The Prediction Of The Movement Risk On Mawo Landslide After The Destruction In Yiliang County,Yunnan

Landslide is a common natural disaster and can cause serious consequences, inthe event of deformation and failure, the consequences are catastrophic. The landslidemovement distance is a direct reflection of its risk assessment indicators afterdestroyed and deformation, and is an important basis for division of the danger zone,also is basic to carry out landslide disaster prediction and prevention work.Therefore,it is significance to predict of the landslide movement hazard after destruction.This paper takes Mawo landslide as the research object in YiLiang county Yunnanprovince, on the basis of site investigation, It is discussed about engineeringgeological conditions, deformation and failure characteristics and formationmechanism of this landslide. Based on the kinematics and the energy conservation lawto establish a computational model of landslide movement distance, by collecting65cases of typical disaster rainfall landslide data, and MATLAB is used to curve fitting ofthe calculation model,and use BP neural network model technology to cause thenonlinear mapping and prediction about landslide characteristics influence parametersand the sliding distance. Meanwhile by means of3DEC and centrifuge to analyze theMawo landslide risk of movement. The following is main conclusion results:(1)Through on-site investigation and engineering geological conditions oflandslide, identify characteristics of the landslide shape and size, material composition,structural features and characteristics of deformation and failure, infer the landslidebelonged to Weak Base Type Creep-Crack Failure Mode.(2) Established an sliding distance simplified kinematic model, which is based onthe basic principles of energy conservation and kinematic and characteristics of terrainconditions after landslide unstable, the model takes into account the impact of theirown factors and site factors landslide of landslide sliding distance, and gives aspecific formula and prediction methods. by collecting65cases of typical disasterrainfall landslide data, and MATLAB is used to curve fitting of the calculation model,finally gives the correction equation. (3) Take catastrophic rainfall typical landslide sliding distance data as sample, forthe suitability test of the prediction of kinematics equations, based on the basicprinciples and algorithm of BP neural network. Select the four factors as the inputmode,which is landslide height, angle of internal friction venue, the venue terrain,landslides slope, and select landslide calculated distance value from fitted formula asthe output mode, build4-9-1three layer BP artificial neural network model.Realization of nonlinear mapping and prediction about landslide sliding distancelandslide parameters and calculations. The results showed that: The errors of BPneural network model and the target value test is in the range of0to8.9%, and themost of the data is less than5%, the error value is small, description the applicabilityof the kinematic equations is good.(4)Mawo landslide movement distance was calculated by two theoreticalmethods, fitting formula of the kinematics and the BP neural network model. Theresults show, the final sliding movement distance in gentle section was283.5mcalculated by fitted formula,while BP was295.2m, therefore deviation was4.12%.(5) They, the landslide failure characteristics in natural state and the fittingformula suitability, were explored by centrifuge model test. The results showed: thecrack distribution in the centrifuge model test was similar to realistic landslide, anddeformation mechanism was Weak Base Type Creep-Crack Failure Mode.Themovement distance in gentle section was0.41m, restore to the actual sliding distancewas82m,therefore deviation was5.03%. Indicating the applicability of fitting formulawas good.(6) By3DEC, Mawo landslide movement risk was simulated. The results showed:The lateral displacement (x direction) at left of landslide (facing slope outside) is103.69,right (facing slope outside) is64.6m. The maximum displacement in thevertical direction (y direction) is342.51m, and in the longitudinal direction (zdirection) is354.03m. Thus, we can obtain the Mawo landslide hazard area afterdeformation and failure: On the left side of landslide (facing slope outside) was105m,and left was65m, the range of distance is345m from Longitudinal horizontal ofthe landslide.