Abstract: In recent years, seismic damage caused by sand liquefaction occurred frequently in earthquakes. Sand liquefaction has become the focus of earthquake engineering. It is of great significance to develop a method for evaluating liquefaction based on the cone penetration test (CPT) with good application prospects to prevent liquefaction damage. This paper tests the CPT liquefaction evaluation methods through the collected 171 sets of liquefaction data of 15 earthquakes, and analyzes the problems of these evaluation methods, and proposes a new liquefaction evaluation formula of hyperbolic model based on data regression optimization. Extracting 147 sets of liquefaction data of the 2011 New Zealand earthquake to test the new formula, the results show that the new formula has a high and balanced discriminant success rate for different seismic intensities and buried depth of sand, and also overcomes the disadvantages in the existing liquefaction evaluation methods. The new formula proposed in this paper has been adopted into the revised content of the latest The General Rule for Performance-based Seismic Design of Buildings and can provide theoretical basis and support for the revision of relevant specification and engineering applications in China in the future.

Key words: evaluation of liquefaction, cone penetration test, hyperbolic model, New Zealand earthquake