Analysis And Prediction Of Mayang Landslide Stability Based On BP Neural Network

Mayang Miao Autonomous County is located in south-central Yuanling-Mayang Basin in Hunan Province, between Xuefeng mountain and Wuling mountain. It is one of the geological hazards prone areas in Hunan province. In order to find out the potential landslide geological disaster in Mayang, and scientifically and reasonably take disaster prevention and disaster mitigation measures to protect people’s lives and properties, a reliable model of landslide stability analysis and prediction are established, based on the project of “1:50000 Geological Hazard Detailed Investigation in Mayang Miao Autonomous County of Hunan Province”. And the main research achievements are as follows.(1) According to the characteristics of landslide geological disaster of Mayang Miao Autonomous County, this project takes the academic thought of “Mechanism analysis of geological process-quantitative evaluation” as the core, combines with the detailed field investigation of geological disaster, establishes a set of methods about landslide stability analysis and prediction, and forms the complete technology roadmap of landslide stability analysis. The research result is used to guide the regional geological hazards prevention and control, and achieves great application achievement. This technology roadmap and method enrich the research contents of landslide stability analysis and prediction, as well as has great influence on the landslide stability analysis under similar geological conditions.(2) Based on system analysis of the data of landslide in Mayang, and through the methods of mathematical statistic and numerical analysis, this project develops a detailed study of hazard type and distribution characteristics of landslides in Mayang. The results show that the hilly region where is 175-275 meters above the sea level is the high prone area of landslide geological disaster, and the landslide in red-mudstone-layer accounts for 88.37% of the total number, while consequent landslide accounts for 44.96%. The relation between landslide development probability and slope approximately presents normal distribution with μ=33.1,σ2=10.22, and when the slope is 23°~43°, the occurring probability of landslide reaches the peak position.(3) Based on the study of landslide hazard type and distribution characteristics, this project screens out the average grade, rainfall intensity and other six disaster-causing factors, establishes landslide stability analysis and prediction model of BP neural network in Mayang under different working condition of sun and storm, and verifies seven slope stability and its high accuracy rate, and proves that the prediction model has reference significance.(4) By using Geo Studio, this project carries out numerical simulation research of the rigid equilibrium in Mayang, and verifies that it is feasible to apply the Slope/W of Geo Studio to carry out the numerical simulation research of the rigid equilibrium.(5) This project explores the advantages and disadvantages of stereographic projection stability analysis method. Through the deduction, it can be found that the formula of ÷÷ is applicable to rock slope planar and wedge-shaped which are two different deformation and failure modes.Through the analysis of landslide hazard type and distribution characteristics in Mayang, and based on study of of BP neural network of Matlab for landslide stability analysis and prediction model in Mayang, this project establishes landslide stability analysis and prediction model which is applicable to Mayang Miao Autonomous County, and guides the work of prevention of landslide geological disaster. At the same time, it also provides other researchers with new research foundation and theory basis.