Abstract:

In order to analyze sand liquefaction problem efficiently, the stepwise discriminant method was applied to its prediction; the factors with significant discriminant ability were selected to establish discriminant function. A number of site cases were used to verify its validity and accuracy. The results show that: stepwise discriminant analysis model has excellent liquefaction predicting performance; and it can choose effective factors which play a leading role in sand liquefaction. Compared to the distance discriminant analysis, discriminant function based on stepwise discriminant analysis is more stable; and fewer factors investigated or tested are needed, which saves human and material resources for the project construction. Thus it is worthwhile to apply stepwise discriminant analysis to liquefaction prediction widely.

Key words: sand liquefaction, stepwise discriminant, prediction