Neural Network Modeling of Shear Wave Velocity of Macau Soils Using SPT and CPT Data

Abstract

Shear wave velocity of soil is an important parameter for earthquake analysis and civil engineering designs. Correlations between shear wave velocity and SPT-N value based on the simple power-law regression model were provided by previous studies. However, due to the inherent uncertainties of the SPT-N values, estimation of shear wave velocity may be improved by incorporating other site investigation data. In this study, artificial neural network models were used for the correlation analysis between the shear wave velocity with SPT (Standard Penetration Test) and CPT (Cone Penetration Test) data, based on recent field test data obtained from the site investigation program of the Macau major transportation project. The analysis results of this study indicated that incorporating the additional soil parameters in the correlation model would improve the prediction performance. When combined with SPT and CPT to form neural network models, a better prediction would be obtained than that using SPT or CPT alone.