Phase correction improved multiple signal classification for impact source localization under varying temperature conditions

Abstract

Lamb waves are commonly used to locate low velocity impacts in composite structures. However, it is readily influenced by varying temperature conditions that the phase drift of uniform linear array signals will quite difference between steering vector and actual signal subspace. Therefore, the traditional multiple signal classification (MUSIC) method might fail to localize the impact sources. Consider the influence of varying temperatures on Lamb waves, a phase correction improved multiple signal classification (PC-MUSIC) method is proposed. The phase errors are calculated by comparing the varying temperature conditions with the normal work condition to obtain a correction matrix, which is employed to correct the steering vector. The impact source location represented by the angle and distance in polar coordinates is finally estimated by scanning the whole monitoring area using the corrected steering vector. The experiment investigations show that the present method has higher location estimation accuracy in composite structures than traditional MUSIC method for the varying temperature conditions.