Numerical-simulation-based Landslide Warning System and Its Application

Abstract

Landslide early warning is a systematic project involving multidisciplinary integration, i.e. geology, mechanics, engineering, monitoring technology, and information technology. When conducting landslide studies, due to the limitation of failure theory of geological body, landslide warning system is purely based on monitoring data nowadays, so it is difficult to integrate interdisciplinary techniques. A trigger condition based prediction theory and disaster stage judgment approach are introduced, with which the landslide time prediction is converted to the disaster stage judgment. The definition and significance of fracture degree is presented, by which the disaster stage judgment is convert to inner fracture state analysis. At last, the numerical-simulation-based landslide warning system is discussed in detail. This system contains four parts, namely parameter acquisition part, disaster kernel analysis system, current state back analysis part, and landslide reliability evaluation part, respectively. In the first part, the essential parameters and its acquisition method is presented, and then the necessity and advantages of dimensional analysis is discussed. In the second part, the new numerical method named continuous discontinuous element method (CDEM) is introduced, and main features, i.e. Lagrange equation, strain strength distribution criteria, element crack strategy, and indented point & indented edge contact model is introduced. In third part, the inverse analysis method of current parameters of geological body based on monitoring data and numerical simulation is discussed. In the fourth part, the reliability evaluation steps are presented in detail. Finally, the sliding occurrence probability of Liangshuijing landslide in Chongqing, China is discussed, which demonstrated to show the precision and rationality of the proposed landslide warning system.