of the virgin beams. The same beams showed energy
absorption indices of 118, 104, and 88% of that of the virgin beams, respectively. On the other
hand, beams strengthened with FRCM composites showed ductility indices that ranged between
86 to 118% of that of the virgin beams whereas their energy absorption indices improved by 11 to
55%.
Figure 5.14 and Figure 5.15 show the change in the normalized indices with the stiffness factor
𝛽𝑓. Recall that 𝛽𝑓 represents the ratio of the axial stiffness of the FRCM system to that of the steel
reinforcement and therefore represents the contribution of the FRCM composites to the beam’s
performance. Beams strengthened with two layers of PBO-FRCM in Scheme I showed average
ductility and energy absorption indices of 103.5 and 116.5 % of those of the virgin beam,
respectively, whereas beams strengthened with four layers of PBO-FRCM showed average indices
of 100 and 134.3%, respectively. For beams strengthened with four layers of PBO-FRCM in
Scheme II, the average indices were 93 and 130% of those of the virgin beam, respectively. On
88
the other hand, beams strengthened with three layers of C-FRCM had similar average indices to
those of their counterparts strengthened with four layers of PBO-FRCM and Scheme II.
Table 5.5: Ductility and energy absorption of the tested beams
Specimen
Midspan
deflection, mm Ductility index Energy absorption index
δy δu ΔI ΔINor.** ψ ψ Nor.
**
UUa, UUb* 11.7 32.9 2.81 1 1972 1
CUA 10.9 35.4 3.2 1.14 2326 1.18
CSA-2P-I 11.83 33.02 2.8 1 2314 1.17
CSA-4P-I 12.86 31.48 2.4 0.86 2346 1.19
CSA-4P-II 12.64 35.17 2.8 1 3059 1.55
CSA-3C-II 12.55 30.1 2.4 0.86 2491 1.26
CUB 8.6 28.26 3.3 1.18 2059 1.04
CSB-2P-I 11.28 33.72 3 1.07 2292 1.16
CSB-4P-I 12.54 37.41 3 1.07 2933 1.49
CSB-4P-II 11.83 29.48 2.5 0.89 2184 1.11
CSB-3C-II 10.57 34.75 3.3 1.18 3012 1.53
CUC 9.8 32.85 3.4 1.21 1735 0.88
CSC-4P-I 12.04 35.53 3 1.07 2657 1.35
CSC-4P-II 12.69 31.32 2.5 0.89 2426 1.23
CSC-3C-II 12.32 30.16 2.4 0.86 2212 1.12
* Average values reported **Normalized with respect to the virgin beam
89
Figure 5.14: Normalized ductility index versus stiffness factor 𝛽𝑓 %
Figure 5.15: Normalized energy absorption index versus stiffness factor 𝛽𝑓 %
5.5 Discussion
The effect of corrosion damage on the flexural response of the strengthened beams is discussed
in the light of the presented test results. As previously mentioned, almost all the strengthened
beams restored the yield and ultimate capacities of the virgin beams. Figure 5.16 shows the effect
of corrosion level on the load-carrying capacities of the strengthened beams.
Figure 5.16 reveals that beams strengthened with four PBO-FRCM layers in Scheme I (beams
CSA-4P-I, CSB-4P-I, and CSC-4P-I) showed almost similar capacities despite the different
90
degrees of corrosion in each group of beams. The comparison between the ultimate capacity of
beam CSA-4P-I that suffered 12.6% mass loss due to corrosion and that of beam CSC-4P-I that
suffered almost double of that mass loss (21.1%) revealed the negligible effect of corrosion on the
load-carrying capacities of the strengthened beams. The former beam (beam CSA-4P-I) showed a
load-carrying capacity of 99.6 kN whereas the latter one (beam CSC-4P-I) showed a load-carrying
capacity of 102.8 KN, with a variation in capacity gain not exceeding 3%. This finding was also
confirmed from the strength results of the tested beams having different degrees of corrosion
damage and strengthened with four plies of PBO-FRCM and with three plies of C-FRCM in
Scheme II. The variation in gain in capacities did not exceed 7 and 9% for the former and the latter
beams, respectively.
Figure 5.16: Effect of corrosion damage on the ultimate strength of strengthened beams
The insignificant impact of corrosion levels on the load-carrying capacities of the strengthened
beams was previously confirmed with the capacities obtained for the unstrengthened beams after
corrosion. Beams CUA, CUB, and CUC showed almost similar capacities ranging between 91 and
95% of the capacity of the virgin beams. It was also attributed to the same mode of failure
encountered in all beams that were strengthened with the same amount and same configuration of
FRCM composites. Based on test observations, the three modes of failure reported for the
strengthened beams showed that failure was primarily governed by the FRCM scheme and the
type of the FRCM composite rather than the damage level to which the beam was subjected due
to corrosion. Knowing that the maximum mass loss in the steel reinforcing bars was 22.7%, which
is quite significant, the reported load-carrying capacities of the corrosion-damaged beams
91
confirmed that strengthening with FRCM composites had counterbalanced the impact of the
corrosion damage.
Figure 5.17 illustrates the effect of corrosion damage on the load-carrying capacities of the
strengthened beams normalized with respect to those of the virgin beams. Recall that the degree
of corrosion damage in the steel reinforcing bars was represented by a reduction in the axial
stiffness 𝐾s = 𝜌𝑠𝐸𝑠. The contribution of FRCM composites to the beam strength was compared to
that of steel reinforcement using the stiffness factor, 𝛽𝑓 = 𝐾𝑓
𝐾𝑠.
Figure 5.17: Normalized strength versus stiffness factor 𝛽𝑓 %
In Figure 5.17, a best fit curve was plotted to address the variation in the ultimate capacities with
the factor 𝛽𝑓. A linear trend was obtained with a least-square coefficient 𝑅2 = 0.75. As can be seen
in Figure 5.17, a tendency of increase in the ultimate capacities is observed with the increase in
the stiffness factor, 𝛽𝑓, from 2.58% (for beam CSA-2P-I) to 8.31% (for beam CSC-3C-II). Based
on the test results, it can be concluded that the increase in 𝛽𝑓 values was largely affected by the
increase in the equivalent stiffness 𝐾𝑓 rather than the decrease in the equivalent stiffness 𝐾𝑠 that
resulted from the mass loss in the steel bars, which can explain the ascending trend in Figure 5.17.
92
5.6 Predicted Strength Results
Analytical calculations were conducted according to ACI 318-14 [98] and ACI 549.4R-13 [54]
provisions to predict the flexural response of the tested beams. The FRCM composites were
assumed to behave linearly up to failure while the steel bars were assumed to behave elastic-
perfectly plastic. The yield strengths of the longitudinal reinforcing steel bars of diameter 15 and
8 mm were taken equal to 466 and 573 MPa, respectively, with elastic modulus 𝐸𝑠 = 200 GPa. The
maximum compression strain in concrete was taken equal to 0.003 mm/mm. Prefect bond between
the concrete substrate and the FRCM system and between the FRCM fabric and the matrix was
assumed during the analysis. However, this assumption was governed by the strain limit of 0.012
mm/mm imposed by ACI 549.4R-13 provisions [54]. Therefore, the design effective tensile strain,
εfe, in FRCM was assumed equal to the experimental ultimate strain, εfu, as obtained from the
coupon test results minus one standard deviation or 0.012 mm/mm, whichever was lower. The
design effective tensile strength, ffe was taken equal to 𝐸𝑓휀𝑓𝑒, where Ef is the cracked tensile
modulus of elasticity of FRCM shown in Table 5.4. The reduction in the cross section of the steel
bars due to corrosion was considered according to the actual average mass loss encountered in
each beam as reported in Table 5.1. The flexural contribution of the longitudinal fibers bonded to
the sides of the beams strengthened in Scheme II was considered in strength calculations.
Following these assumptions, the cracking, yielding, and ultimate load-carrying capacities and the
corresponding deflections were calculated.
The predicted responses were compared to the experimental ones in Figures 5.18a to 5.18c. It
can be depicted that, prior to yielding of steel bars, the predicted and experimental responses of all
beams followed almost a similar trend. However, the predicted responses were lower than the
experimental ones after yielding. The theoretical ultimate capacities, 𝑃𝑢𝑡ℎ, and the ratios of the
experimental to the theoretical values, 𝑃𝑢
𝑃𝑢𝑡ℎ , for all the tested beams are shown in Table 5.1. The
ratios 𝑃𝑢
𝑃𝑢𝑡ℎ ranged between 1.03 for beam CSA-2P-I and 1.23 for beam CSB-3C-II. The ratios
𝑃𝑢
𝑃𝑢𝑡ℎ
indicated that the theoretical formulations of ACI 549.4R-13 [54] reasonably predicted the ultimate
capacities of the corrosion-damaged RC beams strengthened with FRCM composites in Scheme I
but underestimated the capacities of those strengthened in Scheme II. This finding was attributed
to the fact that ACI 549.4R-13 provisions [54] don’t consider the strengthening scheme in their
93
formulations. It was also related to the different modes of failure that occurred in the strengthened
beams. While fabric debonding occurred in beams strengthened with Scheme I and II as previously
explained, the U-shaped FRCM layer in Scheme II played a major role in confining the bottom
fabric layers and in delaying their delamination from the matrix. This explains the enhancement in
the load-carrying capacities of the beams strengthened in Scheme II as compared to their
counterparts strengthened in Scheme I. Since the ACI formulations don’t consider the effect of
scheme in their predictions, they tend to underestimate the capacities of beams strengthened in
Scheme II.
The obtained results suggested increasing the predicted ultimate capacity of the beams
strengthened in Scheme II by 10% to consider the effect of the continuous anchoring. The
comparison between the theoretical and experimental results also indicated that the assumption of
prefect bond suggested by ACI 549.4R-13 [54] while limiting the design effective strain, εfe, in
the FRCM systems is justified.
(a)
94
(b)
(c)
Figure 5.18: Predicted versus experimental flexural response for beams strengthened with a) two
layers of PBO-FRCM in Scheme I, b) four layers of PBO-FRCM in Scheme II, and c) three
layers of C-FRCM in Scheme II
95
5.7 Conclusions
The flexural behavior of RC beams strengthened with PBO- and C-FRCM composites after being
exposed to three different degrees of corrosion damage was presented. The following conclusions
can be drawn from the test results:
• Corrosion of steel reinforcing bars in the moment zone with an average mass loss up to
22.7% had a marginal impact on the flexural response of the tested beams. The maximum
decrease in the yield and ultimate strengths of the corrosion-damaged beams were 15 and
9%, respectively, of those of the virgin beams.
• The use of PBO- and C-FRCM systems enhanced the flexural response of the corrosion-
damaged beams. The type, amount, and anchoring scheme of the applied FRCM composite
governed the failure mechanism and the load-carrying capacities of the strengthened beams.
• Beams strengthened with PBO-FRCM showed strength gain that ranged between 7 and 44%
of that of the virgin beam. Beams strengthened with PBO-FRCM in Scheme I failed by
FRCM delamination while those strengthened in Scheme II failed due to fabric slippage
within the matrix.
• The strength gain in beams strengthened with C-FRCM in scheme II ranged between 39 to
55 % of that of the virgin beams and failed due to excessive premature matrix cracking.
• The mechanical properties of the cementitious matrix and their bond to the fabric layers
played a dominant role in defining the failure mechanism of the PBO- and C-FRCM
strengthened beams and consequently the strengthening efficiency of the FRCM system.
• Although the equivalent stiffness of the FRCM composites, Kf, and the stiffness factor, 𝛽𝑓,
are believed to be good indicators of the strengthening effectiveness of the FRCM systems,
they should not be solely used to compare the strength gain in beams without considering
the mechanical characteristics of the matrix, the number of fabric plies, and the anchoring
scheme used.
96
• Beams strengthened with FRCM composites showed ductility indices and energy absorption
indices that ranged between 86 and 118% and between 111 and 153%, respectively, of those
of the virgin beams.
• The theoretical formulations of ACI-549.4R-13 reasonably predicted the capacities of the
beams strengthened in Scheme I (with end anchors) but underestimated the capacities of
those strengthened in Scheme II (continuous U-shaped). This was attributed to the fact that
ACI-549.4R-13 formulations don’t consider the effect of the continuous anchorage on
delaying the delamination of the FRCM system.
• The theoretical capacities of beams strengthened in scheme II should be increased by 10%
to consider the effect of the continuous anchoring.
97
6. Chapter 6
Post-repair Flexural Performance of Corrosion-Damaged
Beams Rehabilitated with Fabric-Reinforced Cementitious
Matrix (FRCM)
Mohammed Elghazy, Ahmed El Refai, Usama Ebead, and Antonio Nanni
Journal of Construction and Building Materials. Date of acceptance: January 16,2018
https://doi.org/10.1016/j.conbuildmat.2018.01.128
Résumé
Cet article présente les résultats d'un programme de recherche examinant la performance post-
réparation en flexion des poutres en béton armé endommagées par la corrosion et réparées avec
différents systèmes MCRF. Neuf poutres ont été testées, y compris deux poutres qui n'ont été ni
corrodées ni réparées, une poutre corrodée et non réparée et six poutres corrodées et réparées en
deux phases. Les poutres de la phase I ont été soumises à un processus de corrosion accéléré
pendant 210 jours avant d'être réparées alors que les poutres de la phase II ont été soumises à une
corrosion accélérée pendant 70 jours, puis réparées et exposées une autre fois à la corrosion
pendant 140 jours. Les résultats des essais de flexion ont montré que l'exposition des poutres
réparées par MCRF à la corrosion après réparation a entraîné une réduction de 23% de la perte de
masse de l'acier. Les couches MCRF en forme de U étaient plus efficaces à réduire le taux de
corrosion et à augmenter la résistance ultime des poutres réparées que les couches MCRF ancrées
à l'extrémité. Les poutres réparées par la MCRF au PBO ont montré une plus faible rigidité post-
plastification et une plus grande ductilité à l’ultime que celles de leurs contreparties réparées par
MCRF au carbone. Les poutres soumises à un environnement corrosif après réparation présentaient
des capacités de charge qui se situaient entre 14 et 65% au-dessus de celles des poutres vierges.
Les dispositions de l'ACI 549.4R-13 prévoient de façon conservatrice les capacités ultimes des
poutres réparées par MCRF et exposées à un environnement corrosif après réparation.
Mots-clés des auteurs : Corrosion; Matrice cimentaire renforcée de fibres; Flexion; Béton armé;
Réparation; Renforcement; Réhabilitation; Durabilité; Performance à long terme.
98
6.1 Abstract
This paper presents the results of a research program examining the post-repair flexural response
of corrosion-damaged reinforced concrete (RC) beams repaired with different FRCM systems. A
total of nine RC beams were tested, including two beams that were neither corroded nor repaired,
one beam that was corroded and not repaired, and six corroded-repaired beams that were prepared
in two phases. Beams of phase I were subjected to an accelerated corrosion process for 210 days
before being repaired whereas beams of phase II were initially subjected to accelerated corrosion
for 70 days, then repaired and exposed to further corrosion for 140 days. Flexural test results
showed that exposing the FRCM-repaired beams to corrosion after repair resulted in 23%
reduction in steel mass loss. The use of U-shaped FRCM layers was more efficient in reducing the
corrosion rate and increasing the ultimate strength of the repaired beams than the end-anchored
FRCM layers. The PBO FRCM-repaired beams showed lower post-yielding stiffness and more
ductility at failure than those of their carbon FRCM-repaired counterparts. Beams that experienced
post-repair corrosive environment showed load-carrying capacities that ranged between 14 and
65% above those of the virgin beam. ACI 549.4R-13 provisions conservatively predict the ultimate
capacities of the FRCM-repaired beams exposed to post-repair corrosive environment.
Authors’ keywords: Corrosion; Fabric-reinforced cementitious matrix; Flexure; Reinforced
concrete; Repair; Strengthening; Rehabilitation; Durability; Long-term performance.
6.2 Introduction and Background
Repair/strengthening of reinforced concrete (RC) structures is motivated by several factors
including aging, change in use, increased loads, code compliance, and environmental damage (e.g.
corrosion). Corrosion of steel reinforcement is one of the major durability concerns for concrete
structures especially in coastal areas and cold regions where de-icing salts are heavily used. Pitting
corrosion reduces the cross-sectional area of the steel bars and may lead to decrease in ductility
of the steel bars [1,34]. The expansion of corrosion products causes concrete cracking and impair
the composite action of steel and concrete. As a result, the load-bearing capacity and the service
life of the corroded member are considerably jeopardized [4,23,102].
99
Existing externally-bonded repair technologies based on organic matrices referred to as fiber-
reinforced polymer (FRP) have proven success in restoring the serviceability and strength of RC
structures [10–12]. Using FRP products in repair applications was driven by their non-corrosive
properties, lightweight, and high tensile strength. However, FRP matrices are flammable, prone to
deterioration at high temperatures, and have poor thermal compatibility to the concrete substrate
[57,59,92,93]. In order to overcome these drawbacks, fabric-reinforced cementitious matrix
(FRCM) systems have been introduced as promising alternatives.
FRCM systems consist of one or more layers of textiles made of carbon, glass, or
Polyparaphenylene benzobisoxazole (PBO) grids that are sandwiched between layers of
cementitious mortars. Its lightweight, high tensile strength, and ease of application makes the
system appealing. The technique also surmounts the epoxy-bonded FRP systems that lack fire
resistance as the embedded grid is shielded between the mortar layers thus minimizing its
vulnerability hazard as the organic matrix is no longer present. In addition, the compatibility
between the mortar used and the concrete substrate is inherited since both materials have the
cement as a common “base”. FRCM systems, with their innovative features, ensure the endurance
of the rehabilitation process and consequently the sustainability of the strengthened structure.
Much research has been reported on reinforced and prestressed concrete structures strengthened
with FRCM subjected to monotonic loading [58,84]. Previous research studies have proven the
success of FRCM in enhancing the performance of RC structures [103,104] and masonry structures
[105]. D’Ambrisi and Focacci [81] demonstrated that the performance of FRCM materials was
strongly dependent on the fabric type and the matrix constituents. Schladitz et al. [55] reported an
increase in the load-carrying capacity of RC slabs strengthened with one and four layers of C-
FRCM by 67% and 245%, respectively. Fabric rupture was observed at failure in all of the
strengthened slabs even when high amount of FRCM layers was used. Significant decrease in
deflection was reported when four layers of FRCM were used [55]. Elsanadedy et al. [106]
reported that providing U-anchorage at the FRCM ends was efficient in alleviating the FRCM end
debonding in flexure tests. More recently, Ebead et al. reported that the failure mechanism of
carbon-and PBO-FRCM-strengthened beams was governed by inter-laminar shear [97].
Only one study [86] has documented the effectiveness of using basalt and carbon FRCM systems
to restore the ultimate capacity and serviceability of T-beams after a mass loss of 22% in their
100
steel reinforcement due to corrosion. It was reported that the basalt-FRCM system could not restore
the original flexural capacity of the beam whereas the C-FRCM system restored 109% of the
capacity. The authors reported that the use of a combination of internally-embedded and
externally-bonded C-FRCM layers was more effective in improving the strength and ductility of
the beams than the use of the same amount of FRCM layers internally embedded within the
corroded-repaired region.
Corrosion-damaged structures are often vulnerable to the same deterioration mechanism after
repair, which may require further repair during their service life. To date, no data is available in
the literature concerning the post-repair performance of FRCM-repaired beams that might be
vulnerable to further corrosive environments while in service. This paper aims at filling this gap
by reporting on the flexural performance of post-repair RC beams. The test program investigated
the type of the FRCM used (carbon and PBO) and the FRCM repair scheme (end-anchored and U-
wrapped). The paper also reports on the failure modes, the load-carrying capacities, the ductility,
and the straining actions as observed and quantified during the tests. The ultimate load-carrying
capacities of the tested beams were calculated in accordance with the provisions of the ACI
549.4R-13 [54] and compared to the experimental results.
6.3 Experimental Program
The test matrix of the experimental program is given in Table 6.1. Nine large-scale RC beams,
including two virgin uncorroded beams (UU), one corroded unrepaired specimen (CU), and six
corroded and repaired beams were prepared in two phases. In phase I, three beams (referred to as
the short-term beams) were subjected to an accelerated corrosion process for 210 days. The beams
were then tested immediately after being repaired with different FRCM systems and
configurations. In phase II, three beams (referred to as the long-term beams) were initially
subjected to an accelerated corrosion process for 70 days. The beams were then repaired with
different FRCM systems. After curing, the beams were exposed to further corrosive environment
for 140 days prior to testing. The schematic shown in Figure 6.1 describes the two phases of the
experimental program.
The beams were labeled following the A-B-C format. ‘A’ represents the beam condition (UU,
CU, and CR referring to Uncorroded-Unrepaired, Corroded-Unrepaired, and Corroded-Repaired,
101
respectively), while ‘S’ and ‘L’ refer to the short-term and long-term testing procedure,
respectively. ‘B’ denotes the number and type of the FRCM layers applied (4P, and 3C referring
to four layers of PBO-FRCM and three layers of C-FRCM, respectively). Finally, ‘C’ describes
the FRCM strengthening schemes (I and II) as will be detailed in the following sections.
Table 6.1: Test matrix
* 𝐴𝑠 was determined based on the measured average tensile steel mass loss due to corrosion
Figure 6.1: Schematic of the testing procedure of the short- and long-term beams
6.3.1 Test Specimen and Materials
Specimen Ave. Mass
loss (%)
Max. Mass
loss (%) 𝐴𝑠
*
(mm2)
𝜌𝑓
(%)
𝐾𝑓 = 𝜌𝑓𝐸𝑓
(MPa)
Control specimens
UUa, UUb - - 400 - -
CU 22.5 25.9 310 - -
Phase I: Short-term specimens
CRS-4P-I 22.7 25.6 309.2 0.08 95.3
CRS-4P-II 21.1 23.9 315.6 0.08 95.3
CRS-3C-II 21.5 23.2 314 0.185 139.1
Phase II: Long-term specimens
CRL-4P-I 18.7 20.3 325.2 0.08 95.3
CRL-4P-II 18.1 19.2 327.6 0.08 95.3
CRL-3C-II 17.5 18.7 330 0.185 139.1
102
All of the beams were 2.8 m long with similar cross section of width, b = 150 mm and height, h
= 250 mm. The beams were reinforced with two 8M (diameter 8 mm) and 15M (diameter 15 mm)
deformed rebars at top and bottom, respectively. The test specimen was designed to fail in flexure
under four-point load configuration. Thus, the shear spans were reinforced with 10M (diameter 10
mm) deformed stirrups spaced at 100 mm. A hollow stainless-steel tube with external and internal
diameters of 9.5 mm and 7 mm, respectively, was placed at 80 mm from the specimen tension face
to act as a cathode during the accelerated corrosion process. Typical dimensions and reinforcement
details of the test specimen are shown in Figure 6.2.
Normal and salted ready-mixed concrete batches having similar water/cement ratio were used to
cast the beams. Six standard concrete cylinders (150×300 mm) were prepared from each batch and
were tested in compression after 28 days and on the day of testing. Table 6.2 lists the compressive
strengths of both mixes. Prior to FRCM application, the corroded beams were repaired using local
commercial cementitious repair mortar (Sikacrete-08SCC) having a compressive strength of 55.4
MPa (standard deviation of 5 MPa) and flexural strength of 3.4 MPa (standard deviation of 0.3
MPa) as tested by the authors. The yield strengths of the longitudinal reinforcing steel bars of
diameter 15 and 8 mm were 466 MPa (standard deviation of 4.2 MPa) and 573 MPa (standard
deviation of 17.7 MPa), respectively, as tested by the authors.
Figure 6.2: Typical dimensions and reinforcement details of the test beam (all dimensions in
mm)
103
Table 6.2 : Concrete compressive strengths
Compressive strength,
MPa
Standard deviation,
MPa
Coefficient of variation
%
28-day Normal concrete 37.9 0.8 2
Salted concrete 33.5 1.1 3.2
Testing day Normal concrete 41.8 4.8 11.4
Salted concrete 41.2 0.6 1.6
6.3.2 Accelerated Corrosion Process
A potentiostatic technique was used to accelerate the corrosion of the tensile steel reinforcement.
The steel bars were connected to the positive terminal of a DC power supply to work as anode
whereas the stainless-steel tube was connected to the negative terminal of the power supply to act
as cathode. The specimens were electrically connected in series as shown in Figure 6.3 to obtain a
current density of 180 µA/cm2 that was impressed in the reinforcing bars. Corrosion of the steel
bars was localized in the middle 1200 mm of the beam’s span. Salt (NaCl) measured as 5% of the
cement weight was added to the concrete mix used to cast the middle-bottom of the corroded
specimens with a height of 100 mm (Figure 6.2). Similar amount of salt was added to the repair
mortar used to repair the long-term specimens. During the accelerated corrosion process, the
specimens were subjected to wet-dry cycles that consisted of 3 days wet followed by 3 days dry in
a large environmental chamber. The wet-dry cycles provided water and oxygen necessary for the
corrosion process and simulated the environmental conditions of a beam in-service. According to
Faraday’s law, the mass loss in the reinforcing bars was estimated as 30% based on the intensity
of the applied electrical current and the duration of the corrosion process.
104
Figure 6.3: A schematic of the electrical connection
6.3.3 FRCM Composites
Two commercial FRCM systems (PBO and carbon) were used to strengthen the corroded
specimens (Figure 6.4). The fabric properties in the primary direction as reported in the
manufacturers’ data sheet are shown in Table 6.3. The PBO fabric consists of an unbalanced net
of fiber rovings spaced along two orthogonal directions as shown in Figure 6.4a. The associated
inorganic cementitious matrix had a compressive strength of 43.9 MPa (standard deviation of 0.4
MPa) and a flexural strength of 3 MPa (standard deviation of 0.3 MPa) after 28 days as tested by
the authors. On the other hand, the carbon FRCM composite consists of a unidirectional net made
of carbon-fiber strands (Figure 6.4b) and impregnated in an inorganic cementitious matrix of
compressive strength of 42.1 MPa (standard deviation of 4.3 MPa) and flexural strength of 3.2
MPa (standard deviation of 0.3 MPa) after 28 days as tested by the authors.
105
Figure 6.4:FRCM systems; a) PBO-FRCM (unbalanced PBO fabric) and (b) C-FRCM
(unidirectional carbon fabric) - all dimensions in mm
Table 6.3: Fabric properties in the primary direction as given in the manufactures’ data sheet
Fabric Area per unit width
𝐴𝑓, mm2/m Tensile strength,
GPa
Elastic modulus,
GPa
Ultimate strain,
%
PBO 50 5.8 270 2.15
Carbon 157 4.3 240 1.75
According to ACI 549.4R-13 [54], the tensile stress-strain behavior of a FRCM composite
coupon can be expressed by the bilinear relationship shown in Figure 6.5. The initial linear portion
corresponds to the uncracked behavior of the FRCM matrix while the second linear portion
represents its cracked behavior up to failure. Table 6.4 lists the properties of the FRCM composite
systems as reported in [97].
106
Figure 6.5: Idealized tensile stress-strain curve of FRCM coupon specimen ACI 549.4R-13 [54]
Table 6.4: Mechanical properties of FRCM coupons as reported in [97]
FRCM system Cracked tensile modulus
of elasticity Ef, GPa
Ultimate tensile
strength ffu, GPa
Ultimate strain
εfu, %
PBO-FRCM 121 1.55 1.4
Carbon-FRCM 75 0.97 1.25
An equivalent axial stiffness, Kf, given in Equation (6.1), was utilized to compare between the
two FRCM systems based on their cracked elastic modulus, 𝐸𝑓, and the cross-sectional area of the
fabric. The equivalent axial stiffness of each strengthened specimen is shown in Table 6.1. It is
important to mention that for beams repaired with continuous U-shaped PBO-FRCM layer
(Scheme II), the fiber strands on the lateral sides of beams were neglected in estimating 𝜌𝑓 (and
consequently 𝐾𝑓). Their contribution to the flexural strengths of the beams was considered in the
analysis as will be explained later.
𝐾𝑓 = 𝜌𝑓𝐸𝑓 Eq. (6.1)
Where, 𝜌𝑓 and 𝐸𝑓 are the fiber reinforcement ratio and the cracked elastic modulus of the FRCM
composite, respectively.
6.3.4 FRCM Schemes
Two FRCM strengthening schemes were utilized in this study as shown in Figure 6.6. Scheme I
consisted of four PBO-FRCM flexure plies having 150 mm width (equal to the width of the beam)
107
and applied to the soffit of the beam over a length of 2.4 m. The fabric was oriented so that its
primary direction was parallel to the longitudinal axis of the beam. The flexural plies were
anchored at each end using one U-shaped transverse strip of 300 mm width (Figure 6.6a). Scheme
II consisted of bottom flexural plies as in Scheme I but wrapped with one U-shaped continuous
ply along the beam’s clear span (Figure 6.6b). The primary direction of the U-wrapped PBO ply
in Scheme II was oriented parallel to the longitudinal axis of the beams and was counted as an
additional ply. On the other hand, the carbon fabric is a unidirectional fabric. Therefore, the bottom
flexural plies of C-FRCM composite in Scheme II were oriented parallel to the longitudinal axis
of the beams whereas the U-shaped layer was oriented in the transverse direction and therefore did
not contribute to the flexural resistance of the beam (Figure 6.6b).
Figure 6.6: FRCM repair schemes; a) Scheme I and b) Scheme II
108
6.3.5 Repair Methodology
Following the initial corrosion, the corroded beams were placed upside down with the beam soffit
at the top to facilitate the repair process. The deteriorated concrete was first removed using a
hydraulic hammer. The corroded steel bars were totally exposed and cleaned using a steel brush.
The grooved area was then filled with local cementitious repair mortar (Sikacrete-08SCC). After
the curing of the repair mortar, the beam surface was sandblasted prior to FRCM application. The
hand lay-up method proposed by the ACI 549.4R-13 [54] and the manufacturers was followed
while installing the FRCM systems. The beam’s substrate was first damped in water for 2 hours
before applying the first layer of the cementitious matrix with a thickness of 3 to 4 mm. The fabric
was then installed in place and was gently impregnated into the cementitious matrix and covered
with a second layer of matrix with similar thickness. The procedure was then repeated until the
specified number of layers was achieved. All repaired beams were cured for 28 days in laboratory
conditions before being tested or further exposed to corrosion.
6.3.6 Test Setup and Instrumentation
The specimens were tested to failure in a four-point bending configuration as shown in Figure
6.1. All tests were conducted under displacement control at a rate of 2 mm/minute using a MTS
actuator. A spreader steel beam was used to spread the load equally to two loading points spaced
800 mm apart. Beam deflections were measured by means of three linear variable differential
transducers (LVDTs) located at mid-span and at the loading points. All beams were instrumented
at mid-span with a 60 mm long strain gauge bonded to the top surface of concrete and 5 mm strain
gauges bonded to the tensile steel bars. The FRCM-repaired beams were instrumented with 5 mm
strain gauges installed directly on the outer fabric of the FRCM composite at mid-span and at the
loading points. The measuring instruments were connected to a 20-channel data acquisition to
capture the readings at all stages of loading.
6.4 Test Results and Discussion
6.4.1 Corrosion Crack Pattern
Crack patterns that resulted from corrosion were reported for all specimens at the end of the
accelerated corrosion process. For specimens in Phase I (short-term), longitudinal cracks extending
109
parallel to the corroded reinforcement were observed on the lateral surface and the soffit of the
beams. Figure 6.7a shows the cracks pattern due to corrosion as observed for the corroded-
unrepaired beam (CU). Those cracks had a maximum and average widths of 3.5 mm and 1.6 mm,
respectively. All specimens in this phase failed to meet the ACI 318-14 [98] service requirements
that limits the maximum crack width during service to 0.40 mm. For the specimens exposed to the
post-repair corrosive environment in Phase II (long-term), longitudinal cracks parallel to the
corroded reinforcement on one or both lateral surfaces of the specimens were observed. The
corrosion cracks had maximum and average crack widths of 1.3 and 0.8 mm for specimen CRL-
4P-I (Figure 6.7b), 0.7 and 0.45 mm for specimen CRL-4P-II (Figure 6.7c), and 0.25 and 0.16 mm
for specimen CRL-3C-II (Figure 6.7d), respectively. These findings imply that the PBO-FRCM
systems were less effective in decreasing the corrosion rate as compared to the C-FRCM system.
However, both systems significantly enhanced the serviceability of the repaired beams in terms of
crack widths.
6.4.2 Steel Mass Loss
Five steel coupons (200 mm long each) were extracted from the corroded steel bars of the
damaged beams after testing. The actual mass loss of the corroded bars were determined according
to the ASTM G1-03 standards [27]. The average and maximum tensile steel mass loss for each
specimen are listed in Table 6.1. It can be noticed that the steel bars of the long-term specimens
exhibited lower mass loss than those of the short-term ones despite the same duration of corrosion
exposure in both cases. This was attributed to the reduction in the amount of the diffused water
and air to the tensile reinforcing bars due to due to the presence of the FRCM layers and the use
of repair mortar with lower permeability than that of concrete, which also explains the reduction
in the corrosion rate in the long-term specimens, which also explains the reduction in the corrosion
rate in the long-term specimens. For instance, specimens CRL-4P-II and CRL-3C-II showed an
average mass loss 18.1 and 17.5 %, respectively. However, their counterparts in Phase I (short-
term) had an average mass loss of 21.1% (beam CRS-4P-II) and 21.5% (beam CRS-3C-II),
respectively. On the other hand, the FRCM repair scheme had a marginal effect on the corrosion
rate in the long-term beams. Beam CRL-4P-II repaired with Scheme II had an average mass loss
of 18.1 %, which was comparable to the steel mass loss of beam CRL-4P-I repaired with Scheme
I. It is important to mention that the mass loss measurements are highly variable due to the random
110
nature of the corrosion phenomena. Therefore, more research is needed to quantify the effect of
using FRCM composites on the rate of corrosion.
Figure 6.7: Corrosion cracks patterns; a) typical corrosion cracks pattern for short-term
specimens (beam CU); b) beam CRL-4P-I; c) beam CRL-4P-II; and d) beam CRL-3C-II
a)
b)
c)
d)
111
6.4.3 Flexural Cracks Pattern and Failure modes
The virgin beams (UU) and the corroded-unrepaired beam (CU) showed typical crack patterns
and failure modes of under-reinforced beams. They failed at ultimate due to the yielding of steel
bars followed by concrete crushing at the top. Concrete spalling was observed in beam CU due to
crossing of the vertical flexural cracks with the longitudinal corrosion cracks (Figure 6.8a). The
crack patterns and failure modes of the FRCM-repaired beams were highly dependent on the type
of the FRCM system and the repair scheme used. All of the repaired beams failed due to loss of
strengthening action followed by concrete crushing. Based on the test observations, the loss of the
strengthening action took place due to one or a combination of the following modes of failure:
a) Mode A: FRCM delamination at the fabric/matrix interface adjacent to the concrete substrate.
This mode of failure is shown in Figure 6.8b and Figure 6.8c for the beams CRS-4P-I and CRL-
4P-I, respectively.
b) Mode B: partial debonding of the fabric from the matrix accompanied by large slip (more than
3 mm) of the fabric as shown in Figure 6.8d for the beam CRL-4P-II.
c) Mode C: wide flexural cracking (up to 4 mm width) in the cementitious matrix with extensive
fabric slippage at the beam’s soffit in the maximum moment zone. This mode of failure was
encountered in the beams reinforced with C-FRCM and it is shown in Figure 6.8e for the beam
CRL-3C-II.
112
Figure 6.8: Failure mode of a) beam CU due to steel yielding and concrete crushing; b) beam
CRS-4P-I due to FRCM delamination; c) beam CRL-4P-I due to premature FRCM delamination;
d) beam CRL-4P-II due to PBO-fabric debonding from matrix; and e) beam CRL-3C-II due to
fabric slippage
a)
d)
c)
b)
e)
TOP
Side
Side
Side
Bottom
Concrete cover
spalling
Concrete
crushing
FRCM
delamination FRCM delamination
FRCM
delamination Corrosion
cracks
FRCM
delamination
Corrosion
cracks
Partial debonding
Matrix cracking
113
The FRCM delamination (mode A) encountered in beam CRS-4P-I (short-term) was attributed
to the propagation of the flexural cracks in the thin layer of the FRCM matrix adjacent to the
concrete substrate. Its counterpart specimen CRL-4P-I (long-term), having similar FRCM system
and configuration but exposed to post-repair corrosive environment, failed in a more brittle manner
due to the premature delamination of the FRCM composite that was initiated by the longitudinal
corrosion cracks (Figure 6.8c). It should be pointed out that the post-repair corrosive environment
resulted in horizontal corrosion cracks parallel to the reinforcing steel, which notably impaired the
concrete cover and consequently limited the strengthening effectiveness of the FRCM system. The
effect of the post-repair exposure to corrosion was also reflected on the load-carrying capacity of
the long-term beam and on its flexural response as will be presented in the following sections.
On the other hand, the short-term beam CRS-4P-II, which was repaired in the U-shaped FRCM
scheme, failed due to the partial debonding of the PBO fabric from the FRCM matrix (mode B).
As the applied load increased, vertical flexural cracks were formed in the FRCM matrix. The
extension of the cracks in the shear spans was followed by gradual slippage of the fabric from
within the matrix until the strengthening action of the FRCM system was lost. This mode of failure
was also reported at failure for the long-term beam CRL-4P-II.
Similarly, both beams repaired with carbon FRCM system (CRS-3C-II and CRL-3C-II) failed in
a similar mode (mode C). Vertical flexural cracks were observed in the matrix of the U-shaped C-
FRCM layer after steel yielding. As the applied load increased, new flexural cracks developed in
the maximum moment zone accompanied with the widening of the existing cracks. At ultimate, it
was observed that wide flexural cracks formed within the FRCM matrix followed by a noticeable
fabric slippage at the beam’s soffit as shown in Figure 6.8e.
Based on these test observations, it can be concluded that the exposure to the post-repair corrosive
environment had a slight impact on the mode of failure of the beams repaired with scheme II
regardless of the type of fabric used. This can be attributed to the wrapping effect of the FRCM
layer that offset the effect of corrosion cracks in weakening the concrete cover and prevented the
premature delamination of FRCM. On the contrary, the long-term beam CRL-4P-I repaired in
scheme I demonstrated large corrosion cracks that significantly affected the strengthening action
of the FRCM system and led to the premature delamination of FRCM (Figure 6.8c).
114
6.4.4 Flexural Response
The load-displacement curves of the repaired specimens are shown in Figure 6.9. The flexural
response of the virgin beam (UU) and the corroded-unrepaired beam (CU) are also shown for
comparison. Based on the test observations, corrosion of the main reinforcement by an average
mass loss of 22.5% did not have a notable effect on the beam stiffness. For all of the tested beams,
the load-deflection curves consisted of three segments with two turning points indicating the initial
cracking in concrete and the yielding of the tensile steel bars. The flexural response of the repaired
beams was highly dependent on the FRCM repair scheme and its fabric type. It is important to
note that the equivalent axial stiffness of the C-FRCM (3 layers) was approximately 1.5 times of
that of PBO-FRCM (4 layers).
Figure 6.9a and 6.9b illustrate the load versus the mid-span deflection for the short-term and
long-term beams, respectively. The use of FRCM had a slight influence on the stiffness of the
repaired beams prior to steel yielding. However, the post-yielding stiffness significantly increased
in the repaired beams in comparison to the control ones. It can be noticed that beams repaired with
PBO-FRCM (CRS-4P-I and CRS-4P-II) showed lower post-yielding stiffness than that of their C-
FRCM repaired counterparts (Figure 6.9a).
Exposing the beams to corrosion after repair did not affect their load-deflection response as
shown in Figure 6.9c and 6.9d. This can be depicted by comparing the response of the short-term
beams and their long-term counterparts. Both beams repaired with C-FRCM (CRS-3C-II and CRL-
3C-II) showed a sudden drop at ultimate, which indicated their brittle mode failure due to the
sudden cracking in the FRCM matrix and the large fabric slippage within the matrix. The PBO-
repaired beams exhibited different trends that were characterized by a gradual declining branch
after reaching the ultimate load as shown for beams CRS-4P-I and CRS-4P-II and their long-term
counterparts. This was attributed to the different modes of failure of the PBO and C-FRCM
repaired beams as previously described.
115
a)
b)
116
c)
d)
Figure 6.9: Load-deflection relationships of a) short-term beams; b) long-term beams; c) beams
repaired with PBO-FRCM (short-term and long-term); and d) beams repaired with C-FRCM
(short-term and long-term)
6.4.5 Load-carrying Capacities
The load-carrying capacities of the tested beams are shown in Table 6.5. The test results indicate
that an average mass loss of 22.5 % due to corrosion resulted in reduction in the yield and ultimate
strengths by 15% and 10%, respectively. The insignificant effect of corrosion on the yield and
ultimate strengths of RC beams might be attributed to the fact that the corroded steel bars lost their
117
lugs, which increased the measured mass loss without notable loss of the effective cross section
area of the bars. Repairing the corroded RC beams with externally bonded FRCM composites had
restored, and in some cases exceeded, the yield and ultimate strengths of the virgin beam. The
enhancement in the ultimate load of the repaired beams ranged between 14% and 65% of that of
the virgin beam. The experimental results of the yield load, 𝑃𝑦𝑒𝑥𝑝
, and the ultimate load, 𝑃𝑢𝑒𝑥𝑝
, were
normalized with respect to those of the virgin specimens as presented in Table 5.
It is important to note that the efficiency of the FRCM systems in strengthening the damaged
beams was evaluated based on their equivalent axial stiffness and the gain in the ultimate capacity
of the repaired beams. Four layers of PBO-FRCM had an equivalent stiffness of 95.3 MPa
compared to 139.1 MPa for three layers of C-FRCM. For the short-term beams, the use of four
PBO-FRCM layers with repair Scheme I in beam CRS-4P-I restored 110 and 129% of the yield
and ultimate strengths of the virgin beam, respectively. The use of similar number of PBO layers
in Scheme II restored 108 and 139% of the yield and ultimate strengths of the virgin beam,
respectively. The enhancement in the ultimate capacity when Scheme II was used was attributed
to the effect of the U-shaped layer on delaying the delamination of the flexural FRCM plies and
the contribution of the PBO strands on the lateral sides. Consequently, the load-carrying capacity
of the beam CRS-4P-II exceeded that of its counterpart CRS-4P-I.
Table 6.5: Strength results of the tested beams
Specimen 𝑃𝑦
𝑒𝑥𝑝
KN
𝑃𝑢𝑒𝑥𝑝
KN
Normalized loads** 𝑃𝑢𝑝𝑟𝑒𝑑
KN
𝑃𝑢𝑒𝑥𝑝
𝑃𝑢𝑝𝑟𝑒𝑑
𝑃𝑦𝑒𝑥𝑝
𝑃𝑢𝑒𝑥𝑝
UUa, UUb* 75.1 79.7 1 1 81.9 0.97
CU 64.19 72.2 0.85 0.90 64.7 1.11
CRS-4P-I 82.47 102.8 1.1 1.29 86.9 1.18
CRS-4P-II 80.87 111.1 1.08 1.39 93.1 1.19
CRS-3C-II 75.86 109.3 1 1.37 97.8 1.11
CRL-4P-I 79.52 91.1 1.06 1.14 89.9 1.01
CRL-4P-II 83.15 108.1 1.11 1.36 94.7 1.14
CRL-3C-II 81.87 131.9 1.09 1.65 100.8 1.3
* Average values reported **Normalized with respect to the yield and ultimate loads of the virgin beam
118
Exposing the beam CRL-4P-I to further corrosion after repair decreased its yield and ultimate
strengths compared to those of beam CRS-4P-I. It is worth noting that these beams had average
steel mass losses of 18.7 and 22.7%, respectively. The beam CRL-4P-I restored 106 and 114% of
the yield and ultimate strengths of the virgin beam, respectively (Table 6.5 and Figure 6.8c). This
reduction in the strength gain can be attributed to the presence of longitudinal corrosion cracks
with widths up to 1.3 mm in the beam as observed during the test. These wide corrosion cracks
weakened the concrete substrate and caused premature delamination of the FRCM layer, which
significantly limited the strengthening contribution of the FRCM system.
On the contrary, exposing the PBO-repaired beam in Scheme II to post-repair corrosion had no
effect on its yield and ultimate strengths. The use of four layers of PBO-FRCM with Scheme II in
specimen CRL-4P-II increased its yield and ultimate strengths by 11 and 36%, respectively, in
comparison to 8 and 39% for its counterpart short-term specimen CRS-4P-II. A similar trend was
reported for the specimens repaired with CFRCM in Scheme II. Specimen CRL-3C-II showed an
increase of 9 and 65% of its yield and ultimate strengths, respectively, compared to 0 and 37% for
its counterpart specimen CRS-3C-II. The higher enhancement in the ultimate load-carrying
capacity reported in the former specimen (beam CRL-3C-II) can be attributed to the lower mass
loss reported in the steel bars compared to the mass loss in specimen CRS-3C-II (17.5% mass loss
versus 21.5%, respectively). It can also be attributed to the long curing period of the long-term
beam during the post-repair corrosion exposure.
6.4.6 Ductility and Energy Absorption
The ductility index, ΔI, and the energy absorption index, ψ, for each beam are listed in Table 6.6.
The ductility index is defined as the ratio of the mid-span deflection at ultimate, δu, to its mid-span
deflection at yielding, δy, whereas the energy absorption index, ψ, is defined as the area under the
load-deflection curve up to the ultimate load.
119
Table 6.6: Ductility and energy absorption of the tested beams
Specimen
Midspan
deflection, mm Ductility index
Energy absorption
index, KN.mm
δy δu ΔI ΔInorm** ψ ψnorm
**
UUa, UUb* 11.7 32.9 2.81 1 1736 1
CU 9.8 27.06 2.8 1 1735 1
CRS-4P-I 12.04 33.53 2.78 0.99 2657 1.53
CRS-4P-II 12.06 31.32 2.6 0.94 2426 1.4
CRS-3C-II 12.11 30.16 2.49 0.89 2207 1.27
CRL-4P-I 12.74 26.07 2.05 0.73 1780 1.03
CRL-4P-II 12.04 39.99 3.32 1.18 3360 1.94
CRL-3C-II 10.03 31.81 3.17 1.13 2895 1.67
* Average values reported **Normalized with respect to the virgin beam
The corroded-unrepaired specimen (CU) had a ductility index similar to that of the control
specimen. The short-term beams repaired with PBO-FRCM (CRS-4P-I and CRS-4P-II) showed a
ductility index almost similar to that of the virgin beam. The use of C-FRCM in specimen CRS-
3C-II reduced the ductility index by 11% of that of the virgin specimen. On the other hand, long-
term specimens (CRL-4P-II and CRL-3C-II) showed an increase in their ductility indices by 18%
and 13%, respectively, in comparison to that of the control specimen. This was attributed to the
long curing period of the repaired specimens in Phase II. On the contrary, a significant reduction
in the ductility index (27% of that of the control specimen) was determined for the long-term
specimen CRL-4P-I. This was attributed to the premature FRCM delamination as previously
described.
The control and the corroded-unrepaired specimens had similar energy absorption indices. For
the short-term specimens, those repaired with PBO-FRCM (CRS-4P-I and CRS-4P-II) exhibited
an average energy absorption index 15% higher than that of specimens repaired with C-FRCM
(CRS-3C-II). For specimens exposed to post-repair corrosion (long-term specimens), the energy
absorption indices of specimens CRL-4P-II and CRL-3C-II were higher than those reported for
their short-term counterparts by 39 and 31%, respectively. However, specimen CRL-4P-I showed
120
a notable reduction in the energy absorption index by 33% of that of its short-term counterpart
CRS-4P-I. These results are consistent with the modes of failure reported as previously discussed.
6.4.7 Strain Response
Table 6.7 lists the strains measured at mid-span in the fiber, the steel bars, and the compressed
concrete at ultimate for all of the tested specimens. At failure, the strains in the steel bars were
2428 μɛ for the control specimen (UU) while strains ranged from 1946 to 3490 μɛ for the beams
repaired with PBO-FRCM. Higher tensile steel strains (3402 to 4906 μɛ) were reported for the
specimens repaired with C-FRCM.
Figure 6.10a and Figure 6.10b show the load versus the recorded strains in both the concrete and
fiber at mid-span for specimens repaired with PBO-FRCM and C-FRCM, respectively. Prior to
yielding, the concrete strain increased with the same rate in all of the tested specimens. Following
the steel yielding, the virgin beam (UU) and the corroded-unrepaired beam (CU) showed an almost
plastic response until failure occurred by concrete crushing. This response was indicated by the
increase in the recorded strains in concrete in both beams. The use of FRCM composite caused a
notable strain hardening in concrete of the repaired specimens as depicted in Figure 6.10. Concrete
strain hardening in specimens repaired with C-FRCM was higher than that in those repaired with
PBO-FRCM. This finding was consistent with the increase in the post-yielding stiffness
encountered in the former specimens, as discussed earlier. The highest absolute concrete
compressive strain was reported in specimen UU (-3318 μɛ) while the lowest absolute strain was
observed in specimen CRL-4P-I (-1520 μɛ). This observation might be attributed to the premature
FRCM delamination initiated from the post-repair corrosion cracks that resulted in a sudden failure
in specimen CRL-4P-I. The concrete strains recorded in other specimens ranged between -2291 to
-3239 μɛ at ultimate.
121
Table 6.7: Strains at ultimate
Specimen Fiber strains,
µ𝜖
Concrete strains,
µ𝜖
Steel strains,
µ𝜖
UUa, UUb* -3318 2428
CU - -2338 -
CRS-4P-I 8446 -3239 2422
CRS-4P-II 9653 -2992 2928
CRS-3C-II 9777 -2469 4906
CRL-4P-I 7409 -1520 1946
CRL-4P-II 8928 -2296 3490
CRL-3C-II 13876 -2291 3402
* Average values reported
On the other hand, the outer fabric tensile strain increased as the applied load increased. Once
the tensile steel bars reached the yielding point, the fiber strains increased at higher rate than that
observed prior to yielding (Figure 6.10). This result was consistent with the limited effect of the
FRCM system on the beam performance prior to yielding. Exposing the beams to post-repair
corrosion had a marginal effect on the ultimate tensile fiber strain reported for specimens repaired
with PBO-FRCM. Specimens CRS-4P-II and CRL-4P-II showed ultimate fiber strains of 9653 and
8928 μɛ, respectively. Also, the fiber strains measured in specimens CRS-4P-I and CRL-4P-I were
8446 and 7409 μɛ, respectively. The low strain that was encountered in beam CRL-4P-I was
attributed to the premature FRCM delamination initiated from the post-repair corrosion cracks,
which negatively affected the FRCM system strengthening potency. On the other hand, specimens
CRS-3C-II and CRL-3C-II showed ultimate fiber strains of 9777 and 13879 μɛ, respectively. The
increase in the measured fiber strains in the long-term beam CRL-3C-II was consistent with its
higher ultimate load than its short-term counterpart.
122
a)
b)
Figure 6.10: Load-strains relationships for a) beams repaired with PBO-FRCM and b) beams
repaired with C-FRCM
123
It is important to note that the ultimate fiber strain, 휀𝑓𝑢, obtained from the direct tensile tests
(Table 4) conducted on coupons made of the same FRCM used in this study was 14000 μɛ for
PBO-FRCM and 12500 μɛ for C-FRCM [97]. These fiber strains are obviously higher than those
measured during the beam tests. Therefore, relying on the ultimate strains in the fabric may be
misleading in predicting the performance of the repaired beams while assuming perfect bond
between the concrete substrate and the FRCM system. Unlike the fabric slippage that occurred in
the FRCM tensile coupons at ultimate, different failure mechanisms were encountered in the
FRCM-repaired beams. Therefore, the assumption of prefect bond suggested by the ACI 549.4R-
13 [54] while limiting the design effective strain, 휀𝑓𝑒, in the FRCM system is a simplification that
appears justifiable and easy to implement by engineers.
6.5 Predicted Strengths
The predicted ultimate loads, 𝑃𝑢𝑝𝑟𝑒𝑑
, were determined according to the provisions of ACI 318-14
[98] and ACI 549.4R-13 [54]. Perfect bond between FRCM composites and the concrete substrate
was assumed while the fiber strain was limited to 0.012 mm/mm as recommended by ACI 549.4R-
13 [54]. Therefore, the design effective tensile strain, 휀𝑓𝑒, in FRCM was assumed equal to the
experimental ultimate strain, 휀𝑓𝑢, as obtained from the coupon test results minus one standard
deviation or 0.012 mm/mm, whichever was lower. The design effective tensile strength, 𝑓𝑓𝑒, was
taken equal to 𝐸𝑓휀𝑓𝑒, where 𝐸𝑓 is the cracked tensile modulus of FRCM as listed in Table 6.4. For
the beams repaired with PBO-FRCM in Scheme II, the longitudinal fiber strands on the lateral
sides of the U-shaped layer were considered in estimating the flexural strength. The corrosion
damage of the bottom steel bars was presented by a reduction in the cross-section areas of the bars,
calculated based on the measured average mass loss for each beam as given in Table 6.1. The yield
strengths of the longitudinal reinforcing steel bars of diameter 15 and 8 mm were taken equal to
466 and 573 MPa, respectively, with elastic modulus 𝐸𝑠 = 200 GPa. The concrete compressive
strength was taken equal to 41.8 MPa with maximum compression strain equal to 0.0035 mm/mm.
The predicted ultimate load, 𝑃𝑢𝑝𝑟𝑒𝑑
and the ratios of the experimental to predicted ultimate
loads 𝑃𝑢𝑒𝑥𝑝/𝑃𝑢
𝑝𝑟𝑒𝑑, for all of the tested beams are presented in Table 6.5. The ratio 𝑃𝑢
𝑒𝑥𝑝/𝑃𝑢𝑝𝑟𝑒𝑑
for all repaired specimens ranged between 1.01 and 1.3 indicating a very good agreement between
124
the experimental and predicted values. This finding revealed that the load-carrying capacities of
the repaired beams in-service, i.e. those exposed to further corrosive environment after repair, can
be conservatively estimated using ACI 549.4R-13 provisions [54].
6.6 Conclusions
The short and long-term flexural performances of concrete beams repaired with FRCM systems
are presented in this study. The test results have evidenced the following conclusions:
• An average mass loss of 22.5% in the tensile steel bars reduced the yield and the ultimate loads
of the corroded beams by 15% and 10%, respectively, without a notable impact on the beam’s
stiffness or mode of failure.
• The use of FRCM systems reduced the corrosion rate in the steel bars with no evidence on the
effect of the repair scheme on such rate. Exposing the FRCM-repaired beams to post-repair
corrosion resulted in 23% reduction in the steel mass loss.
• Most of the corroded-repaired specimens that were exposed to post-repair corrosive
environment failed to meet the serviceability provisions of the ACI-318-14 for crack widths.
• Repairing the corroded beams with FRCM systems enhanced their flexural behavior and
increased their load-carrying capacities between 14 to 65% of that of the virgin beam.
• The U-wrapped scheme was more efficient than the end-anchoring scheme in delaying the
delamination of the FRCM plies in the short-term repaired beams. It also mitigated the effect
of the longitudinal corrosion cracks and consequently increased the post-repair strengthening
effectiveness of FRCM systems.
• Short-term beams repaired with PBO-FRCM exhibited lower post-yielding stiffness and more
ductility at failure than those of their carbon-repaired counterparts. Average ductility indices
were 97 and 89% of that of the control specimen with average energy absorption indices of
147 and 127%, respectively.
• Long-term beams repaired with scheme II demonstrated higher ductility and energy absorption
indices than those of their short-term counterparts. The short-term beam repaired with scheme
I (CRL-4P-I) failed prematurely due to premature FRCM delamination.
125
• Strain values recorded on the FRCM layer indicated that the assumption of prefect bond
suggested by the ACI 549.4R-13 while limiting the design effective strain in the FRCM system
is justifiable.
• The ACI 549.4R-13 provisions conservatively predict the ultimate capacities of the FRCM-
repaired beams exposed to post-repair corrosive environment.
It is important to note that the results of this study are only applicable to the FRCM systems used
and should not be extrapolated to other systems. Further experimental and analytical studies using
other commercially available FRCM systems are recommended. Tests on RC beams subjected to
higher corrosion levels corrosion levels are also needed to assess the proposed repair technique in
more severe environments.
126
7. Chapter 7
Fatigue and Monotonic Behavior of Corrosion-damaged
Reinforced Concrete Beams Strengthened with FRCM
composites
Mohammed Elghazy, Ahmed El Refai, Usama Ebead, and Antonio Nanni
Journal of Composites for Construction, ASCE. Submitted in the revised form: Fabraury 12, 2018
Status: Under review
Résumé
Cet article fournit un compte-rendu complet de l'utilisation de composites à matrice cimentaire
renforcés de fibres (MCRF) pour renforcer les structures en béton armé endommagées par la
corrosion et soumises à la fatigue. Douze poutres ont été construites et testées à la rupture dans
une configuration de chargement à quatre points. Avant les essais, dix poutres ont été soumises à
une corrosion accélérée pendant 140 jours, ce qui a entraîné une perte de masse moyenne de 19%
dans le renforcement en acier. Huit poutres endommagées par la corrosion ont été renforcées et
testées tandis que les deux autres poutres n'ont pas été renforcées. Deux autres poutres vierges qui
n'ont pas été soumises à la corrosion ont été utilisées comme témoins. Les paramètres d'essai
comprenaient le matériau de fibre (PBO et carbone), le nombre de couches MCRF, la configuration
de renforcement et le type de chargement (monotone et fatigue). Les résultats des tests ont montré
que la corrosion des barres d'acier diminuait considérablement la durée de vie en fatigue des
poutres. Après le renforcement, les poutres endommagées par la corrosion ont entièrement restauré
la capacité de charge des poutres vierges. Les poutres renforcées par MCRF ont subi plus de cycles
de fatigue que leurs contreparties non renforcées, mais n'ont pas pu restaurer la durée de vie en
fatigue des poutres vierges. L'effet de la configuration de MCRF était plus prononcé en fatigue
que dans les tests monotones. Le PBO-MCRF était plus efficace que le C-FRCM à améliorer la
performance en fatigue des poutres endommagées par la corrosion.
Mots clés des auteurs : Corrosion; Matrice cimentaire renforcée de fibres; Fatigue; Flexion;
Réparation; Renforcement.
127
7.1 Abstract
This paper provides a comprehensive account of using fabric-reinforced cementitious matrix
(FRCM) composites to strengthen corrosion-damaged reinforced concrete (RC) structures
subjected to fatigue. Twelve beams were constructed and tested to failure under four-point loading
configuration. Prior to testing, ten beams were subjected to accelerated corrosion for 140 days
leading to an average mass loss in the steel reinforcement of 19%. Eight corrosion-damaged beams
were strengthened and tested while the other two beams remained unstrengthened. Two other
virgin beams that were not subjected to corrosion were used as benchmarks. The test parameters
included the fabric material (PBO and Carbon), the number of FRCM plies, the strengthening
configuration, and the type of loading (monotonic and fatigue). Test results showed that corrosion
of steel bars dramatically decreased the fatigue life of the beams. After strengthening, the
corrosion-damaged beams fully restored the load-carrying capacity of the virgin beam. The
FRCM-strengthened beams endured more load cycles than their unstrengthened counterpart, but
could not restore the original fatigue life of the virgin beam. The effect of FRCM configuration
was more pronounced in fatigue than in monotonic tests. PBO-FRCM was more effective than the
C-FRCM composite in enhancing the fatigue performance of the corrosion-damaged beams.
Authors’ keywords: Cementitious Materials; Corrosion; Fabric-Reinforced Cementitious
Matrix; Fatigue; Flexure; Repair; Strengthening.
7.2 Introduction and Background
Transportation infrastructures such as bridges are prone to corrosion of their steel reinforcement
due to the harsh environment and the heavy use of the de-icing chemicals in cold regions. These
structures are continuously subjected to cyclic loads, which may cause fatigue distress and
consequently reduce their anticipated life. It has been established that the fatigue life of concrete
structures is usually governed by the endurance of the steel reinforcing bars under repetitive loads
and is rarely controlled by concrete [50–52]. While fatigue failures are not common in concrete
structures, the corrosion of steel reinforcement combined with fatigue stresses significantly
128
reduces the fatigue life of the structure [53,107], which necessitates immediate intervention with
appropriate strengthening measures.
In the last decades, externally-bonded strengthening technologies based on organic matrices and
referred to as fiber-reinforced polymers (FRP) have become a common practice. Numerous studies
have investigated the behavior of corrosion-damaged RC beams strengthened with FRP
composites under monotonic and fatigue loads [90,108]. Test results demonstrated that the use of
FRP composites successfully restored the yield and ultimate strengths of the beams, delayed the
cracks propagation and reduced their widths, and increased the fatigue life of the strengthened
beams. The enhancement in fatigue life of the FRP-strengthened beams was attributed to the ability
of FRP composites to decrease the stresses in the steel reinforcing bars [10].
Recently, fabric-reinforced cementitious matrix (FRCM) composites were proposed as
promising alternatives to FRP composites. FRCM composites consist of one or more layers of
carbon, glass, or Polyparaphenylene benzobisoxazole (PBO) fabrics that are sandwiched between
layers of cementitious mortars. Unlike FRPs, FRCM composites are believed to have
environmental acceptability, good thermal and fire resistance, and the advantage of application on
wet surfaces and at low temperature. They are also compatible with the concrete substrate and are
characterized by their ease of installation and long-term durability [57,92,93,106].
Multiple studies have documented the effectiveness of FRCM composites in enhancing the
flexural response of undamaged RC structures under monotonic loads [55,58,59,82,85]. However,
very few studies have been devoted to investigate the fatigue performance of the FRCM-
strengthened structures. Aljazaeri and Myers (2015) [87] reported that PBO-FRCM composites
not only improved the fatigue performance of RC beams but also enhanced their residual flexural
strength after fatigue. Moreover, exposing the PBO-FRCM strengthened beams to high
temperature and humidity did not affect their fatigue performance. Yin et al. (2014) [109]
demonstrated that the use of FRCM composite with hybrid (carbon and E-glass) fabric
significantly increased the fatigue life of undamaged RC beams. The enhancement in the fatigue
life was dependent on the fiber reinforcement ratio. More recently, Pino1a et al. (2016) [88]
demonstrated the effectiveness of PBO-FRCM composites in improving the fatigue life of
undamaged RC beams. However, the fatigue life of the strengthened beams decreased by
increasing the maximum applied loads due to the rupture of the steel reinforcement.
129
The feasibility of using FRCM composites to strengthen corrosion-damaged RC structures has
received little attention. Surface deterioration due to corrosion and the absence of a sound concrete
substrate represent major challenges to the application and the long-term durability of the FRCM
composites. Therefore, the flexural behavior of corrosion-damaged structures strengthened with
FRCM composites has rarely been investigated, not to mention their behavior when subjected to
repetitive loads. In a recent study, El-maaddawy and Refai (2016) [86] used basalt and C-FRCM
composites to restore the flexural capacity of RC beams after being subjected to corrosion. The
FRCM composites were either internally embedded within the corrosion-damaged region or
externally-bonded along the beam span. It was reported that externally-bonded composites were
more effective in restoring the ultimate capacities of the corrosion-damaged beams. Nevertheless,
the authors reported that other investigations should be conducted before field applications could
be recommended.
The present study is part of a large research program that aims at investigating the behavior of
corrosion-damaged RC structures after being strengthened with FRCM composites. It provides
insight into the monotonic and fatigue performance of corrosion-damaged beams strengthened
with FRCM composites having different materials and configurations. The test results reported
herein provide unique experimental data and represent the first research work on the use of FRCM
in strengthening corrosion-damaged beams subjected to fatigue.
7.3 Experimental Program
Twelve large-scale RC beams were fabricated and divided into two groups. Group ‘M’ consisted
of six beams that were tested under monotonic loading up to failure whereas group ‘F’ consisted
of six beams that were tested in fatigue. The beams of each group included four corrosion-damaged
beams strengthened with different FRCM composites and configurations. The other two beams
included one beam that was corroded but not strengthened while the other beam was neither
corroded nor strengthened and acted as control.
Table 7.1 and Table 7.2 list the matrix of both the monotonic and fatigue tests, respectively. The
beams were labeled following the A-B-C format. ‘A’ represents the loading regime (F for fatigue
and M for monotonic) and the beam condition (UU, CU, and CS referring to Uncorroded-
Unstrengthened, Corroded-Unstrengthened, and Corroded-Strengthened, respectively). ‘B’
130
denotes the number and type of the FRCM layers applied (2P, 4P, and 3C referring to two layers
of PBO-FRCM, four layers of PBO-FRCM, and three layers of C-FRCM, respectively). Finally,
‘C’ describes the FRCM strengthening schemes (I and II) as will be detailed in the following
sections. For instance, the label FCS-4P-II refers to a corrosion-damaged beam (C) strengthened
(S) with four layers of PBO-FRCM composites (4P) in scheme (II) and tested under fatigue loading
(F).
7.3.1 Test Specimen
The geometry and the reinforcement details of the test specimen are shown in Figure 7.1. All
beams were 2.8 m long with a rectangular cross section 150 mm in width and 250 mm high. The
beams were reinforced with 2-15M deformed steel bars placed at a cover distance of 25 mm from
the soffit. At the top, 2-8M deformed bars were used as stirrups hangers. All beams were provided
with hollow stainless-steel tubes with external and internal diameters of 9.5 mm and 7 mm,
respectively, placed at 80 mm from the soffit of the beam as shown in Figure 7.1. The shear spans
were reinforced with double-leg steel stirrups of 10 mm diameter each spaced at 100 mm in order
to prevent premature shear failure.
Normal and salted concrete mixes having the same water/cement ratio were used to cast the
beams. Concrete cylinders (150×300 mm) were also cast from both mixes to evaluate their
compressive strengths. The average compressive strengths were 41.8 MPa (standard of deviation
of 4.8 MPa) for the normal mix and 41.2 MPa (standard deviation of 0.6 MPa) for the salted one.
The reinforcing steel bars had a nominal yield strength of 400 MPa and an elastic modulus of 200
GPa as reported by the manufacturer. The yield strengths of the longitudinal reinforcing steel bars
of diameter 15 and 8 mm were 466 MPa (with a standard deviation of 4.2 MPa) and 573 MPa
(with a standard deviation of 17.7 MPa), respectively, as tested by the authors.
131
Table 7.1: Monotonic test matrix and results
* Nornalized to the control specimen (MUU)
** SY-CC = Steel Yielding followed by Concrete Crushing; FD = FRCM Delamination; FS-PED = Fabric Slippage followed by Partial Debonding; MC-FS = Matrix Cracking followed by Fabric Slippage.
Table 7.2: Fatigue test matrix and results
* SY-CC = Steel Yielding followed by Concrete Crushing; FRS = Fatigue Rupture of Steel bars
Specimen Avg. mass
loss (%)
𝜌𝑓
(%)
𝐾𝑓
(MPa)
𝜌𝑆
(%)
𝛽𝑓
(%)
𝑃𝑦
(KN)
𝑃𝑢 (KN)
𝑃𝑦
Nor.*
𝑃𝑢
Nor.* εfu
(µϵ)
Mode of
Failure**
MUU - - - 1.07 - 75.1 79.7 1 1 - SY-CC
MCU 18 - - 0.87 - 64.5 74.2 0.86 0.93 - SY-CC
MCS-2P-I 19.6 0.04 48 0.86 2.8 71.8 85.6 0.96 1.07 8180 FD
MCS-4P-I 19.4 0.08 95.3 0.86 5.54 79.6 102.6 1.06 1.29 10659 FD
MCS-4P-II 19.5 0.08 95.3 0.86 5.55 80.7 102.9 1.07 1.29 8253 FS-PFD
MCS-3C-II 18.6 0.19 139.1 0.87 8.01 78.8 123.3 1.05 1.55 5530 MC-FS
Specimen Avg. mass
loss (%)
𝜌𝑓
(%)
𝐾𝑓
(MPa)
𝜌𝑆
(%) 𝛽𝑓 (%)
Fatigue life
(cycles)
Mode of
Failure*
FUU - - - 1.07 - Over 2
million SY-CC
FCU 19.8 - - 0.86 - 396,000 FRS
FCS-2P-I 18.4 0.04 48 0.87 2.76 545,600 FRS
FCS-4P-I 19.3 0.08 95.3 0.86 5.54 984,800 FRS
FCS-4P-II 18.1 0.08 95.3 0.87 5.48 1,493,300 FRS
FCS-3C-II 18.6 0.19 139.1 0.87 7.99 834,500 FRS
132
Figure 7.1: Geometry and reinforcement details of the test specimen (all dimensions in mm)
7.3.2 Accelerated Corrosion Technique
An accelerated corrosion technique was used to corrode the tensile steel reinforcement. A
constant electric current of 380 milliamps was impressed through the steel bars. The current
density was of 180 µA/cm2 to obtain corrosion products that resemble to those found in natural
corrosion process [20]. Salt (NaCl) measured as 5% of the cement weight was added to the
concrete mix used to cast the middle-bottom of the beams at a height of 100 mm as shown in
Figure 7.1. Corrosion of steel bars was restricted to the middle 1200 mm of the beam’s span. The
steel bars and the stirrups outside the salted zone were coated with anti-corrosion epoxy for
protection.
During the corrosion process, the salted concrete acted as electrolyte while the bottom steel bars
and the hollow stainless-steel tube acted as anode and cathode, respectively. The beams were
placed in a large environmental chamber and were connected in series with a DC galvanostatic
power supply to ensure that the induced current is uniform in all specimens (Figure 7.2). The
beams were subjected to wet and dry cycles that consisted of 3 days wet followed by 3 days dry.
The wet and dry cycles provided water and oxygen necessary for the corrosion process. The
accelerated corrosion process lasted for 140 days to obtain a theoretical steel mass loss 20% in the
main reinforcement according to Faraday’s law.
133
Figure 7.2: Specimens connected in series inside the corrosion chamber
7.3.3 FRCM Composites
Two different FRCM composites (PBO and carbon) were used. PBO-FRCM consisted of an
unbalanced PBO-fabric impregnated in a cementitious matrix with a low dosage of dry polymer.
The fabric was made of spaced fiber rovings prearranged along two orthogonal directions as shown
in Figure 7.3a. The width of rovings was 5 mm and 2.5 mm in the main and secondary directions,
respectively and the nominal thickness is 0.046 mm in the main direction, and 0.011mm in the
secondary direction. The gap opening between the rovings was approximately 5 mm in the main
direction and 15 mm in the secondary direction. Table 7.3 lists the properties of the fabric as
reported in the manufacturers’ data sheet. The cementitious matrix of the PBO-FRCM composite
had a compressive strength of 43.9 MPa (standard deviation of 0.4 MPa) and a flexural strength
of 3 MPa (standard deviation of 0.3 MPa) after 28 days as determined by the authors.
The C-FRCM composite consisted of carbon fabric impregnated in an inorganic cementitious
matrix. The fabric was a unidirectional net made of carbon strands oriented in one direction as
shown in Figure 7.3b. The strands were uniformly distributed at a density of 59 strands per meter
width. The properties of the carbon fabric according to the data sheet provided by the manufacturer
are given in Table 7.3. The cementitious matrix had a compressive strength of 42.1 MPa (standard
deviation of 4.3 MPa) and flexural strength of 3.2 MPa (standard deviation of 0.3 MPa) after 28
days as tested by the authors.
134
Figure 7.3: a) Unbalanced PBO fabric and b) Unidirectional carbon fabric.
Table 7.3: Fabric properties in the primary direction as given in the manufacturers’ data sheet
Fabric Area per unit width
𝐴𝑓 (mm2/m) Tensile strength
(GPa)
Elastic modulus
(GPa)
Ultimate strain
(%)
PBO 50 5.8 270 2.15
Carbon 157 4.3 240 1.75
Table 7.4 lists the mechanical properties of the PBO- and C-FRCM composites used in this study
as obtained from direct tensile tests that were conducted on FRCM coupons by Ebead et al. (2016)
[97]. An axial stiffness coefficient, Kf, given by Equation (7.1), characterized the FRCM
composites based on their cracked elastic modulus, 𝐸𝑓, as listed in Table 7.4, and the fiber
reinforcement ratio, 𝜌𝑓 . Accordingly, the coefficient Kf varied with the number of fabric layers
used. Table 7.1 and Table 7.2 list the fiber reinforcement ratio, 𝜌𝑓, and the axial stiffness
coefficient, Kf, for each of the strengthened beams.
𝐾𝑓 = 𝜌𝑓𝐸𝑓 (MPa) Eq. (7.1)
Similarly, the axial stiffness coefficient for steel reinforcement was calculated as follows:
𝐾𝑠 = 𝜌𝑠𝐸𝑠 (MPa) Eq. (7.2)
Where, 𝜌𝑠 is the tensile steel reinforcement ratio and 𝐸𝑠 is the elastic modulus of steel bars (200
GPa). For corrosion-damaged beams, Table 7.1 and Table 7.2 list the values of 𝜌𝑠 for each beam
taking into account the actual mass loss in the steel bars after corrosion. The contribution of FRCM
135
composites to that of steel reinforcement was expressed by the stiffness factor, 𝛽𝑓, as given in
Equation (7.3). Table 7.1 and Table 7.2 list the values of 𝛽𝑓 for each of the strengthened beams.
𝛽𝑓 = 𝐾𝑓/𝐾𝑠 Eq. (7.3)
Table 7.4: Mechanical properties of FRCM coupons (Ebead et al. 2016) [97]
FRCM
composite
Cracked tensile modulus
of elasticity, Ef (GPa)
Ultimate tensile
strength ffu (GPa)
Ultimate strain
εfu (%)
PBO-FRCM 121 1.55 1.4
Carbon-FRCM 75 0.97 1.25
7.3.4 FRCM Strengthening Configuration
Two schemes of FRCM were used in strengthening the corrosion-damaged beams. In both
schemes, one or more flexural plies of FRCM composites were applied to the beam’s soffit over
the middle 2.4 m of its span. The fabric was oriented such as its main direction was parallel to the
longitudinal axis of the beam. In scheme I, U-shaped transverse strips of FRCM of 300 mm width
were used as end anchors for the flexural plies as shown in Figure 6.4a. According to Ombres
(2011) [80] and Hashemi and Al-Mahaidi (2012) [93], the end anchors avoided the premature
delamination of the concrete cover at the beam ends. In scheme II, the beams were wrapped with
a continuous U-shaped FRCM ply as shown in Figure 6.4b. D’Ambrisi and Focacci (2011) [81]
reported that wrapping the flexural plies with a continuous U-shaped ply delayed the premature
delamination of FRCM composite. The main direction of the continuous U-shaped layer of the
PBO fabric was oriented parallel to the longitudinal axis of the beams and therefore contributed to
the flexural capacity of the strengthened beams MCS-4P-II and FCS-4P-II. On the contrary, the
carbon fabric was a unidirectional fabric and therefore, the continuous U-shaped layer was oriented
in the transverse direction and did not contribute to the flexural capacities of the beams MCS-3C-
II and FCS-3C-II.
136
Figure 7.4: FRCM strengthening schemes: a) Scheme I and b) Scheme II
`
a)
b)
137
7.3.5 Strengthening Procedure
Figure 7.5 shows the strengthening procedure of the corrosion-damaged beams with FRCM
composites. At the end of the corrosion process, the deteriorated concrete was first removed using
a hydraulic hammer as shown in Figure 7.5a. The corroded steel bars were then brushed, and the
beams were repaired using a commercially available cementitious mortar (Figure 7.5b). The repair
mortar used had an average compressive strength of 55.4 MPa (standard deviation of 5 MPa) and
flexural strength of 3.4 MPa (standard deviation of 0.3 MPa) as determined by the authors. After
7 days of curing at ambient temperature, the beam surface was roughened by sandblasting. The
FRCM composites were then installed using the hand lay-up method as proposed by ACI 549.4R-
13 [54]. As recommended by the FRCM manufacturers, the beam substrate was soaked in water
for 2 hours before applying the first layer of the cementitious matrix with a thickness of 3 to 5 mm.
The fabric was then installed and coated with a second layer of matrix of similar thickness (Figure
7.5c and 7.5d). The procedure was then repeated until the specified number of layers was achieved.
All of the strengthened beams were cured for 28 days in laboratory conditions before being tested.
Figure 7.5: FRCM strengthening procedure: a) removing the deteriorated concrete after
corrosion, b) patch repair, c) PBO-FRCM application, and d) C-FRCM application
138
7.3.6 Test Setup and Instrumentation
The beams were tested in a four-point bending configuration using a MTS actuator as shown in
Figure 7.6. A steel beam was used to apply the load equally to two loading points placed 800 mm
apart. The beams had an effective span of 2.8 m and shear spans of 880 mm. Additional supports
were used to prevent the out-of-plane movement of the beam during cycling (Figure 7.6). All
beams were instrumented with 60 mm strain gauges bonded to the top surface of concrete in the
compression zone and 5 mm strain gauges bonded to the tensile steel bars at mid-span. Moreover,
the strengthened beams were instrumented with 5 mm strain gauges installed directly on the outer
fabric of the FRCM composite at mid-span and under the point loads. Beam deflections were
measured by means of three linear variable differential transducers (LVDTs) located at mid-span
and under the point loads. A 20-channel data acquisition system was used to capture the readings
of the measuring gauges.
All beams of group M were tested to failure under monotonic loading under displacement control
at a rate of 2 mm/min. The beams of group F were subjected to cyclic loading at a frequency of 2
cycles per second. The minimum and maximum fatigue loads were 17 and 48 KN, respectively,
representing 21 and 60% of the load carrying capacity of the control beam. This load range was
not only chosen to simulate the actual service loading in the field, but also to challenge the fatigue
strength of the corroded steel reinforcement and FRCM composites. The beams were subjected to
fatigue until failure occurred or 2 million cycles were reached. Beams that lasted 2 million cycles
were tested to failure under monotonic loading to determine their residual strengths.
Figure 7.6: Test setup
Spreader beam
Auxiliary support LVDTs
139
7.4 Test Results and Discussion
7.4.1 Corrosion Crack Patterns and Actual Steel Mass Loss
Figure 7.7 shows the crack patterns at the end of the corrosion process for Beam FCU.
Longitudinal cracks formed on the beam sides along the corroded bars and on the beam soffit. The
widths of the corrosion cracks were measured and recorded for all of the damaged beams. The
maximum and average crack widths were 2.8 and 1.28 mm, respectively. It was therefore
concluded that all beams failed to meet ACI 318-14 [98] serviceability requirements that limit the
maximum crack width for a beam in service to 0.40 mm.
Figure 7.7: Corrosion crack pattern for specimen FCU
To determine the actual steel mass loss, the corroded steel bars were extracted after testing.
Several corrosion pits were observed to be randomly distributed on the bars’ surfaces as shown in
Figure 7.8. A steel brush was used to remove the adhered mortar around the bar. The bars were
then cut into coupons of 200 mm length each. The coupons were chemically cleaned and the mass
loss was determined according to ASTM G1-03 provisions [26]. Table 7.1 and Table 7.2 list the
average mass loss as determined for all of the corrosion-damaged beams. An average mass loss of
19% was reported, which was 1% less than that predicted by Faraday’s law.
140
Figure 7.8: Profile of the steel bars: a) Uncorroded bar, b) corroded steel bar extracted from
beam MCS-4P-II, and c) fatigue rupture of a corroded steel bar extracted from FCS-3C-II
7.4.2 Monotonic Test Results
7.4.2.1 Modes of Failure
Figure 7.9 illustrates the modes of failure of the beams tested under monotonic loads. The control
Beams MUU and MCU failed due to yielding of the steel bars followed by concrete crushing in
compression (Figure 7.9a). On the other hand, PBO-strengthened Beams MCS-2P-I and MCS-4P-
I failed due to the delamination of the FRCM composites at the fabric/matrix interface. This mode
of failure can be depicted in Figure 7.9b for Beam MCS-4P-I. From the test observations, FRCM
delamination was attributed to the propagation of the flexural cracks to the matrix layers adjacent
to the concrete substrate.
The use of continuous U-shaped layer of PBO-FRCM in Scheme II changed the mode of failure
of Beam MCS-4P-II. As the load increased, cracks formed in the matrix of the continuous U-
shaped layer followed by gradual slippage of the fabric from the surrounding matrix. Close to
ultimate, partial debonding of the fabric was observed at the fabric/matrix interface (Figure 7.9c).
On the other hand, Beam MCS-3C-II strengthened with three layers of C-FRCM in Scheme II
failed due extensive cracking in the matrix of the FRCM composite as the applied load increased.
Matrix cracking resulted in slippage of the carbon fabric from the matrix as shown in Figure 7.9d.
c)
a)
b)
141
This mode of failure explained the sudden failure of Beam MCS-3C-II when compared to that of
Beam MCS-4P-II.
As previously stated, fabric strains were measured by means of strain gauges located at midspan.
Table 7.1 lists the ultimate fabric strain, εfu, for all of the strengthened beams. Strain values should
be looked at based on the ultimate capacity and the mode of failure encountered for each beam.
The ultimate fabric strains reported for the PBO-strengthened beams ranged between 8180 and
10659 μɛ, while the strains reported at ultimate for Beam MCS-3C-II was only 5530 μɛ, which
was consistent with the mode of failure of the Beam MCS-3C-II where premature matrix cracking
and fabric separation were observed.
Figure 7.9: Failure modes of the monotonically tested beams: a) Steel yielding followed by
concrete crushing; b) FRCM delamination; c) Fabric slippage and partial debonding; and d)
Matrix cracking followed by fabric slippage
7.4.2.2 Load-deflection and Ultimate Strengths
The load-deflection response of the monotonically tested beams (Group M) are shown in Figure
7.10. The yield load, 𝑃𝑦, and the ultimate load, 𝑃𝑢, of the tested beams are listed in Table 7.1.
Corrosion of steel bars had a marginal effect on the load-deflection response of the unstrengthened
Beam MCU. However, the yield and ultimate strengths after corrosion were reduced by 14 and
7%, respectively.
142
Figure 7.10: Load-deflection relationships of the monotonically tested beams
The use of FRCM composites enhanced the flexural response of the corrosion-damaged beams
with a notable increase in their load-carrying capacities. The strengthened beams showed post-
yielding stiffness that was higher than that of the unstrengthened beams. This finding was
concluded from the slope of the load-deflection curves of Beams MCS-2P-I, MCS-4P-I, MCS-4P-
II, and MCS-3C-II (Figure 7.10). The enhancement in the post-strengthening flexural response
was highly dependent on the type and amount of the applied FRCM composites as follows.
Beam MCS-2P-I, with 𝛽𝑓 = 2.8%, reached an ultimate capacity of 85.6 kN, which corresponds
to a strength gain of only 7% when compared to the ultimate capacity of the control beam. On the
other hand, Beam MCS-4P-I with 𝛽𝑓 = 5.54% (almost double that of Beam MCS-2P-I), reached
an ultimate capacity of 102.6 KN, which was 29% higher than that of the control specimen. These
results indicated that the contribution of FRCM composites increased with an increase in the
number of fabric layers. However, the increase in strength was not linearly proportional to the
increase in the stiffness factor, 𝛽𝑓. This finding was confirmed by comparing the results of Beams
MCS-4P-II and MCS-4P-I having similar 𝛽𝑓 of 5.54%. Both beams showed similar ultimate loads
despite the different schemes used.
The stiffness factor, 𝛽𝑓, should not be used solely to compare the strengthening effectiveness of
different FRCM composites without considering their cementitious matrix bond characteristics.
For instance, Beam MCS-3C-II with 𝛽𝑓 = 8.01% that is 144 % of that of Beam MCS-4P-II, reached
143
an ultimate strength of 123.3 KN compared to 102.9 kN for Beam MCS-4P-II. This variation in
the ultimate strength of Beam MCS-4P-II from that of Beam MCS-3C-II (approximately 20%)
compared to the large difference between their stiffness factors (𝛽𝑓) was attributed to the inferior
properties of the cementitious matrix of the C-FRCM with respect to those of the PBO-FRCM
counterparts.
7.4.3 Fatigue Test Results
7.4.3.1 Fatigue Life
The fatigue test results of beams of Group F are shown in Table 7.2. The virgin Beam FUU
successfully survived 2 million cycles without failure. The test was then halted and the beam was
tested monotonically up to failure to examine its residual strength as will be detailed later.
Corrosion of steel bars reduced the fatigue life of the unstrengthened Beam FCU to 396,000 cycles,
that is only 20% of the fatigue life of Beam FUU.
The use of FRCM composites increased the fatigue life of the corrosion-damaged beams.
However, none of the FRCM-strengthened beams could restore the fatigue life of the virgin beam.
Beam FCS-2P-I had the shortest fatigue life among the FRCM-strengthened beams with 545,600
cycles survived. Beams FCS-4P-I and FCS-4P-II strengthened with the same amount of PBO-
FRCM (same 𝛽𝑓) but with different strengthening schemes survived 984,800 and 1,493,300 cycles,
receptively. On the other hand, Beam FCS-3C-II survived 834,500 cycles before failure.
Figure 7.11 shows the variation of the number of fatigue cycles survived and the stiffness factor,
𝛽𝑓 , for the tested beams. It was noted that increasing 𝛽𝑓 from 2.76 to 5.54 % for beams FCS-2P-I
and FCS-4P-I, respectively, increased the fatigue life by 80% (545,600 cycles for the former beam
versus 984,800 cycles for the later one). It is worth mentioning that both beams were strengthened
with PBO-FRCM in Scheme I. On the contrary, Beam FCS-3C-II had 𝛽𝑓 of 7.99, which was 146
that of Beam FCS-4P-II. However, the fatigue life of the former was 80% shorter than that of the
later (834,500 cycles for Beam FCS-3C-II versus 1,493,300 cycles for Beam FCS-4P-II). While
this finding can be attributed to the scatter that commonly encountered in fatigue tests, it matched
well the monotonic results of both beams and confirmed the conclusion previously obtained that
the stiffness factor, 𝛽𝑓 , should not be used solely to compare the strengthening effectiveness of
different FRCM composites without taking into account their matrix properties. It also confirmed
144
the superior characteristics of the PBO-FRCM composite compared to those of C-FRCM
counterpart. Beam FCS-4P-II demonstrated exceptional bond characteristics between the fabric
and the surrounding matrix and also between the matrix and the concrete substrate. Good bond
ensured the transfer of stress to the FRCM composite thus reducing the stresses in the reinforcing
bars, which allowed the beam to restore about 75% of the fatigue life of the virgin beam and to
exceed that of the corrosion-damaged beam by 277%.
PBO-strengthened beams having similar values of 𝛽𝑓 but with different configuration of their
FRCM composite showed significant scatter between the number of cycles survived. This was
shown from the test results of beams FCS-4P-I and FCS-4P-II having 𝛽𝑓 of 5.54. The use of U-
shaped continuous layer of FRCM in Beam FCS-4P-II improved its fatigue life by nearly 52% of
that of Beam FCS-4P-I. The enhancement in fatigue life of Beam FCS-4P-II was ascribed to the
confinement effect of the U-shaped layer and its contribution in reducing the stress level in the
steel bars. Recall that both beams showed similar load-carrying capacities when tested under
monotonic loading, which suggested that the effectiveness of the U-shaped continuous layer of
Scheme II was more pronounced in fatigue rather than in monotonic tests.
Figure 7.11: Variation of fatigue life of the strengthened beams with the stiffness factor 𝛽𝑓
145
7.4.3.2 Fatigue Behavior and Modes of Failure
The control Beam FUU was tested monotonically to failure after surviving 2 million cycles. The
beam failed due to steel yielding followed by concrete crushing. The beam exhibited a load-
deflection response similar to that of reference MUU specimen with residual deflections at the
onset of loading and reduced stiffness as a result of the cracks formed during the cyclic loading.
The residual yield and ultimate load capacities of Beam FUU were 73.4 and 79.3 KN, respectively,
compared to 75.1 and 79.7 kN for Beam MUU.
Figure 7.12a to Figure 7.12d present the hysteresis loops for the unstrengthened Beam FCU and
for the strengthened Beams FCS-4P-I, FCS-4P-II, and FCS-3C-II, respectively. Test observations
indicate that the fatigue response of all tested beams consisted of three stages. The first stage is
characterized by the initiation of vertical cracks accompanied by the increase in midspan
deflections. In this early stage of the fatigue life, flexural cracks tend to increase in height and
number before they stabilize.
The second stage describes most of the fatigue life of the beams. In this stage, the growth of
cracks is remarkable due to the reduction in modulus of elasticity of concrete as reported by several
authors [52]. As the number of cycles increase, a primary flexural crack (referred to as the fatigue
crack) appeared in the region of maximum moment. Experimentally, the fatigue crack steadily
propagate towards the compression zone until failure occurs at its location. This phenomenon was
common in beams strengthened with both schemes. However, the fatigue crack in beams
strengthened with Scheme II was smaller in width than that observed in beams strengthened with
Scheme I. Fatigue cracks for Beams FCS-4P-I, FCS-4P-II and FCS-3C-II are depicted in Figure
7.13. Small dispersed cracks were also observed in the FRCM cementitious matrix without any
evidence of fabric slippage or FRCM delamination. It was also notable that FRCM-strengthened
beams showed more gradual and steady degradation in their fatigue strength during this stage as
compared to that of the unstrengthened Beam FCU. This might be related to the potential of FRCM
composites to mitigate the crack propagation during cycling.
As per Figures 7.12a through 7.12d, the final stage of the fatigue life is characterized by a sudden
degradation in the beams’ stiffness and a considerable increase in deflection. Failure occurred
suddenly due to the abrupt rupture of one of the bottom steel bars. All of the strengthened beams
failed by fatigue rupture of the corroded steel bars as shown in Figure 7.14 for Beams FCU and
146
FCS-4P-I. Rupture of the steel bars in the FRCM-strengthened beams led to the complete falling-
out of the FRCM composite due to the excessive deformation occurred. The visual inspection of
the extracted corroded bars after failure revealed that rupture of steel bars occurred at the location
of the corrosion pits where the greatest mass loss occurred (Figure 7.8c).
a)
b)
0
10
20
30
40
50
60
0 2 4 6 8 10
Load
(K
N)
Deflection (mm)
0 cycles (0%) 7200 cycle (2%)
100800 cycles (25%) 302400 cycles (75%)
396000 cycle (100%)
0
10
20
30
40
50
60
0 5 10 15
Lo
ad (
KN
)
Deflection (mm)
0 cycles (0%) 7200 cycle (0.7%)201600 cycles (20%) 604800 cycles (60%)806400 cycle (80%) 984800 cycles (100%)
147
c)
d)
Figure 7.12: Load-deflection Hysteresis for a) beam FCU; b) beam FCS-4P-I; c) beam FCS-4P-
II; and d) beam FCS-3C-II
0
10
20
30
40
50
60
0 5 10 15
Lo
ad (
KN
)
Deflection (mm)
0 cycles (0%) 7200 cycles (0.5%)201600 cycles (14%) 806400 cycles (54%)1,209600 cycles (81%) 1,493000 cycles (100%)
0
10
20
30
40
50
60
0 5 10 15
Load
(K
N)
Deflection (mm)
0 cycles (0%) 7200 cycle (0.9%)201600 cycles (24%) 403200 cycles (48%)720000 cycles (86%) 834500 cycle (100%)
148
Figure 7.13: Fatigue cracks at midspan of the strengthened beams (Side view)
Figure 7.14: Fatigue rupture of steel bars in a) beam FCU and b) beam FCS-4P-I
7.4.3.3 Stiffness Degradation
Figure 7.15 shows the variation in stiffness of the tested beams with the number of fatigue cycles
survived. The flexural stiffness of the tested beams is calculated as the ratio of the maximum
fatigue load to the corresponding midspan deflection. The initial stiffness for all tested beams were
measured after 7,200 cycles. Test results show that the virgin Beam FUU lost 5% of its initial
stiffness after 200,000 cycles. The beam suffered from gradual and steady stiffness degradation as
the number of cycles increased. After 2 million cycles, the beam lost about 10% of its initial
stiffness. Beam FCU exhibited a similar trend of stiffness degradation before failing suddenly after
396,000 cycles with flexural stiffness about 84% of its initial stiffness.
Fatigue cracks (0.25 mm) after 0.98M
cycles
Fatigue cracks (0.12mm) after 1.49M
cycles
Fatigue cracks (0.15mm) after 0.83M
cycles
a) Specimen FCS-4P-I b) Specimen FCS-4P-II c) Specimen FCS-3C-II
Bar rupture Bar rupture
149
Figure 7.15: Effect of fatigue cycles on the stiffness of the tested beams
FRCM-strengthened beams showed initial stiffness that was comparatively higher than that of
the unstrengthened ones. As the number of cycles increased, different rates of stiffness degradation
were encountered. The average rate of stiffness degradation in the beams strengthened with PBO-
FRCM was 3.33% per 100,000 cycles for the first 200,000 cycles of fatigue loading. This rate then
decreased to 0.64% per 100,000 cycles until a significant loss of stiffness was observed following
the rupture of one of the steel bars.
Test results indicate that the type of the FRCM composite rather than the number of fabric layers
and the strengthening scheme have a notable effect on the rate of stiffness degradation. This was
depicted in the comparable plots shown in Figure 7.15 for Beams FCS-2P-I, FCS-4P-I, and FCS-
4P-II, having different number of fabric layers and/or different schemes. On the other side, beam
FCS-3C-II demonstrated a higher rate of stiffness degradation than that recorded for the PBO-
FRCM strengthened beams. This finding might be attributed to the superior characteristics of the
PBO-FRCM composite as previously demonstrated in the monotonic test results. It also suggests
that the type of FRCM composite played a major role in defining the stiffness degradation during
fatigue.
150
7.4.3.4 Strain Response
Figure 7.16 shows the variation of strains in concrete and in the outer fabric with the number of
cycles survived. All of the tested beams showed a gradual increase in the measured concrete strains
with an increase in the number of cycles. This is attributed to the concrete softening that results in
an increase in the beam deflections as illustrated in Figure 7.12. The maximum compressive
concrete strains ranged between -763 to -961 μɛ for Specimen FCU while the maximum strains
recorded for the strengthened beams ranged between -603 to -984 μɛ. Similarly, the outer fabric
strains increased during the fatigue life due to the propagation of cracks until failure occurred.
Figure 7.16: Effect of fatigue cycles on the concrete and fabric strains
The maximum fabric strains measured for Beam FCS-2P-I increased from 2707 μɛ at 7200 cycle
to 2827 μɛ at 500,000 cycles and reached 4051 μɛ at failure. Beam FCS-4P-I showed maximum
fabric strains that grown from 2350 μɛ at 7200 cycle to 3875 μɛ at failure. Beam FCS-4P-II
exhibited relatively lower initial fabric strains than that of Beam FCS-4P-I (1972 versus 2350 μɛ)
while the strains reached 2791 μɛ after 1,493,000 cycles. The maximum carbon fabric strains in
Specimen FCS-3C-II were lower than those recorded for the PBO-FRCM and ranged between
1480 to 1850 μɛ. It is important to note that the measured stains in concrete and FRCM did not
exceed their ultimate strain values. These results indicate that the residual fatigue strength of the
steel bars after corrosion control the fatigue strength of the strengthened RC beams rather than the
151
fatigue strength of concrete or FRCM composites. This assumption is valid as long as the
externally bonded composites are able to reduce the stress level in the corroded steel bars to that
level of stress prior to corrosion.
7.5 Conclusions
In this study, the monotonic and fatigue performances of corrosion-damaged RC beams
strengthened with PBO- and C-FRCM composites were experimentally investigated. The
following conclusions can be drawn from the test results:
Monotonic Tests
• An average mass loss of 19% in the steel reinforcing bars due to corrosion reduced the
yield and ultimate strengths of the beams by 14 and 7%, respectively, without changing its
ductile mode of failure.
• Strengthening corrosion-damaged beams with PBO- and C-FRCM composites enhanced
the flexural response and restored/exceeded the load capacities of the virgin beam. The
ultimate loads of the PBO-strengthened beams ranged between 107 and 129% of that of
the control beam and was 155% for the carbon-strengthened beam.
• The modes of failure of the FRCM-strengthened beams varied with the type and scheme
of the FRCM composite used. PBO-strengthened beams failed by fabric delamination in
scheme I and by fabric slippage and debonding in scheme II. Carbon-strengthened beams
failed by the premature cracking of the matrix followed by the fabric slippage.
• The contribution of FRCM composites, expressed by the stiffness factor 𝛽𝑓, increased with
an increase in the number of fabric layers. However, the associated gain in capacity was
not linearly proportional to the increase in stiffness factor, 𝛽𝑓.
• The stiffness factor 𝛽𝑓 should not be used solely to compare the strengthening effectiveness
of different FRCM composites without considering the mechanical properties of its
constituents. Corrosion-damaged beams strengthened with PBO- and C-FRCM composites
having same 𝛽𝑓 showed different load-carrying capacities and different modes of failure.
152
Fatigue Tests
• Corrosion of steel bars dramatically decreased the fatigue life of the RC beam. While the
virgin beam sustained 2 million fatigue cycles, the corrosion-damaged beam failed after
396,000 cycles. Therefore, corrosion shortened fatigue life to 20% of that of the virgin
beam.
• Rupture of steel bars at the locations of the highest corrosion pits intensity was the
governing mode of failure for all of the unstrengthened and strengthened beams.
• Strengthening with FRCM composites increased the fatigue life of the corrosion-damaged
beams by 38 to 377% of that of the unstrengthened beam depending on the type and
configuration of the FRCM composite used. However, strengthening did not restore the
fatigue life of the virgin beam.
• PBO-FRCM composite was more effective than the C-FRCM counterpart in restoring the
fatigue life of the corrosion-damaged beams. Beam FCS-3C-II (with 𝛽𝑓 = 7.99) survived
80% less cycles than those survived by beam FCS-4P-II (with 𝛽𝑓 =5.48).
• The effect of FRCM configuration was more pronounced in fatigue than in monotonic tests.
Having the same number of PBO fabric layers, beam FCS-4P-II strengthened with scheme
II (U-shaped continuous layer) exhibited longer fatigue life than beam FCS-4P-I
strengthened with scheme I (U-shaped end-anchors). Both beams showed similar
monotonic load-carrying capacities.
• The type of the FRCM composite had a notable effect on the rate of stiffness degradation
of the strengthened beams in fatigue rather than the number of fabric layers and the
strengthening scheme applied. The beam strengthened with C-FRCM demonstrated a
higher rate of stiffness degradation with cycling than the beam strengthened with PBO-
FRCM.
The test results presented in this paper demonstrate that strengthening with FRCM composites is
an effective technique to restore the fatigue life of corrosion-damaged concrete beams. However,
the outcome of this study should not be extrapolated to other FRCM composites before being
153
validated with experimental tests. Future work should investigate the fatigue performance of other
FRCM composites and for various levels of corrosion of the reinforcing steel bars.
154
8. Chapter 8
Finite Element Modeling and Experimental Results of Corroded Concrete Beams Strengthened with Externally-
bonded Composites
Mohammed Elghazy, Ahmed El Refai, Usama Ebead, and Antonio Nanni
Journal of Engineering Structures. Date of submission: November 4, 2017
Status: Under review
Résumé
Cet article présente les résultats de la modélisation en éléments finis 3D (ÉF) de poutres en béton
armé endommagées par la corrosion et renforcées en flexion avec des composites collés à
l'extérieur. Les modèles ont été validés par rapport aux résultats d'essais expérimentaux effectués
sur dix poutres non renforcées et renforcées. Les paramètres étudiés comprenaient les niveaux de
corrosion (10% et 20% de perte de masse de l'armature en acier), le type de composite (matrice
cimentaire renforcée de fibre, MCRF et polymère renforcé de fibre, PRF) et le nombre de couches
de fibres (une, deux et quatre couches). Les résultats prédits ont montré un bon accord avec ceux
des tests expérimentaux. Les modèles ÉF ont pu capturer le comportement non linéaire des poutres.
Les modèles de lien-glissement aux interfaces des MCRF/matrice et PRF/béton et le nombre de
couches de fibres ont eu l'impact le plus significatif sur la réponse prédite des poutres renforcées
alors que le niveau de corrosion, modélisé comme une réduction de la section des armatures, a
montré un léger effet sur leur performance. Une étude paramétrique a examiné l'effet de la
variation de la résistance du béton à la compression sur la performance des poutres renforcées.
L'abaissement de la résistance à la compression du béton a diminué les capacités ultimes des
poutres renforcées par les composites, quel que soit le système de renforcement utilisé (MCRF ou
PRF). Le mode de défaillance des poutres renforcées par MCRF était indépendant de la résistance
à la compression du béton et était gouverné par le glissement des fibres de la matrice. Cependant,
pour les poutres renforcées en PRF, l'abaissement de la résistance à la compression du béton a
modifié leur mode de rupture, de la rupture de PRF au décollement à l'interface PRF/béton.
Mots-clés des auteurs : Corrosion; Matrice cimentaire renforcée de fibres; Polymères renforcés
de fibres; Analyse des éléments finis; Flexion; Béton armé; Réparation; Renforcement.
155
8.1 Abstract
This paper presents results of 3D finite element (FE) modeling of corrosion-damaged reinforced
concrete (RC) beams strengthened in flexure with externally-bonded composites. The models were
validated against the results of experimental tests conducted on ten unstrengthened and
strengthened beams. The investigated parameters included the corrosion levels (10% and 20%
mass loss of steel reinforcement), the type of composite (fabric-reinforced cementitious matrix
(FRCM) and fiber-reinforced polymers (FRP)), and the number of fabric layers (one, two, and four
layers). The predicted results showed good agreement with those of the experimental tests. The FE
models were able to capture the non-linear behavior of the beams. The interfacial bond stress-slip
models at the FRCM/matrix and FRP/concrete interfaces and the number of fabric layers had the
most significant impact on the predicted response of the strengthened beams whereas the corrosion
level, modeled as a reduction in the reinforcement cross-section, showed a slight effect on their
performance. The validified models were used in parametric studies to investigate the effect of
varying the compressive strength of concrete substrate and the thickness of concrete cover on the
flexural performance of the strengthened beams. Lowering the compressive strength of the
concrete substrate or increasing the thickness of the concrete cover decreased the load-carrying
capacities of the strengthened beams regardless of the strengthening system used (FRCM or FRP).
Failure of FRCM-strengthened beams was independent of the compressive strength of concrete or
the thickness of concrete cover and was governed by fabric slippage within the matrix, unlike in
the case of FRP-strengthened beams.
Authors’ keywords: Corrosion; Fabric-reinforced cementitious matrix; Fiber reinforced
polymers; Finite element analysis; Flexure; Reinforced concrete; Repair; Strengthening.
8.2 Introduction and Background
Strengthening of corrosion-damaged reinforced concrete (RC) structures has become one of the
most imperative activities in the construction industry. Corrosion impairs the structural integrity
and the serviceability of the structures and can lead to unexpected collapses [44,45,53]. In the past
decades, previous research has documented the effectiveness of fiber-reinforced polymers (FRP)
as reliable strengthening materials for concrete structures [10,11,111]. More recently, cement
mortars reinforced with fabrics made of carbon, glass, or polyparaphenylene benzobisoxazole
156
(PBO), known as fabric-reinforced cementitious matrix (FRCM) or textile-reinforced mortar
(TRM), have been introduced as promising, sustainable, and durable alternatives to FRP
composites. FRCM systems own all the merits of FRPs in terms of corrosion resistivity, light
weight, and ease of installation but with the use of inorganic cementitious matrices as bonding
materials in lieu of the epoxy-bonding agents for FRPs to overcome some of the drawbacks
associated with epoxies [16,59,103,112,113].
Finite element packages have been used to examine the structural performance of strengthened
concrete beams and slabs [114–116] with the majority of the literature body focusing on modelling
the behavior of beams strengthened using FRP sheets or plates. The bond behavior between
concrete and FRP have been developed through rigorous numerical modelling of concrete-FRP
joints in direct shear tests, a basic application that provides insight into FRP-concrete interfacial
behavior [117,118]. The motivation for such existing numerical work was the fact that, despite the
large amount of experimental data available on FRP strengthening of concrete structures, a full
understanding of the various load-deformation behaviors and debonding phenomenon was still
lacking.
On the other hand, recent experimental studies involving FRCM systems have focused on
strengthening undamaged RC members. Test results have demonstrated their efficiency in
restoring the capacities and serviceability of the deficient members [55,81,119]. Elsanadedy et al.
(2013) [83] developed a FE model to predict the flexural behavior of six beams strengthened with
basalt FRCM (B-FRCM) and carbon-FRP systems (C-FRP) using LS-DYNA software. Bond
between FRCM and concrete was modeled through the tiebreak surface-to-surface contact
definition of LS-DYNA to account for both normal and shear forces at the interface. A parametric
study was also conducted by altering the type of mortar and the number of B-FRCM layers used.
A bond-stiffness coefficient, defined as the ratio of the B-FRCM stiffness to its tensile bond
strength, was introduced and recommended not to be less than 225 to avoid premature debonding
failure.
Al-Salloum et al. (2012) [92] evaluated both experimentally and numerically the efficacy of B-
FRCM system for shear strengthening of deficient concrete beams using long woven, knitted or
even unwoven fiber rovings in two orthogonal directions. A prefect bond between FRCM
composites and concrete was assumed in the FE model. The number of textile layers and the
157
orientation of the textile material were altered through a parametric study using FE models. The
models were able to accurately capture the shear strengths of the strengthened beams.
It was observed that most of the previous studies have reported that debonding of FRCM
composites governs the failure of the flexural strengthened RC members [80,84,85], which
highlights the significance of bond of FRCM to the concrete substrate. Debonding of FRCM
usually occurs at the fabric/matrix interface rather than the matrix/concrete interface, unlike what
is typically reported for FRPs, in which debonding occurs at the epoxy/concrete interface
[72,76,117]. D’Ambrisi et al. [120] developed bond-slip models to describe bond between FRCM
fabric and the surrounding matrix. These models were then calibrated with data obtained from
experimental tests in which RC elements were strengthened with PBO-FRCM layers. This model
will be described later in the FE model developed in this study. Ombers [75] proposed another
bond-slip relationship for PBO-FRCM composites bonded to concrete. However, Ombres’ model
was reported to be more conservative than that introduced by D’Ambrisi et al. [120].
The conducted literature shows that little attention has been devoted to investigating the
feasibility of using FRCM composites in strengthening corrosion-damaged RC members.
Corroded structures are characterized by the deterioration of concrete and the loss of structural
integrity as a result of expansive corrosion products. Using cement-based mortars in repair might
be a challenge from the practical and technical points of view. To the authors’ knowledge, only
two studies [86,121] have documented the potential of using FRCM composites to restore the
ultimate capacities and the serviceability of corrosion-damaged beams. However, many
parameters that might affect the performance of FRCM-strengthening have not been fully
documented and yet need to be thoroughly investigated.
The aim of the current study was twofold, namely:
a) expand the understanding of flexural behavior of corrosion-damaged RC beams strengthened
with FRCM composites by i) including parameters that were not included in previous studies
[86,121] and ii) comparing their performance with that of FRP-strengthened beams.
b) validate newly-developed FE models that utilize the bond stress-slip model proposed by
D’Ambrisi et al. [120] to describe the bond behavior between PBO-FRCM and concrete.
158
Details about the test specimens, the accelerated corrosion process, the test setup, and the test
results of some of the beams have been reported in [121] and are reiterated here for convenience.
The FE models were then extended in parametric studies to examine the effect of varying the
compressive strengths of the concrete substrates and the thickness of concrete cover on the flexural
enhancement of the strengthened beams.
8.3 Experimental Investigation
8.3.1 Test Matrix
The test matrix of the experimental program is given in Table 8.1. The beams were subdivided
into two groups A and B and were subjected to accelerated corrosion process to obtain theoretical
mass losses of 10 and 20%, respectively, in the middle third of their tensile steel reinforcement.
Details about the accelerated corrosion process can be found in [121]. At the end of the corrosion
process, one beam in each group was not strengthened (beams CU-A and CU-B) and were used as
benchmarks while other beams were strengthened with the designated externally-bonded
composite system. In addition, two virgin beams (i.e., not corroded nor strengthened: beams UUa
and UUb) were used as controls. The test parameters included the level of corrosion damage (10
and 20%), the type of the externally-bonded composite system (FRCM and FRP), and the volume
fraction of the fabric used in the FRCM composite (1, 2, or 4 layers).
8.3.2 Test Specimen
The geometry and details of reinforcement of the test specimen are shown in Figure 8.1. The
beams were 2.8 meters long with a cross section of 150×250 mm and a clear span of 2.2 meters.
The bottom and top reinforcement consisted of two 15M (diameter 15 mm) and 8M (diameter 8
mm) deformed bars, respectively, placed at a clear cover of 25 mm. Steel stirrups of 10 mm
diameter were provided along the shear spans to avoid shear failure. After testing the beams, the
corroded steel bars were extracted from the beams and the actual mass loss of the corroded bars
were determined according to ASTM G1-03 provisions [26]. Table 8.1 lists the actual average
steel mass loss for each beam.
The cylinder compressive strength of concrete used was 41.8 MPa with standard deviation of 4.8
MPa. The splitting tensile strength of concrete was 3 MPa with standard deviation of 0.3 MPa.
159
The yield strength of the longitudinal reinforcing steel bars of diameters 15 and 8 mm were 466
MPa (standard deviation of 4.2 MPa) and 573 MPa (standard deviation of 17.7 MPa) as determined
by the authors.
Table 8.1: Test matrix
Beam ID* Avg. mass loss
(%)
As
(mm2)
Strengthening regime
Reference Type
No. of
layers
Af
(mm2)
𝐾𝑓
(KN)
Control beams: Uncorroded unstrengthened beams
UUa, UUb - 400 - - - - [121]
Group (A): Theoretical steel mass loss of 10%
CU-A 12.9 348.4 - - - - [121]
CS-A-1C 11.6 353.6 C-FRP 1 19.5 4485 -
CS-A-1P 13.7 345.2 P-FRCM 1 6.9 1421 -
CS-A-2P 13 348 P-FRCM 2 13.8 2842 [121]
CS-A-4P 12.6 349.6 P-FRCM 4 27.6 5684 [121]
Group (B): Theoretical steel mass loss of 20%
CU-B 18 328 - - - - [121]
CS-B-2P 19.6 321.6 P-FRCM 2 13.8 2842 [121]
CS-B-4P 19.4 322.4 P-FRCM 4 27.6 5684 [121]
* UU, CU, and CS refer to Uncorroded-Unstrengthened, Corroded-Unstrengthened, and Corroded-Strengthened beams,
respectively. A and B refer to beam groups A and B, respectively. 1, 2, and 4 in the beam’s label refer to the number of
composite layers. P and C refer to P-FRCM and C-FRP, respectively.
Figure 8.1: Geometry and details of steel, P-FRCM, and C-FRP reinforcement of the tested
beams (all dimensions in mm)
160
8.3.3 Externally-bonded Composite Systems
Two types of externally-bonded composites were used to strengthen the beams namely, the P-
FRCM composite and the C-FRP composite. The P-FRCM composite consisted of a fabric made
of polyparaphenylene benzobisoxazole (PBO) fibers impregnated and bonded to the concrete
surface by a polymer-modified cementitious matrix having compressive and flexural strengths of
43.9 and 3 MPa (standard deviations of 0.4 and 0.3 MPa), respectively, as determined by the
authors. The PBO fabric consisted of an unbalanced net of spaced fiber bundles organized along
two orthogonal directions as shown in Figure 8.2a. The fabric openings were 5 and 15 mm wide
in the primary and the secondary directions, respectively. The bundles were 5 and 2.5 mm wide
with thicknesses of 0.046 and 0.011 mm in the main and secondary directions, respectively. The
tensile tests conducted by D’Antino et al. [78] on the same fabric indicated a tensile modulus of
206 GPa, a tensile strength of 3014 MPa, and an ultimate elongation of 14.5‰. The P-FRCM
composite had a cracked tensile modulus of 121 GPa, a tensile strength of 1550 MPa, and an
ultimate elongation of 14‰ as determined by Ebead et al. [97].
Figure 8.2: a) PBO fabric used in the FRCM composite system and b) Carbon sheets used in the
FRP composite system
On the other hand, the C-FRP composite consisted of flexible unidirectional carbon fiber sheet
(Figure 8.2b) impregnated and bonded to concrete with an epoxy resin having a tensile modulus
of 3.8 GPa and a tensile strength of 30 MPa according to the manufacturer’s data sheet. The dry
carbon fiber sheet had a tensile modulus of 230 GPa, a tensile strength of 3.22 GPa, an ultimate
elongation of 14.5‰, and a nominal thickness of 0.13 mm. The cured laminate had a tensile
161
modulus of 65.4 GPa, a tensile strength of 0.894 GPa, an ultimate elongation of 13.3‰, and a
nominal thickness of 0.38 mm, as reported in the manufacturer’s data sheet.
To compare between the two strengthening systems, the equivalent axial stiffness, 𝐾𝑓, for each
composite system was determined based on their tensile modulus, 𝐸𝑓, and the cross-sectional area
of the fibers embedded within the composite, 𝐴𝑓, as given in Equation (8.1). Table 8.1 lists the
values of 𝐴𝑓 and 𝐾𝑓 for all of the strengthened beams.
𝐾𝑓= 𝐴𝑓𝐸𝑓 Eq. (8.1)
8.3.4 Strengthening Procedure and Configuration
The strengthening procedure is shown in Figure 8.1. At the end of the corrosion process, the
deteriorated concrete was first removed using a hydraulic hammer (Figure 8.3a). The corroded
steel bars were then brushed and a cementitious repair mortar (Sikacrete-08SCC) was used to fill
the damaged zone. The repair mortar had a compressive strength of 55.4 MPa (standard deviation
of 5 MPa) and a flexural strength of 3.4 MPa (standard deviation of 0.3 MPa). After curing, the
beam surface was sandblasted before applying the externally-bonded composite system (Figure
8.3b). The P-FRCM was installed using the hand lay-up method. The first layer of the cementitious
matrix was applied to the concrete substrate with a thickness of 4 to 5 mm (Figure 8.3c). Then, the
fabric was installed and coated with a second layer of matrix of similar thickness. The procedure
was then repeated until the specified number of layers was attained according to the test matrix in
Table 8.1. Similar procedure was followed while applying the C-FRP composite system (Figure
8.3d).
Each layer of the designated composite system was 150 mm wide (width equal to that of the
beam) and was applied to the soffit of the beam over a length of 2400 mm. The fibers were oriented
so that their primary directions were parallel to the longitudinal axis of the beam. The flexural plies
were anchored at each end using one U-shaped transverse strip of 300 mm width and 200 mm
height as shown in Figure 8.1.
162
Figure 8.3: Strengthening procedure of corroded beams: a) removing the deteriorated concrete,
b) patch repairing and sandblasting, c) installation of P-FRCM composite, and d) installation of
C-FRP sheets
8.3.5 Test Setup and Instrumentation
The beams were tested under four-point loading configuration as shown in Figure 8.1. The load
was applied in displacement control at a rate of 2 mm per minute using a MTS actuator. Deflections
were measured by means of three linear variable differential transducers (LVDTs) located at mid-
span and under the point loads. All beams were instrumented at mid-span with a 60 mm long
concrete strain gauge bonded to the top surface of the beams and 5 mm steel strain gauges bonded
to the tensile reinforcing bars. The strengthened specimens were instrumented with 5 mm strain
gauges installed directly on the outer fabric of the composite system at mid-span.
8.4 Numerical Simulation
Three-dimensional finite element (FE) models have been developed to simulate the nonlinear
flexural behavior of the tested beams using the software package ATENA 3D [122]. Due to
symmetry of loading and support conditions, only half of the beam was modeled and the
appropriate symmetry conditions were applied in order to reduce the computational time. The C-
163
FRP composite system was modeled as a discrete reinforcement bonded directly to the beam soffit
while the P-FRCM composite was modeled using a more detailed approach that involved modeling
the fabric and the matrix layers. The mass losses in steel bars due to corrosion were represented
by reductions in the cross-section area of the bars, As, according to the actual mass losses given in
Table 8.1. The mechanical properties of the materials reported earlier were used as input data in
the 3D model. The materials constitutive laws, elements type, and boundary conditions adopted in
the models are presented in the following sections.
8.4.1 Constitutive Laws
8.4.1.1 Concrete and Cementitious Matrix
The 3D nonlinear cementitious material model of the FE package (CC3DNonLinCementitious2)
is used to simulate concrete and the cementitious matrix associated with the P-FRCM system. The
model consists of two parts representing the compressive (plastic) and the tensile (fracturing)
behaviors. The compressive model is based on the Menétrey-Willam failure surface [123] while
the tensile model is based on the classical orthotropic smeared crack formulation and crack band
model. The compressive model consists of an ascending branch that represents the hardening phase
as shown in Figure 8.4a and a descending branch that represents the softening phase shown in
Figure 8.4b. In the hardening phase, the material behaves linearly up to a compressive stress value
of 𝑓𝑐𝑜= 2𝑓𝑡, where fco is the compressive stress at the onset of the nonlinear compressive behavior
and ft is the tensile strength of the material. The nonlinear behavior is governed by Equation (8.2)
as follows:
𝜎𝑐 = 𝑓𝑐𝑜 + ( 𝑓𝑐′ - 𝑓𝑐𝑜 ) √1 − {
ɛ𝑐𝑝− ɛ𝑝
ɛ𝑐𝑝}
2
Eq. (8.2)
where 𝜎𝑐 = the compressive stress in the nonlinear hardening part, 𝑓𝑐′ = cylinder compressive
strength, ɛ𝑝 = plastic strain, and ɛ𝑐𝑝 = plastic strain corresponding to the compressive strength.
164
Figure 8.4: Constitutive laws of concrete and cementitious matrix: a) compressive hardening law,
b) compressive softening law, and c) tensile softening law
In the softening phase, the material is assumed to behave linearly as shown in Figure 8.4b. The
plastic strain, ɛ𝑝, is transformed into displacement, 𝑤𝑐, through the length scale parameter, 𝐿𝑐,
which corresponds to the projection of the element size in the direction of minimal principal stress
as shown in Figure 8.4b. The plastic displacement, 𝑤𝑑, shown in Figure 8.4b is used to define the
end of the softening phase in compression and is assumed equal to 0.5 mm [124].
The compressive model also includes a reduction in the compressive strength after cracking in
the direction parallel to the cracks in a way similar to that proposed by Vecchio and Collins [125].
The reduced strength after cracking is estimated using Equation (8.3) and Equation (8.4) as
follows:
𝑓𝑐𝑒𝑓𝑓
= 𝑟𝑐 𝑓𝑐′ Eq. (8.3)
𝑟𝑐 = 1
0.8 +170ɛ1 , 𝑟𝑐
𝑙𝑖𝑚 ≤ 𝑟𝑐 ≤ 1.0 Eq. (8.4)
where 𝑓𝑐𝑒𝑓𝑓
= the effective compressive strength in a direction parallel to the direction of cracks,
ɛ1= the strain in a direction normal to the direction of the crack, 𝑟𝑐 is a reduction factor due to
cracking, and 𝑟𝑐𝑙𝑖𝑚= maximal compressive strength reduction factor taken as 0.8.
The tension model is assumed linear elastic up to the material tensile strength, 𝑓𝑡, followed by an
exponential softening shown in Figure 8.4c. The slope of the linear ascending branch is taken equal
to the material elastic modulus. In the softening branch (Figure 8.4c), the fixed crack model
proposed in [126] is adopted. In this model, the crack direction is given by the principal stress
direction at the moment of crack initiation. During further loading, this direction is assumed fixed
165
and represented the material axis of orthotropy. The exponential function of the crack opening is
given in Equation (8.5) and Equation (8.6) as derived experimentally by Hordijk [127]:
𝜎𝑡
𝑓𝑡 = {1 + (𝑐1
𝑤𝑡
𝑤𝑡𝑐)
3
} 𝑒𝑥𝑝 (−𝑐2𝑤𝑡
𝑤𝑡𝑐) −
𝑤𝑡
𝑤𝑡𝑐(1 + 𝑐1
3) 𝑒𝑥𝑝(−𝑐2) Eq. (8.5)
𝑤𝑡𝑐 = 5.14 𝐺𝑓
𝑓𝑡 Eq. (8.6)
where 𝑤𝑡 = crack opening displacement, 𝑤𝑡𝑐= crack opening at the complete release of stress, 𝜎𝑡
= the normal stress in the crack (crack cohesion), 𝐺𝑓 = fracture energy of the material needed to
create a unit area of stress free crack, and the constants 𝑐1and 𝑐2 taken equal to 3 and 6.93,
respectively [127]. The crack opening, 𝑤𝑡, is derived from the material strain, ɛ𝑐𝑓, and the crack
band length, 𝐿𝑡, is assumed equal to the size of the element projected into the crack direction.
The shear strength of the cementitious material is calculated using Equation (8.6), which is based
on the modified compression field theory developed by Vecchio and Collins [125] as follows:
𝜏𝑒𝑓𝑓 = 0.18√𝑓𝑐
′
0.31+24𝑤
𝑎𝑔 + 16
Eq. (8.7)
where 𝜏𝑒𝑓𝑓 = effective shear strength of a cracked cementitious material, w = maximum crack
width at a given location, and 𝑎𝑔 = maximum aggregate size. The values of the parameters used in
the constitutive model of the cementitious materials are presented in Table 8.2.
Table 8.2: Characteristics of concrete and P-FRCM matrix used in FE models
Parameter Concrete P-FRCM
Cementitious
matrix C41.8 C30 C20
Cylinder compressive strength, 𝑓𝑐′, (MPa) 41.8 30 20 43.9
Cube compressive strength, 𝑓𝑐𝑢, (MPa) 49.2 35.3 23.5 51.6
Tensile strength, 𝑓𝑡, (MPa) 3.2 2.6 1.9 3.2
Elastic modulus, E, (GPa) 36.7 32.4 27.3 37.35
Poisson’s ratio, ʋ 0.2 0.2 0.2 0.2
Specific fracture energy, 𝐺𝑓, (MN/m) 8.05×10-5 6.45×10-5 4.93×10-5 8.32×10-5
Critical compressive displacement, 𝑤𝑑, (m) -5 ×10-4 -5×10-4 -5×10-4 -5×10-4
Minimum compressive strength reduction factor, 𝑟𝑐𝑙𝑖𝑚 0.80 0.80 0.80 0.80
166
8.4.1.2 Steel and Fabrics
The stress-strain law of the steel reinforcing bars is assumed elastic perfectly plastic (i.e. bilinear)
Prefect bond is assumed between the steel bars and concrete and between the steel bars and the
repair material in all of the strengthened beams. The carbon and the PBO fabrics are assumed to
behave linearly elastic up to failure.
8.4.1.3 Interfacial Bond Stress-slip Models
Interfacial bond stress-slip models have been adopted at the interfaces of C-FRP/concrete and
PBO-fabric/matrix to simulate the experimental observations in which debonding of the
externally-bonded system occurred at the C-FRP/concrete interface in beam CS-A-1C while it took
place at the fabric/matrix interface in the beams strengthened with P-FRCM.
8.4.1.4 C-FRP/Concrete Interfacial Bond Stress-slip Model
The bilinear bond stress-slip model proposed by Lu et al. [117] has been adopted to model the
interfacial behavior between the C-FRP and concrete. The model is governed by Equation (8.8) to
Equation (8.12) and is shown in Figure 8.5.
𝜏𝑚𝑎𝑥 = 1.5𝛽𝑤𝑓𝑡 Eq. (8.8)
𝑆0 = 0.0195𝛽𝑤𝑓𝑡 Eq. (8.9)
𝑆𝑓 = 2𝐺𝑓𝑡 𝜏𝑚𝑎𝑥 ⁄ Eq. (8.10)
𝐺𝑓𝑡 = 0.308𝛽𝑤2 √𝑓𝑡 Eq. (8.11)
𝛽𝑤 = √2.25 – 𝑏𝑓 𝑏𝑐⁄
1.25 + 𝑏𝑓 𝑏𝑐⁄ Eq. (8.12)
where 𝜏𝑚𝑎𝑥 = maximum bond stress, 𝑆0 = slip at maximum bond stress, 𝑆𝑓 = slip at failure, 𝐺𝑓𝑡
= interfacial fracture energy of concrete, 𝛽𝑤 = width coefficient factor, 𝑏𝑓 = width of the C-FRP
sheet, 𝑏𝑐 = width of the concrete substrate, and 𝑓𝑡 = concrete tensile strength.
167
Figure 8.5: C-FRP/concrete interfacial bond stress-slip model according to Lu et al. [117]
adopted for various concrete mixes
8.4.1.5 PBO-fabric/matrix Interfacial Bond Stress-slip Model
The bond between the PBO-fabric and the associated cementitious matrix is described by
Equation (8.13) and is shown in Figure 8.6, as proposed by D’Ambrisi et al. [120]. Equation
(8.13) is a modification of the bond-slip relation proposed in [128] for FRP sheets, where 𝜏 (s) =
bond strength, s = the corresponding slip, 𝜏𝑖 = initial bond strength, 𝑆𝑓 = slip at failure. A, 𝛼, and
𝛽 are curve-fitting parameters and their values were evaluated as 0.92, 8.94, and 43.85,
respectively, for P-FRCM [120]. In their study, D’Ambrisi et al. [120] also proposed the values of
𝜏𝑖 = 0.15 MPa and 𝑆𝑓 = 1.18 mm for P-FRCM.
𝜏 (s) = [ 𝜏𝑖+ A ( 𝑒−𝛼𝑆 - 𝑒−𝛽𝑆 )] . (1 − 𝑆
𝑆𝑓) 0≤ 𝑆 ≤ 𝑆𝑓 Eq. (8.13)
0
0.5
1
1.5
2
2.5
3
3.5
0 0.05 0.1 0.15 0.2 0.25
𝜏(s)
(M
Pa)
S (mm)
C 41.8
C 30
C 20
168
Figure 8.6: PBO-fabric /matrix interfacial bond stress-slip model according to D’Ambrisi et al.
[120]
8.4.1.6 Element Type
To model the concrete and the cementitious matrix, 3D solid brick elements with 8 nodes were
used. The size of the concrete and matrix elements were 30 and 10 mm, respectively, based on a
mesh sensitivity study conducted by the authors. Based on test observations, perfect bond was
assumed between the concrete and the cementitious matrix. The steel, carbon, and PBO were
modeled using 3D truss elements. The typical mesh discretization of the beams along with the
details of the steel and the externally bonded reinforcements are shown in Figure 8.7.
8.4.1.7 Loading, Boundary Conditions, and Monitoring Points
Loading procedure and support steel plates similar to the ones actually used in the tests were
adopted in the FE model. The beams in FE models were loaded by means of prescribed
displacements located at the midpoint of the top surface of the steel loading plate, at an increment
of 0.1 mm. The support plate was restrained from movement in the vertical and transverse direction
(z and y directions, respectively) by means of a line support placed at the middle line of the bottom
surface of the plate. Since only half of the beam was modeled, the vertical surface along the axis
of symmetry was restrained from the horizontal movement (i.e., in the x direction).
The load was monitored at the midpoint of the top surface of the load plate where the prescribed
displacements were applied. Another point at the beam mid-span on the bottom surface of the
concrete or the composite system was used to record the beam deflections. The strains in the
169
bottom steel bars and the outer fiber layer of the composite system were monitored by means of
monitoring points at the beam mid-span. Strain monitoring point was also used to measure the
concrete compressive strains at mid-span.
Figure 8.7: a) Meshing of P-FRCM strengthened beams, b) reinforcement layout for beams
strengthened with 4 layers of P-FRCM, and c) reinforcement layout for beams strengthened with
C-FRP sheet
(a) 1
2 Front view 3
4 Bottom View 5
(b) 6
Front view 7
Bottom View 8
(c) 9
Steel reinforcement
Loading plate
Support plate
Concrete mesh
Externally-bonded P-FRCM
P-FRCM-end anchor
170
8.5 Results and Discussion
In this section, the experimental and numerical results are presented and discussed. The finite
element models were verified by comparing the results of the numerical analysis with the
experimental results. More details about the experimental results can be found in [121].
8.5.1 Crack Pattern at Failure
The predicted cracking pattern at failure was compared to that obtained from the tests in Figures
8.8a, 8.8b, and 8.8c for beams UUa, CS-A-1C, and CS-A-4P, respectively. Both the experimental
and numerical investigations indicated that the control beams (UU) and the corroded
unstrengthened beams (CU-A and CU-B) failed due to yielding of the steel bars followed by
concrete crushing. This mode of failure can be depicted in Figure 8.8a for the control beam UUa.
This finding indicated that the corrosion damage of the bottom steel bars with mass loss up to 18%
(group B) did not change the ductile mode of failure of the RC beams [121]. Beam CS-A-1C failed
due to a longitudinal crack that developed within the concrete layer close to the C-FRP
laminate/concrete interface, which was followed by a sudden rupture of the C-FRP laminate. A
thin layer of concrete was attached to the laminate at failure as shown in Figure 8.8b. The same
mode of failure was captured also by the FE model. A good correlation between the predicted and
experimental crack patterns was achieved as shown in Figure 8.8b.
On the other hand, all beams strengthened with two or four layers of P-FRCM in groups A and
B failed due to FRCM delamination at the fabric/matrix interface adjacent to the concrete substrate
[121] as shown in Figure 8.8c. This was similar to the mode of failure obtained from the numerical
simulation of beam CS-A-4P, which occurred due to the slippage of fabric within the cementitious
matrix. It is important to note that this mode of failure was not similar to that observed in beam
CS-A-1P, which was strengthened with one layer of P-FRCM composite, in which no bond
degradation between the P-FRCM composite and the concrete substrate was observed during the
test. Beam CS-A-1P failed by steel yielding followed by concrete crushing prior to the
delamination of the P-FRCM composite.
For all of the strengthened with two and four layers of P-FRCM, good correlation between the
predicted and experimental crack patterns as shown in Figure 8.8c. These finding indicate that the
inclusion of the interfacial bond stress-slip models between the C-FRP and concrete and between
171
the PBO-fabric and the matrix enabled the FE models to detect the appropriate failure modes of
the strengthened beams.
a)
b)
172
c)
Figure 8.8: Numerical and experimental crack patterns at failure for a) beam UUa, b) beam CS-
A-1C, and c) beam CS-A-4P
8.5.2 Load-deflection Response
The experimental load-deflection responses are compared with those obtained numerically in
Figure 8.9 and Figure 8.10 for beams of groups A and B, respectively. Both experimental and
numerical load-deflection responses consisted of three stages with two turning points indicating
concrete cracking and yielding of steel bars. As can be seen in Figure 8.9 and Figure 8.10, there is
a good agreement between the numerical and experimental load-deflection plots, which verifies
the accuracy of the FE models in capturing the nonlinear responses of the strengthened beams. The
corrosion damage of the bottom steel reinforcement, which was represented in the numerical
analysis by a reduction in the cross-section area of the steel bars, was offset after strengthening
and most of the beams restored their initial capacities as will be detailed later. It is important to
note that the use of externally-bonded composites (i.e. C-FRP or P-FRCM composites) had no
notable impact on the stiffness of the strengthened beam prior to steel yielding. However, all of
the strengthened beams exhibited various degrees of enhancement in their post-yielding stiffness,
which was mainly dependent on the type and the amount of fiber used as has been confirmed
numerically.
The main experimental and numerical results of the load-deflection responses of the tested beams
are summarized in Table 8.3 and Table 8.4. For the unstrengthened beams, the results given in
Table 8.3 indicated that the yield and ultimate loads measured experimentally for beams CU-A
173
and CU-B were 7 and 5% and 14 and 7% lower than the corresponding values of the control
uncorroded beams (UU). The numerical results of the same beams were in agreement with those
measured experimentally. The ratio of the yield loads predicted numerically to those measured
experimentally (𝑃𝑦𝐹𝐸 𝑃𝑦
𝐸𝑋𝑃⁄ ) was 0.99 and 1.02 for beams CU-A and CU-B, respectively. The
corresponding ratios of the ultimate loads (𝑃𝑢𝐹𝐸 𝑃𝑢
𝐸𝑋𝑃⁄ ) for the same beams were 0.94 and 0.93,
respectively.
a) Beam UU, CU-A, and CS-A-1C
a) Beam UU, CU-A, and CS-A-1C
Figure 8.9: Numerical and experimental load-deflection responses for beams of group A
174
Table 8.3: Predicted and experimental strength and fiber strain results
Beam Yield load (kN) Ultimate load (kN) Ultimate fiber strain (‰) Reference
𝑃𝑦𝐸𝑋𝑃 𝑃𝑦
𝐹𝐸 𝑃𝑦𝐹𝐸 𝑃𝑦
𝐸𝑋𝑃⁄ 𝑃𝑢𝐸𝑋𝑃 𝑃𝑢
𝐹𝐸 𝑃𝑢𝐹𝐸 𝑃𝑢
𝐸𝑋𝑃⁄ ɛ𝑓𝑢𝐸𝑋𝑃 ɛ𝑓𝑢
𝐹𝐸
Control beams: Uncorroded unstrengthened beams
UUa, UUb* 75.1 78.22 1.04 79.7 81.7 1.03 - - [121]
Group (A): Theoretical steel mass loss of 10%
CU-A 69.5 68.7 0.99
76.1 71.8 0.94 - - [121]
CS-A-1C 77.9 77.6 1 96.5 103.4 1.07 13.8 14.48 -
CS-A-1P 71.1 74.3 1.04 82.9 79.5 0.96 14.92 10.96 -
CS-A-2P 79.5 75.3 0.95 86.4 89.6 1.04 8.49 12.47 [121]
CS-A-4P 83.3 82.3 0.99 99.6 106 1.06 9.44 13.11 [121]
Group (B): Theoretical steel mass loss of 20%
CU-B 64.5 65.7 1.02
74.2 68.7 0.93 - - [121]
CS-B-2P 71.8 71 0.99 85.6 85.1 0.99 8.18 13.28 [121]
CS-B-4P 79.6 78.6 0.99 102.6 103.8 1.01 10.66 13.32 [121]
* Average values are reported
175
Table 8.4: Predicted and experimental deflections and ductility indices
Beam Deflection at yield load Deflection at ultimate load
Ductility index
Reference 𝛿𝑦
𝐸𝑋𝑃 (mm) 𝛿𝑦𝐹𝐸 (mm) 𝛿𝑦
𝐹𝐸 𝛿𝑦𝐸𝑋𝑃⁄ 𝛿𝑢
𝐸𝑋𝑃 (mm) 𝛿𝑢𝐹𝐸 (mm) 𝛿𝑢
𝐹𝐸 𝛿𝑢𝐸𝑋𝑃⁄ 𝜇 𝐸𝑋𝑃 𝜇 𝐹𝐸 𝜇 𝐹𝐸/𝜇 𝐸𝑋𝑃
Control beams: Uncorroded unstrengthened beams
UUa, UUb* 11.7 10.7 0.91 32.9 31.3 0.95 2.81 2.93 1.04 [121]
Group (A): Theoretical steel mass loss of 10%
CU-A 10.9 10.5 0.96
35.4 29.5 0.83
3.25 2.81 0.86 [121]
CS-A-1C 12.3 11 0.89 30.4 38.7 1.27 2.47 3.52 1.42 -
CS-A-1P 11.7 10.7 0.91 32.9 26.1 0.79 2.81 2.44 0.87 -
CS-A-2P 11.8 10.5 0.89 33 33 1 2.8 3.14 1.12 [121]
CS-A-4P 12.8 11.2 0.88 31.5 33.7 1.07 2.46 3.01 1.22 [121]
Group (B): Theoretical steel mass loss of 20%
CU-B 9 10.1 1.12
28.4 34.9 1.23
3.16 3.46 1.1 [121]
CS-B-2P 11.3 10.6 0.94 33.7 31.1 0.92 2.98 2.93 0.98 [121]
CS-B-4P 12.5 11.51. 0.92 37.4 35.6 0.95 2.99 3.1 1 [121]
* Average values are reported
176
Figure 8.10: Numerical and experimental load-deflection responses for beams of group B
For the strengthened beams, both the experimental and numerical results listed in Table 8.3
revealed that flexural strengthening of the corrosion-damaged RC beams with C-FRP or P-FRCM
enhanced their load capacities. As can be seen in Table 8.3, the ratio of the predicted to measured
yield loads of the tested beams (𝑃𝑦𝐹𝐸 𝑃𝑦
𝐸𝑋𝑃⁄ ) ranged between 0.95 and 1.04 with an average ratio =
1.0, a standard deviation of 2.8%, and a coefficient of variation of 2.8%. The ratio of the predicted
to measured ultimate loads (𝑃𝑢𝐹𝐸 𝑃𝑢
𝐸𝑋𝑃⁄ ) of all beams ranged between 0.93 and 1.07 with an average
ratio = 1, a standard deviation of 5%, and a coefficient of variation of 5%.
Table 8.4 compares the deflections of the tested beams at both yield and ultimate loads. The
deflections predicted numerically were in a good agreement with those measured experimentally.
The predicted deflections at ultimate loads for all beams strengthened with P-FRCM were within
7% error band except in the case of beam CS-A-1P for which the numerical model underestimated
the deflection at ultimate by 21%. This was attributed to the discrepancy in the experimental and
predicted modes of failure observed for the beam CS-A-1P. While test observations indicated
prefect bond between the P-FRCM composite and the concrete substrate up to failure, the FE
model predicted a premature failure due to the fabric slippage within the cementitious matrix. This
finding can be depicted in Figure 8.9b from the sudden drop in the predicted load-deflection curve.
On the contrary, the FE model overestimated the defection at ultimate load for the beam CS-A-1C
strengthened with C-FRP laminate by approximately 27% as can be observed in Figure 8.9a. This
177
was related to the actual premature rupture of the C-FRP laminate at ultimate, which was
confirmed by the measured fiber strain at ultimate as will be explained later.
8.5.3 Load-carrying Capacity
The decline in the ultimate load, 𝑃𝑢, in the corroded unstrengthened beams due to corrosion of
the steel bars and the capacity gain experienced by the strengthened beams are shown in Figure
8.11. As previously noted, the measured ultimate loads for the corroded unstrengthened beams
CU-A and CU-B were 5% and 7% lower than that of the control virgin beam (UU), respectively.
The corresponding values predicted numerically were 12.1 and 15.9%, respectively, which means
that the numerical model tended to overemphasize the effect of corrosion damage on the load-
carrying capacities of the corroded beams. This could be explained by the fact that the FE models
didn’t accurately represent the mass loss that physically occurred during the corrosion process.
While the loss of the bars’ lugs due to corrosion contributed significantly to the mass loss measured
in the lab, it didn’t affect significantly the cross-section of the steel bars and hence didn’t affect
the load-carrying capacities of the corroded beams. However, the mass loss adopted in the FE
models significantly reduced the cross-section area of the bars and therefore contributed directly
to the decline in the load-carrying capacities of the beams.
Figure 8.11: Gain and decline in % in the ultimate loads, 𝑃𝑢, with respect to that of the control
beam (UU)
178
For the strengthened beams, both the experimental and numerical results indicated the
effectiveness of the externally-bonded composite systems (C-FRP and P-FRCM) in restoring the
yield and ultimate loads of the corroded beams as given in Table 8.3. The gain in strengths was
highly dependent on the amount and type of fibers used rather than the level of corrosion damage
in the tensile steel bars [121]. Beam CS-A-1P that was strengthened with one layer of P-FRCM
exhibited load-carrying capacity of 82.9 kN, which was 104% of that of the control beam. The
failure load predicted by the numerical model was 79.5 kN, representing a decline of -4%
compared to the experimental value. Increasing the number of P-FRCM layers further increased
the load-carrying capacities of the strengthened beams. For instance, beams CS-A-2P and CS-A-
4P failed at 86.4 and 99.6 kN, respectively, due to the delamination of the P-FRCM composite.
The predicted ultimate loads for the same beams were 89.6 and 106 kN, respectively, representing
differences of +4 and +6% compared to the experimental values. Both the experimental and
numerical results revealed that the gain in the load-carrying capacities of the strengthened beams
was approximately proportional to the number of P-FRCM layers used (i.e., the cross-section area
of the composite, 𝐴𝑓). The predicted gain in the ultimate strengths for the beams strengthened with
one, two, and four layers of P-FRCM were 7.7 kN (for CS-A-1P), 17.8 kN (for CS-A-2P), and
34.2 (for CS-A-4P), respectively. This finding was also confirmed by the similar ultimate fiber
strains reported for these beams.
Beams CS-B-2P and CS-B-4P encountered load-carrying capacities of 85.6 and 102.6 kN,
respectively, compared to 85.1 and 103.8 kN predicted numerically, representing differences of -
0.6 and +1.1% compared to the experimental values. The measured and predicted ultimate loads
for the beams of group B were approximately similar to their counterparts in group A (CS-A-2P
and CS-A-4P) that experienced lower level of corrosion. This finding indicated that approximately
doubling the level of corrosion had an insignificant impact on the load-carrying capacities of the
strengthened beams. More details about the effect of corrosion level on the performance of the
FRCM-strengthened beams are reported in Elghazy et al. [121].
The beam strengthened with C-FRP laminate failed at 96.5 kN, which was 21.1% higher than
that of the control beam. The finite element model predicted a load-carrying capacity of 103.4 kN,
representing a difference of +7% compared to the experimental value. This slight discrepancy
between the numerical and experimental capacities is generally acceptable. It is important to note
179
that the equivalent axial stiffness, 𝐾𝑓, of one layer of C-FRP, which strengthened beam CS-A-1C,
was about 80% of that of four layers of P-FRCM that’s strengthened beam CS-A-4P. It was
therefore expected to experience a higher load-carrying capacity in the latter beam than its former
counterpart. However, experimental and numerical data confirmed that beam CS-A-1C (with the
lower axial stiffness) showed an ultimate capacity of 97.5% of that of beam CS-A-4P (with the
higher axial stiffness). This finding indicated that the strengthening effectiveness of the P-FRCM
system was slightly lower than that of the C-FRP system.
8.5.4 Strain Response
The experimental and predicted relationships between the applied loads and the strains measured
in the outer fiber and concrete are shown in Figure 8.12 for all of the tested beams. Tensile steel
strain responses for beams CS-A-1C and CS-A-4P are also shown in Figure 8.13. Representative
numerical strain profiles of the internal and external reinforcement at ultimate are shown in Figure
14 for beams CS-A-1C and CS-A-4P.
Similar to the load-deflection responses, the numerical and experimental load-strain curves
consisted of three segments with two turning points that indicated the concrete cracking and the
yielding of the tensile steel bars. It can be observed that the numerical models reasonably predicted
the strain responses in the fiber, the concrete, and the tensile steel bars, which further validated the
accuracy of the FE models.
Table 3 lists the strain values in the outer fiber of the composite systems at ultimate as determined
experimentally, ɛ𝑓𝑢𝐸𝑋𝑃, and predicted numerically, ɛ𝑓𝑢
𝐹𝐸. Both the experimental and numerical
results indicated that beam CS-A-1C failed due to fiber rupture, with strain values ɛ𝑓𝑢𝐸𝑋𝑃 and ɛ𝑓𝑢
𝐹𝐸
reaching 13.8 and 14.48‰, respectively. This finding suggested that the measured strains in the
C-FRP laminate at failure was lower than that reported in the manufacturer’s data sheet and utilized
in the FE model. This finding explained why the FE model slightly overestimated the ultimate load
of beam CS-A-1C by 7%. The predicted ultimate strain profile of the internal and external
reinforcement shown in Figure 14a for beam CS-A-1C indicated that the maximum C-FRP strain
took place within the maximum moment zone and decreased gradually outside this zone, which
confirmed the rupture of fibers approximately at mid-span as observed in the tests.
180
Figure 8.12: Fiber and concrete strain response
Experimental results
Experimental results
Numerical results
Numerical results
a) Group A
a) Group B
181
Figure 8.13: Tensile steel strain response for beams CS-A-1C and CS-A-4P
a) Beam CS-A-1C
b) Beam CS-A-4P
Figure 8.14: Strain profile of the internal and external reinforcement at ultimate
182
Beam CS-A-1P experimentally recorded the maximum tensile strains in the outer fabric of the
P-FRCM layer (14.92‰) of all the beams. This was attributed to the mode of failure of beam CS-
A-1P, in which no fabric delamination was observed until failure occurred. On the other hand, the
predicted fiber strain at failure for beam CS-A-1P was 10.96‰ before failure occurred numerically
due to fiber slippage within the matrix. This discrepancy between the experimental and numerical
results can be attributed to the accuracy of the adopted interfacial bond-slip model between the
PBO-strands and the matrix in the FE model.
The measured fiber strains, ɛ𝑓𝑢𝐸𝑋𝑃, for beams strengthened with two and four layers of P-FRCM
ranged between 8.18 and 10.66‰ while the predicted fiber strains, ɛ𝑓𝑢𝐹𝐸, ranged between 12.47 and
13.32‰. Both the experimental and numerical results of such beams showed that the strains in the
PBO-fiber at peak loads were lower than the PBO-fiber rupture strain, which confirmed the modes
of failure of these beams due to the fiber slippage within the matrix. This finding can also be
depicted from the strain profile of the internal and external reinforcements predicted at ultimate
load as shown in Figure 8.14b for beam CS-A-4P. The slippage of the PBO-strands occurred within
the maximum moment zone (Figure 8.14b), which was consistent with the test observations.
8.5.5 Ductility
The deflection-based ductility index, μ, has been employed in the present study to evaluate the
effect of the externally-bonded composites on the ductility of the strengthened beams. The ductility
index is the ratio of the mid-span deflection at ultimate, 𝛿𝑢, to the mid-span deflection at yielding,
𝛿𝑦. A higher ductility index indicates a higher ability for the beam to provide sufficient
deformation, and hence ample warning, prior to failure. The experimental and numerical ductility
indices of the tested beams are given in Table 8.4. The predicted ductility indices were in
reasonable agreement with the experimentally obtained ones with a tendency of the FE model to
overestimate the ductility indices of the strengthened beams.
The ductility indices obtained experimentally ranged between 0.87 and 1.07 times that of the
control beam, which was adequate to guarantee satisfactory ductility. All beams strengthened with
P-FRCM showed ductility indices approximately similar to that of the control beam except beam
CS-A-4P that showed a ductility index 13% less than that of the control beam. Beam CS-A-1C
183
exhibited a ductility index equal to that of the beam CS-A-4P. These results indicated that the
externally-bonded composites slightly reduced the ductility of the strengthened beams.
Numerically, the ductility indices ranged between 0.83 and 1.2 times that of the predicted index
of the control beam. As previously noted, the FE model tended to overestimate the indices of the
strengthened beams except in case of beam CS-A-1P, in which the predicted index was 13% less
than the experimental index. This was explained by the different mode of failure predicted by the
FE model.
8.6 Parametric Studies
The verified FE models were used to investigate the effect of varying the compressive strength
of concrete substrate and the thickness of concrete cover on the flexural behavior of FRCM- and
FRP-strengthened beams. In each case, three FE models corresponding to three beams of the
experimentally-tested beams namely, beams CS-A-1C, CS-A-2P, and CS-A-4P, were selected as
baseline models for comparison. Then, the parameter of interest was varied to study its influence
on the flexural behavior of each beam. The results of the parametric study are summarised in Table
8.5 and Table 8.6 and discussed in the following sections.
8.6.1 Effect of Concrete Compressive Strength (𝒇𝒄′ )
This investigation aims to assess the potential of using the externally-bonded strengthening
technique to restore the service and ultimate capacities of RC beams having low compressive
concrete strengths. The tested beams had concrete compressive strength of 41.8 MPa (C41.8). Two
other concrete models C30 and C20 with compressive strengths 30 and 20 MPa, respectively, were
used in the numerical analysis to predict the flexural performance of beams CS-A-1C, CS-A-2P,
and CS-A-4P. The parameters used in the concrete models C41.8, C30, and C20 are given in Table
8.2.
The results of varying the concrete strength are summarized in Table 8.5 and the predicted load-
deflection responses are shown in Figure 8.15. The load-deflection response was similar in all
models with a noticeable lower stiffness response in beams with low concrete strengths. The use
of concrete with low compressive strengths decreased the load-carrying capacities of the
strengthened beams and decreased their mid-span deflections at ultimate. For example, the use of
184
C30 and C20 to model beam CS-A-4P reduced its ultimate load by 5.9 and 12.1% of that of the
baseline model (i.e. beam with concrete C41.8). Similar trend was observed in beams CS-A-1C
and CS-A-2P as can be depicted from Table 8.5.
Table 8.5: Summary of the parametric study results (effect of fc’)
FE Model Concrete 𝑃𝑢 (kN) Change in 𝑃𝑢
(%)** 𝛿𝑢 (mm) ɛ𝑓𝑢 (%) Mode of failure
CS-A-2P
C41.8* 89.6 0 32.97 12.47 Fiber slippage
C30 83.1 -7.2 28.63 9.535 Fiber slippage
C20 79.4 -11.3 26.41 11.06 Fiber slippage
CS-A-4P
C41.8* 106 0 33.69 13.11 Fiber slippage
C30 99.8 -5.9 31.37 12.02 Fiber slippage
C20 93.2 -12.1 30.68 11.57 Fiber slippage
CS-A-1C
C41.8* 103.4 0 38.69 14.48 laminate rupture
C30 97.7 -5.5 38.46 13.21 laminate debonding
C20 90 -12.9 35.03 11.59 laminate debonding
* Baseline FE model
** Relative to that of the baseline FE model
Slippage of the PBO fabric within the matrix governed the mode of failure in all of the FRCM-
strengthened beams regardless of their concrete compressive strengths. This suggested that the
interfacial bond stresses encountered at the fabric/matrix interface were higher than those
encountered at the concrete/matrix interface, which initiated slippage at the fabric level rather than
the separation between the matrix and the concrete substrate.
For beam CS-A-1C strengthened with CFRP laminates, it was remarkable to observe that the use
of low concrete compressive strength changed its mode of failure from laminate rupture (as
observed in the experimental test with concrete C41.8) to separation at the laminate/concrete
interface. This was attributed to the fact that the bond strength at the C-FRP/concrete interface was
significantly affected by the mechanical properties of the concrete substrate as shown in Figure
8.5. At failure, the recorded strains in the outer fabric in almost all of the models with lower
compressive strengths were lower than those recorded in the baseline models (Table 8.5).
185
Figure 8.15: Effect of the concrete compressive strength on the load-deflection response for
beams a) CS-A-1C, b) CS-A-2P, and c) CS-A-4P
186
8.6.2 Effect of Concrete Cover
Different thicknesses of the clear concrete cover were modeled to investigate their influence on
the strengthening effectiveness of FRCM and FRP systems. The clear concrete cover of the tested
beams was 25 mm. Two other concrete covers of 10 and 50 mm were modeled to predict the
flexural performance of beams CS-A-1C, CS-A-2P, and CS-A-4P.
The results of this study are summarized in Table 8.6 and the predicted load-deflection responses
are shown in Figure 8.16. All modeled beams showed similar load-deflection responses. However,
increasing the concrete cover from 25 to 50 mm significantly decreased the flexural stiffness of
the strengthened beams compared to that of the baseline model with concrete cover 25 mm (Figure
8.16). In addition, both yield and ultimate capacities of the strengthened beams were affected by
varying the concrete cover. This was attributed to the change in the effective depth of the steel
reinforcing bars as can be depicted from Figure 8.16. For example, the use of 50 mm concrete
cover in beams CS-A-4P and CS-A-1C decreased their ultimate capacities by 9.6 and 17.4 % of
those of the baseline models, respectively, whereas the use of a concrete cover of 10 mm increased
the ultimate capacities by 5.4 and 4.7% of those of the baseline models, respectively.
Table 8.6: Summary of the parametric study results (effect of concrete cover)
FE Model Concrete cover
(mm) 𝑃𝑢 (kN)
Change in 𝑃𝑢
(%)** 𝛿𝑢 (mm) ɛ𝑓𝑢 (‰) Mode of failure
CS-A-2P
25 89.6 0 32.97 12.47 Fiber slippage
10 90.2 +0.7 27.63 9.084 Fiber slippage
50 79.4 -11.3 34.51 12.85 Fiber slippage
CS-A-4P
25 106 0 33.69 13.11 Fiber slippage
10 111.8 +5.4 34.95 12.31 Fiber slippage
50 95.8 -9.6 33.55 12.09 Fiber slippage
CS-A-1C
25 103.4 0 38.69 14.43 laminate rupture
10 108.2 +4.7 37.23 14.47 laminate rupture
50 85.4 -17.4 27.43 10.08 laminate debonding
* Baseline FE model
** Relative to that of the baseline FE model
For the beams strengthened with P-FRCM system, varying the concrete cover did not affect their
mode of failure, which was characterised by the slippage of the PBO fabric within the matrix. This
finding was consistent with the measured fabric strains at failure shown in Table 8.6. On the other
hand, increasing the concrete cover in beam CS-A-1C (beam strengthened with C-FRP laminates)
187
to 50 mm changed its mode of failure from laminate rupture (in case of using 25 and 10 mm
concrete covers) to premature debonding at the laminate/concrete interface. This premature failure
significantly reduced the strengthening effectiveness of the FRP system. The results of this
parametric study indicated that the strengthening effectiveness of P-FRCM system was higher than
that of C-FRP system in the presence of thick concrete cover. This finding can be attributed to the
fact that failure of P-FRCM system is mainly dependent on the bond characteristics at the
fabric/matrix interface rather than the FRP/concrete bond strength in the case of FRP system.
b)
a)
188
Figure 8.16: Effect of the concrete cover on the load-deflection response for beams a) CS-A-1C,
b) CS-A-2P, and c) CS-A-4P
8.7 Conclusions
This paper discussed the numerical simulation of corrosion-damaged RC beams strengthened
with C-FRP and P-FRCM composites under flexural loading using ATENA. The load carrying
capacities, load-deflection responses, and load-strains responses were evaluated and compared
with the experimental results to validate the accuracy of the model. The validated models were
used in parametric studies to investigate the effect of varying the concrete compressive strength
and the thickness of concrete cover on the flexural behavior of the strengthened beams. The
following conclusions can be drawn from this study:
• The experimental and numerical studies indicated that the virgin and the corroded
unstrengthened beams failed due to yielding of steel bars. Corrosion up to 18% mass loss in
the steel bars showed no effect on the ductile mode of failure of the beams. However, the
numerical models tended to overestimate the effect of corrosion damage on the load-carrying
capacities of the corroded beams.
• Corrosion had an insignificant effect on the mode of failure of the C-FRP and P-FRCM-
strengthened beams. All beams strengthened with two and four layers of P-FRCM failed due
189
to FRCM delamination at the fabric/matrix interface adjacent to the concrete substrate
regardless of the corrosion level.
• Beam CS-A-1P strengthened with one layer of P-FRCM composite failed after steel yielding
prior to the delamination of the P-FRCM composite while beam CS-A-1C strengthened with
one layer of C-FRP laminate failed due to the sudden rupture of the laminate.
• Both the experimental and numerical results approved the ability of C-FRP and P-FRCM
composites to restore/increase the load-carrying capacities of the corrosion-damaged beams.
However, C-FRP composites were slightly more efficient than P-FRCM with similar axial
stiffness in restoring the beams strength.
• FE models accurately predicted the mode of failures of all of the strengthened beams except
that of beam CS-A-1P. While test observations indicated prefect bond between the P-FRCM
composite and the concrete substrate up to failure, the numerical model predicted a premature
mode of failure due to fabric slippage within the cementitious matrix.
• The ratios of the predicted to experimental yield loads of the tested beams (𝑃𝑦𝐹𝐸 𝑃𝑦
𝐸𝑋𝑃⁄ ) ranged
between 0.95 and 1.04 whereas the ratios of the predicted to measured ultimate loads
(𝑃𝑢𝐹𝐸 𝑃𝑢
𝐸𝑋𝑃⁄ ) ranged between 0.93 and 1.07. These results indicated the accuracy of the FE
models to predict the flexural performance of the tested beams.
• The predicted deflections at ultimate of all beams strengthened with P-FRCM were within 7%
error band except in the case of beam CS-A-1P in which the FE model underestimated its
deflection at ultimate by 21%. On the contrary, the FE model overestimated the defection at
ultimate for the C-FRP strengthened beam by 27%.
• The numerical results indicated the feasibility of using the bond slip models of D’Ambrisi et
al. [120] and Lu et al. [117] to simulate the behavior of P-FRCM and C-FRP-strengthened
beams at the composite/concrete interface.
• The results of the parametric study indicated that lowering the concrete compressive strength
or increasing the concrete cover decreased the load-carrying capacities of the strengthened
beams regardless of the strengthening system used (C-FRP or P-FRCM).
190
• The parametric study suggested that failure of FRCM-strengthened beams was independent
of the compressive strength of the concrete substrate or the thickness of the concrete cover.
Failure of such beams was governed by fabric slippage within the matrix. For beams
strengthened with C-FRP, lowering the compressive strength of the concrete substrate or
increasing the concrete cover changed their mode of failure from laminate rupture to fiber
slippage within the laminate.
Finally, the results of this study showed a promising potential of using FRCM composites in
restoring the load-carrying capacities and the serviceability of corrosion-damaged beams.
However, the experimental and numerical work presented in this paper is limited to the composites
used in the research reported in this paper and should not be extended to other types. The long-
term performance and durability of FRCM-strengthened beams are currently under investigation.
More research is needed to develop refined bond-slip models for FRCM composites, which
incorporate various parameters such as the number of fabric layers, the fabric type, and the fabric
orientation. Accurate correlation between the actual mass loss in the steel bars due to corrosion
and the reduction of reinforcement cross-section in FE models is also needed to better represent
the effect of corrosion level on the flexural performance of the strengthened members.
191
Chapter 9
Conclusions and Recommendations
8.8 Summary
Corrosion of steel reinforcement is the main cause of RC structure deterioration. Hence, durable
and costless repair technologies are vastly needed. FRCM composites have been introduced
recently as a new strengthening and repair technology for RC structures. The current study aimed
at investigating the monotonic and fatigue flexural behaviors of corrosion-damaged RC beams
strengthened with FRCM composites. Moreover, the long-term performance of the FRCM-
strengthened beams were investigated by exposing the strengthened beams to further corrosive
environment. The study includes both experimental and numerical investigations. In addition, the
analytical predictions provided by the current designs have been verified against the experimental
results.
The conducted work herein expands on the current knowledge of FRCM applications and
confirms the potential of FRCM composites to optimize the structural performance of corrosion-
damaged RC structures. The outcomes of this work are presented in five journal articles. A
summary of the findings and conclusions are presented in the following section.
8.9 Conclusions
8.9.1 Effect of Corrosion on RC Beams
• Corrosion of steel reinforcement initiated longitudinal cracks parallel to the corroded steel
bars. Wider corrosion cracks indicated higher steel mass loss due to corrosion. The
maximum widths of the observed corrosion cracks were 1.5, 2.8, and 3.5 mm for average
steel mass losses of 12.5, 19, and 22%, respectively. All of the corroded beams failed to
meet the provisions of ACI-318 design code for crack width criteria.
• Corrosion of steel reinforcing bars in the moment zone with an average mass loss up to
22.7% had a marginal impact on the flexural response of the tested beams. The maximum
decrease in the yield and ultimate strengths of the corrosion-damaged beams were 15 and
9%, respectively, of those of the virgin beams.
192
• Corrosion of steel reinforcement reduced the yield and ultimate strengths of the corrosion-
damaged beams as compared to those of the virgin (non-corroded) beams. The yield and
ultimate strengths of the damaged beams decreased at average rates of 0.66 and 0.40%,
respectively, per 1% of steel mass loss. This finding was attributed to the fact that the
corroded bars lost their lugs, which increased the measured mass loss without affecting
their effective cross-sectional areas.
• Corrosion of steel reinforcement severely affected the fatigue life of RC beams. Corrosion
damage of 19.8% mass loss reduced the fatigue life of the corroded beam by 80% of that
of the virgin (non-corroded) beam. Moreover, the corroded beam failed abruptly due to
the sudden rupture of the corroded steel bars.
8.9.2 Short-term Performance of Corroded Beams Strengthened with
FRCM
• The use of FRCM composites improved the flexural response of the corrosion-damaged
beams. FRCM significantly enhanced the post-yield stiffness of the corroded beams
compared to that of the corroded unstrengthened beams. The enhancement in the flexural
response was highly dependent on the FRCM strengthening scheme, its type, and the
number of FRCM layers used rather than the level of corrosion damage.
• The use of PBO-FRCM and C-FRCM increased the yield and ultimate strengths of the
corroded beams damaged at various levels of corrosion (up to 22.7% steel mass loss).
After 22.5% steel mass loss, the ultimate strengths of beams strengthened with PBO- and
C-FRCM were 54 and 51% higher than those of the corroded unstrengthened beam and
about 39 and 37% higher than those of the virgin (uncorroded unstrengthened) beam,
respectively. While, the yield strength of the beams strengthened with PBO-and C-FRCM
were 26 and 17 % higher than those of the corroded unstrengthened beam and about 0.8
and 0.1% higher than those of the virgin beam, respectively
• All of the corroded-strengthened beams failed due to the loss of the strengthening action
of the FRCM layers regardless of their corrosion damage level. The only exception was
for the beam strengthened with one layer of PBO-FRCM in which no distress of the
193
FRCM layer was observed until failure occurred due to concrete crushing. Four distinct
failure mechanisms were observed in the strengthened beams:
a) FRCM delamination: Delamination occurred between the fabric and the first layer of
the matrix adjacent to the concrete substrate. This mode of failure was observed in all
beams strengthened with PBO-FRCM in Scheme I (end-anchored layers).
b) Fabric slippage: This mode of failure was observed in all beams strengthened with
PBO-FRCM of Scheme II (U-shaped wrapping). Slippage occurred between the PBO
fabric and the cementitious matrix accompanied by slight delamination at the
fabric/matrix interface.
c) Matrix cracking and fabric separation from the matrix: This mode of failure was
reported in all beams strengthened with C-FRCM. It can be described by a progressive
cracking in the cementitious matrix associated with the sudden debonding between the
carbon fabric and the matrix. This mode of failure was more brittle than that observed
in the PBO-strengthened beams, which can be attributed to the superior characteristics
of the cementitious matrix of the PBO-FRCM compared to those of the C-FRCM
counterparts.
d) C-FRP laminate rupture: This mode of failure was reported for the beam strengthened
with C-FRP sheets. A longitudinal crack initiated at the laminate/concrete interface
followed by the sudden rupture of the laminate.
• Increasing the number of FRCM layers increased the yield and ultimate loads of the
strengthened beams. The increase in strength was approximately proportional to the
added number of layers. For instance, the use of one, two, and four layers of PBO-FRCM
increased the ultimate strengths of the strengthened beams by 9, 14, and 30% of that of
the corroded beam with12.5 % steel mass loss due to corrosion, respectively.
• Anchoring the FRCM flexural layers with a continuous U-shaped layer (Scheme II) was
more effective than the use of end anchors (Scheme I) in increasing the yield and ultimate
loads of the strengthened beams. This observation was attributed to the effect of the
continuous U-shaped layer in mitigating and delaying the delamination of the FRCM and
consequently increasing their strengthening potential. The use of two layers of PBO-
194
FRCM in Scheme I enhanced its yield and ultimate loads by 6 and 8%, respectively, while
the use of the same number of layers but in Scheme II improved its yield and ultimate
strengths by 14 and 28%, respectively.
• PBO-FRCM composites were more efficient than C-FRCM in improving the flexural
performance of the corroded beams. The beam strengthened with four layers of PBO-
FRCM in Scheme II showed an ultimate strength 15% higher than that of beam
strengthened with two layers of C-FRCM with the same Scheme and corrosion-damage
despite the fact that two layers of C-FRCM had similar axial stiffness of the four layers
of PBO-FRCM. Moreover, the PBO-FRCM strengthened beams showed more ductile
failure than their counterparts strengthened with C-FRCM. These observations were
related to the superior bond characteristics of the matrix of the PBO-FRCM system as
compared to those of C-FRCM.
• C-FRP and PBO-FRCM composites restored/increased the load-carrying capacities of the
corrosion-damaged beams. However, C-FRP composites were slightly more efficient
than the PBO-FRCM with similar axial stiffness in restoring the beam strengths.
• Although the equivalent stiffness of the FRCM composites, Kf, and the stiffness factor,
𝛽𝑓, are believed to be good indicators of the strengthening effectiveness of the FRCM
systems, they should not be solely used to compare the strength gain in beams without
considering the bond characteristics at the matrix/fabric interface and the anchoring
Scheme used.
• Beams strengthened with FRCM composites showed ductility indices and energy
absorption indices that ranged between 86 and 118% and between 111 and 153%,
respectively, of those of the virgin beams.
8.9.3 Long-term Performance of Corroded Beams Strengthened with
FRCM
• The use of FRCM systems reduced the corrosion rate in the steel bars with no evidence
on the effect of the FRCM strengthening scheme on such rate. Exposing the FRCM-
strengthened beams to post-strengthening corrosion resulted in 23% reduction in the steel
195
mass loss. This was attributed to the reduction in the amount of water and air diffused to
the tensile reinforcing bars due to the presence of the FRCM layers, which also explained
the reduction in the corrosion rate in the long-term beams.
• For beams exposed to the post-strengthening corrosive environment, longitudinal cracks
parallel to the corroded reinforcement on one or both lateral surfaces of the beams were
observed. However, the use of continuous U-shaped layers (Scheme II) reduced the width
and the length of the corrosion cracks compared to the use of end anchors (Scheme I).
• Exposing the beams strengthened in Scheme II to post-strengthening corrosive
environment had no impact on their mode of failure regardless of the type of fabric used.
However, corrosion cracks resulted in premature delamination of FRCM for beams
strengthened in Scheme I, which significantly affected the strengthening action of the
FRCM system. This can be attributed to the wrapping effect of the FRCM layer in Scheme
II that offset the effect of corrosion cracks in weakening the concrete cover and prevented
the premature delamination of FRCM flexural layers.
• Exposing the beams to corrosion after FRCM strengthening had a marginal impact on
their load-deflection response and strength. The long-term beams strengthened in Scheme
II demonstrated higher ductility and energy absorption indices than those of their short-
term counterparts. Short-term beam strengthened in Scheme I showed higher ductility
and energy absorption indices than those of their long-term counterparts due to the
premature FRCM delamination caused by corrosion cracks.
8.9.4 Validation of ACI-549.4R-13 Design Equations
• Strain values recorded on the FRCM layers indicated that the assumption of prefect bond
suggested by the ACI 549.4R-13 while limiting the maximum design strain in the FRCM
system to 0.012 mm/mm is a simplification that appears justifiable and easy to implement
by engineers.
• The theoretical formulations of ACI 549.4R-13 reasonably predicted the ultimate
capacities of the corrosion-damaged RC beams strengthened with FRCM composites in
Scheme I (end-anchored bottom layers) but underestimated the capacities of those
196
strengthened in Scheme II (with a continuously-wrapped layer). The ratios of the
experimental to the theoretical ultimate capacities (𝑃𝑢
𝑃𝑢𝑡ℎ⁄ ) was in the range of 1.03 to1.18
and 1.1 to 1.23 for the beams strengthened in Scheme I and Scheme II, respectively. This
finding was attributed to the fact that ACI 549.4R-13 provisions don’t consider the effect
of the U-shaped FRCM layer present in Scheme II in confining the bottom fabric layers
and in delaying their delamination from the matrix.
• The obtained results suggested increasing the ultimate capacities predicted by ACI
549.4R-13 for beams strengthened in Scheme II by 10% to consider the effect of the
continuous anchoring.
• ACI 549.4R-13 provisions conservatively predict the ultimate capacities of the FRCM-
strengthened beams exposed to post-strengthening corrosive environment. The ratios of
the experimental to the theoretical ultimate capacities (𝑃𝑢
𝑃𝑢𝑡ℎ⁄ ) in the range of 1.21 to 1.3
for the beams strengthened in Scheme II, while it was 1.02 for the specimen strengthened
in Scheme I.
8.9.5 Fatigue Performance of Corroded Beams Strengthened with
FRCM
• Strengthening corrosion-damaged beams with FRCM composites increased their fatigue
life by 38 to 377% of that of the corroded-unstrengthened beams without restoring the
fatigue life of the virgin beams. The enhancement in fatigue life was dependent on the
FRCM type, amount, and strengthening Scheme.
• Rupture of steel bars at the locations of corrosion pits was the governing mode of failure
in all of the unstrengthened and strengthened beams tested under fatigue. However, the
presence of FRCM composites mitigated this brittle failure in the strengthened beams.
• PBO-FRCM composite was more effective than the C-FRCM counterpart in restoring the
fatigue life of the corrosion-damaged beams. Beam (FCS-3C-II) strengthened with three
layers of C-FRCM (with 𝛽𝑓 = 8) survived 80% less cycles than those survived by the
beam (FCS-4P-II) strengthened with four layers of PBO-FRCM (with 𝛽𝑓 =5.48).
197
• The effect of FRCM configuration was more pronounced in fatigue than in monotonic
tests. With the same number of PBO fabric layers, the beams strengthened in Scheme II
exhibited fatigue life 150% more of that of its counterpart beam strengthened in Scheme
I.
• The FRCM material had a notable effect on the rate of stiffness degradation of the
strengthened beams tested in fatigue rather than the number of fabric layers and the
strengthening Scheme applied. The beam strengthened with C-FRCM demonstrated a
higher rate of stiffness degradation with cycling than that strengthened with PBO-FRCM.
8.9.6 Numerical simulation
• The comparison between the experimental and numerical results revealed that the
developed finite element (FE) models were able to predict the nonlinear flexural response
of the corroded beams strengthened with C-FRP and PBO-FRCM from initial loading to
ultimate with good accuracy. In addition, the numerical model was able to detect the
failure modes of the strengthened beams.
• The ratios of the predicted to experimental yield loads of the tested beams (𝑃𝑦𝐹𝐸 𝑃𝑦
𝐸𝑋𝑃⁄ )
ranged between 0.95 and 1.04 whereas the ratios of the predicted to experimental ultimate
loads (𝑃𝑢𝐹𝐸 𝑃𝑢
𝐸𝑋𝑃⁄ ) ranged between 0.93 and 1.07. These results indicated the accuracy of
the FE models to predict the flexural performance of the tested beams.
• Ultimate deflections predicted for beams strengthened with PBO-FRCM were within 7%
error band except in the case of beam strengthened with one layer of PBO-FRCM in
which the FE model underestimated its ultimate deflection by 21%. On the contrary, the
FE model overestimated the ultimate deflection for the C-FRP strengthened beam by
27%.
• The numerical results indicated the feasibility of using the bond slip models of D’Ambrisi
et al. [120] and Lu et al. [117] to simulate the behavior of PBO-FRCM and C-FRP-
strengthened beams at the composite/concrete interface.
198
• The results of the parametric study indicated that the use of concrete with low
compressive strengths decreased the load-carrying capacities of the strengthened beams
regardless of the strengthening system used (C-FRP or PBO-FRCM).
• The parametric study suggested that failure of FRCM-strengthened beams was
independent of the compressive strength of the concrete substrate. Failure of such beams
was governed by fabric slippage within the matrix. For beams strengthened with C-FRP,
lowering the compressive strength of the concrete substrate changed their mode of failure
from FRP rupture to debonding at concrete/laminate interface.
8.10 Recommendation for Future Work
The findings and conclusions of the current study approved the feasibility of FRCM composites
to strengthen corrosion-damaged RC beams. However, some topics still require further
investigations. Recommendations for future studies are:
• Studying the structural performance of RC beams damaged at higher corrosion levels
(more than 23% mass loss) and strengthened with FRCM.
• Examining the post-strengthening performance of FRCM-strengthened beams under
harsh environmental conditions including freeze/thaw and high temperatures.
• Investigating the ability of various FRCM systems to reduce the corrosion rate in the steel
reinforcement when applied at various levels of corrosion damage and quantifying the
change in corrosion rates.
• Investigating the possibility of using FRCM composites in strengthening continuous RC
beams and quantifying their ability to redistribute moments between the sections.
8.11 Impact of Current Research
The results and outcomes of this research work can be used as a fundamental step to include
design provisions for the flexural strengthening using FRCM composites in the Canadian codes
(S6 and S806). It can also be used to update the current design guidelines ACI 549.4R-13.
199
9. References
[1] Almusallam AA. Effect of degree of corrosion on the properties of reinforcing steel bars.
Construction and Building Materials 2001;15(8):361–8. https://doi.org/10.1016/S0950-
0618(01)00009-5
[2] Du YG, Clark LA, Chan AHC. Effect of corrosion on ductility of reinforcing bars.
Magazine of Concrete Research 2005;57:407–19.
https://doi.org/10.1680/macr.2005.57.7.407
[3] Ou YC, Nguyen ND. Influences of location of reinforcement corrosion on seismic
performance of corroded reinforced concrete beams. Engineering Structures
2016;126:210–23. https://doi.org/10.1016/j.engstruct.2016.07.048
[4] Fang C, Lundgren K, Chen L, Zhu C. Corrosion influence on bond in reinforced concrete.
Cement and Concrete Research 2004;34(11):2159–67.
https://doi.org/10.1016/j.cemconres.2004.04.006
[5] Broomfield JP. Corrosion of Steel in Concrete: Understanding, Investigation and Repair,
Second Edition. Taylor & Francis Group, London and New York. 2006.
[6] Sun J, Huang Q, Ren Y. Performance deterioration of corroded rc beams and reinforcing
bars under repeated loading. Construction and Building Materials 2015;96:404–15.
[7] Zhang W, Ye Z, Gu X, Liu X, Li S. Assessment of fatigue life for corroded reinforced
concrete beams under uniaxial bending. Journal of Structural Engineering 2017;143(7).
https://doi.org/10.1061/(ASCE)ST.1943-541X.0001778
[8] NACE International. Corrosion costs and preventive strategies in the united states.
Houston, USA. 2002. https://www.nace.org/uploadedFiles/Publications/ccsupp.pdf.
[9] International Union of Painters and Allied Trades. Pre-Budget Submission to the House of
Commons Standing Committee on Finance, Ottawa, Canada. 2014.
[10] Masoud S, Soudki K, Topper T. CFRP-strengthened and corroded RC beams under
monotonic and fatigue loads. Journal of Composites for Construction 2001;5(4):228–36.
https://doi.org/10.1061/(ASCE)1090-0268(2001)5:4(228)
200
[11] Dong JF, Wang QY, Guan ZW. Structural behaviour of RC beams externally strengthened
with FRP sheets under fatigue and monotonic loading. Engineering Structures
2012;41:24–33. https://doi.org/10.1016/j.engstruct.2012.03.024.
[12] Baggio D, Soudki K, Noël M. Strengthening of shear critical RC beams with various FRP
systems. Construction and Building Materials 2014;66:634–44.
https://doi.org/10.1016/j.conbuildmat.2014.05.097
[13] Bisby LA, Green MF, Kodur VKR. Response to fire of concrete structures that
incorporate FRP. Progress in Structural Engineering and Materials 2005;7:136–49.
[14] Fib bulletin 14. Externally bonded FRP reinforcement for RC structures. Technical Rep.,
Int. Federation for Structural Concrete, Luasanne, Switzerland. 2001.
[15] Chowdhury EU, Eedson R, Bisby LA, Green MF, Benichou N. Mechanical
Characterization of Fibre Reinforced Polymers Materials at High Temperature. Fire
Technology 2011;47(4):1063–80.
[16] Hashemi S, Al-Mahaidi R. Flexural performance of CFRP textile-retrofitted RC beams
using cement-based adhesives at high temperature. Construction and Building Materials
2012;28:791–7. https://doi.org/10.1016/j.conbuildmat.2011.09.015
[17] American Concrete Institute (ACI). Protection of metals in concrete against corrosion.
ACI 222R-01. Farmington Hills, MI, 2010.
[18] Zhang WP, Chen H, Gu XL. Bond behaviour between corroded steel bars and concrete
under different strain rates. Magazine of Concrete Research 2016;68(7):364–78.
https://doi.org/10.1680/jmacr.15.00174
[19] Zhang W, Song X, Gu X, Li S. Tensile and fatigue behavior of corroded rebars.
Construction and Building Materials 2012;34:409–17.
https://doi.org/10.1016/j.conbuildmat.2012.02.071
[20] El Maaddawy TA, Soudki KA. Effectiveness of impressed current technique to simulate
corrosion of steel reinforcement in concrete. Journal of Materials in Civil Engineering
2003; ;15(1):41–7 .https://doi.org/10.1061/(ASCE)0899-1561(2003)15:1(41)
201
[21] Malumbela G, Alexander M, Moyo P. Interaction between corrosion crack width and steel
loss in RC beams corroded under load. Cement and Concrete Research 2010;40(9):1419–
28. https://doi.org/10.1016/j.cemconres.2010.03.010
[22] Ou YC, Susanto YTT, Roh H. Tensile behavior of naturally and artificially corroded steel
bars. Construction and Building Materials 2016;103:93–104.
https://doi.org/10.1016/j.conbuildmat.2015.10.075
[23] Mancini G, Tondolo F, Iuliano L, Minetola P. Local reinforcing bar damage in RC
members due to accelerated corrosion and loading. Construction and Building Materials
2014;69:116–23. https://doi.org/10.1016/j.conbuildmat.2014.07.011
[24] Mangat PS, Elgarf MS. Flexural strength of concrete beams with corroding reinforcement
ACI Structural Journal,1999:149–58.
[25] Almusallam AA, Al-Gahtani AS, Aziz AR. Effect of reinforcement corrosion on bond
strength. Construction and Building Materials 1996;10:123–9.
[26] ASTM G1-03. Standard practice for preparing, cleaning, and evaluating corrosion test
specimens, ASTM G1-03. West Conshohocken, PA, 2011.
[27] Saifullah M, Clark LA. Effect of corrosion rate on the bond strength of corroded
reinforcment. International Conference on Corrosion and Corrosion Protection of Steel in
Concrete, Sheffield: Sheffield Acadamic Press; 1994, p. 591–602.
[28] Alonso C, Andrade C, Rodriguez J, Diez JM. Factors controlling cracking of concrete
affected by reinforcement corrosion. Materials and Structures 1998;31(7):435–41.
[29] Andrade C, Alonso C. On-site measurements of corrosion rate of reinforcements.
Construction and Building Materials 2001;15:141–5. https://doi.org/10.1016/S0950-
0618(00)00063-5
[30] Andrade C, Alonso C, Molina FJ. Cover cracking as a function of bar corrosion: Part I-
Experimental test. Materials and Structures 1993;26(8):453–64.
[31] Bhargava K, Ghosh AK, Mori Y, Ramanujam S. Analytical model for time to cover
cracking in RC structures due to rebar corrosion. Nuclear Engineering and Design
202
2006;236(11):1123–39. https://doi.org/10.1016/j.nucengdes.2005.10.011
[32] El Maaddawy T. Performance of corrosion-damaged reinforced concrete beams repaired
with cfrp laminates. PhD. Thesis, University of Waterloo, 2004.
[33] Andisheh K, Scott A, Palermo A. Preliminary estimation of reduction factors in
mechanical properties of steel reinforcement due to pitting simulated corrosion. NZSEE
Conference, Wellington, New Zealand, 2014.
[34] Du YG, Clark LA, Chan AHC. Residual capacity of corroded reinforcing bars. Magazine
of Concrete Research 2005;57(3):135–47. https://doi.org/10.1680/macr.2005.57.3.135
[35] Apostolopoulos CA. Mechanical behavior of corroded reinforcing steel bars S500s
tempcore under low cycle fatigue. Construction and Building Materials 2007;21(7):1447–
56. https://doi.org/10.1016/j.conbuildmat.2006.07.008
[36] Apostolopoulos CA, Papadopoulos MP. Tensile and low cycle fatigue behavior of
corroded reinforcing steel bars S400. Construction and Building Materials
2007;21(4):855–64. https://doi.org/10.1016/j.conbuildmat.2005.12.012
[37] Mancini G, Tondolo F. Effect of bond degradation due to corrosion - A literature survey.
Structural Concrete 2014;15(3):408–18.
[38] Wu YZ, Lv HL, Zhou SC, Fang ZN. Degradation model of bond performance between
deteriorated concrete and corroded deformed steel bars. Construction and Building
Materials 2016;119:89–95. https://doi.org/10.1016/j.conbuildmat.2016.04.061
[39] Lv HL, Wu YZ, Fang ZN, Zhou SC. Deterioration behavior of reinforced concrete beam
under compound effects of acid-salt mist and carbon dioxide. Construction and Building
Materials 2015;77:253–9. https://doi.org/10.1016/j.conbuildmat.2014.12.101
[40] Wang X, Liu X. A strain-softening model for steel-concrete bond. Cement and Concrete
Research 2003;33(10):1669–73. https://doi.org/10.1016/S0008-8846(03)00137-6
[41] Zandi K, Coronelli D, Lundgren K. Bond capacity of severely corroded bars with
corroded stirrups. Magazine of Concrete Research 2011;63:953–68.
[42] Sæther I. Bond deterioration of corroded steel bars in concrete. Structure and
203
Infrastructure Engineering 2011;7(6):415–29.
[43] Almusallam AA, Al-Gahtani AS, Aziz AR, Dakhil FH. Effect of Reinforcement
Corrosion on Flexural Behavior of concrete slabs. Journal of Matreial in Civil Enginering
1996; 8:(3)123–7. https://doi.org/10.1061/(ASCE)0899-1561(1996)8:3(123)
[44] Vidal T, Castel a., François R. Corrosion process and structural performance of a 17 year
old reinforced concrete beam stored in chloride environment. Cement and Concrete
Research 2007;37(11):1551–61. https://doi.org/10.1016/j.cemconres.2007.08.004
[45] Torres-Acosta AA, Navarro-Gutierrez S, Terán-Guillén J. Residual flexure capacity of
corroded reinforced concrete beams. Engineering Structures 2007;29(6):1145–52.
https://doi.org/10.1016/j.engstruct.2006.07.018
[46] Gu XL, Zhang WP, Shang DF, Wang XG. Flexural behavior of corroded reinforced
concrete beam.12th Biennial International Conference on Engineering, Hawaii, USA,
2010. https://doi.org/10.1061/41096(366)339
[47] Malumbela G, Alexander M, Moyo P. Variation of steel loss and its effect on the ultimate
flexural capacity of RC beams corroded and repaired under load. Construction and
Building Materials 2010;24(6):1051–9. https://doi.org/10.1016/j.conbuildmat.2009.11.012
[48] Xia J, Jin WL, Li LY. Effect of chloride-induced reinforcing steel corrosion on the
flexural strength of reinforced concrete beams. Magazine of Concrete Research
2012;64(6):471–85. https://doi.org/10.1680/macr.10.00169
[49] Dang VH, François R. Prediction of ductility factor of corroded reinforced concrete beams
exposed to long term aging in chloride environment. Cement and Concrete Composites
2014;53:136–47
[50] Heffernan PJ, Erki MA, Du-Quesnay DL. Stress redistribution in cyclically loaded
reinforced concrete beams. ACI Structural Journal 2004;101(2):261–8.
[51] Parvez A, Foster SJ. Fatigue behavior of steel-fiber-reinforced concrete beams. Journal of
Structural Engineering 2014;141(4). https://doi.org/10.1061/(ASCE)ST.1943-
541X.0001074
204
[52] Schläfli M, Brühwiler E. Fatigue of existing reinforced concrete bridge deck slabs.
Engineering Structures 1998;20:991–8. doi:10.1016/S0141-0296(97)00194-6.
[53] Yi W, Kunnath SK, Sun X, Shi C, Tang F. Fatigue behavior of reinforced concrete beams
with corroded steel reinforcement. ACI Structural Journal 2010;107:526–33.
[54] American Concrete Institute. Guide to design and construction of externally bonded
fabric-reinforced cementitious matrix (frcm) systems for repair and strengthening concrete
and masonry structures, ACI 549.4R-13. Farmington Hills, MI, 2013.
[55] Schladitz F, Frenzel M, Ehlig D, Curbach M. Bending load capacity of reinforced
concrete slabs strengthened with textile reinforced concrete. Engineering Structures
2012;40:317–26. https://doi.org/10.1016/j.engstruct.2012.02.029
[56] Triantafillou T, Papanicolaou C. Textile reinforced mortars (trm) versus fibre reinforced
polymers (frp) as strengthening materials of concrete structures. 7th International
Symposium of the Fiber-Reinforced Polymer Reinforcement for Reinforced Concrete
Structures. ACI Special Publication, 2005:99–118.
[57] Tetta ZC, Koutas LN, Bournas DA. Textile-reinforced mortar (TRM) versus fiber-
reinforced polymers (FRP) in shear strengthening of concrete beams. Composites Part B:
Engineering 2015;77:338–48. https://doi.org/10.1016/j.compositesb.2015.03.055
[58] Brückner A, Ortlepp R, Curbach M. Textile reinforced concrete for strengthening in
bending and shear. Materials and Structures 2006;39(8):741–8.
[59] Täljsten B, Blanksvärd T. Mineral-based bonding of carbon frp to strengthen concrete
structures. Journal of Composites for Construction 2007;11(2):120–8.
https://doi.org/10.1061/(ASCE)1090-0268(2007)11:2(120)
[60] American Concrete Institute (ACI).Thin fiber and textile reinforced cementitious
systems,SP-244CD, Technical Rep.by ACI Committee 549.Farmington Hills, MI, 2007
[61] Peled A, Bentur A. Fabric structure and its reinforcing efficiency in textile reinforced
cement composites. Composites Part A: Applied Science and Manufacturing
2003;34(2):107–18. https://doi.org/10.1016/S1359-835X(03)00003-4
205
[62] Peled A, Mobasher B, Sueki S . Technology methods in textile cement-based composites.
RELEM Proceeding PRO36. 2004:187-202
[63] Mobasher B, Peled A. Effect of processing on mechanical properties of textile - reinforced
concrete. Textile Reinforced Concrete (TRC)-Symposium sponsored by Committees 549-
544, ACI Special Publications,2007.
[64] Peled A, Bentur A. Geometrical characteristics and efficiency of textile fabrics for
reinforcing cement composites. Cement and Concrete Research 2000;30(5):781–90.
https://doi.org/10.1016/S0008-8846(00)00239-8
[65] AC434. Acceptance criteria for masonry and concrete strengthening using fabric-
reinforced cementitious matrix (FRCM) composite systems. ICC-Evaluation Service,
Whittier, CA, 2013.
[66] International Code Council. International Building Code (IBC). Country Club Hills, IL
2009.
[67] Weiland S, Ortlepp R, Bruckner A, Curbach M. Strengthening of RC Structures with
textile reinforced concrete (TRC). ACI Journal Proceedings 2007;244:157–72.
[68] Pino VA. Fabric reinforced Ccementitious matrix ( FRCM ) composites as a repair system
for transportation infrastructure. Ph.D. Thesis, Univerity of Miami, 2016.
[69] Arboleda D. Fabric reinforced Ccementitious matrix (FRCM) composites for
iInfrastructure strengthening and rehabilitation : characterization methods. Ph.D thesis
University of Miami, 2014.
[70] Mechtcherine V. Novel cement-based composites for the strengthening and repair of
concrete structures. Construction and Building Materials 2013;41:365–73.
https://doi.org/10.1016/j.conbuildmat.2012.11.117
[71] Hartig J, Häußler-Combe U, Schicktanz K. Influence of bond properties on the tensile
behaviour of Textile Reinforced Concrete. Cement and Concrete Composites
2008;30(10):898–906. https://doi.org/10.1016/j.cemconcomp.2008.08.004
[72] Ortlepp R, Hampel U, Curbach M. A new approach for evaluating bond capacity of TRC
206
strengthening. Cement and Concrete Composites 2006;28(7):589–97.
https://doi.org/10.1016/j.cemconcomp.2006.05.003
[73] Gartner A, Douglas EP, Dolan CW, Hamilton HR. Small beam bond test method for cfrp
composites applied to concrete. Journal of Composites for Construction 2011;15(1):52–
61. https://doi.org/10.1061/(ASCE)CC.1943-5614.0000151
[74] Banholzer B, Brockmann T, Brameshuber W. Material and bonding characteristics for
dimensioning and modelling of textile reinforced concrete (TRC) elements. Materials and
Structures 2006;39:749–63.
[75] Ombres L. Analysis of the bond between Fabric Reinforced Cementitious Mortar (FRCM)
strengthening systems and concrete. Composites Part B: Engineering 2015;69:418–26.
https://doi.org/10.1016/j.compositesb.2014.10.027
[76] D’Ambrisi A, Feo L, Focacci F. Experimental analysis on bond between PBO-FRCM
strengthening materials and concrete. Composites Part B: Engineering 2013;44(1):524–32.
https://doi.org/10.1016/j.compositesb.2012.03.011
[77] Sneed LH, D’Antino T, Carloni C, Pellegrino C. A comparison of the bond behavior of
PBO-FRCM composites determined by double-lap and single-lap shear tests. Cement and
Concrete Composites 2015;64:37–48. https://doi.org/10.1016/j.cemconcomp.2015.07.007
[78] Antino TD, Carloni C, Sneed LH, Pellegrino C. Matrix – fiber bond behavior in PBO
FRCM composites : A fracture mechanics approach. Engineering Fracture Mechanics
2014;117:94–111. https://doi.org/10.1016/j.engfracmech.2014.01.011
[79] Arboleda D, Babaeidarabad S, Hays CD, Nanni A. Durability of fabric reinforced
cementitious matrix (FRCM) composites. The 7th International Conference on FRP
Composites in Civil Engineering, Vancouve, Canada: 2014.
[80] Ombres L. Flexural analysis of reinforced concrete beams strengthened with a cement
based high strength composite material. Composite Structures 2011(1);94:143–55.
https://doi.org/10.1016/j.compstruct.2011.07.008
207
[81] D’Ambrisi A, Focacci F. Flexural strengthening of RC beams with cement-based
composites. Journal of Composites for Construction 2011;15:707–20.
https://doi.org/10.1061/(ASCE)CC.1943-5614.0000218
[82] Si Larbi A, Contamine R, Hamelin P. TRC and hybrid solutions for repairing and/or
strengthening reinforced concrete beams. Engineering Structures 2012;45:12–20
https://doi.org/10.1016/j.engstruct.2012.06.002.
[83] Elsanadedy HM, Almusallam TH, Alsayed SH, Al-Salloum YA. Flexural strengthening of
RC beams using textile reinforced mortar–Experimental and numerical study. Composite
Structures 2013;97:40–55. https://doi.org/10.1016/j.compstruct.2012.09.053
[84] Babaeidarabad S, Loreto G, Nanni A. Flexural strengthening of rc beams with an
externally bonded fabric-reinforced cementitious matrix. Journal of Composites for
Construction, 2014;18(5). https://doi.org/10.1061/(ASCE)CC.1943-5614.0000473
[85] Loreto G, Leardini L, Arboleda D, Nanni A. Performance of RC slab-type elements
strengthened with fabric-reinforced cementitious-matrix composites. Journal of
Composites for Construction, 2014;18(3). https://doi.org/10.1061/(ASCE)CC.1943-
5614.0000415
[86] El-maaddawy T, EL Refai A. Innovative repair of severely corroded T-beams using
fabric-reinforced cementitious matrix. Journal of Composites for Construction 2017;21(1).
https://doi.org/10.1061/(ASCE)CC.1943-5614.0000726
[87] Aljazaeri ZR, Myers JJ. Fatigue and flexural behavior of reinforced concrete beams
strengthened with a fiber reinforced cementitious matrix. Journal of Composites for
Construction 2017;21(1). https://doi.org/10.1061/(ASCE)CC.1943-5614.0000726
[88] Pino V, Akbari Hadad H, De Caso Basalo F, Nanni A, Ali Ebead U, El Refai A.
Performance of FRCM-strengthened RC beams subject to fatigue. Journal of Bridge
Engineering, 2017;22(10). https://doi.org/10.1061/(ASCE)BE.1943-5592.0001107
[89] Masoud S, Soudki K. Evaluation of corrosion activity in FRP repaired RC beams. Cement
and Concrete Composites 2006;28(10):969–77.
https://doi.org/10.1016/j.cemconcomp.2006.07.013
208
[90] Al-Saidy A. H, Al-Jabri KS. Effect of damaged concrete cover on the behavior of
corroded concrete beams repaired with CFRP sheets. Composite Structures
2011;93(7):1775–86. https://doi.org/10.1016/j.compstruct.2011.01.011
[91] Malumbela G, Alexander M. Load-bearing capacity of corroded , patched and FRP-
repaired RC beams. Magazine of Concrete Research Research 2011;63(11):797–812.
https://doi.org/10.1680/macr.2011.63.11.797
[92] Al-Salloum YA, Elsanadedy HM, Alsayed SH, Iqbal R. Experimental and numerical
study for the shear strengthening of reinforced concrete beams using textile-reinforced
mortar. Journal of Composites for Construction 2012;16(1):74–90
https://doi.org/10.1061/(ASCE)CC.1943-5614.0000239
[93] Hashemi S, Al-Mahaidi R. Experimental and finite element analysis of flexural behavior
of FRP-strengthened RC beams using cement-based adhesives. Construction and Building
Materials 2012;26(1):268–73. https://doi.org/10.1016/j.conbuildmat.2011.06.021
[94] Blanksvärd T. Strengthening of concrete structures by the use of mineral based
composites. Ph.D Thesis, Luleå University of Technology 2007.
[95] Wang DY, Wang ZY, Asce AM, Yu T, Li H. Shake table tests of large-scale substandard
RC frames retrofitted with CFRP wraps before earthquakes. Journal of Composites for
Construction 2017;21(1). https://doi.org/10.1061/(ASCE)CC.1943-5614.0000720
[96] Elghazy M, Refai AE, Ebead UA, Nanni A. Performance of corrosion-aged reinforced
concrete (RC) beams rehabilitated with fabric-reinforced cementitious matrix (FRCM).
Proceedings of the 4th International Conference in Sustainable Construction Materials and
Technologies (SCMT4), Las Vegas, USA, 2016.
[97] Ebead UA, Shrestha KC, Afzal MS, Refai A El, Nanni A. Effectiveness of fabric-
feinforced cementitious matrix in strengthening reinforced concrete beams. Journal of
Composites for Construction 2017;21(2). https://doi.org/10.1061/(ASCE)CC.1943-
5614.0000741m
[98] American Concrete Institute (ACI). Building code requirements for structural concrete (
ACI 318M-08 ) and Commentary. Farmington Hills, MI, 2014.
209
[99] Awani O, El-Maaddawy T, Ismail N. Fabric-reinforced cementitious matrix: A promising
strengthening technique for concrete structures. Construction and Building Materials
2016;132:94–111. https://doi.org/10.1016/j.conbuildmat.2016.11.125
[100] Elghazy M, El Refai A, Ebead U, Nanni A. Corrosion-damaged reinforced concrete beams
repaired with fabric-reinforced cementitious matrix (FRCM). Journal of Composites for
Construction, 2017, In press.
[101] liu, Y. and Weyers R. Modeling the time-to-corrsion cracking in chloride contaminated
reinforcd concrete structures . ACI Materials Journal 1998;95:675–81.
[102] Azad AK, Ahmad S, Al-Gohi BHA. Flexural strength of corroded reinforced concrete
beams. Magazine of Concrete Research 2010;62(6):405–14.
https://doi.org/10.1680/macr.2010.62.6.405
[103] Bournas DA, Lontou P, Papanicolaou, C. G. Triantafillou TC. Textile-reinforced mortar
(TRM) versus FRP confinement in reinforced concrete columns. ACI Structural Journal
2007;104:740–8.
[104] Jesse F, Weiland S, Curbach M. Flexural strengthening of RC structures with textile-
reinforced concrete. Textile-Reinforced Concrete (ACI-SP250), ACI Special Publications
2007:41–50
[105] Harajli M, ElKhatib H, San-Jose JT. Static and cyclic out-of-plane response of masonry
walls strengthened using textile-mortar system. Journal of Materials in Civil Engineering
2010;22(11):1171–80. https://doi.org/10.1061/(ASCE)MT.1943-5533.0000128
[106] Elsanadedy HM, Almusallam TH, Alsayed SH, Al-Salloum YA. Flexural strengthening of
RC beams using textile reinforced mortar-Experimental and numerical study. Composite
Structures 2013;97:40–55. https://doi.org/10.1016/j.compstruct.2012.09.053
[107] Ma Y, Xiang Y, Wang L, Zhang J, Liu Y. Fatigue life prediction for aging RC beams
considering corrosive environments. Engineering Structures 2014;79(15):211–21.
https://doi.org/10.1016/j.engstruct.2014.07.039
[108] Ekenel M, Myers JJ. Fatigue Performance of CFRP strengthened RC beams under
environmental conditioning and sustained load. Journal of Composites for Construction
210
2009;13(2):93–102. https://doi.org/10.1061/(ASCE)1090-0268(2009)13:2(93)
[109] Yin S, Sheng J, Wang X, Li S. Experimental investigations of the bending fatigue
performance of TRC-strengthened RC beams in conventional and aggressive chlorate
environments. Journal of Composites for Construction 2014;20(2).
https://doi.org/10.1061/(ASCE)CC.1943-5614.0000617
[110] Ebead U, Shrestha KC, Afzal MS, Refai A El, Nanni A. Effectiveness of fabric-reinforced
cementitious matrix in strengthening reinforced concrete beams. Journal of Composites
for Construction 2017;21(2). https://doi.org/10.1061/(ASCE)CC.1943-5614.0000741
[111] Dong J, Wang Q, Guan Z. Structural behaviour of RC beams with external flexural and
flexural–shear strengthening by FRP sheets. Composites Part B: Engineering
2013;44(1):604–12. https://doi.org/10.1016/j.compositesb.2012.02.018
[112] Wu HC, Sun P, Teng J. Development of fiber-reinforced cement-based composite sheets
for structural retrofit. Journal of Materials in Civil Engineering 2010;22(6).
https://doi.org/10.1061/(ASCE)MT.1943-5533.0000056
[113] Al-salloum YA, Siddiqui NA, Elsanadedy HM, Abadel AA, Aqel MA. Textile-reinforced
mortar versus FRP as strengthening material for seismically deficient RC beam-column
joints. Journal of Composites for Construction 2011;15(6):920–33.
https://doi.org/10.1061/(ASCE)CC.1943-5614.0000222
[114] Elsayed W, Ebead UA, Neale KW. Studies on mechanically fastened fiber-reinforced
polymer strengthening systems. ACI Structural Journal 2009;106:49–59.
[115] Ebead UA, Saeed H. Modeling of reinforced concrete slabs strengthened with fiber-
reinforced polymer or steel plates. ACI Structural Journal 2010;107:218–27.
[116] Abdel Baky H, Ebead UA, Neale KW. Nonlinear micromechanics-based bond-slip model
for FRP/concrete interfaces. Engineering Structures 2012;11-23.
https://doi.org/10.1016/j.engstruct.2012.01.010
[117] Lu XZ, Teng JG, Ye LP, Jiang JJ. Bond-slip models for FRP sheets/plates bonded to
concrete. Engineering Structures 2005;27(6):920–37.
https://doi.org/10.1016/j.engstruct.2005.01.014
211
[118] Chen GM, Chen JF, Teng JG. On the finite element modelling of RC beams shear-
strengthened with FRP. Construction and Building Materials 2012;32:13–26.
https://doi.org/10.1016/j.conbuildmat.2010.11.101
[119] Ombres L, Verre S. Structural behaviour of fabric reinforced cementitious matrix ( FRCM
) strengthened concrete columns under eccentric loading. Composites Part B
2015;75:235–49. https://doi.org/10.1016/j.compositesb.2015.01.042
[120] D’Ambrisi A, Feo L, Focacci F. Bond-slip relations for PBO-FRCM materials externally
bonded to concrete. Composites Part B: Engineering 2012;43(8):2938–49.
https://doi.org/10.1016/j.compositesb.2012.06.002
[121] Elghazy M, El Refai A, Ebead UA, Nanni A. Effect of corrosion damage on the flexural
performance of RC beams strengthened with FRCM composites. Composite Structures
2017;180:994-1006. https://doi.org/10.1016/j.compstruct.2017.08.069
[122] ATENA [Computer software]. Červenka Consulting s.r.o. Prague, Czech Republic, 2016.
[123] Menetrey P, Willam KJ. Triaxial Failure Criterion for Concrete and its Generalization.
ACI Structural Journal 1995;92:331–318.
[124] Van Mier JGM. Multi-axial Strain-softening of Concrete, Part I: Fracture. Materials and
Structures 1986;19(3):190–200.
[125] Vecchio FJ, Collins MP. The modified compression-field theory for reinforced concrete
elements subjected to shear. ACI Journal Proceedings 1986;83:219–31.
[126] Cervenka V. Constitutive model for cracked reinforced concrete. ACI Journal Proceedings
1985;82:877–82.
[127] Hordijk DA. Local approach to fatigue of concrete, Doctor dissertation. Delft University
of Technology, 1991.
[128] Dai J, Ueda T, Sato Y. Development of the nonlinear bond stress–slip model of fiber
reinforced plastics sheet–concrete interfaces with a simple method. Journal of Composites
for Construction 2005;9(1):52–62. https://doi.org/10.1061/(ASCE)1090-
0268(2005)9:1(52)
212
10. Bibliography
During this research work at Laval University, the candidate has authored the following
publications:
Journal Articles:
1. Elghazy, M., El Refai, A., Ebead, U., and Nanni, A. “Corrosion-Damaged Reinforced
Concrete Beams Repaired with Fabric-Reinforced Cementitious Matrix (FRCM).” Journal of
Composites for Construction, ASCE. Under review. In review. Submitted in the revised form:
February 2018.
2. Elghazy, M., El Refai, A., Ebead, U., and Nanni, A. (2017). “Effect of Corrosion Damage on
the Flexural Performance of RC Beams Strengthened with FRCM Composites.” Journal of
Composite Structures. Date of acceptance: https://doi.org/10.1016/j.compstruct.2017.08.069
3. Elghazy, M., El Refai, A., Ebead, U., and Nanni, A. “Post-repair Flexural Performance of
Corroded Beams Rehabilitated with Fabric-Reinforced Cementitious Matrix (FRCM) under
Corrosive Environment.” Journal of Construction and Building Materials. Date of acceptance:
23 January 2018. https://doi.org/10.1016/j.conbuildmat.2018.01.128
4. Elghazy, M., El Refai, A., Ebead, U., and Nanni, A. “Fatigue and Monotonic Behavior of
Corrosion-damaged Reinforced Concrete Beams Strengthened with FRCM Composites.”
Journal of Composites for Construction, ASCE. In review. Submitted in the revised form:
February 2018.
5. Elghazy, M., El Refai, A., Ebead, U., and Nanni, A. “Finite Element Modeling and
Experimental Results of Corroded Concrete Beams Strengthened with Externally-Bonded
Composites.” Journal of Engineering Structures. In review. Date of submission: November
2017.
213
Refereed Conference Papers:
1. Elghazy, M., El Refai, A., Ebead, U. and Nanni, A. “Post-Repair Flexural Performance of
Corrosion-Damaged Beams Repaired with Fabric-Reinforced Cementitious Matrix.” The 9th
International Conference on Fibre-Reinforced Polymer (FRP) Composites in Civil
Engineering - CICE 2018, Paris, France, July 2018.
2. Elghazy, M., El Refai, A., Ebead, U. and Nanni, A. “Experimental Results and Modeling of
Corroded Concrete Beams Strengthened with FRCM.” The 10th International Conference on
Short and Medium Span bridges, Quebec City, Canada, August 2018.
3. Elghazy, M., El Refai, A., Ebead, U. and Nanni, A. “Corrosion-Damaged Beams Repaired
with Fabric-Reinforced Cementitious Matrix under Fatigue Load.” The 10th International
Conference on Short and Medium Span bridges, Quebec City, Canada, August 2018.
4. Elghazy, M., El Refai, A., Ebead, U. and Nanni, A. (2017). “Corrosion-Damaged Beams
Repaired with Carbon-Fabric-Reinforced Cementitious Matrix,” The 5th International
Conference on Durability of Fiber Reinforced Polymer (FRP) Composites for Construction
and Rehabilitation of Structures, CDCC 2017, Sherbrooke, Canada, July 2017.
5. Elghazy, M., El Refai, A., Ebead, U. and Nanni, A. (2016). “Performance of Corrosion-Aged
Reinforced Concrete Beams Rehabilitated with Fabric-Reinforced Cementitious Matrix
(FRCM),” The 4th International Conference on Sustainability Construction Materials and
Technologies, SCMT4, Las Vegas, USA, August 2016.
214