摘要：In order to study the change rules of interior acoustic field with the development of cavitation, this article presented a method that through comparing the experiment results with numerical simulation results in cavitation bubbles’ distribution images under different cavitation coefficients to optimize the accuracy of acoustic field simulation using computational fluid dynamics combined with the Lighthill acoustic analogy. First, a closed visual testing system was established based on pump product test system and high-speed photography. Second, the cavitation performance under the rated operating condition was calculated using different cavitation models, in order to obtain both pump cavitation performance curve and distribution of cavitation bubbles. Next, based on the external characteristics experiment and high-speed photography experiment, the appropriate cavitation model for unsteady numerical calculation was selected. On the basis of vapor volume fraction distribution and the cavitation performance curve, four different typical points representing different cavitation coefficients were selected for further analyses. The direct boundary element method was used to calculate the variation characteristics of cavitation-induced noise at different cavitation coefficients. Finally, the peak-to-peak value of pressure fluctuation coefficient is defined as Д, and the pressure pulsating frequency domain signals were analyzed to further study the influence of pressure pulsation on acoustic field. The results show that the effect of pressure pulsation on noise is mainly focused on the discrete eigenvalues, and as for the broadband noise, the influence is not obvious. With the development of cavitation, axial passing frequency, 10- to 100-Hz frequency band, and 1000- to 3000-Hz frequency band show an increasing trend, while blade passing frequency and its harmonic frequencies show a decreasing trend. At the onset of cavitation, 1000- to 3000-Hz frequency band has the highest sensitivity for cavitation detection.
Loading...
