Neural network approach to prediction of bending strength and hardening behaviour of particulate reinforced (Al–Si–Mg)-aluminium matrix composites

In this study, Al2O3/SiC particulate reinforced (aluminium matrix composites) AMCs, which was produced by using stir casting process, bending strength and hardening behaviour were obtained using a back-propagation neural network that uses gradient descent learning algorithm. First, to prepare the training and test (checking) set of the network, some results were experimentally obtained and recorded in a file on a computer. In the experiments, dual ceramic powder mix was prepared by chemical route then inserted in A332 alloy in melt condition by stir casting process. While SiC particles were supplied commercially, Al2O3 was chemically produced from aluminium sulphate. Al2O3/SiC powder mix was prepared, and heated up in ceramic crucible in the furnace. Then Al2O3/SiC ceramic cake was produced . This ceramic cake was milled to adjust particle size before stir casting. This dual ceramic powder with different SiC particle size range was added into liquid matrix alloy. The effect of the reinforcing particle size on bending strength and hardness resistance of Al2O3/SiC reinforced AMCs was investigated. Mechanical tests showed that bending strength and hardness resistance of 10 vol% Al2O3/SiC dual ceramic powder composites decrease with increasing reinforced SiC particle size. In neural networks training module, different SiC (μm) particles size range was used as input and bending strength, hardening behaviour and also porous properties in the produced AMCs were used as outputs. Then, the neural network was trained using the prepared training set (also known as learning set). At the end of the training process, the test data were used to check the system accuracy. As a result the neural network was found successful in the prediction of bending strength, hardness behaviour and also porous properties for any given SiC (μm) particles size range in the produced AMCs.