Effects of the number of simulation iterations and meshing accuracy in monte-carlo random finite-difference analysis

Abstract

Monte-Carlo random finite-difference analysis (MCRFDA) can incorporate the spatial variability of soil properties into the analysis of geotechnical structures. However, two factors, namely, the fineness of the generated elements (reflected by the number of elements, Ne) and the number of MC simulation iterations, NMC, considerably affect the computational efficiency of this method, creating a barrier to its broad use in real-world engineering problems. Hence, an MCRFDA model of a circular underground cavern is developed in this study. The convergent deformation of the cavern is analyzed while considering the spatial variability distribution of the elastic modulus. Moreover, the effects of NMC and Ne on random FD calculations are investigated. The results show the following. An NMC greater than 500 is desirable for the FD analysis of a conventional structure. For a specific structure, Ne does not have a significant impact on the mean of the simulated values but appreciably affects the standard deviation (SD) of the simulated values, where reducing Ne increases the SD of the simulated values.