Nonlinear Mechanism Of Slope Evolution And Landslide Prediction

The paper concludes the research of slope engineering and landslide forecast.in the base, nonlinear theory and relative theory and methods are used to nonlinear mechanism of slope evolution and the characteristic chaos of slope and landslide prediction and forecast methods. In this paper, nonlinear theory and methods are used to study the predicting methods of slope deformation destruction. The formulations of relative dimension Da and Kolmogorov entropy K2,which reflect the dynamic characteristics of slope system, are obtained. The analyses of examples show that the deformation and failure process of landslide is chaotic and fractal. The fractal theory can be used to study landslide prediction.On the basis of displacement-time series of the slope,a nonlinear dynamic model is set up according to backus generalized linear inversion theory in this paper.due to the equivalence beween autonomous gradient system and catastrophe model,a standard cusp catastrophe model can be obtained through variable substitution.The method is used in analysis of displacement data of Huangci landslide and Wolongsi landslide and in understanding how slopes evolve before sliding.The result shows that the nonlinear dynamic model can make satisfactory prediction result.is it most important that there is a sudden fall of d,which indicates the occurrence of catastrophe(when D=0).The judgement formula for high and low speed landslide and a new theory of stiffness-Effect instability are presented for the instability of planar-sliding slope by the catastrophe theory. It is pointed out that the limit equilibrium stability analysis method of rigidity body has a lot of defects. The relationship between the action of outside environment and the response of slope system is complicatedly nonlinear according to the nonlinear dynamics model of slope evolution built. When the nonlinear action of slope itself is equivalent to the action of outside environment,the evolution process of slope is of chaos phenomenon and its road to chaos is realizedwith bifurcation of period doublings.By the characteristic analysis of slope displacementtime series,the largest Lyapunov exponent and saturation inseted dimension are obtained,and the longestforecast time is given. On this base,the number of inputing node of neural network isgiven,moreover,the method of forecasting slope displacement is presented based onadaptive neural network. Through the slope displacement forecasting of the Three Gorges Project list slope and Xintan slide,the results are satisfied. And this has muchmeaning to established slope displacement forecast-timely.