Studies on characteristics of cavity and cavitation damage behind circular cylinders in water with a venturi

Abstract

Cavitation is a complex hydrodynamic phenomenon involving the appearance, growth and collapse of vapour- and/or gas-filled bubbles. During the collapse phase, it generates forces sufficient to erode even the strongest materials. It manifests itself in engineering systems connected with many specialised fields of knowledge. Despite many investigations which have been carried out using various approaches which reflect the differing theories and experimental procedures, the fundamental mechanics of cavitation and cavitation damage are still far from being entirely understood. The basic objective of the present investigations is to achieve a better understanding of the phenomenon. For this purpose, the following aspects of cavitating flows past sources mounted in the test section of a two-dimensional venturi are studied under different hydrodynamic conditions: A. The effect of the bounding walls on the incipient and choked cavitation numbers. B. The shape, the characteristic dimensions and the angle of detachment of the 'steady' cavity that forms behind the cavitation source. C. The mechanics of 'unsteady' cavity including the periodic shedding of vortices behind the cavitation source and the correlation between Strouhal number and cavitation number. D. Cavitation damage behind sources of different diameters and the correlation of damage with size of source and cavitation number. Experimental data was obtained in a venturi with water as the test liquid. The free stream velocity was varied in the range 13.03 – 27.5 m/sec and the pressure in the range 1.03 – 6.65 kg/cm² gauge. Six cavitation sources of diameters ranging from 0.95 to 2.54 cm were used for most of the experiments. Larger cavitation sources of diameters up to 5.08 cm were used for determining only the effect of blockage ratio on cavitation numbers. The ambient temperature of the test water was 29.2°C. The incipient and desinent cavitation numbers remained approximately constant up to a blockage ratio d/B of about 0.30, after which they showed an increasing trend. The choked cavitation number increased at a low rate up to a blockage ratio of 0.30, after which it increased rapidly. The boundaries of the test section did not seem to affect the experimental data in the present investigations for which the blockage ratios were less than 0.30. The increase in the normalized length of cavity l/d with reduction in cavitation number was slow up to l/d less than or equal to k and thereafter the increase was rapid even for small. Decreases in cavitation number. The change in the growth rate of the length of cavity occurred at different values of cavitation number for sources of different diameters. This value of cavitation number increased with increase in diameter. The choked cavitation number is a meaningful parameter controlling the length of cavity, because the normalized length of cavity varied with the effective cavitation number K–K? within a small band for all sources. The length of cavity was dependent only on the cavitation number and not individually on the free-stream velocity and the ambient pressure. The use of the maximum velocity in the test section in the definition of the cavitation number resulted in unified correlation of the normalized length of cavity with modified cavitation number K?. The variations of the area, the partial length and the maximum width of cavity with cavitation number showed trends similar to those of the length of cavity with cavitation number. The angle of detachment varied from 62° at choked conditions to a limiting value of about 84° as cavitation number increased, while Roshko’s theoretical value for zero cavitation number is 55°. Reynolds number plays a meaningful role in non-cavitating flows past bluff bodies, but in cavitating flows its significance is found to be secondary to that of cavitation number. Preliminary computations for the drag coefficient indicated a linear variation between drag coefficient and the ratio of cavitation numbers K/K?. This variation can be expressed by the equation: C?(K) = C?(K = 0)(1 + K) The nature of vortex shedding phenomenon behind circular cylinders was found to be qualitatively similar to that behind wedges reported by Young and Holl. It was observed that a minimum Strouhal number S??? occurred at cavitation numbers which increased with increase in the diameter of source. The magnitude of the minimum Strouhal number increased with increase in the diameter of source. As cavitation number approached the value for cavity-formation conditions, the Strouhal number S reached a limiting value of 0.30. Since the Strouhal number in non-cavitating flows is only a function of Reynolds number, a sudden change was indicated as the flow conditions change from cavitating to non-cavitating. While the variation of frequency of shedding of cavities with cavitation number was a function of free-stream velocity, the Strouhal number–cavitation number relationship for a particular source was found to be independent of velocity. The modified Strouhal number S attained a limiting value at different values of cavitation number for different sources. Cavitation damage indicated a peak at a critical value of cavitation number. The magnitude of the peak damage increased with increase in the diameter of source and the cavitation number corresponding to peak damage also increased with the diameter. This is partly due to the different ranges of cavitation numbers possible in the venturi for different sources. The cavitation numbers for peak damage agreed with those at which the length of cavity started increasing rapidly. The weight losses for the various cavitation sources did not indicate a unified variation with the effective cavitation number. However, the possibility of the peaks in damage occurring around a value of K–K? = 0.35 was indicated. Thus the choked cavitation number seems to be significant in damage studies also. Attempts to unify damage data of the different series of experiments by using weight losses computed at the beginning of the maximum weight-loss rate zone in correlating the mean depth of penetration with the cavitation number yielded encouraging results. With the experimentally determined weight losses, the maximum mean depth of penetration varied in the range 0.13–0.81 mm. The range for the maximum modified mean depth of penetration using computed weight losses was narrower, namely 0.33–0.68 mm. Thus, modification of the actual weight losses on some logical basis may be a step in the right direction for obtaining a unified weight loss–cavitation number relationship. The maximum damage conditions did not correspond with the minimum Strouhal number and no correlation was apparent between cavitation damage and Strouhal number. Thus, the vortex shedding phenomenon seems to have only a secondary influence on cavitation damage patterns.