Fracture mechanics analysis of cracked structures using weight function and neural network method

2018 International Conference on Material Strength and Applied Mechanics

材料科学;力学

Chen, J.G.^1 ; Zang, F.G.^1 ; Yang, Y.^1 ; Shi, K.K.^1 ; Fu, X.L.^1

Science and Technology on Reactor System Design Technology Laboratory, Nuclear Power Institute of China, China^1

关键词: Fracture mechanics analysis;    Neural network method;    Probabilistic characteristics;    Probabilistic fracture mechanics;    Probabilistic methodology;    Probability of failure;    Thermal-mechanical load;    Weight function method;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/372/1/012013/pdf
DOI  :  10.1088/1757-899X/372/1/012013

来源: IOP

【 摘 要 】

Stress intensity factors(SIFs) due to thermal-mechanical load has been established by using weight function method. Two reference stress states sere used to determine the coefficients in the weight function. Results were evaluated by using data from literature and show a good agreement between them. So, the SIFs can be determined quickly using the weight function obtained when cracks subjected to arbitrary loads, and presented method can be used for probabilistic fracture mechanics analysis. A probabilistic methodology considering Monte-Carlo with neural network (MCNN) has been developed. The results indicate that an accurate probabilistic characteristic of the KIcan be obtained by using the developed method. The probability of failure increases with the increasing of loads, and the relationship between is nonlinear.