Study On Prediction Model Of High And Steep Slope Deformation Based On Fuzzy Time Series

Mining of mineral resources has always been an important part of the national economy,many high and steep slopes will be formed in the process of open pit mining,the mine slope is usually in the dynamic change project of continuous excavation.If the slope deformation exceeds a certain range,its stability will directly affect the construction and safety of the project.Therefore,it is necessary to establish and optimize the deformation prediction model of high and steep slope.The deformation of the high and steep slope of open pit is affected by the sudden change factors such as blasting and heavy mechanical pressure.Therefore,in this thesis,aiming at the special working condition of the high and steep slope of open pit,the fuzzy time series with obvious prediction effect on the sudden change factors is selected as the basic algorithm to establish the prediction model.Firstly,the single variable fuzzy time series model is studied,and the domain division of fuzzy time series model is improved by using optimization algorithm.Then,the multivariable fuzzy time series model is studied to establish the prediction model of slope monitoring.In this thesis,an improved fuzzy time series prediction model based on global distribution optimization algorithm is proposed.The whole distribution optimization algorithm is used to divide the domain,which is based on the characteristics of historical data.It overcomes the subjectivity of the method of average theory domain,improves the accuracy of data fuzzification stage,and makes the fuzzy relation more reasonable.Moreover,it is more accurate in the calculation of the prediction value,which improves the prediction accuracy.In this thesis,an improved two type fuzzy time series prediction model is proposed.In this thesis,the relationship among zenith distance,slant distance and azimuth angle is established by two-type fuzzy time series model,and the weight of each observation in the prediction process is allocated according to its correlation degree,so that the two-type observation can assist the one-type observation in prediction.The higher the correlation degree,the better the prediction effect of the first-order observation.Then,the whole distribution optimization algorithm is used to optimize the region division,which improves the prediction accuracy of the model.The simulation results verify the effectiveness and accuracy of the algorithm.