A polynomial model for predicting liquefaction potential from cone penetration test data

Liquefaction is a serious geotechnical hazard leading to catastrophic damage to life and property. In many instances, it may be preferable to predict liquefaction susceptibility indirectly by common in-situ tests, such as the Cone Penetration Test (CPT). A new approach for prediction of liquefaction susceptibility is proposed, which presents a polynomial model to correlate the Cyclic Resistance Ratio (CRR) predicated on subsoil geotechnical properties from CPT tests, that is, normalized cone tip resistance (qC1) and friction resistance (fs). The derived model is applied to a total of 182 data sets, including eld investigation records from eighteen earthquakes. The performance of the proposed approach is compared to other available methods within a quantitative validation framework (e.g., precision, recall, and F-score). Results indicate the accuracy and generalization of the proposed new approach in predicting liquefaction susceptibility.