Improved Time Series Analysis Model Applied To Subway Construction Settlement Monitoring And Forecasting

The scale of China’s urban subway will rapidly grow during "Twelve Five" period.Therapid development of the metro-scale causes the accidents in subway construction occurfrequently. In order to ensure the safety of subway construction. Prevent not timely data analysisand prediction, forecasting inaccurate triggered subway construction accident. It is essential toestablish a system of deformation data monitoring and forecasting during subway construction.In response to this situation, the article develops a system which meets the practical applicationof the system of subway construction. The system is developed in accordance with themonitoring of the construction phase of the project, data handling processes as well as thetechnical specifications strictly. Short-term prediction of subsidence monitoring data is a corepart of the system,there is a better short-term prediction of time series analysis model tosubsidence monitoring data, therefore, from the basic theory of the time series, the combinationof subsidence monitoring data and to explore the shortcomings of time series analysis model.Study found that the advantages of wavelet analysis theory in signal processing to make up forlack of time-series analysis model: First, after the settlement data after pretreatment hierarchicalthreshold wavelet denoising method denoising, in time series analysis model use the theexponential smoothing data to predict, this method can extract data effectively at the same time,to ensure the smoothness of the curve and the local high-frequency similarity is not lost, theexperiments show that, compared with data after hierarchical threshold wavelet denoisingforecast separate time series model predictions more accurate data, the method is effectivesubsidence monitoring data processing can be applied to actual engineering practice, with theactual value.