Prediction Methods for the Maximum Surface Settlement of Subway Tunnels Based on Machine Learning
It has always been a hot topic to study the settlement caused by subway tunnel construction in tunnel re-search.However,due to the complexity of the geological strata and the variability of construction parameters,it is difficult to balance ease of use and accuracy to predict the maximum ground settlement caused by tunneling by em-pirical formulas or numerical simulations.In recent years,with in-depth research on the machine learning theory and the rapid development of computer software and hardware technology,emerging algorithms such as machine learning are increasingly being applied to predict maximum ground subsidence.This article uses two machine learn-ing algorithms(the random forest and the BP neural network)to predict and analyze the maximum ground settle-ment caused by tunnel construction by taking the geometric parameters of the tunnel,the construction parameters of the shield machine and geological parameters as inputs.The results show that the two machine learning algo-rithms can both achieve higher-quality prediction,and that the stability of the random forest model is better than that of the BP neural network model.
NETL
NSTL
万方数据
