Development of Liquefaction Index Prediction Equations from Post-liquefaction CPT Data Using ANN and GEP

Abstract

In this paper, Artificial Neural Network (ANN) and Genetic Expression Programming (GEP) have been used to develop two Liquefaction Index (LI) equations which will be able to predict effectively whether a soil layer at any depth would liquefy or not in case of an earthquake. 226 post-liquefaction Cone Penetration Test (CPT) data (133 are liquefied cases and the rest 93 are non-liquefied cases) have been collected from published literature and using the collected data, an ANN and a GEP model have been built. From each developed model, a LI equation has been developed which uses CPT data of soil and Peak Ground Acceleration (PGA) as inputs and returns either 1 or 0 (1 means liquefaction may occur and 0 means liquefaction may not occur). A comparative study between both the models has also been conducted in this study.