Research On Metro Tunnel Deformation Prediction Based On Wavelet And Time Series Analysis Combined Model

The major cities in China are in the construction of efficient subway tunnel network nowadays. In the construction and operation of subway tunnel, the deformation will be affected by many factors. If the deformation is not controlled will lead to serious consequences. It is of great significance to establish an effective deformation prediction model, while each model has its own limitations, the single prediction model sometimes can't meet the accuracy requirement. It is necessary to integrate and optimize the existing models. The subway tunnel deformation data has the characteristics of dynamic,relatively stable and containing noise. Time series analysis has a good effect in dealing with the dynamic stationary signal and wavelet analysis can effectively remove the noise in the original signal, that can be used as a preprocessing tool to improve the accuracy of prediction. In this paper the combination of wavelet analysis and time series analysis is used to analyze and predict the deformation data of subway tunnel. The main contents of this paper are as follows:(1) Prediction method and deformation analysis of subway tunnelStudy the deformation data processing and prediction methods, analyze and compare the characteristics of common methods from the theoretical basis, analysis methods, data requirements and research priorities. Study on the factors influencing the settlement of metro tunnel structure, method for laying reference network, the technical requirement of survey. The monitoring points, the main structure of the station,the deformation of each section, and the characteristics of single point deformation data of subway tunnel are analyzed based on Nanjing metro line ten tunnel construction settlement data.(2) Research on wavelet analysis and time series analysis modelThe basic theory of wavelet transform and threshold denoising is studied. The denoising effect is evaluated by changing the wavelet function and threshold estimation method in order to select the wavelet function and the estimation method of de-noising threshold which is suitable for the data of this paper. Study the classification and characteristics of time series analysis and basic principles of AR, MA, ARMA model.Focusing on the method of model identification and parameter estimation in modeling.Modeling and prediction of subway tunnel deformation data by time series analysis.(3) Construction of combined model and case verificationTwo combined models are constructed with the characteristics of wavelet analysis and time series analysis. The superiority of the combined model is verified by comparing with a single model, because the combined model removes the noise, makes the signal smoother. Time series analysis give full play to the advantages. The optimized model has better prediction effect by evaluating indicator. Predict the settlement of Zhongsheng station, Longhualu station and region between Zhongsheng station and Yuantong station to analyze the deformation trend and the cause of deformation, which is conducive to timely detection of problems and take appropriate measures.