Band-Stop Filtering Method of Combining Functions of Butterworth and Hann Windows to Ultrasonic Guided Wave

Abstract

In view of the presence of a large amount of noise in the measured ultrasonic signals and the small contribution of existing filtering methods to the improvement of the signal-to-noise ratio (SNR), the band-stop filtering method of the combining functions of the Butterworth and Hann windows (B–H band-stop filtering method) to ultrasonic guided waves was developed by analyzing the spectral features of the ideal and measured ultrasonic signals after the discrete Fourier transform has been applied. Its application to a measured signal explained the B–H band-stop filtering method more clearly, and indicated the feasibility to maximum noise removal. The application of the method for structural-damage localization proved that the method had the advantages of greatly reducing noise; hence, the maximum value of the Wigner–Ville distribution (WVD) could be identified more easily and the localization accuracy was improved. Compared with the Butterworth band-stop filtering method and the extended Kalman filtering method, the B–H band-stop filtering method has the best characteristics in terms of improving the SNR of signals, the clarity of the WVD of the effective signals, and the damage-localization accuracy.