Analysis And Prediction Of Settlemeng Of High-rise Buildings Under Construction Of Subway Tunnel

With the gradual progress of urbanization,the scale of cities continues to expand and the flow of people to and from these city is increasing,which cause that the traffic jams have become the normality in many cities during rush hours.The subways,with many advantages like large volume,fast speed,low noise and no occupation of ground space,have become the first choice for many big cities to solve the traffic problems.During the construction of the subways,the original formation stability is destroyed,which will inevitably lead to the surface subsidence deformation.Moreover,the subways are often built in the bustling area with amount of ground buildings in these cities.So in order to ensure the smooth construction of the project and the safety of personnel property,the deformation monitoring of the buildings must be completed well.The prediction analysis of building deformation monitoring data can be used to know the building deformation trend in the coming period,which can help to take some safeguard measures in time to avoid some engineering accidents when abnormalities are found.At present,the common methods for data processing and prediction analysis of deformation monitoring include:regression analysis,time series analysis,wavelet analysis,gray system,artificial neural network,etc.In this paper,wavelet analysis and BP neural network are combined to predict the building settlement data.According to the tunnel construction project of a certain section of Zhengzhou Metro Line 3,this paper introduces the general situation of construction and the monitoring scheme of the project,and the Peck formula is used to simulate and calculate the surface deformation after construction.Besides,the general rules of the deformation of the surface building during the subway construction are analyzed based on the measured data of the construction subsidence monitoring JZ31 point.Finally,the combined model of wavelet analysis and BP neural network is used to make prediction analysis of the settlement data of the JZ31 point.In the actual deformation monitoring work,the obtained data often contain noise,which means that it’s necessary to carry out data noise reduction processing before conducting simulation prediction by using measurement data.The wavelet analysis method is used to detect the gross error of the settlement monitoring data of building deformation Monitoring Point JZ31,and the gross error detected is processed by linear interpolation method.And then,the influence of wavelet threshold selection mode,wavelet basis function and decomposition level on denoising effect is analyzed step by step,and the wavelet denoising model with best effect is selected in the end.According to the measured settlement data of Monitoring Point JZ31,the BP neural network prediction model,wavelet denoising BP neural network prediction model and wavelet transform BP neural network prediction model are established respectively,and the prediction results are compared and analyzed.The results show that the prediction accuracy of the combined model is higher than that of the single model,and the wavelet transform BP neural network prediction model is more suitable for the construction settlement analysis caused by subway tunnel construction by reason that it has the best effect.