Physics-data fusion modeling for lateral friction prediction of large-diameter PHC piles: integrating finite element analysis with XGBoost optimization

In response to the insufficient accuracy of predicting the lateral friction distribution of large-diameter PHC pipe piles under complex foundation conditions, this paper proposes a coupled construction method combining ANSYS finite element simulation calculation and XGBoost machine learning, which uses FEM-AI coupling points to achieve deep integration of physical and data models, ensuring the physical fusion of the physical model and improving its prediction accuracy. This article selects the actual experimental environment of a certain dock project, and the results show that the prediction error of this article is only an average of 3.1%, which is about 2.3 times faster than traditional methods in terms of calculation speed. The research results of this article break through the limitations of traditional general models with low universality and applicability, and provide an efficient and intelligent design method for PHC pipe piles under large-scale foundation structure conditions.