Abstract: At present, there is no unified, systematic cognition regarding the mechanism of karst collapse as research achievements are mainly concentrated on the causes and influential factors of karst collapse in qualitative sense rather than quantitative approach. In this paper, the process of karst collapse covered with sand layer was investigated quantitatively, and the barrier effect of clay layer covering the karst was studied by sampling (keeping the pressure head unchanged during the process of sampling) and pressure monitoring on a physical model. Results showed that: 1) in the structure of rock covered with sand, three times of karst collapse happened during the steady decline of groundwater level, with the critical seepage-breakage gradients rcaching 0.525, 2.500 and 3.400, respectively; the maximum collapse rate gradually decreased, yet the durations and degrees of collapse gradually increased; moreover, the process of the first karst collapse was more complicated than that of the second and the third collapse; 2) the seepage theory of liquefaction has been verified by the results of karst collapse of rock covered with sand layer; 3) groundwater level fluctuation exacerbates the formation of karst collapse; 4) the seepage-breakage gradient increased from 0.525 to 48.300 in the presence of clay layer, indicating that clay layer could significantly improve the safety of karst overburden layer and effectively inhibit karst collapse. The results of this study could improve the understanding of karst collapse mechanism in a quantitative sense.

Key words: karst collapse, physical model test, hourglass, seepage theory of liquefaction, seepage-breakage gradient, quantitative research