Research On Foundation Pit Deformation Monitoring Based On Multiple Linear Regression Analysis Model

With the continuous advancement of urbanization,the city scale expands rapidly.With the high-rise buildings,large shopping malls and underground projects constructing,the scale of foundation pit project also keeps increasing.It is inevitable that the horizontal and vertical displacement of foundation pit occurs during the whole excavation process.The monitoring of foundation pit deformation is an essential requirement of construction.The processing and analyzing of the data by collecting monitoring data,the safety of foundation pit construction process can be effectively guaranteed and the safety of subsequent excavation can be predicted and analyzed.Therefore,valuable suggestions will be put forward for the subsequent construction decision by modeling the measured data and predicting the deformation.After years of development,the monitoring methods of foundation pit excavation include regression analysis method,wavelet analysis theory,kalman filtering model,gray theory model,artificial neural network and time series model.In this paper,multiple regression models are mainly used as the basic model for research.Through the optimizing parameters of independent variables and optimizing coefficient of multiple regressions,the overall optimization of multiple regression models is finally realized.Through the comparison of the optimization sequence of the process multiple regression model,the prominent characteristics of the optimized multiple regression model are analyzed,and the optimization method can be brought into full play in the real project according to the actual construction environment.Based on the above data processing process and analysis method,the multiple regression analysis model is applied to a foundation pit monitoring project in a certain city.The datum network of horizontal and vertical displacement is studied,and the above model is built in both horizontal and vertical directions.The experimental data show that the deformation of the point data far from the bottom of the foundation pit in the horizontal direction is relatively large,which is affected by the construction after analysis.The long-term horizontal displacement statistics are distributed symmetrically around the starting datum,indicating that there is no obvious displacement at the monitoring point.Most of the subsidence values obtained at the vertical deformation point are within 6mm.After deleting the outliers,the detection data basically satisfies the linear law.In conclusion,the displacement of each monitoring point in the horizontal direction of the foundation pit is within the normal range,while the vertical monitoring point slightly sinks and gradually tends to be stable,and the overall foundation pit is safe and stable.The theoretical analysis of multiple regression model has a very reliable reference for the safety of the whole foundation pit construction process.