Modeling and Optimization of Roll-bonding Parameters for Bond Strength of Ti/Cu/Ti Clad Composites by Artificial Neural Networks and Genetic Algorithm

Abstract

This paper deals with modeling and optimization of the roll-bonding process of Ti/Cu/Ti composite for determination of the best roll-bonding parameters leading to the maximum Ti/Cu bond strength by combination of neural network and genetic algorithm. An artificial neural network (ANN) program has been proposed to determine the effect of practical parameters, i.e., rolling temperature, reduction in thickness, post-annealing time, post-annealing temperature and rolling speed on the bond strength of Ti/Cu composite. The most suitable model with correlation coefficient (R2) of 0.98 and mean absolute error (MAPE) 3.5 was determined using genetic algorithm (GA) and the optimum practice condition are proposed. Moreover, the sensitivity analysis results showed the post-annealing temperature with the negative effects is the most influential parameter on the strength of bonding.