Modelling of Process variables for fly ash based Al- 6063 composites using Artificial Neural Network

In this paper predictive model for Metal Matrix Composite (MMCs) has been developed with the use of Artificial Neural Network. Stir casting process has been used to fabricate the fly ash based AL-6063 particulate MMC. The hardness of fly ash based AL-6063 MMC is taken as output variable, however fly ash(FA) percentage of reinforcement in MMC, stirring speed of stirrer and pouring temperature of liquid phase of particulate reinforced MMC are considered to be input variable. This work is divided into two phases, in first phase twelve set of experiments have been performed with above mentioned input-output variables. Using these results artificial neural network has been trained with the help of feed forward back propagation technique in second phase. Maximum hardness of value 44.24 at 9 % of FA percentage at 730° C pouring temperature with 350 rpm stirring speed of stirrer was predicted through this model.