NEURAL NETWORK SIMULATION ON MECHANICAL PROPERTIES OF ZA22/Al₂O_3f COMPOSITES

The typical model of neural network e.g.B-P algorithm was adopted to analyze as well as predict the mechanical properties of ZA22/Al₂O_3f composites.The results show that B-P model is not only suitable for the regression and prediction for the single variable output non-linear system,but also suitable for the multi-variables output non-linear system.Perfect fault tolerance ability was also found in the B-P model while it was used to analyze the mechanical properties of ZA22/Al₂O_3f composites.