NEURAL-NETWORK MODELING OF CPT SEISMIC LIQUEFACTION DATA. TECHNICAL NOTE

The use of cone-penetration-test (CPT) resistance data as a field index for evaluating the liquefaction potential of sands is gaining attention because of its popularity as an in situ test technique for site characterization. This paper investigates the feasibility of using neural networks to evaluate liquefaction potential from actual CPT field data. A back-propagation neural-network algorithm was employed to model actual field-liquefaction records. Results suggest that neural networks can successfully model the complex relationship between seismic parameters, soil parameters, and the liquefaction potential. This simple model is as reliable as the conventional method of assessing liquefaction potential. Calibration and normalization of the cone resistance is not required, unlike the traditional method. With additional field case records, data can be readily included in the neural-network training and testing data for further improvements of modeling of liquefaction potential.