Experimental and FE analysis of composite RC beams with encased pultruded GFRP I-beam under static loads

Abstract

Encasing glass fiber reinforced polymer (GFRP) beam with reinforced concrete (RC) improves stability, prevents buckling of the web, and enhances the fire resistance efficiency. This paper provides experimental and numerical investigations on the flexural performance of RC specimens composite with encased pultruded GFRP I-sections. The effect of using shear studs to improve the composite interaction between the GFRP beam and concrete was explored. Three specimens were tested under three-point loading. The deformations, strains in the GFRP beams, and slippages between the GFRP beams and concrete were recorded. The embedded GFRP beam enhanced the peak loads by 65% and 51% for the composite specimens with and without shear connectors, respectively. Moreover, a non-linear Finite Element (FE) model was developed and validated by the experimental results to conduct a parametric study. The peak loads of the composite specimen without shear studs increased by 14% and 31% and of the composite specimen with shear studs increased by 20% and 32% for the compressive strength of 35 MPa and 45 MPa, respectively.

Keywords

GFRP I-beam RC beams flexural tests deformations strains slip numerical analysis

Introduction

Glass fiber reinforced polymer (GFRP) I-beam is a new FRP composite manufactured by pultrusion technology. GFRP pultruded profiles exhibit low self-weight, low maintenance cost, and high durability, which have allowed them to become a competitive replacement as a primary structural material instead of steel reinforcement reinforced concrete. Compared with some typical FRP materials, such as GFRP sheets and GFRP bars, GFRP I-beam has higher compressive and flexural strength. Due to these superior material properties, I-beam or other pultruded profiles can be used directly as a standard construction product such as a cooling tower or offshore platform or applied in hybrid structures. In particular, GFRP I-beam was recommended for use in pedestrian bridge construction owing to its high corrosion resistance (Neagoe et al., 2015; Yuan and Hadi, 2017). There is an interesting potential for the use of GFRP-pultruded profiles in hybrid GFRP-concrete structural elements, either for new constructions or for the rehabilitation of existing structures (Hadi and Youssef, 2016).

Various aspects of RC composite beams consisting of a concrete block at the top and an I-beam of GFRP protrusions at the bottom were explored previously. The ultimate capacity, ductility, and stiffness were improved (Allawi and Ali, 2020). However, the web instabilities during loading were the disadvantages for this composite beam. Moreover, the fire performance of such a composite beam was poor, because the I-beam is exposed to air without the protection of concrete covering. The alternative form of the composite beam was proposed by encasing the GFRP I-beam in the RC cross-section. Previous experimental studies were conducted on these types of composite beams. Concrete slabs were tested to investigate the benefit of replacing steel reinforcing concrete slabs with pultruded GFRP grating sections (Hadi and Youssef, 2016). The local buckling failure of the embedded grid sections was reduced due to the confinement of concrete. The structural behavior of GFRP prismatic beams bonded at the top to concrete slabs using various GFRP (rectangular and I-sections) in composite action with the concrete was investigated (Evbuomwan, 2013). The plain GFRP I-section beam had a higher stiffness than the GFRP rectangular section beam but attained a lower ultimate failure load. The encased GFRP I-beam provided flexural strength as well as increased shear strength (Hadi and Yuan, 2017). Moreover, the slip between the concrete and the I-beam had decreased the load-carrying capacity. The connection between the FRP profile and concrete is the most essential component influencing flexural behavior (Xue and Zhang, 2014). The effect of shear connectors on the flexural behavior of composite beams with GFRP I-section was explored previously (Neagoe et al., 2015). The structural response of hybrid GFRP I-section mechanically connected to RC slabs with different cross-section geometry was experimentally investigated (encased GFRP I-section in concrete and GFRP I-section connected to RC slab by shear studs). The profile’s lateral confinement was provided with a more stiff mechanical connection and the slip values were half those without shear studs (Neagoe et al., 2015).

Finite Element (FE) models were developed to extensively study and estimate the behavior of FRP structures as well as their collapse using various failure criteria. Aziz and Tsai (1965) and Tsai and Wu (1971) developed the first generation of failure criteria for FRP structures. These simple criteria based on isotropic materials, such as metals, extended the scope of the von Mises criterion to composite materials and were capable of identifying the failure of a material point but were unable to influence that failure on the degradation of material stiffness (Duarte et al. 2017). The second generation of failure criteria was developed to influence the stiffness and strength of the FRP by degrading these properties beyond material failure (Hashin and Rotem, 1973; Hashin, 1980). The Hashin-based criterion contained a progressive damage model. The stiffness (constitutive matrix) of a damaged point was decreased by considering the previously calculated failure mode indexes (Duarte and Díaz Sáez, 2017).

This paper provides experimental and numerical investigations on the flexural performance of RC specimens composite with encased pultruded GFRP I-sections. The effect of using shear studs to improve the composite interaction between the GFRP beam and concrete was explored. Three specimens were tested under three-point loading. Moreover, a non-linear Finite Element (FE) model was developed and validated by the experimental results to conduct a parametric study.

Experimental program

Dimensions and Reinforcement

The total specimens’ lengths were 3000 mm with supports at 125 mm from each end of the specimen and a clear span of 2750 mm. The cross-sectional dimensions were 200 mm and 300 mm for the width and thickness, respectively. The details of the tested specimens are illustrated in Figure 1 and listed in Table 1. Two steel rebars with a diameter of 16 mm were provided as flexural reinforcement, as shown in Figure 1(a). Moreover, two steel rebars with a diameter of 10 mm were used to support the transverse reinforcement. The longitudinal and transverse reinforcement was designed according to ACI 318-19 (2019) to ensure flexural failure and prevent any premature shear failure. The first specimen, NR, was used as a reference specimen without GFRP I-beam. GFRP I-beams were used to strengthen the specimens CG and CGC, as shown in Figure 1(a). The geometrical dimensions of the GFRP Profile are shown in Figure 1(a). Shear studs were fabricated on the top surface of the GFRP I-beam in specimen CGC to enhance the composite interface between the GFRP beam and concrete in the compression zone since concrete is very weak in tension and works perfectly in compression. Moreover, the GFRP profile’s lateral confinement was provided with a more stiff mechanical connection under compression. These connectors had a diameter of 12 mm and a height of 70 mm. The longitudinal spacings between these studs were 375 mm to create a fully composite interaction between the GFRP beam and concrete according to AASHTO (2017). The minimum required number of studs to obtain the full shear connection was 16 studs with a spacing of 375 mm, as provided in the current study.

Figure 1.

Details of the tested specimens (a) Cross-sections of the specimens and (b) Elevation of the specimens. (Dimensions are in mm).

Material properties

A local company supplied the concrete with a 100 mm slump test result of the fresh concrete. Standard cubes (150 mm × 150 mm × 150 mm) and standard cylinders (150 mm × 300 mm) were tested to explore the mechanical properties of concrete. The cylindrical and cubic compressive strengths were 23.43 MPa and 29.29 MPa, respectively. The indirect tensile strength of concrete was determined by conducting splitting tests according to ASTM C496-96 (2017). Two standard concrete prisms (100 mm × 100 mm × 500 mm) were tested according to ASTM C78/C78M (2018) to determine the modulus of rupture of concrete. The splitting tensile strength and modulus of rupture were 2.46 MPa and 2.41 MPa, respectively. The modulus of elasticity was 20754.64 MPa as determined according to ASTM C469 (2010). Three steel rebars with 500 mm gauge length were tested according to ASTM A370-19 (2019). The yield stress, ultimate tensile strength, and elongation percentage of flexural steel reinforcement were 520.7 MPa, 687.1 MPa, and 23%, respectively.

Table 1.

Tested specimens scheme.

Specimens	Cross-section (mm)	Top bars		Bottom bars		Stirrups		Specimen weight (kg/m)	Type of encased beam	Connection
Specimens	Cross-section (mm)	Diameter (mm)	No. of bars	Diameter (mm)	No. of bars	Diameter (mm)	Spacing (mm)	Specimen weight (kg/m)	Type of encased beam	Connection
NR	200 × 300	10	2	16	2	10	125	503.89	N/A
CG	200 × 300	10	2	16	2	10	125	509.83	GFRP I-beam
CGC	200 × 300	10	2	16	2	10	125	510.51	GFRP I-beam	Shear studs

The Relative Density range of the GFRP beam was 1.6-2.1, as provided by the manufacturer (DURA composites, United Kingdom). To obtain the longitudinal and transverse compressive strengths of the GFRP I-section, fifteen coupon specimens of dimensions 10 mm × 12.7 mm × 38.1 mm were prepared and tested according to ASTM D695 (2015). Strain gauges were used in these tests to calculate the modulus of elasticity. The results of these tests are listed in Table 2. Moreover, according to ISO 527-4 (2019), tensile tests were conducted on coupon specimens with dimensions of 10 mm × 25 mm × 250 mm to determine the longitudinal tensile strengths of the flange and web, as listed in Table 2.

Table 2.

Compressive and tensile mechanical properties of the GFRP beam.

Mechanical Properties, Position, and Test Direction	Compression			Tension
	Flange	Web	Web	Flange	Web
	Longitudinal	Longitudinal	Transverse	Longitudinal	Longitudinal
Strength (MPa)	305.18	347.1	118.3	355	340
Ultimate strain (mm/mm) %	0.145	0.305	0.93	2.97	2.5
Elastic modulus (GPa)	28.5	25.68	6.8	19.5	20.1

The bond stress-slip characteristics between the GFRP beam and concrete

Push-out test was conducted to investigate the bond stress-slip characteristics between the GFRP beam and concrete. A GFRP I-beam of 350 mm in length was embedded in a concrete block with cross-sectional dimensions of 200 mm × 300 mm and a height of 300 mm, as shown in Figure 2(a). The bond length between the two components was 250 mm. A Linear Variable Differential Transformer (LVDT) was used to measure the relative displacement between the GFRP beam and concrete. The specimen was loaded at a rate of 0.1 mm/min using a displacement-controlled machine. The bond stress-slip curve of the tested specimen is shown in Figure 2(b).

Figure 2.

The bond stress-slip characteristics (a) Push-out test and (b) The bond stress-slip curve.

Test Setup and Instrumentations

An electric hydraulic jack with a 1000 kN capacity was used to apply the mid-span concentrated load (see Figure 3). The applied load had been increased gradually at an increment of 5 kN/minute. LVDTs were used to measure the mid-span deflections of the relative horizontal slip between the GFRP beam and concrete. Moreover, Strains on the top and bottom surfaces of concrete were measured using strain gages (S1 and S2) attached to the mid-spans of the tested specimens (see Figure 4(a)). For the GFRP beam, strains were measured at the top and bottom flange as well as the web at the mid-span (see Figure 4(b)).

Figure 3.

Test setup.

Figure 4.

Arrangement of strain gauges setup (a) for concrete; (b) for GFRP beam.

Experimental results

Load –deformation behavior

The experimental results are listed in Table 3 and the deformations of the tested specimens are illustrated in Figure 5. The tested specimens exhibited linear increment in the ascending region until yielding of the steel reinforcement. The embedded GFRP beam improved the stiffness of the composite specimens relative to the reference one. Moreover, the steel reinforcement of the composite specimens CG and CGC started to yield at loading levels of 38% and 50.3% higher than the reference specimen NR, respectively. Following that, the two curves increased with similar slopes until the peak loads were reached. The composite specimens experienced peak loads of 134.62 kN and 147.23 kN. The embedded GFRP beam enhanced the load-carrying capacity by 51% and 65% for the specimens CG and CGC, respectively. Providing shear studs improved the degree of shear connection between the GFRP I-beam and concrete, which increased the load capacity of specimen CGC by 9.5% in comparison to the case without shear studs.

Table 3.

Summary of the experimental results.

Specimen	Crack load Pcr (kN)	Deflection at crack load ∆cr (mm)	Yield load Py (kN)	Deflection at yield load ∆y (mm)	Peak load Pu (kN)	Deflection at peak load ∆u (mm)	Failure mode
NR	19.82	1.3	79.5	13.5	89.17	40	Yielding of tensile steel rebars and crushing in concrete
CG	20.62	0.445	110.4	20.1	134.62	66	Delamination and transverse shear failure of the GFRP I-beam
CGC	20.89	0.475	120.2	18.1	147.23	45	Delamination and transverse shear failure of the GFRP I-beam

Figure 5.

Load-deformation relationships of the tested specimens.

After reaching the peak loads of the tested specimens, the applied loads dropped down due to the excessive slip of the GFRP profile relative to the surrounding concrete as well as the delamination at the interfacial layers. These composite specimens exhibited enough ductility to reach maximum deformations exceeding 200 mm, two times the ultimate deformation of the reference specimen (see Figure 5). The ductility index was calculated as the ratio between the deformation (Δ_u) corresponding to the ultimate load and the yielding deformation (Δ_y) (Maghsoudi and Akbarzadeh Bengar, 2011). The ductility index of the tested specimens was 2.96, 3.28, and 2.48 for specimens NR, CG, and CGC, respectively. The GFRP beam increased the composite specimen CG ductility by 10.8% relative to the reference specimen. However, providing the shear connectors to specimen CGC decreased ductility by 16.2%.

Cracks propagation and failure modes

At the mid-spans, initial flexural cracks were formed when the cracking loads were 4% and 5% for specimens CG and CGC, respectively, relative to the reference one. The cracks propagated towards the loading point and through the longitudinal axis of each specimen. The strains in concrete at the same level as the longitudinal steel rebars (Figure 4) were used as a good indication of the yielding of steel rebars. Moreover, starting the non-linear relationships between the applied load and deformation (Figure 5) was a good indication of the yielding of these rebars. Yielding in steel reinforcement followed by crushing in concrete was the failure mode of the reference specimen NR (see Figure 6(a)).

Figure 6.

Failure modes of the tested specimens (a) the reference specimen NR, (b) the composite specimen CG, and (c) the composite specimen CGC.

For composite specimen CG, the crushing of concrete started at the load of 110 kN. The deflection started to increase rapidly when the applied load reached 134.62 kN. The failure started symmetrically as crushing in concrete. However, as the applied load increased, the specimen exhibited an excessive deformation, which caused a horizontal displacement of one of the roller supports and led to the fracture of the GFRP profile at the post-failure stage in an as asymmetrical behavior. On the other hand, the deflection increased rapidly for the composite specimen CGC when the applied load reached 147.23 kN. Figure 6(b) and (c) show the failure modes of the composite specimens. After reaching the peak load, concrete was crushed and local buckling of the steel rebars in compression occurred. After the crushing of concrete, the failure in the GFRP profile occurred. After failure and the test was stopped, the crushed concrete was removed directly at the mid-span (see Figure 7). It was visually observed that the failure in the GFRP profile was by delamination and transverse shear failure of the web and crushing in the top flange. The denomination and longitudinal shear failure of the web of the GFRP beam caused large slips between the GFRP profile and concrete.

Figure 7.

Buckling of the compression steel rebars and rupture of the GFRP profile (a) specimen CGC and (b) specimen CG. (The pictures were taken after removing the crushed concrete).

Strain measurements

Figure 8 illustrates the recorded strains as a function of the applied loads for the tested specimens. The positive values indicate tensile strains, and the negative ones are compressive strains. For specimen NR, the strain increased linearly up to 60 kN with maximum strain values of 0.0035 and −0.001 for concrete’s bottom and top faces, respectively. After that, the tensile reinforcement yielded, and the flexural cracks propagated, which led to reductions in the strain values, and then concrete crushing occurred. On the other hand, the composite specimen CG exhibited a maximum compressive strain of −0.004 at an applied load of 130 kN, while the maximum tensile strain was 0.0025 at 118 kN applied load when the tensile reinforcement yielded. The recorded strains in the top and bottom flanges of the GFRP beam increased linearly up to the yielding of the tensile steel rebars, and then the strains increased steadily up to failure (see Figure 8(b)). The same behavior was observed for the composite specimen CGC with shear connectors. However, higher levels of strains were recorded in the top and bottom flanges of the GFRP beam because there was no observed slipping due to the presence of the shear connectors.

Figure 8.

Recorded strains as a function of the applied loads for the tested specimens (a) specimen NR, (b) specimen CG, and (c) specimen CGC.

The axial strains as a function of the depth of the GFRP I-section are shown in Figure 9 at various loading levels for the composite specimens CG and CGC. The neutral axis (NA) position moved to the upper portion of the cross-section due to the concrete cracking and steel reinforcement yielding. After reaching failure, an increased slip strain developed between the GFRP profile and concrete, leading to the crushing of the GFRP. Therefore, the NA moved down, as shown in Figure 9(a). Whereas the top flange of the GFRP profile remained stable even after reaching the failure load due to the presence of the shear connectors, which provided a full connection with the concrete. Therefore, the position of the NA approximately remained at the same depth.

Figure 9.

Axial strains as a function of the GFRP beam depth at different loading levels (a) specimen CG and (b) specimen CGC.

Slippage and Split-up of the GFRP Profile

The relative horizontal slip between the GFRP beam and concrete was measured in two stages. In the first stage, two LVDTs were used to monitor the slippage. However, there was no significant variance in the slip between the two components. After reaching the ultimate load, the second slip stage started to measure the slip using a steel ruler. The slip increased significantly after the rupture of the GFRP profile, as shown in Figure 10. The slip for the composite specimen CG slowly increased at the roller support during the test, with a final slip about 35 mm out of the section at the top fractured piece and 15 mm inside the section at the bottom fractured piece, as illustrated in Figure 10(a). On the other hand, the ultimate slip was 20 mm for the composite specimen CGC, as illustrated in Figure 10(b). The connection between the GFRP beam and concrete is the most essential component influencing the beam capacity and slippage (Xue and Zhang, 2014). The main difference between the two specimens CG and CGC was that the profile’s lateral confinement was provided with a more stiff mechanical connection with slip values around half that without shear connectors.

Figure 10.

Slip and splitting of the GFRP profile (a) composite specimen CG and (b) composite specimen CGC.

An analytical approach was used to calculate the horizontal slip between the GFRP beam and concrete. The support rotation was calculated according to the following equation:

ϑ = \tan^{- 1} (\frac{∆}{L / 2})

(1)

Where:

ϑ

is the support rotation,

∆

is the maximum deflection of the tested specimen, and

L

is the clear span of the tested specimen. Based on the ultimate deflection of 275 mm and the horizontal distance of 1375 mm for specimen CG, the ultimate support rotation was calculated as 11.3^o, as shown in Figure 11. The horizontal displacement of 33 mm was obtained at the longitudinal splitting position of the GFRP beam (165 mm from the bottom fiber of the cross-section). The measured slip was 35 mm for the top part and it was close to the calculated one. The same approach was used for the composite specimen CGC. The support rotation was 8.89^o, as shown in Figure 12. The calculated horizontal slip was 24.2 mm, which was close to the measured slip of 20 mm.

Figure 11.

The rotation angle and the GFRP failure of the CG specimen.

Figure 12.

The rotation angle and the GFRP failure of the CGC specimen.

Finite Element model

Selection of Elements and Mesh Sensitivity

FE analysis using Abaqus (2019) software was carried out. Figure 13 illustrates the FE model for the tested specimen with GFRP I-beam and shear studs. Concrete was modeled using three-dimensional (3-D) linear 8-node elements (C3D8R). The 2-node linear 3-D truss elements (T3D2) were used to model steel reinforcement. The 8-node shell element S8R with reduced integration was used to simulate the GFRP I-beam. The two supported steel plates, the loading steel plate, and the shear connectors were modeled using general-purpose linear brick elements (C3D8R).

Figure 13.

FE mesh of the composite specimen CGC.

Mesh sensitivity analysis approaches were used to represent the load-displacement behavior of the beam specimens. Different mesh sizes (50 mm, 25 mm, and 20 mm) were utilized to characterize the mesh sensitivity of the simulation. The comparison showed that the refinement of the mesh size does not affect the displacement and load. Thus, a mesh size of 25 mm was chosen to be used in further simulations.

Material constructive models

The behavior of concrete under uniaxial tension was modeled using the tensile strain-softening of concrete, also referred to as the tension stiffening of steel bars, due to the tensile resistance of the concrete layer surrounding the bar simulates stress transfer between the reinforcement and concrete. The concrete damage plasticity (CDP) model, which is included in Abaqus, was employed to simulate the damage in concrete during loading. The behavior of concrete under tension was characterized by a linear–elastic relationship until reaching the tensile strength of concrete. After cracking initiation, a softening stress-strain relationship was used to represent the progressive convergence of micro–cracks. This latter relationship simulates concrete’s tensile cracking behavior and can be specified by a post-failure stress-strain relationship. The relationships of stress-strain curves of concrete under compression and tension are shown in Figure 14 (EN 1992-1-1, 2004). The different parameters of the CDP model were 31^o, 0.1, and 1.16 for the dilation (φ), eccentricity (ε), compressive strength to uniaxial pressure ratio biaxial (f_bo⁄f_co), respectively. Moreover, the coefficient K and viscosity parameters (μ) were 0.667 and 0.001, respectively.

Figure 14.

Relationships of stress-strain curves of concrete.

The bi-linear relationships were used to model the stress-strain relationship of steel reinforcement (Gadamchetty et al., 2016). The modulus of elasticity and yielding stress were used to define the linear elastic part. Otherwise, the ultimate stress f_u and plastic strain were used to define the strain hardening part. The degradation in the GFRP I-beam was modeled according to Hashin’s criteria (Hashin, 1980). According to the damage evolution law, the material stiffness of the GFRP I-beam degraded after satisfying the damage initiation criteria. The used mechanical properties of the GFRP material and progressive damage parameters are listed in Table 4 (Barbero et al., 2013; Barbero, 2013).

Table 4.

Mechanical properties and progressive damage parameters of the GFRP material.

	Definition		Value
Engineering elastic constants	Longitudinal modulus of elasticity (Ez)E		27.1 GPa
	Transverse modulus of elasticity (Ex = Ey)		6.8 GPa
	Transverse shear modulus of elasticity (gxy)		17.5 GPa
	In-plane shear modulus of Elasticity(Gzx = Gzy)		2.7 GPa
	Major Poisson’ ratio (υ_zx = υ_zy)		0.23
	Minor Poisson’ ratio (υ_xy)		0.1
Strength values	Tensile strength	Longitudinal	347.5MPa
	Tensile strength	Transverse	50 MPa
	Compressive strength	Longitudinal	326.14 MPa
	Compressive strength	Transverse	118.3 MPa
	Shear strength	Transverse	8.04 MPa
	Shear strength	In-plane	104.23 MPa
Damage evolution	Tensile fracture energy	Longitudinal	18.3
	Tensile fracture energy	Transverse	5
	Compressive fracture energy	Longitudinal	5.8
	Compressive fracture energy	Transverse	5.5

Boundary conditions

The experimental boundary conditions were adopted in the FE analysis as a simply supported beam. The first supports were constrained in Y- and Z-directions, representing hinged support. At the same time, the second support was constrained only in the Y-direction, which expressed roller support. The whole model was constrained in the X-direction. The full bond technique was assumed to simulate the connection between the concrete and steel rebars. However, the bond between the GFRP beam surface and the surrounded concrete was simulated using surface-to-surface contact pairs. The contact property was represented by the tangential behavior with a penalty friction formulation. The tangential shear stress was adopted from the push-out test at 0.422 MPa, and the friction coefficient was used equally at 0.55 according to the test of Hadi and Yuan [6]. However, the full bond between the shear studs and concrete was assumed.

Validations of the FE results

The conventional load-deformation curves for the tested specimens were used to validate the FE results, as shown in Figures 15 and 16. Initially, the FE deformations exhibited linear elastic behaviors with a higher stiffness than the experimental records. This difference in behavior could be attributed to the used constitutive models for materials and the full bond assumed between the concrete and steel rebars as well as GFRP I-beam. As the applied load gradually increased, cracks were formed and the non-linear deformation behaviors were obtained in good agreement with the experimental results. Table 5 summarizes the experimental and FE results regarding the maximum deflections and loads. The comparisons show that the difference between the maximum applied loads reached about 5.9% for the reference specimen NR, while they were 0.4% for the composite specimens.

Figure 15.

Comparisons between deformations of the FE and experimental results.

Figure 16.

Crack patterns of the experimental specimens and FE models (a) reference specimen NR, (b) composite specimen CG, and (c) composite specimen CGC.

Table 5.

Comparison between the experimental and FE results.

Specimen	NR		CG		CGC
Specimen	Max. Deflection (mm)	Max. Load (kN)	Max. Deflection (mm)	Max. Load (kN)	Max. Deflection (mm)	Max. Load (kN)
Experimental	109	89.17	260	134.6	215	147.26
FE	110	94.43	270	135.1	220	147.83
% Difference	0.92	5.9	3.8	0.37	2.3	0.38

The FE crack patterns for concrete were represented in Abaqus by defining the post-cracking damage properties for the CDP model by visualizing the tensile damage at the integration point DAMAGET. Figure 16 shows a closed agreement between the FE and experimental crack patterns, indicating that the FE program Abaqus effectively predicted the failure behavior of the tested specimens. Based on these comparisons, the proposed FE model can simulate the tested specimens with and without the GFRP beam. Also, the model simulated the slip progress and the de-bonding between the GFRP profile and concrete with very convergent behavior in failure, as shown in Figure 15. Therefore, the proposed model was used to evaluate a comparative parametric study.

Failure progressive in the GFRP profile

As shown in Figure 17, three stages of failure progression of the GFRP profile were specified in the FE results according to the criterion of Hashin’s damage for fiber tensile HSNFCCRT, fiber compressive HSNFCCRT, and tensile matrix damage HSNFTCRT. As seen in Figure 17(a), the damage began in the pultruded profile for the CGC specimen during the early loading stage at about 60% of the ultimate load capacity. The second stage of failure was at the peak load, where the damage was progressive in the tensile matrix of the GFRP’s web, as shown in Figure 17(b). After reaching the peak loads, fluctuations in the load-deflection curves occurred due to the formation of more tensile cracks in concrete, which experienced tensile stresses instead of compressive stresses at this level of loading. The last stage was at the maximum deflection, in which the accessive fracture accrued in the web and at the top and bottom flanges of the GFRP beam, as shown in Figure 17(c). As a result, the proposed FE model was able to capture the post behavior of the composite beams after reaching the ultimate capacities.

Figure 17.

FEM output of GFRP failure for specimen CGC; (a) at 60% of ultimate load; (b) at the ultimate load; (c) at the maximum deflection.

Bending moment capacities

In order to identify the bending moment capacities of the tested specimens and to exclude the effect of shear force, the validated FE model was used to conduct analysis for the tested specimens under the effect of four-point loading to create pure bending zones at the mid-span of the beams. Two concentrated loads at quarter spans were implemented in the FE analysis. With three-point loading, there was no pure bending section at every section of the beams. Therefore, there were combined effects of the applied bending moments and shear forces, which reduced the bending moment capacities. Comparisons between the two behaviors for each tested specimen are illustrated in Figure 18. Significant increases in the moment capacities after excluding the effect of shear force were obtained. The moment capacities were 206.2 kN, 234.4 kN, and 242.1 kN for the specimens NR, CG, and CGC, respectively. The effect of shear force reduced these capacities by 54%, 42%, and 39% for the specimens NR, CG, and CGC, respectively. These results confirmed the effectiveness of the GFRP beam in relieving the effect of the shear force in reducing the moment capacity of the composite beams. This effectiveness improved after adding shear studs to increase the composite interface between the GFRP beam and concrete.

Figure 18.

Comparisons between the behaviors for the tested specimen under three-point and four-point loading.

Parametric Study

Effect of the concrete compressive strength

The effect of the concrete compressive strength of (25 MPa, 35 MPa, and 45 MPa) on the flexural behavior of the tested specimens was investigated. The different compressive strengths were implemented in Abaqus through the stress-strain curves as well as the CDP model. The deformation behaviors of the analyzed specimens are shown in Figure 19. Enhancements in the peak loads, service load, and maximum deflections of the analyzed specimens were obtained as the concrete compressive strength increased, as listed in Table 6. The percentages were estimated for the reference concrete compressive strength of 25 MPa. Moreover, the service load was considered to be 60% of the maximum load.

Figure 19.

Effect of the concrete strength on the deformations of the tested specimens (a) refernce specimen NR, (b) composite specimen CG, and (c) composite specimen CGC.

Table 6.

Influence of the concrete compressive strength.

Specimens	Compressive Strength (MPa)	Peak Load (kN)	Increase in Max. Load (%)	Deflection at Service Load (mm)	Decrease in Deflection at Service Load (%)	Max. Deflection (mm)	Increase in Max. Deflection (%)
NR	25	94.43	—	4.30	—	110.00	—
	35	107.00	13.31	4.10	4.65	150.00	36.36
	45	115.64	22.46	3.90	9.30	167.16	51.96
CG	25	135.104	—	8.06	—	270.00	—
	35	154.316	14.22	7.70	4.47	290.00	7.41
	45	177.085	31.07	7.50	6.95	328.00	21.48
CGC	25	147.83	—	6.40	—	220	—
	35	178.15	20.51	6.14	4.06	230.92	4.96
	45	195.96	32.56	5.9	7.81	243.47	10.67

The peak load of the CG specimen was increased by 14% and 31% and for the CGC specimen was 20% and 32% for the compressive strength of 35 MPa and 45 MPa, respectively. However, the peak load of the reference specimen increased by 13% and 22%. The effectiveness of the concrete compressive strength in enhancing the peak load of the composite specimens relative to those without GFRP beams is confirmed by these results. Moreover, the mid-span deflections at the service loads were reduced by 4% and 7% for the concrete compressive strength of 35 MPa and 45 MPa, respectively.

Effect of the steel reinforcement ratio

In this section, the effects of the longitudinal steel reinforcement ratio on the flexural behavior and deflection were investigated. Three ratios of 0.44%, 0.78%, and 1.2% were implemented in this analysis. The steel reinforcement ratio was changed using two steel rebars with different diameters (12 mm, 16 mm, and 20 mm for the 0.44%, 0.78%, and 1.2% ratios, respectively). The deformation behaviors of the analyzed specimens are shown in Figure 20. The load-deflection curves confirmed that increasing the reinforcement ratio caused noticeable effects on the analyzed specimens' peak loads, service load, and maximum deflections.

Figure 20.

Effect of the steel reinforcement ratios (a) reference specimen NR, (b) composite specimen CG, and (c) composite specimen CGC.

Table 7 summarizes the effects of the reinforcement ratio on the peak loads and mid-span deflections. The reference reinforcement ratio used in this table was 0.44%. The percentage of the maximum load for the CG specimen increased by 38% and 66%, and the CGC specimen increased by 25% and 44% for the 0.78% and 1.2% reinforcement ratio respectively. However, the reference specimen without a GFRP beam increased by 40% and 83% for the reinforcement ratio of 0.78% and 1.2%, respectively. These results confirmed that the contribution of the steel reinforcement to enhance the flexural behavior was less effective when providing the GFRP beam to work compositely with concrete. However, the mid-span deflections at the service were increased by about 10% for composite specimens for the reinforcement ratio of 0.78% and 1.2%, respectively. At the same time, the increases were 30% and 53% for the reference specimen without the GFRP beam.

Table 7.

Summary of the effect of the reinforcement ratio.

Specimens	Reinforcement Ratio (%)	Peak Load (kN)	Increase in Max. Load (%)	Deflection at Service Load (mm)	Increase in Deflection at Service Load (%)	Max. Deflection (mm)	Increase in Max. Deflection (%)
NR	0.44	67.37	—	3.30	—	100	—
	0.78	94.43	40.16	4.30	30.30	110	10
	1.2	123.26	82.95	5.05	53.03	150	50
CG	0.44	97.8	—	7.3	—	207.4	—
	0.78	135.1	38.14	8.06	10.41	270	30.16
	1.2	162.73	66.38	8.1	10.96	297.6	43.16
CGC	0.44	118.24	—	6.0	—	200	—
	0.78	147.83	25.02	6.52	8.67	220	10
	1.2	169.76	43.57	6.6	10.0	236.7	18.35

Effect of the spacing of the shear connectors

In this section, the effect of the longitudinal spacing of the shear connectors on the flexural behavior of the composite specimen CGC is discussed. Three spacings of 250 mm, 375 mm, and 500 mm were implemented in the FE analysis to represent degrees of shear connection of 100% and 75%. As shown in Figure 21, the load-deflection curves confirmed no changes in the behaviors before reaching the ultimate capacities. These results are consistent with the experimental results since no significant variance in the slip between the GFRP beam and concrete during this stage of loading. However, slight changes existed in the behavior after reaching the ultimate capacities due to the differences in the slip at the final failure stage.

Figure 21.

Effect of the spacing of the shear connectors on the specimen CGC.

Conclusion

This paper provides experimental and numerical investigations on the flexural performance of RC specimens composite with encased pultruded GFRP I-sections. The following conclusions can be drawn based on the practical and FE results.

(1) The embedded GFRP beam enhanced the peak loads by 51% and 65% for the specimens CG and CGC, respectively. The improvement in the composite interaction between the GFRP beam and concrete by using shear connectors increased the peak load of specimen CGC by 9.5% relative to the composite specimen without shear connectors.

(2) The GFRP beam increased the ductility of the composite specimen CG by 10.8% relative to the reference specimen. However, providing shear connectors to specimen CGC decreased ductility by 16.2% relative to the reference specimen without a GFRP beam.

(3) The top flange of the GFRP profile remained stable even after reaching the failure load due to the presence of the shear connectors, which provided a complete connection with concrete. Therefore, the position of the NA approximately remained at the same depth.

(4) Delamination and longitudinal shear failure of the web of the GFRP beam were the modes of failure of the composite specimens, which caused large slips between the GFRP profile and concrete.

(5) The developed FE model provided good agreement with the experimental results regarding deformations and damaged patterns.

(6) The effect of the applied shear force reduced the bending capacities by 54%, 42%, and 39% for the specimens NR, CG, and CGC, respectively. These results confirmed the effectiveness of the GFRP beam in relieving the effect of the shear force in reducing the moment capacity of the composite beams. This effectiveness improved after adding shear studs to increase the composite interface between the GFRP beam and concrete.

(7) The peak load of the CG specimen was increased by 14% and 31% and for the CGC specimen was 20% and 32% for the compressive strength of 35 MPa and 45 MPa, respectively. However, the peak load of the reference specimen increased by 13% and 22%, which confirmed the effectiveness of the concrete compressive strength in enhancing the flexural behavior of the composite specimens.

(8) The contribution of the steel reinforcement to enhance the flexural behavior was less effective when providing the GFRP beam to work compositely with concrete. However, the service mid-span deflections increased by about 10% for composite specimens for the reinforcement ratio of 0.78% and 1.2%, respectively.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Ayman El-Zohairy

References

Abaqus (2019) Abaqus 6.2: Computer Software for Finite Element Analysis; Dassault Systems Simulia: Johnston. USA: Rhode Island.

ACI Committee 318 (2019) Building Code Requirements for Structural Concrete. USA: American Concrete Institute: Farmington Hills, MI, pp. 318–319.

AASHTO (2017) ASSHTO LRFD Bridge Design Specifications. 8th edition. AASHTO

Allawi

Ali

(2020) Flexural behavior of composite GFRP pultruded I-section beams under static and impact loading. Civil Engineering Journal 6(11): 2143–2158. DOI: 10.28991/cej-2020-03091608.

ASTM-C496/C496M (2017) Standard Test Method for Splitting Tensile Strength of Cylindrical Concrete Specimens. West Conshohocken, PA: ASTM International.

ASTM C78/C78M (2018) Standard Test Method for Flexural Strength of Concrete (Using Simple Beam with Third-Point Loading). West Conshohocken, PA: ASTM International.

ASTM- C469/C469M (2010) Standard Test Method for Static Modulus of Elasticity and Poisson’s Ratio of Concrete in Compression. West Conshohocken, PA: ASTM International.

ASTM A370 (2019) Standard Test Methods and Definitions for Mechanical Testing of Steel Products. West Conshohocken, PA: ASTM International.

ASTM D695 (2015) Standard Test Method for Compressive Properties of Rigid Plastics. West Conshohocken, PA: ASTM International.

10.

Azzi

Tsai

(1965) Anisotropic strength of composites - investigation aimed at developing a theory applicable to laminated as well as unidirectional composites, employing simple material properties derived from unidirectional specimens alone. Experimental Mechanics 5(9): 283–288. DOI: 10.1007/BF02326292.

11.

Barbero

(2014) Finite Element Analysis of Composite Materials Using Abaqus. Boca Raton, FL: Taylorfrancis. DOI: 10.1201/b14788.

12.

Barbero

Cosso

Roman

, et al. (2013) Determination of material parameters for Abaqus progressive damage analysis of E-glass epoxy laminates. Journal of Composites Part B: Engineering 46: 211–220. DOI: 10.1016/j.compositesb.2012.09.069.

13.

Duarte

APC

Díaz Sáez

Silvestre

(2017) Comparative study between XFEM and Hashin damage criterion applied to failure of composites. Thin-Walled Structure Journal 115: 277–288. DOI: 10.1016/j.tws.2017.02.020.

14.

Evbuomwan

NFO

(2013) Flexural behaviour of GFRP prismatic beams in composite action with concrete. In: Research and Applications in Structural Engineering, Mechanics and Computation, Chapter: 267. DOI: 10.1201/b15963-289.

15.

Gadamchetty

Pandey

Gawture

(2016) On practical implementation of the ramberg-osgood model for FE simulation. SAE International Journal of Materials and Manufacturing 9(1): 200–205. DOI: 10.4271/2015-01-9086.

16.

Hadi

MNS

Youssef

(2016) Experimental investigation of GFRP-reinforced and GFRP-encased square concrete specimens under axial and eccentric load, and four-point bending test. Composites for Construction 20(5): 11. DOI: 10.1061/(ASCE)CC.1943-5614.0000675.

(2017) Experimental investigation of composite beams reinforced with GFRP I-beam and steel bars. Construction and Building Materials Journal 144: 462–474. DOI: 10.1016/j.conbuildmat.2017.03.217.

18.

Hashin

(1980) Failure criteria for unidirectional fiber composites. Journal of Applied Mechanics 47(2): 329–334. DOI: 10.1115/1.3153664.

19.

Hashin

Rotem

(1973) A fatigue failure criterion for fiber reinforced materials. Journal of Composite Materials 7(4): 448–464. DOI: 10.1177/002199837300700404.