Deformation Monitoring And Prediction Analysis Of Foundation Pit In Soft Soil Area

Dig deep excavation,small area,no support or less support in engineering applications,this type of foundation pit slope stability is poor,often exist in the process of excavation supporting structure deformation and other unfavorable factors,the deformation data by construction conditions,precipitation condition parameters significantly affected.Therefore,if a foundation pit design scheme,construction plan combined with field monitoring data,a time series analysis and prediction method of foundation pit deformation can be put forward,which can greatly improve the safety of foundation engineering,and the results can also be used for the optimization design of foundation pit support scheme.Aiming at the special soft soil conditions in a certain area,this paper collects and sorts out the measured data of a foundation pit project,takes the horizontal displacement of the supporting structure'top as an example,constructs the BP neural network deformation prediction model,elaborates the network sample establishment,function selection,parameter setting,network training,inspection and operation,and predicts the deformation of the foundation pit in the later period.The main conclusions are as follows:In the BP neural network deformation prediction model,the optimal intermediate layer parameters corresponding to different input layer parameters are also different.Setting reasonably the input layer hidden layer parameters and proper training of the network,this model can predicted successfully the variation law and variation of the foundation pit.The error between predicted and measured values is small,and the deformation law is consistent.The model is suitable for foundation pit projects with similar hydrogeological conditions and similar excavation depth.The better results are obtained by finite element simulation analysis.