The Landslide Risk Evaluation Of Research Based On Comprehensive Reliability Analysis And Particle Flow Simulation

Landslide disasters are the most common and harmful among many geological disasters.Tens of thousands of losses are caused by landslide disasters every year.It not only endangered the safety of human life and property,but also caused a certain degree of damage to the resources and environment.Carrying out in-depth risk assessment and research work on landslide disasters will help relevant departments improve the standard of disaster prevention and mitigation,then minimize the losses caused by disasters.This paper comprehensively analyzes the research results of landslide risk assessment at home and abroad,and solves the problems of complex or large amount of calculation in the current research methods,risk assessment theory based on single landslide,establishes a complete set of single landslide risk assessment technical solutions.The following are the main research :(1)On-site investigation was conducted on Banling landslide,by analysing the basic feature of the landslide,and the rock and soil mass parameter values in the area were collected.Based on the principle of strength reduction,the slope model was constructed using finite element software,and the stability coefficients of the slope under natural and Heavy rain(saturated)were calculated to be 1.14 and 0.81,respectively.(2)Comprehensive reliability analysis method is used to analyze slope stability.RBF neural network is mainly used to fit slope function,and PSO algorithm is used to optimize RBF neural network model parameters.In addition,in order to improve the optimization accuracy of PSO algorithm,to adjust the algorithm the weight value and learning factor are improved,through the comparative analysis shows that the improved PSO algorithm has better search ability.According to the probability distribution of rock and soil parameters,combined with Monte Carlo simulation method to calculate the slope instability probability and reliability index in natural and saturated conditions.(3)By writing a particle flow analysis program to simulate the process of damage landslide,which can predict the disaster area of the landslide.The establishment of geometric model is based on terrain elevation data and spatial interpolation algorithm,and imported into particle flow software(Particle Flow Code,PFC)to generate particles.Through the uniaxial compression simulation experiment,obtaining the meso parameters required for PFC simulation calculation.Obtaining the particle movement speed,displacement and final accumulation form from the set monitoring points to calculate the hazard range of the landslide(4)We are counted the types and number of disaster-bearing bodies in detail,according to the international calculation formula,analysed the risks of population casualty and property loss.The results show that under saturated conditions,the maximum population casualties of Banling landslide is 0.41 person /year,and the maximum property loss is 4.836 million RMB / year.Among them,the risk of population casualties has exceeded the standards allowed by the international community.