Subway Wall Horizontal Displacement Monitoring And Prediction Model Research

With the accelerated pace of urbanization in China, subway are rising a nationwide construction boom. In order to protect the safety of subway construction and operation, deformation monitoring and prediction should be conducted. The aim of deformation monitoring obtains the information of the deformation observations, with deformation information analysis deformation conditions scientifically and accurately, to grasp the deformation characteristics and regular pattern, thus, to accurately analysis and forecast the deformation trend, in order to provide decision-making basis for project construction. Currently, Deformation monitoring and forecasting methods are regression analysis, time series, gray model, Kalman filter, neural network, combined model and so on.All the above single predictor model are used to predict the deformation trend of the deformation body with the original deformation monitoring data. In order to achieve improvements on a single prediction model, getting more accurate and longer prediction, in this paper, the traditional predictable pattern that only using raw monitoring data of the deformed body is changed, while it starts with the deformation itself and the factors which are caused deformation. Considering it has received and will be received deformation data in the future, which are subject to the impact of external conditions. It uses the polynomial least squares fitting and self-regression, by adding the impact of external conditions change to deformation monitoring data of deformable body, in order to achieve the amendment of the original monitoring data, which belongs to the deformation body. Using the method of time series predict the amended deformation monitoring data, it finds that the effects are better than the predicted effects of the original deformation monitoring data with time series.In this paper, it uses the first-phase project of Jinxing Station excavation monitoring data as the basis in Kunming Subway Line No.2, due to the top wall of subway enclosure structure are effected by the force of the steel structure support. Therefore, the experimental data adopt the horizontal displacement point SP5and the axial force measured point ZL1-4, the horizontal displacement point SP4and the axial force measured point ZL1-3, the points are all at northern end of the pit, the two sets of measurement points are located at the same point location; to the same period data, using sixty period datas predict ten period datas, as to the same period data, using seventy period datas predict ten period datas, it has four experiments in all. It applies polynomial least squares fitting and regression to amend horizontal displacement of the top wall, that is, by adding the effect of steel support. The amended horizontal displacement uses the time series to predict, and find the effect, which is superior to the uncorrected top wall horizontal displacement time-series prediction. Experimental process and results all achieve adopting the MATLAB computer program, the workload are reduced, and achieve to the desired effect.