Astm stp 474 1970

Foreword This publication is a collection of most of the papers presented at the Symposium on Characterization and Determination of Erosion Resistance, organized by Committee G-2 on Eros

CHARACTERIZATION AND

DETERMINATION OF

EROSION RESISTANCE

A symposium presented at the Seventy-second Annual Meeting AMERICAN SOCIETY FOR TESTING AND MATERIALS Atlantic City, N J., 22-27 June 1969

ASTM SPECIAL TECHNICAL PUBLICATION 474

List price $28.75

~ l j ~ AMERICAN SOCIETY FOR TESTING AND MATERIALS

1916 Race Street, Philadelphia, Pa 19103

(~) BY A M E R I C A N SOCIETY FOR T E S T I N G AND ]V[ATERIALS 1970 Library of Congress Catalog Card Number: 72-108625

ISBN 0-8031-0063-9

NOTE

The Society is not responsible, as a body, for the statements and opinions advanced in this publication

Printed in Baltimore, Md

October 1970

Foreword

This publication is a collection of most of the papers presented at the Symposium on Characterization and Determination of Erosion Resistance, organized by Committee G-2 on Erosion by Cavitation or Impingement

in co-operation with the U S Office of Naval Research, at the Seventy- second Annual Meeting of the American Society for Testing and Materials held at Atlantic City, N J., 22-27 June 1969 A Thiruvengadam, Hydro-

nautics Inc., now at Catholic University of America, presided as chairman

of the symposium, and F J Heyman, Westinghouse Electric Corp., pre- sided as co-chairman of the symposium

Three of the papers presented at this symposium have been published

in other ASTM publications as follows: "Surface Damage from High Velocity Flow of Lithium" by L G Hays and "Dynamic Response and Adhesion Failures of Rain Erosion Resistant Coatings" by A F Corm and A Thiruvengadam, Journal of Materials, Vol 5, No 3, Sept 1970;

and "ASTM Round-Robin Test with Vibratory Cavitation and Liquid Impact Facilities of 6061-T 6511 Aluminum Alloys, 316 Stainless Steel, and Commercially Pure Nickel" by C Chao, F G Hammitt, C L Kling,

T M Mitchell, and D O Rogers, Materials Research & Standards, Vol

10, No 10, Oct 1970 They therefore do not appear in this volume

Related ASTM Publications

Erosion by Cavitation or Impingement, STP 408

(1967), $20.00

Environmentally Controlled Cavitation Test (Improvements in a Cavitating

Cavitation Damage Mechanism and I t s Correlation to Physical Properties

Effect of Temperature and Pressure on Cavitation Damage in S o d i u m - -

Experimental and Analytical Investigations on Liquid I m p a c t ErG~ion

A Statistically Verified Model fur Correlating Volume Loss Due to Cavitation

Erosion Rate-Velocity Dependence for Materials at Supersonic S p e e d s - -

Rain and Sand Erosion, Phenomena of Material Destruction Caused b y

Hydrodynamic Model of Correlation of Metal Removal Rates from Repeti-

Discussion 408 Analogy Between Erosion Damage and Pitting of Machine Component

Cavitation Damage Resistance and Adhesion of Polymeric Overlay Materials

The most significant contribution of these symposia has been the establishment of a common forum for scientists and engineers working

on cavitation and impingement erosion and the exchange of useful knowl- edge As a further step we focused the attention to a specific problem and selected the theme for this symposium as characterization and determina- tion of erosion resistance This was reflected in several papers presented

in this symposium Some have made significant advances toward defini- tion and characterization of erosion resistance in its own right and toward establishing statistical and physical correlations between this and other material properties

The ASTM round-robin tests using vibratory cavitation and liquid impact facilities and the comparative erosion tests of steam turbine blade materials in Europe are examples of the recent attempts toward standardization of existing test methods Besides, new testing techniques also have been advanced Hopefully these efforts will lead toward the generalization of erosion test results so that they can be applied to the prediction of service performance and toward quantitative and qualita- tive understanding of the erosion process A significant number of papers have concerned themselves toward the understanding of the process of erosion itself It was highly encouraging to note the international response

to this symposium and the most enlightened discussions that followed the presentations of papers These discussions are documented fully in this volume It is our belief that this symposium has contributed greatly

to the scope of the ASTM-G-2 Committee on Erosion which includes

"the promotion of knowledge in the area of erosion of materials by cavita- tion or impingement; the development, evaluation, and correlation of test methods; and the establishment of standards."

Copyright* 1970 by ASTM International www.astm.org

2 CHARACTERIZATION AND DETERMINATION OF EROSION RESISTANCE

A Thiruvengadam

Associate professor of Mechanical Engineering, The Catholic University

of America, Washington, D.C 20017; symposium chairman

F J Heymann

Senior engineer, Technology Development, Large Turbine Division, Westinghouse Electric Corp., Lester, Pa 19113; symposium co-chairman

Phillip Eisenberg 1

Cavitation and Impact Erosion-Concepts, Correlations, Controversies

REFERENCE: Eisenberg, Phillip, " C a v i t a t i o n a n d I m p a c t E r o s i o n - -

C o n c e p t s , C o r r e l a t i o n s , C o n t r o v e r s i e s , " Characterization and Determination

of Erosin Resistance, A S T M S T P ~7~, American Society for Testing and

Materials, 1970, pp 3-28

A B S T R A C T : Increasingly severe requirements for materials in modern high performance applications have re-emphasized the need for fundamental under- standing of erosion processes associated with cavitation and liquid impinge-

ment In recent, years, there has been disclosed a number of concepts which are

proving helpful in the description of both types of erosion in a wide range of hostile environments (high temperatures, corrosive liquids, etc.) Such probings are providing a framework for development of useful methods of analysis and prediction of the response of materials under cavitation attack and liquid impact In spite of the ingenuity of these ideas and investigations, however, the nature of the complex interactions among the m a n y parameters involved inevitably has resulted in controversies which have yet to be resolved

Perhaps the most important concepts which have made possible rational cor- relation attempts and show promise of a unified treatment of erosion are those

of energy absorption In both types of erosion, very useful results have been obtained based on static strain energy, but a more complete and critical evalua- tion of these ideas must await the accumulation of data on behavior of materials

at high strain rates Strain rates of interest are those associated with cavitation bubble collapse and with liquid droplet velocities typical of rain impact on high- speed aircraft and droplets in wet steam or vapor turbines

Whether cavitation damage descriptions can be treated independently of the manner in which the pressures of collapsing cavities are applied shock waves

or internal jet formation still requires investigation Internal jet impact is analogous to droplet impingement These problems are connected intimately with the hydrodynamics of cavitation bubbles, droplet deformation, and the physical properties of the liquid environment In the latter connection, recent work in hot liquid alkali metals has added to the background information needed to achieve useful correlations

The dependence of rate of erosion on exposure time and the existence of definite " z o n e s " of erosion (incubation, accumulation, attenuation, steady state,) now seem clearly established for both cavitation and impingement attack The mechanisms which account for this behavior, however, whether hydrodynamic, mechanical, metallurgical, or combinations of these, still require clarification Very impressive correlations have been achieved for data in the so-called steady-state zone using strain energy concepts, but even here there

is a question about the "steadiness" of this zone Phenomenological fatigue

President, Hydronautics, Inc., Laurel, Md 20810

Copyright* 1970 by A S T M International www.astm.org

4 CHARACTERIZATION AND DETERMINATION OF EROSION RESISTANCE

theories have been adapted to describe the history of damage progression with good results Here again further research is required to clarify whether the role assigned to fatigue is truly governing in the context of these theories or whether the correlations based on statistical fatigue theories are physically correct

KEY WORDS: cavitation, erosion, impingement, evaluation, tests

It is a great pleasure and indeed an honor to be given the privilege of aelivering the opening paper to this Symposium on Characterization and Determination of Erosion Resistance The fact that these meetings, sponsored by the ASTM Committee on Erosion by Cavitation or Impinge- ment, can attract the type of participation that it does is indicative of the importance of the field and the great activity now being devoted to cavitation and impingement erosion The international character of the meeting is a tribute to the American Society of Testing and Materials for its foresight in sponsoring these stimulating forums The Committee also deserves praise for its recognition of the intimate connection between cavitation erosion and liquid impingement and the benefits to be derived from the intersection of those working in the separate subjects for ex- change of views, opinions, and criticisms The frequency of the conference also guarantees that each family does not become so enamored of its ideas that it loses sight of shortcomings, and the exposure thus ensured

is crucial to continuing progress toward understanding of the physical phenomena as well as application in engineering contexts What has happened, of course, is the stimulation of those previously devoted to impingement erosion to give their attention also to cavitation erosion phenomena and vice versa, with the result that, properly, an attempt to achieve unification of treatment is underway

Increasingly severe requirements for materials in high performance applications have re-emphasized the need for fundamental understanding

of the erosion processes associated with cavitation and liquid impinge- ment I believe it is fair to say that recent probings and concepts are providing a framework for development of useful methods of analysis, characterization, and prediction of the response of materials under cavita- tion attack and liquid impact in a wide range of hostile environments In spite of the ingenuity of these ideas and investigations, however, the nature of the complex interactions among the very many parameters involved inevitably has resulted in controversies which will undoubtedly rage for some time to come

It is not my intention, nor would it be feasible, to try to give a complete review and critique of the status of knowledge in these fields The subject

is just too complex and merits more than the superficial treatment that would be the outcome in the time available In fact, I would not have the temerity to make such an attempt before this audience Furthermore,

EISENBERG ON CAVITATION AND IMPACT EROSION 5

the remarkable scope and excellence of the papers and discussions of the previous symposium sponsored by the ASTM Committee in 1966 [1], 2

the meeting on Deformation of Solids by Impact of Liquids sponsored

by the Royal Society of London in 1966 [2], and the Second Meersburg Conference on Rain Erosion and Allied Phenomena in 1967 [3] make a new review rather premature Consequently, it is my purpose to touch on just a few points that I believe are particularly interesting and important and which require further clarification and research I think it also would

be of interest to trace the development of the ideas which have brought our understanding of these problems to its present state of activity and which, coincidentally, are the sources of some of the severest controversy

at the present time Yet, these are just the topics which must be under- stood if rational characterization methods are to be achieved

More than three decades ago, De Haller and Ackeret [4] pointed out the similarity between cavitation damage and liquid impact erosion For many years, starting with Rayleigh's computation, it was assumed that cavitation bubbles, as long as they remain spherical, collapse with suffi- cient force to produce pressures high enough to cause damage Ackeret [5], however, pointed out that in general cavitation bubbles do not col- lapse spherically, at least in the flow conditions that he observed, and consequently the collapse pressures must be too small to cause damage Even in the case of spherical collapse, it has been shown that collapsing bubbles must be so close to the attacked surface that it is unlikely that very many will ever reach a strategic position Knapp [6] estimated that only one in 30,000 cavitation bubbles reached the surface of the body

on which he made his observations of the dependence of damage on stream velocity More recently, Ivany [7] and Hickling and Plesset [8] calculated that bubbles, in order to cause damage, must be well within a distance

of one bubble radius from the wall before the pressures developed by spherical collapse will be high enough to cause damage In any event, collapsing bubbles become unstable and do not collapse spherically Thus, they produce much smaller pressures, and various attempts have been made to prove that the collapse of cavities cannot be the cause of damage (Parenthetically, it might be said at least that the early contro- versies concerning nonmechanical origins seem to have been laid to rest.)

A very closely reasoned discussion of this problem is contained in an outstanding paper by Benjamin and Ellis [9] in the proceedings of the Royal Society meeting mentioned earlier, and I recommend that to you

In addition, it is a beautifully written paper

During an investigation of ultrasonic cavitation, Kornfeld and Suvorov

in 1944 [10], observed that air bubbles underwent forced oscillations in the acoustic pressure field which resulted in distortions leading to insta-

The italic numbers in brackets refer to the list of references appended to this paper

6 CHARACTERIZATION AND DETERMINATION OF EROSION RESISTANCE

solid boundary (after Naude and Ellis, Ref 12)

bility and disintegration In some cases, these bubbles broke apart, and

it was inferred t h a t the flow of water between the new bubbles produced jets T h e y found further t h a t cavitation bubbles b e h a v e d in the same way and again a t t r i b u t e d the formation of jets to unstable and non- spherical collapse T h e y deduced t h a t such jets could cause damage Since this is a completely random process, it seems unlikely t h a t the jets thus formed, t h a t is, due to.unstable collapse, even if close to a wall, v e r y often will be directed toward the wall Consequently, there seems to be a v e r y small probability t h a t such a mechanism can account for v e r y much

of the observed damage in a cavitating flow In the late 1940's, it occurred

to me [11], on the basis of underwater explosion bubble phenomena, t h a t

it was more likely t h a t directed jets could be formed in pressure gradients and near a wall (which is equivalent) and an instability mechanism need not be dominant or present, at least initially Evidently, this was a bold idea at the time; it remained for Naude and Ellis [12] about ten years later to show t h a t this does indeed happen Their photographs of a cavity collapsing near a wall are now classical One frame of a series of high-

EISENBERG ON CAVITATION AND IMPACT EROSION 7

speed photographs of a spark-generated bubble collapsing on a wall is shown in Fig 1 The internal jet directed toward the wall can be seen plainly More recent photographs of a cavity collapsing in a gravity gradient taken by Benjamin and Ellis [9] is shown in Fig 2 The photo-

FIG 2 Photographs taken by Benjamin and Ellis (Ref 9) during collapse (A and B) and rebound (C and D) of a cavity under gravity far from boundaries of liquid

8 CHARACTERIZATION AND DETERMINATION OF EROSION RESISTANCE

graphs at the top show successive stages during collapse The lower photo- graphs show the growth cycle Between B and C (near the minimum radius) an internal jet was formed that was so strong that it continued

to persist even during the growth cycle Such jets would be expected to form in any pressure gradients

These discoveries opened a new era of investigations of cavitation damage mechanisms While there is now no doubt that such jets can be formed, there still remains some question about the damage potential of these jets in flowing systems, as opposed to the idealized laboratory ex- periments in which the cavity can be placed at will relative to the bound- ary or in magnetostriction oscillators in which the wall is placed near the cavity Measurements of jet velocity in bubbles collapsing under ambient pressures comparable to atmospheric pressure show velocities sufficiently high to cause fatigue failure of metals

Unfortunately, this is not the end of the story The same objections can be raised against this mechanism that was raised earlier about the damage potential of the shock waves generated by collapsing bubbles Again, can the bubbles get close enough to the boundary for the jet to

do its damage? There are various hydrodynamic phenomena which can explain how cavitation bubbles can be forced toward the boundary Maximum damage seems to occur in the region of the downstream stag- nation point of the average cavity flow Obviously, the transient cavities are convected toward this region Transient cavities (bubbles) flowing along the boundary (in cases where fully developed cavity flow has not yet occurred) are attracted toward the boundary both by proximity to its image (that is, by the pressure reduction between the cavity and the wall) and by the Bjerknes force which arises from the collapse motion

of the cavity However, there are also forces present which tend to drive the cavity away from the boundary

Depending upon the manner in which the cavity is first formed, it may

be driven away early in its motion by the pressure gradient forward of the low pressure region (the so-called screening effect first suggested by Johnson [13]) Cavities convected along the boundary of a well-estab- lished cavitation region and which generally are assumed to be those causing the damage are slowed down in their motion by the adverse pressure gradients entering the high pressure regions Consequently, there must be just the right balance between the pressure gradient forces retarding the bubbles to prevent too early collapse and formation of the jet and motion of the bubble and the effect of the wall in finally forcing collapse and jet formation In any event, there remains the mystery of which mechanism, shock waves, or jet formation is responsible for actual damage It seems not unlikely that both mechanisms contribute mutually

to the rapid damage that is observed

Another possibility connected with jet formation but unrelated to the

EISENBERG ON CAVITATION AND IMPACT EROSION 9

O o s,ou !o

I ~ " " ~ ~ CAVITY NEAR BOUNDARY

y

FIG 3 Tulin's suggested mechanism of cavitation damage by ultra jets (Ref 14)

mechanisms discussed so far has been suggested by Tulin [14] in a paper

published in the Sedov 60th anniversary volume Tulin arrived at this suggestion from a theoretical investigation motivated by an interest in the generation of so-called "ultra jets" by means of weak shocks impinging

on a bubble Ultra jets are jets having a velocity greater than half the sound speed in the liquid Such ultra jets have been demonstrated by

Bowden and Brunton [15] Tulin demonstrated that weak shocks can

indeed give rise to ultra jets and that such jets have an upper bound approximately equal to twice the sound speed in the liquid His suggestion

is that a weak shock which may not be damaging in itself, because it originates too far from the boundary, may induce an ultra jet in a cavity

on or near enough to the wall for the jet thus formed to strike the bound- ary, Fig 3 This question remains to be investigated

One final word about the damage of collapsing cavities This is a rela-

tively new observation, I believe Van Manen [16] observed a severe

bending of the trailing edge of a propeller blade in the region where cavi- tation damage to the blade material also occurred, Fig 4 He postulated that this bending failure was associated with the loading induced by the

collapsing mass of transient cavities Wijngaarden [17] has shown theo-

retically that a water-bubble mixture of cavitation bubbles containing

a permanent gas is highly dispersive and can give rise to very high average pressures which could account for the failure seen in Fig 4 I point out this observation not because it contributes very much to our knowledge

of impact phenomena but because thin foils often are used in the labora- tory, and care must be taken that erosion measurements are not obscured

by the failure and loss of material associated with this mechanism From the material damage standpoint, the existence of internal jets and their damage potential brings together the diverse interests in damage

] 0 CHARACTERIZATION AND DETERMINATION OF EROSION RESISTANCE

by cavitation and by liquid impact One might say that as far as the fluid dynamics of erosion is concerned, the problem has now found quite com- mon ground

The fluid dynamics of impacting jets and droplets seems to present problems which are even more difficult to solve and phenomena which still defy adequate quantitative description The starting point is still the one-dimensional water-hammer pressure pcv as first discussed by Honeg- ger [18] The work by Ackeret and De Hailer considered similar ideas More detailed analyses and contributions have been made by Heymann

[19, 20], Jolliffe [21], and Jenkins and Booker [22]

Extensive discussions and descriptions of the various physical phe- nomena occurring during droplet impingement and jet impact have been given by Olive Engel in an impressive series of papers over the course of about fifteen years (see, for example, Ref 23), and, in recent years Brun- ton [24], Bowden and Field [25], Bowden and Brunton [15], Hoff et al

[26], Heymann [27], and Fyall [28] to name only a few, and it would be superfluous to try to give a more up-to-date story here Suffice it to say that much more detailed and sophisticated studies of the fluid dynamics

of impacting and spreading droplets and jets remains to be done Even

EISENBERG O N CAVITATION AND IMPACT EROSION | 1

then we may not have the full story In addition to the direct impact pressures produced by droplets, secondary hydrodynamic phenomena may contribute to erosion I refer to the possibility of inducing cavitation either within the droplet during its motion following initial impact or in the radial flow over surface roughnesses during spreading I believe that these possibilities were first pointed out by Olive Engel At least I became aware of them during discussions with her some 15 years ago In the first case, cavitation bubbles may be formed because of the reflection of rare- faction waves at the surface of the solid after impact and propagation of the initial compression wave In the second case, surface roughness of sufficient height to cause local pressure reduction in the radial flow or locally separated regions must be present The first mechanism seems conceptually possible with each impact The second would seem to be possible only after some initial roughening takes place Whether any of this occurs is still a matter of conjecture

Cavitation in liquid films impacted by droplets or jets is, however, not

a matter of conjecture any longer It has been demonstrated by Brunton

pressure wave propagated through the liquid layer by the impacting droplet Furthermore, although liquid layers are often cited as providing

a cushion against damage by impingement, Brunton showed that craters can be formed by the collapse of such cavitation bubbles even when held

at a distance from the surface Brunton showed further that the collapse

of the small air bubbles indicated a motion from which he deduced that internal jets would be formed, which can perhaps be described by the type of analysis suggested by Tulin

Turning now to the problem of response of materials to cavitation or impingement attack and erosion, it is necessary to distinguish between the basic physical phenomena occurring in the materials during the pro- cesses which lead to failure and the correlating parameters which hope- fully will ultimately allow a complete description method suitable for characterization and prediction, whether based on basic principles or phenomenological deductions For ductile materials at least, there seems now to be little controversy that initial failure is associated with fatigue mechanisms It is only necessary to recall the X-ray diffraction patterns obtained by Plesset [30] some years ago as a convincing demonstration Otherwise, tension or compression failures in brittle materials still endorse the ideas of mechanical action in both cavitation and liquid impact ero- sion It is easy enough to make a general remark such as this one, but, having said this, I hasten to add that a great deal has been ]eft unsaid There is still a considerable gap and disagreement concerning the proper- ties of materials which are most important in determining whether a material will be susceptible to massive erosion under the types of attack

we are discussing Much has been written about this subject, and you are

12 CHARACTERIZATION AND DETERMINATION OF EROSION RESISTANCE

In this connection, however, I would like to refer to two rather exciting events which occurred early in this decade These have been the basis for much of the work on the subject of correlations since then They also have provided a source of research stimulation, on the one hand, and controversy, on the other, that has not only kept us entertained but also has led to some enlightenment about the subject I refer to the sug- gestion and demonstrations by Thiruvengadam [37, 38] that strain energy

to fracture provides the highest degree of correlation of material response

to cavitation attack and the discovery of the so-cMled zones of cavitation damage first clearly identified by Thiruvengadam and his collegues [39,

zones occur for many types of metallic materials and flow configurations, both for cavitation damage and liquid impact configurations It is suffi- cient to illustrate this sequence by data on erosion of a specimen rotated through a falling liquid jet (liquid cylinder), Fig 5 These data are for cold drawn Type 316 stainless steel passing through a falling liquid jet

of water [40] The axis of the jet is parallel to the face of the specimen

so that impact is analogous to a flat surface striking a generator of a cylinder The characteristic curves of the zones of erosion occur just as

EISENBERG ON CAVITATION AND IMPACT EROSION 1 3

in cavitation damage The early part of the process has been called the incubation zone in which little or no loss is evident Thiruvengadam named the rising part of the curve the accumulation zone; the falling part, the attenuation zone; and the tail, the steady-state zone It has been suggested by the ASTNI Committee that the latter three zones be designated acceleration, deceleration, and steady periods, respectively Experiments on various metallic materials have shown that in the steMy- state zone the relative rates of erosion are uniquely ordered, unlike the behavior in the other zones In very careful experiments at the National Aeronautics and Space Administration and Hydronautics, it has been possible to achieve a high degree of correlation of the material response

in the steady zone on the basis of strain energy to failure [31, 32] Never- theless, others have concluded that hardness, "ultimate resilience" or other combined property parameters seem to provide more convincing correlations [33, 34, 35] And this question is still controversial

There are further troubles While Thiruvengadam found that a very well-defined steady zone existed in all of his cavitation erosion tests in the magnetostriction oscillator (and in fact demonstrated that very similar results are observed in rotating disk machines and water tunnels), others

1 4 CHARACTERIZATION AND DETERMINATION OF EROSION RESISTANCE

have found that the characteristic curve seems to show a steadier zone over a region spanning the peak rate shown in this figure Consequently, there have been two opposing views: that of Thiruvengadam supported,

among others, by Evans and Robinson [41], and that of Plesset and Devine

[~2], Pearson and Hobbs (see Heymann, Ref 27), for example Arguments have been advanced that the behavior of the curve in the steady zone is only an incidental effect associated with the sparsity of cavitation clouds generated in magnetostriction oscillators on a very rough surface Other arguments countered that this effect is far from incidental and is asso- ciated closely with the hydrodynamics of the situation based on the ob- servation that this occurs also in other types of erosion machines in which

the cavitation cloud is generated independently of the eroded surface [31]

Now, the steady zone is not always steady, it has been found Heymann first pointed out an oscillating behavior of the erosion rate data taken at Hydronautics in cavitation damage experiments Although the Hydro- nautics investigation recognized a significant scatter, no significant peri- odicity or amplitude variation could be identified Dr Engel, I under- stand, is studying this problem further However, Thiruvengadam has shown such behavior in liquid impact studies on 1100 aluminum, Fig 6, Ref ~0 This experiment was of the same type as in Fig 5 but over a lower

FIG 7 Erosion of granite by cavities convected in high-speed water jet (Ref 43)

EiSENBERG ON CAViTATiON AND iMPACT EROSION "~ 5

range of specimen speeds The oscillation was not found in similar impact tests on titanium and commercially pure nickel One can invent various explanations for this behavior, but they are all conjecture and, in any event, may not be crucial in view of more recent correlation attempts based on adaptation of statistical fatigue theories and descriptions such

as those of Thiruvengadam and Heymann discussed later

I should like to mention just one other set of experiments having to

do with combined impact and cavitation erosion which is of interest because of velocity dependence In connection with drilling through rock

by means of high-speed jets, Johnson at Hydronauties has suggested the u~e of a high-speed jet in which the erosion action is augmented by inducing cavitation and eonveeting the collapsing bubbles in such a way

as to ensure collapse at the working face Kohl [43] has carried out the experiments, illustrated in Fig 7, in which, a jet (in this ease up to 500 ft/s) was directed against the face of rock specimens which could be ro- tated to simulate a rotating drill and could be placed at an angle to the direction of the jet At the nozzle, various devices are placed to induce cavitation The bubbles are then eonveeted within the liquid jet to the rock face By properly adjusting the distance between nozzle and working surface, the cavities could be made to collapse at the face The optimum spacing between nozzle and surface was found for maximum rate of ero- sion The data shown were all taken under these conditions The first two photographs in Fig 7 show the appearance of eroded granite after

30 and 60 min with the distance between the nozzle and specimen not varied The third photograph shows the erosion after 35 rain during which the distance to the surface was adjusted for optimum drilling as erosion progressed (The last photograph shows the erosion after 7 min when thermal shock also was applied and is not of direct interest in the present context.) Figure 8 shows the characteristic curve of erosion depth rate versus testing time duration for various jet speeds

In summary, the situation is about as follows: there is little doubt that, in ductile materials, initial damage (without weight loss) takes place immediately upon application of the cavitation or impingement load Cold working and fatigue mechanisms are evident In fact, it has been possible to show fairly good correlation between threshold damage and fatigue data taken at the frequencies characteristic of magnetostrie- tion oscillators [44] Various rules have been developed for predicting the threshold of damage for droplet impingement, all of which have short- comings and which will not lead to accurate prediction methods until more detailed investigations of droplet impingement physics are carried out At the other end of the damage process,where massive damage has taken place, a great deal of attention has been devoted to an understand- ing and description of what happens Although the controversy rages unabated, nevertheless useful results have been forthcoming; for example,

16 CHARACTERIZATION A N D DETERMINATION O F EROSION RESISTANCE

NOZZLE TO SPECIMEN DISTANCE : 4.25 inches

SPECIMEN MATERIAL : 1100-F-AL

FIG 8 Rate of depth of erosion versus testing time in experiments illustrated in Fig 7

(after Kohl, Ref 43)

the ability to select replacement materials by means of Thiruvengadam's strain energy criteria, for a given machine which is known to be eroding

at a given rate His nomogram [45, 31] for such selection is now well known and, although not the last word on the subject by any means, is the only method available to the engineer for correction of serious situa- tions by means of replacement materials Of course, other well-known

EISENBERG O N CAVITATION AND IMPACT EROSION 1 7

methods of correction are available (aeration, coatings, redesign, etc, [11]), but their description is beyond the scope of this discussion

In between these extremes, there is still very little known nor does there seem to be any solid clues on which to base a rational description

of the transition between zones and the development of the central zones What has been happening, however, is an attempt to make a phenomeno- logical description of the entire process, at least for metals, based on statistical fatigue theories Also, from a very practical engineering point

of view, and reflective of the urgency of the problem in a number of fields, effort has continued to be devoted to the relation of peak rate of erosion

to velocity of the stream (as in rotating machinery) or to velocity of droplet impact (as in rain erosion of aircraft or droplet erosion in wet turbines) Starting with the experiments of Knapp who found a sixth power variation of cavitation erosion in the cavity flow about a body of revolution, followed by the measurements of Kerr and Rosenberg [46]

in an hydraulic turbine which also showed the sixth power rule, a number

of investigations have been made in both cavitating and droplet impinge- ment environments Results have been observed varying from a fifth power result to seven and even higher Various arguments can be made

1 8 CHARACTERIZATION AND DETERMINATION OF EROSION RESISTANCE

FIG lO Peak erosion intensity and jet horsepower versus jet velocity, l~-in, jet (Ref 43)

in either the case of cavitation or droplet impingement to give anywhere from a fifth power law to a t e n t h power law depending upon the mecha- nism assumed to be dominant in the hydrodynamical behavior of the system This is a question t h a t is v e r y far from being settled and is of

v e r y i m p o r t a n t consequence in engineering applications I t again points

up the necessity for an understanding of the complex hydrodynamics of these systems

In the case of droplet or jet impingement, most experimenters seem

to find a fifth power variation of peak rate of erosion with impingement velocity An example of such results is shown in Fig 9 which refers to the impingement experiments mentioned earlier [~0] Here, a fifth power variation has been fitted to the data, although one might argue for a sixth power variation just as well In experiments on g r a n i t e - - t o illustrate the behavior of a brittle m a t e r i a l - - t h e variation seems to be either a fifth power as in the case of a 89 jet, Fig 10, or a 6.5 power, as in Fig

11, of a z/~-in, jet In these experiments, of course, there is the combined action of liquid droplet impact and cavitation bubble collapse

EISENBERG ON CAVITATION AND iMPACT EROSION 19

FIG l l - - P e a k erosion intensity and jet horsepower versus jet velocity, 1/~-in jet (Ref 43)

Finally, I would like to mention a correlation based on Thiruvengadam's definition of erosion intensity for several types of flow configurations and materials In Fig 12, there is shown the intensity of cavitation dam- age as a function of velocity for the rock drill data mentioned previously, data by Kohl on a hydrofoil of l l 0 0 - F aluminum in a rotating facility,

by Rasmussen in a rotating disk apparatus, by Thiruvengadam on soft aluminum in the wake of a cavitating circular cylinder in a rotating disk apparatus, and by Shalnev on lead in the wake of a circular cylinder spanning a two-dimensional venturi section When the data are reduced

in this way it is difficult to derive anything but a sixth power rule It can

be argued that in the impingement experiment, "gross" variation of erosion rate as a function of the velocity with which the target moves through a given field of droplets will give a rate one power higher, but this would mean the seventh power in the data cited The question there- fore remains: If the cavitation damage mechanism is indeed associated with the high-speed jets induced in collapsing bubbles, why should not

20 CHARACTERIZATION AND DETERMINATION OF EROSION RESISTANCE

both cavitation damage and impingement erosion show the same variation with speed?

After so many years of what appeared to be floundering in attempts

to achieve reasonable and rational correlation and prediction methods,

it is heartening to see the rather bolder approaches now being taken Particularly of interest are those which bear a resemblance to the proce- dures that have proved useful in the application of fatigue data to engi- neering applications Several attempts are now underway to adapt sta- tistical fatigue ideas to the correlation of impact data These include the work by Heymann [27], 3 Thiruvengadam [47], and Hammitt et al, 4

8 See p 212

4 See p 288

EISENBERG ON CAVITATION AND IMPACT EROSION "~ |

among others At the ASTM Conference in 1966, Heymann proposed a framework for correlation based on treatment of small elements of a surface subjected to an impact environment The removal of such ele- ments was assumed to be described by some statistical distribution and showed that the ideas do indeed give qualitative representations of the form of the erosion rate curve including, if necessary, the oscillations sometimes observed in the so-called steady zone

Thiruvengadam, also in 1966 (Jet Propulsion Laboratory Conference

on Turbine Erosion), proposed a theory of erosion [47] which too has given encouraging indication of being very useful in describing erosion rate based on descriptions of failure probability He assumes that the intensity of collapse or collision is inversely proportional to the Nth power of the radial distance and that the intensity of erosion is related

to the intensity of impact by an efficiency factor He used a Weibull distribution to describe this factor, which he relates to the efficiency of the material in absorbing the transmitted energy In effect, each material then has a characteristic Weibull shape parameter While it can be argued that the attenuation of the pressure pulse depends upon the mechanism assumed (shock wave transmission from a collapsing cavity, an internal jet, or direct droplet impingement), his result is rather insensitive to the rate of attenuation of the pressure pulse and accounts for the quite fair agreement he achieves An example of this type of analysis is shown in Fig 13 from Ref 48 Here are shown the results of cavitation erosion experiments and the comparison using a Weibull distribution to describe the fatigue life, and normalized with respect to the maximum rate of erosion Since this is discussed by Thiruvengadam et al in more detail in their paper for this symposium, 5 I will stop here It can be argued that all of the distribution functions have sufficient degrees of freedom or arbitrary coefficients to enable fitting any type of experimental data Some are perhaps more attractive than others, being based on physical arguments such as the Gumbell distribution, but they all will give about the same results In fact, the physics, once more thoroughly understood, will probably disclose that there is not a great deal of freedom in selecting the characteristic parameters

It is quite clear that the subject matter of this lecture has been very restricted and shows my personal bias and experience primarily with the problem of cavitation damage I have tried, however, to point up the areas in which the physical phenomena inevitably forces a convergence

of ideas in the topics of interest to this symposium Clearly, the hyrdo- dynamical problems and phenomena in all aspects of liquid impingement and cavitation damage require vigorous attention The response of mater- ials just to impact alone is evidently not clearly understood beyond the

5 See p 249

EISENBERG ON CAVITATION AND IMPACT EROSION ')3

fact t h a t fatigue p h e n o m e n a and fracture mechanics at high cycling rates play a d o m i n a n t role I n this connection, I believe t h a t much depends upon further u n d e r s t a n d i n g and accumulation of d a t a on the b e h a v i o r

of materials at high strain rates which m u s t i n e v i t a b l y play an i m p o r t a n t role in all of the various aspects of erosion I h a v e not mentioned the host of problems associated with erosion in corrosive environments Here, again, m u c h remains to be done, a n d the role of corrosion fatigue and stress corrosion cracking m u s t receive a t t e n t i o n not only for scientific reasons b u t also for urgent and wide-spread applications

Finally, it is well to emphasize t h a t all of the factors I h a v e discussed

m u s t be understood clearly and t a k e n into account if rational and useful, generalized m e t h o d s for characterization of materials are to result

Acknowledgments

T h e author is i n d e b t e d to Mr F J H e y m a n n for several constructively critical c o m m e n t s which were m o s t helpful in the p r e p a r a t i o n of the manuscript

T h a n k s are due the Office of N a v a l Research for partial support under

C o n t r a c t No Nonr-3755(00)

R e f e r e n c e s

[1] Erosion by Cavitation or Impingement, A S T M S T P $08, American Society for Test- ing and Materials, 1967

[P] "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 Stream Turbines, and

to Cavitation Erosion," Philosophical Transactions, Royal Society of London, Series

A, No 1110, Vol 260, 28 July 1966, pp 73-315

[3] Proceedings of the Second Meersburg Conference on Rain Erosion and Allied Phe- nomena, Fyall, A A., and King, R B., eds., Royal Aircraft Establishment, Farn- borough, England, 1968

[$] Ackeret, J and de Haller, P., "~ber WerkstoffzerstSrung durch Stosswellen in Flussigkeiten," Schweizer Archiv fiir angewandte Wissenschaft und Technik, Vol 4,

[7] Ivany, R D., "Collapse of a Cavitation Bubble in Viscous, Compressible Liquid- Numerical and Experimental Analysis," University of Michigan, Department of Nuclear Engineering Technical Report No 15, April 1965

[8] Hickling, R and Plesset, M S., "The Collapse of a Spherical Cavity in a Compress- ible Liquid," California Institute of Technology, Report No 85-24, Pasadena, Calif., 1963

[9] Benjamin, T B and Ellis, A T., "The Collapse of Cavitation Bubbles and the Pres- sure Thereby Produced Against Solid Boundaries," Philosophical Transactions,

Royal Society of London, Series A, No 1110, Vol 260, 28 July 1966, pp 221-240

[10] Kornfeld, M and Suvorov, L., "On the Destructive Action of Cavitation," Journal Applied Physics, Vol 15, 1944, pp 495-506

24 CHARACTERIZATION AND DETERMINATION OF EROSION RESISTANCE

[11] Eisenberg, Phillip, "On the Mechanism and Prevention of Cavitation," David Tay-

fiir Schiffstechnik, No 3, 1953 and No 5, 1954

Hemispherical Cavities Collapsing in Contact with a Solid Boundary," California

1961, pp 648-656

[13] Johnson, V E., Jr., and Hsieh, T., "The Influence of the Trajectories of Entrained

Office of Naval Research, Washington, D C., 30 Sept.-3 Oct 1966 (Also Hydro- nautics, Inc., Technical Report No 707-1, 1967, Laurel, Md.)

[15] Bowden, F P and Brunton, J H., "The Deformation of Solids by Liquid Impact

pp 433-450

[16] Van Manen, J D., "Bent Trailing Edges of Propeller Blades of High Powered Single

of the I A H R Symposium, Sendal, Japan, 1962

[17] Van Wijngaarden, L., "Linear and Nonlinear Dispersion of Pressure Pulses in

Naval Research, Washington, D C., 28 Sept.-3 Oct 1966

[18] Honegger, E., "Uber Erosionsversuche," Brown Boveri Review, Vol 14, No 4,

1927, pp 74-95

[19] Heymann, F J., "On the Shock Wave Velocity and Impact Pressure in High-

Society of Mechanical Engineers, Vol 90D, 1968, pp 400-402; discussion and clos- sure, Vol 91D, 1969, pp 134-135

[20] Heymann, F J., "High-Speed Impact Between a Liquid Drop and a Solid Surface,"

Journal of Applied Physics, Vol 40, No 13, Dec 1969, pp 5113-5122

[21] Jolliffe, K H., "The Development of Erosion Damage in Metals by Repeated Liquid

[22] Jenkins, D C and Booker, J D., "The Impingement of Water Drops on a Surface

Pergamon Press, New York, 1960

American Society for Testing and Materials, 1962, pp 3-16

[25] Brunton, J H., "Deformation of Solids by Impact of Liquids at High Speed,"

Erosion and Cavitation, A S T M S T P 307, American Society for Testing and Materi- als, 1962, pp 83-98

[25] Bowden, F P and Field, J E., "The Brittle Fracture of Solids by Liquid Impact,

1964, pp 331-352

[26] Hoff, G., Langbein, G., and Rieger, H., "Material Destruction Due to Liquid Im-

for Testing and Materials, 1967, pp 42-69

American Society for Testing and Materials, 1967, pp 70-110

sophical Transactions, Royal Society of London, Series A, No 1110, Vol 260, 28 July 1966, pp 161-167

[~9] Brunton, J H., unpublished results, personal communication

[30] Plesset, M S and Ellis, A T., "On the Mechanism of Cavitation Damage," Trans- actions, American Society of Mechanical Engineers, Vol 77, No 7, Oct 1955, pp 1055-64

EISENBERG ON CAVITATION AND iMPACT EROSION "~5

[31] Eisenberg, Phillip, Preiser, H S., and Thiruvengadam, A., "On the Mechanisms

of Cavitation Damage and Methods of Protection," Transactions, Society of Naval

Architects and Marine Engineers, Vol 73, 1965, pp 241-279

[32] Young, S G and Johnston, J R., "Accelerated Cavitation Damage of Steels and Superalloys in Sodium and Mercury," Erosion by Cavitation or Impingement, A S T M

S T P 408, American Society for Testing and Materials, 1967, pp 186-219 [33] Heymann, F J., "Erosion by Cavitation, Liquid Impingement, and Solid Impinge-

ment; A Review," Report E-1460, Westinghouse Electric Corp., Lester, Pa.,

can Society of Mechanical Engineers, Vol 89D, 1967, pp 90-110

[36] Garcia, R., Hammitt, F G., and Nystrom, R E., "Correlation of Cavitation Dam- age with Other Material and Fluid Properties," Erosion by Ca~,itation or Impinge- ment, A S T M S T P 408, American Society for Testing and Materials, 1967 [37] Thiruvengadam, A., "Prediction of Cavitation Damage," Ph.D thesis, Depart-

ment of Hydraulic Engineering, Indian Institute of Science, Bangalore, India, 1961

[38] Thiruvengadam, A., 'CA Unified Theory of Cavitation Damage," Transactions, American Society of Mechanical Engineers, Vol 85, Journal of Basic Engineering,

Sept 1963, pp 365-377

[39] Thiruvengadam, A and Preiser, H S., "On Testing Materials for Cavitation Dam-

age Resistance," Hydronautics, Inc., Technical Report 233-3, Oct 1963; see also

Journal of Ship Research, Vol 8, No 3, Dec 1964, pp 39-56

[$0] Thiruvengadam, A and Rudy, S L., "Experimental and Analytical Investigations

of Multiple Liquid Impact Erosion," Hydronautics, Inc., Technical Report 719-1, June 1968; see also Technical Report 719-2, Aug 1969

[$1] Evans, A and Robinson, J., "Erosion Experiments Related to Steam Turbine Blad-

inK," Parsons Journal, Vol 10, No 61, 1965

[$2] Plesset, M S and Devine, R E., "Effect of Exposure Time on Cavitation Damage,"

Journal of Basic Engineering, Transactions, American Society of Mechanical Engi-

neers, Vol 88D, No 4, 1966, pp 691-705

[$3] Kohl, R E., "Rock Tunneling with High Speed Water Jets Utilizing Cavitation Damage," Hydronautics, Inc., Technical Report 713-1, June 1968

[$$] Thiruvengadam, A., "High Frequency Fatigue of Metals and Their Cavitation Damage Resistance," Hydronautics Inc., Technical Report 233-6, Dec 1964 [$5] Thiruvengadam, A., "Intensity of Cavitation Damage Encountered in Field In- stallations," Hydronautics, Inc., Technical Report 233-7, Feb 1965

[$6] Kerr, S L and Rosenberg, Kjill, "An Index of Cavitation Erosion by Means of Radioisotopes," Transadions, American Society of Mechanical Engineers, Vol 80,

STP474-EB/Oct 1970

DISCUSSION

J H Brunton 1 (written discussion) It is important to stress the point

made by the author, that the commonly looked for correlation between the erosion behavior of materials and their mechanical properties, as determined in quasistatic tests, is likely to have only a limited success The reason for this is that the time scale of the loading in erosion is very different from that at present attainable in even the more sophisticated methods of dynamic testing

Suppose we consider a small water droplet, 100 tLm in diameter, impact- ing against a hard surface at say 500 m/s Repeated impacts of this kind are known to cause the eventual breakup of high-strength solids The duration, T, of the central water-hammer pressure is determined by the time it takes for release waves to move into the central zone It can be shown that T is given by,

T = ~ 1 - - 1 - -

where :

R = radius,

V = impact velocity, and

C = compression wave velocity for the appropriate pressure in the liquid

For the values given above, T is found to be of the order of 10 -9 An estimate of the peak peripheral pressure, associated with jetting of the liquid at the instant the outward flow of liquid begins, can be made, and this too, for the chosen conditions, is again of the order 10 -9 These extremely short loading times result from the small dimensions of the drop and the high impact velocities and stress wave velocities

What differences in material behavior can be expected under these loading conditions? Among engineering materials metals are the least strain-rate sensitive, but even here, with steels, for example, increases in yield strength by factors of two to three are found for microsecond loading times Increases of this kind are not related simply to the statically deter- mined mechanical properties For the much shorter loading times found

in drop impingement and in bubble collapse more extreme changes can

1 University Engineering Department, Cambridge, England

26

Copyright s 1970 by ASTM International www.astm.org

DISCUSSION ON CAVITATION AND IMPACT EROSION 2 7

be expected For example, 10 -9 to 10 -s is about the time needed for dis- location loops to be generated by the Frank-Read mechanism The failure

of this process to bring about a rapid increase in plastic strain could prevent relaxation of the impact stress in the metal Further differences could be considered the prevalence of fine slip lines at high rates of strain and of coarse slip bands at normal strain rates Again, there could

be a size effect in the solid In erosion very small areas of the surface are loaded individually, a very different situation from that found in most mechanical tests where large volumes are under stress In this respect it commonly is found to be the case that failure processes which depend on flaw distributions, and in metals this includes dislocation networks which require higher stresses as the dimensions of the loaded area are reduced

If differences of the type outlined above are important, then it would

be as well to treat erosion as a special type of failure in the sense that brittle fracture, creep, and fatigue are special types of failure Instead of attempting to classify materials in terms of known mechanical properties

it is probably more useful to work directly in terms of their erosion proper- ties as determined in a standardized test Perhaps a vibratory test or a wheel and jet test would be suitable for this purpose Over the long term

it seems probable that erosion studies will follow a similar pattern of development to that taken in the related subject of fatigue Erosion tests will provide quick answers to practical problems, while an elucidation

of the basic mechanisms in terms of material constants is likely to be a lengthy process

F G Hammitt 2 (written discussion) This introductory paper is interest-

ing in pointing out both those questions upon which agreement exists as well as those where it does not In the former category, much has been said recently regarding the importance of microjet impact in cavitation damage, and I believe it generally is agreed that this is indeed an impor- tant, though not exclusive, mechanism An interesting point which l believe needs consideration is the fact that rapid and extensive damage

is caused upon materials such as stainless steels and even stellites by rotating wheel impact facilities wherein the impact velocity is only about

300 ft/s which corresponds to a "water-hammer" pressure of about 20,000 psi Since this is well below even the fatigue strength of such ma- terials, it is apparent that either actual pressures are well in excess of the water-hammer pressure, or other mechanisms importantly are involved, such as perhaps secondary cavitation in this case

In the nature of controversies, the assumption of a proportional relation between cavitation damage and strain energy, reported by Hydronautics,

is in my opinion a serious error (as further discussed in my paper at

Professor, University of Michigan, North Campus, Ann Arbor, Mich 48104

28 CHARACTERIZATION AND DETERMINATION OF EROSION RESISTANCE

this symposium) For some types of materials such as tool steels, the relation between cavitation resistance and strain energy is in fact inverse I believe t h a t the b o d y of data available from all sources t o d a y shows a much b e t t e r proportionality between ultimate resilience and damage resistance, as is reported in both m y own paper t o d a y and t h a t

b y Frank Heymann However, the correlation with ultimate resilience also leaves much to be desired

F u r t h e r in the nature of controversy I do not believe in the existence

of a final "steady-state zone" of damage rate as reported by the H y d r o - nautics investigators, and in fact I have never seen d a t a which in my opinion supports this hypothesis R a t h e r the damage appears to contin- ually drop off, depending on the length of test In some cases (as shown, for example, in the paper b y Young and Johnston3), it again increases after a minimum at a rapid rate in some cases I believe t h a t most of these effects are due to changes in "flow g e o m e t r y " W h e n the damage becomes large, the flow field is affected and the damage rate becomes un- predictable Its behavior depends v e r y much on the particular material- fluid combination Thus I feel t h a t the m a x i m u m rate attained in a test

is the only practical one to be used, both for the reasons cited above and for the usual economic necessity of conserving test time to practical limits Finally, I would like to comment on the relation between damage and flow velocity in cavitation tests While in some cases damage is indeed very sensitive to velocity as reported by Dr Eisenberg, there are other cases where it is not Such a case has been found in our own tests with a simple conical diffuser and was previously reported 4 1 believe this relation- ship depends upon the sensitivity of pressure to velocity in the collapse zone, which is much greater for some flow geometries and cavitation conditions

t h a n for others

For related surveys of present thinking on cavitation damage, I would like to mention the v e r y comprehensive recent American Society of Mechanical Engineers book on this and other cavitation subjects 5

T e s t ) , Characterization and Determination of Erosion Resistance, A S T M S T P

47~, American Society for Testing and Materials, 1970, pp 29-47

ABSTRACT: Early attempts to produce cavitation erosion on a stationary specimen close to a vibrating probe showed promise as a means of testing weak materials and coatings However, the geometry affected the damage rate, the condition of the cavitating liquid was not accurately known, and the temperature in the cavitation zone rose appreciably An improved version of this test has been devised at the University of Strathclyde, Glasgow, and subse- quently further developed at the National Engineering Laboratory, East Kilbride

The specimen is mounted a small distance from the vibrating probe, and liquid is fed continuously through a central hole in the stationary specimen With water the temperature rise could be kept easily below 2 C In oil, although the temperature rise was considerably higher (10 C) it was under complete control and depended mainly on the flow rate and specimen separation The erosion pattern obtained on the stationary specimen was well distributed and fairly uniform The rate of erosion reach an early maximmn value and then decreased with time It decreased as the separation was increased Optimum test conditions were obtained with 0.5 mm (0.020 in.) separation from a 20 kHz vibrator having a peak-to-peak amplitude of 50 ~m (0.002 in.) The specimen and vibrating probe were both 16 mm (5/8 in.) diameter Effects of probe amplitude, test liquid flow rate, temperature, and static pressure were examined

This technique has been used with considerable success for the evaluation

of white-metal layers in lubricating oils, fiber-reinforced epoxy resins in water, and a range of typical engineering alloys

K E Y W O R D S : cavitation, erosion, liquids, metals, physical properties, testing, vibration, wear, evaluation, tests

30 CHARACTERIZATION AND DETERMINATION OF EROSION RESISTANCE

known to warrant the exclusion of a historical account of this develop- ment T h e real step forward in the technology of such tests came with the introduction of the solid accelerator horn for mounting the specimen, while the frequency of 20 k H z was chosen, in G r e a t Britain, primarily because of the availability of an "ultrasonic drill" unit, which could be adapted easily for the purpose This development once and for all freed the experimenter from the drawbacks inherent in the use of a nickel tube transducer and opened up new possibilities to improve the testing technique b y concentrating on the testing environment The fundamental improvements were a good control of vibration amplitude and a positive location of the vibrating probe in space In the meantime, the need for testing environment and materials has become complicated b y the ex- tension of the range of fluids to be tested to hydrocarbons and of materials

to reinforced plastics, rubber, bearing metals, and various types of surface coatings A cursory examination of the behavior of various fluids, in particular of viscous hydrocarbons, shows up the futility of an open beaker

~est as the circulation of the fluid is of primary importance in the forma- tion of pressure fields necessary for bubble collapse On the other hand,

it also is known t h a t mechanically weak materials, such as copper crys- tals, can show slip lines after 20 s of vibration at 20 kHg in open air, and it is thus doubtful whether such vibration stresses do not obscure cavitation damage effects Consideration of the above mentioned factors has led to the joint development over a period of some years at the Uni- versity of Strathclyde and the National Engineering L a b o r a t o r y ( N E L )

of a technique, which should enable us to expose materials to cavitation attack in a controlled fluid environment without subjecting t h e m to additional mechanical stressing

Because similar attempts were pursued over the same period b y Schra- der [1], 3 Endo et al [2], Rieger [3], and much earlier work had been done

b y J P Paul at the University of Strathclyde, their relevance justifies a brief mention T h e essence of all these tests lies in inducing cavitation bubble formation and collapse in a film of fluid contained between two parallel surfaces In the case of Paul's experiments the film was introduced through a centrally located hole in one specimen between two vibrating

b u t t faces of an Avery-Schenk fatigue tester In the case of Schrader and others, the vibrating end of the transducer was separated b y a static film of fluid from a test specimen positioned in a beaker and fully sub- merged

The results of tests b y Paul were disappointing in producing little damage but were explainable from much later experience in other fields The specimens were not submerged in the fluid, and this resulted in air being drawn into the film during the receding stroke, thus providing

a The italic numbers in brackets refer to the list of references appended to this paper

HOBBS AND RACHMAN ON CONTROLLED CAVITATION TEST 3 1

cushioning on the compression stroke and reducing bubble collapse shock

I n addition the 50 Hz frequency was far too low to yield quick results

T h e test on engine coolants by Schrader and the use of various fluids

b y other experimenters were quite successful in producing damage, but also have shown up the presence of a considerable t e m p e r a t u r e gradient

in the fluid film between the center and edge of the specimen In addition the gas content of the fluid in the film itself was v e r y uncertain Thus,

it m a y be concluded that, apart from relieving the material of the sta- tionary specimen from inertial stressing, none of the variants mentioned above has been successful in controlling adequately the cavitation en- vironment

T h e , m e t h o d of testing described below was a combination of all factors mentioned above, and its adoption was directed in particular towards the testing of lubricating oils, bearing metals, and plastic coatings The physical features of this arrangement ensure a continuous renewal of the cavitating fluid film, no mechanical stressing of the specimen, and the possible use of two dissimilar materials which m a y simulate electrolytic effects

The studies and test results reported in this paper are a summary of investigations of the effects of flow, gap width, and pressure distribution

on the characteristics of this type of test as well as on the damage pat- terns obtained in lubricating oils on white-metal specimens The impor- tance of the latter is t h a t ordinary tests in oils usually do not produce a coherent damage pattern, and thus the results are an i m p o r t a n t contribu- tion to the present state of the art

Equipment

The basic equipment comprised a commercial water-cooled magneto- strictive transducer driven at a nominal frequency of 20 kHz A d u m m y specimen was fitted to the stepped velocity transformer, and this could

be vibrated at an amplitude of 50 #m (approximately 0.002 in.) The

s t a t i o n a r y specimen was mounted on an anvil which was located immedi- ately below the transducer The test fluid was fed through a hole in the center of the stationary specimen, and the test gap between the speci- mens was drowned by the outflow filling a Perspex ring surrounding the anvil and overflowing into a collecting trough whence it returned to a sump for repeated recirculation, Fig 1 T h e means of recirculation are immaterial as long as the pump will not cavitate and produce a change

in the condition of the fluid apart from raising its pressure In the tests described, a Mono P u m p was used for circulating water and a gear pump for oil T e m p e r a t u r e distribution was measured by embedding thermo- couples in three positions 120 deg a p a r t at 2.3, 4.3, and 6.35 mm (0.090, 0.170, and 0.250 in.) radius The thermocouple leads were taken out of

32 CHARACTERIZATION AND DETERMINATION OF EROSION RESISTANCE

FIG 1 Diagram of anvil apparatus

the side of the specimen leaving the hot junction exposed to the film but flush with the specimen surface

Temperature Measurements

Distribution of Temperature over Specimen

In the work described by Endo et al [2] the temperature at equilibrium was a maximum at the center of the specimen and decreased with increas- ing radius It also was affected by the nature of the fluid and the thick- ness of the film

With the flowing film, measurements indicated that the temperature generally increased as the fluid flowed radially outwards This distribu- tion was affected by the nature of the fluid, the film thickness, and also the flow rate With only three thermocouples it was not possible to obtain detailed temperature profiles, but the readings obtained showed that the fluid temperature rose very quickly on entering the film and thereafter changed relatively little, Fig 2 Hence for the remainder of the analysis the film mean temperature rise will be taken as the difference between the inflowing temperature and the mean of the temperatures indicated

by the three thermocouples

HOBBS AND RACHMAN ON CONTROLLED CAVITATION TEST 3 3