Genetic algorithm-enabled quantitative characterization of planar defect using local resonance frequency and attenuation.

Driven by the applications of advanced manufacturing technologies which enable complex designs, the nondestructive evaluation of damage in complex structures is playing an increasingly important role across various industries. The local defect resonance (LDR) has demonstrated greater applicability to defects in complex thin-walled structures than traditional methods. However, existing LDR-based methods suffer from the low accuracy in the quantitative evaluation of defect owing to the difficulty in determining the defect boundary. A method based on the frequency and attenuation of LDR is proposed in this investigation to quantify the diameter and thickness of circular defects simultaneously using the genetic algorithm. In this method, the reflections of guided ultrasonic waves at defect boundaries are analyzed using a normal mode expansion method, and thereby the relations between the FBH parameters (i.e., diameter and thickness) and LDR attributes (i.e., the frequency and attenuation rate) are obtained. On this basis, a method based on a genetic algorithm is proposed to inversely determine the defect parameters using the LDR attributes. The proposed method is validated through numerical investigation and experimental evaluations of a series of flat bottom holes in plate structures. The proposed method enhances the accuracy and efficiency for the quantitative evaluation of defects in complex structures, advancing the application of LDR-based nondestructive evaluation techniques and providing basis for developing structural health monitoring techniques using LDR.