Minimum variance distortionless response guided-wave imaging with particle swarm optimization

Abstract

Ultrasonic guided-wave detecting techniques are promising for wide application in structural health monitoring. A guided-wave imaging method based on improved MVDR (minimum variance distortionless response) is proposed for the defect detection of aluminium sheets by Lamb waves, utilizing the differences between baseline and defect signals to extract an envelope curve containing defect position information. Specifically, an imaging algorithm with a position weight vector is introduced to highlight the characteristics of pixels distorted by the presence of the structural defect. QPSO (quantum particle swarm optimization) is used for weight vector optimization. The QPSO improves upon the standard particle swarm optimization (PSO) algorithm by relaxing the requirement that either symbolic manipulation or constraint conditions of the second-order differentials are required in the Lagrangian optimization for further judgment. Diagonal loading compensation and windowing methods are used for the further improvement of imaging quality. Experimental comparison results and the optimizing procedure of the particle swarm’s weighting factor demonstrate the effectiveness of the proposed algorithm.