Analysis Of Landslide Stability And Landslide Prediction Based On Neural Network

Our country have Mountainous area ratio of major and complex weather conditions. geological disasters occur frequently in my country, that makes our country to become a serious disaster losses country. Landslides disasters have a large proportion of disasters. It seriously affected the normal operation and regional economic development. Casualties and economic losses results from Landslides disasters is very large. Establishing an effective landslides stability analysis system model and further achieving stability evaluation and prediction of landslide, is important to disaster prevention and mitigation in the project.On the basis of analysis of stability evaluation and prediction of landslide research status, discusses the factors influencing the stability of landslide and slope stability analysis methods; analyzing advantages and disadvantages of neural network research method, operability of evaluation of landslide stability analysis, and select BP neural network to analysis landslide stability.By analysis of factor, select average slope, slope of the invading front, sliding surface slope, the maximum annual average rainfall, the recent slide situation, human engineering activities as indicators of slope stability evaluation, factors influencing of landslides were extracted by engineering examples.On the basis of constructed the database, has established based on BP neural network landslide stability analyses and the prediction model; applying Matlab toolbox to design neural networks, weight initialization, data normalization and network training, based on that has carried on the training to the landslide stability analyses and the prediction model. The error analysis result indicate the sample training result and the actual situation tally basically, prove this model can tally with the project reality.Also, the model established in the paper has openness, as data continued to add and improve, the model can constantly optimize, and the accuracy of the model assessment may be further improved, have the important guiding sense to the landslide stable appraisal and the prevention decision-making.