抄録
This paper describes a study on an inverse analysis method to identify the crack parameters (crack position and crack depth) of a plate using the genetic algorithm (GA). The identification method is based on the eigenvalue change of the cracked plate. The eigenvalues of the plate are calculated by FEM based on Bogner element and are tabulated in advance. Using these tabulated data, spline surfaces are formed and the eigenvalues under arbitrary crack parameters are approximately evaluated on the surfaces. As a result, we can dispense with a large amount of FEM calculation that is necessary to evaluate the fitness function. By maximizing the fitness function in GA, we can predict the crack parameters. To demonstrate the effectiveness of the proposed method, the several identified results are presented.
