Multi-Objective Genetic Algorithm Based Optimization of Age Hardening for AA6063 Alloy

The present article attempts to optimize the process parameters of artificial ageing for an AA6063 Al-Mg-Si alloy using multi-objective genetic algorithm (MOGA) to simultaneously achieve the maximum ultimate tensile strength (UTS) and percentage of elongation (%El). For this, a feed-forward multi-layered perceptron artificial neural network (ANN) has been developed which is trained by the scale conjugate gradient back propagation algorithm. The dataset required for the model has been compiled from the experimental results of this study, as well as, from the open literature. The network consists of solutionizing time and temperature, storage time/pre-ageing, rate of quenching, ageing time and temperature as input variables and UTS, %El as their outputs. The developed ANN model establishes the interrelationships between the input and output variables which can serve as objective functions for the optimization, following the theory of Pareto-optimality. The Pareto solution generated from MOGA between UTS and %El assists to conclude that the desired combination of high strength and ductility has been achieved through slow cooling after solutionizing, high pre-ageing time and high temperature of ageing. Furthermore, the designed heat treatment schedule through MOGA has been applied to the selected alloy on an experimental basis which shows satisfactory results.