Research on Settlement Prediction of Tunnel Engineering Based on Machine Learning
The settlement analysis of tunnel engineering is of great significance for the overall stability and safety of tunnel structures.The settlement caused by tunnel construction can be effectively predicted by machine learning al-gorithms.The specific application process of machine learning algorithms(BP,GA-BP neural network models)in tunnel engineering settlement prediction was studied,which was generally divided into four stages of pre analysis of tunnel engineering related parameters,clarification of the optimal parameters of machine learning models,accuracy analysis of machine learning models and prediction of tunnel engineering settlement volume.By studying the distri-bution characteristics of related data of tunnel engineering,the predictive performance of BP and GA-BP neural network models was compared.The results show that there is a relatively significant correlation between the palm surface condition,on/off state,thickness,groundwater level and surface subsidence.And these four parameters are also the main factors affecting surface subsidence;The influencing factors can be effectively characterized by the PC A principal component analysis method.The distribution patterns of 12 features can be comprehensively charac-terized through principal component 1 and 2;The relationship between the influencing factors of shield tunneling and surface settlement can be better described by BP and GA-BP neural network models.The GA-BP neural network model has higher prediction accuracy.It is recommended to use the GA-BP neural network model to characterize and analyze the surface settlement.
国家科技期刊平台
NSTL
万方数据
