ASSESSING SEISMIC SOIL LIQUEFACTION POTENTIAL USING MACHINE LEARNING APPROACH

Abstract

The liquefaction vulnerability of soil is generally related to a few soil parameters which are ordinarily measured by laboratory tests on distributed and undistributed tests under distinctive test conditions. This study uses methods based on a standard penetration test to assess liquefaction criteria to appraise the liquefaction vulnerability for soil deposits of Chalus City placed in a high seismic area. To overcome the deficiencies of these experimental strategies an ANN-based model has been created utilizing the Artificial Intelligence technique to anticipate liquefaction. The proposed model is a function of the plasticity index, liquid limit, water content, and some other geotechnical parameters. Reliability index (β) and probability of liquefaction (PL) have also been determined for both the proposed methods for a superior understanding of their accuracies and strength. First-order second moment (FOSM) reliability analysis has been embraced in the present paper. The observation drawn from the study illustrates a reliable and conventional expectation rate of the regression as compared to the experimental strategy. A strong regression shown for assessing the liquefaction vulnerability, which is based on field test information for preparatory prediction, would be of extraordinary help within the field of geotechnical designing.