Reliability Analysis of RC Slab-Column Joints under Punching Shear Load Using a Machine Learning-Based Surrogate Model

Reinforced concrete slab-column structures, despite their advantages such as architectural flexibility and easy construction, are susceptible to punching shear failure. In addition, failure is a typical brittle failure, which introduces difficulties in assessing the functionality probability of structures. Therefore, prediction resistance corresponding reliability analysis critical issues design reinforced RC order enhance computational efficiency (RC) joints, database containing 610 experimental data used for machine learning (ML) modelling. According nonlinear mapping between selected seven input variables four ML models, artificial neural network (ANN), decision tree (DT), random forest (RF), extreme gradient boosting (XGBoost) established. With assistance three performance measures, root mean squared error (RMSE), absolute (MAE), coefficient determination (R2), XGBoost best model; its RMSE, MAE, R2 32.43, 19.51, 0.99, respectively. Such also reflected comparison with five empirical models introduced this paper. The process visualized by SHapley Additive exPlanation (SHAP); importance sorting feature dependency plots explain globally. Furthermore, paper adopts Monte Carlo simulation learning-based surrogate model (ML-MCS) calibrate joints real engineering example. A total 1,000,000 samples were obtained through sampling, index ? practical building was calculated simulation. Results demonstrate that target requirements under provisions can be achieved. sensitivity stochastic then conducted, impact on structural deeply examined.