Efficient Structural Analysis: Predicting Punching Shear Strength in Two-Way Concrete Slabs Using Gradient Boosting Machine Learning

Abstract

This study delves into the application of machine learning (ML), specifically a Gradient Boosting (GB) model, for predicting the punching shear strength (PSS) of two-way reinforced concrete flat slabs. Leveraging a dataset comprising 241 experimental observations from reputable sources, the research investigates the influence of critical factors on PSS, including slab thickness, column section width, effective slab depth, reinforcement ratio, concrete compressive strength, and reinforcement yield strength. Hyperparameter optimization techniques are employed to fine-tune the model's parameters, leading to enhanced predictive performance. Monte Carlo simulations are utilized to validate the model's reliability and generalizability. The results demonstrate that the GB model achieves high precision and reliability, reducing the need for resource-intensive experimentation in predicting PSS for two-way slabs. Furthermore, the study compares the performance of the developed model with that of conventional design codes, highlighting the model's superior accuracy. This research contributes to the broader application of ML in structural engineering, offering an efficient and accurate approach to analyzing and designing structural elements.