アルミニウム鋳造合金における介在物群の幾何学情報に基づく粒子破損寿命評価

抄録

In cast aluminum alloys, it is found from an experiment that a fatigue crack is initiated from the fracture of Si particles. We have been investigating a methodology to evaluate the particle fracture life from the geometrical parameters of Si particles. By using the synchrotron radiation CT, the geometrical parameters of Si particles were measured, and the image-based finite element analyses were also performed to evaluate the mechanical parameters of Si particles. A few effective geometrical and mechanical parameters were selected to determine the fracture of Si particles. The particle fracture life of Si particle was determined from the chronological observation of CT images. Then, the relationship between the particle fracture life and the geometrical and mechanical parameters were learned by the artificial neural network to predict the particle fracture life of unknown Si particles. In practice, forty three Si particles including the fractured and nonfractured particles were selected from five fatigue specimens, and used for the finite element analysis to evaluate the mechanical parameters. Their geometrical and mechanical parameters were used as the training data for the neural network. The validity of the predicted particle fracture life was examined through the comparison with the actual parameters of fractured particles.