Nonlinear Time Series Predication of Slope Displacement based on Smoothing Filtered Data

Abstract

According to the slope in geotechnical engineering, many displacement monitoring points are usually set to obtain the displacement data to ensure slope stability, these data are typical nonlinear time series, and it has high value about how to make use of displacement monitoring data to do the next step forecast analysis. Due to a certain degree of error, smoothing filter method is used to pretreat the displacement data, eliminate the influence of the error on the results and ensure the rationality. Based on smoothing filter data, these two methods are proposed to predict the displacement of the slope: exponential smoothing and chaos neural network. Both methods are used to make predictive analysis of the displacement monitoring data of outer monitoring point TP/BM27 in high slope of Three Gorges Ship-Lock, forecasting results show that: predictive values are close to measured values, chaos neural network prediction method is better than exponential smoothing method. At the same time, the displacement data with higher reliability and smoothing filter processing are used to make predictive analysis, the results can be more reasonable, so smoothing filter processing plays an important role in the analysis of displacement prediction.