Abstract: The collapsibility of loess can be evaluated rapidly through physical and mechanical indexes. On the basis of field immersion tests and corresponding indoor tests, we analyze the relationship of loess collapsibility against depth, water content, dry density, void ratio, plasticity index, and compressibility coefficient using mathematical statistics, theoretical analysis, and data mining. The thick Q₃ loess along the Zhengzhou-Xi’an Highway is taken as research background. The research conclusions are as follows the coefficient of loess collapsibility is well correlated with physical and mechanical indexes, among which depth, moisture content, and dry density are linearly negatively correlated with collapsibility; void ratio is linearly positively correlated with collapsibility. The limit moisture content for loess collapse is 20%, limit dry density, 1.4 g/cm³, and limit void ratio, 0.9. In comparison, the collapsibility coefficient is mostly influenced by void ratio, followed by dry density, moisture content, plasticity index and coefficient of compressibility. In addition, we put forward an empirical formula of evaluating loess collapsibility by using multiple physical indexes based on data mining.

Key words: loess, collapsibility evaluation, physical indexes, coefficient of collapsibility, correlation