Analysis And Prediction Of High-slope Model's Monitoring Data Based On Nonlinear Method

With more and more large-scale hydropower projects in mountainous areas and canyons, the height of slope project is becoming higher and higher, and the problems in it are becoming more and more complex and difficult. The instability of slope will not only affect the progress of the project which may bring about huge economic losses, but also cause casualties. So the high slope stability monitoring is one of the most important parts in hydropower projects, and the analysis on monitoring data and forecasts of high-slope behaviors have important social and economic significance.In this thesis, high slope is regarded as a dynamic system, and nonlinear science such as fractal theory and chaos theory are introduced to the prediction and analysis of monitoring data. The main contents of research are as follows:Analyzing the high-slope monitoring time series by Rescaled Rangle Analysis (R/S), high rocky slope is found to be a complicated nonlinear dynamic system with fractal character. Based on fractal theory, variable dimension fractal of forecasting model is set up and improved, whose results show that forecasting model is anti-noisy, reliable to small data sets, accurate and adaptable. Then according to chaos theory, there is the study of the phase space reconstruction of high-slope dynamic system and the selection methods of parameter. Finally, by analyzing the monitoring data, chaos character in system is proved, and high-slope monitoring prediction model by maximal Lyapunov exponent is set up.