Shear capacity prediction for reinforced concrete deep beams with web openings using artificial intelligence methods: Fid-Bau Portal

Shear capacity prediction for reinforced concrete deep beams with web openings using artificial intelligence methods

Saleh, Mona / AlHamaydeh, Mohammad / Zakaria, Mohamed

Highlights • Predicting ultimate shear strength of reinforced concrete deep beams is considered complex for their non-linear behavior. • Traditional prediction methods incapacitate to solve this issue precisely, especially if there are openings. • Artificial intelligence algorithms provide the facilities to establish robust predicting model for these structural members. • Support vector regression, multilayer perceptron regressor, gradient boosting regressor, and Ensemble algorithms were tested. • Stacking Ensemble model was the most fitting model among the examined models.

Abstract The prediction of nominal shear capacity ( $V_{n})$ of reinforced concrete (RC) deep beams with web openings is quite complex due to their highly nonlinear behavior. In this paper, Artificial intelligence (AI) methods have been applied to overcome that complexity by providing a reliable predicting model for RC deep beams having web openings. Support vector regression (SVR), multi-layer perceptron regressor (MLP), gradient boosting regressor (GBR), and Ensemble algorithms have been examined on the ten most influential input parameters. Input data includes both experimental results of 179 specimens with various opening shapes and simulated results of 5032 specimens. Among the explored AI algorithms, it was found that Stacking Ensemble showed the best results with a determination coefficient (R²) of 0.998. Finally, it can be concluded that AI algorithms are considered a sufficient and powerful tool for predicting the nominal shear strength ( $V_{n}$ ) of RC deep beams with web openings.