RESEARCH

Evolutionary Design of Finite Element Meshes for Injury Biomechanics Research

Lewis Payton, Auburn University
Brian Carnahan (2003-2004)

The objective of this project is to develop an affordable, accurate, responsive, and timely tool that will markedly facilitate and enhance injury biomechanics research.

Historically, Finite Element Analysis (FEA) has been used in lieu of human cadaver impact and loading studies, but it is expensive both in terms of computational time and the expertise required of the operator. This study seeks to simplify the input parameterization through the use of Evolutionary Computation (EC) techniques for finite element mesh design.

In an EC search algorithm, an initial population of parameterized solutions (i.e. mesh design rules) is randomly generated. A quantitative measure of each mesh’s performance (i.e. fitness) is determined through FEA. Through stochastic selection, those meshes with the highest fitness undergo crossover (parameter values between meshes are totally or partially exchanged) and mutation (parameter values within a single mesh are randomly altered).

The new mesh designs created through crossover/mutation replace the original meshes in the population, and then the process of selection, crossover, and mutation is repeated. Over a series of generations, the EC search algorithm discovers solutions that are highly fit in terms of the performance criteria.

In regard to injury biomechanics, the proposed method could be used to design innovative and effective mesh structures for modeling anatomical structures under impact conditions.

The general parameterization method developed should be applicable to a variety of FEA software packages, permitting widespread use in other biomechanical applications.

More ...

Back To Current Research ...