Analysis Of Stress At Crack Tip Based On Neural Network

In this paper, we will find the nonlinear relationship between the coordinates(r, 9) and the stress at crack tip with the neural network, collecting the simples around the crack tip. Then we can get the precise stress field. With the field we get, we can describe the plastic field of the crack belongs to type I. This proves the feasibility and correctness of the application of neural network in the analysis of the stress at crack tip. Also we can get the plastic field of the symmetrical material and the bi-material with central perfoliate crack.After this, we can calculate the stress tri-axcity with the stress field. At a circle containing the plastic field we can get the point with the max stress tri-axcity. Then we can get the degree of the point. Comparing the degrees of rupture with the ones coming from experiments we find the results are in accord with the ones we get from experiment. Based on this, the author gives the result of the symmetrical material and the bi-material.In conclusion, the author thinks that the neural network can analyse the stress at crack tip and the stress tri-axcity degree can be the criterion in analysis of fracture.