Artificial Neural Network Model for Evaluating Gravelly Soils Liquefaction Using Shear Wave Velocity

Evaluation of liquefaction potential of soil is important in geotechnical earthquake engineering. Significant phenomena of gravelly soil liquefaction were reported in 2008 Wenchuan earthquake. Thus, further studies on the liquefaction potential of gravelly soil are needed. This paper investigates the potential of artificial neural networks-based approach to assess the liquefaction potential of gravelly soils form actual shear wave velocity field data. The success rates for occurrence and nonoccurrence of liquefaction cases in Wenchuan earthquake could be up to 100% and 94%, respectively. The study suggests that neural networks can successfully model the complex relationship between seismic parameters, soil parameters, and the liquefaction potential of gravelly soils.