On this page

Research on the Optimization of the Load-bearing Capacity of Short Steel-Concrete Columns Using Genetic Algorithms

By: Qingyun Ge 1, Jing Zheng 1, Fulian Yang 1, Caimei Li 2
1School of Architecture and Civil Engineering, West Anhui University, Lu’an, Anhui, 237012, China
2Gates Winhere Automobile Water Pump Products (Yantai) Co., LTD., Yantai, Shandong, 712000, China

Abstract

This paper conducts a systematic study on the axial compressive strength of short columns in steelconcrete composite structures (CE-CFST), revealing their mechanical properties and failure mechanisms. A refined finite element model was established using ABAQUS to analyze the interaction between the steel tube and concrete interface and the material constitutive relationships. A bearing capacity prediction model based on a BP neural network was proposed, and its interpretability was verified using the SHAP method. The research results show that for specimens with a steel tube thickness of t mm  3 , the circumferential tensile prestress of the CE-CFST-3 steel tube is 183 , which is only about 0.17 times the yield strain of the steel tube, with a strength ratio of 1.05. For specimens with a steel tube thickness of t mm  5 , the bearing capacity also increases significantly. The BP neural network model performed best among the comparison models, with the contribution of each parameter to the results decreasing in order from ρx, fc, t, L, fy, to ρy.