A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Interpretable Ensemble Machine Learning Models for Shear Strength Prediction of Reinforced Concrete Beams Externally Bonded with FRP
Shear strengthening is a complex phenomenon that garnered significant attention in the structural engineering community. Due to the catastrophic nature of shear failures, several attempts have been made in retrofitting reinforced concrete (RC) beams out of which the incorporation of externally bonded fiber reinforced polymer (FRP) layers offer a remarkably fast, economical, and reliable solution. This paper presents an approach to predict the shear capacity of FRP strengthened RC T-beams using interpretable ensemble machine learning models. The study covers a comprehensive databank comprising a wide array of parameters including concrete design, FRP composition as well as beam cross sections. The efficiency of the developed models in predicting the shear capacity of FRP retrofitted RC T-beams is evaluated by comparing the results with several design guidelines. It is observed that the random forest and CatBoost models provide the most precise shear capacity estimations of the FRP retrofitted RC T-beams. The R2 and MAE values obtained from the random forest model were 0.897 and 0.128 kN, respectively, whereas those by the CatBoost model were 0.899 and 0.127 kN, respectively. The best performing model CatBoost is made interpretable using the Shapley Additive exPlanations which reveal that the most important input parameter contributing to shear capacity of the FRP strengthened RC T-beams is the height of the FRP layers used in the retrofit process. The proposed ensemble models presented in this paper are proved to be superior to the existing mechanics-driven models currently being used for design practices.
Interpretable Ensemble Machine Learning Models for Shear Strength Prediction of Reinforced Concrete Beams Externally Bonded with FRP
Shear strengthening is a complex phenomenon that garnered significant attention in the structural engineering community. Due to the catastrophic nature of shear failures, several attempts have been made in retrofitting reinforced concrete (RC) beams out of which the incorporation of externally bonded fiber reinforced polymer (FRP) layers offer a remarkably fast, economical, and reliable solution. This paper presents an approach to predict the shear capacity of FRP strengthened RC T-beams using interpretable ensemble machine learning models. The study covers a comprehensive databank comprising a wide array of parameters including concrete design, FRP composition as well as beam cross sections. The efficiency of the developed models in predicting the shear capacity of FRP retrofitted RC T-beams is evaluated by comparing the results with several design guidelines. It is observed that the random forest and CatBoost models provide the most precise shear capacity estimations of the FRP retrofitted RC T-beams. The R2 and MAE values obtained from the random forest model were 0.897 and 0.128 kN, respectively, whereas those by the CatBoost model were 0.899 and 0.127 kN, respectively. The best performing model CatBoost is made interpretable using the Shapley Additive exPlanations which reveal that the most important input parameter contributing to shear capacity of the FRP strengthened RC T-beams is the height of the FRP layers used in the retrofit process. The proposed ensemble models presented in this paper are proved to be superior to the existing mechanics-driven models currently being used for design practices.
Interpretable Ensemble Machine Learning Models for Shear Strength Prediction of Reinforced Concrete Beams Externally Bonded with FRP
Lecture Notes in Civil Engineering
Gupta, Rishi (editor) / Sun, Min (editor) / Brzev, Svetlana (editor) / Alam, M. Shahria (editor) / Ng, Kelvin Tsun Wai (editor) / Li, Jianbing (editor) / El Damatty, Ashraf (editor) / Lim, Clark (editor) / Rahman, Jesika (author) / Muntasir Billah, A. H. M. (author)
Canadian Society of Civil Engineering Annual Conference ; 2022 ; Whistler, BC, BC, Canada
Proceedings of the Canadian Society of Civil Engineering Annual Conference 2022 ; Chapter: 85 ; 1265-1278
2024-02-06
14 pages
Article/Chapter (Book)
Electronic Resource
English
| British Library Online Contents | 2017
| British Library Conference Proceedings | 2000
Reinforced Concrete Beams with Externally Bonded FRP Plates
| British Library Conference Proceedings | 1999