Research On Deformation Prediction Of Underground Engineering Based On Time Series Analysis

Since the reform and opening up,underground mining engineering,underground traffic engineering and underground integrated pipe gallery have been continuously developed.Underground excavation will inevitably lead to the redistribution of groundwater level and stress field,which may lead to different degrees of settlement and deformation displacement.When the formation moves and the surface deformation exceeds a certain limit,accidents such as roadway or foundation pit destruction,ground subsidence and surface building collapse will occur.In the underground excavation construction,how to accurately monitor and predict the occurrence of deformation is particularly important.In this paper,a underground mine in a underground project and a city underground project as the research object,the time series of monitoring data during excavation process using time series analysis method,the use of R/S analysis model,ARIMA model,BP neural network model,ARIMA-BP integrated model separately for a coal mine and a deformation of underground engineering time series analysis and forecast,and achieved good results.The main research work is as follows:(1)In view of the non-linear features in the time series of deformation data of underground engineering,and R/S analysis has good non-linear descriptive ability,the R/S analysis model is used in the research.At the same time,by using the BP neural network model,Combining capabilities ARIMA model,the ARIMA-BP integrated model is constructed to fully exploit the effective information in the time series of measured data.(2)Taking a coal mine as an example,the settlement trend of the key monitoring area of the surface subsidence in the mining area is predicted by using the R/S analysis model.The precision leveling data and actual mining conditions are used to verify the results.The prediction results are in good agreement with the measured data.At the same time,the ARIMA model,BP neural network model and ARIMA-BP integrated model are used to predict and analyze the six representative monitoring points in the surface subsidence area.The results show that the prediction accuracy of the integrated model is higher than the single model.(3)Taking an underground project as an example,the ARIMA prediction model,BP neural network prediction model and ARIMA-BP integrated model are used to analyze the settlement of ground surface,the vertical displacement of continuous diaphragm wall and the horizontal displacement of continuous diaphragm wall Monitoring data for predictive analysis.The results show that the above prediction model has certain predictive ability for the above monitoring data,and the same integrated model is better than the single model in predicting the results.However,the integrated model of ARIMA-BP error correction has the best prediction accuracy for the monitoring data.