NEURAL NETWORK PREDICTION OF THE EFFECT OF SEMISOLID METAL (SSM) PROCESSING PARAMETERS ON PARTICLE SIZE AND SHAPE FACTOR OF PRIMARY a-AI ALUMINUM ALLOY A356.0

Problems such as the difficulty of the selection ofprocessing parameters and the large quantity of experimental work exist in the morphological evolutions of Semisolid Metal (SSM) processing. In order to deal with these existing problems, and to identify the effect of the processing parameters, (i.e. shearing rate-time-temperature) combinations on particle size and shape factor, based on experimental investigation, the Artificial Neural Network (ANN) was applied to predict particle size and shape factor SSM processed Aluminum A.356.0 alloy. The results clearly demonstrated that, the ANN with 2 hidden layers and topology (4, 2) can predict the shape factor and the particle size with high accuracy of94%. The sensivity analysis also revealed that shear rate and solid fraction had the largest effect on shape factor and particle size, respectively. The shear rate had a reverse effect on particle size.