Optimization design of ultrasonic flowmeter flow channel based on frequency spectrum analysis

Abstract: Ultrasonic flowmeter is a non-contact flow measurement method which is widely used in various regions from agricultural irrigation to food processing. The accuracy and stability of ultrasonic flowmeter can be affected by large scale vortex because the measurement basically depends on average line velocity on the ultrasonic path rather than the whole intersection scanning. For U-shape transit-time ultrasonic flowmeter, the reflection columns extend into flow area and generate vortexes. The influence on the ultrasonic measurement from different scale vortexes is known as turbulence error. By using constriction design, commercial ultrasonic flowmeters can reduce the unfavorable impact of turbulence fluctuation and then increase its signal to noise ratio (SNR). But the pressure loss caused by the necking design is correspondingly large. In order to replace the constriction and keep applicable measurement accuracy, the discussion on the sources and correction of different measurement errors of ultrasonic flowmeter is emphasized. This paper developed a numerical simulation model for ultrasonic flowmeter based on the large eddy simulation (LES) theory and also validated it. Upon the obtained LES data, the frequency spectrum analyses are firstly practiced to study the relationship between measurement accuracy and turbulent diffusion on the basic U-shape ultrasonic flowmeter without optimization. It is found that the mean flow rate at the second half of flowmeter is relatively high and the turbulent fluctuating scale is comparatively large. Breaking the large scale vortexes at the second half is probably a good way to stabilize the turbulent fluctuation. This manuscript designed three new types of U-shape ultrasonic flowmeters with grid structure, which canceled the constriction part in the U-shape ultrasonic flowmeter. The statistical characteristics of turbulent error based on 6 different U-shape ultrasonic flowmeters are compared. The best optimized design is Case 3 which can potentially replace the U-shape ultrasonic flowmeter with constriction design due to low pressure loss. It can be found that, compared to the ultrasonic flowmeter with constriction, Case 3 reduces the pressure loss by 55% and 61% under the Reynolds number of 5 000 and 50 000 respectively. The turbulent error of Case 3 under the Reynolds number of 5 000 is 0.01%, which is as small as U-shape ultrasonic flowmeter with constriction. As for large flow rate under the Reynolds number of 50 000, the turbulent errors of Case 3 and U-shape ultrasonic flowmeter with constriction are very close, which are 0.18% and 0.17%, respectively. In order to analyze the turbulence reducing effects of Case 3, the flow characteristics of Case 3 is studied. The distribution of grid structure of Case 3 has little influence on the averaged velocity in the measured path. The velocity in Case 3 is layered without good mixing, which can be regarded as the sign of less large scale fluctuation. The introduction of grid structure can restrict the low frequency pulsation, while increasing the high frequency components. As the high frequency components are easier to be smoothed out by time average, the ultrasonic measurement is optimized. If finer grid structure is introduced, the measurement error would decline and pressure loss would correspondingly increase with high possibility. It can potentially be customized by the engineering requirements in the future.