Localization of defects in magnetic permeable bar by multiple signal classification approach using rotating dipole model

Abstract

A simplified search algorithm based on multiple signal classification (MUSIC) approach with rotating dipole model has been proposed to identify the defect locations inside a magnetically permeable bar. This new algorithm identifies the defect locations and the moments of the modelled anti-dipoles without the necessity to decompose the single column lead vector by singular value decomposition (SVD). Here, we find the location of the first dipole by using the rotating dipole model of MUSIC approach and the rest of the defect locations are identified using the orthogonal projection method coupled with rotating dipole model of MUSIC. Three deeply buried defects were modelled as anti-dipoles and the leakage fields due to these three deeply buried defects were simulated by direct forward calculation. Using the calculated data as an input, it has been possible to identify the number of defects and their locations by this new approach with significant reduction in computations, even in the presence of reasonable levels of additive noise.