Numerical Analysis And Artificial Neural Network Prediction For Deep Foundation Pit Retaining Structure Displacement

The foundation pit engineering incorporates the geotechnique and structure engineering into an organic integrity influenced by many factors. Not only is there the problem of safety and stability of foundation, but also the danger of the neighboring buildings, under ground pipes, municipal equipment due to the displacement of stratum under excavation.It is the key to the success of foundation pit design that whether the displacement of the foundation and surrounding. Soils can be forecast and controlled.This thesis is based on the monitoring project of Xinguang lightway first section pumping station foundation. In allusion to the complexity and non-linearity of the deep foundation system. Based on the real-time monitoring on the foundation pit retaining distortion, the foundation pit retaining distortion is simulated and analyzed by numerical analysis and artificial neural network technology. The main researches of this thesis are as follows:(1) The construction process of the deep foundation excavation was simulated and analyzed using the special geotechnical software FLAC. With numerical simulation, optimize the number of the steel support layers; discuss the foundation distortion influenced by beyond pressure.(2) The composing and the working principles of automatic monitoring system on foundation were dissertated. The Xinguang first section pumping station foundation monitor program was taken as an example, the monitor basis, the content and the result were discussed.(3) To predict the foundation pit retaining structure displacement as time goes on by the improved BP nerve network model. The practice suggested that the results of forecasting are satisfied with neural network models. Sequentially, it is proved that the method is reliable and practicable for predicting in underground engineering.