Improved Time-Frequency Distribution Series for Ultrasonic Nondestructive Evaluation

Time-frequency distribution series (TFDS) is introduced to represent ultrasonic signals, detect and locate defects, because TFDS can effectively decrease the cross-terms with a minimum degradation of the desired properties. Moreover, considering the time-frequency concentration of the time-frequency energy atoms, the improved TFDS reduces the computational complexity, which makes TFDS more practical. Both simulated and experimental ultrasonic signals are represented by TFDS and Wigner-Ville distribution (WVD) and their performance are compared in terms of cross-terms, resolution and the ability to character defects. Results show that TFDS has the advantages of characterizing and localizing defects in ultrasonic nondestructive evaluation