Detection and quantification of flaws in structures by the extended finite element method and genetic algorithms

This paper investigates the extended finite element method (XFEM)-GA detection algorithm proposed by Rabinovich et al. (hit. J. Nuttier Meth. Engng 2007; 71(9):1051-1080; Int. J. Numer Meth. Engng 2009; 77(3):337-359) on elastostatic problems with different types of flaws. This algorithm is designed for non-destructive assessment of structural components. Trial flaws are modeled using the XFEM as the forward problem and genetic algorithms (GAs) are employed as the optimization method to converge to the true flaw location and size. The main advantage of the approach is that XFEM alleviates the need for re-meshing the domain at every new iteration of the inverse solution process and GAs have proven to be robust and efficient optimization techniques in particular for this type of problems. In this paper the XFEM-GA methodology is applied to elastostatic problems where flaws are considered as straight cracks, circular holes and non-regular-shaped holes. Measurements are obtained from strain sensors that are attached to the surface of the structure at specific locations and provide the target solution to the GA. The results show convergence robustness and accuracy provided that a sufficient number of sensors are employed and sufficiently large flaws are considered. Copyright (C) 2009 John Wiley & Sons, Ltd.