Researchon The Intelligent Diagnosisof Stability Of High Slope Under The Condition Of Rainfall Based On GIS

This thesis mainly study high slope area intelligent regional evaluation underrainfall condition and its related problems, through the uniform designmethod(UDT),saturated-unsaturated theory, the limit equilibrium method, the neural network theoryand GIS, the rainfall situation based on the data of Fengjie to evaluate the stability ofhigh slope area. In this thesis, the high slope stabilityunder rainfall as the researchobject, established a intelligent regional assessment system based on GIS platform andprovides a positive and practical way to explore the intelligent GIS and got the highslope stability intelligent regional diagram of Fengjie based on GIS rainfall, which havecertain guiding significance of Fengjie disaster prevention. In this thesis, the mainprogress is as follows:①On the basis of rainfall infiltration mechanism of slope stability studies, researchthe catastrophic factors of high slope instability under the typical rainfall condition, todetermine six input parameters of the experimental sample library: Rock mechanicsparameters (cohesion, angle of internal friction), physical parameters (severe), thegeometric characteristics factors (slope height, inclination), rainfall factors (dailyrainfall). To determine an output parameter of Neural network training sample database:the safety factor.②To establish numerical experiments to determine the sample database approach:uniform design.③numerical experimental models based on numerical experiments sample library.Based on Geo-studio limit equilibrium method, the slope safety factors werecalculated.Taken into account of the effects of rainfall infiltration, combined withsaturated-unsaturated theoretical, numerical experiments models and parameters wereestablishedand selected.④On the basis of numerical experimental models and numerical sample databaseto establish neural network training samplesand organize data, deal with the singularityand convergence problems.⑤BP neural network established by MATLAB, try to Establish a mapping bytraining. Through neural network learning training data, the neural network systemmapping rules automatically generated, based on the expected error so that the neuralnetwork weights and thresholds to automatically match their map. ⑥based on GIS to establish disaster factors raster data andon the basis of"Fengjie detailed geological disasters investigation report" Creating a data source rasterdata, reclassified on this basis, established different levels of Catastrophe factorsregional map.⑦Combined with neural network system to take GIS spatial analysis, achieve GISintelligence. Made use of Fishnet and Spatial Join to extract GIS model data, importneural network to compute, the calculation results by Access GIS processing re-importthe model, and then reclassified raster conversion and output of different colors regionalassessment diagram.In this thesis, the high slope stabilityunder rainfall as the research object, combinedwith uniform design, saturated-unsaturated theory, limit equilibrium method, neuralnetwork theory and GIS, and compare previous studies to improve the effectiveness ofthe high slope stability intelligent regional assessment in case of rain.