Feature Extraction of Ultrasonic Flaw Signals of Pipelines Using Hilbert-Huang Transform

Ultrasonic nondestructive testing technique is widely used to detect material flaws. When flaws were classified using collected ultrasonic echo signals, the features of the ultrasonic flaw signals play an important role in ultrasonic flaw classification systems. In this paper, a new feature extraction method was proposed. Hilbert-Huang Transform was applied to ultrasonic flaw signals to obtain the marginal spectrum, the instantaneous energy density and HHT time frequency spectrum. According to the marginal spectrum, the instantaneous energy density and HHT time frequency spectrum, the features of flaws were extracted in the frequency domain, time domain and time-frequency domain. This is a new feature extraction method. From these new features the flaws can be classified easily.