Reliability Analysis of Liquefaction Utilizing Monte Carlo Simulation Based on Simplified Stress Method

Liquefaction analysis is one of the most challenging issues in seismic geotechnical engineering. The unknown factors and pertinent uncertainties involved in the evaluation of liquefaction potential make the problem to be complicated. Liquefaction evaluation include deterministic and probabilistic methods. Deterministic methods are simple but they are not capable to consider the uncertainties. With regard to heterogeneous nature of the soil and probabilistic nature of earthquake loading, it seems that deterministic method is not sufficient for evaluation of liquefaction. Reliability methods are able to capture the uncertainties depending on variability of soil parameters and also to determine the factor of safety proportional to the acceptable risk. In recent years, reliability analysis of liquefaction has been done using approximated method. In the present research, reliability analysis of liquefaction triggering has been discussed using Monte Carlo simulation that is an accurate method. For this purpose, the parameters earthquake magnitude (Mw ), maximum horizontal acceleration (amax /g), total stress (sv ), effective stress (s'v ), fines content percent (FC), and SPT blow count (NSPT) are selected as stochastic parameters and the probability of liquefaction has been estimated. Application of the proposed method to the 233 well-documented case studies verify that deterministic method is not accurate enough to predict the liquefaction and reliability analysis should be used instead. Besides, the sensitivity tests indicate that the SPT blow count is the most influential parameter in liquefaction evaluation and large number of iterations is not required in Monte Carlo Simulation and the results converge after a specific number of iterations.