Abstract
Composite foundation bearing capacity is the main basis of composite foundation design, and how to accurately predict the composite foundation bearing capacity is of great significance to foundation engineering. In order to analyze the main influencing factors of composite foundation bearing capacity, and predict the corresponding foundation bearing capacity. Based on the actual measurement data of vibration replacement stone column composite foundation project, the main factors affecting the bearing capacity of its foundation are identified by using gray correlation analysis (GRA) and stepwise regression (SR). On this basis, the support vector machine regression model (SVR) is constructed to predict the bearing capacity of the composite foundation. And the prediction results are compared with those of the BP neural network and GRA-SVR model. The results show that the main factors affecting the bearing capacity of vibration replacement stone column composite foundation include diameter, effective pile length, dense current, filling coefficient, natural density, replacement rate, bedding thickness, and pore ratio. The prediction accuracy of the GRA-SR-SVR, BP neural network, and GRA-SVR model are 98.23%, 97.08%, and 97.63% in order, and the GRA-SR-SVR model has the highest prediction accuracy, and it can accurately and effectively predict the composite foundation bearing capacity.