Multi-Objective Genetic Algorithm to Optimize Variable Drawbead Geometry for Tailor Welded Blanks Made of Dissimilar Steels


Abstract

Formability of a tailor welded blank (TWB) is affected by the strength ratio of the base metals joined. In this paper, formability of TWB with very high strength ratio made by joining twinning-induced plasticity (TWIP) and low carbon steels is numerically studied using a limiting dome height test. The drawbead geometry at the weaker side is modified to increase the dome height. The design of drawbead is optimized by treating it as a multi-objective problem with maximum dome height and minimum weldline movement as objectives, which were constructed as metamodels through a genetic algorithms based approach. The necessary data for the metamodeling are generated by finite element (FE) simulation using the commercial solver, LS-DYNA (R). The multi-objective optimization is carried out using a predator-prey genetic algorithm. The Pareto front estimated using this evolutionary approach is validated using FE simulations and a good correlation is obtained.