Determination of Material Constants of Khan-Liu Yield/ Fracture Criterion by Genetic Algorithm and Particle Swarm Optimization Methods

Optimal determination of material constants of behavior criteria with minimal number of experimental test data is of interest to designers. Khan-Liu yield/fracture criterion is one of the comparatively accurate and user-friendly criteria to predict behavior of alloys such as Ti-6Al-4V alloy. This criterion with ten constants can take into account effects of parameters such as asymmetry in tension and compression, anisotropy, hydrostatic pressure, strain rate and temperature as uncoupled. Evolutionary algorithms are optimally suitable methods for determining the materials behavior equations constants. In this article, the genetic algorithm and particle swarm optimization methods are used to determine the material constants of Khan-Liu criterion. Experimental results of uniaxial tension and compression tests in two directions, rolling and transverse direction of Ti-6Al-4V sheet, at different temperatures with 1 sec-1 strain rate are used. From these data equal-biaxial points were calculated. After applying two algorithms on these data, results showed that particle swarm optimization is better than genetic algorithm. Therefore, this method is suggested to determine the constants of this criterion. The material constants extraction methodology with two genetic algorithm and particle swarm optimization, as well as aspects of the development and improvement of Khan-Liu criterion are results of this paper.