Characterized model uncertainties of CPT-based simplified procedures for assessing soil liquefaction and its application to Taiwan offshore wind farms

This paper aims to calibrate model uncertainties of the current Cone Penetration Test (CPT) methods for assessing soil liquefaction potential and introduce a new and concise one called the hyperbolic function (HBF) method based on the unique Chi-Chi earthquake database in Taiwan. The development and features of the HBF method were introduced at first. The performance and model uncertainty of each CPT method is then evaluated with a database of historical and Chi-Chi earthquake cases using the indices of the predicted success rate and bias-factor models, respectively. The success rate evaluations show that the HBF is one of the most accurate methods. The calibrated bias-factor models can indicate the conservatism of each CPT method applied to the Taiwan sites. Afterward, the liquefaction hazard curves associated with these CPT methods for Taiwan sites are derived based on the case histories of liquefaction and non-liquefaction in Taiwan. A real case study to illustrate the applications of these calibrated models evaluating liquefaction hazards to an offshore wind farm near Taiwan western coast is also demonstrated in this paper. These above results bring not only recommendations on CPT liquefaction assessment methods incorporated into the Technical Specifications for Site Investigation and Design of offshore wind farms in Taiwan, but procedures to calibrate regional biased models for other offshore wind farms in the world.