Search for content and authors
 

Calibration and evaluation of a combined fracture model for the microvoid formation that competes with shear in the polycrystalline microstructure by means of evolutionary algorithms

Jacek Jackiewicz 

University of Technology & Agriculture, Department of Mechanical Engineering, Prof. S. Kaliskiego 7, Bydgoszcz 85-791, Poland

Abstract

Changes of stress or strain ratios during an operation can have an effect on cracking during metalworking processes. Therefore, there is a substantive interest of the aluminum and steel industries of the automotive and aviation manufacturing sectors in numerical simulations of fracture processes of typical structural materials. The motivation of the present contribution is to establish a detailed calibration procedure for a combined fracture model for the microvoid formation that competes with shear in the polycrystalline microstructure. The combined fracture model is hierarchically arranged according to the structure of the material. The established model is related to the common fracture criteria. Therefore, the various competitive physical modes of fracture can be taken into consideration (i.e. (a) shear fracture, (b) material interface decohesion, (c) reduction in material strength caused by the presence of the tensile hydrostatic stress that may lead to the brittle mode of fracture or to the ductile damage evolution), from which the only one with the greatest fracture likelihood determines the overall fracture limit. A set of experimental tests is taken as a reference to computer simulations, which have been carried out by means of evolutionary algorithms. The applied and developed evolutionary algorithms well match to computer simulations of this kind because the members of a population of individuals, which represent identifying material parameters of the combined fracture model, adapt to its environment in order to find the fracture mode with the greatest likelihood, after many generations. In other words, the members of the population of the identifying material parameters tend to disappear if they cannot adjust, but the others, which can adjust well enough, will live to reproduce, and in consequence they will give better adjustable parameters of the model.

 

Legal notice
  • Legal notice:
 

Related papers

Presentation: Invited at E-MRS Fall Meeting 2007, Symposium G, by Jacek Jackiewicz
See On-line Journal of E-MRS Fall Meeting 2007

Submitted: 2007-05-14 20:50
Revised:   2009-06-07 00:44