Determination of Punching Shear Capacity of Concrete Slabs Reinforced with FRP Bars Using Machine Learning

Abstract

The prevention of the shear damage that may occur in the immediate surroundings of the columns on flat slabs due to punching is an important subject. In the literature, several experiments have been carried out by using the FRP composite bars to increase the punching strength of the flat slabs. In this study, firstly, an extensive literature review has been carried out, and the experimental data for the 141 slabs, which were produced with GFRP bars, CFRP bars and the traditional reinforced concrete steel bars and which were damaged by punching, has been gathered. Parameters were adjusted for the collected data, and afterwards, prediction models were developed for the punching strength of the slabs by using the relevant algorithms in five different machine learning techniques (Multiple Linear Regression, Bagging-Decision Tree Regression, Random Forest Regression, Support Vector Regression and Extreme Gradient Boosting (MLR, Bagging-DT, RF, SVR, XGBoost). In addition to the effect of each parameter in the data and the testing of the algorithms' convergence performance in relation to the results, the study intuitively discussed the extent that the ACI 440 and other approaches in the literature predict the punching strength. The prediction value of especially the building codes was more conservative than the experimental results. The best results were achieved by the SVR among the five different algorithms. SVR achieved a predicted success of for the strength of slabs produced with GFRP bars. After analysis, R-2 values, MAE and RMSE performance metrics were found to be well above the empirical correlations with 96.23%, 0.16 and 0.19 for slabs produced with GFRP bars, respectively.