Sequential Gaussian simulation predicts soil liquefaction potential

Abstract

The Yellow River Delta is covered with a large number of pipelines, but due to the complex soil composition in the region, ensuring that pipelines are not damaged by soil liquefaction is an important issue at present. Based on the simplified method of the cone penetration test (CPT), the sequential Gaussian simulation (SGS) can probabilistically simulate the liquefaction potential index (LPI) in the study area to solve the problem of the smoothing effect occurring in the kriging method. In this study, 10 experiments were conducted in the Yellow River Delta to evaluate soil liquefaction within the site using uncertainty analysis by the SGS method. The results indicate that (1) All LPI values in the study area are less than 5, with an overall sub-moderate liquefaction potential. (2) The results of the variogram model show that the Gaussian function model has the best fit with a Root Mean Squared Error of 0.429. The results of the e-type simulation realizations illustrate that the soils around the three sites S1, S5, and S10 exhibit high LPI values, distributed in a band in the middle of the western and eastern parts of the site. (3) Uncertainty analysis was performed using LPI = 2 as a threshold to explore the distribution of areas of moderate liquefaction potential and areas of low liquefaction potential in the study area. (4) Improvements were made to address the current problem of inappropriate values of liquefaction thresholds and the lack of medium liquefaction potential thresholds by proposing when LPI = 20 as the liquefaction threshold, LPI = 10 and 16 as the thresholds for low liquefaction potential, medium liquefaction potential and high liquefaction potential.