Application of weighted sensitivity for identification analysis of defect topologies in structure based on level-set based topology optimization

Abstract

In this study, defect topology identification analysis based on the level-set-based topology optimization is carried out by using oscillation waveforms obtained by hammering test. In defect identification analysis, sensitivity may cause numerical oscillations that affect the identified shape. To solve this problem, we employed a weighted sensitivity. Weighted sensitivity is a method to suppress sensitivity oscillations between iterations. In this study, weighted sensitivity was extended to element. In addition, comparison was made with conventional analysis by solving a defect identification problem. As a result, identification shape was better with weighted sensitivity than conventional analysis. This is presumably due to the fact that the weighted sensitivity suppressed the sensitivity oscillation, and the structural renewal progressed slowly. This paper reports the effect of weighted sensitivity for the defect identification problem based on numerical results.