Construction and Building Materials 75 (2015) 112–120

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier .com/locate /conbui ldmat

Flexural performance of a precast concrete slab with steel fiber concretetopping

http://dx.doi.org/10.1016/j.conbuildmat.2014.09.1120950-0618/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author. Tel.: +60 75531602.E-mail address: suhaimibakar@utm.my (S. Abu Bakar).

Farnoud Rahimi Mansour, Suhaimi Abu Bakar ⇑, Izni Syahrizal Ibrahim, Abdul Kadir Marsono,Bahram MarabiDepartment of Structure and Materials, Faculty of Civil Engineering, Universiti Teknologi Malaysia, 81310 UTM-Skudai, Johor, Malaysia

h i g h l i g h t s

� Increasing the flexural capacity of concrete slab by applying a steel fiber topping.� Try to overcome certain problems by using steel fiber in concrete.� The mechanism relies on a good bonding between the two layers.� To provide a better results, roughness for interface has used.� Flexural performance depends mainly on the type of roughness.

a r t i c l e i n f o

Article history:Received 25 February 2014Received in revised form 30 July 2014Accepted 27 September 2014

Keywords:Cement-base concrete overlaySteel fiber reinforced concreteFlexural capacityComposite action

a b s t r a c t

The positive effects of various types of fibers on concrete ductility and other engineering properties, suchas the tensile, flexural, fatigue, and load-bearing capacity after cracking and toughness, are well known.Steel fiber-reinforced concrete (SFRC) has been used increasingly in recent years and has been applied tovarious structural components. Considerable interest has been developed in using steel fibers in concreteto increase the load-carrying capacity of the structural members in service. It has been used recently toincrease the flexural capacity of concrete slabs by applying a thin layer of SFRC onto an existing slab, atechnique known as cement-base bonded overlay. The objective of this research is the investigation ofthe flexural behavior of a precast concrete slab with a steel fiber concrete topping. To reinforce theconcrete overlay, hooked-end steel fibers with a length of 30 mm and a diameter of 0.75 mm were used.Because the performance of this composite slab depends on the bonding between the old and newconcrete, different types of roughness at the interface has used to provide good bonding between thetwo layers. Based on experimental tests, the flexural performance was shown to depend not only on add-ing the steel fibers to the topping but also on the type of interface roughness. To examine the compositebehavior of the specimens, the interface slip was also measured throughout the test. The results showed agood reliability of roughness in providing bonding strength at the interface. It was also found that rough-ness in the transverse direction provides the best bonding strength at the interface. Although the resultsshowed interface slip at mid-span, slip was not detected at either end of the specimen.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

There are a number of situations where it may become neces-sary to increase the structural performance of concrete members.In recent years, the cement-base bonded overlay technique hasbeen used to strengthen and enhance the structural performanceof precast concrete slabs by adding a thin layer of cast in-situreinforced concrete to the existing slab. The primary purpose is

to improve the load-carrying capacity and stiffness of the slab byincreasing its thickness. The efficiency of this strengtheningtechnique strongly depends on the bonding between the addedlayers with the substrate. Debonding becomes possible becauseof cracking of the overlay and because of the tendency of theoverlay not to follow the curvature of the substrate.

It is generally accepted that the main cause of overlay debond-ing is the cracking of the overlay due to one of the causes, or thesuperposition of the two causes, sketched in Fig. 1, designated as‘‘mechanical origin’’ and ‘‘differential length change.’’ Debondingof mechanical origin is shown in Fig. 1a. It is the consequence of

F. Rahimi Mansour et al. / Construction and Building Materials 75 (2015) 112–120 113

the flexural straining of the structure by the applied load. Debond-ing of differential length change is illustrated in Fig. 1b. It resultsfrom the effects of the different length changes of the overlayand base, especially because of shrinkage. In both cases, debondingresults from a full depth separation of the overlay (e.g., a crack),preventing continuous stress transmission within the overlay. Atthe crack location, sharp built-in peak stresses develop at the inter-face, which subsequently cause debonding.

Reinforcement of the overlay enables force to be transmittedthrough the crack and decreases the intensity of the mechanicaldiscontinuity at the crack location. Consequently, the built-inpeak stresses at the interface decrease, delaying overlay debond-ing. Steel reinforcing in the form of tied bars and welded mats isnormally used for overlay reinforcement. However, using steelbars or welded fabric behind thin covers lead to wide cracksand can only be used in relatively thick overlays [1]. Substitutionof fibers in concrete is expected to overcome these problems anddelay overlay debonding. Fibers are not affected by such con-straints and, as a bonus, they present a reduced risk of corrosion[1]. Fibers also reduce the amount of steel bars required andaccelerate the construction time because they are added directlyto the concrete batch [2]. Although all types of reinforcementact in similar ways, the literature shows that fiber reinforcementhas a higher effectiveness. The amount of fibers required toachieve the desired result will depend on the type of fiber andits bond characteristics.

It has been recognized that the behavior of concrete remarkablyimproves by reinforcing it with discontinuous fibers [3]. Studieshave shown that fibers transfer loads to the uncracked parts ofthe concrete, prevent crack propagation, reduce brittleness and sig-nificantly improve the concrete’s tensile strength, compressivestrength, toughness, energy absorption capacity and durability[3,4]. Compressive tests showed that fibers in concrete had onlya small effect on the compressive strength of concrete [5,6]. A mainadvantage of fiber reinforced concrete (FRC) is its high energyabsorption capacity and high toughness due to its higher ductility[7]. According to a statement by Zollo, the fibers have more effecton the energy absorption capacity and crack control comparedwith the ultimate flexural strength [8]. This has also been con-firmed by Khaloo and Afshari based on their investigation of theflexural strength, deflection and energy absorption of small con-crete slabs [9].

Among other fibers, steel fibers (SF) have become popular inrecent years because of their structural properties [3]. In 1963,Romualdi and Batson reported significant improvement in con-crete properties through the use of a randomly distributed SF.Improved durability has been demonstrated by the previous

Fig. 1. (a) Debonding of mechanical origin and (b

research of cement-based bonded overlays with added SF. Today,hooked-end steel fiber (HSF) is mostly used due to its significantlyimproved reinforcement effects compared to straight SF [10]. Astrong connection between the hydrated cement matrix and theHSF has been demonstrated based on experiments performed byAckermann and the microscopy images of steel fiber reinforcedconcrete (SFRC) by Uygunoglu [11].

The performance of the composite slab depends on the bondingbehavior between the old and new concrete. The most reliable andeconomical solution to increase the bonding strength of the inter-face is the roughening the surface of the substrate. Although sev-eral studies have been carried out, there is still a lack ofinformation on the effect of various types of roughness on theoverall performance of the composite slab. The aim of this researchis to evaluate the flexural capacity of precast slabs with a SFconcrete topping related to different roughnesses of the substratesurface. To examine the composite behavior of the specimens,interface slip was also measured throughout the experiments.

2. Previous work

2.1. Mechanical properties of SFRC

The mechanical properties of SFRC mainly depend on the SFgeometry, aspect ratio, volume fraction and concrete strength. Ithas been observed that concrete mixtures with waved SF exhibithigher compressive strengths than concrete mixtures with HSF[6]. Additionally, it has been shown that specimens with HSF exhi-bit higher toughness than specimens with waved SF [6]. Uygunoglushowed that flexural strength is directly related to SF volume frac-tion, concrete age and aspect ratio, while crack development inbending is inversely related to these parameters [11]. A study byKhaloo and Afshari on small SF concrete slabs showed that theenergy absorption increased as the SF volume, aspect ratio andconcrete strength increased [9]. However, a reduction in theincreasing rate of energy absorption was also shown when the con-tent of SF was increased. The ductility of SFRC also improves withincreasing SF content in volume; ductility was also shown toimprove at the same SF content when the SF length was increased[5]. These phenomena are caused by the higher deformability andenergy absorption of SFRC during the cracking phase and the factthat post-cracking behavior is affected by different SF lengths [5].Although the higher SF aspect ratio may result in lower compres-sive strengths and elastic moduli, the toughness and peak strainsof SFRC increase, which lead to more energy absorption and bettercrack control. A different study by Altun et al. on SFC with different

) debonding of differential length change [1].

114 F. Rahimi Mansour et al. / Construction and Building Materials 75 (2015) 112–120

classes of concrete showed higher flexural toughness as the con-crete class changed from C20 to C30 [2].

The major criteria that influence the flexural performance ofSFRC are the volume fraction and aspect ratio of SF. Therefore, itis necessary to determine the optimum aspect ratio and SF volumeto achieve the highest flexural strength. Based on the literature, anoptimal aspect ratio [8,10], and SF volume [2,9] has been found forthe SFRC strength. The optimum volume fraction for SF was foundto be in the range of 1.0–2.5%; less than 1.0% does not have mucheffect, and beyond 2.5% has negative effects, including a reductionof the compressive strength due to difficulties in providing uniformfiber distribution. When an SFRC specimen has a lower volumefraction of SF, a large number of micro-cracks may be initiatedunder compression, easily resulting in specimen failure [3]. Easyworkability is another reason for a dosage restriction of 2.5%because a higher amount makes it difficult to spread the concrete[10]. However, it has been shown by Topco et al. that replacingsome part of the cement by fly ash can increase the workabilityof SFRC [4].

2.2. Bonding behavior of the topping and substrate

The cement-based bonded overlay technique is frequently usedto repair a concrete slab surface or improve the mechanical capac-ity of a structure by increasing its thickness [1]. This method con-sists of adding a thin layer of a cast in-situ topping to the existingslab to enhance its structural performance. At first, a normal rein-forced concrete topping was used, but research shows that usingfiber reinforced concrete produces better results, despite its addi-tional expense. Girhammar and Pajari investigated the use offiber-reinforced concrete versus conventionally reinforced con-crete [12]. Test results showed that the ultimate load-bearingcapacity was approximately the same for both types of topping(i.e., plain concrete or fiber reinforced concrete) [12]. The tensilestrength was somewhat higher for the fiber reinforced concretetopping [12], however.

Whatever the aim of a cement-based overlay (i.e., smoothing orstrengthening), the substrate-overlay combination is expected towork monolithically [1]. Nevertheless, as the member is bent inflexure, the precast member and in-situ concrete topping tend toslide in relation to each other, as shown in Fig. 2, and cause the fail-ure of the member [13]. In this case, the durability of the structuralsystem depends not only on the durability of the material used butalso and primarily on the composite action of the two members [1].According to a study performed by Lemieux et al., the new fiberreinforced concrete layer can be added to the bottom face (i.e., ten-sion zone) or the top face (i.e., compression zone) of the structuralmember without any debonding [14]. They also found that deb-onding usually does not start during the first monotonic loadingbut gradually develops under cyclic loading [14]. Overall, accordingto cyclic loading, no debonding was observed with the overlaylocated in the compression zone, while some debonding occurredfor the overlays located in the tension zone [14]. The resultsstrongly suggest that in the case of an overlay located in the com-pression zone, it is possible to develop and maintain monolithicstructural action in the substrate/overlay composite. Additionally,

Fig. 2. Sliding of the substrate and topping [15].

placing the overlay in the compression zone saves constructiontime and labor.

There are many factors that affect the composite action of thetwo members. To ensure good adsorption, the moisture contentof the substrate plays an important role. Emmons and a statementby Chorinsky that was mentioned by Austin et al. stated that a con-crete substrate surface that is too dry or too wet always results in aweak bond strength of the interface: an excessively dry substratemay absorb too much water from the repair material, which resultsin an incomplete hydration of the cement and reduces the adher-ence, while excessive moisture in the substrate may clog the poresand prevent the absorption of the repair material as well as reducethe adherence. Opinions diverge about pre-wetting the substratesurface [16]. A statement by Talbot et al. and Omar et al. reportedthat pre-wetting the surface (i.e., saturated with a dry surface)before applying the new concrete layer is the most appropriatesolution [16,17]. However, the influence of pre-wetting seems tobe insignificant according to Julio et al., Saucier and Pigeon, whostate that wetting the substrate surface did not influence the dura-bility of the bond using low W/C pastes as bonding agents, butimproved the durability of those made with high W/C pastes.

To achieve the full composite action of the precast slab with aconcrete topping, good bond strength between the two portionsis a key factor to have a monolithic system [12,13]. The bondstrength is the adhesion between the overlay and substrate. Inpractice, some type of reinforcement crossing the interface is usedto increase adhesion between both surfaces (Fig. 3). It is mandatorythat the reinforcement be sufficiently anchored in both the sub-strate concrete and the overlay. The complicated detail of this tech-nique made it costly and difficult to implement.

The bond and shear strength between the contact surfaces mustbe strong when no mechanical connection is provided [13]. Analternative to enhance the adhesion of the two layers requiresusing a bonding agent. In this case, epoxy-based resins are themost commonly adopted in bond fresh to hardened concrete parts,but the resulting benefits are not widely accepted by researchers[19,20]. Some suggest that an adequate bond can only be achievedby combining the use of bonding agents with a proper technique toincrease the substrate roughness, mainly when the substrate has asmooth surface [20]. Others state that bonding agents are not nec-essary if the substrate concrete is dry and properly roughened toexpose its aggregates [20].

In previous research, conventionally reinforced concrete hasoften been used for a topping with some type of fabric reinforce-ment and with special shear reinforcement at joints to increasethe shear capacity of the interface [12]. Today, it is well known thatthe most reliable and economical solution to increase the shearcapacity of the interface is roughening the surface by variousmethods before the placement of the concrete topping. Themechanical connection between the overlay and its substrate canbe achieved by the roughness of surface. According to a researchdone by Girhammar and Pajari, earlier works on pre-stressed hol-low core slabs with a concrete topping include those performed byScott, Ueda and Stitmannaithum [12]. Scott found that compositeaction could be achieved up to the ultimate load without anyreinforcing steel projecting from the precast slab onto the toppingconcrete. Ueda and Stitmannaithum observed complete composite

Fig. 3. Example of dowel bars crossing the interface [18].

F. Rahimi Mansour et al. / Construction and Building Materials 75 (2015) 112–120 115

action with only a small interface slip and without failure at theinterface by only providing roughness at the interface. Later,research performed by Dowell and Smith [21] demonstrated thatprecast concrete panels act compositely in flexure with a cast-in-place topping slab. The concrete topping had no reinforcing barscrossing the interface between the slabs and only different rough-nesses applied to the top of the precast concrete slab. They verifiedthat no horizontal shear slip occurred between the precast con-crete panels and cast-in-place slabs and that the deck acted as afully composite member to failure. Tests carried out by Izni et al.on hollow core slabs (HCU) with concrete toppings ensured themonolithic behavior up to the ultimate loading capacity [13]. Later,they observed higher shear strengths at the interface by roughen-ing the top surface of the HCU. High interface shear strengths ofunreinforced interfaces with simple methods of roughening thesurface of the precast concrete have been achieved by Beushausenbased on a statement by Girhammer and Pajari [12].

It should be note that if the surface roughness is produced by avigorous method, the surface will be very rough, and micro cracksmay be induced just beneath the prepared surface; this will reducethe bond area and bond strength substantially. Therefore, rough-ness by itself does not have significant influence, but the methodof the surface preparation is more important. In this research, athin layer of SFRC was applied on a precast slab to improve its flex-ural performance. The substitution of SF in the topping is expectedto delay overlay debonding and improve the energy absorption andductility of the member. Additionally, to provide a monolithicmember, roughness on the surface was created after casting thesubstrate to prevent micro cracks created by other more vigorousmethods.

Specimen with a SFRC topping Controlled specimen

Fig. 5. Nominal geometry of the specimens.

3. Testing program

As discussed, the new reinforced concrete layer can be added tothe bottom or top face of the structural member. From theseoptions, it seems that adding a new reinforced concrete layer ismuch more effective and easier when the work is performed onthe top face of the member. However, in many practical cases, itis necessary to add the new reinforced concrete to the bottom faceof the member. To investigate the actual behavior of the compositeslab under a combination of flexural and shear loads, a semi-fullscale laboratory test was carried out.

Substrate-overlay composites should work monolithically, andthe bond at the interface ensures the continuity of deformationbetween the base and the overlay. In these conditions, the durabil-ity relies not only on the durability of the material used but also onthe durability of its bond with the base. Poor bonding leads todetachment of the two layers, and the composite member doesnot act as a monolithic structure when the load achieves at a

Fig. 4. Roughness condit

certain value. Therefore, investigations should be performed onthe bonding strength at the interface.

The results and interpretation of the interface bond testsdepends to a large extent on the test method used. Common bondtest methods include interface shear, torsion and tension tests witha wide range of possible test setups [18]. Interface shear strengthvalues obtained by different test methods may differ substantiallybecause test results depend on specimen size, test setup and load-ing rate [18]. A comparison of test results obtained with differenttest methods, therefore, is problematic.

3.1. Test specimens

Four frameworks, including a control specimen, were madewith dimensions of 500 mm in width, 3200 mm in length and100 mm in thickness. Orthogonal mesh reinforcements with a25 mm cover were placed at the bottom part of the substrate toascertain a minimum 3D spacing according to BS8110. These rein-forcements were designed to sustain an ultimate failure load of31.88 kN. The precast member was cast by ready mix concretewith a 40 MPa compressive strength after 28 days. After casting,the surfaces of the fresh concrete were roughened in both the lon-gitudinal and transverse directions by a stiff brush and left in placefor 7 days, as shown in Fig. 4. To cast the topping, a framework500 mm in width, 3000 mm in length and 50 mm in thicknesswas made above the precast layer. Before casting the topping,strain gauges were installed on the mid-span of the substrate inter-face; the surface of substrate was cleaned of any dust and pre-wetfor better bonding between the topping and precast member.Finally, the topping layer was cast by SFRC, including a 1% SF vol-ume, which was shown to be the optimal percentage according toprior research [22]. Hooked-end SF 30 mm in length and 0.75 mmin diameter were used in this experiment. For the control specimenthat did not have SF in its topping, mesh reinforcement was pro-vided at the bottom of the topping layer and cast with normal con-crete. Using mesh reinforcement helps to control the shrinkage andload distribution within the concrete topping. In the design ofthese composite specimens, consideration has been given tochoose a thickness such that the maximum interface shearstrength can be studied at a location closer to the neutral axis of

ion at the interface.

Table 1Concrete properties of precast member and topping.

SlabNo

Type of interfaceroughness

Percent of SFintopping (%)

Precast member Topping

Compressivestrength (MPa)

Splitting tensilestrength (MPa)

Flexural strength(MPa)

Compressivestrength (MPa)

Splitting tensilestrength (MPa)

Flexural strength(MPa)

28 days Testday

28 days Testday

28 days Testday

28 days Testday

28 days Testday

28 days Testday

1 Smooth as cast 1 33.3 40.5 2.4 2.4 12.5 13.8 26.9 26.9 2.9 2.9 13.5 142 Transverse

direction30.2 32 2.9 3.1 12.6 13.2

3 Longitudinaldirection

36.1 38.3 2.9 3 14.9 15.3

4 Transversedirection

0 36.9 38.5 2.7 2.8 12.4 12.5

Fig. 7. Load versus the mid-span deflection.

116 F. Rahimi Mansour et al. / Construction and Building Materials 75 (2015) 112–120

the member. The overall nominal geometry of the specimens isshown in Fig. 5.

The characteristics of the specimens and concrete properties forboth the precast unit and topping after 28 days and on the testday (i.e., day of the flexural tests on the slab specimens) areshown in Table 1. These properties are based on an average ofthree cubes with dimensions of 150 mm � 150 mm � 150 mm,three cylinders with dimensions of 300 mm in height and150 mm in diameter, and three prisms with dimensions of500 mm � 100 mm � 100 mm.

3.2. Testing setup

All of the specimens were whitened for better observation ofthe development of cracks. The specimens were simply supportedand subjected to a pair of point loads at a distance of 1.0 m fromboth ends. Two linear variable differential transducers (LVDT) wereplaced at both ends of the precast slab and topping to measure theinterface slip. The differences in the LVDT measurements quanti-fied the amount of interface slip. In addition, three other LVDTswere installed under the specimens in line with the mid-spanand the applied loads to measure the deflection of the specimen.To measure strain, eight Demec gauges were installed the speci-mens at different depths. The overall test layout is illustrated inFig. 6.

The load applied increased by 2 kN increments at each step andcontinued until failure. The failure load was defined as a load thatcaused the specimen to fail in flexure or that caused failure at theinterface between the substrate and overlay. Mid-span deflections,strain values and interface slips were recorded for every loadincrement and will be discussed in the next parts.

Fig. 6. Test

4. Test results and discussion

4.1. Load versus mid-span deflection

The load versus mid-span deflection curves for all slab speci-mens are shown in Fig. 7. Throughout the test, the deflectionsrecorded at 1 m from both supports were the same, which showedthat the loading and test rig position were centrally aligned. Ini-tially, as the load increased, the deflection also increases linearlyup to a certain load (i.e., the yield load), and after that point, themid-span deflection varies non-linearly and reaches a maximumvalue. Beyond the ultimate load point, the deflection starts increas-ing appreciably as the load increases. Slab No. 2 shows the highest

layout.

Table 2Ductility of the specimens.

Slab No. Yield load (kN) Ultimate load (kN) Displacement at yield load Displacement at ultimate load Ductility factor

1 60 72 43.06 99.88 2.312 60 77.7 35.06 128.41 3.663 62 72 38.61 102.78 2.664 68 84 40.8 105.84 2.56

Fig. 8. Load versus strain for slab No. 1.

Fig. 9. Load versus strain for slab No. 2.

Fig. 10. Load versus strain for slab No. 3.

F. Rahimi Mansour et al. / Construction and Building Materials 75 (2015) 112–120 117

deflection, which is not comparable to other specimens. The max-imum and minimum deflections occur in slab No. 2 and slab No. 1as 128.41 mm and 99.88 mm, respectively.

From the load versus deflection curves, it is evident that theultimate load carrying capacity of slab No. 4, which does not haveSF in its topping, is higher than those slabs with SF reinforced top-pings. The ultimate load for slab No. 4 is 84 kN, while the ultimateload for slab No. 2 is recorded as the second highest with 77.7 kN.The ultimate load capacity for both slab No. 1 and 3 are 72 kN,which are lower than the other specimens. However, the resultsare still higher than Pcal.

The ability of a member to deform without a significant loss ofits strength is known as ductility. One method of quantifying duc-tility is by the ductility factor (i.e., the displacement ductilityindex), which is defined by the ratio of the ultimate deflection tothe deflection at yielding of the tensile reinforcement [23]. Thedisplacement ductility factors calculated according to the abovedefinition are shown in Table 2.

It is shown that slab No. 2 exhibits the highest ductility with aductility factor of 3.66. The lowest ductility factor is shown in slabNo. 1 at 2.31. Comparing slabs No. 2 and 3, which are different onlyin the type of roughness at their interface, a 28% improvement inductility is achieved when the roughness is changed from longitu-dinal to transverse. The results show the positive effects ofinterface roughness and SF in the topping. The ductility factorincreases by approximately 36% when only the interface roughnessof the specimens changes from smooth as-cast to the transversedirection. However, ductility factor change from 2.56 to 3.66 whichis nearly 30% when both specimens have roughness in the trans-verse direction and only 1% SF added in their toppings. Therefore,the ductility of a slab depends on both the roughness conditionat the interface and the existence of SF in the topping. A good resultis achieved when the interface roughness and SF topping is appliedto a specimen without considering the direction of roughness, asshown by slab No. 3, which had longitudinal roughness and SFtopping.

Comparing the energy absorption of the specimens is also pos-sible by investigating the load versus deflection graph. Because thearea under the load–deflection diagram defines energy absorption,slab No. 2, which contains SF in its topping and has a roughness inthe transverse direction, is capable of dissipating more energy. Inaddition, the energy dissipation of the specimen without SF butwith roughness in the transverse direction is even better thanthe SF specimens with any type of roughness at the interface. Thisshows that the performance of the composite member depends notonly on adding SF to the topping and mainly depends on the typeof interface roughness.

4.2. Load versus strain

The load was applied gradually to all of the specimens, andstrain along the mid-span was measured at various stages of load-ing. Figs. 8–11 show the load versus strain curves of the specimensin 5 different levels that are defined in Fig 6. Strain at the interfacewas measured by a strain gage installed at the interface beforecasting the topping. The strains at other levels were found by

Demec gauges installed within the specimens at various depths.The positions of the strain gage and Demec gauges are shown inFig. 6. Negative and positive strains recorded at the top and bottomsurface, respectively, indicate that the specimens are in sagging.

As shown in the figures, the strain at the mid-span graduallyincreases during the test by increasing the applied load in all fourspecimens. Significant changes in the slope of the curves indicatethat multiple cracks started to occur at the mid-span. The maxi-mum positive strain occurred in the tension region at level 5, while

Fig. 11. Load versus strain for slab No. 4.

118 F. Rahimi Mansour et al. / Construction and Building Materials 75 (2015) 112–120

the maximum negative strain occurred in the compression zone atlevel 1. When the specimens are close to their ultimate capacity,large bottom strain increments were observed for a small incre-ment of load. The results of these figures will be used later to findthe interface slip and efficiency of the system.

4.3. Interface slip

The aim of this study is to characterize the effects of a thin layerof SFRC on the flexural capacity of an existing slab. When the com-posite members are subjected to an increasing load, they initiallydeform with the load at a very low rate. However, after the criticalvalue of stress is reached, the specimens undergo a large deforma-tion with relatively small increases in the applied load. This defor-mation is caused by the slippage of the material along contactsurfaces. Interface slip can be measured by strain distributiondiagram along the depth of slab and will be explained later.

As discussed in previous sections, the efficiency of this compos-ite member depends on the bonding between the substrate and the

Fig. 12. Theoretical strain

Fig. 13. Strain distribu

topping. To evaluate the bonding condition, slippage between thetwo layers has been measured at the ends and the mid-span ofall specimens. There are several methods to detect debonding: typ-ically, LVDTs or strain gages spanning the interface are placed andwill be activated when debonding reaches their location. Based ontheir low cost and high reliability, LVDT sensors are oftenrecommended.

4.3.1. Interface slip at the endInterface slip at each end of the specimens was measured using

two LVDTs. Any movement between the two layers was observedas differences in the LVDT values and defined as interfaceslip. The maximum and minimum movements obtained were0.04 mm for slab No. 3 and zero for slab No. 2, respectively.

4.3.2. Interface slip at the mid spanStrain describes the deformation of the specimen and can be

measured as changes of the specimen’s length divided by the spec-imen’s original length. The compressive region yields negativestrains, while the tension region yields positive strains. Interfaceslip at the mid-span is calculated using a strain distributiondiagram along the depth of the specimens using the following def-inition. Fig. 12 shows the strain distribution diagram of a specimenwith Demec gauges installed at points A, B, C and D. In the case ofno slippage, the strain distribution diagram follows Fig. 12a, whileslippage causes the strain to change between points B and C, asshown in Fig. 12b. To calculate the interface slip, the gradient ofthe bottom is considered as the true strain gradient and is thenapplied for the overlay’s strain gradient as illustrated in Fig. 12c.The dotted line shows the true strain gradient applied to theoverlay’s strain gradient. The difference between B0 and C0 showsthe interface slip at the center of interface. The diagram with alower gradient shows less interface slip and therefore the bestmonolithic behavior.

distribution diagram.

tion for slab No. 1.

F. Rahimi Mansour et al. / Construction and Building Materials 75 (2015) 112–120 119

Strain distributions along the depth of the specimens weremeasured with strain gages and Demec gauges and are shown inFigs. 13–16.

The interface slip is zero in the case of equal strain gradients forboth layers; the gradient changes if any slippage occurs. Figs. 13–16 show changes in strain gradient for the substrate and overlayfor all of the specimens. The test results show that the strain gra-dient starts to change for the substrate and the overlay by applyingonly 10 kN load. However, the rate of changes at this load is veryless for slab No. 2 comparing to the other specimens. Thesechanges become more evident as the load is increased towardsthe ultimate load. The bottom strains are slightly larger than thetop strains, indicating that that the precast member is undertension while the topping is in compression.

Fig. 14. Strain distribut

Fig. 15. Strain distribut

Fig. 16. Strain distribut

Interface slip results at the end show that interface slip wasonly observed in slab No, 3. However, for the other specimens withstrain distribution diagrams pattern shown in Figs. 13–16, thestrain gradients above and below the neutral axis were not same.It suggests an internal slip occurred in the mid-span region (wherethe strain gauge was positioned).

According to the previous explanation, the substrate’s straingradient was applied to the overlay’s strain gradient, allowingthe interface slip to be measured. Interface slip for all specimensis shown in Fig. 17. Large slip at the interface is a sign of poor bond-ing between the concrete topping and precast slab.

The largest slip occurred for slab No. 4 at 0.38 mm, indicatingpoor bonding strength at the interface. This result proves thatthe substitution of SF in concrete is expected to delay overlay deb-

ion for slab No. 2.

ion for slab No. 3.

ion for slab No. 4.

Fig. 17. Interface slip at the mid-span.

120 F. Rahimi Mansour et al. / Construction and Building Materials 75 (2015) 112–120

onding. Slab No. 1, with 0.1 mm of slippage, showed the bestinterface slip. It increased rapidly to 0.006 mm by applying only10 kN, remained constant up to 50 kN and reached 0.1 mm at60 kN. However, slab No. 2 showed more slippage at the interface;its slip also changed in a markedly different way, remaining nearzero until a under a load of 30 kN and remained superior to theother specimens up to 40 kN but then increased suddenly. Inter-face slippage was observed in all of the specimens; however, nonefailed at the interface.

5. Conclusion

In this research, a thin layer of SF concrete was added to a con-crete slab to increase its flexural performance. Using SF in concreteis expected to overcome certain problems. The performance of thiscomposite slab depends strongly on the full composite action ofthe precast member and the topping. Therefore, good bondingstrength at the interface is required. According to previousresearch, roughness on substrate surface could create goodbonding between the two layers. Compared to other bondingtechniques, roughness on substrate surface is a cost effectivemethod and does not require intensive labor.

Based on the experimental tests, ductile performance dependsnot only on adding SF to the topping but mainly on the type ofinterface roughness. Higher energy absorption and ductility werefound in slab No. 2, in which SF was introduced to the compositeslab and the interface surface was roughened in the transversedirection.

The largest slip occurred in slab No. 4. This result proves thatthe substitution of SF in concrete is expected to delay overlay deb-onding and result in good bonding at the interface. However, asmall interface slip occurred even for the specimens that used SFRCin the topping. The fracture surface occurred in the concrete sub-strate without any failure at the interface. The overall resultsshowed that the best toughness and ductility occurred in slabNo. 2, despite a small slippage at the interface.

References

[1] Turatsinze A et al. Durability of bonded cement-based overlays: effect of metalfibre reinforcement. Mater Struct 2005;38(3):321–7.

[2] Altun F, Haktanir T, Ari K. Effects of steel fiber addition on mechanicalproperties of concrete and RC beams. Constr Build Mater 2007;21(3):654–61.

[3] Wang Z, Wu L, Wang J. A study of constitutive relation and dynamic failure forSFRC in compression. Constr Build Mater 2010;24(8):1358–63.

[4] Topçu _IB, Canbaz M. Effect of different fibers on the mechanical properties ofconcrete containing fly ash. Constr Build Mater 2007;21(7):1486–91.

[5] Olivito R, Zuccarello F. An experimental study on the tensile strength of steelfiber reinforced concrete. Compos B: Eng 2010;41(3):246–55.

[6] Soulioti D et al. Effects of fibre geometry and volume fraction on the flexuralbehaviour of steel-fibre reinforced concrete. Strain 2011;47(s1):e535–41.

[7] Özcan DM et al. Experimental and finite element analysis on the steel fiber-reinforced concrete (SFRC) beams ultimate behavior. Constr Build Mater2009;23(2):1064–77.

[8] Wang Z, Wu J, Wang J. Experimental and numerical analysis on effect of fibreaspect ratio on mechanical properties of SRFC. Constr Build Mater2010;24(4):559–65.

[9] Khaloo AR, Afshari M. Flexural behaviour of small steel fibre reinforcedconcrete slabs. Cem Concr Compos 2005;27(1):141–9.

[10] Xu B, Shi H. Correlations among mechanical properties of steel fiber reinforcedconcrete. Constr Build Mater 2009;23(12):3468–74.

[11] Uygunoglu T. Investigation of microstructure and flexural behavior of steel-fiber reinforced concrete. Mater Struct 2008;41(8):1441–9.

[12] Girhammar UA, Pajari M. Tests and analysis on shear strength of compositeslabs of hollow core units and concrete topping. Constr Build Mater2008;22(8):1708–22.

[13] Ibrahim IS, Kim SE, Copeland S. Bending capacity of precast prestressed hollowcore slabs with concrete toppings. Malaysian J Civil Eng 2008;20(2):260–83.

[14] Lemieux M et al. Behavior of overlaid reinforced concrete slab panels undercyclic loading-effect of interface location and overlay thickness. ACI Struct J2005;102(3).

[15] Gohnert M. Horizontal shear transfer across a roughened surface. Cem ConcrCompos 2003;25(3):379–85.

[16] Julio EN, Branco FA, Silva VtD. Concrete-to-concrete bond strength. Influenceof the roughness of the substrate surface. Constr Build Mater2004;18(9):675–81.

[17] Omar B, Fattoum K, Maissen B. Influence of the roughness and moisture of thesubstrate surface on the bond between old and new concrete. Contemp Eng Sci2010;3:139–47.

[18] Bissonnette B, et al. Bonded cement-based material overlays for the repair, thelining or the strengthening of slabs or pavements. State-of-the-art report ofthe RILEM technical committee.

[19] Bonaldo E, Barros JA, Lourenço PB. On bonding repairing steel fibre reinforcedconcrete to hardened concrete, 2005.

[20] Santos DS, Santos P, Dias-da-Costa D. Effect of surface preparation and bondingagent on the concrete-to-concrete interface strength. Constr Build Mater2012;37:102–10.

[21] Dowell RK, Smith JW. Structural tests of precast, prestressed concrete deckpanels for California freeway bridges. PCI J 2006;51(2):76–87.

[22] Mansour FR, Parniani S, Ibrahim IS. Experimental study on effects of steel fibervolume on mechanical properties of SFRC. Adv Mater Res 2011;214:144–8.

[23] Dhasarathan A, Thenmozhi R, Shree MSD. Experimental study on the ductilecharacteristics of hybrid ferrocement slabs. A research work of GovernmentCollege of Technology.