Experimental behaviour of reinforced concrete elements repaired with polymer-modified cementicious mortar

1 23

Materials and Structures ISSN 1359-5997Volume 44Number 2 Mater Struct (2010) 44:517-527DOI 10.1617/s11527-010-9646-0

Experimental behaviour of reinforcedconcrete elements repaired with polymer-modified cementicious mortar

1 23

Your article is protected by copyright and all

rights are held exclusively by RILEM. This e-

offprint is for personal use only and shall not

be self-archived in electronic repositories.

If you wish to self-archive your work, please

use the accepted author’s version for posting

to your own website or your institution’s

repository. You may further deposit the

accepted author’s version on a funder’s

repository at a funder’s request, provided it is

not made publicly available until 12 months

after publication.

ORIGINAL ARTICLE

Experimental behaviour of reinforced concrete elementsrepaired with polymer-modified cementicious mortar

Carlo Pellegrino • Francesca da Porto •

Claudio Modena

Received: 11 October 2008 / Accepted: 7 July 2010 / Published online: 24 July 2010

� RILEM 2010

Abstract This paper presents the results of experi-

mental tests on flexural behaviour of reinforced

concrete beams repaired by polymer-modified mortar.

Tests were repeated varying repair thickness, which

included or did not include the steel reinforcement.

Also the position of repair mortar was varied, as it was

carried out either in the tension or compression region.

Finally, the reinforcement ratios of beam sections were

also varied. Results were compared with those from

control beams, which were tested in non-damaged,

non-repaired conditions. Thick repairs that include

the longitudinal reinforcement can restore the load-

bearing capacity and ductility of non-damaged control

beams, whether they are applied in the tension or in the

compression region. Thin repairs in tension, which

substitute the concrete cover, but do not include the

reinforcement, can even decrease the load bearing

capacity and ductility of the beam, whereas thin repairs

in compression are more effective than those in

tension.

Keywords Reinforced concrete �Repair mortars � Interface � Cracking

1 Introduction

Rehabilitation and strengthening of reinforced con-

crete structural elements is a common task for

existing constructions. Strengthening of a structural

element is aimed at increasing or restoring the load

bearing capacity, due to changes in conditions of use

(e.g. increased loading) or deterioration and damage

of the concrete structure (for example due to envi-

ronmental conditions or seismic events). Several

materials and methods are available for strengthening

reinforced concrete elements, such as adding or

applying mortar; spraying concrete or mortar; inject-

ing or filling cracks, voids and interstices; adding

reinforcing steel bars; installing bonded rebars; post-

tensioning; bonding steel plates or fibre reinforced

polymers (FRP) sheets/plates and others [4]. In

particular, externally bonded FRP sheets/plates have

found increasingly wide applications in civil engi-

neering due to their high strength-to-weight ratio and

high corrosion resistance. A number of experimental

programs and analytical studies have been developed

in the last few years at the University of Padova on

flexural [14, 20], shear [11–13] and bond behavior

[15, 16] of FRP strengthened elements.

Recently, Jumaat et al. [6] made a review of various

repair materials and techniques for reinforced concrete

beams, and Engindeniz et al. [2] dealt with the same

subject in the particular case of non-seismically

designed reinforced concrete beam-column joints.

The formers concluded that the effectiveness of repairs

C. Pellegrino (&) � F. da Porto � C. Modena

Department of Structural and Transportation Engineering,

University of Padova, Via Marzolo 9, 35131 Padova, Italy

e-mail: [email protected]

Materials and Structures (2011) 44:517–527

DOI 10.1617/s11527-010-9646-0

Author's personal copy

depends on the characteristics of repairing materials,

the method of application, the property matching

between old concrete and new material, the preparation

of substrates and use of bonding agents. According to

them, cement based materials, new concrete or mortar,

are more suitable than other materials for concrete

repair [6].

The main aims of cementicious repair of beams in

tension or compression regions under static loadings

are the increase of stiffness to reduce deflections

under serviceability loading conditions, and the

reduction of crack amplitudes to increase beam

durability. However, some issues related to this kind

of technique, such as the choice of the repair material

properties and its position and thickness, with the aim

of improving the compatibility and effectiveness of

the intervention, are still subjects of research.

Emberson and Mays [1] carried out one of the first

experimental studies on the influence of mechanical

and physical properties of repair systems, applied on

either the compression or tension regions of rein-

forced concrete beams subjected to short term static,

long term creep and cyclic loading. They used nine

different repair materials, among which epoxy,

cementicious, and polymer-modified mortars. The

results showed that properties of applied materials

should be as close as possible to substrate concrete.

Modules of elasticity should lay within a range of

±10 kN/mm2, whereas tensile strength of the repair

materials should be higher than substrate concrete.

Hassan et al. [5] tested the compatibility of cementi-

cious, polymer, and polymer modified mortar repairs

to concrete. They also concluded that mismatch in

elastic modules reduces the load carrying capacity of

the combined system. High shrinkage of cementi-

cious repairs also affects the effectiveness of repairs,

whereas polymer-modified mortars, due to reduced

shrinkage and compatibility of elastic modules, are

the most appropriate repair materials. Mangat and

O’Flaherty [8] applied seven different ordinary and

polymer-modified cementicious materials, on two

existing bridges. The case studies showed that repairs

with stiff materials were more efficient than others,

which is in partial disagreement with the results of

other authors. However, differences in elastic mod-

ules were all included in the range given by

Emberson and Mays [1]. Furthermore, the most

performing materials were characterized by the

lowest shrinkage properties.

Rı́o et al. [18] tested beams designed to fail in

flexure, after localised artificial corrosion at midspan

and localised patch-repair with three types of mortar

(cement based, epoxy resin binder, and polymer mod-

ified mortar). They demonstrated that load-bearing

capacity of the repaired elements is generally slightly

lower than the reference, non-damaged beam, but

higher than beams damaged by corrosion. In this

case, repair should be deemed effective. They also

concluded that the mechanical properties of the repair

materials should be close to those of the parent

concrete. Park and Yang [10] tested eight beams

repaired in the tension region with ordinary Portland

and polymer-modified cement mortar. They varied

reinforcement ratio and repair length. Although

ordinary Portland cement repairs gave acceptable

results, the polymer-modified repairs offered the best

results in terms of load bearing capacity, failure mode

and ductility. Shannag and Al-Ateek [19] tested 30

under-reinforced concrete beams, repaired in the

tension region with five materials: ordinary Portland

cement and four types of fiber-reinforced cementi-

cious materials. Once repaired, the beams were tested

as they were or after accelerated corrosion. The

results showed that repairs with ordinary Portland

cement had the poorest performance if compared to

fiber-reinforced materials, both before and after

induced corrosion. This confirmed also the results

of Nounu and Chaudhary [9], who compared ordinary

Portland cement with free flowing micro-concretes,

obtaining better results with the latter. Kim et al. [7]

applied fiber-reinforced cementicious materials at the

intrados of reinforced concrete beams with and

without stirrups. Applying a repair that includes

reinforcement, the beams designed to fail in shear

reached the same strength and centre deflection of

beams with stirrups, but when the repair had double

thickness, the capacity of beams without stirrups was

halved because of interface failure. Beams designed

to fail in flexure could re-establish original beam

capacity with both types of repair thickness.

In this framework, the effect of the geometric

configuration of the repair material (position and

thickness) on the efficiency of the rehabilitation inter-

vention [17], in relation to the existing steel reinforce-

ment configuration (both in tension and in compression

regions), was not comprehensively studied for flexural

elements but only some aspects have been analysed

separately. In this work an experimental investigation

518 Materials and Structures (2011) 44:517–527


to control the effectiveness of polymer-modified

cementicious mortar repairs applied to beams under

flexural loads is carried out. The repair material had

similar mechanical properties and slightly higher

tensile strength than the concrete substrate and was

applied over the entire length of one face of the beams.

The aim is to give some new insights on validating the

effectiveness of such materials in recovering the

properties of non-damaged, non-repaired beams, ver-

ifying the effect of (a) the repair position (tension or

compression region), (b) the percentage of reinforce-

ment in the repaired region, and (c) the repair thickness

(including the steel reinforcement or not), on cracking

pattern and, in general, structural behaviour of flexural

element. Results were compared with those from

control beams, which were tested in non-damaged,

non-repaired conditions, in terms of first cracking and

yielding load, ultimate capacity, cracking pattern and

crack amplitudes, stiffness and ductility.

2 Experimental program

2.1 Design and preparation of specimens

The main objective of the experimental program was

to assess the static behaviour in bending of eight

reinforced concrete beams repaired by polymer-mod-

ified cementicious material. Beams were made with

rectangular sections with area of 150 9 310 mm,

concrete cover of about 20 mm and effective depth of

about 274 mm. The total length of beams was 1.9 m

and their effective span length (distance between

supports) was 1.6 m. All beams were designed to

obtain flexural failure. Four beams (A-type) had longi-

tudinal reinforcement constituted by three 12 mm

diameter reinforcing bars in the tension region and

two 12 mm diameter bars in the compression region,

while the other four beams (B-type) had two 12 mm

diameter reinforcing bars in the tension region and

three 12 mm diameter bars in the compression region.

Shear reinforcement was constituted for all beams by

stirrups with 8 mm diameter reinforcing bars and

200 mm spacing. Two specimens (T0, one for each

type of beams) were used as control beams for the

corresponding beam type and were tested in non-

damaged/non-repaired conditions. The other speci-

mens were cast leaving the reinforcement non-covered

with a curing process of 28 days. All beams repaired at

the tensile side were cast turned upside-down to avoid

problems due to weak concrete areas between the

longitudinal reinforcement and the bottom form.

After this period, the non-covered surface was

prepared and repair material was applied. The prep-

aration of the surface included roughening, cleaning

of dust, powders and any impurities to improve the

adhesion between concrete core and mortar, and

wetting, trying to reproduce, as close as possible, the

typical field situation in laboratory conditions. The

polymer-modified cementicious mortar for repair was

applied after eventual evaporation of water in excess.

Among these six specimens, two beams (T50t, one

for each beam type) were repaired with mortar having

50 mm thickness in the tension region and other two

beams (T50c, one for each beam type) were repaired

with mortar having 50 mm thickness in the compres-

sion region (in this cases repair included steel

reinforcement). Repair mortar having 20 mm thick-

ness was applied in the tension region (T20t_a, with

three reinforcing bars in tension) and mortar having

30 mm thickness was applied in the compression

region (T30c_b, with two reinforcing bars in tension)

on the two remaining beams since the concrete cover

of the longitudinal reinforcement was slightly differ-

ent at the two beam sides. In these two cases repair

substitutes only the concrete cover and, in the latter

case, it partially includes the transversal reinforce-

ment, but not the longitudinal one. Hence the bond

surface has similar characteristics in both cases with

the only difference consisting in the partial inclusion

of the extreme part of the stirrups at the compressive

side. Figure 1 shows the details and Table 1 lists the

data of the tested beams.

2.2 Materials

The main mechanical properties of concrete were

experimentally evaluated after 28 days curing. Mean

cubic compressive strength, measured on four sam-

ples cast during specimen construction and with

dimension 150 9 150 9 150 mm, was 34.8 N/mm2.

Cylinder compressive strength, measured on samples

with diameter of 150 mm and height of 320 mm, was

33.6 N/mm2. Mean tensile strength, measured with

splitting tests on three cylindrical samples of the

same dimensions, was 3.2 N/mm2. Elastic modulus

was not measured, but according to the measured

cylinder compressive strength and Eurocode 2

Materials and Structures (2011) 44:517–527 519


formulation [3], it was assumed to be around

32,500 N/mm2.

Ribbed bars used for longitudinal reinforcement

(diameter 12 mm) and for transversal reinforcement

(diameter 8 mm) were both tested in tension.

Mechanical properties were similar for the two types

of bars, with mean yield stress of 532 N/mm2 and

mean tensile strength of 628 N/mm2.

Finally, the cementicious material, used for repair-

ing all beams, was premixed, tixotropic, polymer-

modified mortar with high-strength hydraulic binders

and aggregates with maximum thickness of 4 mm.

This product has high bond properties, low CO2 and

vapour permeability, limited shrinkage. It is generally

used for cover repair in reinforced concrete structures.

Mechanical properties of the repair mortar were

measured on samples with dimensions of 40 9

40 9 160 mm, cast during the repair interventions.

These samples were tested after 28 days curing.

Density of hardened mortar was 2,170 kg/m3. Mean

tensile strength deducted from flexural tests was equal

to 3.5 N/mm2. Mean cubic compressive strength was

39.6 N/mm2, mean elastic modulus was 26240 N/mm2.

Table 2 compares the mechanical properties of the

concrete support and the repair material. It can be seen

that the measured mechanical properties differ for no

more than 10%, and concrete and mortar elastic

modules differ for less than 10 kN/mm2.

2.3 Testing procedures

Flexural tests on beams were carried out under

typical four-point testing configuration (Fig. 1). The

2 φ 12

3 φ 12

T0_a

270 1600 150

P

150

665 270 665

310

150

310

2 φ 12

3 φ 12

T0_b

2 φ 12

3 φ 12

T50t_b

2 φ 12

3 φ 12

T50t_a

50

50

2 φ 12

3 φ 12

T20t_a

20

2 φ 12

3 φ 12

T30c_b30

2 φ 12

3 φ 12

T50c_a

2 φ 12

3 φ 12

T50c_b

50

50

Fig. 1 Dimensions, rebars and repairs arrangement of beams

Table 1 Details of specimens

Type of

element/test

Section

(mm2)

Longitudinal reinforcement ql (%) Transversal

reinforcement

qw (%) Condition Designation

Tension Compression

Beam flexural 150 9 310 3U12 2U12 0.73 1U8/200 mm 0.33 Control beam T0_a

Repair 50 mm ten. T50t_a

Repair 50 mm co. T50c_a

Repair 20 mm ten. T20t_a

2U12 3U12 0.49 1U8/200 mm 0.33 Control beam T0_b

Repair 50 mm ten. T50t_b

Repair 50 mm co. T50c_b

Repair 30 mm co. T30c_b



shear span was 665 mm (ratio between shear span

and effective depth of the cross section equal to about

2.5). Tests were carried out monotonically, with loads

increased between 0.1 and 0.5 kN/s. Load cell with

300 kN capacity was used to measure applied loads.

Control beams were instrumented with five strain

transducers (DD1; 100 mm measuring base), close to

beam midspan. Three DD1 were placed sequentially

on the lateral face of the beam, in the tension region,

along the longitudinal reinforcement, to measure

opening and propagation of cracks. One DD1 was

placed on the beam intrados, in the tension region;

the other was placed on the beam extrados, in the

compression region. Three linear variable differential

transducers (LVDT) were placed at the beam extra-

dos at supports and at the beam intrados at midspan to

measure displacements and deflections. For the

repaired beams, other three strain transducers (DD1)

were placed on the lateral face of the beam along the

mortar repair layer, and across the repair layer–

concrete beam interface, to gather information on the

behaviour of the strengthening material, and the

interface. Figure 2 shows a detail of the transducers

used to measure crack amplitudes on the concrete

element and on the mortar repair layer in the tension

region near midspan.

3 Test results

3.1 Failure modes and ultimate loads

All beams presented flexural failure with yielding of

reinforcement in the tension region and ductile

behaviour up to failure. Vertical and sub-vertical

cracks developed uniformly and symmetrically on the

constant bending moment area, and crossed the beam

intrados in the tension region to connect each other

on the two lateral faces of the beams. Differences in

failure modes were determined by thickness and

position of the repair. Figure 3 shows the crack

patterns of the tested beams.

In the control beams T0_a (3U12 mm in tension)

and T0_b (2U12 mm in tension) first cracks occurred

at about 25 kN. Subsequently, vertical cracks devel-

oped on the constant bending moment area. Larger

cracks were exactly located at midspan (critical crack)

and under the loading points. Debonding of cover or

concrete portions did not occur. Ultimate load for

T0_a was 161 kN, corresponding to bending moment

of 53.4 kNm. Ultimate load for T0_b was 115 kN,

corresponding to bending moment of 38.2 kNm.

Beams T50t_a (3U12 mm in tension) and T50t_b

(2U12 mm in tension) were both repaired with

50 mm of mortar in the tension region. First cracks

occurred at about 30 kN. In T50t_a, larger cracks

were exactly located at midspan (critical crack) and

under the loading points. These main cracks were

alternate with smaller cracks. Debonding of repair did

not occur, resulting in behaviour very similar to that

of the control beams. Ultimate load was 162 kN,

corresponding to bending moment of 53.7 kNm. In

T50t_b, three cracks developed at midspan and under

the loading points. Two of them crossed the beam

intrados to connect each other on the two lateral faces

of the beam, whereas the third (on the right, critical

crack), reached the repair-concrete beam interface,

and developed along the interface in the horizontal

direction. This crack split the mortar repair layer far

away from the loading points, and close to the

support. Figure 4a shows the critical crack at failure

for T50t_b. However, the repair did not debond from

the beam. Ultimate load was 112 kN, corresponding

to bending moment of 37.4 kNm.

Table 2 Mechanical properties of concrete and repair mortar

Property Concrete Mortar

Density of hardened material (kg/m3) 2,380 2,168

Mean compressive strength (N/mm2) 34.8 39.6

Mean elastic modulus (kN/mm2) 32.5a 26.2

Mean tensile strength (N/mm2) 3.19 3.48

a Evaluated on the basis of EN 1992-1-1

Fig. 2 Transducers to measure crack amplitudes on concrete

element and on mortar repair layer in the tension region near

midspan



Beams T50c_a (3U12 mm in tension) and T50c_b

(2U12 mm in tension) were both repaired with

50 mm of mortar in the compression region. First

cracks occurred at about 30 kN for T50c_a and

25 kN for T50c_b. More than in other cases, cracks

were developed also outside the constant bending

moment area, defined by the two loading points. Two

critical cracks for T50c_a and one critical crack for

T50c_b were located at midspan. In both cases,

various cracks raised until reaching the repair–

concrete beam interface, in the compression region.

However, debonding of repair did not occur. Ultimate

load for T50c_a was 164 kN, corresponding to

bending moment of 54.5 kNm. Ultimate load for

T50c_b was 117 kN, corresponding to bending

moment of 38.9 kNm.

Beam T20t_a (3U12 mm in tension) was repaired

with 20 mm of mortar in the tension region. First

cracks occurred at about 30 kN. The crack at midspan

crossed the beam intrados in the tension region to

connect with the corresponding crack on the other

lateral face of the beam, whereas the two lateral

cracks (developing under the loading points, critical

cracks), reached the repair–concrete beam interface,

and followed the horizontal interface direction

(Fig. 4b). For loads higher than 145 kN, debonding

of repair occurred, then load rose up to 153 kN

(ultimate load), corresponding to bending moment of

50.9 kNm (95% of control beam).

T0_a

T0_b

T20t_a

T50t_b

T50t_a

T30c_b

T50c_a

T50c_b

Fig. 3 Crack pattern of beams at failure

Fig. 4 Critical crack at failure for T50t_b (a) and T20t_a (b)



Finally, beam T30c_a (2U12 mm in tension) was

repaired with 30 mm of mortar in the compression

region. First cracks occurred at about 25 kN. Critical

crack was located at midspan, but cracks were

developing also outside the constant bending moment

area. For loads higher than 110 kN, debonding of

repair occurred, with horizontal cracks propagating

along the repair-concrete beam interface at the beam

ends (see Fig. 5). Load then rose up to 112 kN

(ultimate load, 97% of control beam), corresponding

to bending moment of 37.3 kNm.

3.2 Loads and deflections

Figure 6 shows the load deflection curves of A-type

(above) and B-type (below) beams. Figure 7 shows

the load strain curves of beams repaired in tension

(T50t_a) and in compression (T50t_c) with thick

repair layer, and beams repaired in tension (T20t_a)

and in compression (T30t_c) with thin repair layer.

The diagrams show strains measured over the beam

extrados, under the beam intrados, and on the beam

lateral sides close to the intrados. In addition, for

beams repaired in tension, the diagrams give strains

measured on the lateral side of the repair. In these

diagrams, extrados strains (compression) are plotted

positive as side and intrados strains (tension).

In A-type beams (3U12 mm in tension), the

control beam (T0_a) and the beam repaired in

compression (T50c_a) had the same elastic stiffness

before cracking, two times higher than initial stiffness

of beams repaired in tension region (T50t_a and

T20t_a). In the control beam, cracking occurred at

23.4 kN, that is 78% of the evaluated theoretical

value (30 kN). In the repaired beams, cracking

occurred around 33 kN, that is 10% higher than the

theoretical value and 41% higher than the actual

value of the control beam (see Table 3). This is due

to the higher strength of the repair mortar (see

Table 2).

According to the initial stiffness, the midspan

deflection at cracking of beams repaired in tension

(T50t_a and T20t_a, 0.85 mm on average) was about

two times higher than that of the beam repaired in

compression (T50c_a, 0.38 mm; see Table 4). This

value was lower in the control beam, as cracking

occurred earlier than in repaired beams (T0_a,

0.21 mm). After cracking, differences in stiffness

decreased.

Yielding was evaluated to occur after completion

of the cracking phase, on the basis of strains measured

by instruments placed on the beam intrados and lateral

faces. Reinforcement yielding is also shown by the

bent of load deflection curves in Fig. 6. Yield loads in

the control beam (T0_a, 144 kN) and the beamFig. 5 Debonding of mortar repair layer for T30c_a

Load-deflection, A-type beams

0

30

60

90

120

150

180

Deflection [mm]

Lo

ad [

kN]

T0_aT50t_aT50c_aT20t_a

Load-deflection, B-type beams

0

20

40

60

80

100

120

0 3 6 9 12 15 18

0 2 4 6 8 10 12

Deflection [mm]

Lo

ad [

kN]

T0_bT50t_bT50c_bT30c_b

Fig. 6 Load deflection curves of A-type and B-type beams



repaired in compression (T50c_a, 153 kN) were

higher than the theoretical value evaluated for the

control beam (138 kN) and higher than the experi-

mental values measured on the other beams (T50t_a

and T20t_a, 139 kN on average; see Table 3). The

corresponding values of midspan deflection were

similar in all beams (on average, 7.28 mm) except for

the beam repaired in tension by substituting only the

cover (T20t_a). In this case, deflection at yielding was

8.95 mm.

Ultimate load was measured in all beams, but in

two beams (T50c_a and T0_b), LVDTs measuring

midspan deflection were not working until the end of

the test. All A-type beams practically reached or

Load-strain, T50t_a

0

30

60

90

120

150

180

Strain [10-3]

Lo

ad [

kN]

extrados strain

beam strainrepair strain

intrados strain

Load-strain, T50c_a

0

30

60

90

120

150

180

Strain [10-3]

Lo

ad [

kN]

extrados strain

beam strain

intrados strain

Load-strain, T20t_a

0

30

60

90

120

150

180

0 2 4 6 8 10

0 2 4 6 8 10

Strain [10-3]

Lo

ad [

kN]

extrados strainbeam strainrepair strainintrados strain

Load-strain, T30c_b

0

30

60

90

120

0 2 4 6 8 10

0 2 4 6 8 10

Strain [10-3]

Loa

d [k

N]

extrados strain

beam strain

intrados strain

Fig. 7 Load strain curves of beams repaired in tension (T50t_a; T20t_a), and in compression (T50c_a, T30c_b)

Table 3 Results of flexural tests

Beam Cracking load (kN) (a)/(b) Yield load (kN) (a)/(b) Ultimate load (kN) (a)/(b)

Theory (a) Test (b) Theory (a) Test (b) Theory (a) Test (b)

T0_a 30.0 23.4 1.28 138 144 0.96 162 161 1.01

T50t_a – 32.9 0.91 – 140 0.99 – 162 1.00

T50c_a – 32.4 0.93 – 153 0.90 – 164 0.99

T20t_a – 33.8 0.89 – 138 1.00 – 153 1.06

T0_b 29.4 27.4 1.07 93 93 1.00 109 115 0.95

T50t_b – 31.1 0.95 – 93 1.00 – 112 0.97

T50c_b – 24.7 1.19 – 96 0.97 – 115 0.95

T30c_b – 25.4 1.16 – 93 1.00 – 112 0.97



overcame the theoretical value of the ultimate load

evaluated for the control beam (162 kN), except for

the beam repaired in tension by substituting only the

cover (T20t_a, ultimate load of 153 kN, that is 94%

of the theoretical value). In the case of thick repairs,

performance of repaired beams (T50t_a, 162 kN;

T50c_a, 164 kN) were even better than that of the

control beam (T0_a, 161 kN; see Table 3).

At ultimate load, deflections reached about 15 mm

in the control beam (T0_a, 15.09 mm), and the beam

repaired in tension (T50t_a, 14.87 mm). Even if it

was not possible to have this measure for the beam

repaired in compression (T50c_a), it is very likely

that its midspan deflection was in the same range, as

the experimentally observed behaviour was very

similar. Instead, the beam repaired in tension by

substituting only the cover (T20t_a) showed lower

midspan deflection at ultimate load (11.49 mm, see

Table 4). Therefore, if beam behaviour is evaluated

by means of ductility index, that is the ratio of

deflection at ultimate load to yielding, the beam

repaired with thick mortar layer in tension (T50t_a)

had ductility index of 2.03, very similar to the

ductility index of the control beam (T0_a, 2.00). The

beam repaired in compression also had similar

behaviour. Instead, the beam repaired in tension

by substituting only the cover (T20t_a) showed

worse behaviour, with ductility index of only 1.28

(see Table 4).

In B-type beams (2U12 mm in tension), all beams

had similar elastic stiffness before cracking. In the

control beam (T0_b), cracking occurred at 27.4 kN,

that is 93% of the evaluated theoretical value

(29.4 kN). Similar values were reached by beams

repaired in compression (T50c_b, 24.7 kN, and

T30c_b, 25.4 kN). In the beam repaired in tension

(T50t_b), cracking occurred at 31.1 kN, that is 6%

higher than the theoretical value and 14% higher than

actual value of control beam (see Table 3). This

increase is due to the higher tensile strength of the

repair mortar. According to the initial stiffness,

midspan deflection at cracking in all beams was in

the range between 0.36 and 0.40 mm. After cracking,

the two beams repaired in compression (T50c_b and

T30c_b) presented the same stiffness as the control

beam (T0_b), whereas the beam repaired in tension

(T50t_b) had the lowest stiffness and the highest

deflections.

Yield loads, in all beams, almost reached the

theoretical value evaluated for the control beam

(93 kN) except for the beam repaired in compression

(T50c_b), in which yield load was slightly higher

(96 kN, see Table 3). The corresponding values

of midspan deflection were similar in all beams

(on average, 4.5 mm). The beam repaired in com-

pression by substituting only the cover (T30c_b) had

the lowest deflection, 4.32 mm; the beam repaired in

tension (T50t_b) had the highest deflection, 4.81 mm

(see Table 4).

All type-B beams overcame the theoretical value

of ultimate load evaluated for the control beam

(109 kN). The control beam and the beam repaired in

compression (T0_b and T50c_b) reached 115 kN

(that is 105% of the theoretical value), whereas the

beam repaired in tension and that repaired in

compression by substituting only the cover (T50t_b

and T30c_b) reached 112 kN (103% of the theoret-

ical value; see Table 3).

Table 4 Deflections and ductility index of flexural tests

Beam Cracking load Yield load Ultimate load Ductility

index (-)Load (kN) Defl. (mm) Load (kN) Defl. (mm) Load (kN) Defl. (mm)

T0_a 23.4 0.21 144 7.54 161 15.09 2.00

T50t_a 32.9 0.81 140 7.34 162 14.87 2.03

T50c_a 32.4 0.38 153 6.96 164 (8.95)a –

T20t_a 33.8 0.90 138 8.95 153 11.49 1.28

T0_b 27.4 0.37 93 4.45 115 (7.73)a –

T50t_b 31.1 0.36 93 4.81 112 10.13 2.11

T50c_b 24.7 0.39 96 4.67 115 10.94 2.34

T30c_b 25.4 0.40 93 4.32 112 10.06 2.33

a Last value measured before LVDT removal



At ultimate load, deflections were all around

10 mm. The beam repaired in compression with thick

mortar layer (T50c_b) gave the highest value

(10.94 mm); the beam repaired in compression with

thin mortar layer (T30c_b) gave the lowest value

(10.06 mm), followed by the beam repaired in tension

(T50t_b, 10.13 mm). Even if it was not possible to

have this measure in the control beam (T0_b), it is

very likely that its midspan deflection was in the same

range, as the experimentally observed behaviour was

very similar. Taking into account the ductility index, it

can be seen that the two beams repaired in compres-

sion (T50c_b and T30c_b) had almost the same value

(2.33), whereas the beam repaired in tension (T50t_b)

had slightly lower ductility (2.11, see Table 4), due to

the lower stiffness shown in cracked phase and the

higher deflection at yielding.

Finally, strain measurements confirmed that beam

substrate and repair layer, in beams repaired with

thick mortar layer in tension (see T50t_a in Fig. 7),

were working together, i.e. collaboration between

the two materials occurred with negligible sliding.

Figure 7 shows that strains measured on close posi-

tions on the lateral faces of beam and repair, and at

beam intrados (thus on repair), had similar trends and

were consistent with position of instruments and

distance from neutral axis. The same can be also said

for beams repaired with mortar in compression (see

T50c_a and T30c_b in Fig. 7). The strains measured

on the beam extrados (thus on repair), were consistent

with the beam behaviour and were very similar to

those measured on the beams non-repaired at the

extrados (compare, for example, T50c_a and T50t_a).

Instead, in the beams repaired with thin mortar layer

in tension (see T20t_a in Fig. 7), the strains measured

on the repair were almost zeroed.

4 Discussion

Mortar repair in the tension region of beams could

increase the value of loads that induce the first crack

from about 14 to 44%, although the tensile strength of

the repair mortar was only 10% higher than the tensile

strength of concrete. The position of the repair had

greater influence on the beam behaviour in the case of

beams with higher percentage of reinforcement in

tension. In any case, the repairs in compression

increased the beam stiffness or re-established the same

stiffness of the non-damaged control beams, whereas

the repairs in tension always decreased the stiffness.

In the case of thick repairs (50 mm) including

longitudinal reinforcement, applied on the compres-

sion or tension region of beams, experimental results

showed that the repair technique and material could

re-establish the load bearing capacity of the non-

damaged control beams, evaluated by experimental

testing and theoretical calculation. Even though some

cracks opened along the repair–concrete beam inter-

face, sliding of the two materials did not occur and

the original reinforced concrete beam and the repair

material kept working together. The crack patterns

and the strain measurements demonstrated this fact.

In the case of thin repairs, substituting only the

concrete cover, debonding of the repair occurred and

caused the beam failure. For thin repair (20 mm) in

tension, collapse was anticipated at 94% of the

control beam experimental and theoretical ultimate

loads, for thin repair (30 mm) in compression, this

effect was less marked. In the latter case, ultimate

load was 97% that of the control beam and 103% of

the theoretical value. In addition, strain analysis of

thin repair in tension showed that the repair was not

collaborating to the load bearing capacity, whereas

thin repair in compression was loaded.

Analysis of the ductility index confirmed these

results. Even though displacements at ultimate load

were not measured for two beams, it was possible

to conclude that thick repair, in general, could

re-establish the available ductility of the non-damaged

control beams, that is, to ensure the same beam

deflection capacity after repair. Thin repair in com-

pression had the same effect, whereas thin repair in

tension decreased the available ductility to only 64%

that of the control beam.

5 Conclusions

In this work an experimental investigation to control

the effectiveness of polymer-modified cementicious

mortar repairs applied to beams under flexural loads

is carried out. The aim is to give some new insights

on validating the effectiveness of such materials

in recovering the properties of non-damaged, non-

repaired beams, verifying the effect of (a) the repair

position (tension or compression region), (b) the

percentage of reinforcement in the repaired region,



and (c) the repair thickness (including the steel

reinforcement or not), on cracking pattern and, in

general, structural behaviour of flexural element.

The main conclusions arising from the experimental

tests show that polymer-modified cementicious mor-

tars, with mechanical properties like those obtained in

the present experimental investigation and similar to

those of the concrete substrate (compressive strengths

differing less than 10%, and elastic modules differing

less than 10 kN/mm2), can be effective for the repair of

beams. The effectiveness of the intervention depends

also on the position and thickness of the repair layer in

relation to the position of the steel reinforcement.

In repaired beams, first cracking occurs later than

in control beams. Thick repairs that include the lon-

gitudinal reinforcement can restore the load-bearing

capacity and ductility of non-damaged control beams,

whether they are applied in the tension or in the

compression region. Thin repairs in tension, which

substitute the concrete cover, but do not include the

reinforcement, can even decrease the load bearing

capacity and ductility of the beam. Thin repairs in

compression are more effective than those in tension.

It must be observed that this study uses a polymer-

modified repair material for the rehabilitation of a non

damaged member built in laboratory conditions. The

specimens have been built in conditions as close as

possible to field members which are typically repaired

by chipping away bad concrete, sandblasting rebar, and

then filling the void with repair material. Therefore, even

if the interface of the specimens has been roughened

similarly to field conditions, bond, stress distribution,

and fracture of the repair material could be slightly

different in field situations with respect to that obtained

for the specimens of this study in laboratory conditions.

Acknowledgements The authors gratefully acknowledge

Tassullo S.p.A. that provided materials and specimens for

experimental testing, and Eng. Fabio Mazzucco for his

contribution to the experimental investigation developed

during his MSc thesis. The experimental tests were carried out

at the Laboratory of Structural Materials Testing of the

University of Padova, Italy.

References

1. Emberson NK, Mays GC (1996) Significance of property

mismatch in the patch repair of structural concrete. Part 3:

reinforced concrete members in flexure. Mag Concr Res

48(174):45–57

2. Engindeniz M, Kahn LF, Zureick AH (2005) Repair and

strengthening of reinforced concrete beam-column joints:

state of the art. ACI Struct J 102:187–197

3. European Committee for Standardization (2004) Eurocode

2—design of concrete structures. Part 1–1: General rules

and rules for buildings. EN 1992-1-1, Brussels, Belgium

4. European Committee for Standardization (2005) Products

and systems for the protection and repair of concrete

structures—definitions, requirements, quality control and

evaluation of conformity. Part 3: structural and non-

structural repair. EN 1504-3, Brussels, Belgium

5. Hassan KE, Brooks JJ, Al-Alawi L (2001) Compatibility of

repair mortars with concrete in a hot-dry environment.

Cem Concr Compos 23:93–101

6. Jumaat MZ, Kabir MH, Obaydullah M (2006) A review of

the repair of reinforced concrete beams. J Appl Sci Res

2(6):317–326

7. Kim JHJ, Lim YM, Won JP, Park HG, Lee KM (2007)

Shear capacity and failure behaviour of DFRCC repaired

RC beams at tensile region. Eng Struct 29:121–131. doi:

10.1016/j.engstruct.2006.04.023

8. Mangat PS, O’Flaherty FJ (2000) Influence of elastic

modulus on stress redistribution and cracking in repair

patches. Cem Concr Res 30:125–136

9. Nounu G, Chaudhary Z-UL-H (1999) Reinforced concrete

repair in beams. Constr Build Mater 13:195–212

10. Park SK, Yang DS (2005) Flexural behaviour of reinforced

concrete beams with cementitious repair materials. Mater

Struct 38:329–334

11. Pellegrino C, Modena C (2002) FRP shear strengthening of

RC beams with transverse steel reinforcement. J Compos

Constr 6(2):104–111

12. Pellegrino C, Modena C (2006) FRP shear strengthening of

RC beams: experimental study and analytical modelling.

ACI Struct J 103(5):720–728

13. Pellegrino C, Modena C (2008) An experimentally based

analytical model for shear capacity of FRP strengthened rein-

forced concrete beams. Mech Compos Mater 44(3):231–244

14. Pellegrino C, Modena C (2009) Flexural strengthening of

real-scale RC and PRC beams with end-anchored pre-

tensioned FRP laminates. ACI Struct J 106(3):319–328

15. Pellegrino C, Modena C (2009) Influence of FRP axial

rigidity on FRP-concrete bond behaviour: an analytical

study. Adv Struct Eng 12(5):639–649

16. Pellegrino C, Tinazzi D, Modena C (2008) An experi-

mental study on bond behavior between concrete and FRP

reinforcement. J Compos Constr 12(2):180–189

17. Pellegrino C, da Porto F, Modena C (2009) Rehabilitation

of reinforced concrete axially loaded elements with poly-

mer-modified cementicious mortar. Constr Build Mater

23(10):3129–3137

18. Rı́o O, Andrade C, Izquierdo D, Alonso C (2005) Behaviour

of patch-repaired concrete structural elements under increas-

ing static loads to flexural failure. J Mater Civ Eng 17(2):

168–177. doi:10.1061/(ASCE)0899-1561(2005)17:2(168)

19. Shannag MJ, Al-Ateek SA (2006) Flexural behaviour of

strengthened concrete beams with corroding reinforce-

ment. Constr Build Mater 20:834–840

20. Valluzzi MR, Grinzato E, Pellegrino C, Modena C (2008)

IR thermography for interface analysis of FRP laminates

externally bonded to RC beams. Mater Struct 42(1):25–34



http://dx.doi.org/10.1016/j.engstruct.2006.04.023

http://dx.doi.org/10.1061/(ASCE)0899-1561(2005)17:2(168)

Experimental behaviour of reinforced concrete elements repaired with polymer-modified cementicious mortar

Documents