Estimation Study of Stress Concentration Factor of Crack Structure Based on BP Neural Networks

摘要

Stress concentration factor is an importantparameter to describe the stress concentration phenomenonand crack propagation process of structure. In order torespond the fatigue characteristic of crack structure, quantitative prediction model of stress concentration factor isbuild. The stress of box structure is analyzed by using theFEM. The stress state used for the calculation of stressconcentration factor is extracted from the path between rootsof crack and the stress release area according to the FEMresult. Then the stress concentration factor is solved accordingto the extracted stress state. The prediction model of stressconcentration factor based on BP neural networks is achievedby nonlinear training the data, the parameter of crack is theinput of BP neural networks and the stress concentrationfactor is the output. The average prediction accuracy up to91.4% is achieved by using the nonlinear mapping network, which makes the precise estimation study of relationshipbetween stress concentration factor and crack parameter come true.