INVERSE ANALYSIS OF INTERNAL DEFECTS IN STRUCTURES USING EXTENDED FINITE ELEMENT METHOD AND IMPROVED ARTIFICIAL BEE COLONY ALGORITHM

The information of defect in a structure is determined using the extended finite element method combined with a kind of improved artificial bee colony algorithm, based on real structural response. As searching optimal value may appear in an arbitrary direction in traditional artificial bee colony (ABC) algorithm, weighted average mutation and a cross operator are introduced to avoid the local optimum in the optimizing. The presented inverse method is also used to determine the location of a single-circular-like defect and an elliptical-like defeat and two irregular defects, and the robustness of the algorithm under the condition of measuring error is also studied. The numerical results indicate that the adapting artificial bee colony (AABC) algorithm proposed can present the real information of defects accurately. The convergence speed of AABC is faster than that of traditional ABC, and it is unlikely that the local optimum will appear in the optimizing. The presented method can locate the defects accurately and show a high robustness.