System reliability analysis of soldier-piled excavation in unsaturated soil by combining random finite element and sequential compounding methods

The stability of excavation in unsaturated soil is closely related to the variation of soil properties and matric suction. Determination of different failure modes and identifying the contribution of each one in an unsaturated soldier-piled excavation are the vital aspects of system reliability analysis. To address these issues, this paper provided a stochastic framework with a random elasto-plastic finite element-based program coded in MATLAB to evaluate the reliability indices of individual failure modes with considering the inherent uncertainty of real site soil properties and unsaturated state. In the next step, the sequential compounding method (SCM) was utilized to obtain the system reliability index by compounding the reliability indices of individual failure modes. Numerical results of a case study showed that in all failure modes, considering unsaturated state not only increases the mean value of factor of safety (FS) but also decreases the related standard deviation, which can be counted as a goal of reliability analysis. Among the reliability indices of the components, the most critical one is attributed to the lateral displacement. Furthermore, the safety ratio concerning the shear force has the maximum reliability index compared to the others. Moreover, based on the coefficient of variation (COV) of the components, it was found that the uncertainty of the soil parameters has the most significant effect on the global safety factor of the excavation.

Automated summary

This study establishes a stochastic framework to evaluate the reliability indices of individual failure modes in unsaturated soil excavation, considering the uncertainty of soil properties and unsaturated state. The system reliability index is obtained using the sequential compounding method, with non-saturated state affecting the mean and standard deviation of the safety factor.