COMPARISON OF TIME-FREQUENCY DISTRIBUTIONS FOR PULSE-ECHO ULTRASONIC EVALUATION AND FILTERING

Michal Kubínyi, Radislav Šmíd

Abstract

For a direct evaluation of ultrasonic signals it is necessary to analyse data with an acceptable level of noise. Ultrasonic signals represent a specific category of time domain signals to be analysed. In order to increase difference between the level of noise and the amplitude of the ultrasonic pulse a proper method for data differentiation has to be used. Within this article are discussed and compared two time-frequency distributions. The first approach is via wavelet transform. The second proposed method is S-transform. This transformation has been recently introduced for ultrasonic echo analyses. This tool represents intermediate stage between the Fourier transform analysis and the Wavelet transform analysis. In order to filter ultrasonic signals from Electromagnetic Acoustic transducer with a high level of noise, new, different approach was developed. The result showed complexity of the ultrasonic signal and noise in case of the Electromagnetic Acoustic Transducer. We designed the new adaptive signal processing method, which performs well on studied ultrasonic signals.

Keywords

EMAT, S-transform, correlation, signal filtering

Download

IMEKO-TC4-2010-083.pdf

DOI

-

IMEKO TC

TC4 - Measurement of Electrical Quantities

Back to the proceedings