Buy article PDF
The purchased file will be sent to you
via email after the payment is completed.
US$ 35
|
Structural Engineering and Mechanics Volume 97, Number 2, January25 2026 , pages 183-215 DOI: https://doi.org/10.12989/sem.2026.97.2.183 |
|
|
|
Hybrid random forest-based models for estimating punching shear strength of flat slabs |
||
Mosbeh R. Kaloop, Furquan Ahmad, Pijush Samui, Jong Wan Hu, Mohamed Rezaik, Basem S. Abdelwahed
|
||
| Abstract | ||
| Punching shear failure in reinforced concrete (RC) slabs is a brittle and critical phenomenon; therefore, reliable prediction of punching shear strength (PSS) is essential for safe design. Conventional design code equations (e.g., ACI and EC2) often involve simplified assumptions that may not fully capture the nonlinear interactions among geometric, material, and reinforcement parameters, especially in complex loading scenarios. To address this limitation, this study develops a novel hybrid machine-learning paradigm for accurate PSS estimation. The approach integrates the Random Forest (RF) algorithm with three advanced meta-heuristic optimization techniques: the dragonfly algorithm (DA), sparrow search algorithm (SSA), and whale optimization algorithm (WOA), to enhance RF hyperparameter tuning and predictive accuracy. A comprehensive dataset containing eight influential parameters was used to construct two modelling cases that account for variations in slab cross-sectional characteristics. The hybrid RF-DA and RF-SSA models achieved the highest predictive performance, reaching a correlation coefficient of 0.98 during the testing phase, outperforming conventional RF, SVR, ELM, and code-based predictions. Sensitivity analysis revealed that slab and column geometry, as well as concrete strength, exert the strongest influence on PSS. This study introduces a novel integration of Random Forest with DA, SSA, and WOA for PSS prediction, enabling the superior modeling of complex, nonlinear structural behavior. The hybrid framework provides a reliable, datadriven alternative to traditional code equations, with the RF-DA model demonstrating exceptional potential for broader application in concrete strength prediction and structural design optimization. | ||
| Key Words | ||
| hybrid intelligence modelling; metaheuristic optimization; punching shear strength; random forest; RC-slab | ||
| Address | ||
| Mosbeh R. Kaloop: Department of Civil and Environmental Engineering, Incheon National University, Incheon, Korea; Incheon Disaster Prevention Research Center, Incheon National University, Incheon, Korea; Public Works Engineering Department, Mansoura University, Mansoura, Egypt Furquan Ahmad: Department of Civil Engineering, National Institute of Technology Patna, Patna, India Pijush Samui: Department of Civil Engineering, National Institute of Technology Patna, Patna, India Jong Wan Hu: Department of Civil and Environmental Engineering, Incheon National University, Incheon, Korea; Incheon Disaster Prevention Research Center, Incheon National University, Incheon, Korea Mohamed Rezaik: Appout ITs, Tanta, Egypt Basem S. Abdelwahed: Structural Engineering Department, Mansoura University, Mansoura, Egypt | ||