Dresden 2003 – scientific programme
Parts | Days | Selection | Search | Downloads | Help
M: Metallphysik
M 41: Postersitzung
M 41.19: Poster
Thursday, March 27, 2003, 14:30–16:30, HSZ
Stochastic Modelling of Microscopic Particle Distributions — •Jürgen Gegner1, Christoph Henninger2, and Andreas Öchsner2 — 1SKF GmbH, Department of Materials Physics, Ernst-Sachs-Strasse 5, D-97424 Schweinfurt — 2Institute of Applied Mechanics, University of Erlangen-Nuremberg, Egerlandstrasse 5, D-91058 Erlangen
Advanced process simulation, e.g. in the framework of material design, often has realistic model distributions of spatial particle arrangements as a prerequisite. The input data generally stems from micrograph series evaluated by digital image analysis. The authors present a versatile mathematical tool that derives two- or three-dimensional representative model distributions of definite size by means of a stochastic search method. A certain particle arrangement can be modelled by positioning objects step by step by a random constructive heuristic. This technique requires rather short computational time and works well when only nearest-neighbour distances are considered. Higher neighbourhood levels, however, favour an improvement heuristic, which transforms an arbitrary particle arrangement into the desired one by moving the existing objects in small steps. In both methods, the task is formulated as an optimization problem, where the objective function expresses the difference between actual and rated distance distribution. The task of setting the sizes of the objects is solved separately.