Experimental investigations of cavitation erosion in the incubation period

INVESTIGATIONS Of CAVITATION EROSION IN THE

INCUBATION PERIOD

BY J. VARGA* AND G. SEBESTYÉN **

Introduction

Cavitation erosion investio-ations carried out h"

• t> .'

varI~us n!etl.lOds, using the conventional hydrody- nannc prmclples (flow devices) or the accelerated n!ethods (magn.e.tostriction devices) are primarily aIIl1ed to classIfy the relative cavitation erosion properties of different materials. Such experimen- tal investigations involve the measurement of weight or volume losses of difTerent materials as the function of the arbitrarily choosen experimen- tal time, at a given state of cavitation. Considerino'

that the weight or volume loss pel' unit time

doe~

not represent a constant value but chano-es with the time and furthennore the physical

stat~

^{of the}

mate rial will similarly change as the damage conti- nues, none of the difTerent measurements used to

ch~rac1erizeerosion damage is perfect and wc are obhged to search for the available criteria.

In carrying out a large number of experimental studies of this type, !ittle attention was paid to the more minute investigation of the so-called "incu- bation" period. Materials exposed to cavitation effect will exhihit considerable damage only artel' the incubation period. In that stage the collapse energy of the bllhhles IH'oduce only the fatirJuin_{... b b}o'

on the surface of the test specimen, without the appearance of appreciable damage and hence loss of weight. Isolated small pits are found on the

(') Professor, Department of Hydraulic lvlachinery, Budapest Teclmical University, Hungary.

(") Hesearch Engineer, Department of Hydraulic Machi- nery, Budapest Teclmical University, Hunga~·y.

surface only. Beyond that stage at the same cavi- tation condition it ,viII produce a quick destruction of the test specimen. This character of cavitation èrosion damage is confirmed repeatedly.

This nature of cavitation erosion damage pro- cess brought about the necessity of subjecting the incubation period to a close examination. A few results of these investigations 'will be discussed below. The experimental investigations have been carried out in a test section built into the closed circuit water tunnel in which the cavitation-pro- duced behind circulaI' cylinder model-acted upon lead test specimens arranged on the side wall of the test section. The test equipment has alI'eady heen discussed in detail in earlier papers [1], [21,

[3]. -

1/Loss of weight (G) due to cavitation damage plotted versus the time <-r) of experiment. - 1. 'l'he incubation period. - 2. The period of total destruction.

Perte de poids (G) provoquée par la cavitation, en fonc- tion de la durée (,) de l'expérience. - 1. Période « d'in- cubation ». - 2. Période de destruction totale.

911

Article published by SHF and available at http://www.shf-lhb.org or http://dx.doi.org/10.1051/lhb/1966058

12 15 Tthl

Perte de poids (G) en j'onction de la durée (T) de l'essai. Les dimensions de l'éclwntillon d'essai sont ^{.18 X} 200^1I1l1l;!, la vitesse d' é^{C 01/}1^l'men t e s t v^C/J = 12 m /s, et la longuellr relative de la zone de cavitation est ),

=

^3.

Perte de poids (G) en j'onction des chocs (N) subis par l'échantillon d'essai, pOlir des vitesses d'éclwntillon variables.

Données basées slIr le 7ll<!moire de Kozirev. - 1.Période «d'in- cubation ». - 2. Zone de tran- sition. -- 8. Phase de destnlc- tion rapide.

[fl ,

...i·

1

¹

" !

Il

G-G(T) AMt,8;t 200mm^f

500 1---· d-t,8mm

,

V.. - f2m/s

,·3

1

Nr.300

'<00

il

J<>O

,

200

1

/

100 /

Il

0

0

G-G{T)

9 A-48;o:200mm^t

d·1tBmm

7

).,3 8 , . . . .. VQO-12m/s

Pb

1

7_ ...

61-- ... !

1

5

j

4

il

3

2

/

1

1/

,

1

4/

\Veight. 10ss pel' unit time (G') versus the lime (T) of experiment.

Perte de poids par unité de temps (G') en j'onction de la durée (T) de l'essai.

G.IO,3 [mg]

1

2/

Weight loss (G) as the funetion of impacts (N) on the test speei- men, at ditIerent specimen speeds. Based on Kozirev's paper. ' - 1. Incubation period.

- 2. Transition zone. -- il.

Stage of rapid destruction.

3/

Weight Joss (G) plotted versus the time (-r) of expel'iment.

Dimensions of the test section are 48 X 200 s'l.mm, the now velocity is VC/J = 12 m/sec., the relative lengt.h of t.he cavitation zone is ).= il.

The critical damage period J. VARGA and G. SEBESTYÉN

A number oi' earlier experimental studies carried out with difTerent flow velocities have confirmed the statement that the curves of weight loss versus lime, in the early stages of testing are composed oi' t\vo clearly distinct zones (apart l'rom a short tran- sition section). An idealized weight loss versus Urne curve is illustrated in Figure 1. The zone marked 1 in the Figure is the incubation period, in which the deformation of the test speeîmen surface is predominant, accompanied by a moderate loss oi' weight. Zone 2 of the curve is the period of quick des trudion, the so-calied ..total destruction"

or accumulation zone. The intersection point oi' the two lines represents the end of the inëubation perioc!. In the i'ollo"\ving it will he referred to as the critical point, and the whole period as the cri- tical damage period ("""'it). The weight or volume loss pertaining to it will be called the critical weight (or volume) loss. The adjective "critical" involves the statement that from the point as the damage con tinues the physical state of the material and similarly the mechanical strength properties of the test specimens are changing, and therefore, the resistance of the test specimen to cavitation erosion decreases.

There is a concept according to which the" pure"

elIect oi' cavitation erosion manifests itself during the incubation period only hecause the damage becomes cumulative artel' a major damage occurs on the surface. ft may weil he assmned that the physical state of the material does not undergo any basic change "\vithin the incubation (critical damage) period, consequently the incubation period might be considered as the truc cavitation damage phase without any secondary phenomena. It may i'urthermore he assumed that the so-called contrac- tion work [4] i. e. specific fracture energy (the energy absorbed pcr unit volume of the material up to fracture) at the end of the critical damage period would he exhausted. This is confirmed by the i'ael that small cracks appear under the surface of the test specimen as was verified by electron- microscopie investigations [5 J. Such an interpre- tation of the critical point offers a basis for compa- ring the resistance of different test specimens 10 cavitation damage.

In the literature the division of the weight-loss curves into typical zones i'requently appears. Kozi- l'ev [6 J divides the curves obtained with the rota- ting-disk equipment into three zones. He calls the first zone with no loss oi' weight, incubation period

"\vhich is i'ollowed by a short transition section i'eaturing slight loss of weight. The third zone is the period oi' rapid destruction (Fig. 2). Bogatchev and Mints [7Jin agreementwith Figure 1 dividethe weight-Ioss curves into two sections also obtained by accelerated method. Leith [8J defined the incubation period as the time interval during which considerable plastic deformation of the test spe- cimen takes place without any apparent weight loss.

Govinda Rao

i91

approaches the subject i'rom the energetical point of view stating that the incubation period is characterized by an energy level causing plastic deformation. Thiruvengadam [10J defined 912

Influence of the flow velocity on the critical period It is difIicult to determine the critical time l'rom weight-Ioss curves plotted versus time of experi- ment because the curves dilTer l'rom the idealized curve of Figure 1

.i

ust by the presence of a short transition period between the incuhation period and the zone of total destruction (Fig. 3). Thus the intersection point of the two curves cannot he determined unambiguously. ln view of this, a method of evaluating the experimental results has been sought for permitting the determination of critical weigh t loss and critical time ,vith satisfac- tory accuracy. The transition point between the two curve sections may be regard cd the one to which a tangent circle of minimum radius can he relegated. However, it cannot be unamhiguously determined hecause the test points are usually not availahle in the desired density; even if more points were plotted, the accuracy could not he increased on account of the unavoidahle scattering of the test resulls. Therefore, a more accu rate lnethod was searched for to determine the critical value.

Plotting the quantity of material eroded during unit ti me versus the time of experimen t, the curve of Figure 4 will he obtained. Here the incuhation periml appears in the form of a straight line parallel to the abscisse. The critical point can he detennined even more perfectly and accurately by carrying out the sarne procedure with the points of Figure 4. Thus, the curve of Figure 5 will be obtained l'rom which it is possible to locate the cri- tieal point with salisfactory aeeuracy.

LA HOUILLE BLANCHE/N° 8-1966

The determination of the critical time

Plotting the eritical periods (ohtained by the ahove methods) versus the speeds pertaining to theln, the curve of Figure (j will he obtained.

Plotting same in a eo-ordinate system of logarith- mie sc ale, a straight line can he laid across tile test points (Fig. 7) that can he described by the equa-

tion ,,/yn

=

con st. in which the power exponent is

111= 5. This relationship is identical with the one fOlll1d in a previous paper [11

J

for the accumula- tion zone following the incubation zone with a given eonstan t quantity of material eroded. The points plotted in Figure 7 represent the test results with two dilTerent thieknesses

Ca = ;)

mm and 8 mm) of lead test specimens. Obviously, the cri- tical period is the same with both types of test spe- 913 the incubation period as that period after 'which the first permanent plastic dent is formed. It is also evident l'rom the statements of the ab ove authors that the opinions are not uniform in connection with the incubation period. The authors who claim that the incubation period involves no \veight loss at aIl, are probably unable to measure the very small weight losses involved.

Ihl

P,!riode critique des!rllctive par l'érosion dlle à la cavitation ('cru)' en fonction de la vitesse de l'écOlzlemen! (v",,J. Les épais- se/us des plaqlles de plomb son!

a

=

3 e! 8mm.

Yariation de G" = G/,2 en fonc- tion de la dllrée (c) de l'essai.

7/

Criticn! period of cavitation damage ('crit) versus the Ilo\\"

veloeity (v",).

Période critique de la des!rzlc- tion dlle à la cllvitation ('('ril) en fonction de la vitesse de l'écolliement (v",J.

Ic,iI"T,;rit(V_) A ..~8x200mm~

d"',Bmm f.1o=3mm

<:-o=8mm ),3

12

.'J _{C'· G'lr)}

A-48x200mm'

/

V_- 12mfs )-3

J\

~ Nr.300

1/

/

'\ ^{/ Il}

~ Il

~'JI

2

•

^KJ

_"

_'" ^{16 T}

"

32

28

"

7c,d

[hJ 1 . .

10

1'"9·^h

5

~

o \

,

. .

lait -1ërd(""~

A-48x 200mm 5 , _~^d-48mm_a-3mm

.a"8mm

"-3

0

i

9

\

,

8 i ^{i .}

7 i 6 ,

1\

i !

,

1\

J,

,

l~

3 7cr/t [hl

9 10 Il 12 13 II, 15 V_[mJ'!

6/

Critica! period of cavitation crosiou dalnage Ct('l'it) versus the no\\" velocity (voc )' The thic!mess of the Ieae! sheds are

[[ = 3 nnd 8 mm.

5/Variation of G"

=

Gfl;2 as the funetion of the time (,) of experiment.

J. VARGA and G. SEBESTYÉN

G.70-3 G-6(r)

[mg]77--- A·48x200mm2 d·48mm lJ i_ l!;72mls -

1 )..,=3 9 0h_q=45)J-

oh_g-37)J-

e---- .

he;-72^jL-loi' ./-._--;

7

4 - 3

21-·--1-··· II-Ik-·-··· ; _~

72 16

3/Erasiona1 weight 10ss (G) yel'SUS the dura- tian (T) of experiment.

Perte de poids par érosion (Gj en l'onc- tion de la dllrée (T)

de l'expérience.

T[h]

Table that with experiments carried out at a cons- tant cavity length

Clo =

3d),

G"l'it= 290mg on the average with a lead sheet of 3lnm,

G"l'it= 420 mg on the average \Vith a lead sheet of 8 mm.

'1'0 explain this phenomenon requires further experiments to be carried out.

Apart l'rom the data given in the Table, it eould he ascertained in a large number of tests that the cri- tical quantity of Jnaterial may be regarded as indc- pendent of the Jlow velocity under constant cavita- tion conditions, i.e. constant cavity length. The critical material quantity depends only on geome- trical dimensions such as model dialneter, test sec- tion dimensions and cavity length, which appear to represent a unique function of cavitation num- ber [l2J. This was verified by experiments condueted \Vith lead and aluminium test specimens.

It may be noted that em'lier the authors ^1:11 J had found in their experiments condueted in order to determine the velocity-damage exponent, that the exponent value was shown to be less than 5 and of variable value in the period of incubation and to reach the constant value 5 only hy the end of criti- cal period.

cimens. Itmay be noted here that the experimcnts on the total destruction (accumulation) period have also been made "\vith the ahove test specimens of two different thiclmesses and the same resuHs have been obtained. On the other hand, oddly enough, the critical quanti tics of material pertaining to the critical times differ with the thickness of the test specimen (see Table 1). It is evident l'rom the 914

Influence of surface roughness on the incubation period

The literature of cavitation research ineludes only a small numher of pa pers dealing with inves- tigations on the relationships hetween the surface roughness of hydraulic machines and cavitation erosion. In this respect, the expcriments by Shalnev [1:-3J and, more recently, Numachi et al. [14J deserve mentioning.

In our experimental studies, the objective was to determine the inJluence of the roughness of test specimen on the intensity of cavitational erosion.

The tests "\vere carried out by the methods dis- cussed in the Introduction. Also, the quality of test specimen was the same. The surfaces, initially sm.ooth, have been made rough by rolling a weight- loaded screwed spindle over them. Then the pre- cise value of roughness \Vas measured using a pro- filometer-type instrument. The grooves were set at right angles to the direction of 110w. The value of roughness was expressed in conventional n11S

tenns. The roughness values of the th l'ce test spe- cimens were 111/

= ]

2IJ., :37 IJ. and 45 IJ..

The weight losses of test specimens of tluee different roughness values are shown (versus the time of experiment) in Figure 8. The graphs will be made clearer by plotting those weight-Ioss curves in a logarithmic co-ordinate system (Fig. 9). From the Figures, and l'rom comparisons with experi- ments involving smooth-surface test specimens, the following conclusions can be drawn:

1. Roughening the specimen surface shifts the location of the critical point, its value decreasing with roughness (under the same flow conditions) -- Le. the critical period becomes shorter;

2 3

91

Erosional weight 105s (G) versus the duration (T) of experiment in 10garith- mie co-ordinate system.

s par ero- onction de de l'expé- oorclonnées Perte de poid

G[mg] sioIl (G) en l'

GaG (T) la durée (Tj

A=1r8 x 200mm d=48mm rienee, en c

).=3 v..= 12m/s loua ri/hm iqIleS

ohq-=J,.5p l,- Clhq-=37;.L ₁

r--

=

<!>hCj=f2;.t ^l'_i i

! 't,.>

1

1 J

1

1

1

^{1 1!}

1

LJ ¹

Il

3 ₁

ri

1---

f-+ ,_.-

.'-- 1---- 1-'-

1

1

J- -1

1 1

.

. /

"'~ T-t-

^{c -f--}

. / V

....

'"

frh]

10 98 7 6 5 ft 3

2 3 . . 5 6 789lOf 2

2 10 9 8 7 6 5 ft 3

specimen.

Photo(Jraphie de la surface d'érosion SUI' Ull échantillon lisse.

specimen.

PllOto(Jrllphie de la s1/rfllce d'érosion S1/r lin échantillon lisse.

12/Photograph of eroded surface on a smooth test specimen.

Photo(Jraphie de la sllrfllce d'érosion sllr un échantillon lisse.

13! Photograph of cavitational caverns on a grooved test specimen.

Photo(Jraphie de cavités de cavitation SUl' lin écIHln- tillon rainul'é.

14/Photograph of cavitational caverns on a grooved test specimen.

PllOto(Jl'(lphie de cavités de cavitation SUI' ^Ill! éclu1Jl- iillon rainuré.

15/ Photograph of cavitation al caverns on a grooved test specimen.

PllOto(Jraphie de cavités de cavitation sur llIl écIHln- iillon rainuré.

915

J. VARGA and G. SEBESTYÉN

tational erosion is increased by additional efl'eets coming from the grooved surface. Figures 10, 11 and 12 illustrate the cavitation damage of smooth test specimens; Figures 13, 14 and 15 ill ustrate the eroded surfaces of grooved test specimens. It is evident from the latter figures that the pits deveJop primarily in the grooves.

The variation of incubation period as the function of surface roughness is shown in Figure Hî. The Figure suggests the conclusion that, und el' the conditions of our experiments, roughness up to

11'1

=

:35IJ. has no appreciabJe influence on the length of incubation period. At a roughness higher than that, the incubation periocl tends to shorten rapidly followed hy total rlestruction.

T,n' • Teril (hl{)

il A ='t8X200mm²

v=12mjs ,H

1 1

- " - -

~!

1

\

¹

~."J-_

1

\

2 8

10 20 30 50

161 ^{Variation of incubation period ('nit) with surface}

roughness (h'1),

1Tariation de !a période «d'incllbation » ('nit) en fonc- tion de !a rlluosilé sllperficielle (h,).

List of references

[lJ VAHGA (J.), SEIlESTYl~N (Gy.), SCHALNEW (K.K.) und TSCHEHNAWSKI.J (B. A.). - Untersuchung des Mass- stabel1'ektes der Kavitationserosion. Acta Technica ,1c.

Sci. HunfJ., 51 (1H(i5), No. a-4.

Table 1

[

"

NI'~lllEH

OF TEST Voo Tcrit G('rit a h

'1 SPECDIEN

(m/s) (h) (mg) (mm) _(ft)

- - -

- - - -

200 13',60 4,5 340 :3

<

¹

201 14,43 3,2 275 3

<

¹

204 13,95 Ll,2 305 3

<

¹

205 10,40 14,5 270 3

<

¹

20(i 13,05 5,7 265 3

<

¹

207 9,05 28,0 290 :3

<

¹

average 291

:300 12,00 7,8 320 8

<

¹

301 14,35 3,2 43'0 8

<

¹

303 12,74 7,0 480 8

<

¹

:-;'() 4 12,00 8,4 450 8

<

¹

average ,120

:ml 12,00 7,:3 430 8 37

311 12,00 8,4 450 8 12

312 12,00 4,0 53 8 45

2. In the period of total destruction following the incubation period, the curves are practically parallel to one another. Hence, the conclusion can he drawn that at this stage the surface roughness has no longer influence on weight losses due to erosion.

The ahove statements are confinned bv some additional observations. During the expe;'iments, it was found that the erosion pits OCCUl' invariably along the grooves. The rougher the surface is, the sooner will the pits appear. This phenomenon may he conceived in such a way that the efIect of cavi- 916

[2J VAHGA (.L) und SEIlESTVÉN (Gy.). - Beitriige zur Hydro- dynamik der Kavitationserosion und des Skalenef- feUes. Mill. der [(onl- fiir Wasser[;raftmasch. Timi- soara (1 %4).

[:l] VAHGA (J.) and SEllESTYI\N (Gy.). - Determination of the Frequeneies of \Vakes Shedding l'rom CirculaI' Cylin- ders. Acta Technica Ac. Sei. Hllnu., 5:3 (1\J(j(i) , No. 1-2.

[4J Gn.LIDIOT (L.) und SINAY (G.). -- Die Brucharbeit ais

\Vel'kstoffkenngriisse. Acta Technica 11c. Sci. llUIlU., XXII. (1()58), Fase. 1-2.

[5] VASV.Ùll (1".). - Ujabb adatok a fémek kavitàciôs ron- esolàsi elméletéhez. (New data relative to the theorv of cavitation destruetion of metals.) Thesis for. obtain'- ing the degree of Candidate of Teehn. Sc. Budapest (1%1).

[(i1K03blpeB (c.ru. - fUJl:poa6pa3IlBllblH 1l3HOC Me1'l.'l.'lOB llpll KaBIlTaIUIU. MOCKBa (964).

[7] Gora'leB (11.Ii.) H MHlIL( (P. 11.). - nOBblweHHe I\aBHT,ll(- lIOIIHO-3p03HOIILIOH CI'OHKOCTII JreTaJ1eH MaIUHa. MOCKBa

(964) .

[8] LElTlI (\\'. C.). - Cavitation Damage of "[etais. The EIlyilleerillY .TourIla! (Canada), vol. 42-1 0()5H).

[H] GOVINDA Iho (N. S.) and THlHUVENGADAM (A.). - Predic- tion of Cavitation Damage. .Tourn. of the HUdr. Div., l'roc. of the A.S.C.E., HY 5 (1%1).

[10] THlHUVENGADAM (A.). - A Unilied TheOl'V of Cavitation Damage. l'raIlS. 0/ A.S.M.E . .TollrIl. of Iiasic EllY. (sept.

1(Hia).

[11J VAl\GA (.L) and SEIlESTY1\N (Gv.). -- Observations on Cavitation Velocity-Damage hxponent in a Flowing System. Periodica Po!ytecllIl., 8 (1HM), No.

a.

[12J SILIlElOlANN CE.) and SONG (C. S.). - lnstabilitv of Ventilated Cavities. .Tourll. of Shlp. Res. (June, Ù)(il).

[lBJ SCHALN.JEW (K.IU. --- Kavitation bei Umstriimen einer Fliiche mit Hauheiten dreieckigen Umrisses. Mill. der

[(0111. /iir Wasserhra/'tmasch, Timisoara (1%,1).

[14J NUIAcm (l'.), OIlA (H.) and CHlDA (.L). Effeet of Sur- face Boughness on Cavitation performance of Hydro- foils, Hep. 1. l'raIlS. of the A.S.JI.H., vol. 87 (l!J(i5), sel'. D, No. 2.

(French summaryonpage 910.)