Prediction of axial bearing capacity of concrete-filled steel tube columns based on the artificial neural network
The axial bearing capacity of concrete-filled steel tube(CFST)columns is one of the im-portant indicators in structure design,and accurately predicting the capacity is a significant prerequi-site for ensuring structure safety.A dataset including 394 samples by collecting literature data and numerical simulation data is established in this paper.The bearing capacity of CFST columns under axial compression is predicted based on the artificial neural network.A BP neural network model with good prediction accuracy and generalization is established to analyze and compare the predicted values with the experimental data values.SHAP interpretable analysis is adopted to analyze the influ-ence of model parameters on the prediction results.The results are as follows:The BP neural network model established in this paper can quickly and accurately predict the axial bearing capacity of CFST columns with different stirrup ratios under different sizes.Compared with traditional experimental and numerical simulation methods,BP neural network model can effectively save time and costs.The re-search results can provide certain data reference for the prediction and application of artificial neural network technology in the axial bearing capacity and other mechanical properties of CFST columns.
NETL
NSTL
万方数据
