Experimental Investigation of Composite Beams with Asymmetric Steel Sections Featuring Expanded Web Openings

Steel-concrete composite beams are extensively used in structural engineering due to their high strength-to-weight ratio and efficient structural performance. Among these systems, castellated or expanded web steel beams have gained significant attention as they provide increased section depth and flexural capacity without a corresponding increase in material consumption [1]. Compared with traditional I-section building components, web-opening steel beams offer numerous advantages in contemporary structures. One of the advantages is the ability to integrate utilities such as ducting, electrical wiring, or piping without compromising the beam's structural integrity, as well as enhanced visual appeal and a better strength-to-weight ratio. Notably, this beam type is favored in projects with extensive spans, such as stadiums, multistory buildings, and bridges, due to its ability to facilitate installation and optimize space usage [2]. Silva et al. [3] and Morkhade and Gupta [4] introduced the Vierendeel mechanism in steel beams, including web holes through analytical and numerical methodologies. The simply supported castellated beams (CB) under a central focused load were the subject of a study by Zirakian [5]. Many researchers, including Khan et al. [6] and Osmani et al. [7], have presented finite element analyses of CBs with and without holes. When holes are punched into the steel, the beams become weaker, their deflection increases, their shear capacity drops, their failure modes alter, and more intricate design considerations are required. The cost of manufacturing steel beams with extended webs is higher than that of standard webbed beams. Additionally, opening the steel beams introduces residual stresses that may affect the structural behavior of the beams [8-11]. Asymmetrical steel sections in composite concrete beams are an efficient means of enhancing structural performance and accommodating complex loading scenarios. Welded plate sections facilitate the creation of tailored geometries appropriate for asymmetric behavior. The use of welded steel plates along or within the web has been demonstrated to markedly improve stiffness, shear resistance, and load-carrying capacity. This is especially true in beams featuring openings or non-uniform stress distribution. In line with this, the composite interaction between steel and concrete enhances flexural performance and overall structural behavior [12]. A numerical study of the structural performance of composite and non-composite concrete castellated double-channel beams using alternate opening strategies was conducted by Abbas and Alshimmeri [13]. Ahmed and Said [14] investigated the use of additional plates to deepen wide-flange beams made of hot-rolled steel. Furthermore, an ABAQUS-created finite element numerical model was used to verify the experimental results. A study by Naji and Al-Shamaa [15] assessed the structural behavior of castellated steel beams, focusing on how design parameters, such as the number of apertures and the angle of cut, affect stiffness, deflection, and load-bearing capacity. Thus, by transforming the web into circular, rectangular, or hexagonal openings, CBs aim to improve the strength-to-weight ratio of conventional I-beams. This, in turn, increases load-bearing capacity and reduces weight while preserving or increasing stiffness, making them a good choice for long-span structures.

In the current study, the experiment investigates the expansion ratio and performance differences of innovative expanded web plates in asymmetric steel-profile concrete composite beams under static loading.

3.1 Properties of hardened concrete

Table 3 presents the characteristics of hardened concrete. The cube compression test followed BS 1881: Part 116, while the cylinder compression test and the splitting tensile strength were conducted in accordance with ASTM C39M-17 and ASTM C496/C496, respectively.

Table 3. Properties of hardened concrete

3.2 Load deflection

During the test, the applied load and mid-span deflection of each beam were recorded at each load increment and illustrated as load-versus-mid-span deflection curves (Figures 13–15). The results of the experimental investigation revealed that increasing the web expansion to 30% enhanced the maximum load by 34.67% and 23.56% for models B-2S.30 and B-1M.30, respectively. In comparison, at a web expansion of 50%, the load increased by 31.65% and 31.22% for models B-2S.50 and B-1M.50, respectively. Additionally, the load increased by 10.81% and 41.95% for models B-2S.70 and B-1M.70, respectively, compared to the control sample without expansion, as depicted in Table 4. This is likely due to the increased effective depth of the steel section in resisting bending.

As for secant stiffness, it increased across all models, ranging from 11.87% to 18.52% relative to the reference sample. However, at 50% web expansion ratio for both models, the secant stiffness was almost identical, indicating stress redistribution along different paths.

Table 4. Ultimate load and mid-span deflection results

tu_pian_42.png

Figure 13. Load the verse deflection curve for Group 1

tu_pian_43.png

Figure 14. Load the verse deflection curve for Group 2

tu_pian_44.png

Figure 15. Load the verse deflection curve for Group 3

3.3 Mode of failure

In the reference (solid) sample without an expanded web, the failure mechanism was flexural. In the first group (specimen B-2S.30), although failure was governed by local buckling at the web post, the web remained relatively stable. In contrast, in the second group (specimen B-1M.30), the failure was a flexural mechanism, as displayed in Figures 16–18. When the web expansion was increased to 50%, the failure occurred in the B-2S.50 specimens due to local buckling at the web post., whereas B-1M.50 exhibited fracture at the web post weld, as presented in Figures 19 and 20. At 70% web expansion, while the failure rate of B-2S.70 specimen decreased, the web post buckling occurred. Conversely, B-1M.70 presented an increase in tolerance until web post buckling failure occurred. This indicates the slenderness ratio of the web plate in model B-2S.70, since the web plate thickness was divided on both sides. In model B-1M.70, it was placed in the middle to connect the two asymmetrical sections, providing greater stiffness, as displayed in Figures 21 and 22.

tu_pian_45.png

Figure 16. Failure mode of specimen B-NN

tu_pian_46.png

Figure 17. Failure mode of specimen B-2S.30

tu_pian_47.png

Figure 18. Failure mode of specimen B-1M.30

tu_pian_48.png

Figure 19. Failure mode of specimen B-2S.50

tu_pian_49.png

Figure 20. Failure mode of specimen B-1M.50

tu_pian_50.png

Figure 21. Failure mode of specimen B-2S.70

tu_pian_51.png

Figure 22. Failure mode of specimen B-1M.70

3.4 Effect of type connection

At 30% web expansion ratio, the maximum load decreased by 8.25% for specimen B-1M.30 compared to specimen B-2S.30. This results in greater stability for the web plates connected on both sides of the asymmetrical steel section. At 50% web expansion ratio, the difference in maximum loads was significantly small, allowing the web plates connected on both sides to reach their maximum load capacity. Although the two connection types exhibited nearly identical ultimate load capacities, their failure modes differed. This indicates that comparable strength can be achieved through different internal load-transfer and stress redistribution mechanisms. In one connection configuration, the applied load was transferred through a relatively direct stress path, leading to localized damage concentration in the critical region. In the other configuration, the connection details allowed stresses to redistribute over a wider region, altering the failure mode while maintaining a similar ultimate resistance. Therefore, the close values of load capacity should not be interpreted as evidence of identical behavior. Instead, they suggest that the two connection types mobilized distinct resistance mechanisms, yielding comparable ultimate strengths. This difference in failure behavior highlights the notable role of connection detailing in controlling stress redistribution, local stiffness, and the progression of damage. In response, further investigation, including detailed strain measurements and numerical modeling, is recommended to confirm the stress redistribution patterns and better understand the governing mechanisms. Conversely, at 70% web expansion ratio, the load capacity of specimen B-1M.70 increased by 28.1% compared to that of B-2S.70 specimens. This is likely due to the increased section stiffness with increasing depth, as presented in Table 5 and Figure 23.

tu_pian_52.png

Figure 23. Comparison of the type of connection of the tested specimens

Table 5. Results of the type connection effect

3.5 Effect of expansion ratio

Based on Table 6, the asymmetrical steel sections are connected by a web plate distributed on both sides, with a thickness comparable to the web thickness. When the web expansion ratio is raised to 50% and 70%, respectively, the maximum load capacity of models B-2S.50 and B-2S.70 drops by 2.24% and 17.71%, respectively, compared to model B-2S.30. This is due to a slenderness web plate and an increased effective depth. Conversely, when two asymmetrical sections are joined by a web plate with the same thickness as the steel section's web and is welded between them, increasing the web expansion ratio from 50% to 70% increased the load by 6.2% and 14.88% for models B-1M.50 and B-1M.70, respectively, in comparison to model B-1M.30. This is due to the increased depth and, consequently, the increased moment of inertia, as displayed in Figure 24.

Table 6. Results of the expansion ratio effect

tu_pian_53.png

Figure 24. Comparison of the expansion ratio of the tested specimens