Introduction

Need for strengthening of reinforced concrete (RC) structures is becoming more apparent, particularly, when there is an enhancement in peak load, a change in use, degradation problem and some design and/or construction deficiency. Use of externally bonded fibre-reinforced polymer (FRP) sheets/plates/strips/warps is a modern and easy way to strengthen RC beams [1, 2]. Several studies have stated that FRP composites when applied to RC members improve the performance, durability and cost-effectiveness in rehabilitation [3,4,5]. Among the all fibres, glass fibres have high-temperature resistance, high strength and low cost, which makes glass fibre-reinforced polymer (GFRP) the most suitable FRP reinforcement for various civil engineering applications [6, 7]. Many researchers have reported that FRP-strengthened RC beams exhibit an increase in strength and decrease in ductility [8,9,10]. For successful strengthening, adequate bond between the FRP laminates and existing concrete members is a significant factor, whereas others reported that damage of interfacial bond due to ‘concrete cracking’ played a more important role [11, 12]. On the other hand, incorporating fibres in concrete effectively controls the crack development [13].

In order to solve the above problems, an effective approach is to include short discrete fibres into cement-based materials to increase its structural integrity. Addition of such fibres not only enhances the requisite properties of RC, but also changes the characteristics of the material from brittle to ductile [14, 15] and improves fatigue performance [16, 17]. A variety of fibres are existing such as natural, synthetic, steel and glass fibres [18, 19]. Steel fibre has high toughness and tensile strength [20], but workability reduction due to balling and corrosion are the key challenges in steel FRC [21]. Though glass fibre improves the concrete toughness, long-term strength of concrete is reduced due to alkali attack. High strength, chemically inert carbon fibres have a disadvantage of high cast. Advanced materials offer the assurance of innovative applications in concrete composite. Basalt fibre (BF) is an inorganic fibre which is derived from the molten basalt rock. The processing method of this type of fibre is like glass fibre, but less energy is used and no supplement is required, which makes it cheaper than carbon or glass fibres [22, 23]. The inclusion of basalt fibre extensively increases the energy absorption capacity and deformation of geo-polymer concrete [24]. It also enhances the tensile strength, fracture energy and flexural strength [25, 26].

The new synthetic polyolefin fibre, which is ductile and more flexible, is expected to improve strain capacity and toughness. Addition of polyolefin fibres improves the fracture energy, tensile strength [27] and restrains the shrinkage cracking in concrete [28]. It has been shown recently that HFRC composites can offer more desirable engineering properties than single FRC [29]. The flexural capacities of strengthened beam elements have also been investigated by the introduction of finite element (FE) modelling technique [30,31,32,33].

On the basis of a comprehensive literature review reported on the RC beam strengthening, external attachment of FRP lamination is found to be more effective. However, the use of fibres and external strengthening in combination is uncommon. Therefore, studies which focus on both the strength and ductility of structural members are required to be undertaken. Hence, the key of this research is to assess the performance of GFRP strengthened RC beams along with the addition of polyolefin and basalt fibres. The purpose is to study the ductility and strength parameters of RC beams strengthened with HFRC and compare the same with the RC strengthened beam and control beam. Additionally, finite element models were developed using ANSYS software (version 15) for simulating the observed experimental behaviour.

Experimental study

Materials

OPC 53 grade was used; coarse aggregate and fine aggregate having a specific gravity of 2.63 and 2.63 and maximum size of 20 and 4.75 mm were used in this study. Throughout the study, tap water was used for manufacturing concrete. A commercial superplasticizer (SP) ‘classic super flow’ conforming to IS: 9103-1999 [34], with the dosage of 0.5–1.5% by weight of cement, was used to enhance the concrete workability. In order to fabricate workable concrete, a water/cement (w/c) ratio of 0.46 has been used in this study [35]. Both basalt and polyolefin fibres, acquired from commercial establishments, were used in concrete. In the present study, two fibres were used to form the hybrid fibres. The properties of above-mentioned fibres are presented in Table 1.

Table 1 Salient properties of basalt and polyolefin fibres
Full size table

The longitudinal reinforcement for all the beams consisted of high yield strength deformed (HYSD) bars of 10 mm and 8 mm diameters, conforming to the relevant IS code. Commercially available unidirectional GFRP laminates were used for flexural strengthening of beams on the tension side. The GFRP coupons were prepared using the wet (hand) layup system as per ASTM [36]. The rupture strain and tensile strength of GFRP fabric determined from wet layup specimens were found to be 1.15% and 1300 MPa, respectively. Two-part epoxy resin (mixing ratio of resin to hardener = 4:1 by weight) was used, for bonding the laminates to the beams.

Mix proportioning and casting of beams

The proportioning of the concrete mixture has been carried out in accordance with IS: 10262-2009 [37], with a 31.6 MPa target mean strength, and this formed the ‘control mix’. The proportions of the control concrete mixture are given in Table 2. Basalt fibre with 1% fibre volume fraction (Vf) was incorporated into ‘control mix’ for casting fibre-reinforced concrete (FRC) beam specimens. Basalt along with polyolefin fibres has been mixed in the ratio of 70:30 at 1% fibre volume fraction (Vf) and incorporated into control mix for casting hybrid fibre-reinforced concrete (HFRC) beams. Three groups of RC beams were casted, which were ordinary RC beams (B0P0 L0 and B0P0 L5) in group A, basalt FRC beams (B100P0 L0 and B100P0 L5) in group B, whereas the beams in group C were basalt and polyolefin HFRC beams (B70P30 L0 and B70P30 L5). In each group, a beam was strengthened using externally bonded unidirectional GFRP laminates (B0P0 L5, B100P0 L5 and B70P30 L5). The test program was aimed to examine the strength and deflection behaviour of RC, FRC and HFRC beams with and without laminates. All the beams had a size of 150 mm depth, 100 mm width and 1000 mm length. All beams were built in compliance with IS 456:2000 [38]. The reinforcement, dimension and geometry details of tested beams are shown in Fig. 1. The details of designation of beams are summarized in Table 3.

Table 2 The proportions of the control concrete
Full size table
Fig. 1
figure 1

Reinforcement details of beam

Full size image
Table 3 The designation and details of various beams
Full size table

FRP strengthening of beams

All the beams were casted and then cured for 28 days, and after that, the strengthening process was carried out on the tension face of the beams using the ‘wet layup’ system in accordance with the American Concrete Institute specifications 440.2R-2008 [1]. Before bonding the laminates, the tension surface of all the beams was cleaned, dried and greased with contaminates, if any, removed. The tension side of the beams was prepared to a minimum rough profile (CSP 3) as defined in International Concrete Repair Institutes guidelines, (ICRI 03730) [39]. Abrasion disc was used to remove any loose cement paste layer such that one could see clearly the exposed aggregate particles. The various stages of preparation of surface are shown in Fig. 2.

Fig. 2
figure 2

Step-by-step procedure for strengthening of beams: a Surface preparation, b GFRP fabric cutting, c application of epoxy on concrete surface, d squeezing out of entrapped air bubbles

Full size image

The prepared surface was then uniformly coated with a primer and allowed to cure. After that, the resin and hardener were mixed in the predetermined ratio 4:1 with the mixing carried out according to the manufacture’s recommendations, until a uniform colour was attained. The reinforcing glass fibres were cut according to the required size, cleaned to remove grease or contaminates, if any, present on the fibres. The epoxy resin was applied uniformly onto the prepared surface, and then, the reinforcing glass fibres were pressed onto the above surface with gentle pressure being applied using a rubber roller. This was done to make sure that the FRP sticks well to the surface, ensure expelling of entrapped air and to remove excessive epoxy that may be present between the FRP and concrete surface. Laminated beams thus prepared were cured for a week before testing them. The procedure for beam strengthening is shown in Fig. 2a−d.

Test set-up and instrumentation

All beams have been tested on a universal testing machine (UTM) available in the Department of Civil Engineering, PEC. The beams were simply supported with 50 mm offset from both the supports, having an effective span of 900 mm. Loads were applied with an increment of 2.5 kN under the four-point method. The load from the load cell was transferred to the beam with the help of a steel I-section spreader-beam. This load from the spreader beam got transferred onto the two loading points on the beam. A mechanical dial gauge having an accuracy of 0.01 mm was used to measure the deflections at the mid-span of beams, and crack-widths were measured using a crack detection microscope having a least count of 0.02 mm. Figure 3 shows the photographic view of the test set-up along with the instrumentation used. The loads were gradually increased until failure of the beams. Load and maximum deflection were recorded at each load interval. Further, crack propagation and distribution were marked and noted at every 10-kN loading interval. Load (in kN) and deflections (in mm) formed the basic experimental data, based on which load–deflection curves for all the beams were developed, and the ductility ratios were calculated to evaluate the relative ductility performance of the various beams.

Fig. 3
figure 3

The experimental test set-up

Full size image

Experimental results and discussion

Effect on load carrying capacity

The experimental outcomes are shown in Table 4, in terms of the measured ultimate load, yield load and primary crack load and their corresponding deflections. The enhancement in the load carrying capacity of ‘laminated RC beam’ over the control beam at the above three stages was 31.1%, 41.2% and 33.4%, respectively. It shows that there is an effect of lamination on the strengthened beam at all the above three stages with the increase being very high at the first crack and ultimate stage, the highest being at the ultimate stage. Comparing the loads at the three stages of ‘laminated FRC beam’ with that of ‘laminated RC beam’, it was found that the percentage increase of loads was 9.86%, 27.6% and 27.7%. This shows the significant role played by adding basalt fibres in the strength enhancement of ‘laminated RC beam’. Comparing the ultimate, yield and first crack loads of ‘laminated HFRC beam’ with respect to that of ‘control beam’, the percentage increase is 79.13%, 108.82% and 81.86%, respectively. This shows that there is a ‘combined effect’ of fibres which extensively influences the above loads of ‘laminated HFRC beams’. Similar findings were observed by Syed et al. [40], Ramesh et al. [41, 42]. Figure 4 shows the load comparisons at various stages of tested beams.

Table 4 Experimental results of various tested beams
Full size table
Fig. 4
figure 4

Comparisons of loads at various stages

Full size image

Effect on deflection

The recorded deflections for all tested beams are listed in Table 4. The deflection at first crack, yield and at ultimate stages recorded for control beam was 1.25 mm, 3.61 mm and 8.6 mm. The maximum deflection at all above three stages for ‘laminated RC beam’ was 1.30 mm, 3.76 mm and 6.3 mm, respectively. The deflection at all stages of ‘laminated RC beam’ decreases over the ‘control beam’ which indicates the contribution of stiffness due to laminates. Using basalt and polyolefin fibres for ‘laminated HFRC beams’ led to an increase in the deflections. The maximum percentage increase in the deflection at primary crack, yield and ultimate stages of ‘laminated HFRC’ beams over the ‘control beam’ is about 96%, 38.5% and 46.51%, respectively. Further, the increase in deflection at all above three stages of ‘laminated HFRC’ over the ‘laminated RC beam’ is about 88%, 32.97% and 100% respectively. This indicates that there is a ‘combined effect’ of fibres. Deflection at ultimate stage for laminated HFRC beam was found to be 16.33% less when compared to HFRC beam.

Effect on ductility

The ductility is a measurement of the capacity of the structural member to absorb energy and can be expressed as deformity resistance when transforming from the elastic to the plastic zone and failure. Ductility in a reinforced concrete member can be measured by using any of the following methods: (1) rotational ductility index, (2) curvature ductility index or (3) deflection ductility index. In the present study, ductility is calculated using deflection ductility index method. Deflection-based ductility (DΔ) is defined as the ratio between ultimate deflection (Δu) to deflection during yield stage (Δy) [43]. That is,

$$D_{\Delta } \, = \,\Delta_{{\text{u}}} /\Delta_{{\text{y}}}$$
(1)

Camata et al. [44] suggested the ductility in terms of energy (DE). It is defined as the energy (area under load–deflection curve) of the beam at ultimate stage (Eu) divided by the energy of the beam at yield stage (Ey). That is,

$${{D}}_{{\text{E}}} \, = \,{{E}}_{{\text{u}}} /{{ E}}_{{\text{y}}}$$
(2)

The measured ductility in terms of energy and deflection is shown in Table 5 and Fig. 5

Table 5 Ductility indices of beams
Full size table
Fig. 5
figure 5

Comparison of ductility of beams

Full size image

After examining the outcomes, it can be observed that the deflection and energy ductility were maximum for ‘HFRC beam’ [B70P30 L0]. As seen in Table 5, the deflection ductility dropped from 2.38 to 1.67 for control beam when compared to laminated RC beam. This indicates that a greater reduction in the ductility of the beam occurs when the beam is strengthened with GFRP laminates as is an external attachment to the RC beams and decreases the potential of redistribution of moments in the hogging region. Similar findings were reported by Attari et al. [45], Jason duic [46]. Using ‘laminated RC’ sample as a reference beam, the deflection and energy ductility were enhanced by 25.14% and 56.88% and 24.87% and 49.74% for ‘laminated FRC’ and ‘laminated HFRC’ beams, respectively. This shows that there is a combined effect, which extensively enhances the ductility ratios of laminated HFRC beam.

Finite element analysis (FEA)

FEA software ANSYS (version 15) was used to model all the beams.

Element type

Selection of element type is more important in any FEA. The following elements were selected for modelling the beams.

  • Concrete Concrete was modelled using the 3-D ‘SOLID65’ element, which is defined by 8 nodes having 3 degree of freedom (DOF) at each node as shown in Fig. 6a. This element is capable of handling plastic deformation and cracking.

  • Steel reinforcement The longitudinal and transverse reinforcements were modelled using the two nodes ‘link180’ spar element with three DOF, i.e. translations at each node, as shown in Fig. 6b.

  • Steel Plates To model the steel plates at the point of application of load and at supports, ‘solid185 structural’ element was used (Fig. 6c).

  • GFRP laminates The GFRP laminates used for beam strengthening were modelled using ‘solid185-layered’ element, which has eight nodes and three DOF, i.e. translations (Fig. 6d). The connection between the concrete and GFRP was modelled using the (basic) concept that ‘the bond between the concrete and FRP is considered perfect’. The perfectly bonded model has been widely adopted by many researchers such as Hassan et al. [47], Jayajothi et al. [48].

Fig. 6
figure 6

a–d Details of the type of elements used for modelling of various components of beams [49]

Full size image

Real constants

Real constants are properties that depended on the cross-sectional properties and the element type. Not all materials require elements of the same type and real constants may have different values. ‘Real constant set 26’ was used for the solid 65 elements. Values were entered for the volume ratio and material number. Values for cross-sectional (C/S) area and initial strain were entered. Zero strain value was adopted at the starting of the experiment because there is no primary stress in the beam. C/S area in sets 8 and 10 refers to the reinforcement of 8 mm (main) and 10 mm (hanger) diameter bars, respectively.

Material properties

Concrete is a brittle material, which exhibits a nonlinear behaviour both in compression and tension. The material properties used for plain concrete model are as follows: elastic modulus = 28062 MPa, Poisson’s ratio = 0.33, compressive strength = 33.23 MPa. Steel reinforcement for the model was assumed to exhibit an ‘elasto-plastic’ behaviour which is identical both in compression and tension. The modulus of elasticity, yield stress and Poisson’s ratio for steel were taken as 200,000 MPa, 415 MPa and 0.3, respectively. The behaviour of the GFRP materials was considered to be linear elastic up to failure, and the modulus of elasticity and tensile stress were considered as 1620 MPa, 515 MPa, respectively.

Modelling and gluing

Solid blocks were used to model concrete beam, supporting steel plates and GFRP laminate, while line elements were used to model steel reinforcement. ‘Glue’ command was used to attach the GFRP layer with the concrete beam.

Boundary conditions

Boundary conditions were applied at the loading points and supports. For all beams, the supports were modelled as simply supported, and the loading was applied using the four-point method. At the supports, displacements in all directions were constrained in the model to get a unique solution. To simulate simply support condition, X, Y and Z directions at the nodes of one support were given constant values of zero. Similarly, only Y direction at the nodes of the other support was given a zero value. The beams were coupled at one-third of effective span, and two equal forces were applied on the coupled nodes.

Meshing

After developing the beam models with supporting plates, it has to be meshed before carrying out the process of analysis. Mesh tool which is available in preprocessor helps to establish mesh attributes, size control and mesh shape which are used for meshing the solid models. No reinforcement mesh is required because individual elements have been generated via the nodes in the modelling process. The GFRP laminates were meshed as brick elements. Square shape mesh of size 25 mm was chosen for all beams. Solution controls command was utilized to run the static analysis of all beams. The detail of overall meshing with boundary conditions for RC beam is shown in Fig. 7.

Fig. 7
figure 7

Typical FE modelling of RC beam

Full size image

Comparison of experimental and numerical results of various beams

Experimental outcomes have been compared with FE model analysis. Comparisons were partial to main characteristics of beams such as primary crack, yield load, ultimate load, deflection and load–deflection response. Tables 6 and 7 summarize the load carrying capacity and their corresponding deflections at all above three stages of the various beams, acquired from the evolved FE models and experiments. The FE models results for loads are within ± 12%, and the deflections are within ± 8% of the experimental results. This variation may be due to the reason that there is ‘perfect bond between the steel reinforcing and concrete’ as is assumed in the FE modelling, but this assumption is not true for the experimentally tested beams. The percentage difference is considerably reasonable and acceptable; similar observations were made by Choobbor et al. [33].

Table 6 Comparison of loads at various stages: experimental versus numerical
Full size table
Table 7 Comparison of deflections at various stages: experimental versus numerical
Full size table

Load–deflection relationship

The developed finite element model (FEM) of full-scale beams results was compared with the outcomes obtained from the experimental investigations. Figure 8a−f shows the numerical and tested load–deflection response of beams at all loading stages. From the graph, it is observed that the first crack occurs at the linear region. Beyond that the curve becomes nonlinear and the yielding of steel takes place in this region with the failure occurring at the ultimate stage. The typical nodal displacement of beams at ultimate stage is shown in Figs. 9 and 10. It is clear from Fig. 8a−f and Tables 6 and 7 that the load–deflection response of all beams from FE model analysis agrees well with the experimental results.

Fig. 8
figure 8

a–f Load–deflection curves of all beams: experimental versus numerical

Full size image
Fig. 9
figure 9

Typical nodal displacement of control beam (B0P0L0) at ultimate stage

Full size image
Fig. 10
figure 10

Typical nodal displacement of laminated HFRC beam (B70P30L5) at ultimate stage

Full size image

Delamination behaviour and cracking pattern

The crack pattern obtained from the FE modelling at each load step is compared with the tested beams. Figures 11 and 12 show the typical crack pattern of the tested and modelled beams. The control beam had initial flexural cracking accompanied by the concrete crushing in compression. The crack pattern for ‘laminated RC’ beams was observed by the development of vertical cracks at the mid-span in the tension zone. With applying further loads, a greater number of diagonal cracks appeared and the initial cracks propagated towards the loading points. At the ultimate load, due to the sliding between the laminate and concrete, development of horizontal cracks in the tension zone was observed. This beam had a greater load carrying capacity as compared to the reference beam specimen, but brittle failure occurred due to debonding of laminates with a sudden decrease in load carrying capacity. More number of cracks was observed in laminated HFRC beam compared to the control and RC beam. Incorporation of fibres in concrete increased the load carrying capacity, resulting in ductile failure. The load at initiation of FRP delamination of beams increased with the addition of fibres into the concrete. The ANSYS software shows circles at crushing and cracking positions. Cracking and crushing are highlighted with a circle and octahedron outline, respectively. The first crack at an integration point is shown as a red outline, whereas the second and third cracks are shown as green and blue outlines. Flexural cracks were observed at the mid-span regions which were formed vertically up the beam. Finally, at the last load step, compression cracks appeared.

Fig. 11
figure 11

a–b Crack pattern of control beam (B0P0L0) at the ultimate load

Full size image
Fig. 12
figure 12

a–b Crack pattern of laminated HFRC beam (B70P30L5) beam at the ultimate load

Full size image

Conclusions

  • The fibre volume fraction and combination of fibres (basalt: polyolefin) influence the load at various stages. The loads at above all stages are maximum for the laminated HFRC beam.

  • The percentage enhancement in the load capacity at first crack, yield and ultimate stage of laminated HFRC beam, over the control beam, is about 79.13%, 108.82% and 81.86%. This shows that there is a combined effect of hybrid fibres and GFRP lamination.

  • The deflection at ultimate load of laminated RC beam decreases over the control beam. The mid-span deflection of ‘HFRC beam’ is maximum. The percentage increase in deflection of HFRC beam over control beam is about 75.11% at the ultimate stage.

  • Laminated HFRC beam has the highest ductility. The maximum increase in the deflection ductility of laminated HFRC beams ranges from 1.39 to 1.70 times the control beam and 1.59 to 1.76 times the laminated RC beam

  • Yielding of steel reinforcement, followed by crushing of concrete in compression, is typical failure mode of the ‘control beam’. Failure due to debonding of laminate and rupture of laminate were typical failure modes in the case of laminated RC beam and HFRC beam, respectively.

  • The ultimate loads of all beams obtained by the FE model are slightly higher than the results from experimental study. The percentage difference between the above two methods ranges from 5 to 12%. This shows a very close agreement between the above two methods.