CAVITY DETECTION USING GENETIC ALGORITHM AND BOUNDARY ELEMENT METHOD IN ELASTOSTATICS PROBLEMS

    Authors

    D. Ojeda; E. Divo; A. Kassab;M. Cerrolaza

    Comments

    Authors: contact us about adding a copy of your work at STARS@ucf.edu

    Abbreviated Journal Title

    Rev. Int. Metod. Numer. Calc. Dise.

    Keywords

    Boundary element method (BEM); cavity detection; genetic algorithm; elastostatic; OPTIMIZATION; BEM; Engineering, Multidisciplinary; Mathematics, Interdisciplinary; Applications

    Abstract

    A method for the efficient solution of the inverse optimization problem of cavity detection using a point load superposition technique in elastostatics boundary element methods is presented in this paper. The superposition of point load clusters to simulate the presence of cavities offers a tremendous advantage in reducing the computational tine in the elastostatics field solution with the boundary element method as no boundary re-discretization is necessary throughout the optimization process. An objective function is formulated to minimize the difference between BEM-computed and measured deformations along some of the field boundaries. A Genetic Algorithm is employed as the optimization method to guarantee a global optimal solution within the design space of a highly non-linear and multi-parameter objective function. In addition, the Genetic Algorithm is parallelized and dynamically balanced to mitigate its commonly known efficiently issues. Results of cavity detection problems simulated using numerical experiments and added random measurement errors validate the approach in regular and irregular geometrical configurations with single and multiple cavities.

    Journal Title

    Revista Internacional De Metodos Numericos Para Calculo Y Diseno En Ingenieria

    Volume

    23

    Issue/Number

    4

    Publication Date

    1-1-2007

    Document Type

    Article

    Language

    Spanish

    First Page

    363

    Last Page

    377

    WOS Identifier

    WOS:000261304400002

    ISSN

    0213-1315

    Share

    COinS