Parametric investigation of rectangular CFRP-confined concrete columns reinforced by inner elliptical steel tubes using finite element and machine learning models

Abstract

Nowadays, due to the structural advantages gained by combining three different materials' properties, columns made of carbon-fiber reinforced polymer (CFRP)-confined concrete with inner steel tube have received researchers' interest. This article presents the nonlinear finite element analysis and multiple machine learning (ML) model-based study on the behavior of round corner rectangular CFRP-confined concrete short columns reinforced by the inner high-strength elliptical steel tube under the axial load. The reliability of the proposed nonlinear finite element model was verified against the existing experimental investigations. The effects of the parameters such as the concrete grade, thickness of reinforcing steel tube, cross-sectional size of inner steel tube, and thickness of CFRP on the behavior of the columns are comprehended in this study. Furthermore, multiple ML models were proposed to predict the ultimate axial load, ultimate axial strain, and lateral strain of the test specimens. The reliability of the proposed ML models was evaluated by six distinct performance metrics. From the parametric investigation, it was found that concrete with lower compressive strength gained more strength enhancement because of confinement between CFRP and the inner steel tube than high-strength concrete relative to its unconfined compressive strength. The proposed ML models of extreme gradient boosting and random forest provided the best-fit results than the artificial neural network and Gaussian process regression models in predicting the axial load and axial and lateral strains of the columns.

Keywords

• Confined concrete
• Steel tube
• Dilation angle
• Lateral strain
• Strength enhancement
• Machine learning