Flow Measurement Wavelet Filtering for Ultrasonic Flow Meter

Abstract:

This paper introduces the wavelet filtering to the digital signal processing of ultrasonic flow meter. The noise sources in ultrasonic flow meter are analyzed, and they are divided into three groups: power input, integrated circuit, and other acoustic ways. Some of the noise is not white noise, so traditional noise reduction filter is not effective. Original signal from ultrasonic flow meter is sampled, and four wavelet filters (Haar wavlet, Daubechies wavelet, Symlets wavlet, and discrete Meyer wavelet) with three threshold methods (SURE threshold, Universal threshold, and Minimax threshold) on the signal is compared through experiment. The result shows Haar wavelet with Universal threshold has the best filtering ability for ultrasonic flow meter.

• Periodical: Applied Mechanics and Materials (Volume 390)
• Pages: 445-449
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.390.445
• Online since: August 2013
• Authors: Yu Yang, Wen Wen, Guang Hua Zong, Feng Lin Ding, Yong Li
• Keywords: De-Noise, Flow Measurement, Ultrasonic Flow Meter, Wavelet Filtering

References

[1] Y. Yu, G.H. Zong: Review of Scientific Instruments, Vol. 83(2) (2012): pp.026107-026103.

[2] I. Daubechies, W. Sweldens: Journal of Fourier Analysis and Applications, Vol. 4(3) (1998): pp.247-269.

[3] A.G. Bruce, H. -Y. Gao: Biometrika, Vol. 83(4) (1996): 727-745.

[4] D.L. Donoho: IEEE Transactions on Information Theory, 41(3) (1995): 613 - 627.

[5] H. -Y. Gao, A.G. Bruce: Statistica Sinica, 7(469-488) (1997): 855-874.

[6] D.L. Donoho, I.M. Johnstone: Journal of the American Statistical Association, 90(432) (1995): 1200-1224.