A discussion on deformation of solids by the impact of liquids, and its relation to rain damage in aircraft and missiles, to blade erosion in steam turbines, and to cavitation erosion - The collapse of cavitation bubbles and the pressures thereby produced against solid boundaries

Our object is to present a broad review of this subject as a branch of hydrodynamics, referring both to the well known ‘implosion’ mechanism first analysed by Lord Rayleigh and, more particularly, to the recently perceived possibility that effects of equally great violence, such as to damage solid boundaries, may arise through the impact of liquid jets formed by collapsing cavities. In §2 a few practical facts about cavitation damage are recalled by way of background, and then in §3 the significance of available theoretical and experimental information about cavity collapse is discussed. The main exposition of new ideas is in §4, which is a review of the factors contributing to shape changes and eventual jet formation by collapsing cavities. Finally, in §5, some new experimental observations on the unsymmetrical collapse of vapour-filled cavities are presented.

This paper describes the early stages of cavitation damage observed in cavitating venturi tunnels. The cavitating fluids were water and mercury, and a wide range of specimen materials were used. The damage was found to consist of single-event symmetical craters and irregular fatigue-type failures. The degree of damage was highly sensitive to minor flow perturbations, and this is discussed. The effect of stress level in the specimen before testing, and relations between cavitation resistance and the mechanical properties of the materials are considered.


The determination of the stresses produced by cavity collapse has been of interest since Rayleigh’s discussion of the problem. One theoretical calculation relating to this problem is the magnitude of the pressure pulse which is radiated when a spherical bubble collapses and rebounds in a liquid. A calculation of this kind has been made although it was necessary to idealize the physical situation. The peak pressures predicted by this treatment were of the order of some thousands of atmospheres and could, therefore, furnish a mechanism for the damage of solid surfaces. Since these peak pressures decrease rapidly with distance from the centre of the bubble, the solid boundary must be in the immediate neighbourhood of the bubble in order that damage may be produced by this mechanism. In this situation spherical collapse or rebound cannot be expected to take place. An additional disturbance from spherical symmetry arises because the spherical shape is unstable. There is now both theoretical and experimental evidence that jet formation may develop from this unstability, and could under suitable conditions give rise to cavitation damage. This evidence is briefly discussed.


If a small cavity or bubble in a liquid is subject to impact or to shock, tiny Munroe jets may be formed on its concave surface. The velocity of these microjets may be high. A short film illustrating the formation of these small jets in cavities and in coalescing drops was shown.


The behaviour of established and potential turbine blade and erosion shield materials subject to impact erosion by water droplets of controlled size has been investigated over a range of impact velocities up to 1040 ft./s. Both the topographical form and the microstructural characteristics of damage have been studied, and correlated with the conditions of the test and the mechanical properties and phase constitution of the materials. It has been found that the rate of erosion, as measured by mass loss, changes during the course of a test. An initial incubation period is generally followed, successively, by periods of increasing, constant, and then decreasing rates of erosion, possibly culminating in a second steady, but lower, rate of erosion.


An investigation of the erosion of solids by repeated liquid impact at relatively low velocities has been carried out. The work has shown that even at low velocities compressible behaviour of the liquid is important in determining the impact pressure. An attempt has also been made to determine the distribution of the impact load. The mechanism of erosion in brittle polymers and in ductile metals has been studied. The effect of altering the conditions of impact on the erosion behaviour is described.


Efficient expansion of steam in turbines cools the vapour to the point where it becomes wet. As turbines become larger the higher blading speeds employed lead to erosion damage of the blading as a result of impact with accumulated water in the form of drops. The distribution of this damage in the turbine is discussed. The processes of drop formation, release and subsequent motion before impact with the moving blades are described and the application of this knowledge to practical design is illustrated by particular examples.


Damage produced by cavitation under field conditions can be a serious problem. The main causes of this damage and its characteristics are discussed briefly and possible remedial measures are examined. Accelerated laboratory tests are found to play an important part in cavitation erosion research, but interpretation of results needs care. Most past investigators have tended to treat cavitation damage and droplet erosion as unrelated phenomena and only qualitative correlations between the respective simulated tests have been possible. This paper presents an attempt to correlate quantitatively the results of three different erosion tests. A broad correlation between results of the drop impact erosion and constricted tube cavitation tests shows general agreement. A more detailed, but restricted, correlation has been obtained between results of drop impact and vibratory cavitation erosion tests. In both correlations, however, there is evidence of some discrepancies between corrodible and incorrodible materials. A number of factors which govern the rate of damage in the various laboratory tests are of interest. In particular, in the drop impact test the velocity of collision and the jet diameter are shown to have significant effects. There is a marked similarity between the behaviour of materials in this test and in fatigue tests and also evidence of a threshold velocity below which measurable damage ceases. The other laboratory tests were found to have their own particular controlling parameters, but the general phenomenon of cavitation erosion is more complex and is not discussed in detail. By conducting comparative tests under reproducible conditions it has been possible to classify a variety of new and traditional materials in order of relative erosion resistance and thus provide some guide to their selection for service. While the results add to the evidence that hardness is the major attribute controlling erosion resistance other properties such as ductility, elasticity and fatigue strength are seen to be significant.


A great variety of methods for the evaluation of cavitation erosion are now available. However, because of a lack of correlation with the hydromechanics of cavitation, they fail to provide a solution to the problem of scale effects. Studies are described of the hydromechanical aspects of cavitation erosion induced in the wake of a circular profile model for plane flow conditions in a water tunnel. An energy parameter for the estimation of resistance of materials to cavitation damage will be introduced. An explanation of cavitation effects will be reached by varying the parameters likely to affect the intensity and development of erosion. These parameters include: the state and structure of the cavitation zone, relative dimensions of the model, cavitation layer thickness, the characteristics model dimension, flow velocity, specimen erosion volume, experimental duration, and Reynolds and Weber numbers. The energy parameter is derived in terms of the erosion volume and the work done by the cavitation drag forces. The reciprocal value of the energy parameter gives the erosion resistance of the material in terms of the amount of work done in damaging unit volume of the material. This parameter is used to explain the power law obeyed by dimensional and velocity scale numbers. It will be shown that erosion volume is proportional to flow velocity to the fifth power, and to the characteristic model dimension to the third power. The conditions for prognosis of erosion volume from the experiments with models will be specified.


Sign in / Sign up

Export Citation Format

Share Document