Research On Data Processing Methods Of Transit-time Ultrasonic Flowmeter

Flow measurement takes an important position in the fields of industry and agriculture. As an important tool of flow measurement, ultrasonic flowmeter has many outstanding advantages and is getting more and more attentions. But because of the disturbance of external and internal environment, the data measured by ultrasonic flowmeter often contains a lot of noises, which will certainly affect the accuracy and stability of flowmeter. Therefore, it is necessary for us to reduce the noise from the collected data.Considering the problems of low accuracy and bad stability of transit-time ultrasonic flowmeter designed by our lab, it is necessary to denoise the data obtained from it. So this paper adopts three data processing methods:mean rolling filter, Fourier transform method and wavelet transform method, respectively, from time domain, frequency domain and time-frequency domain to analyze and process the collected signals, and mainly emphasizes the application of Fourier transform and wavelet transform in data processing of ultrasonic flowmeter and explains their denoising effects by the Matlab simulation.This paper transforms the transit-time signals from time domain to frequency domain by Fourier transform, and by analyzing the frequency spectrum of signals finds there are many high-frequency components, so decides to adopt the low-pass filter to denoise the signals. Fourier transform method could filter out some of noises, to a certain extent, improve the accuracy and stability of ultrasonic flowmeter, but the effects are still not very satisfactory.Wavelet transform denoises the signals from time-frequency domain and signals can be decomposed in the different scales. According to the changes of wavelet coefficients in each scale, this paper uses the wavelet threshold method, and analyzes the importance of a proper mother wavelet, decomposition scale, threshold method for the denoising effects. This paper adopts db6 and db3 mother wavelet, sqtwolog soft-threshold method and 6 decomposition scales to denoise the signals, which can obtain the good data-processing effects.By comparing the denoising results of Fourier transform and wavelet transform, wavelet transform method is more feasible for the data processing of ultrasonic flowmeter, which can effectively improve the accuracy and stability of ultrasonic flowmeter.