A new method of reliability-based liquefaction probability for sandy soils-using the knowledge nested partition algorithm

This paper employed the knowledge nested partition algorithm to minimize the Hasofer-Lind reliability index. Accordingly, it developed a new probabilistic model which is based upon Seed simplified method for assessing the earthquake induced liquefaction potential of sandy soils. Six major parameters for liquefaction evaluation were expressed using a radial space system. This study used 360 SPT-N based case histories, including liquefied and non-liquefied, to verify the model accuracy. Through the establishment of a liquefaction probability evaluation process, this study examined the variance of parameters that affect the liquefaction potential. Local liquefaction hazard analysis, if having a great amount of field data, could advantage from the model developed in this study.