Experimental investigation and ANN prediction of Al-SiC nano-composite properties produced using dynamic compaction method

Application of metal matrix composites reinforced by ceramic phase is increasing rapidly in different industries. Recently, dynamic compaction has been considered as one of the complementary methods of powder metallurgy in production of composite materials. Moreover, aluminum, as one of the most important industrial metals, is widely used in manufacture of metal matrix composites. In this study, the effects of dynamic compaction on pure and composite aluminum powders are studied. To produce composite specimens, pure aluminum and nano-silicon carbide powders are used as matrix and reinforcement respectively. Dynamic compaction experiments are performed using gas-gun apparatus in different velocities. Afterwards, density, strength and microstructure of compacted samples are evaluated. The obtained results show density, strength and spring back of pure samples are increased with increasing compaction velocity. Ceramic particles increase porosity in microstructure of composite samples. Adding 1 weight percentage of ceramic increases the strength of samples and leads to uniform distribution of particles into metal matrix, whereas adding more than 1% causes significant decrease in strength. In the following of the study, experimental results obtained from dynamic compaction and the difference between experimental and predicted results are considered as input and objective function of artificial neural networks optimization method. The obtained results show that this empirical-numerical method is able to predict density of specimens obtained from dynamic compaction with a good accuracy.