Prediction of Shallow Buried Pipeline Stress Based on BP Neural Network

• 1、College of Science,Wuhan University of Science and Technology,Wuhan 430065

Abstract：In this paper, the finite element analysis of shallow buried underground pipeline under the action of vertical uniform load is carried out. Considering the influence of soil characteristics around shallow buried pipeline, the interaction between pipeline and soil is analyzed, and the Pipe-soil interaction model based on the elastoplastic criterion of Drucker-Prager is established.And the stress of different sizes of pipelines under different loads are obtained by ABAQUS parametric modeling. Based on the large amount of data obtained, BP neural network is used to predict the stress state of shallow buried underground pipeline under the vertical uniform load. In addition some suggestions for the selection of BP neural network training samples are made for good prediction result. The results show that the prediction results of the elastic and small deformation stage pipeline are in good agreement with the finite element calculation results.And it can make reference for urban pipeline area construction transportation and planning.

Keywords： Mechanical simulation ABAQUS BP neural network finite element