engleski

Genetic algorithm implementation in material behavior simulation system

Constitutive modeling of material behavior is becoming increasingly important in prediction of possible failures of highly loaded engineering components, and consequently, optimization of their design. Material behavior is characterized by material parameters, which need to be identified for material model definition. Material models differ in range of material properties they can describe and proportionally in their complexity to define. Complex material models, such as Chaboche's, are characterized by large number of material parameters that have to be carefully identified to follow material behavior as accurately as possible. Therefore modern numerical methods for parameter identification, such as genetic algorithm, proved to be necessity in modeling material behavior today. Because of the complexity of the presented problem, unique genetic operators’ routines in genetic algorithm calculation procedure are developed to enable fast and reliable convergence of the results. These routines, chosen to build an optimal genetic algorithm procedure are proposed in this paper. Within it the material response of the steel 42CrMo4 is simulated by using an automated system for parameter identification to verify the proposed procedure.

Genetic algorithm; Material behavior; Simulation

nije evidentirano