ASSESSMENT OF NON-LINEAR REGRESSION APPROACH FOR BACK-ANALYSIS ON TUNNELLING-INDUCED SURFACE SETTLEMENT– A CASE STUDY IN HO CHI MINH CITY

Abstract

Previous researchers proved that surface settlement induced by tunnel constructions can be described by a Gaussian curve with the two key parameters K, the trough width factor, and V_L, the volume loss. Knowing K and V_L values enables surface settlement trough to be calculated which is essential to determine the potential effects of tunnelling to soil and surrounding buildings. A non-linear approach has been widely used to estimate K and V_L values from field measurement using the sum of absolute errors. However, the reliability of the determined results could not be quantified. This paper uses surface settlement data from a case study in a tunnelling project in Ho Chi Minh city to determine K and V_L using the non-linear regression approach. Then, a method was proposed to quantify the goodness of fit and the reliability of the determined K and V_L. The results show that knowing the reliability of the K and V_L is essential for the designers and researchers to determine if these values should be used as reference for their calculation in a similar tunnelling project to predict surface settlement.