Wear 278–279 (2012) 53–61

Contents lists available at SciVerse ScienceDirect

Wear
journal homepage: www.elsevier.com/locate/wear

Scaling of cavitation erosion progression with cavitation intensity and cavitation

source
Jin-Keun Choi ∗ , Arvind Jayaprakash, Georges L. Chahine
Dynaflow, Inc., 10621-J Iron Bridge Road, Jessup, MD 20794, USA

a r t i c l e i n f o a b s t r a c t

Article history: A simple mathematical expression is presented to describe cavitation mean depth of erosion versus time
Received 3 May 2011 for cavitating jets and ultrasonic cavitation. Following normalization with a characteristic time, t*, which
Received in revised form occurs at 75% of the time of maximum rate of erosion, and a corresponding material characteristic mean
27 December 2011 2
erosion depth, h*, the normalized erosion depth is related to the normalized time by h̄ = 1 − e−t̄ + e−1 t̄ 1.2 .
Accepted 3 January 2012
This was obtained by conducting systematic erosion progression tests on several materials and varying
Available online 12 January 2012
erosion field intensities. Both a modified ASTM-G32 method and Dynaflow's cavitating jets techniques
were used and the jet pressures were varied between 1000 and 7000 psi. The characteristic parameters
Keywords:
Cavitation erosion
were obtained for the different configurations and the correlation was found to be very good, exceeding
Erosion testing an R2 of 0.988 for all cases. Relationships between these parameters and the jet pressure were obtained
Steel and resemble familiar trends presented in the literature for mass loss. The study allowed a comparative
Non-ferrous metals evaluation and ranking of the various materials with the two accelerated erosion testing methods used.
Erosion modeling While several materials ranked the same way with the different erosion intensities and testing method,
the relative ranking of erosion resistance of some materials was seen to be dependent on the cavitation
intensity.
© 2012 Elsevier B.V. All rights reserved.

1. Introduction With the continual desire to increase ship speeds and carry-
ing capacity motivated by increased economic benefits of higher
Prediction of cavitation erosion performance of a new mate- speed transportation or larger payload, the hydrodynamic loading
rial is a very difficult endeavor as it involves good knowledge of on propellers has significantly increased over the past decades [15].
both the material and the cavitation environment to which it will As a result, potential for cavitation erosion on various parts of the
be subjected. This prediction is, however, commonly expected or ship control and propulsion system, such as propeller blades, hub,
required as one designs a new turbo machinery blade or propeller or rudders and nearby ship stern sections continues to increase along
addresses whether a new claimed ‘advanced’ material will provide with the search for better erosion resistant materials.
the promised performance. To do so, the industries have to rely Proper evaluation of new materials for their resistance to cav-
on laboratory testing, using accelerated erosion testing methods itation erosion requires a comprehensive effort contrasting the
and comparative tests between the new material and previously “intensity” of the cavitation field with the “resistance” of the
used materials. This raises questions, such as: (a) How to trans- material. In the absence of historical data on the performance
pose the accelerated test results to the operation at full scale of of a proposed new material in the target cavitating flow fields,
the new design? (b) How accurate is it to accept that ranking and the designer and the decision maker have to rely on laboratory
quantitative erosion rate ratios remain the same between the accel- experimental studies. Field erosion studies have been conducted
erated method erosion tests and the full scale erosion, especially for hydraulic turbines and pumps (e.g. [15–19]), but for marine
that previous studies indicate that the erosion resistance of mate- applications small scale laboratory tests are more common. The lab-
rials sometimes depends on the intensity of the cavitation field oratory experimental studies aim at determining within required
[1–6]. There have been numerous recent studies to better under- short time periods an evaluation of the new material, whereas in
stand the cavitation erosion and attempts to model the physical the real field cavitation erosion is expected to not occur but after
process involved [7–14]. a long duration of exposure. Such accelerated erosion test tech-
niques include the utilization of ultrasonic vibration to generate
the cavitation [20–22], cavitation flow loops with strong flow sep-
∗ Corresponding author. Tel.: +1 301 604 3688; fax: +1 301 604 3689. aration or venturi effects [23–26], and submerged cavitating jets
E-mail address: jkchoi@dynaflow-inc.com (J.-K. Choi). [6,27–29] among other methods. There are also attempts to test the

0043-1648/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
doi:10.1016/j.wear.2012.01.008
54 J.-K. Choi et al. / Wear 278–279 (2012) 53–61

model propeller in the water tunnel [30]. Some of these techniques

are standardized and follow the American Society for Testing and
Materials (ASTM) Standards [31]. The ultrasonic technique and the
liquid jet technique are the two most popular laboratory tech-
niques for testing cavitation erosion characteristics of materials.
In the present study, the ultrasonic cavitation and cavitating jets at
different intensity were used.
In this paper, we follow the progression of erosion (i.e. time
history of material loss) and represent the progression using a
mathematical function. This is comparable to previous attempts
to use Weibull functions to describe the mass loss curves [32]. This
was not used here as is, since Weibull functions have the limita-
tion that the terminal erosion rate (as time goes to infinity) has
to be zero. In this paper, we propose an improved mathematical
model to erosion progress, and show that it is useful to compare
different erosion progressions for different materials under a large Fig. 1. Erosion progression curve for aluminum 1100-0 obtained in a cavitation jet
range of cavitation intensities for both cavitating jet and ultrasonic erosion test. Inserts show a picture during the incubation period and another picture
cavitation. during the acceleration period.
The aim of the current work is to understand the relative aggres-
sivity of the cavitation fields generated by ultrasonic cavitation
and jet cavitation of various driving pressures and to identify the hardening of the surface. Cavitation peening techniques take
relative erosion resistance ranking of the tested materials. The advantage of this phase to render the material more resistant to
dependency of such ranking on the cavitation intensity is also stress. During this initial phase, permanent deformation may occur,
addressed in this study. sometimes accompanied with plastic flow and local displacement
of material micro particles, as well as the development in the later
stage of micro-cracks for brittle materials. On a weight loss versus
2. Background: materials response to cavitation loads time curve (Fig. 1) this is the initial very short period where lit-
tle material loss is observed. This can be difficult to observe in
Cavitation erosion, no matter where and how it is generated, some accelerated tests, but its duration is actually very important
results from the repeated impulsive loading of the material by to the determination of the life extent of the cavitating device (e.g.
high intensity short duration pressures loads, due to shock waves propeller in the full scale application). Following this period, the
and bubble reentrant jet impacts [23,24,34–37]. These are diffi- erosion process accelerates.
cult to measure but can be inferred from acoustic signals and pit It is known that the weight loss curve has an S shape, as illus-
measurements [7,27,29,33]. Statistical correlations can be obtained trated in Fig. 2, which shows an erosion acceleration phase during
between these measurements and can be associated with the facil- which the erosion rate increases until attaining a maximum. This
ity producing the erosion and with empirically accepted cavitation is called the Accumulation or Acceleration Period. In this phase, the
intensity indicators, such as flow speed, ambient pressure, ampli- material experiences increased fracture and weight loss following
tude and frequency of ultrasonic horn. While “weak” materials may the end of hardening in the incubation period. The extent of this
fail rapidly under the repeated shock waves and jet impacts, a more zone depends upon the strain-hardening properties of the mate-
“resistant” material will accumulate stain and experience over a rial and involves microscopic chunks of material being removed
long period the symptoms of fatigue. Initially the material sur- following propagation of large cracks in between the grains of the
face gets deformed and is modified microscopically without any material. The accumulation period ends once the surface proper-
loss of material (incubation period). This is accompanied by work ties of the material have changed so much that an interaction begins

Fig. 2. Typical G32 test erosion curves: weight loss S-curve and erosion rate versus time curve.
 J.-K. Choi et al. / Wear 278–279 (2012) 53–61 55

Fig. 3. Ultrasonic technique eroded samples pictures. Left: tested G32 metallic button sample; right: eroded composite material sample from the alternative G32 method.

to occur between the new material surface shape and the cavita- 3.1. Ultrasonic cavitation erosion testing – ASTM G32
tion region. The new roughness affects the cavity dynamics and
entrapped gas and liquid in the deep craters start cushioning the In ultrasonic cavitation tests, the cavitation is generated by a
bubble collapse. This time period is known as the Attenuation Period vibratory device employing a magnetostrictive ultrasonic horn. A
(or Deceleration Stage). Finally, a local equilibrium between the ero- sample “button” of the material being tested is affixed to the end
sive power of the cavitation field and the response of the material of the horn and is subjected to cavitation resulting from the vibra-
occurs and the erosion process enters the Steady-State Period (or tions of the horn. A cavitation hemispherical cloud forms at the tip
Terminal Stage), where the rate of weight loss reaches a quasi con- of the horn and executes severe dynamics resulting in bubble cloud
stant value, or a linear behavior of the weight loss S-curve. The clear growth and collapse. In an “alternative” G-32 test configuration
separation between the four regions described above depends on (also known as a stationary specimen method), the horn tip is placed
the type of materials. at a small distance from the stationary material sample and a rather
Since the first three stages of the cavitation erosion are quite cylindrical cavitation cloud is generated in between the sample
unsteady and dynamics, correlations between various scales, var- and the face of tip of the horn equipped with a strongly cavita-
ious materials, or between accelerated and real field tests have tion resistant “button” (e.g. Titanium). In the standard G-32 test the
concentrated in previous work mostly on the characteristics of the temperature, liquid beaker volume, horn tip submergence beneath
steady-state period [1,2]. Since the erosion rate in that region is the free surface, frequency, and amplitude of the oscillations are all
constant it has been used to characterize the material. This does prescribed by the ASTM method [31]. Our tests presented in this
not however work for a weak coating which may fail and delami- paper deviated from the ASTM G32 in selecting (for both direct and
nate before getting to this stage. A new approach we present in this alternative methods) a sample diameter of 0.5 inch instead of 0.626
paper is to fit the full time history curve with a single mathematical inch.
function. The usual test procedure is to expose the sample to cavitation
for a selected period of time, interrupt the test, remove the sample,
and record weight to enable calculation of weight loss as a function
3. Accelerated erosion tests
of time. The sample is then returned for additional time intervals of
erosion. In this study, we also recorded other characteristics such as
Several laboratory techniques to generate cavitation have been
volume of erosion imprint, maximum width and depth, and took
used conventionally to study cavitation erosion in a controlled envi-
photographs of the evolution of the eroded region. Fig. 3 shows
ronment and in an accelerated manner. Accelerated erosion tests
examples of eroded samples tested by the ultrasonic cavitation.
involve subjecting the considered material to an erosion field that
is significantly more “intense” than the actual cavitation that the
studied material will be subjected to (so far, using either much 3.2. Cavitating jets – ASTM G134 and others
higher ‘repetition rate’ of the cavitation events, more energetic
events, more developed cavitation, or a combination of these). Cavitation intensity produced by cavitating jets can be var-
However, there are reasons to believe that this choice has to be ied in a very wide range through adjustment of the type of jet,
done much more scientifically than is often practiced, since the the jet velocity, the jet diameter, the jet angle, the standoff dis-
objective should be to accelerate the erosion while not subjecting tance, the ambient pressure in which they are discharged [28]. This
the material to a different cavitation regime or another range of flexibility makes a cavitating jet a great research and test tool to
load levels. study parametrically the effect of cavitation intensity on materi-
Accelerated erosion laboratory techniques include ultrasonic als behavior. The cavitation generated by a cavitating jet provides
flows, cavitation flow loops with strong flow separation, rotating realistic cavitation bubble clouds with distribution of various size
disks, cavitating venturi flows, vortex generators, and submerged micro bubbles, shear flows with vortices, and dense bubble clouds,
cavitating jets [20–29]. Some of these techniques are standard- which collapse on the sample. With the control of the operating
ized by the American Society for Testing and Materials (ASTM) pressure, the jet angle, and the stand-off, the testing time can be
Standards and include Standard G-32 “Test Method for Cavitation adjusted to provide either quick erosion for initial screening or
Erosion Using Vibratory Apparatus” and G-134 “Test Method for Ero- time-accelerated erosion more relevant to the real flows.
sion of Solid Materials by a Cavitating Liquid Jet” [31]. The ultrasonic The cavitating jet erosion test setup used in this study consisted
technique and the liquid jet technique are the two most popular of a DynaJets® , cavitating jet nozzle, a sample holder, a water tank,
techniques for testing cavitation erosion characteristics of materi- and a pump. The sample holder ensured that the sample was placed
als, and the results from both methods will be studied here when back precisely at the same location when it was returned from any
eroding the same materials. examination. The nozzle and the sample were submerged in the
56 J.-K. Choi et al. / Wear 278–279 (2012) 53–61

Table 1
Materials tested in this study.

Substrate Coating

Aluminum 7075 None

Stainless steel 316 None
Stainless steel A2205 None
Nickel aluminum bronze (NAB) None
High yield stress steel (HY-80) None
Aluminum 1100-O None
Aluminum 7075-T651 Anodized
Low temperature colossal super-saturation (LTCSS) on LTCSS
stainless steel A2205 [38]

water tank so that the sample was subjected to only cavitation

bubbles but not to water droplet impacts. A photograph showing a
typical setup of the jet nozzle and the sample in its holder is shown
in Fig. 4. The overall test procedure is similar to that of the G32 tests
other than using a cavitating jet. The time intervals are appropri-
ately selected to capture a cumulative weight loss curve displaying
as much as possible the characteristic S-curve.
Fig. 5. Cavitation erosion average depth versus time on AL7075 resulting from a
DynaJets® nozzle cavitation at different nozzle pressures (jet velocities). The solid
4. Erosion progression for different cavitation intensities
lines are analytical curves obtained from the mathematical erosion evolution model
shown in Eq. (6).
In pursuit of definitions of the “intensity of the cavitation” in the
flow filed and the “resistance of the material” to the cavitation ero-
Fig. 5 shows an example of the mean depth of erosion of Al 7075-
sion, we have conducted systematic tests and obtained erosion data
T651 subjected to the G32 ultrasonic cavitation field and to a set of
on several materials under various cavitating jet pressures (speeds)
cavitating jets with different pressures across the nozzle. The ero-
and ultrasonic cavitation. Table 1 shows the materials tested, while
sion curve marked with G73 is obtained from a liquid impingement
the various cavitation intensity fields are described further down.
erosion test on the same material following the ASTM G73 stan-
dard [31]. The figure illustrates that both the jets and the ultrasonic
4.1. Mathematical representation of long duration erosion
device result in similar shape MDE curves with the G32 falling a lit-
progression history
tle below the 1000 psi cavitating jets used in these tests (other tests
indicate that G32 is more similar to a 700–800 psi DynaJets® ). Data
In order to study the effect of the imparted cavitation field
points with symbols are from the actual measurements, while the
energy on the evolution of the cavitation erosion over a long period
curves are representation of the erosion data using the following
of time, a set of erosion tests were conducted at different jet pres-
mathematical expression:
sures. The test data were expressed by mathematical erosion curves
so that the jet pressure effect can be sought for the whole erosion h = 0, t  = t − tinc ≤ 0,
n (2)
history of different cases. To account for differences in densities of h = h1 (1 − e−t ) + h2 t  ˇ ,
different materials and the size of the cavitation area, the erosion
time history is presented in terms of mean depth of erosion (MDE), where tinc is the incubation time, h1 , h2 are depth constants and n
h, versus time, defined as: and ˇ are time parameters which characterize the cavitation ero-
sion progression.
m(t)
h(t) = , (1)
A(t) 4.1.1. Definition of the characteristic scales
where m(t) and A(t) are the mass loss and eroded area at time t, and Eq. (2) can be normalized using a characteristic depth, h*, and
 is the material density. a characteristic time, t*, and defining h̄ as the normalized erosion
depth by and t̄ as the normalized time:
h t − tinc
h̄ = , t̄ = . (3)
h∗ t∗
This leads to the following MDE normalized equation:
n
h̄ = 1 − e−t̄ + ˛t̄ ˇ . (4)

The constants, n, ˛, and ˇ, are parameters which characterize the

cavitation erosion progression and were found in our tests to date,
illustrated by the examples in Fig. 5, as having the values:
1
n = 2, ˛= = 0.37, and ˇ = 1.2. (5)
e
The erosion evolution was therefore found to be characterized by:
2 1 1.2
h̄ = 1 − e−t̄ + t̄ . (6)
e
All three parameters (h*, t*, and tinc ) are characteristics of the
Fig. 4. Nozzle and sample holder in the Dynaflow’s “7 ksi – 5 gpm” cavitation jet response of the specific material to the particular cavitation ero-
erosion test loop. sion field. For all times less than tinc there is no weight loss of the
 J.-K. Choi et al. / Wear 278–279 (2012) 53–61 57

Fig. 6. Cavitation erosion mean depth of erosion rate, dh̄/dt̄ (left) and its time derivative, d2 h̄/dt̄ 2 (right) versus nondimensional time, t̄, as described by the mathematical
erosion evolution model shown in Eqs. (6)–(8). Maximum erosion depth rate, (dh̄/dt̄)max , occurs when t = 0.75t*.

material. h* and t* are such that h = h* at t − tinc = t*, or h̄ = 1 at t̄ = 1. Fig. 7 shows all Fig. 5 data in a normalized format and illus-
The first and the second derivatives of the normalized MDE can be trates the good quality of the matching of the mathematical model,
derived for this case: Eq. (6), to the experimental data. Notice that specific tests cover
different ranges of the erosion curve, but all test data fall on
dh̄ 2 1.2 0.2 one standardized shape of erosion curve. Fig. 8 shows the corre-
= 2t̄e−t̄ + t̄ , (7)
dt̄ e sponding normalized erosion rate (dh̄/dt̄) curves, which fall on
one curve as expected. Table 2 shows the standard deviation of
d2 h̄ 2 0.24 −0.8
= 2e−t̄ (1 − 2t̄ 2 ) + t̄ . (8) the error when the measured data are expressed by Eq. (6). The
dt̄ 2 e normalized standard deviation is small and is seen to be of the
2
These two functions are shown in Fig. 6, which illustrates that the order
   coefficient of determination (R =
 of 1–3% in all cases. The
2 2
mean erosion depth rate achieves its maximum when t = 0.75t*. 1− (yi − fi ) / (yi − ȳ) ) represents how well the prediction
This provides a good interpretation for t*, as being 4/3 the time at (fi ) describes the actual data (yi ). The same data and curves are
which the rate of the mean depth of erosion attains its maximum. shown in log–log scale in Fig. 9 in order to enhance the various
It would have been more eloquent to change variables to have the regions of the history. It is noticeable that the empirical represen-
maximum at t*, but this would affect the simplicity of Eq. (6), which tation curve fits are good from the acceleration stage to the terminal
is very easy to remember in the present form. erosion stage, but obviously do not capture the incubation period
where the erosion depth is almost zero.

Fig. 7. Normalized cavitation erosion average depth versus normalized time on

AL7075 resulting from DynaJets cavitation at different nozzle pressures (jet veloci-
ties) and G32. The solid line is obtained from the mathematical erosion progression Fig. 8. Comparison of normalized erosion rate curves of AL7075 erosion data
model shown in Eq. (6). obtained by various test methods.
58 J.-K. Choi et al. / Wear 278–279 (2012) 53–61

Table 2
Normalized standard deviation or error in expressing the erosion data with Eq. (6).

Test method and intensity h*, mm t*, min h*/t*, mm/min Standard deviation in h/h* Coefficient of determination, R2

G32 0.041 232 1.8e−4 2.4% 0.9985

1000 psi jet 0.593 689 8.6e−4 3.2% 0.9920
2000 psi jet 0.482 373 1.3e−3 2.2% 0.9972
5000 psi jet 16.3 533 3.1e−2 1.4% 0.9919
7000 psi jet 22.0 274 8.0e−2 1.4% 0.9882

−0.3547
t ∗ = 132.7(pjet ) , t ∗ in hour and pjet in psi. (10)

Using the relationship between the jet pressure and the jet veloc-
ity, the above scaling relationship can be expressed using the jet
velocity as follows:

h∗ = 2.4 × 10−7 (Vjet )4.4326 , h∗ in ␮m and Vjet in m/s, (11)

t ∗ = 337(Vjet )−0.7094 , t ∗ in hour and Vjet in m/s. (12)

The exponent 4.43 over the jet velocity in the erosion depth expres-
sion is consistent with the typical values of 4–5 reported in the
literature [1,2].

4.2. G32 and cavitating jet erosion for different materials

This section considers a set of erosion tests on different mate-

rials and examines whether the proposed mathematical model
applies to these different materials. An example of G32 raw test
data – erosion depth versus time – for various materials is shown in
Fig. 11. The corresponding curves using the mathematical expres-
Fig. 9. Cavitation erosion average depth versus time on AL7075 resulting from a
DynaJets nozzle cavitation at different nozzle pressures (jet velocities) shown in sions from Eq. (6) with the best fits are also shown. Fig. 12 shows
log–log scales. The solid lines are analytical curves from the mathematical erosion the model curve with all the normalized data fallen on it. This
model shown in Eq. (6). The almost horizontal initial 2 ksi dots correspond to the graph shows the mass loss normalized by the characteristic mass,
incubation period. m* which corresponds to the characteristic MDE, h*. It is notice-
able that the individual erosion curves cover different portions of
The characteristic erosion depth, h*, and the characteristic time, a common S-shaped normalized curve depending on the erosion
t*, are directly related for a given material to the cavitating jet speed resistance and the test duration. As we can see, the erosion progres-
(or upstream pressure). This is illustrated in Fig. 10. As expected, sion data for all materials is very well fitted with the mathematical
the characteristic erosion depth increases and the characteristic expression (6).
erosion time decreases as the cavitation intensity (i.e. the jet pres- Similarly, DynaJets® cavitating jets erosion tests were con-
sure) increases. The scaling relationship from this data set can be ducted with a 5000 psi jet for the same materials presented in
expressed as: Fig. 11 for G32. The cavitating jet results are shown in Fig. 13.
Here again, all of the data are also relatively very well fitted with
h∗ = 8 × 10−5 (pjet ) h∗ in ␮m and pjet in psi, (9)
2.2163
, the mathematical expression (6). Fig. 14 shows the normalized

Fig. 10. Scaling of the characteristic erosion depth, h* (left) and the characteristic time, t* (right) with the cavitating jet nozzle pressure.
 J.-K. Choi et al. / Wear 278–279 (2012) 53–61 59

Fig. 11. Mean depth of erosion (based on mass loss) versus time curves obtained from G32 tests for various materials. Solid lines are fits of the data with the mathematical
model of Eq. (6). Modified ASTM G32 method was used in these tests; (reg) represents the direct method, (alt) represents the alternative method.

erosion time history of the three selected materials (aluminum,

nickel aluminum bronze, and stainless steel) under various ero-
sion intensities. A nice collapse onto one normalized S-curve is well
demonstrated.

5. Discussion

A mathematical description of the erosion time history, as pre-

sented above, is very useful in the task of comparing cavitation
erosion behavior of different materials exposed to the same or dif-
ferent known cavitation intensity, and where individual erosion
tests were conducted under different cavitation intensities or dif-
ferent duration of erosion exposures. It appears that the overall
cavitation erosion resistance to a cavitation field is characterized
by the two parameters: h*, the characteristic mean erosion depth
and t*, the characteristic erosion time, with the characteristic erosion
rate defined as h*/t*.
Tables 3 and 4 show the numerical coefficients used to fit the
erosion data of the different materials tested by the G32 method

Fig. 13. Mean depth of erosion (based on mass loss) versus time curves obtained
from a 5000 psi jet cavitation and curve fits of the data to the mathematical model
shown in Eq. (6).

Table 3
Parameters of the erosion mathematical model and characteristic erosion rates for
the different materials tested with G32 ultrasonic tests. The materials are ordered
by increasing value of the characteristic erosion depth rate.

Material tested Characteristic Characteristic Characteristic

with G32 erosion depth, h*, time, t*, min erosion depth rate,
alternative method mm h*/t*, ␮m/min

LTCSS 0.0014 1100 0.001

NAB 0.0498 4190 0.012
SS A2205 0.0693 4520 0.015
HY80 0.0622 1180 0.053
SS316 0.0621 771 0.081
AL7075 0.197 2100 0.094
AL7075 0.0419 232 0.181
Anodized AL7075 0.0186 98 0.190
Fig. 12. Curves of the normalized mass loss versus normalized time. m* is the char-
AL1100 0.0139 35 0.392
acteristic mass loss. All curves start from the origin, but for the same physical time,
the tested materials reached different erosion stages.
60 J.-K. Choi et al. / Wear 278–279 (2012) 53–61

Table 4
Parameters of the erosion mathematical model and characteristic erosion rates for the different materials tested with 5000 psi DynaJets® cavitating jet tests. The materials
are ordered by increasing value of the characteristic erosion depth rate.

Material tested with Characteristic erosion Characteristic time, t*, min Characteristic erosion
5000 psi cavitating jet depth, h*, mm depth rate, h*/t*, ␮m/min

NAB 4.3 941 4.57

LTCSS 4.0 348 11.5
SS A2205 14.2 776 18.3
SS316 17.9 804 22.2
Anodized AL7075 15.0 533 28.3
AL7075 11.0 374 29.5
HY80 9.7 223 43.8

Table 5
Comparison of erosion rates between G32 tests and 5000 psi cavitating jet erosion tests. Materials are presented ranked by the jet erosion results.

Material G32 ultrasonic 5000 psi cavitating Ratio of erosion 750 psi jet, deduced Ratio of erosion
alternative jet (␮m/min) rates (5 ksi jet/G32) from 5000 psi results rates (750 psi
(␮m/min) (␮m/min) jet/G32)

NAB 0.012 4.6 381 0.0351 1.2

LTCSS 0.001 11.5 11,500 0.0878 38.5
SS A2205 0.015 18.3 1220 0.140 4.1
AL7075 Anodized 0.190 28.3 149 0.216 0.5
AL 7075 0.094 29.5 314 0.225 1.1
HY80 0.053 43.8 826 0.334 2.8

and the 5000 psi jet cavitation. In these tables, the materials are Table 5 shows a comparison of the erosion rates of the materials
listed in the order of increasing characteristic erosion rate. The as obtained by the two accelerated erosion methods using differ-
rankings by the two test methods in Tables 3 and 4 agree in general, ent levels of cavitation intensity. The ratios of the erosion rates
but a couple of materials did not rank the same way. For example, between the jet at 5000 psi and the G32 method, for example, vary
HY80 was more resistant than SS 316 in G32 tests, but the order vastly from material to material. This ratio varies by two orders
reversed in the 5000 psi cavitating jet tests. From the observed ero- of magnitudes between the weakest and the strongest erosion
sion rates of the two test methods, it is obvious that the cavitation resistant materials and is, as expected, very high for the strongest
field of a 5000 psi jet is much more intense and erosive than the materials as for these the high intensity jets produce significant
cavitation field generated in G32 ultrasonic tests. This tells that, rel- erosion, while the G32 is just at the limit of solicitation of the
ative to SS 316, HY80 has greater resistance in a weaker cavitation material.
field but is less resistant in an intense cavitation field. Comparing A better comparison is to run the two methods – cavitating
NAB and LTCSS is also interesting. The LTCSS shows higher ero- jets, ultrasonic cavitation – at comparable cavitation intensity. This
sion resistance than NAB in the weaker cavitation field of the G32 would be when the jet intensity is in the range of 700–800 psi. The
test, but not in the intense cavitation field of 5000 psi jet cavitation last two columns in Table 5 show such a comparison using the pres-
test. These tell that the material erosion response of some materials sure correlations in Fig. 10, to deduce the jet erosion rates as these
depends on the intensity of the cavitation field. were not obtained experimentally. The erosion rates correspond-
ing to the 750 psi jet were estimated from the 5000 psi jet results
using the exponents 2.22 and −0.35 in (9) and (10). These estimated
erosion rates are comparable to those obtained from G32 tests, as
the ratios in the last column are mostly of order 1. This indicates
that the exponents on the jet pressure in (9) and (10) obtained from
tests using Al 7075 may be applied to other materials, and that the
cavitation intensity of G32 tests is similar to that of the 750 psi
cavitating jet. One exception is noticed for LTCSS, which showed
extremely low erosion rate from the G32 test. A possible expla-
nation is that the cavitation aggressiveness of G32 is very low for
this especially hardened surface of the stainless steel and that it is
able to barely exceed the incubation period, while the 5000 psi jet
cavitation proceeds well beyond into the erosion of the material
penetrating the hard surface layer after which the erosion proceed
in the unhardened steel.

6. Conclusions

In pursuit of practical definitions of the “intensity of the cav-

itation” of a given cavitating flow field and of the “resistance of
the material” to cavitation erosion, we have conducted systematic
long term erosion tests (i.e. much beyond the incubation period)
Fig. 14. Curves of the normalized erosion depth versus normalized time for alu-
on several materials using two different accelerated erosion test-
minum, NAB, and stainless steel tested with cavitating jets at various pressures.
Curves are the mathematical expressions (6) and the symbols are the measured ing methods. The test methods used were the modified ASTM G32
data points. Ultrasonic Cavitation Erosion tests and the DynaJets® cavitating
 J.-K. Choi et al. / Wear 278–279 (2012) 53–61 61

jets run at various jet speed or nozzle pressures. Through accel- [9] S.M. Ahmed, K. Hokkirigawa, Y. Ito, R. Oba, Scanning electron microscopy obser-
erated erosion testing and analysis a comparative evaluation and vation on the incubation period of vibratory cavitation erosion, Wear 142
(1991) 303–314.
energy ranking of the materials and the erosion testing methods [10] G. Bregliozzi, A. Di Schino, S.I.-U. Ahmed, J.M. Kenny, H. Haefke, Cavitation
were established. wear behavior of austenitic stainless steels with different grain sizes, Wear
This was done by expressing the erosion progression using a 258 (2005) 503–510.
[11] M. Dular, O. Coutier-Delgosha, Numerical modeling of cavitation erosion, Int. J.
simple mathematical equation with characteristic parameters. A Numer. Meth. Fluids 61 (2009) 1388–1410.
characteristic time, t*, which occurs at 75% of the time of maximum [12] H. Soyama, M. Futakawa, K. Homma, Estimation of pitting damage induced by
rate of erosion, and a corresponding material characteristic mean cavitation impacts, J. Nucl. Mater. 343 (1–3) (2005) 116–122.
[13] R. Fortes Patella, J. Reboud, A. Archer, Cavitation damage measurement by 3D
erosion depth, h*, were defined. Using these to normalize the mean
laser profilometry, Wear 246 (2000) 59–67.
depth erosion and time, the simple erosion depth time evolution [14] J.-P. Franc, Incubation time and cavitation erosion rate of work-hardening
2
expression found is: h̄ = 1 − e−t̄ + e−1 t̄ 1.2 . This equation was able materials, J. Fluids Eng. 131 (2009) 021303.1–021303.14.
[15] M. Billet, The special committee on cavitation erosion on propellers and
to represent all our experimental erosion data for the various cavi-
appendages on high powered/high speed ships, in: 24th International Towing
tation sources and the various materials within only 2–3% deviation Tank Conference (ITTC), Vol. III, UK, 2005.
(Coefficient of determination, R2 > 0.988). Such a mathematical rep- [16] M. Grekula, G. Bark, Experimental study of cavitation in a Kaplan model turbine,
resentation of the erosion time evolution enables one to compare in: 4th Int. Symposium on Cavitation, CAV2001, Pasadena, CA, 2001.
[17] M. Farhat, P. Bourdon, Extending repair intervals of hydro turbines by miti-
various materials tested under cavitation fields of different intensi- gating cavitation erosion, in: CEA Electricity ’98 Conference and Exposition,
ties following determination of the two characteristic parameters: Toronto, 1998.
characteristic mean depth of erosion and characteristic erosion [18] M. Farhat, P. Bourdon, P. Lavigne, R. Simoneau, The hydrodynamic aggressive-
ness of cavitating flows in hydro turbines, in: ASME Fluids Eng. Div. Summer
time. The various tested materials were ranked based on their cav- Meeting, FEDSM’97, 1997.
itation resistance. Comparisons of the erosion rates from the two [19] Guideline for Prediction and Evaluation of Cavitation Erosion in Pumps, Turbo-
accelerated erosion methods using very different levels of cavita- machinery Society of Japan, 2010.
[20] F.G. Hammitt, C. Chao, C.L. Kling, T.M. Mitchell, D.O. Rogers, Round-robin test
tion intensity showed that the erosion response of some materials with vibratory cavitation and liquid impact facilities of 6061-T 6511 aluminum
depends on the cavitation intensity. Several other approaches to alloy, 316 stainless steel and commercially pure nickel, materials research and
quantify and understand the dynamics of cavitation intensities are standards, ASTM 10 (1970) 16–36.
[21] C. Chao, F.G. Hammitt, C.L. Kling, ASTM Round-Robin Test with Vibratory Cav-
presented in companion publications. These included pitting tests
itation and Liquid Impact Facilities of 6061-T6 Aluminum Alloy, 316 Stainless
during the incubation period [33] and cavitation field pressure mea- Steel, Commercially Pure Nickel, The University of Michigan Report MMPP-
surements [39]. All these efforts will be eventually combined to 344-3-T/01357-4-T, 84, 1968.
[22] K.H. Light, Development of a Cavitation Erosion Resistant Advanced Material
provide a practical tool for erosion predictions.
System, M.S. Thesis, Mechanical Eng., University of Maine, 2005.
[23] M.A. Dominguez-Cortazar, J.P. Franc, J.M. Michel, The erosive axial collapse of
Acknowledgements a cavitating vortex: an experimental study, J. Fluids Eng. 119 (1997) 686–691.
[24] F.G. Hammitt, Damage to solids caused by cavitation, Philos. Trans. Roy. Soc.
Lond. A, Math. Phys. Sci. 260 (1110) (1966) 245–255.
Support for this work was provided by Office of Naval Research [25] X. Escaler, F. Avellan, E. Egusquiza, Cavitation erosion prediction from inferred
(ONR) under Contract Number N00014-08-C-0450, monitored by forces using material resistance data, in: 4th Int. Symposium on Cavitation,
Dr. Ki-Han Kim. We are thankful for his support, suggestions, and CAV2001, Pasadena, CA, 2001.
[26] J.S. Baker, Cavitation Resistant Properties of Coating Systems Tested on a Ven-
discussions. The authors would also like to acknowledge Dr. Mar- turi Cavitation Testing Machine, Bureau of Reclamation, Research Laboratory
tin Donnelly, NSWC Carderock Division, Dr. Jean-Pierre Franc, LEGI, and Services Division, Denver, CO, 1994.
Grenoble, France, Dr. Ayat Karimi, EPFL, Lausanne, Switzerland, and [27] T. Momma, A. Lichtarowicz, A study of pressures and erosion produced by
collapsing cavitation, Wear 186–187 (Part 2) (1995) 425–436.
Dr. Farrell Martin, NRL, for useful discussion. Finally, the authors [28] G.L. Chahine, P. Courbière, Noise and erosion of self-resonating cavitating jets,
would also like to recognize the long hours of efforts by Emmanuel J. Fluids Eng. 109 (1987) 429–435.
Coleman and Patrick Aley of Dynaflow, Inc. in conducting the labo- [29] M.K. Lee, W.W. Kim, C.K. Rhee, W.J. Lee, Liquid impact erosion mechanism and
theoretical impact stress analysis in TiN-coated stream turbine blade materials,
rious erosion experiments. Metall. Mater. Trans. A 30A (1999) 961–968.
[30] W. Pfitsch, S. Gowing, D. Fry, M. Donnelly, S. Jessup, Development of mea-
surement techniques for studying propeller erosion damage in severe wake
References fields, in: Proc. 7th Int. Symposium on Cavitation, CAV2009, Ann Arbor, MI,
2009.
[1] A. Thiruvengadam, Handbook of Cavitation Erosion, Hydronautics Technical [31] Annual Book of ASTM Standards – Section 3 Material Test Methods and Analyt-
Report 7301-1, 1974. ical Procedures, vol. 03.02, American Society for Testing and Materials (ASTM),
[2] P. Eisenberg, H.S. Preiser, A. Thiruvengadam, On the mechanisms of cavitation 2010.
damage and methods of protection, Trans. Soc. Nav. Archit. Mar. Eng. 73 (1965) [32] Y. Meged, Modeling of the initial stage in vibratory cavitation erosion tests by
241–286. use of a Weibull distribution, Wear 253 (2002) 914–923.
[3] M.A. Dominguez Cortazar, Le Cavermod, Modèle Physique de l’érosion de Cav- [33] J.-P. Franc, M. Riondet, A. Karimi, G.L. Chahine, Material and veloc-
itation: Qualification Expérimentale et Numérique, Thèse Universite Joseph ity effects on cavitation erosion pitting, Wear 274–275 (2012) 248–259,
Fourier Grenoble I, 1992. doi:10.1016/j.wear.2011.09.006.
[4] J.L. Reboud, Réponse Impulsionnelle d’un Milieu Elastoplastique: Application [34] K.A. Morch, Dynamics of cavitation bubbles and cavitating liquids, Treatise
a L’étude de l’érosion de Cavitation, Thèse, Institut National Polytechnique de Mater. Sci. Technol. 16 (1979) 309–355.
Grenoble, 1987. [35] J.R. Blake, B.B. Taib, G. Doherty, Transient cavities near boundaries. Part I. Rigid
[5] F. Pereira, F. Avellan, Ph. Dupont, Prediction of cavitation erosion: an energy boundary, J. Fluid Mech. 170 (1986) 479–497.
approach, J. Fluids Eng. 120 (1998) 719–727. [36] H. Zhang, J. Duncan, G.L. Chahine, The final stage of the collapse of a cavitation
[6] P.A. March, Evaluating the relative resistance of materials to cavitation erosion: bubble near a rigid wall, J. Fluid Mech. 257 (1993) 147–181.
a comparison of cavitating jet results and vibratory results, in: Proc. Cavitation [37] G.L. Chahine, T.O. Perdue, Simulation of the three-dimensional behavior of an
and Multiphase Flow Forum, ASME, Cincinnati, 1987. unsteady large bubble near a structure, drops and bubbles, in: T.G. Wang (Ed.),
[7] H. Soyama, A. Lichtarowicz, T. Momma, E.J. Williams, A new calibration method Third International Colloquium, American Institute of Physics, Monterey, CA,
for dynamically loaded transducers and its application to cavitation impact 1988, pp. 188–199.
measurement, J. Fluids Eng. 120 (1998) 712–718. [38] S. Collins, P. Williams, Low-temperature colossal supersaturation, Adv. Mater.
[8] S. Hattori, M. Takinami, O. Tomoaki, Comparison of cavitation erosion rate Process. (September) (2006) 32–33.
with liquid impingement erosion rate, in: 7th Int. Symposium on Cavitation, [39] J.-P. Franc, M. Riondet, A. Karimi, G.L. Chahine, Impact load measurements in
CAV2009, Ann Arbor, MI, 2009. an erosive cavitation flow, J. Fluid Eng. 133 (December) (2011).