Flexural Performance of Prefabricated Steel-Fiber-Reinforced Concrete Wall Panels: Finite Element Analysis

Abstract

This study proposes and evaluates a prefabricated steel-fiber-reinforced concrete (SFRC) wall-panel system for flexural performance. Material tests were used to calibrate and validate a compression stress–strain model for SFRC with good predictive accuracy. Finite element analyses quantify the panels’ flexural capacity and the effects of wythe thickness, connector spacing, and connector layout. Results show that adding glass-fiber grids produces synergy with steel fibers, improving the composite wall system’s flexural performance. Relative to plain-concrete panels, SFRC panels exhibit 29.5% lower peak strain and 27.2% higher peak stress. FE analyses indicate that shortening the connector length reduces flexural capacity. Within the studied range, a 200 mm connector spacing delivers the best structural performance. A full-height connector layout is recommended to ensure structural integrity.

1. Introduction

Prefabricated components are a major research focus [1,2,3,4,5]. Owing to shorter construction time, improved quality, and reduced energy use, prefabricated systems are widely adopted across diverse structural applications [6,7,8,9,10]. As key building envelope components, concrete wall panels are evolving toward lower weight, higher energy and environmental performance, and greater load-carrying capacity and toughness [11,12]. Common insulated wall types include interior insulation, exterior insulation, and sandwich panels.

Prefabricated concrete sandwich panels comprise two precast concrete wythes and a core insulation layer connected by mechanical connectors. These panels deliver economic, social, and environmental benefits and align with the trends of industrialized construction and building energy efficiency [13]. The flexural performance of sandwich wall panels has received extensive attention, and numerous studies have been reported. Yu et al. [14] developed an ECC–RC composite sandwich panel and conducted four-point bending tests and FE analyses; ECC markedly improved flexural performance, and maximum crack width remained well below the serviceability limit. Mugaged Amran et al. [15] demonstrated that precast foamed concrete sandwich panels can serve as an alternative to conventional concrete slab systems. Morcous et al. [16] experimentally investigated composite UHPC sandwich wall panels and found that UHPC can enhance panel performance while reducing thickness and mass. Mahdi et al. [17] conducted experimental and theoretical studies on a new composite sandwich wall panel, showing that partially replacing the core with EPS reduces self-weight without compromising flexural capacity. Meng et al. [18] performed experimental and mechanical studies on FRP-fabric-reinforced UHPC panels and reported that GFRP fabrics can further increase flexural capacity. Abeysinghe et al. [19] examined an innovative hybrid composite floor plate system (HCFPS); it showed high stiffness and excellent flexural performance, with self-weight reduced by 50–70% compared with conventional slabs. These studies primarily optimize concrete materials in composite panels, advancing lighter and stronger designs.

Connectors in sandwich panels tie the exterior and interior wythes into a single unit and carry the exterior wythe’s self-weight as well as wind and seismic actions. Therefore, the composite action of sandwich wall panels primarily depends on the type and layout of the connectors [13]. O’Hegarty et al. [20] tested and performed FE analyses on FRC sandwich panels with foam insulation and FRP connectors; all panels failed in ductility, showing high strain energy and large displacements before ultimate failure. Huang et al. [21] confirmed that FRP connectors enhance the composite action of sandwich wall panels. Shin et al. [22] reported that composite action depends on the number of shear connectors; increasing the number improves load-carrying and deformation capacities. Jiang et al. [23] examined the effects of the number and layout of W-shaped steel–GFRP shear connectors on performance. Gombeda et al. [24] experimentally showed that increasing the amount of shear reinforcement improves overall flexural capacity and ductility.

This study combines low-conductivity, low-density XPS with SFRC, incorporates glass-fiber grids, and employs steel-truss connectors to propose a prefabricated SFRC wall-panel system. A constitutive model for SFRC is calibrated from material tests, and FE analyses quantify the effects of wythe thickness, connector spacing, and connector layout on the panels’ structural response. Design recommendations are provided to inform subsequent design and application.

2. Experiment on Steel-Fiber-Reinforced Concrete

An experimental program was conducted to characterize the mechanical properties of SFRC with varying fiber contents. Four groups (12 specimens in total) were cast to evaluate SFRC compressive strength versus fiber content.

2.1. Experimental Program

2.1.1. Materials

Raw materials for SFRC included cement, water, fine aggregate (sand), coarse aggregate, a water-reducing admixture, and steel fibers. The mixtures incorporated P.O 42.5R cement, Grade II fly ash, and silica fume as the primary cementitious materials (Sibirskiy Cement Holding Co., Krasnoyarsk, Russia). Material properties are summarized in

Table 1. Technical parameters of cement.

Specific Surface Area/(m2/kg)	Setting Time/min		Compressive Strength/MPa		Flexural Strength/MPa
	Initial	Final	3 Days	28 Days	3 Days	28 Days
350	95	150	28.5	49.3	5.6	7.9

,

Table 2. Chemical composition of fly ash (%).

SiO2	Al2O3	CaO	Fe2O3	TiO2
58.13	25.14	5.18	7.12	4.43

and

Table 3. Chemical composition of silica fume (%).

Loss	Al2O3	CaO	SiO2
2.52	1.57	0.18	95.73

.

Straight steel fibers (Bekaert Lipetsk LLC, Lipetsk, Russia) were used; geometry and properties are shown in Figure 1 and

Table 4. Properties of steel fibers.

Length/mm	Diameter/mm	Aspect Ratio	Tensile Strength/MPa	Elastic Modulus/GPa
20	0.8	25	1200	200

. Specimens were cast by pouring the fresh mix into molds and compacting them on a standard vibrating table to achieve density; fiber orientation in the matrix was assumed to be random and three-dimensional.

2.1.2. Mix Design

According to ACI 544.1R-96 [25], the fiber volume fraction in concrete is typically not greater than 2%. Consequently, four mixtures were designed with fiber volume fractions of 0%, 0.5%, 1.0%, and 1.5%. The mixture proportions are listed in

Table 5. Mix proportions by group/(kg/m³).

Group	Water	Cement	Sand	Aggregates	Steel Fiber Volume Fraction (%)
1	164	430	780	1024	0%
2	164	430	780	1024	0.5%
3	164	430	780	1024	1%
4	164	430	780	1024	1.5%

.

2.2. Experimental Results

2.2.1. Failure Modes

Figure 2 illustrates the failure modes of the specimens. Plain concrete specimens (0% fiber) developed multiple diagonal surface cracks at early loading stages. As loading continued, crack widths increased. This response was characteristic of brittle failure. Compared with plain concrete, SFRC specimens exhibited delayed crack initiation under compression. SFRC specimens ultimately exhibited fewer cracks, indicating effective fiber bridging and crack restraint.

2.2.2. Experimental Results and Analysis

In accordance with GB/T 50081–2019 [26], the reported strength is the arithmetic mean of three replicate specimens, rounded to the nearest 0.1 MPa. As summarized in

Table 6. Compressive performance of cube specimens.

Specimen ID	Steel Fiber Volume Fraction (%)	Compressive Strength (MPa)	Average Compressive Strength (MPa)	Strength Growth Rate (%)
1-1	0	50.6	51.1
1-2		50.0		—
1-3		52.7
2-1	0.5	53.5	54.0	5.6
2-2		52.6
2-3		55.9
3-1	1.0	54.7	55.1
3-2		55.7		7.8
3-3		54.8
4-1	1.5	56.2	56.0
4-2		55.5		9.5
4-3		56.1

, the cube compressive strength increased with the steel-fiber volume fraction (V_f). Relative to V_f = 0.0%, the increases were 5.7%, 7.8%, and 9.4% for V_f = 0.5%, 1.0%, and 1.5%, respectively. The strength gain exhibited diminishing returns with increasing V_f. This trend is consistent with prior reports [27,28]. The conclusion regarding fiber content is confined to material-level results for the tested mix design and was used solely to calibrate the compressive constitutive parameters of SFRC; it is not extrapolated to member-level flexural analysis.

2.2.3. Fitting of Axial Compression Stress–Strain Curves

Following the mix proportions in Table 5, four groups (12 total) of standard 150 × 150 × 300 mm prismatic specimens were prepared. Per CECS 13:2009 [29], a YAW-5000 (Jinan Hansen Precision Instrument Co., Ltd., Jinan, China) electro-hydraulic servo testing machine (500 t) was used to record complete SFRC stress–strain responses under uniaxial compression. Figure 3 shows representative axial stress–strain curves for the four mix groups.

Building on Guo Zhenhai’s concrete model, a piecewise axial compression stress–strain law was adopted and calibrated for the SFRC mixture used herein, as given in Equation (1) [30]:

$$\left\{\begin{array}{l}y=ax+(3-2a)x^2+(a-2)x^3\hspace{1em}0\le x\le 1\\ y={\displaystyle \frac{x}{b{(x-1)}^2+x}}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em} x>1\end{array}\right.y=\sigma /f_{\mathrm{c}}, x=\epsilon /{\epsilon }_{\mathrm{c}}$$

(1)

where σ is compressive stress; f_c is the uniaxial compressive strength; ε is axial compressive strain; ε_c is the peak strain at f_c; parameters a and b control the pre-peak curvature and post-peak softening, respectively.

The experimental stress–strain data are fitted to Equation (1) via nonlinear regression, and the fitted curves are shown in Figure 4. The fitted curves agree well with the test data and capture the constitutive response of the concrete used.

To quantify relations between the model’s shape parameters and key material properties, least-squares regression was used to derive predictive expressions for parameters a and b as functions of compressive strength and fiber variables:

$$a=-5.21325+5.3839f^{0.07352}+1.170615{\lambda }_{\mathrm{f}}$$

(2)

$$b=-1.4588+1.82517f^{0.09542}+2.81356{{\lambda }_{\mathrm{f}}}^{2.78533}-1.9545{\lambda }_{\mathrm{f}}$$

(3)

$${\lambda }_{\mathrm{f}}={\rho }_{\mathrm{f}}{l}_{\mathrm{f}}/d_{\mathrm{f}}$$

(4)

where f is the uniaxial compressive strength, λ_f is a fiber characteristic value, ρ_f is the steel-fiber volume fraction (by volume), and l_f/d_f is the fiber aspect ratio. For zero fiber content (λ_f = 0), the calibrated parameters are a = 1.50728 and b = 2.51253.

For validation, the fitted shape-parameter relations were used to compute a and b, which were substituted into Equation (1). The resulting curves were compared with the experimental data (Figure 5). Statistically, the predicted-to-measured strength ratio had a mean of 0.97, a standard deviation of 0.025, and a coefficient of variation of 0.017. These validations indicate that the proposed shape-parameter relations and the calibrated constitutive model reproduce the SFRC response with high accuracy. The model is suitable for nonlinear FE analysis (ABAQUS 2020) of SFRC wall panels.

3. Finite Element Analysis

3.1. Wall-Panel Design

Three wall-panel types were designed: (i) SFRC panels, (ii) SFRC panels with glass-fiber grids, and (iii) plain-concrete panels with glass-fiber grids. Plain-concrete panels with glass-fiber grids served as the control. These three panel types isolate the contributions of steel fibers and glass-fiber grids to sandwich-panel mechanics and enable comparison with plain-concrete and non-grid configurations at the macro scale. To ensure consistency, all finite element (FE) models (ABAQUS 2020) used panels with overall dimensions of 4000 × 3000 × 150 mm (length × height × thickness). The glass-fiber grids used a uniform 25.4 mm square mesh. Per Chinese standard 08SJ110-2 [31] for prefabricated concrete façades, these dimensions fall within the full-span panel classification. The vertical edges incorporate 20 mm-thick concrete ribs that protect the integral thermal-insulation layer. The three-dimensional configuration of the wall-panel assembly is illustrated in Figure 6.

3.2. Constitutive Models

3.2.1. Steel-Fiber-Reinforced Concrete (SFRC)

The concrete damage plasticity (CDP) model was adopted to represent the constitutive response of concrete. CDP parameters were as follows: dilation angle ψ = 30°, eccentricity e = 0.1, f_b0/f_c0 = 1.16, ratio on the tensile/compressive meridians K_c = 0.667, and viscosity parameter μ = 0.005. The calibrated compressive stress–strain law from tests (Section 2.2.3) was input as the material model in the FE simulations (ABAQUS 2020). For tension, the uniaxial tensile stress–strain curve proposed by Moradi et al. [32] was adopted. Extensive studies show that fibers enhance concrete tensile capacity, toughness, and crack resistance [33,34,35]. Fibers also enable post-cracking load carrying [36]. Accordingly, fiber pull-out behavior is critical to wall-panel flexural performance. Given the limitations of macro-scale models, only the fiber-induced enhancements in compressive and tensile behavior were considered in the constitutive model, while explicit pull-out mechanisms were not modeled. This simplification may overpredict stiffness and underestimate crack localization [37].

3.2.2. Steel Trusses and Glass-Fiber Grids

Steel-truss members (CRB550 grade) were modeled with the following properties: elastic modulus E_s = 190 GPa, Poisson’s ratio ν = 0.30, and 0.2% proof strength f_y = 510 MPa. An elastic–perfectly plastic law was adopted due to the absence of a distinct yield plateau:

$$\sigma =\left\{\begin{array}{l}{E}_{\mathrm{s}}\epsilon \hspace{1em}\hspace{1em}0\le \epsilon \le {\epsilon }_{\mathrm{y}}\\ f_{\mathrm{y}}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em} \epsilon >{\epsilon }_{\mathrm{y}}\end{array}\right.$$

(5)

where E_s is the steel elastic modulus, ε is strain, ε_y is yield strain, and f_y is yield strength.

The glass-fiber grid (type GGA5050) was modeled with a uniform 25.4 mm square mesh. The longitudinal and transverse breaking strengths were 50 kN/m; elongation at break ≤ 3%; elastic modulus E = 35 GPa; and Poisson’s ratio ν = 0.25. Glass-fiber grids exhibited approximately linear-elastic behavior up to brittle rupture [38,39]; therefore, an ideal elastic model was used.

3.3. Element Types and Meshing

Concrete and XPS insulation were modeled with linear eight-node, reduced-integration solid elements (C3D8R). The global mesh size was 100 mm. Steel trusses were modeled with beam elements at a 25 mm nodal spacing. Glass-fiber grids, embedded in concrete due to their small thickness, were modeled with T3D2 truss elements (25 mm mesh).

3.4. Boundary Conditions and Contacts

Figure 7 illustrates the support configuration between the panel and the primary structure. Boundary conditions were simplified to one roller support and one pin support in the FE model [40,41]. Adhesion–debonding effects at supports and at insulation–concrete interfaces were neglected, and a constrained (tied) contact formulation was applied. Prior studies indicate negligible slip between steel trusses or glass-fiber grids and concrete [42,43]. Accordingly, the steel trusses and glass-fiber grids were coupled to the concrete using the embedded-element technique.

3.5. Finite Element Model Verification

To verify the FE model, the bending test of the wall-panel specimen QB1 reported in [44] was simulated. QB1 comprised plain-concrete wythes, XPS insulation, and steel-truss plus GFRP connectors, with welded wire mesh in concrete; it was selected for validation due to similarity to the present sandwich panel system. The test panel measured 2800 × 1860 × 150 mm, with inner and outer concrete slabs each 50 mm thick. Concrete was C50, and the XPS insulation layer was 50 mm thick. The panel configuration is shown in Figure 8.

Figure 9 compares the experimental and FE load–displacement curves. The simulated response agrees closely with the test data, validating the model for assessing wall-panel flexural behavior.

3.6. Analysis of Typical Members

3.6.1. Load–Displacement Curves

Three wall-panel types are considered: SFRC with glass-fiber grids (G-FCSP), plain concrete with glass-fiber grids (G-CSP), and SFRC without grids (FCSP). All typical panel models used a 50 mm concrete wythe, a 50 mm insulation layer, 500 mm connector spacing, and a longitudinal full-height connector layout. Figure 10 shows the load–displacement curves of the three wall-panel types. The 50 mm marker in Figure 10 indicates the code-based limit deflection of the panels (serviceability limit state, GB 50204-2015) [45].

Figure 11 illustrates the division of the load–displacement curve of model G-FSCP into different stages. Initially, the panels respond elastically, and the load–displacement curves are linear. The initial stiffness of all three panels is nearly identical, indicating that the grids have no significant effect on initial stiffness. Based on the tensile-stress criterion, the cracking loads are 94.4 kN (G-FCSP), 89.6 kN (FCSP), and 70.1 kN (G-CSP), confirming the contribution of steel fibers to flexural capacity. At the serviceability-limit deflection, the loads are 158.4 kN (G-FCSP), 143.1 kN (FCSP), and 137.3 kN (G-CSP); G-FCSP exceeds FCSP and G-CSP by 10.7% and 15.4%, respectively, demonstrating synergy between the glass-fiber grids and steel fibers.

3.6.2. Contour-Plot Analysis

A comparative analysis of damage evolution in G-FCSP, FCSP, and G-CSP at the cracking- and ultimate-load states is presented to elucidate the influence of glass-fiber grids and steel fibers on failure modes. Figure 12, Figure 13, Figure 14, Figure 15, Figure 16 and Figure 17 present contour plots of S33 stress, LE33 strain, tensile damage, and truss von Mises stress for the three wall-panel types at both states. At the cracking load, G-FCSP exhibits pronounced local tensile stress and strain concentrations (S33, LE33) at the mid-span of the bottom face. Meanwhile, the magnitude and extent of tensile damage at the bottom face are smaller than in FCSP and G-CSP. With continued loading, the concrete damage in G-FCSP propagates slowly from mid-span toward both sides of the bottom face. Compared with FCSP and G-CSP, G-FCSP develops a finer and narrower tensile-damage band, indicating effective crack-width control by the glass-fiber grid and high-toughness SFRC. At the ultimate load, the bottom-face damage zone in G-FCSP further enlarges, and the mid-span segment of the truss bottom chord yields.

S33 stress, LE33 strain, and truss von Mises stress are quantified for the three panel types at the cracking and ultimate load states. As shown in Figure 12a,b, Figure 13a,b, Figure 14a,b, Figure 15a,b, Figure 16a,b, Figure 17a,b, at cracking, the G-FCSP cracking stress exceeds that of FCSP and G-CSP by 5.2% and 26.2%, respectively; the corresponding cracking strain is 5.8% and 26.7% lower. At the ultimate load, the maximum tensile stress in G-FCSP is the largest, 5.4% and 24.8% higher than that of FCSP and G-CSP, respectively. These results indicate that the combined use of steel fibers and glass-fiber grids effectively reduces deformation and alleviates local strain concentration.

4. Parameter Analysis

This section analyzes the effects of wythe thickness and connector arrangement on cracking load, ultimate load, and flexural stiffness to inform engineering design. Four concrete wythe thicknesses were considered: 50, 55, 60, and 65 mm. Connector spacings were 200, 350, 500, and 650 mm. As shown in Figure 18, connector arrangements are denoted A1–A4 and illustrated in the schematic. A1 used full-height vertical connectors (length 2800 mm); A2 used segmented vertical connectors (1200 mm per row at 200 mm spacing); A3 followed A2 with added horizontal connectors between two rows (total horizontal length 3600 mm); and A4 used three rows of 800 mm vertical connectors plus two rows of horizontal connectors.

Figure 19 presents load–deflection curves for FCSP and G-FCSP panels under different parameter sets. The analysis focuses on the cracking load and ultimate load extracted from these curves.

4.1. Cracking and Ultimate Load

To quantify the enhancement from glass-fiber grids, Figure 20 compares the cracking and ultimate loads of G-FCSP and FCSP panels under various parameters. Figure 20a,b, Figure 20c,d, and Figure 20e,f compare the two panel types for wythe thickness, connector spacing, and connector arrangement, respectively. In Figure 20a,b, G-FCSP shows average increases of 6.9% in cracking load and 10.5% in ultimate load relative to FCSP across the thickness range. With increasing thickness, the ultimate load of G-FCSP rises by 13.5%, 11.5%, and 12.2%, respectively. Glass-fiber grids act synergistically with steel trusses to share tensile demand. When self-weight and cost are not governing design, increasing wythe thickness effectively enhances flexural performance. In Figure 20c,d, for the same spacing, G-FCSP shows average increases of 4.2% in cracking load and 12.1% in ultimate load compared with FCSP. Reducing spacing increases the ultimate load of G-FCSP. Densifying spacing from 500 mm to 200 mm yields a maximum ultimate-load gain of 20.6% (compared with the 14.4% at 350 mm). The improvement arises because smaller spacing increases connector participation, enhancing ultimate capacity. FE results indicate that a 200 mm spacing delivers superior flexural performance via more effective material synergy. In Figure 20e,f, comparing A1 and A2 shows no notable changes in cracking or ultimate load. Thus, changing truss length alone does not materially improve panel bending performance. A3 adds transverse trusses (3600 mm total), increasing the cracking and ultimate loads by 7.8% and 4.3%, respectively, relative to A2. A4 (three longitudinal rows plus two transverse rows) reduces the cracking and ultimate loads by 8.6% and 4.5% relative to A1. Overly short connectors disrupt continuous load transfer, reducing both cracking and ultimate loads. Therefore, a full-height connector layout is recommended.

4.2. Stiffness at Cracking

As shown in Figure 21, the thickness strongly affects overall panel stiffness. Stiffness increases with thickness. For G-FCSP, relative to 50 mm, 5 mm step increments raise stiffness by 12.4%, 24.2%, and 40.5%. FCSP shows a similar trend. Greater stiffness reduces bending deformation for a given load, so a higher load is required to reach cracking deformation. Accordingly, increased stiffness indirectly elevates both cracking stress and ultimate load. Reducing spacing from 500 mm to 350 or 200 mm increases stiffness by 4.3%/13.5% (FCSP) and 4.5%/13.3% (G-FCSP). Although wider spacing (fewer connectors) lowers elastic stiffness, the overall effect is moderate. Altering layout, e.g., adding horizontal connectors, can improve elastic stiffness, though the effect is limited. A2 (two vertical rows) has a negligible effect relative to A1, with a 4.5% reduction in elastic stiffness. Consequently, a full-height connector distribution is preferred to ensure structural integrity.

5. Conclusions

This study develops a precast SFRC composite wall-panel system that attains superior flexural capacity relative to plain-concrete panels. Based on material tests and finite element (FE) analyses, the following conclusions are drawn:

• At a 1% steel-fiber volume fraction, the compressive strength increases by 7.8% relative to plain concrete. The strength gain diminishes at higher fiber contents, indicating 1% as the optimum content within this study. A compressive stress–strain law for the present SFRC mix was formulated based on Guo Zhenhai’s concrete model. The law agrees well with the measured curves. The fitted shape-parameter relations and the segmented law accurately predict the compressive behavior of SFRC.
• Analyses of typical panels show that steel fibers enhance overall integrity. Steel fibers provide effective restraint and delay crack initiation. At lower strain levels, SFRC panels sustain higher stress. Relative to plain-concrete panels, SFRC panels exhibit a 29.5% lower peak strain and 27.2% higher peak stress.
• Increasing wythe thickness markedly enhances panel load capacity. From 50 to 65 mm, the cracking load increases by 40.4%, with a concurrent 40.5% rise in panel stiffness. At the same thickness, G-FCSP exceeds FCSP by 6.9% in cracking load and 10.5% in ultimate load, confirming the synergy between grids and steel fibers. Greater wythe thickness is recommended where self-weight and cost permit. Connector spacing has a significant effect on overall stiffness. Reducing spacing from 500 to 200 mm raises G-FCSP ultimate load by 20.6%, mainly by mobilizing the steel-truss tension system. A full-height vertical connector layout is recommended to ensure structural integrity.

Further Studies

Develop a macro–micro multiscale modeling framework based on fiber volume fraction and pull-out test data. Explicitly account for fiber orientation and fiber–matrix interfacial debonding/bridging to capture their effects on crack propagation and energy dissipation, thereby providing a mechanistic basis for member-level behavior. In parallel, assess how glass-fiber grids and steel-truss connectors influence the spatial distribution of fibers.

Subject to a fixed steel-quantity constraint, extend the parametric study to truss bar diameter/grade and anchorage length, and perform parametric optimization to identify connector layouts that maximize wall-panel performance.

Conduct component tests to validate and calibrate the numerical model and the proposed design recommendations, thereby demonstrating performance advantages and applicability.