Reliability study on shield tunnel face using a random limit analysis method in multilayered soils
Hong-Zhan Cheng,Jian Chen,Ren-Peng Chen,Guo-Liang Chen
DOI: https://doi.org/10.1016/j.tust.2018.11.038
IF: 6.9
2019-02-01
Tunnelling and Underground Space Technology
Abstract:Tunnels are generally driving in multilayered soils, and natural soils are well known to have high variability in strength. However, few attempts have been made to consider both these two features for stability analysis of tunnel face. To address this problem, this study employs an efficient method named random limit analysis method (RLAM). The method adopts random fields to describe the spatial variability of soil strength parameters, and employs the upper bound theorem of limit analysis to determine the critical state of tunnel face. A homogenous tunnel and three two-layer tunnels were investigated to validate the proposed method compared to the rigorous random finite difference method (RFDM). It is shown that the proposed method can not only explicitly consider soil heterogeneity and parameters variability, but also significantly improve the computational efficiency of probability of failure. RLAM integrating with a Monte Carlo simulation is compiled by MATLAB, thereby allowing the reliability problem of tunnel face to be studied conveniently. Parametric analyses are then performed to explore the effects of the auto-correlation distance on the reliability and failure mechanism for different soil conditions. The results suggest that partial collapse may be observed when tunnel is driving in multilayered soils, especially considering the variability of soil properties. In general, neglecting the variability will result in an overestimation of the stability of tunnel face, and probability of failure is more sensitive to the change of the vertical auto-correlation distance than the horizontal one.
construction & building technology,engineering, civil