Prediction of Slope Deformation Time Series Based on Quasi-Newton
Shiguo Sun, Zhenhua Su, Yinghai Zhang, Bo Tian, Pei Xi Guo · 2015
In mining production, the slope's stability is one of the decisive factors of safe production of mine.Security issues and economic losses caused by slope's instability were immeasurable.Predict the slope's deformation according to the monitoring data of slope's displacement is an effective and feasible way.In order to predict the landslide accurately, ensure mine's safety and reduce the loss of property and casualties, set up mathematical model ARMA(n,n) based on time series method, determined the order of the model by improved Pandit-Wu method, determined parameters by Quasi-Newton method of engineering optimization algorithms, at last calculate the dynamic predictive values of a mine's slope displacement monitoring data.The results showed that, the method can improve the accuracy of prediction, and provided a reliable method for all kinds of slope landslides' forecast and early warning.