Evaluation of ground settlement in response to shield penetration using numerical and statistical methods: a metro tunnel construction case

Abstract

Shield tunnelling can cause ground settlement, which poses significant risks to adjacent structures or facilities. To understand complex soil behaviour in response to shield penetration, a model that can establish the shield–ground relationship and accurately predict tunnelling-induced ground settlement is necessary. The aim of this paper is to combine numerical methods and statistical methods for settlement prediction in the Wuhan (China) metro project. During the pre-construction stage, due to the lack of instrumentation data, the numerical method was applied to simulate the tunnelling process. The relevant factors influencing ground settlements were identified by examining the model’s sensitivity to each parameter. After the shield launch, data of the relevant factors and field measurements were collected. Using these data, a statistical model based on an adaptive relevance vector machine (aRVM) was trained for real-time prediction of the ground settlement development. The simulation results show that a number of factors, including geometrical, geological and shield operational parameters, contribute to ground settlement, and the aRVM model can accurately and effectively predict settlement development. The example application demonstrates that the method is a practical tool for settlement prediction and can be widely used in metro projects.