NASA Techaical Manordum 83352

A Comprehensive Analysis of Cavitation and Liquid Impingement Erosion Data

(UASA-la-83352) A CCflPREBIIEIFE PISPLI5IS C F R E Z - Z B C 3 0 C A V f l A ' I I C R ABD L I Q U I C I C E I E G E P E b l EICSICI LATA (BASA) 18 p RC BOZ/!!E A C 1 CSCL 1 l F

Uncl a E 63/26 C 7 t t l

'--

P. Veerabhadra Rao and S. G. Young Lewis Reswch Center Cl&and, Ohio

Prepared for the Second Conference on Cavitation sponsored by the Institution of Mechanical Eagineers Edinburgh, Scotland, September 6-8, 1983

https://ntrs.nasa.gov/search.jsp?R=19830016363 2020-04-12T11:38:26+00:00Z

r. l E E W n 4 D R A RIU) and 5 G YOUNG Nat ional Aeronautics and Space Administration Lewis Rerearcn Cecter Cleveland. Ohio 44135 A Comprehensive Analysis of Cavitation and Liquid Imping-nt Erosion Data

oRmNAL PAGE t3 OF Pocrrr mm

P. Veerabhadra Rao bid 5. 6. Young

Rational Aeronautics and Space ministration Lewis Research Center Cleveland. Ohio 44135

SYNOPSIS Cavitation-erosion experimental Cata previously repwted by the present authors covering several niaterials tested in a rotating disk device and a magnetostriction apparatus have been ana- lyzed ~sirg normalizaiion and curve-fitting techniques. Fro this Drocess a universal awroach is derived which cen include data frole cavitation and liquid il~pingement studies for specific materials from different test devices.

One of the primary objr:tives a* erosion research has been to w e l laboratory ercsion data to freld conditions with a r e confidence ar.C reiia- b i l ity. Honegger (1). as eariy as 1927, consid- ered 'spec i f ic erosion' in an atterpt to colapare materials and tiwe effects. doever. systematic investigations pertaininn to time effects on erosicn rates were conducted in the mid-1960's ( 2 :o 5). In view of the strong 6-Wndence of the erosion rate on exposure tiae in both cavita- :ion and liquid iwingenent envirmnts. several formulations. mdrls. nomograms, and charts were presented by Citferent investigators ( 4 to 14). The main purpose of these formulatioris was to

\ a ) Iaentify the dmage as well as eros'on mechanisms involved dut ing the erosiqn process with time;

( b ) Characterize and quantify. as precisely as possible. the erosion rate as the exposure tiae increases for long terns;

( c ) Test less resistant materiais in the isboratory for relatively shor: times and extrapolate these data to resistant materials in the field.

In view of the difficulties encountered in the past to characterize and Rodel matorials with d if fcrent devices and laboratory conditions. many investigators agree that coclparisats of test re- sults should be done only if based on the corte- spondir.j stages of :he erosionrate-versus-:isle curves. Specif ica1.y. these stages have been named the incubation period. the acceleration period (accumlation zone). the peak damage rate, the deceleration period (attenuation zone), the steady-state region, and the 1-ng-term erosion period, which is characterize< as either cyclic. decreasing. or increasing. depending On the test method and the erosion resistmce of the material (15). Typical erosion-rate-versus-tlme curves are reproduced in Fig. 1 depicting all periods (2 to 5).

A historical background of work on long-term cavitation erosion prediction and aethods for .mdeling the erosion-rate-versus-time curves are presented in a recent study by the authors (16). Several prediction equations for 1 iquid ilpinge- ment erosion are presented in (15). The impor-

t a t models. forulations, w r m s . and the variables necessary to evaluate the erosion- versus-time curve or erosion in a certain amount of exposure time are presented in T a l e 1. No single .ode1 or prediction attemct has yet been fully precise in.its ability to predict erosion rates either during the initial phases of drage or during the advanced stages cf erosion. Hence. icag-term predictions using earlier formulations differed from actual data by factors of t m or nore. Host prolonged operations of machines re- quire a higher confidence level to operate ma- chinery at o p t i u efficiency. Thus if a method is devised to accurately predict the long-term erosion of a baseline material, and it is found that the prediited erosion would be detrimental. either t'le material M v be changed at the design stage or more accurate overhaul prriods may be established.

A method for erosion-rate-data curve fitting is presented as normal ired c w l a t ive average erosion rate as a function of nomallzed time. This aelhod greatly reduces individual variations of the instantaneous-erosion-rate-versus-tire curves. In this manner, a universal approach to the analysis of data from previous experimental results is presented for prediction purposes. The long-term exposure behavior is discussed and correction factors wtaining to the incubation period are describel. This paper is a condensed version of (16).

2 NOTATION

ER erosfon rate

)rER mariwra erosion rate

P pressure t exposure ti* to cavitation or im-

pingement erasion

ta*tb* incubation periods of curves ir.

tc ,... tn A.8.C ,... W in Fig. 9

i incubation period of a typical erosiorcrate-versus-time curve

P VEERABHADRA RAO 8 2 3 t YOUNG

time t o a t t a i n maxi- w peak r a t e

of erosion on rate-versus-time

curves

ve loc i t y

c w l a t i v e vol lae loss due t o cavi-

t a t i o n erosion corresponding t o

t hours exposwe maxi- c u u l a t i v e voltme loss due

t o r a v l t a t i o n wcs ion corresponb

ing t o the slope o f the erosion-

versus-time curve jo in ing the

o r i g i n and the point of tangency

~t incremental time causing an incre-

mental v o l e loss b v

AV incremental v o l u e loss o f material

i n incremental time dt

Subscripts:

curu 1 a t i ve average

instantaneous

3 3ATA ANALYSIS AND PROCEmE

3.1 Erosion Data Sources

i n the development of t h i s curve- f i t t ing approach fo r long-tem cavi ta t ion erosion-rate prediction. orlgina! data sets obtained independently by each of the present authors were used. One used a ro- ta t lng disk device (17) and the other z magneto- s t r i c t i o n apparatus (18 t o 20). The de ta i l s o f the ro ta t ing disk device and magnetostriction a* paratus have been described i n d e t a i l i n (19. 2i. and 22).

The experimental conditions f o r the r o t a t i n g disk device were velocity. 35 t o 37.3 als; pressure. 0.11 t o 0.17 MPa (abs); d i e t e r of the cav i ta t ion inducer. 25.4 n; and t e s t l iquie. water. The materials tested were a l l r i n u , cop- per. brass I, brass 11, stainless steel. and mi ld steel. ?he coqos i t i ons o f materials and t b e i r properties were reported i n (16. 21. and 22). The experimental conditions pertaining t o the wgnetos t r i c t ion apparatus were frequency. 25 kHzi a q l i t u d e . 44 p.; tes t l iquids. sodim {from 204 t o 649. C) . and water. The materials tested were nickel, a l u i n u , zinc. iron. i-605 cobalt- base alloy, S te l l i t e . and stainless steel; the compositions o f materials and t h e i r mechanical properties were previously reported (18 t o 20).

3.2 Erosion Oata Treatment Wthod

Figure 2 presents cumlativesrosion-versus- erposure-time curves for stainless s tee l tested I n a ro ta t ing disk device a t four d i f f e r e n t velo- c i t 1es (17). Figure 3(a) presents instantaneous- erosion-ratc-versus-time curves for the same material (sec upper curve i n Fig. 2). As erosion resistance increases the incubation period be- comes -re pronounced. Because there are several peaks c l d va: leys i n the erosion-rate-versus-t ime curves. the predic t ion o f erosion r a t e w i th expo- sure t i n e becomes increasing';/ d i f f i c u l t .

A s a f i r s t step t o iaprove the si tuat ion, the cumulative average erosion r a t e i s calculated and

p lo t ted versus time (Fig. 3(b)) f o r the same data presented i n Fig. 2. The osc i l l a t ions observed i n Fig. 3(1) are considerdbly smother i n Fig. 3(b) because o f t h i s treatment. It i s now ev i - dent tha t r o a t e r i a l responds t o erosion i n a s im i la r ranner a t d i f fe ren t ve loc i t i es and each erosion-rate-vefsus-time curve has a maximum ero- sion r a t e i f the tes t has been run f o r a suf f i - c i e n t length o f time.

Using these s i n i l a r i t y pr inciples. each data point o f Figs. 3(a) and (b) r - s normalized w i th r-t t c $ e l erosion r a t e and the time cwre- spomQing t o t h i s p e l . Figures 4(a) and (b) present norralized-instantaneous-erosion-rate- versus-narrralized-ti* and nwaal ized-cuulat ive- avermrosion-rate-vwsus-rw~lral i red-t ime. re- spectively. However. the scatter i n Fig. 4(a) i s too great t o provide an accurate curve o r predic- t i v e equation f o r the f i e l d engineer. Theoreti- ca l and o q i r i c a l .oGels proposed by e a r l i e r in- vestigators (Table 1) do not f i t these p lo ts un- less many assupt ions are made; fu r the ra re . i he scatter bands are large (16). On the other hand, Fig. 4(b) provides a smooth curve without o s r i ll- ations. ind icat ing that a properly n o r r a l i z w e r o s i m ra te follows a ce r ta in natural trend even under d i f fe ren t experimental conditions.

F i g w e 5 presents t yp ica l normalized cumulative erosion r a t e versus normalized t i n e for d i f f e r e n t aa te r ia l s tested i n a ro ta t ing disk device and a magnetostriction apparatus using both v ib ra t ing and stat ionary specimens. A c&arison of nomalized-instantaneous-erosion-rate-versus- normal ized- t ie curves fo r the sane aa te r ia l s (16) indicated that there i s too much scatter and most o f the ind iv idual materials cannot be repre- sented by any s ing le f o r w l a t i o n . Figure 5. on the other hand, shows a considerable reduction i v scat tw. This consistent conf igurat ion was ob- served not only f o r materials tested i n a rota- t i n g disk device w i th water, but also f a r a var ie ty o f materials tested w i th a magnetostric- t i o n device using both water and l i q u i d sodiuca.

The r a t i o o f instantaneouslcunulative peak heights MRi/M€R, varied from 1.0 t o 3.87 f o r aa te r ia l s tested i n the ro ta t ing disk device and from 1.0 t o 1.71 f o r aa te r ia l s tested w i th the Mqbetos t r i c t ion apparatus using water and l i q u i d s o d i u (16). As would be expected. the times t o a t t a i n llaxirm cumulative erosion r a t e (tag) were always longer than the times t o a t t a i n aaxiarca instantaneous erosion r a t e (b i ) . Furthennore, f o r the materials tested. the r a t i o t a i l t a b varied from 0.42 t o 0.960 i n the ro ta t ing disk device and from 0.3 t o 0.88 i n the aagnetostric- t i o n apparatus (16). As the erosion resistance o f the material increased. the r a t i o HERi/#Ra decreased. Mo c lear-cut trend for the r a t i o tm. tm was observed.

"Th: advantages of n o n n a ~ i z e c ~ - ~ ~ ~ ~ a t i v e - eros i on-rate-versus-normal i red-t irae p l t t s are (a) scatter o f the instantaneous-erosion-rate- versus-time curves i s g rea t l y reduced. resu l t i ng i n a consistent. r e l a t i v e l y smooth set o f curves; and (b) the HERa and tma can be evaluated frm the erosion-versus-time curve.

4.1 Comparisons w i th Ear l i e r Invest igat ions

Oata reported by Kerr (23). Thornas and Brunton (24) , and E l l i o t t , e t a l . (25) were analyzed i n the same manner as the present invest igat ions

(16); typical p lots are presented i n Fig. 6. The improveaemt of using c w l a t l v e erosion ra te was clear for data reported i n (23. 24). This data treataent further supports the view that the normal ized c w l ativestwion-rak-versus-tie curves have s i g n i f i c n t advantages f o r erosion prediction with reduced data scatter. It n s noted from figs. 5 and 6, the quantitat ive data i n (161, and the typical data i n Table 2 tbat brass Md stainless steel tested a t dif ferent ex- perimental conditions agreed v e ~ y well on the normalized average basis. Results i n Table 2 In- dicate that MR-I#Ra varied from 1.4 t o 3.2 md tniltma from 0.88 t o 0.82. One may therefore in- fer that quantitat ive correlations exist betwen cavitat ion and l iqu id iwingement irrespective of the W ? device used t o produce erosion.

4.2 Effect of T i w e Increments on Prediction -1s

'How many expwiwntal points are necessary?' and 'Uhat time intervals should be used t o obtal?r the most precise predictions possible?' are fre- quently asked questions. To investigate the ef- fect of the interval length on the accuracy o f the f i na l plots. Fig. 7 was plotted using l-hr intervals for cavitat ion data o f t&e cobalt-base a1 loy L-605 i n l iqu id sodim a t 427. C. The s a e data with time intervals of 5. 10. and 15 min are presented i n Fig. Sib). With fewer points the determination of erosion rate and tr i s affected as shown i n Taicle 3.

W o r differences can be noted by c a p w i n g the two sets of data. The paraeters calculated a t -in intervals are far less accurate than those calculated a t 5. lo-. and 1 M i n inter- vals. Errors o f 50 t o 300 petrent were observed i n determining the parameters HERa and tm As the erosion resistance decreased, the e m 0n- creased with long interval expe r im ts (Table 3). figures S(b) and 7 indicate, however. that these close-interval data need be collected only u n t l l an accurate peak i s attained. Since HER, wd tm are the crucial p a r m t e r s which are used t o cafculate the requisite quantities. errors in- volved i n the i r determination w i l l lead t o great- er ir.accuracies nhen they are used for long-term predictions. This study points out the llpor- tance o f using close intervals fn the early sta- ges of erosion (up t o the peak rate of erosion) t o arr ive a t precise parmeters f o r prediction purposes.

4.3 Effect of Long Exposure- on Erosion Rates The long-term erosion rate has been controversial ever since investigators have been mare of the influence of test time on eroslon rate. Same in- vestigators have reported continuous decrease i n erosion rate after the i n i t i a l peak rate. sare have reported constant f i n a l rates (steady state), while others report cycl ic rates a t even longer exposures. A l l o f these patterns have been well documented i n (2 t o 7) and i n various papers presented a t the ASTn syu@osia (26 t o 29). Ccnrl ative-erosion-rate-versus-time curves are presented i n Fig. 8 i n the nomalized forn. which generally shows a decreasing trend irrespective o f erosion resistance and material (17 t o 19).

The long-term exposwe plots presented i n Fig. 8 are unique, as the r a t i o tlm, ap- proached nearly 400 (the highest r a t i o believed t o be observed t o date). Some of the deviations from a smooth curve i n these plots are believed t o result from the small number o f data points

taken &in9 the incubatfon m d rccelcration periods. As explained 11, the previous section, d i f f i c u l t i e s i n obtaining the t rue values of IER, and ud the sensi t iv i ty o f these parueters t o good long-term predictions are par- t l a l l y responsible for the d i f f i c u l t y i n obtain- ing a single plot.

4.4 Incubation Period Cor rx t i on

A t rp i ca l set of most corrrmly observed cuu la- t ivcsrosio#uate-versus-t 1?1e curves i s schemati - c a l l y represented i n Fig. S(a). The incubation periods are indicated i n the f igure as ta. tb. mi! t . By subtracting these times f raa the t ime sf each experimental point on each of the respective curves a condensed set of plots i s generated (Fig. 9lb)). The new normalized time for peak erosion rate i s nar calculated as (t - t )It&, - ti).

AI! plots of normalized c l ru la t ive erosion ra te and normalized time use the relationships (vlt)l(v.ltm) and tltm, respectively. A correc- t l on factor f o r Incubation period as used i n Fig. 9(b) i s necessary f o r a l l the previous figurer; t h i s correction factor would sh i f t the curves t ~ l a r d the y-aris by the a ~ w t of time equiva- l en t t o the incubation period. I n making the t ransi t ion from the lode1 t o prototype the incu- bation period f o r the prototype re la t ive t o the model should be knam i n order to make th is cor- rect ion for rw'e accurate long-term predictions.

Recently H e w n (30). while analyzing the ASTn 6-2 spansored 'round robin' test progrdp. found that coqarisons and correlatiocs with the ( r a x i u ) cunulative erosion ra te gave w e scatter nd Inconsistency than those using maxi- u instantaneous erosion rate. This i s primar- i l y because the incubation period i s dependent on the average erosion rate, and because there i s mure scatter i n incubation perloas than i n w iu erosion rates (30) . A possible disad- vantage of the incubation period correction sug- gested i s that i t may soaetiaes terove one of the fundmental definit ion of the avera rosion- ra te approach. Sa investigators c.. 31 and 32) used incubation period t o predict erosion rates. These correlations are better than mate- r i a l property correlations w i t h erosion rates (32). (Neither o f these investigators (31. 32) considered long-term erosion-rate predict ions. )

4.5 Erosion Resistance Variation Labaatory and f i e l d devices produce uneven ero- sion over the test speciaen; hence, calculations for eroston resistance are very general and vary considerably even within the same device o r test.

The normalized c u r l a t i v e erosim-rate- versus-t ime curves i n the present i nvest lgat ions, though gemrally smooth. i ndicate s ~ p deviations wi th erosion intensity. As the erosion resis- tance eecreased. the portions o f the curves fol- lowing the peas attained a lower value at long test times. For a s i n ~ l e test device these por- tions o f the curves were lower a t long times fo r more resistant materials, But the height of each curve a t longer test times appears t o be a func- t i o n o f both the device and the material. Thls correlat ion between the level of the long-term- erosion-rate-versus-time curve and erosion resis- tance may be helpful i n applying th is universal p l o t approach t o data from both laboratory and f i e l d devices. An erp i r ica l factor called 'the apparatus severity factor* i s described i n (15) f o r )!quid i~pingement data. This m y sene as an a1 ternate approach t o th is discuss ion.

P VEERABHADRA RAO and S 6 YOUR6 0RK;INAL PAGE 13 OF POOR QUAUTV

4.6 Universal Approach Plots Slrmary plots o f nomallzed erosion ra te versus normalized time are ented i n Fig. 10 for the prev ;ous e x p e r i m t a E a t a f w brass. stainless steel. mild steel, W cobalt a l loy 1-605. Every material tested i n any type o f c r v i t r t i on or l iq- u id ilqln-t device can be represented i n t h i s manner. Depending upon the test device and ute- r i a l tested, a mean curve may be chosen scat- te r bands can be defined or derived. The accu- racy of the derivation o f the hrq p a r a ~ t e r s ma and HER, (including the incubation period) con- tr ibutes to the accuracy of the prediction. The deviation i s greater i n the norrslized tie re- gion from 0 t o 1 than i t i s i n any other port ion o f the curve.

Although the plots o f Fig. 10 are similar to the set of p lots reported by rnothe~ investigator (7). Fig. 10 i s based on experlaental data w i t b - w t any assllptions or direct re lat lon to theary. The plots i n (7) were generated using instantan- eous erosion rate versus tie bile the pment curves of Fio. 10 were developed using cuaulative erosion ra te versus time. Equations proposed i n (9) use both tangent points and f i xed average- depth-of -erosion values coabined wi th a curve-f i t approach which resul ts i n auch wider variations than i n the current study.

The concept o f a normal ired-cuaulativc-ero- sion-rate-verrus-~~calized-tire curve was f i r s t suggested by Heyunn (9) and used by one o f the present authors (17. 22) t o check the va l id i ty of the erosion theory proposed by Thiruvengadm (7) for a rotat ing disk device. By using c u r l a - t i v e erosion rate in;tead of instantaneous ero- sion rate the data scatter was considerably re- duced. and the plots (22) were closer t o the the- oretical curves presented i n (7). The use o f cumulative erosion rate was also considered by Lichtarowicz (12). RoMl ized- instar . t~e~us- erosion-rate-versus-normalized-time curves have also been presented fo r erosion-corrosion model- ing using a magnetostriction apparatus (33) and for steam turbine blade and shield materials using four dif ferent iqingCPent devices (25). L ich tara icz (12) also Suggested that only two paraeters HER, and ta, may be used t o pre- d ic t c w l a t i v e erosion rates fo r a l l r i n l a (Table 1). However. t h i s paper shows the irpor- tance of two additional parmeters, the incuba- t ion period ti and the erosion resistance (IIER).

The erosion process due t o cavitat ion and l i qu id iqhgeaent i s believed t o be a function of the ear l ier history o f the eroded surface (in- cluding work hardening, surface stresses. ad changes i n material properties). Also. a study of the relationship between the surface roughness and the erosion rate would be helpful to gain ad- d i t ional insight into the erosfon process a t 1 onger times.

4.7 Application of the Universal Curve F i t Approach

To check the advantage o f the analysis propored i n th is paper. data reported i n (25) f o r stain- less steel during l i qu id drop impingement a t a velocity o f 305 t o 314 mls were analyzed. O f the three devices used, the English Electr ic Cam- pany (EEC) data have been taken as standard md normallzed time as 2. Table 4 presents tha pa- rameters o f HER$, hi, CPR , ma, and percentage error measured while analyzfng Instantaneous and

c w l r t l v e erosion rate. I n most situations. c ~ l a t i v e erosion ra te provides r better p d i c - t i a n than instantaneous erosion rate. It i s theG pors'ble t o calculate the theoretical cuu la t i ve erosion o f a specimen a t any point i n time. Tnis infomation i s useful i n determining the extent of a colponent's erosion and i t s remaining useful l i fe.

Data f o r a large nuber of materials tested i n both a rotat ing disk device and a magnetostric- t i on osc i l la to r have been analyzed i n a manner h i c h brings the results of the two methods closer ro a universal curve f ~ t .

Normalized c w l a t i v e erosion rate has been plot ted versus normalized time and a curve f i t i s p q m e d rh ich covers a caprehensive variety of materials. tes t conditions, and amices. C u u l a - t i v e erosion ra te nd time are nar.alized t o the peak erosion ra te and time t o peak erosion rate. respectively. Mjustaents are suggested f o r in- cubation periods.

It was shan that the universal approach p lo t i s we accurate I f small t i ae intervals are used before the peak d a g e ra te i s reached.

After the peak damage rate I s passed. a t long exposure times. more resistant materials show a lower nomai f zed average eros ion rate.

The curves and data scatter bands derived from th i s universal curve-fit approach appear t o be useful i n correlat ing di f ferent types of lab- oratory tests w i t h each other and w i t h f i e l d data.

REFERENCES

(1) WYIE66ER. E. Tests on erosion caused by jets. Brown Boveri Review. 1927. g. 96104.

(2) THIRlMWGAMC(, A. and PREIXR. H. S. On testing materials for cavitat ion d a a e re- sistmce. Journal of Ship Research, 4964. 8. 39-56.

(3) HESST.-W. 5. and OEVIWE. R. E. Effect of - expos& tie on cavitat ion damg& &r-

nal of Basic Engineering, 1966. 88, 6 m 5 . (4) m. F. J. On the t i ae depkSaence of -

the ra te Of erosion due t o i a i n o m m t or - 7 a--- - - cavitation. Erosion b i cavitat ion or i t

can e y or W L Z e r T i s % L e E ; a f i L 7 . (5) TICKER, J. u. and rn 6EE. A. Y. J: Time

depetknce of cavitation erosion and the ef- fect o f same material properties. Proceed- ings 3rd International Conference on- Erosron soc a e m a , o . . 5%: Aiza: Esi 2 1 i % e n t , ~a r :bk~gh .

THIRUMIWMNl, A. Handbook of Cavitation W Hydronautics, IK. Tech. Report t 3 X * ~ a u r e l . Hdrrland. 1s5.

-- ~ -..- ceedings-2nd kersbwg-conference on ~ a l i i h ~ l 0 n and iTT ied Phenomena, Vol. 2. ~ o y a l A f x l gh. 1967. HOFF, 6. and UN6BEII. 6. Resistance of na- t e r i r l s towards various types o f aechklcal s t W s . Proceedings 2nd Wrsbu rg Confer- ence on R w vnenooena. Ro a1 Aircraf t Establ ishaient. Farnborough. 19b.

P VEERABWA RAO and S 6 rOllWG

(9) HmAW(, F. J. Tarard quantitative pnd lc - (n) LAKSmrm\ RAD, N. S.. VEERABWUYU\ RAD. P. t ion o f l i qu id iapact erosion. Characteri- uld RU). 8. C. A study o f cavita- zation and Determination of Eros'1on Resfs- t i on erosion i n rotat ing caponents. Pro- B. can WKlety t o r lest ing ceedings o f 5th Confemue on Flu id -3- :nd U a & ? ~ : ~ * P ~ & + l p t ~ i a . 1970. p&,;-lllj5~adealaf Kf*. Budest .

(LO) EWL. 0. 6. A fiat a p p f ~ c h t o a .i~rO- scopic model o f erosion rate i n dror, t-ct (23) KERR. 5. 1. Determination o f the re la t ive

(11) PEREL)WI, R. ~ . - & d OENISOV, V. O. The strength o f materials under the action o f droplet impacts. Royal Aircraf t Establish- ment Library Translation No. 1651. 1971.

(12) LICHTAROUICZ, A. Cavitating j e t apparatus for cavitation erosion testing. Erosion: Prevention and Useful @plfcations. STP 664. k r i c a n Society f o r Testing and tbter ia ls. Philadelphia. 1979.

[ I31 r(O%IEVIC. J. Dvanarics o f the cavitat ion - . d y g e . Proceedings, Jo i Design and Operation o't F Pol. 2. Colorado State Un tions. Fort Collins. Colo.. 1978;

(14) scm~fr. G. F. t iqu id m i - s o l i d par t ic le iapact erosion. Year Control Handbook. A M . New York. 19210.

(IS) A S T M - ~ o a i t t e e - ~ ~ Uorking Docuent. Stand- ard practice fo r l iqu id iqingement erosion testing. Draft no. 6. Mar. 1982.

(16) VEERAflHAORA RAO. P. and WllM16, 5. G. Uni- versal moroach to analvsis o f cavitat ion and I i q u ~ d - i ~ i n ~ e e e n t bos ion dab. NASA TP-2061. 1982.

(17) VEERAW~DUA RM, P. Characteristics. cor- relations, s imi la r i t ies Md prediction of erosion due to cavitat ion and l i qu id im- p -*lent. Ph.0 Thesis. Indian Ins t i tu te of Science. Bangalwe. India. 1975.

(18) YOUNG. 5. 6. Cavitation damage o f stainless steel. nickel and a l u i n u al loy i n water for ASTM round robin tests. NASA TM X-1670. 1968.

(19) YOUNC. 5. G. and JOHISTOII. J. R. Ef fect o f t e -e ra tu re and pressure om cavitat ion d 4 -

i n sodium. characterization and Deter- mination af Erosion Resistance, - Aaerican k c i e t y f o r Testing and Materials, Philadelphia, 1970.

(20) YOUNG. 5. G. Study o f cavitat ion damage t o high pur i ty metals dnd a nickel-base super- a l loy i n water. NASA TN 0-6014. 1970.

(21) V E E R A B W A RAO. P., S V M A RM, B. C. and LAKSmAllA RW), 1. 5. Erosion and cavi ty characteristics i n rotat ing corgonents. Jarrnal of Testing an@ Evaluation. 1980. 8. -

resistmce to cavitat ion e m t o n by the v i - bratory mthod. Transactions o f A X . 1937. 59. 37S397.

(24) m. 6. P. and BRUWTOIY. J. H. Drop i, t erosion o f metals. Proceedin s

=society, L- %lie&

(2s) mion. Ti.. w I o n . J. 8. U* wm. A. Caparison of erosion resistance of standard s t e m turbine blade and shield ma- te r ia ls on four test rigs. Characterization and Determination @f Erosion Resistance,

ICM e y or Jesting and ~ ~ ~ i ~ ? h i l a d e ~ ~ a ~ l k 0 .

(26) Erosion by Cavitation and Iqingeaent. STP

(27) Characterization and Determination of Ero- sim ReSlstMce. STP 414, herscan Society for Testing and Materials, Philadelphia, 1970.

(28) THIRWllGI\Dkn. A. Erosion tiear and Inter- faces wi th Corrosion, -- ciety for Testing and Materials. Philadel- phia. 1974.

(29) ADLEU, Y. F. Erosion: Prevention and Use- p 1 ~ 1 ; c a t i o n s . 51' sTI. - l c n socjefy

or est ng and Haterials. Philadelphia. 1979.

(30) HYWM. F. J. Canclusions frae the ASTH

Scrfpta Pub1 ishing Company. Uashington. D. C.. 1976.

(32) W E E R A ~ I R A RAO. P., e t al. Estimation of cavitat ion erusion w i B Gcubation periods and material properties. Jarrnal of Testing and Evaluatlon, 1981. 9, 1

(33) -and THlRm=*A. Cavi- - ta t ion ekion-corrosion &el ing. ~ r o s i o n

Uear and Interfaces with Corrosion, STP56T; k i c m Society f o r Testing and 6 t e r i a l s . Philadelphia. 1974.

L I (el- l

( I ) k- ra te # W w S k (21 1 s l W r C6m (huceot c t* rl. frm straight-I* mcc af en-tr-1

( I ) C I I . t h e r r q P d s m i a ~ r Mmri.1 1-s a t

(21 -1 -t d%rZitr (1) lhlrc af 1*i. why* w Zit

CCI CQ -it t(r (4) -a l ta r . w a t u l enim

m i s t l c . wmw

(11 lclMC tie (2) Restst- .r-t cat to t i - crash a hparuyucc c w l t i . l r . n uc* k m V i ~ t i R . IL~I~~u

( I ) mtb ot c ras ia a t d i c b dfut af crater tN1i.. - -it-t

(4) P m p w t t u I i t p -tat. S . k t l l * r ) i - i n ~ ~ d w w s ~ ~ -14 be m m w r r ta t- -

1 a f ~ t a U r 8 w m r t ~ i n t 8 m l s u r r r s - m r t d

, (5) R4t.a Of +NO d O l W S i U i n fi-1 I s w p - s t a t e m r t d t a race a* cma 1 im f i r s t s teups t r te **a4

( I ) blrir rrr m . m rate I ( I ) Cnlwha d v l l r t r l (21 C m i u r l m s q d u t . r i a l

1 (4 friction d g t C 1 a l &lap e i a t l c w e b

(31 CmlW6.l rat8 C dnclom C l o Q H cntmtoo .tw

(1) M LIIU~W mrg. -la rate (2) T r u U t a h mab ar r1n t . r rm.)c

a+ha rats I

ORIGINAL PBSZ {(J

OF POOR QUALm

Table 2 M u i ~ r a t r of MDIIW and tbe IWU~S t o at ta tn ~t for VYIOUS u t n ~ a l s - 6ro@ 1-t

, 8 I IUB stainless / 2 . 8 ~ / .R 1 1-21 i 3.51 4.70 I .74 ' , steel I i I

[Ram slze. 1.5 r; n lac l t y . 125 .I=; data source. (24).]

Table 3 U.lY ra te of m4 t1.r lo at ta ln t t for L-605 cobalt al loy I n a m t o s t r l c t l m . ~ 9 u a t u s 11th -11 M Iry lnt.?rwrls of ti-

Mter~rl

Cobrlt

[ht. s w y . (19). test I~pu id . IIOUI~ s o 4 1 ~ at 427. C; q11:dc. 44.5 L; f m u m c y . 25 UI2; -*.en. .tbr.tlnq.]

Iluln ra te of rrnzm. TI- to a t t a ~ n m u r u ra te of woslon. 3!h aln

5 M LO =In 1

75 m 42-73 1 3 3 - a 31.0 22.5

-4 S.80 M 4b.U 4b.a 10 15

U t n a u caIcu lat im ti- ~ o s s ~ b l e l h ~ n calculated by thrs ~ c b c d .

M u i u rate of '

emsim. d ~ q x t -

*it - 'he tnql ish i l a t r r c C a r n y r t g . btlP - T h e C. 1. Pwsms' erosion r i o .

- Tlu mitr erosion t rq? r t p of thc Central i l e c t r t c ~ t y bcnerat~nq Board ( C k G B ) .

Lnstutmeaus. 1

0.93.10'~

~ . r r t s t o at tam u . 1 ~ emston rate

T a l e 4 Prtd ic t ra, of erosion us1119 n o n r l l r e d arwage croslon rrtr an4 l n s t m t r a u s cmr lon rate

Instmtarrous. -1

I l p

tfC'

A w e r q e . @'%

0.37.10'~ 2.52 5.07.10~ , 1 . 4 ~ 1 0 ~ 0.69 ,

Ivwaqc.

?XIl. g l l p

193

==a

2 . U

92.8

!

$2

22.4 1 .25 32

3.8 1 -34 1 79

iI ell g

Percolt -or of

noru! ~ r c d ems lm ra te

1.6 0.19

1.54

4 1

n

41.01

6.05

.97

.81

119

45

71

t R iiiii;

1.78

Pwcrnt error of n ~ r ~ l rzrd

U O % I O ~

rate

0.61 0

ORiGINAL PACE rS OF POOR QUALITY

:F i n t steawstate period I :

A

1 : I 1 L Tendencyilwlard +

+G!Azfinai period

I I (8 , I I -

Exposure time

t

.-Armmulation zone

-

cab Thiruvengadam and Preiser la. (b) Plesset and Devine 13).

cc) Heymann (4). (dl Tichler and de Gee 1%.

I I Exposure time

Figure 1. - Characteristic erosion-rate-versus4irne curves.

t

,r Peak erosion rate

: C

= 5 C 0 z W

Exposure time

4

r c ~ t i n u o u s

,' emion rate

*

~ P O S U ~ ~ time. h r

Figure 2. - Curnulath? erosion versus time for rgiFless steel tested in dting-disk de vice. Pressure. 150 kPa abs; inducer diameter, 25.4 mm. lnstaneneous emsion nte at O equals skpe of local tangent at Q = A V l A t cumulative average erosion rateat Q equals sloped line joining origin and pint Q - V 4

P Q

L 1 B - - au - J

S 0

0 40 60 80 100 120 W u r e time, hr

(2.) Instantaneous erosion rate. b) Cumulative erosion raie.

Figure 1 -Erosion rate versus time for stainless steel $s$d in a M n g disk device at various velocities. Pressure. 150 MPa abs.

Normalhed time

0 1 2 3 4 Normalized time

(a) Instantaneous erosion rate. b) Cumulative erosiov rate.

Figure 4 - Normalized erosion rate versus normalized time for stainless steel tested in a rotating disk device at various velocities. Pressure. 0.150 MPa aba.

ORtGINAL PAGE lS , OF POOR QUALIW

0 3% 60°

.6

:daterial

.4 0 Annealed nickel A Zinc2 O Nickel (distance between

Ipecimen and horn, 0.051 cml .2 0 Nickel (distance between

specimen and horn. 0.064 cm)

I I I I 1 I 0 1 2 3 4 5 6 1 8 9

L E

B a g o 0 8 0 8

al 8 O Oo v B O

8 Pressure, Tempentcre. A

2 MPa "C 0 P = Q o a2 649 v 0 m - a 0 0 .3 649

v 5 . 4 0 .4 649

NorPalind time

U

'E - - 6 z

(a) Aluminum tested with a rotating disk device velocity. 35.8 mls; temperature. 320 i P C: test liquid. water.

(b) L a cobalt alloy tested with a mepnetostriction apparatus. Specimen. vibrating; frequency. 25 kHz: test liquid. liquid sodium.

0 4 . 2 427 0 0 .3 M

. z - A 0 . 4 427 0 . 2 m

0 A . 3 204

with a mrv~nrtostriction appantus. Specimen. stationary: frequency. room temperature; test IIquM. wter.

v - 4 204 I I I I I I I I J

Figure 5 - Normalized average eroslon rate versus noml izud Hme for wrious materials t e r n

Material 0 Coprpr 0 &a- 0 BDl4lBms A Silicon skd 0 1U8 Stainless s M

E

-rial Test liquid

0 B m s F m h rater 13 Costinm 0 Cold rolled sw

A C a r t i m 8z llBE

gZE

1 O Cart iron Saltwater

0 Wim Saltrratw

I 1 I I I I 9 1 2 3 4 5 6 7

Wonnalizedtimo

la) Liid imginpecnent ldata fm yB ). 8) Vibratory cavitation ldata frm (23 1.

Figure 6. - Mwmalind average erosion rate versus normalized time for various materials tested

F i n I. - )loraalizuJ erorian n(e ems nafmaliad EU for L a Jlq IcrW in Ilugw(lKtm 4pparatus at 4@ C in liquid sahum; I-hr ti- intervals; spciarm. v i m @ fregmn~y. 25 kt&?.

1 I0 Normalized time

Fire a - Normalized cumulative average erosion rate versus normalired time for long snposures.

aRcP

;r;t r2.x

fg O

F POO

R Q

3%R

Y