Cavitation Inception Scale Effects: I. Nuclei Distributions in Natural Waters II. Cavitation Inception in a Turbulent Shear Flow

Abstract

Cavitation scale effects can be grouped into two major categories: susceptibility of the water to cavitation, i.e., the amount, size, and type of microbubbles or microparticulates in the water acting as inception nuclei, and flow field effects due to such factors as velocity and pressure distributions, body size and shape, viscous effects, and turbulent phenomena. Experimental investigations into these two aspects of scale effects were performed in the present study. Field investigations of marine nuclei populations were made using underwater holography to observe microbubbles and particulates, including microplankton in oceanic waters of Los Angeles Harbor, San Pedro Channel and near Santa Catalina Island. Holographic detection was shown to be a reliable method of measuring the nuclei number concentration density distributions. Overall, very high concentrations of the various types of potential cavitation nuclei were observed at all of the test sites and depths examined, although the statistical significance of these results is strong only in the smaller size ranges (less than 50 μm), where a significant number of counts were made. Relatively high bubble concentrations during calm sea conditions, and their population inversion below the thermocline where organism activity was high, indicate a possible biological source of bubble production rather than the usual surface mechanisms of breaking waves and whitecaps. The measured population of particulates is somewhat higher than comparable data in the ocean or in cavitation test facilities, and the number density distribution of particulates decreases approximately as the fourth power of the particle size, as often reported in the literature. An increase in particle concentration near the bottom of the thermocline in clear coastal waters is observed. The total concentration of particles and bubbles in a liquid provides an upper bound on the number of potentially active cavitation nuclei. The measured bubble sizes can be used to indicate that the average tensile strength of the ocean waters examined in this study should be on the order of via few thousand Pascals, with a minimum expected value of about one hundred Pascals. The present results support the recommendation of Billet (1985), that a concentration of at least 3 bubbles per cm^3 in the 5 to 20 μm radius range is needed in test facility water in order to model marine conditions. Experimental studies were also made on the inception processes in a large turbulent free shear layer generated by a sharp edged plate in a water tunnel at Reynolds numbers up to 2 X 10^6. Two distinct types of vortex motion were evident in the shear layer, the primary span wise and the secondary longitudinal vortices. Cavitation inception occurs consistently in the secondary shear layer vortices and more fully developed cavitation is visible in both structures, with the streamwise cavities primarily confined to the braid regions between adjacent spanwise vortices. A Rankine vortex model indicates that the secondary vortex strength is always less than 10% of that of the primary structure. Measurements of fluctuating pressures in the turbulent shear layer are made by holographically monitoring the size of air bubbles injected into the non-cavitating flow, showing that pressure fluctuations were much stronger than previously reported, with positive and negative pressure peaks as high as 3 times the freestream dynamic pressure, sufficient to explain the occurrence of cavitation inception at high values of the inception index. Cavitation inception indices display a strong dependence on the dissolved air content and thus on the availability of freestream bubble cavitation nuclei. The present inception data do not display a clear dependence on freestream velocity (or Reynolds number) but do fall into the overall range of data of previous bluff body investigations. The occurrence of inception in the secondary vortices of the shear layer, and previous reports of velocity dependence of these cores (Bernal 1981) may provide the key to explaining the commonly observed Reynolds number scaling of the inception index in shear flows.

Files

Additional details