De-nosing of acoustic emission signal and its application in crack detection
WANG Xing-lu, SHANGGUAN Wang-yi, ZHAO Kun
(School of Vehicle Engineering, Xi‘an Aeronautical University, Xi‘an 710077, China)
Abstract: Aiming at the problem of noise interference when using acoustic emission technology to evaluate the fatigue fracture state of crack propagation, the results are affected by the factors such as threshold selection and modal component when the signal is processed by WPT or EEMD, the working principle of the two de-nosing methods was analyzed. A new de-nosing method combined WPT and EEMD was proposed, the method includes IMF-WPT de-nosing method, WPD-EEMD de-nosing method and EEMD-WPT de-nosing method. A simulated fatigue fracture signal with noise was de-noised by three kinds of de-nosing methods, the signal-to-noise ratio and the distorition of the de-nosing signal were calculated. Fatigue crack propagation tests of metal materials and welding structures were carried out, acoustic emission of fatigue crack propagation was de-noised by WPT-EEMD de-nosing method.The results indicate that de-nosing signal with signal-to-noise ratio of 10.32 and distortion of 0.31% obtained by WPT-EEMD de-nosing method, the WPT-EEMD method has the best de-nosing effect. The high frequency noise can be removed effectively and the reliability of the analysis results can be improved.
Key words: acoustic emission signal; de-nosing; WPT-EEMD; fatigue crack propagation