Determination of Anisotropic Yield Coefficients by a Data-Driven Multiobjective Evolutionary and Genetic Algorithm

Abstract

The texture induced anisotropy of yield strength in cold rolled sheet metals is modeled using anisotropic yield criteria. The classical and other optimization methods used so far to determine the yield coefficients are limited by fixed set of experimental data, initial guess values, and pre-determined weight factors. A robust multiobjective optimization based on evolutionary algorithm proposed in this paper minimizes the error in yield stress and plastic strain ratio simultaneously and thereby overcomes the limitations in the approaches used before. The new approach is tested using Hill48 and Barlat89 yield criteria for five different materials from literature. The new approach is observed to improve the prediction capability of yield coefficients when compared to earlier approaches. The Pareto frontier obtained in the new approach can serve as a comparative tool to evaluate the accuracy of different yield criteria.

Keywords:

• Anisotropy
• Evolutionary algorithms
• Genetic algorithms
• Neural net
• Optimization
• Pareto front
• Plasticity
• Sheet metals
• Yield criterion