FLAW IDENTIFICATION IN ELASTOMECHANICS - BEM SIMULATION WITH LOCAL AND GENETIC OPTIMIZATION

Single and multiple Raw identification problems are considered. Static and steady-state dynamic analysis of structures with flaw(s) is perfo rmed by the boundary element method. Inverse problems are formulated a s output (i.e. measurement) error minimization problems and they are s olved by numerical optimization techniques. As it is shown in this pap er by means of numerical experiments, for elastostatic cases, an appro priate modelling of the structural analysis problem, a good choice of the error measure, and the use of established numerical optimization s oftware are usually sufficient for the solution of the problem. Even m ultiple flaw identification is possible. Elastodynamic loadings lead t o nonconvex problems which are solved here by means of global, genetic optimization algorithms.