Damage identification of a tunnel liner based on deformation data

Abstract

A new system identification approach based on tunnel deformation data is proposed to find the damage in the tunnel liner. For this, an inverse problem in which deformation data and dead load of concrete lining are known a priori is introduced to estimate the degree and location of the damage. Models based on uniform reduction of stiffness and smeared crack concept are both employed and the efficiency and relative advantage are compared with each other. Numerical analyses are performed on the idealized tunnel structure and the effect of white noise, common in most measurement data, is also investigated to better understand the suitability of the proposed models. As a result, model 1 based on uniform stiffness reduction method is shown to be relatively insensitive to the noise, while model 2 with the smeared crack concept is proven to be easily applied to the field situation since the effect of stiffness reduction is rather small. Finally, real deformation data of a rail tunnel in which health monitoring system is in operation are introduced to find the possible damage and it is shown that the prediction shows quite satisfactory result.