Study On Subsidence Prediction And Numerical Simulation Of The Constructions Around The Deep Excavation

With the rapid development of China’s economy, many large deep foundation pit engineering appeared all over the country, of which a large part in the city center, the construction of foundation pit will make the surrounding buildings deformation, also bring some potential safety problems. Real time monitoring of buildings, the deformation data analysis and processing the measured, and the use of appropriate forecasting method to predict the deformation of foundation pit excavation construction, is an important means to ensure the safety of the surrounding buildings in the foundation pit during the construction period.The research content in deformation analysis is generally divided into physical interpretation of deformation and geometric analysis of the two part. Physical interpretation of deformation is the task of the relationship between the deformation body deformation and deformation is established, explain the reasons for deformation. Geometric analysis is the task of the state space description of deformation body under various loading conditions and characteristics. Be directe dagainst the deformation of constructions which cased by the deep excavation, firstly this paper using the modified gray system model deal with the data processing; secondly combining BP neural network model which have strong self learning features, improved the prediction error of the grey model. Through compared with the measured data, the combined forecasting model can achieve high accuracy; thirdly, In this paper, based on the geological data of deep foundation pit excavation, used the finite software FLAC3D to establish a model of deep foundation pit which consisted of supporting structure、surrounding soil and constructions. The parameters of the model are optimized and the deformation monitoring data of deep excavation at the early stage of building, and according to the model to simulate the deep foundation pit excavation, monitoring point the buildings on the prediction, the predicted results is slightly larger than the measured results.