Pareto-based multi-objective hot forging optimization using a genetic algorithm

Abstract

Efficient control techniques must be preceded by well-designed processes. A generally accepted definition of a well-designed process is one that is Pareto optimal, i.e., no design objective can be improved without degrading at least one other design objective. Indeed, optimal design enables effective trade-off of competing design objectives, including controllability and robustness goals. The main goal of the present work is the design of a Pareto-based multi-objective optimization model for multi-stage hot forging processes. The optimization methodology considers a genetic algorithm supported by an elitist strategy. An iterative procedure is considered and Pareto optimal solutions are found managing the drawing of the Pareto front and enabling the extraction of optimal solutions according to selected preferences. The design example consists of a two-stage forging process applied to a pre-heated billet of AISI 1018 steel. Two sets of variables are considered: shape design variables and process variables. The objective functions are related to the minimization of the forging load, the control of the forged shape and the material microstructure.