System reliability analysis of reservoir landslides: Insights from long-term reservoir operation

The reservoir operation awakens numerous landslides with multiple sliding surfaces known as reservoir landslides, and the systematic stability analysis for such landslides is becoming increasingly urgent. Taking the Majiagou landslide as an example, this paper analyses the comprehensive performance of the landslide from a probabilistic point of view. Under a reservoir operation cycle, a series of numerical analyses are carried out to simulate the migration of the seepage field, then the dynamic stability of the landslide is quantified accordingly. Subsequently, the wetting-drying cycles test is used to model the weakening of strength parameters in hydro-fluctuation belt under the long-term reservoir operation. Considering the weakening effect of long-term reservoir operation on the hydro-fluctuation belt, the system reliability is evaluated using the Ditlevsen's bounds. The results suggest that the reservoir operation can affect the stability of the landslide by changing the seepage field. The system failure probability gradually rises as the number of wetting-drying cycles increases. Compared with conventional probabilistic analysis that calculates the failure probability of each sliding surface mechanically, analyzing the landslide in terms of system reliability can effectively narrow the failure probability range, which provides an insightful idea for evaluating the systematic stability of analogous reservoir landslides.