Accepted Manuscript

Fatigue behaviour of damaged RC beams Strengthened with Ultra high per-

formance fibre Reinforced Concrete

A. RamachandraMurthy, P. Vindhya Rani, D. ShanmugaPriya

PII: S0142-1123(18)30281-0
DOI: https://doi.org/10.1016/j.ijfatigue.2018.06.046
Reference: JIJF 4750

To appear in: International Journal of Fatigue

Received Date: 28 December 2017

Revised Date: 28 June 2018
Accepted Date: 29 June 2018

Please cite this article as: RamachandraMurthy, A., Vindhya Rani, P., ShanmugaPriya, D., Fatigue behaviour of
damaged RC beams Strengthened with Ultra high performance fibre Reinforced Concrete, International Journal of
Fatigue (2018), doi: https://doi.org/10.1016/j.ijfatigue.2018.06.046

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Fatigue behaviour of damaged RC beams Strengthened with Ultra high performance fibre
Reinforced Concrete
RamachandraMurthy A*, Vindhya Rani P1 and ShanmugaPriya D2
*CSIR-Structural Engineering Research Centre, Chennai, India, 600113
1
M.E. Student, GVP college of Engineering, Visakhapatnam, India
2
Assistant Professor, Dhanalakshmi College of Engineering, Chennai, India, 601301

Abstract.This paper critically examines the performance of reinforced concrete (RC) beams retrofitted
with a thin strip of ultra-high performance fibre-reinforced concrete (UHPFRC) under cyclic loading. RC
beams were preloaded under static loading approximately to 70%, 80% and 90% of maximum load of
control beams and then tested under fatigue loading with a stress ratio of 0.1 and frequency 2 Hz after
retrofitting with a UHPFRC thin strip of 10 mm thickness. It has been found that the damaged RC beams
can be successfully strengthened and rehabilitated by using a thin precast UHPFRC strip adhesively
bonded to the prepared tensile surface of the damaged beams. No de-lamination of the strip was observed
in any of the retrofitted beams. A finite element model was developed to predict the number of cycles to
failure and load-deflection behavior of the retrofitted RC beams. The model accounts for the (i) degree of
pre-damage, (ii) fracture behavior of concrete and UHPFRC through their respective specific fracture
energy and stress-crack opening relation, and (iii) elasto-plastic behavior of the reinforcing steel. The
model predictions are in very good agreement with the corresponding test results. It can be concluded that
this UHPFRC is an excellent candidate for the repair and rehabilitation of damaged RC flexural elements.
Keywords: RC beam; Static loading; Pre-damage; Ultrahigh strength concrete; Retrofitting; fatigue
loading; Finite element analysis

1. Introduction
In recent years, many reinforced concrete (RC) structures have been found to suffer from various
deteriorations such as cracks, concrete spalling, large deflections, etc., which need to be rectified
to support the design or to resist possible higher loading or to repair existing cracks. These
deteriorations are caused by numerous factors such as aging, corrosion of steel reinforcement,
environmental effects from seawater or accidental impacts on the structure.
*
Corresponding author, Ph.D., E-mail: murthyarc@serc.res.in
1
M. Tech., E-mail: vindhyacivil@gmail.com
2
M.E, E-mail: dspshanmu1992@gmail.com
Strengthening of existing concrete structures is necessaryto enhance their load-carrying capacity
and to satisfy serviceability criteria.Some structural components such as bridge girders, slabs in
parking garages, airport pavements and machine foundations are subjected to repeated loads
during their service lifetime.This type of loading may cause failure even when the nominal peak
loads are smaller than the ultimate capacity of the structure [1-4].The performance of popular
techniques of retrofitting using externally bondedsteel plates and fiber-reinforced plastic (FRP)
laminates has been extensively investigated [5-8].Although the techniqueof retrofitting using
externally bonded steel plates has gained popularity, being relatively quick, causing minimal site
disruption, and producing only minimal change in section size, several drawbacks have been
identified such as the occurrence of undesirable shear failures, difficultyin handling heavy steel
plates, corrosion of the steel, and the need for butt joint systems as a result of limited workable
lengths. Similarly, CFRP materials have gained popularity due to high strength to weight ratio,
stiffness to weight ratio and chemical inertness [8-13]. The limitation of this method is brittle
failure i.e. rapid crack propagation without appreciable deformation due to mismatch of tensile
strength and stiffness with that of concrete. The drawbacks of the above two methods have led to
the development of a new fibre reinforced ultra-high strength concrete with high mechanical and
durability properties making it a suitable choice for retrofitting [14-19]. This material has a
unique combination of low permeability, high strength, high tensile strength and high toughness.
Elimination of coarse aggregate in the matrix improves the homogeneity and the superior
characteristics are due to lower amount of free water. Ductility is the major quality that
differentiates it from other repair materials. Ultra-high performance fibre reinforced concrete
(UHPFRC) has also excellent bond with normal concrete [20- 25].

1.1 Research Significance

Although UHPFRC had been identified as a promising material for repair and retrofitting of RC
structures more than a decade ago, few investigations on retrofitted RC structural members with
UHPFRC overlay under fatigue loading were reported in the literature [14-25]. The novelty of
the present investigation is two-fold. Firstly, the effect of the level of damage of the concrete
element on its load bearing capacity after retrofitting with a thin UHPFRC strip has been
examined. Secondly, a finite element model has been developed for predicting the number of
cycles to failure that accounts for the (i) degree of pre-damage, (ii) fracture behavior of concrete
and UHPFRC through their respective specific fracture energy and stress-crack opening relation,
and (iii) elasto-plastic behavior of the reinforcing steel. The extensive experimental and
numerical investigations confirm that UHPFRC is an excellent prospective candidate for the
repair and rehabilitation of damaged RC flexural elements. The present study involves
experimental and numerical studies on preloaded and retrofitted RC beams under fatigue cyclic
loadingto provide some additional useful information.

2. Experimental Studies
The experimental investigations were carried out on under-reinforced beams cast in the
laboratory. A few test beams were tested to failure to determine the ultimate failure load (named
hereafter control beams) and the remaining beams were subjected to different levels of damage
by static loading and then unloaded. These statically pre-damaged beams were retrofitted with a
UHPFRC strip on the tension face and tested to failure under constant amplitude fatigue loading.
A further set of test beams were subjected to fatigue loading to induce flexural cracks and then
unloaded. These fatigue pre-damaged beams were also retrofitted with a UHPFRC strip and then
tested to failure under fatigue loading.
Details of materials and mixes
Concrete mixes were designed with the grades of compressive strength according to the Indian
standards [26]. The mix was made of ordinary Portland cement 53 grade, natural sand, crushed
aggregate size below 12mm and potable water, (Normal strength concrete, NSC). The NSC mix
proportion by weight of cement, fine aggregates, coarse aggregates and water was taken as
1:1.67:1.86:0.45. The UHPFRC mix consists of cement, silica fume, Quartz sand, Quartz powder
and water in the ratio of 1:0.25:1.1:0.4:0.23, respectively. Brass-coated steel fibers diameter
0.18mm and length 13mm are used to prevent corrosion. The content of steel fibers was 2% by
volume of concrete and the dosage of superplasticizer was 3.5% of dry mass of binder. The NSC
specimens were de-moulded after 1 day and cured in a water tank at ambient temperature for 28
days. The UHPFRC specimens were also de-moulded after 1 day and cured in water at ambient
temperature for 2 days. The specimens were then placed in an autoclave at 90°C for 2 days and
in an oven at 200°C for 1 day. Thereafter the specimens were air cooled for about 6 hours and
kept in water at ambient temperature for a further 1 day before testing. Compression and split
tensile tests were carried out on cylindrical specimens 150×300 mm (diameter x height) in the
case of NSC and 75x150mm in the case of UHPFRC. The average compressive strength and split
tensile strength for NSC were 35MPa and 3.2MPa, respectively. For UHPFRC the values were
122.5MPa and 20.7MPa, respectively. The specific fracture energy values for NSC and
UHPFRC were taken as 185 N/m and 13760 N/m, respectively, as measured by Ramachandra
Murthy et al. (2013) [26,29].
Casting and testing of beams
The dimensions of the RC beam are 1500mm (length) x 100mm (width) x 200mm (depth). The
reinforcement details of a typical RC beam including sectional view is shown in Fig. 1. The
dimensions of UHPFRC are 1500mm (length) x 100mm (width) x 10mm (thickness).Table 1
shows the details of all the beams used in the experimental studies. The details include
designation of the beam, degree of preloading and thickness of UHPFRC strip for retrofitting. In
view of previous experience, it was decided to retrofit all the pre-damaged beams with 10mm
UHPFRC strip.

P/2 P/2
A
6mmØ @100mmc/c 6mmØ @100mmc/c
400 400 400

2-8mmØ

200
2-10mmØ

A 100
1200
1500

Beam Elevation Section at A-A

Fig 1 Geometry of the RC beam

All the beams were cast with NSC mix and were removed from their moulds after one day
followed by water curing at ambient temperature for 28 days. Three beams were tested to failure
as control beams to compare the performance of those retrofitted with UHPFRC strip. The beams
were tested under four point bending over an effective span of 1200mm by simulating simply
supported conditions. Testing was carried out under displacement control with a loading rate of
0.5mm/min. A linearly varying displacement transducer(LVDT)was used to measure the mid-
span vertical displacement. The load-displacement curves of all specimens were automatically
recorded. Typical test set up and failure pattern of a control beam are shown in Fig. 2. Fig. 3
shows the average load-displacement curve (CBA). The average cracking load, yielding load and
ultimate loads are 33.21kN, 71.34kN and 77.84kN, respectively. The average vertical deflection
corresponding to yielding load and ultimate load are 6.67mm and 14.13mm, respectively. The
average maximum deflection is 29.10mm.

Fig 2 Typical test set-up and failure pattern of control beam

Fig 3 Average load-deflection curve of control beams

 Table 1 Details of beams cast and tested
Type of Beam Type of beam Beam Type of beam Beam
Beam designation designation designation
Control CB4 90%Preloaded B9 90%Preloaded RB9
beams CB2 (Static) (2 beams) (Static)& (2 beams)
(Static) CB5 strengthened
Average followed by
CBA fatigue cycling

70%Preloade B6 70%Preloaded RB6 Control fatigue FB2

d (2 beams) (Static)&retrofitt (2 beams) FB3
(Static) ed followed by Average
fatigue cycling FBA
80%Preloade B7 80%Preloaded RB7 Fatigue pre- RE3
d (2 beams) (Static)& (2 beams) loaded to 10000
(Static) Retrofitted cycles &
followed by strengthened
fatigue cycling followed by
fatigue cycling

Fatigue pre- RE4

loaded to 20000
cycles &
strengthened
followed by
fatigue cycling

From Figs. 2 and 3 it is observed that the response of the beam is almost linear elastic up to first
crack load, i.e. about 30kN. Later, flexural and shear cracks were formed with an increase of
load. Yielding of steel occurred at about 65 kN and final failure by crushing of concrete in local
loaded region. In all the tested beams, the maximum crack depth is observed to be in the range of
160 –170mm. All the control beams failed predominantly in the flexural mode. Final failure
occurred close to the loading point. The shear cracks formed during testing were found not to be
responsible for the final failure of the beam.
2.1 Pre-damage of RC beams
To induce pre-damage, a set of beams were pre-loaded to approximately 70%, 80% and 90% of
mean ultimate load of control beams. Table 1 presents the details of pre-loaded beams. Hence, no
repair was required prior to the attachment of the UHPFRC thin strip. Fig. 4 presents the typical
load - deflection curves of various pre-loaded beams under static loading. It can be readily
noticed from Fig. 4 that the steel yielded to some extent at 90% pre-load. Table 2 presents the
salient load and deflection values of pre-loaded beams.

Fig 4 Load vs Deflection curves for Pre-loaded beams

Table 2 Salient Features of Pre-Loaded Beams

Beam details Beam Cracking Maximum Deflection at
designation load, kN load, kN Maximum load,
mm
Approx.70% pre- B6 27.32 54.31 3.83
load B6 28.45 55.62 4.10
Avg. 27.88 54.96 3.96
Approx.80% B7 31.41 68.12 5.99
pre-load B7 32.33 67.94 6.07
Avg. 31.87 68.03 6.03
Approx.90% pre- B9 31.26 74.53 8.92
load B9 30.57 74.11 9.01
Avg. 30.92 74.32 8.97

Two beams were tested under sinusoidal fatigue loading with a frequency of 2 Hz and stress ratio
0.1 (Fig.5). The maximum load is 68.0 kN and the minimum load is 6.8 kN. During testing,
deflection and number of cycles to failure were recorded. These beams are called as control
fatigue beams in Table 1. Fig. 6 shows the plot of number of cycles vs deflection. Average plot
is also shown in Fig. 6.
 Fig 5 Typical fatigue loading spectrum of control beam

Fig 6 Deflection vs No. of cycles to failure (control fatigue beams)

From Fig. 6, it can be seen that the number of cycles to failure and the corresponding deflection
for FB2 are 34598 cycles and 14.82 mm, respectively. The values for FB3 are 36590 cycles and
16.05 mm. The average values (FBA) are 35594 cycles and the corresponding deflection is
15.43 mm. Typical failure pattern of a fatigue control beam is shown in Fig. 7 and it occurs close
to the loading points.
 Fig 7Typical failure pattern of fatigue control beam
Two beams were pre-loaded under fatigue loading to about 30% of number of cycles to failure of
FBA, i.e.to 10000 cycles at same frequency and stress ratio and a further two beams to about
20000 cycles (55% of the number of cycles to failures of FBA).
2.2 Strengthening schemes and instrumentation
UHPFRC strips (1450 mm x 100 mm x 10mm, length x width x thickness) were made using the
mix proportions already mentioned. The strips were cured in moulds for 24 h before demoulding.
After demoulding, they were cured in water for two days, in autoclave at 90°C for 2 days and in
an oven at 200°C for 1 day. Thereafter they were air cooled for 6 hours and kept in water for a
day before use. The tension surface of beam was prepared by carefully cleaning and roughening
using an angle grinder to improve the bond between the UHPFRC strip and the pre-damaged
beam. The strip was bonded to the prepared surface of the pre-damaged beam with a commercial
epoxy adhesive of average 3mm thickness. Fig. 8 shows a typical UHPFRC strip and its
attachment to pre-damaged beam surface.
 Fig 8 Strengthening of pre-damaged RC beam using UHPFRC strip

All the pre-loaded beams were retrofitted with UHPFRC strips 10 mm thick, as per Table 1.
Retrofitted beams were tested under displacement control in four point bending. The retrofitted
beams RB6, RB7, RB9, RE3 and RE4 were tested under fatigue loading with stress ratio 0.1
(maximum load = 68kN and the minimum load = 6.8 kN). The responses such as central vertical
deflection and fatigue cycles were captured during the experiment. The central vertical deflection
was measured with an LVDT. Fig. 9 presents the variation of deflection with the number of
cycles for the retrofitted beams and for the control beam tested under fatigue loading. Table 3
presents the salient features of statically pre-damaged and retrofitted beams tested under fatigue
loading.
 Fig 9 Deflection vs No. of cycles to failure
( fatigue life of statically pre-loaded retrofitted beams and control beam

Table 3 Deflection vs No. of cycles for statically pre-damaged and

retrofitted beams
Beam ID Pre-damage static Total No. of cycles Deflection, mm
load to failure
(%)
FBA ---- 35594 15.44
RB6 70 88229 13.64
RB7 80 82152 13.47
RB9 90 80698 13.61

From Fig. 9 and Table 3, it can be observed that the retrofitted beams have superior fatigue
performance irrespective of static pre-damage load level, with the number of cycles to failure
more than twice that of the control beam without retrofitting. The deflection of all the retrofitted
beams is, as expected, smaller than the control beam because of their higher flexural rigidity.
Fig. 10 shows the typical failure pattern of retrofitted beams.
 Fig 10 Typical failure pattern of the strengthened beams

Fig 11 Deflection vs No. of cycles to failure

(fatigue life of cyclic pre-loaded retrofitted beams and control beam)

Fig. 11 shows the number of cycles Vs deflection for the beams pre-damaged by cyclic loading
and retrofitted and for the un-retrofitted control beam. Table 4 presents the salient features of the
tested beams. Table 4 depicts the No. of cycles vs deflection for the beams pre-damaged by
cyclic loading and retrofitted and for the un-retrofitted control beam.
 Table 4 Deflection vs No. of cycles for the beams pre-damaged by cyclic
loading and retrofitted

Beam ID Pre-damage level Total No. of cycles Deflection, mm

to failure
FBA ---- 35594 15.44
RE3 10000 cycles 74026 13.42
RE4 20000 cycles 50800 11.20

From Fig. 11 and Table 4, it can be noted that the pre-damaged retrofitted beams have superior
fatigue performance to the un-retrofitted beam. However, the fatigue life of the retrofitted beams
reduces as the number of cycles used for inducing pre-damage is increased, attesting to a higher
level of pre-damage leading to lower flexural rigidity. As a result of the latter the deflection at
fatigue failure decreases too.
2.3 Common observations from the tested retrofitted beams
All retrofitted RC beams are found to be stiffer and stronger than the respective control beams
(Fig.9, Fig. 11). Even the beams preloaded to (i) 90% of ultimate load of static test and (ii) 20000
cycles under fatigue load, regained their original load carrying capacity. From Fig. 7 and Table
4, it can be observed that the beam which was preloaded to 70% under static loading, withstood
88229 cycles to failure, which is about 2.5 times the average number of cycles to failure for the
case of a beam tested under fatigue loading (control fatigue beam). Similarly, beam which was
preloaded to 80% under static loading, withstood 82152 cycles to failure, which is about 2.3
times the average number of cycles to failure for the case of a beam tested under fatigue loading.
Further, a beam which was preloaded to 90% under static loading, withstood 80698 cycles to
failure, which is about 2.2 times the average number of cycles to failure for the case of a beam
tested under fatigue loading. The interesting observation is that the maximum deflection of
retrofitted beam among all the damage levels is about 13.5 mm against 15.44 mm obtained for
the control fatigue beam. This confirms that the beams preloaded under static loading and tested
under fatigue loading after retrofitting, performance is excellent and very much comparable with
the beam tested under fatigue loading (control fatigue). From Fig. 8 and Table 5, it can be
observed that the beam (RE3) which was preloaded to 10000 fatigue cycles (about 30% of
control fatigue) and after retrofitting, it withstood 74026 fatigue cycles which is about 2 times
the average fatigue cycles of a control beam. Similarly, the beam (RE4) which was preloaded to
20000 fatigue cycles (about 55% of control fatigue) and after retrofitting, it withstood 50800
fatigue cycles which is about 1.4 times the average fatigue cycles of a control beam. The
maximum deflection for the retrofitted beam RE3 is 13.42 mm and for RE4 is 11.20 mm. The
maximum deflection for the control fatigue beam is 15.44 mm. This gives confidence that the
preloaded beams under fatigue exhibited significant performance after retrofitting under fatigue
load without loss of ductility. For all the retrofitted beams, initially, the fatigue load – deflection
behaviour or load – deflection behaviour is linear elastic. After elastic limit micro-cracks and
visible cracks were initiated which then extended downwards into the UHPFRC strip depend on
the pre-damage level. On further increase of fatigue load/static load, many micro-cracks
developed in the UHPFRC but no de-bonding was noticed. The width of micro-cracks in the
UHPFRC strip was very small compared to the width of cracks in the parent concrete beam. The
retrofitted beams continued to carry more load until the steel reinforcement yielded. Later, the
localized macro-crack in the UHPFRC strip crossed the interface with the parent beam and
reached the reinforcement. This major flexural crack then propagated towards the compression
zone in the post peak stage of loading. Final failure occurred by crushing of concrete in the local
loading region. Although multiple flexural cracks formed in the UHPFRC strip because of fiber
bridging only one major macro-crack propagated upwards into the parent beam with the
remaining closing by fiber bridging (Fig. 10). The retrofitted beam was able to maintain
structural integrity even when the major flexural crack opened by 15 to 20mm. All retrofitted
beams failed in a similar mode to respective control beams, i.e. flexure, followed by concrete
crushing in the compression zone. The retrofitted beams behaved as an integral unit with the
UHPFRC strip with no de-lamination. It was observed that the cracks that had formed in the
parent beams during pre-loading were not active at all during the loading of retrofitted beams.
3. Numerical analysis
It is well known that numerical simulation is preferable as it reduces casting and testing time of
control, preloaded and retrofitted beams. Numerical simulation was performed by using finite
element analysis. ABAQUS, general purpose finite element software was used to model and
predict the behaviour of control and retrofitted beams. Modelling was carried out as close as
possible to the actual test conditions. Firstly, the control RC beam was analyzed and the results
were compared with the corresponding experimental observations. Secondly, the preloading
stage was simulated to the required level for RC beams and the simulation continued for the
retrofitted beams from their pre-damaged states. Pre-damage is induced by applying 70%, 80%
and 90% of ultimate load of CBA under static loading and fatigue loading is applied after
retrofitting them with UHPFRC strip in the existing pre-damaged state.
All the experimental material data of normal strength concrete (NSC), UHPFRC and steel
reported earlier were used for modeling. The post-yield stress-strain relationship of reinforcing
steel used was as given in Table 5.The popularly employed concrete damage model available in
ABAQUS was used to represent the nonlinear behavior of NSC and UHPFRC compression and
tension (Table 6). The data of compressive stress and corresponding strain was obtained as per
the procedure proposed by van Mier et al., (1997). Further, the constitutive post-cracking
relationships for NSC and UHPFRC in tension were taken from the earlier works of the authors
[23, 26-27] and are shown in Fig. 12. The salient points of bi-linear tension softening curve are
also given in Table 6. In this manner, the degradation of NSC and UHPFRC under both tension
and compression was fully captured in the FEA model. The main assumption made in the
analysis was that the bond between the pre-damaged beam and the UHPFRC strip was perfect,
which was already confirmed by experimental observations.

Fig 12 Stress Vs Crack Opening Relation in Tension

The concrete beam and UHPFRC were modeled with the brick element C3D8R (Cube Three
Dimensional eight-node Reduced integration) with 3 degrees of freedom at each node to achieve
a uniform stress distribution. The reinforcing steel bars were modeled as three dimensional truss
elements (T3D2) having three translational degrees of freedom at each node. In order to simulate
the experimental conditions as closely as possible, two steel loading plates were modeled at the
location of two loading points. The typical assembly of different parts and FE model is shown in
Fig. 13.
 Fig 13 Typical assembly and FE model
Table 5Plastic Properties of Steel
Stress, MPa Inelastic strain
332 0
352 0.0001
373 0.0003
394 0.001
435 0.002
435 0.003
440 0.005
435 0.01
400 0.03
370 0.06

The RC beams were analysed using direct cyclic analysis under fatigue loading in FE software
(ABAQUS). The frequency of loading followed in the experiment is incorporated as periodic
amplitude calculated as per the fourier series equation as shown below
Amplitude A = A0+Ƹn=1N(Ancosnw(t-to)+Bnsinnw(t-t0))
Where Ao = Initial amplitude, An and Bn are fourier coefficients
An = aocoswt, Bn = aosinwt
The sine curve that is input in the model in terms of periodic amplitude is represented by the Fig.
14.
 Fig 14 Cyclic load data input in ABAQUS

Table 6 Properties of NSC and UHPFRC in Compression and Tension

Compression Behavior of UHSC Tension Behavior of UHSC
Crack
Comp.Stre Inelastic Damage Tensile Damage Displacement,
opening,
ss, MPa Strain parameter stress, MPa parameter mm
mm
107.33 0 0.000 13.5 0 0 0
114.65 0.0032 0.032 5.5 1.235 0.564 1.983
124.43 0.0038 0.123 0 3.786 0.988 3.786
113.32 0.0044 0.221
93.54 0.0054 0.343
64.76 0.0063 0.564
41.32 0.0074 0.724
25.92 0.0083 0.872
14.41 0.0092 0.954
9.76 0.012 0.972

Concrete Damage plasticity properties for NSC

0 0
18.34
0.00016 0.0513
24.65
0.00042 0.0998
31.78
0.00075 0.1792
35.87 2.7 0 0 0
0.00123 0.2643
34.65 1.3 0.134 0.365 0.184
0.00179 0.3125
33.21 0 0.317 0.988 0.317
0.00226 0.3841
29.23
0.0043 0.5845
21.54
0.0076 0.7854
17.32
0.0098 0.8542
Static non-linear analysis was performed on the control RC beam. Fig. 15 compares the predicted
load–deflection behaviour with average experimental data. The two are found to be in good
agreement.

Fig 15 Predicted Load – Deflection Curve of Control RC Beam Compared With Average
Experimental Curve

The finite element analysis (FEA) of beams which had been retrofitted after being subjected to
pre-loading was carried out in two stages. In the first stage, the beams were analyzed to
appropriate pre-loading level, i.e. 70%, 80% or 90% of the ultimate load of CBA. In the second
stage, FEA was continued from the existing pre-loaded state of the beam but with the UHPRFC
strip attached to its tension face under fatigue loading (Fig. 16). Fig. 17presents the response of
control fatigue beam obtained through test and FEA. Both the curves are in fairly good
agreement with each other. Figures 18, 19 and 20 compare the plot of no. of cycles vs deflection
obtained through FEA and experiment for the specimens RB6 and RB9.
Fig 16 Preloading and Strengthening using UHPFRC incorporated in ABAQUS

Fig 17 Deflection vs No. of cycles to failure (comparison of FEA and experiment)

Fig 18 Deflection vs No. of cycles to failure (retrofitted beam – RB6)

 Fig 19 Deflection vs No. of cycles to failure (retrofitted beam – RB7)

Fig 20 Deflection vs No. of cycles to failure (retrofitted beam – RB9)

Table 7 consolidates the salient features of retrofitted beams, Fig. 21 shows the predicted fatigue
response of control and retrofitted beams.
 Fig 21 Deflection vs No. of cycles to failure
(combined FEA plot of retrofitted beams)

Table 7 No. of cyclesvs deflection for retrofitted beams (Static preload and fatigue
strengthened)
Beam ID Preload Total No. of cycles to failure Deflection, mm
(%) Experiment FEA Experiment FEA
FBA ---- 35594 34126 15.44 13.25
RB6 70 88229 85000 13.64 12.56
RB7 80 82152 80000 13.47 12.01
RB9 90 80698 79000 13.61 11.25

From Table 7, it is clear that there is very good agreement between the experimental and
numerically predicted values and the maximum difference between the predicted and the
experimental values is found to be less than 10%. The crack patterns and tensile damage of the
beam as illustrated in Fig. 22 also shows that the maximum damage occurs at the bottom of the
beam under the loading points with the possible cracks extending up to 80% of the beam depth.
The majority of the flexural tensile damage which extended up to 60% of the depth of the beam
is limited to the span between the loading points (i.e. to the zone of the constant bending
moment).
 Fig 22 a) Flexural damage pattern for typical retrofitted beam

Fig 22 b) Deflection contour superposed with FE mesh

Fig 22 c) Tensile damage

Concluding remarks
The following major conclusions can be drawn from this extensive study:
 The predominant failure mode of retrofitted RC beams is similar to respective
conventional RC beams, i.e. flexural failure comprising of yielding of steel followed
by crushing of concrete;
  No de-bonding/delamination of UHPFC strip was observed during testingindicating
the excellent integral behavior of composite NSC-UHPFRCmember;
 Theplot No. of cycles vsdeflection curves of all pre-loaded (70%, 80% and 90%) and
retrofitted RC beams under fatigue loading are found to be stiffer and stronger than the
control beams;
 The inelastic deformation of all pre-loaded (70%, 80% and 90%) and retrofitted RC
beams was comparable to the control beams without a significant loss of ductility;
 The maximum number of cycles to failure of all pre-loaded (70%, 80% and 90%) and
retrofitted RC beams under fatigue is found to significantly higherthan the average
maximum number of cycles of control beams.
 The maximum number of cycles to failure of all preloaded and retrofitted RC beams
under fatigue is considerably high compared to control fatigue beam. The inelastic
deformation of retrofitted beams is well comparable with the control beam;
 The chosen thickness of UHPFRC strip (10 mm) is found to be adequate for the pre-
loaded and retrofitted beams under various loading conditions to regain their original
behavior prior to pre-loading and retrofitting;
 The response of the pre-loaded and retrofitted beams predicted using an integrated
finite element model is found to be in good agreement with the experimental
observations. The simulated crack patterns, tensile damage zones and crushing of
concrete at loading points nearly reproduced the damage evolution patterns observed
during the tests, thus confirming the accuracy and robustnessof the integrated FEA
model.
It can therefore be concluded with confidence that UHPFRC is an excellent candidate for
retrofitting damaged RC flexural elements as it overcomes many of the drawbacks associated
with steel plates and FRP laminates. The retrofitting methodology presented in the paper is
useful for strengthening of structural components such as bridge girders, slabs in parking
garages, airport pavements and machine foundations are subjected to repeated loads during their
service lifetime
 Acknowledgements
This paper forms part of UGC-UKERI collaborative project between CSIR - Structural
Engineering Research Centre and Cardiff University, UK. Authors thank Professor BL
Karihaloo, Cardiff University for his technical suggestions and corrections. Authors also thank
the staff of the Computational Structural Mechanics Group and Structural Testing Laboratory for
the co-operation and suggestions provided during the investigation.

References
1. G. Manfredi and M. Pecce (1997), "Low cycle fatigue of RC beams in NSC and HSC", Eng
Struct, 19, 3, 217-223.
2. Deng Z. C. (2005), "The fracture and fatigue performance in flexure of carbon fiber
reinforced concrete". Cement Concrete Comp, 27, 131–140.
3. Wang W, Wu S. G, Dai H. Z. (2006), "Fatigue behaviour and life prediction of carbon fiber
reinforced concrete under cyclic flexural loading", Mat Sci Eng A, 434, 347–351.
4. Cheng L. (2011), "Flexural fatigue analysis of a CFRP form reinforced concrete bridge
deck", Compos Struct, 93,2895–2902.
5. M. Shahawy and T. E. Beitelman (1999), "Static and Fatigue performance of RC beams
strengthened with CFRP laminates" J Struct Eng, 125, 613-625.
6. Dong J.F., Wang Q.Y. and Guan Z.W. (2012), "Structural behavior of RC beams externally
strengthened with FRP sheets under fatigue and monotonic loading", Eng Struct, 41, 24-33.
7. Ashraf A. Alfeehan. (2014), “Strengthening of RC beams by external steel plate using
mechanical connection technique,”J. Eng and Development, 18, 202-215.
8. Priyanka Sarker, Mahbuba Begum and Sabreena Nasrin. (2011), “Fiber reinforced polymers
for structural retrofitting: A review,” J. Civ. Eng. (IEB), 39, 49-57.
9. Ahmed, O., and Gemert, D. V., ‘‘Behaviour of R.C. beams strengthened in bending by CFRP
laminates,’’ Proc., Structural Faults and Repair—99:8th Int. Conf, London (CD-Rom),
1999.
10. El-Refai J. and West K. Soudki (2012), "Fatigue of reinforced concrete beams strengthened
with externally post-tensioned CFRP tendons", Constr Build Mater, 29, 246–256.
11. Nanni A. (2003), "Concrete repair with externally bonded FRP reinforcement: examples
from Japan" Concr Int, 97, 22–26.
12. Buyukozturk O., and Hearing B. (1998). "Failure behavior of pre-cracked concrete beams
with FRP", J Compos Constr, 2-3, 138–144.
13. Bakis C. B, L. Brown, V. Cosenza, E. Davalos and J. Lesko, (2002),"Fiber-reinforced
polymer composites for construction- state of art review", J Compos Constr, 6, 73–87.
14. A. Aykac, L. Kalkan and A. Uysal (2012), "Strengthening of reinforced concrete beams with
epoxy-bonded perforated steel plates",Struct Eng Mech, 44, 6, 735-751.
15. Attari N. Amaziane S. and Chemrouk, M. (2012), "Flexural strengthening of concrete beams
using CFRP, GFRP and hybrid FRP sheets", Constr Build Mater, 37, 746-57.
16. A. Rahimipour, F. Hejazi, R. Vaghei and M. S. Jaafar (2016), " Finite element development
of a Beam-column connection with CFRP sheets subjected to monotonic and cyclic loading"
Comput Concrete, 18, 6, 1083-1096
17. Alaee F.J. and Karihaloo B.L. (2003), "Fracture model for flexural failure of beams
retrofitted with CARDIFRC", J Eng Mech, 129, 1028-1038.
18. Alaee F.J. and Karihaloo B.L.(2003), "Retrofitting of Reinforced Concrete Beams with
CARDIFRC". J Compos Constr, 7, 174–186.
19. Xu S. L., Wang N. and Zhang X. F., (2012), "Flexural behavior of plain concrete beams
strengthened with ultra-high toughness cementitious composites layer", Mater Struct, 45,
851-859.
20. Shin H.O, Yoon Y.S., and Cook W.D. (2015), "Effect of Confinement on the Axial Load
Response of Ultrahigh-Strength Concrete Columns". J Struct Eng, 141,6, 04014151-12
21. Kevin. Z, Arash. E. Z, and Kay. W (2015), "Rehabilitation of steel bridge girders with
corroded ends using Ultra-High Performance Concrete", Structures Congress, 1411-1422.
22. P. R. Prem, A. Ramachandra Murthy, G. Ramesh, B.H. Bharatkumar and Nagesh R. Iyer
(2015), "Flexural Behaviour of Damaged RC Beams Strengthened with Ultra High
Performance Concrete", Indian Concrete Journal, 2057-2069.
23. Almusallam T.H. Salloum Y.A. (2001), "Ultimate strength prediction for RC beams
externally strengthened by composite materials". Compos Part B - Eng, 32, 609–619.
24. S. D. P. Benson and B. L. Karihaloo (2005), "CARDIFRC – Development and mechanical
properties. Part I: Development and workability", Mag Concrete Res, 57, 6, 347–352.
25. Leung C. K. Y, Cheung Y. N and Zhang J. (2007), "Fatigue enhancement of concrete beam
with ECC layer", Cement Concrete Res, 37, 743–750.
26. A. Ramachandra Murthy, Karihaloo, B. L., Nagesh R. Iyer and Raghu Prasad, B. K. (2013),
"Bilinear tension softening diagrams of concrete mixes corresponding to their size-
independent specific fracture energy", Constr Build Mater, 47, 1160-1166.
27. T.J. Mohammed, B. H. Abu Bakar and N. M. Bunnori (2015), "Strengthening of reinforced
concrete beams subjected to torsion with UHPFC composites", Struct Eng Mech, 56, 1, 123-
136.
28. J.G.M. van Mier, S. P. Shah, M. Arnaud, J.P. Balayssac, A. Bascoul, S. Choi, D.
Dasenbrock, G. Ferrara, C. French, M.E. Gobbi, B.L. Karihaloo, G. König, M.D. Kotsovos,
J. Labuz, D. Lange-Kornbak, G. Markeset, M.N. Pavlovic, G. Simsch, K-C. Thienel, A.
Turatsinze, M. Ulmer, H.J.G.M. van Geel, M.R.A. van Vliet, D. Zissopoulos. (1997)“Strain-
softening of concrete in uniaxial compression”, Mater Struct, 30, 195-209.
29. A. Ramachandra Murthy, Fatigue and fracture behaviour of ultra high strength concrete
beams. PhD thesis, Indian Institute of Science, Bangalore, India, 2011.
30. G. C. Lykidis and K. V. Spiliopoulos (2014), "An efficient numerical simulation of the cyclic
loading experiments on RC structures", Comput Concrete, 13, 3, 343-359.
Highlights of the manuscript

Fatigue behaviour of damaged RC beams Strengthened with Ultra high performance fibre
Reinforced Concrete

 Ultra-high performance fibre-reinforced concrete - a novel and advanced repair and retrofitting
material
 No delamination was observed during testing
 An integrated finite element model was developed accounting for (i) degree of pre-damage, (ii)
fracture behavior of concrete and UHPFRC and (iii) elasto-plastic behavior of the reinforcing
steel