Prediction of ground settlement due to tunneling using artifical neural networks

Abstract

Ground surface settlement trough associated to tunneling is characterized by two important parameters: the maximum surface settlement at the point above the tunnel centerline (Smax) and the width parameter (i) which is defined as the distance from the tunnel centerline to the inflection point of the trough. The estimation of these settlement parameters is a very complex problem due to uncertain nature of the soil. Over the years, many methods have been proposed to predict the tunneling-induced settlements. Most of these methods are empirical in nature. However, a method with high degree of accuracy and consistency has not yet been developed. Accurate prediction of settlement is essential since settlement is the governing factor in the design process of the tunnels. In this research, the use of artificial neural network (ANN) for the prediction of maximum surface settlement and trough width is explored.