Study On Foundation Pit Deformation Prediction Based On Neural Network And Modeling Analysis

With the rapid development of urbanization, subway projects and high-rise buildings are increasing. Security and stability of foundation pit engineering has an important impact on the surrounding buildings and the superstructure. The deformation of foundation pit needs to be closely monitored and predicted in the excavation process. The study and improvements of foundation pit deformation prediction have an important role in construction control and accident prevention. They also have a wide range of reference significance to other projects.This paper briefly introduces the deformation characteristics and monitoring technology of foundation pit. And then the application of artificial neural network in deformation prediction is studied systematically. The construction method of the prediction model based on the finite element model characteristics and the artificial neural network model is summarized. The approximate modeling and the deformation prediction fusion model of neural network is conducted by using the monitoring method and measured data analysis of a foundation engineering instance and the prediction results of the fusion model are compared an analyzed.In this paper, it is taken into account that the neural network model has deficiency to reflect the mechanical relations and relatively good numerical proximity. In addition, considering the complexity of the actual situation of the foundation pit structure, the structure and basic excavation conditions of a certain foundation pit instance are approximately modeled by using Abaqus finite element software. The changes in characteristics of the model deformation are introduced to neural network model. The BP neural network fusion model construction and prediction processing is achieved by using Matlab development platform. At last, the fusion model deformation predictions are compared with the predictions of neural network model based on the measured deformation data. Comparative results show that the fusion deformation prediction model is relatively more precise which verify the feasibility of the fusion model. And it comes to a conclusion that fusion model can better reflect the mechanical relationship characteristics of the internal excavation structure during the construction process and has a positive reference for the study and improvement of safety control and prediction methods of engineering construction.