Fatigue test and in situ observation were performed using a cast aluminum alloy AC4CH-T6, the fatigue cracking behavior was observed. Then the geometry and distribution of inclusions in the cast aluminum alloy is measured by using the CT images. The relationship between the geometrical parameters of the particles and the particle fracture life is evaluated and learned by the artificial neural network. Using the trained artificial neural network, the particle fracture life of particles in other CT image was predicted.
