OBSERVATIONS ON CAVITATION DAMAGE IN A PLOWING SYSTEBM* Frederick G. Hammitt June 1962 Internal Report No. 14, ORA Project 03424 for National Aeronautlic and Space Admainistration Grant No, Ns 6-39-60, Summary of OlA Technical Report 03424-4-T accepted for. publicatioe in Trans. ASPI.

The author would like to acknowledge the conultation ati asEistance of Pofessors R, no EfDPehlke a Co Ao Siebert of the Chemical aen1 Metal[lurgical Engineerin Departmnt, and tV:~e.sistance of Messrso L. Bri a, Vo Bie P, To Chu, V. P~ Cramerv Ro Do Iany, B.o Rpkte,, and Mo J, Robinson, in the conducting of the research investigations herein repo Bted Sincere thanks also are extended to the personnel of the ecPrometrie Manu~factur^g C.any for the loan of their oLi.o Proftcorder, and assistance in its operation0

Cavitation daage to specimens og stainle$ss steels carbon steel, aluminum, and plexiglas, placed in a cavitating ventur using water and mercury as test fluids is mostly i, the form of irregularly shaped piLts which,do mot change with additio1l eQposaue -to the eavitatig field within the limited d3uratikon'j ltil %izeco Th ate of duzage is. very hig. initia,11, decreases for re1ativelly short period og tiie then increases again up to the maximu test dueratlonS of 150 hours w -th water and 270 ho rs with mercuwylo Observation of damage effects by several indtepemndmt technigues using a vasriety of specimen matierial~s with two different fluids under v riolAs $fludi dynFaic ondleitons u leads t;o a suggestied co(ieratinLg ael in te as of -the c(avlttioa n btbbl&e deriitzgity enr gy c-amd p eciaen iaatexnital atrsngtb,'

ACKxOWLQQeGEg iT i LIST OF TAB LE LIST OF FIGU vS Io IN DUCTION 110 TEST APPARATU', x:~ OPEIATING PROCEDURE -.1;O EXEiRIE- AL flSZkWA lOW A o Dna.ge As A F nction of Time 2o Single Event Pi'ttij Theory Co Types of Pits D0 Pit size Dista:ikcbu tioe Ie Test Material Effets F, UFluid Effects 58 o Fl Parazeter Effgeo s t-, CONiCL US3 4 6, PbXt~~ PaosP~blE~e~: 4~~;i~;;, aXS$ sde fi w,/

Table vps X Pit Count Tabulations I Typical Mechaanical Prop rties of Materials Tested 26 tPb~wPzr

2 Dage Test Veaturi 4 3 Drawing of Test Spimen 5 4 Photograph of Test Spemen 5i b, Cm parilson of Weight Loss by Pit Counting and by Rai oactive Technique 9 ~olw~ite oss Pi, Unit Area vs o Time f@o Stainless Steel In Water U9 7 Volu Loss 1Pe Unit Area vs. Time for Stainltess $t~el in /(;)rCUf y 13 $ ~Typical Cavi7tation Damage oa Carbon Stees with Water 14 9 eveleopat of Cavitation Dage at Two i atmions on 380 Mes2 leja 8Steel Wit5h Wate (~ t;aer 15 and 30 hours) }1!@,,evelopmam of Cavitation Dama.e- a't Two,acations on 302 8tainl~ess t-l 1e 7 1th Water. {afgstr a150 xls) L. TyA,ai l )rofi oerder Trace of Pit Profi ie i1 l12,j~aoactiveg Cavitation Pasrtile Size DXitrsiimtion ai Li' tJ'ypical Sectio i Throuagh Irreular Shaped Pit For Sta. itlemss at, e. in ater 22, 14 Wxoialies d AiKl Pressure Profiles 31 Fo Stainless t~el,'I Wlater 32 Voolae L-s Pe Unit Area vs, Degree of Cavit;ation F;'ox' 8t aiLems ftie 8 i itatr~P a4

Emge Exponent vis,t Test Durantion F.or Several mypotherled i bbLw aergy tctza Hy ~Bpothesized Bubble Snergy Spectra t3s 5

Damage to ~structural materials by cavitatiaon, perhaps assisted by ordinary erosion and corrosion, has long bee a serious problem to the manufacturers and users of fluid-flo componentso It has often been possible in the past to avoi.d the problem to some extent either by suf f iciently reducing the performance-ratings of components, or by increasing system pressures0 In na erous present-day applications9 such as those' in the aero-space field, such compromises my not be /easible, becausiw of the over~-riding necessity of minim'zeig sizes and weights of components0 Hence, there is a strongly renewed interest in understanding the fundamentals of the cavitation damage process, so that meaningful predictions of daage to be anticipated with a variety of fluids and structural materials over a very large temperature, pressure, and velocity range can be adeo While much data on various instances of cavitation damage are available in the liter-.2? ture, it is necessary that further systematic investigations be made under v~ey carefully controlled and well known ~o.ndit-ins, covering a broad range of fluid, material, anrd flowa. parameterso ThiS paper discusses some of the results fr>ca the initial phases ot such an investigation0o Bcause of the coniderable complexities of the damage phen enon, it is desirable that the test conditions under which damage is obtained match as closely as possible the actual operating conditions of applicable fluid componentso Iowever, t'his approach if carrie to anm extree $os noea represengt a practical ideal, beause of the lack of generali ty

As the best cugpromise solution availabile the inesrtion of damage specimens into a cavitating venturi test section ws selected for the present investigationo It was felt that this solution combined the close applicability of flow-inaduced cavitation in a flowing system with a highly-simpliftied flow pattern. In addition, the equipment requirement in order to operate over a considerable range of fluid, flow, and material paramteters is relatively modesto II Test Apparatus (1) The test facility is a closed loop, powered by a, centri fugal pump, and includes a plexiglas venturi test section (Fig 1),0 The venturi (Fig. 29 has a 60 included angle nobzle and diffuser, se:parated by a cylindrical throat of 0o51 in-che diameter and 2035 inches length0 The two damage specimens are inserted ith th:eir midpoint 079 inches donstream of the throat exito Theiy consist of planar sections located para.llel to the strem with tap red leading and trailing edgea (Fig~ 3. 4)o They are 0074 inches long by 0o06 inches wide,, and are subawtrged to a d'epth of about 0o20 inche@s iato the cavitating streaml They ara located symmetrically about a vertical p)lae passing through the venturi centerline, so that they are each at the same elevation, with their axes inclined at 45~ to the vertical0 The faacilJity has been oprated with both Nwater aLznd mey, cryo This paper is coanerned mainly with the water tests, since thc -Pmercury data lare not yet f lly evaluated0 Throat ve. citi s

DRIVE PULLEY BEARING HOUSING THROTTLING VALVE MEASURING VENTURI -TEST SECTION I - lI STUFFING BOX SUMP TANK THROTTLING VALVE89 COOLING WATER IN COOLING WATER y Fig. 1 Overall loop layout.

-2. 1.25~-[1" B I x 3 STUD AV TO NVOSE WEAR SPECIMEN SANAR CAV TO BACK CAV TO I ST MK.11 VISIBLE INITIATION Fii~ ~ ~19 cvT ZDK SCN 11 -.510' 2 —— ~3 —---- 10111 2.247.4 3.013" 5.361 0 -5773' SECTION A -AI 20=5 54 6.~52 E 604' 14. 578" Fig. 2 Damage test venturi, showing locations of specimens and specimen holders. The dotted lines in the diffuser represent locations of cavitation termination for various degrees of cavitation. (Half-inchne venture test section No. II.)

-5I E o.;oo, WA 0. 0.500"- 0.12 d' 0.740" 0.74 5" 28" 28 1 0.060" Fig. 3 Damage test specimen. Fig. 4 Damage test specimen. The upper shadowed surface is the polished surface.

-e av it a t G C4")ndltlon C a Ube a jus@ted betwe5an,Ctiatiionl Lnd'secomd mark" (defined -in Appedix)o Wiater temperature can be varied between about 500 F and 160~ Fo With mercury, the attainable throat velocities range between about 20 and 50 ft/sec, over the sae range of cavi — tation conditions o Test materials so far have included carbon steel, austenitic stainless steel, alu8minum, and plexiglaes in water; all but the last in i.ercury, In all cases, the specimens are metallographicalLy polished prior to a test0 Further deteals of the facility and its operation have been given previouslyo (1, 2) perating Procedure.To the present, the maximum test duration is of the order of 10t to 200 hours However, tkhe specimens are removed at frequent intervaks during this period for examinationo These 0,6tminations allways i nlude: i) Tabulaton of pits according to size and nuber ii) Weight iieasurement of specimen' iii) Phot oicrographs of unusual pit formationso In some cases, the following item have also been includedo. i) Pit tabnlatioa according to type and location (1) ii) Measuretent of pit. profiles using precision'"progicorder (1 3) ur"Likaear Profipc Qder anufactured by icro~,tical Aanbon.f/cig turinig cpanyy.pa nn Arbor, Sic i an.o

structus re (t R$hAe dest-ruective examinai t lo$ prevents further spci men use) iv) Irradiation of test specimens with subsequenit masureent of radioaetive contamination'in water, and of cavitation particle size distribution using sta@ked, precision filters (1, 4) The experimentl observations, to be discussed 1ater, are drawn from all of the above sources, EX rimental Observation The experimntal observations of this investigation are so far largely concerned with the early phases of cavitation damage, iteo, relatively al11l individual pits, primaily, before gros damge has occurred, SineG the key to an eventual understanding of the mechani.sms producing gros daage must lie in an understanding of these initial phases, and since previous precise observations of the initial phases arte not nerous, i it is felt that the present observations aro of substantial interest A o. age as.nanction of Time In the present tests damage has been evaluated uing two techniques o i) Irradiated test specimen (4) i) Calculation, based on pit size and nber tablulation, using a typical pit profile (1) Direct measuremert of weight loss in the water tests has not': been feasib srle ti~e he ls,.as a proportion of tes s ci~eE

measurements hava been achi evd in the maerury tests wh eo-r tfhe geighlt loses area much greaterC, The absolutle magnitudes of daage obtabined rom the irradiated specimen test (only one has so far been ccnducted and from the pit volue calculations do not at present agree, although the sbhpes of the weight loss vso time curves are virtually identical (Fig0 5) RHowever, in the irradiated Specimen test, th~ere are rvarious significant sources of e rror which it is ho>pad will be reduced by subsequent development of the procedureo Oeviously, there are also substantial.nicertainties in the:it volume calculation involving the assumpt:io of a typical pit shape, in the extrapolation of pitting dsnsities from a relatively small monitored surface to a. larger unobserved surface, in the reoation between pit volume and material volume:Temoved/ etc. In the future it will be possible to obta in a direct comparison between the pit ealculation results a.,d;he results of direct weight measurements in the mercury testso However, these data are not as yet available O Even though the absolute magnitudes of volume or weight losses are subjec-At to uncertainty, it is belJeved that the,,onsistsency between tests is goo d so that meaningful c!pa.risons betwe:;en the diffe-rent materials, fluids, and test conditions: can be drawno This statement is based upon the reasonably "~it has so far ben ass ue.id that these a.re eqgul1 ~ weve', f, is krnn that this is not exactly trite fo:~ any of the pit e,,on ~ figarations and is probably in gros error f or s.m~. (~, rat.e~.~'o

r~~~~~~~~~~~~~~~~~ ~~Trans ASME, Dec 1962 F.G. Hammitt 01 Fig. 5 o En n TOTAL SPECIMEN WEIGHT LOSS (GRAMS) moo0 )d O0 C' -.- 0 O OI C (D c+~ I.!! _(D (n 0b ( C+ C) co r~ ~ ~ (0 I> rt (D (Dqq HU) Cb ~~~~~~C.(D C C+ CZ (0 Q0 (_ 0 C+ d Cb H. 0)r Cb ~ +n 01 rt- ~ ~ ~ ~ ~,. P..... (D c, rs cb l~1 o O(D P% If) U c C1 Cb (D r+ ~ ~ \ rr(D C ~ j

SthiO~. Cent~r Obt$lwnad ~r vol ume loss as a kuac-tioi tt, as, well as upon the virtually identical shape of th e u rvS prod uced from th~e irradiated specimen approach and tlh, pit calculation-: two very different methods of observationo The damage has been presented in terms of volume removed per unit area exposed, i.eo, specific volume loss (cm 1/cm or simply cam, representing a mean depth of penetration if the wear were unifor rather than in the form of pits), It is felt that presentation in these terms allows the most meanigful comparison possible between different materials and tes't arrangements, The specific observations are discussed belowE 1Q Initial Rapid Damage Rate Figures 5 and 6 show cavitation damage with wate.r as a function Of time for "standard" cavitation at 65 ftc/s6ec throat velocityo Damage for other cavitation conditions is also shown in Figo 60 Although these curves are for,,:somewhat similar conditiolns, their shape is typical of all the curves for either water or mercuryo It is noted that there "is a very rapid initial ratle of damage (more apparent in Fig 5 where greater detail is shown), showing substantial pitting after no more than one hot.ur of exposure (first examination)o after the first one to three hours, the damage rate decreases substantially, reeaining relatively low for a period (up to the order of 35 to 100 hours, depending upon cavitation condition, material. etco. Fig~ 6), anId then climbs at an accslee-ated rate, at least to the maxima dauration attained for water to

-1160 Symbol Specimen Cavitation Condition 0 3-/8, 3-19 Cavitation to Nose ~55C LA 3-35,3- 36 Visible /nit/ation: o 550 _ 3-4, 3- 5 Standard ~E N 5V 3-16, 3-/ 7 Cavitation to st Mark oE 0 3-3 3-34 Sonic Initiation 0.2 LL 45 L Throat Velocity 64.7 ft/sec u4 5 Cold Water, Air Content-Saturated W 40 aw 35 LOI w 0... z _ w 0 20 40 60 80 100 120 140 160 TEST DURATION - HOURS Fig. 6 Mean volume loss per unit area of specimen exposed to fluid for pairs of stainless steel specimens vs. test duration; for several cavitation conditions with water.

foPr. eazipsim~ dueatio~ of 2T0 haours shows a:en sbseque nt ieve!.... ing4fi f and then a second accelerated climbo it is felt that the above results are significant in thiee i} The absence of any "i~nubation perio d" is noted for these tests~. Suice the initial rapid damage rate does not result in a significant weight loss, it is felt that this initial damage may have been missed in some of the previous investigations which relied only on weight measurements for damage detection, In Fig0 5, the initial weight loss is clearly showvn by the irrmadiated tracer test as well as by pit coun itso It is noted that an apparent incubation period would result Aif the later, rapid damage portion of the curve (Fig. 6o were extrapolated linearly to zero. ii} Large pits, io,, o of the same general size as the largest noted in any of the runs, are formed within the first hour or soe (Fig" 8} tiii) The wea.i rate generally tends to increase with t.ie~ Ha the present investigation, no limit to this trend has yet been foundi The,",e a.re, however, periods of rea tively short duration during which the wear rate appears to reach a mrinitime The above a:fects are believed more a function of the properties of the material surface than of te c avitEtion condittion It is postulated that the initial rap-id wear rate is a result of remov..l of surface defects such as ic l lus' ions or o her "wek spots" Once these relatively few "weak 5pOEs have'been removed~ the wesar rate d.ereases dra;tic.llyo As

-13UZ 60- 0 Specimens 47-3, 48-3, Mercury o Standard Cavitation, Throat r x Velocity = 34 ft/sec L-~ E 55 E 55 O Specimens 3 -04,3-05, Woter a E r | Standard Cavitotion, Throat'J. E 50-. Velocity = 64.7 ft/sec W45CO 0.. w 40 rD- 35 LC a.25 0 20 1 5 0 z 10 0 50 100 150 200 250 300 TEST DURATION - HOURS Fig. 7 Mean volume loss per unit area of specimen exposed to fluid for pairs of stainless steel specimens vs. test duration; for standard cavitation condition with mercury, and comparison with water.

-14Fig. 8 Typical cavitation damage on carbon steel, for standard cavitation condition with water. (Specimen 1-19, after 3 hours; throat velocity 64.7 ft/sec; x 100.)

the -oedurancs 1J it is approachedo In addition, as the amouat of dmage increases, there are flow perturbations due to the damage and additional area is exposedo The final shape of the dmage —,time curve is a result of the interplay of these various influences, and perhaps of others as yet unsuspectedo In the relatively short duration tests of the present investigation, it is doubted that the feedback between surface roughening and foi >w rturbatioi is yet important0 inle-Event Pttitug Theor It has been experimentally verified, within the durations of the present water tests, that the surface outline and depth of a pit, once formed, does not,, in general, change during subsequent testingo This is illustrated by the photcmicrographs of Figures 9 and 10, shoring two particular areas, one after 15 hours of testing, and one after 30 hours, and both again after 150 hours, Interamediate duration pictures are also available,(l) as are other series taken at different locations and on different specinenso (3) iThese have not been included in the interest o~x: brevity0 They all illustrate the fact that while new pits my be formed in the area of an old pit, the old ones are, in genexral not changed, at least in surface appearance~ It has been demonstrated in addition, (3) that their depth profiles are also unchanged (Figo l1)0 Also, as previously mentioned and as illustrated by Figo 8, typical "large" pits are formed as readily early in the test as later, These facts are consi$tent with the assption that the pits, at least i.n this X$~d~8

-16- ~~~~~~ - Il I- |(a) (b ) less steel, for standard cavitation condition with water. (Specimen 3-5; location (a) after 15 hours; location (b) after 30 hours; throat velocity 64.7 ft sec; x 100.)

-17Iis~f j...................................I ~~~~~~~~~~~...':........E........~i..44...i..r.....,.6f >............ zZ* i~i;a:?. v:<:ff-s::'s i-iL:E::gff~i.::.::-:. L~:77iELi::::&:i' i:E:E:*:: s:dE*'s i:L:.:t tr:::::i:vo| =;| _..................................................................... l M ~ * Bz~~~e~~eB~~~g<>>aelaa e e.?~~~~~~~~~ iiiiiiiiiyB~iii:i i -iiiiff ii B a a=t at a (a) 9~~~~~~~~~~~~~~~~~~~~~ii~i~: xli I: — *- - R:S~i.888z?::'.g~i: S:?Si:,.V&i:Bi~R~:'SiT:::::E.EE-:-tiE7:':ET~E:TH::-::'0T::i:::: i:: -TEi:::E -i:.':_:i-i -i:::::7-'::::::7::L:) i~:i E: -- -~ i~ ii ii:::: i:~:L _:- -_::,:::-:_::::: i 7: i iT: t if T i E f:...:.:............ Fig. 10 Development of cavitation damage on type 302 stainless steel, for standard cavitation condition with water. (Specimens 3-5; locations (a) and (b) after 150 hours throat velocity 64.7 ft/sec, x 100.)

i TOTAL'ROCORDER -O LINEAR [I L/ WAVINESS )FICORDER El ROTARY @ RPM. 010 PROFILE: []ROUGHNESS DATE Sp cime 26'I -, ( I Trace rec rded 62/10/ 61 _ r 1.r, wear damage d CHART NO. 22131 MICROMETRICA El MICROCORDER I LINEAR DATE PART NO. - 26 PROFICORDER [ ROTARY C@ RPM 0oo4 L/ / I IiqAI~ ev~len3.100 b a e 6 1 h e dage MICROMETRICAL MANUFACTURING COMPANY, ANN ARBOR, MICH. Fig. 11 Typical Proficorder traces of cavitation pit profiles showing the unchanged profile of the same pit after 10 hours additional testing. (Note horizontal scale change from 0.010 inch/div. to 0.004 inch/div; vertical scales are 100 microinches per division.)

further at'ack Stc ttiastiay it is unlikely that a second large pit would overlap a previous one if their location i entirely randoma since in the longest duration water runs, only about 5% of the exposed area has been affected (1I) If the location is perturbed from the random case by effects upon the material, it is likely that the bias would be such as to tlend to prevent a second pitting of the same area because of the effects of local work-hardeningo Since the pits are not changed after their formation, they must have been formed in a single event, if not n@cessapril.y by a single blow. anay of the pits are virtually symmetrical craterso It is inconceivable to the writer that these could- have been formed in other than a single blow because of their symmetryo However, those of irregular shape could be either the result of a singlelblow or the cumulative effect of many weaker blows (fatigue failure).In any case the actual material removal must have occurred as a single event, or pits would necesarily continue to gror as exposure time is increased. The above romarks are further corroborated by the results of the irradiateRd specimen run where the size distribution of the cavitation d:mage particles was actually measured (4) (Fig 12)o It was found that about 50% of the removed mass would not pass a filter of about 2 mil pore size, Nulserica)ly of course, the gireat majority of particles are then less than 2 mil, but a substantial number are greater 3 This roughly 3The largest singledevent pits observed in the water testas are of th e order of:io mils0o

-200) (D w E 0 0 LX EL. 1 ILL) L,/ L I ~1 I I... 0.01 0.1 1.0 I0. FILTER PORE DIAMETER - MILS Fig. 12 Percentage, by weight, of radioactive stainless steel cavitation particles which pass through filter vs. filter pore diameter.

Cs Types of Pits As mentioned briefly before9 the pits are either appro~i~ mately s eymmtrical craters, or are characterized by an entirely irregular contouro (Figures 8 9, and 10) The profiCorder traces (3) have shown that in all cases ex ined the diameter to depth ratio is large (order of 20)o Hence the pits of irregular contour have been called slabso It is postulated that their shapa is a function of the irregularities of the surface structuere (~a test with an etched specimen shoae~d for exasple that in some cases the pit outlines followed the grain boundaries) rather than of the fluid-dynamic parameterso A plausible mechanism for their formation is that the fatiguing of the underlying material results in the loosening of a smalU. slab which is "peeled off" in the downstream direction by the staeam velocity~ This mechanism was suggested by Boetcher w ho included in his paper (5) a picture of such a slab, apparently ready to leave the surface This mechanism has been somewhat Corroborated in the present investigation by the follwin&ge i) In about 90% of the pits examined with the proficoarder (3) (no true crater has been so exained to the present) there is a ridge of materiaY. on the downstream edge of the pit oulyo This suggests the pe&.ing away of a slab in the downtream dircODeto o ii) Slip lines have been photographed below pits in the present investig.ation (1) (Fig0 13) as well as in previous investigatlons {( and 6 for example), indicating the presence ~f laVrge mechanical streesses These have also beDn indicated

-22Fig. 13 Typical section through irregularly shaped pit in stainless steel, showing slip lines. (X2000, oil immersion, slightly oblique illumination.)

In a limited inumer of cases in the present investigation S a very detailed pit tabulation has been made to discover the proportion of pits of the two types and their location as a function of the various applicable par meters. The results of one such tabulation are presented in Table X It is noted gene2rally that the craters represent only 10% to 20% of the total pits i4a these watei tests with no definite trends evident with the;rloW tost parameters Since pits of this type a.re presumably formed in a single very intense blow, it seems likely that ththis proportion may increase for combinations involving denser fluids, more intense cavitation, or weaker materials,0 The tabulations have not as Fet proceeded far enough to draw any conclusions on theso matterso Do0 Pit Size Diltribution Further reference to Table X indicates that the number of pits in a given aze range is always substantially increased as the size und ea consideration is decreased, For example, there are hundreds of pits in the range 04 to 1G0 mils, whereas there are well less than 100 in either of the larger categories (5 to 10 and 2o5 to 5 milst respectively)o The large number of 4its in. the relatively uncountable swallser ranges does not introduce large errors in the volue loss ~alculation sin$c the pit volure is proportional to the d-aeiZter cubed, and evyen the disproportionate number of pits in th s.allest category has only a relatively negligble influen@e on the volue lo~so

-24TABLE I PIT COUNT TABULATIONS* Tabulation A Pit Count and Various Sizes Pit Size (mils) VVL VL L S (10 > D > 5) (5 > D > 2-1/2) (2-1/2 > D > 1) (1 > D >.4) % Cir. Total % Cir. Total % Cir. Total % Cir. Total Sample No............ 4-3 Polished 10.0 10 23.4 17 17.33 75 12.5 400 Surface Sample Position...... Front Numbered Side 21.75 23 18.7 48 16.05 81 15.1 325 Standard Cavitation Opposite Side 18.75 16 21.1 19 14.55 55 4.9 465 Throat Velocity......64.5fps Subtotal 20.5 39 19.4 67 15.45 136 9.1 790 (2 sides) Duration of Run.......150 hrs. Total All 18.4 49 21.4 84 16.1 211 10.3 1190 Surfaces Sample No........... 5-3 Polished 15.4 13 22.2 18 25.4 110 25.8 480 Surface Sample Position.......Back Numbered Side 11.1 9 21.1 19 15.9 63 16.3 320 Standard Cavitation Opposite Side 6.3 32 7.2 14 10.2 49 12.9 350 Throat Velocity.......64.5fps Subtotal 7.4 41 15.2 33 13.4 112 14.5 670 (2 sides) Duration of Run.......150 hrs. Total All 9.3 54 17.7 51 19.4 222 19.2 1150 Surfaces Sample No............ 18-3 Polished 0 11 4.5 23 6.8 74 15.0 320 Surface Sample Position...... Front Numbered Side 0 29 5.0 40 5.9 51 13.5 200 Cavitation to Nose Opposite Side 0 13 9.1 11 9.1 33 12.3 130 Throat Velocity....... 64.5fps Subtotal 0 42 5.9 51 7.2 84 13.1 330 (2 sides) Duration of Run....... 50 hr. Total All 0 53 5.4 74 7.0 158 14.0 650 Surfaces Sample No............19-3 Polished 9.1 33 17.2 64 15.0 100 10.2 365 Surface Sample Position....... Back Numbered Side 5.3 19 0 14 9.3 43 18.7 150 Cavitation to Nose Opposite Side 11.1 9 0 16 7.7 26 8.6 105 Throat Velocity.......64. 5fps Subtotal 7.1 28 0 30 8.7 69 14.5 255 (2 sides) Duration of Run.......50 hrs. Total All 8.2 61 11.7 94 12.4 169 11.9 620 Surfaces TOTAL ALL 4 SAMPLES 8.75 217 13.9 303 14.3 760 14.1 3610 Tabulation B Pitting Intensity Total pits/in.2 - hr. VVL VL L S (10 > D > 5) (5 > D > 2.5) (2.5 > D > 1) (1 > D >.4) No. Pos. Cav. Condition Vel. Duration Pol. Side Pol. Side Pol. Side Pol. Side 4-3 Front' Standard 64.5fps 150 hrs. 1.79 0.87 3.05 1.49 13.44 3.04 71.7 19.6 5-3 Back Standard 64.5fps 150 hrs. 2.33 0.915 3.22 0.74 19.7 2.50 86.0 14.9 18-3 Front Cav. to Nose 64.5fps 50 hrs. 5.91 2.81 12.38 3.41 39.8 5.62 172.0 22.1 19-3 Back Cav. to Nose 64.5fps 50 hrs. 17.75 1.87 34.4 2.01 53.8 4.62 196.0 17.1 Tabulation C Summary WL VL L S % Cir. Total % Cir. Total % Cir. Total % Cir. Total 4-3 Total All Surfaces 18.40 49 21.40 84 16.10 211 10.26 1190 5-3 Total All Surfaces 9.26 54 17.65 51 19.4 222 19.2 19.2 1150 4-3 & 5-3 Total All Surfaces 13.6 103 20.0 135 17.8 433 14.67 2340 18-3 Total All Surfaces 0 53 5.4 74 6.96 158 14.0 650 19-3 Total All Surfaces 8.2 61 11.7 94 12.4 169 11.9 620 18-3 & 19-3 Total All Surfaces 4.39 114 8.94 108 9.8 327 13.0 1270 4-3 & 5-3 + 18-3 & 19-3 Total 8.75 217 13.87 303 14.34 760 14.07 3610 Memo, L. Barinka, Nov., 1961.

HiStion td abocve0, Ho v, tho is not i port;;.$rar than ti p t oft i the Volume loss calculation0 Bo'lTest BMaterial1 ~Bffects The tests reported herein mainly involve water, at approAiMtely ambient tmperature and at close to saturation air con.-D tent as the working fluid0 In a few cases ercury has b.een ised. The test mterials have beeni) Sitainless $t eel (gtpe 302 annealeod) ii X1010 Carbon Steel (annealed) ii) Alumin: ) Type 1100-0 (annealhed and b) Type 6061T(G61-ST-6 9 age-hardened) iv) Plexigxlas {polymethyl methacrylate) echanical properties of the tested materials are listed in Table 110 On all these materials the general appearance, of the pit ting is similar. There are, however, si gnificant di.ge rnes between rates of pitting, Also, in the case of the carbon ste el9 there is significant corrosion in addition to the mchaaical pttitigo The daAage rates gor the stainless steel and Carbon steel are of the same orderr of magnitude, with the volume loss of the carbon steel being aboat twice that ot f th stainless9 prior to that period where corrosion becomes very s-ignificant (about 2 hourS)o n mercury the rgtio 2s ho t order of 4 at 20 hours but only 5/3 at 100 hours, based on two sp9cimens of each material0

TABLE II TYPICAL MECHANICAL PROPERTIES OF MATERIALS TESTED Hardness Tensile 0.2% Yield Strength Strength % Elong. % Reduction Measured Bending Fatigue Elestic Material Condition psi psi in 2 in. in Area Rockwell B BHN Strength psi Modulus Stainless Steel Annealed 98,ooo 37,000 55 65 76 140* 35,000 28 x 106 Type 302 @ 107 cycles 1010 Carbon Annealed 50,000 30,000 40 71 48 85* 25,000 28 x 10 Steel @ 107 cycles Aluminum Annealed 13,000 5,000 35 - - 23** 5,500 10 x 10 1100-0(2S0) @ 107 cycles Aluminum Age 45,000 40,000 12 - 57 90* 13,500 10 X 10 6061-T6(61-ST6) Hardened @ 107 cycles Plexiglass - lo,445* - 2-7 - - M90-M100 - 0.4 x 10 Rockwell** * Measured value or converted fron measured value ** Typical value

comamrea d to that of the stelsro 1n f~act the, vo1Vm8 loss from the aluinum. in a five-minute test is of the order of lo5 times that of the carbon steel for an hour test under the same lo t condltio, althoughl as shown in Table II the hardnesses and strengths are roughly the sameo No explanation for these somewhat surprising results is available at present. No quantitative test of the soft aluminum has been made, although it was naoted generally that Its pitting rate was many times that of the age-hardened aluminum, The plexiglas was surprising in that its resistence to damage in the water tests was very high, superior even to the steelso Al'though no quantitative tests wPere made with plexiglas in water, it was noted that virtually no visible pitting ocurred on the walls of the plexiglas venturt after hundreds of hours of exposure, and that a plexiglas damage specimen, in a relatively short test, showed considerably less pitting than the stainless steel under the same conditions It was felt that this relative immunity to cavitation da age probably resulted from the low elastic meaulus (about 1/70 that of steel, Table XI), combined with relatively high strength (about 1/5 of carbon steel) in that relatively very large deflections would be required to cause the material strength to be exceeded0 Since the high-pressures generated by collapsing bubbles are extremely local in nature, such deflections might remove the material from the critical areao A similar arg aent has been applie previously to explaiA t.he relative -27

plexiglas ~I bessted in mercury was found to be g::reater th~~ athbXat of carbon steel by a factor of about 100 The reason for the inconsisteicy of the results between water and mercury is not at present knowno F o Fluid Effects 1ly two fluids have been involved so far: approximately ambieast temperature water, and mercury, Whera s an in itia set of water runs has been completed, the initia.l aercury test, arxe still in progress, However, certain significant tretds are evident o i T@he wear rate with mercury is apparenAtly several orders of magnitude greater than that with wa'w;ero At presen~t there is some uncertainty regarding the exact ratio for any given test since the weight losses with mercur7y are 1arge enougb for direct measurement with good precitiion, while,. oso with water were not, so that the scmhat tenz:ous calculat.ons based on pit tabulatio were necessaryo Altlhough the irradiated specimen test with water is a form of direct meaaiu-r. ment, there are various possibilities of sign.Aficant error which cannot be evaluated until additional te:,ts are madeo sence a precise comparison is not possible u'!;il the. teati cedures have been developed to better precisalion, and the pit count data obtained with mercury have been r CFuced ~ and,oiar,-revl with the direct weight lossesa0 However, a "'best gu.esaa da::age ratio bet een rcury and water 4 age rates S-)Xt approgip tSa t.

.*de2taDly bacause of its ability to prduce damage rapidly, mercury is a very ueful cavitation amge test fluid, ii) Thi type of pitting encountered with mercury and water is generally similar, However, the mercury test series and the examination of the data produced are not sufficiently advanced to allow more detailed conclusions, Flow Parameter Effects The flow parameters which have been varied significantly in the present tests are cavitation condition and throat veloity0 Their effects are discussed belowo, I1 Cavitation Condition4 Independent of throat velocity, the cavitation condition ("degree of cavitation") can be varied from "zero cavitation" through initiation to fully-developed conditions (First, and 3econd Mark), wherein the region of visible cavitation teri.-. _tes considerably downstre a of the trailing edge of the test specimenrc From the viewpoint of cavitation damage, the s ignifieant difference between these conditions is the pressure existing in the vicinity of the test specimen, As the degr.e of cavitation is adjusted toward the more developed conditionsa,'the pressures in the vicinity of the test specimen decrease, eventually reaching values close to the vapor pressureo Such a direction of adjustment results in two diverse trends: i 1The number of bubbles in the vicinity of the test 4ftined in appendtix,

i~ he driving pressure differential fs'bubbel colldapv decreases, (Typical axial pressure profiles areo sho in Figo 14 ftor all the cavitation conditionso) When the degree of cavitation for the water tests is reduced to "visual initiation", for example, there is no cavitation visible on the test spe imen0 However, sa11l, l oca cavitation regions, i duced by the test specimen itself maot xisat since the daage in this condition is signifiCanto Thus!it appears that a small number of highly energetic bubble collapses can be more harmful than enoaously greater numbers of less energetic bubbleso hais trend is further evidenced by the fa,.t that very little damage is sustained when the cavitation region entirely envelopes the test speci no On the other hand. "sonic initiationt", wherein no cavitation is visible, aleo prtuces very little dage.e (Even though a large pre~ssur difflerentil is available in the vicinity of the tesat sya 3ei for hbubble collapse, too few bubbles for significant damage penetrate far enougho) Finally, if the throat pressure is raised costiderably above that corresponding to sonic ic.iK.X, tion tJe pitting rate is reduced another order of magni.ude (apparntly pure eroion from single-phase flow also pr.duSe pittin,? of similar appearance, but in considerably reduice& quantity)0 These results are illustrated graphically for rater in Fig0 150 Although not yet entirely evaluaated, the reasults wb ~aercurB y apFoar diafflrent in that thhe bubbles do not penetrate

(D 0.) o 1.4 Reynolds No. = 306x 05 - Sonic E Y _- - o.306 x /05- Visible I._ L L- \Posiftinof -3 =02 /055- st Mark Test Specimen 1.0 -8 -8 Throat OGf/ef ~-].4~ I~ PThrot Inlet o f >.2 - oz +Nozl End of Ent rance Diffbuser o _ — __ __ _s i 1 2 3 4 5 6 7 8 9 10 I 12 13 14 15 16 -PRESSURE TAPS -*- 6 P j4 AXIAL DISTANCE - INCHES Fig. 14 Normalized axial pressure profiles for several cavitation conditions with water (I inch venturi test section No. II).

-32C24 4Stainless Steel Specimens (? Throat Velocity 64. ft/sec -~ oCC 0a~~~) 0 ~~Cold Water, Air Content 22 Saturated U) _E _ <E E20 oW) 0 o~-> (.> C wIU 6 cU~~~ 0 88___ 5l Hours r ~~~~~~~~40 Hours a. 0 I 627 Hours z4:D 12~ ~~~~~2Ho~ LJ o. I 0 U) o 51.5 Hours J 8 w 2 ~~~~~~~~40 Hours J6 0 ZERO SONIC VISIBLE CAVTO STD CAVTO CAVTOISt NOSE BACK END MARK DEGREE OF CAVITATION (ABITRARY SCALE) > ~~~~~~~~27 Hours z 4 Fig. 15 Mean volume loss per unit area of specimen exposed to w fluid vs. degree of cavitation, for pairs of stainless steel specimens in waterTO STD CAVTO CAVTO I NOSE BACK END MARK DEGREE OF CAVITATION (ABITRARY SCALE)

ttsit n iiations$t and even cavitation to the sPcime, "O8 produce relatively little daages, so that the aximum e prouced by the relatively well-developed conditions of stpan-d ard and first mark These results are shown for a typical case in Fig0 l16 The reason for the inconsistency in this respect between water and mercury will have to await a theoretical analysis of the applicable bubble mechanics problem, which is presently being attempted, 2o Throat Veloi y Throat vselocity for either f luid has been varied eovser a factor of about two0 The effects upon wear have not been as great for any of the materials as expected judging from the observations of previous investigators (for ex ple, the approximate 6th power effect5 observed by Knapp (10) and others (11, 12, 13)) Howevers it is not obvious to the writer that in an arrangement such as the venturi of the present tests, an increase of throat velocity should correspond to a large increase in damage, unless the pressures in the vicinity of the test specimen are also substantially increased, without a corresponding dimunition in the number of bubbleso The fulffi1D ing of these conditions is of course a function of the degree of cavitation used0o 2n any case the mechanism in the present fli arrangement whereby the throat velocity affects daage rate is neither clear nor simpleo Assuing Dlamage p(

-3425 V. =34 ft/sec Mercury - Cold 2 Stainless Steel 2E N Specimens LO E E2O 87 Hours El.? E i. L,. L,. 0) LL.I (n x15 CLLL- w 57 Hours 0 D U) 10 a_ W D CLU -J 5 0 z LU 27 Hours, Q~~~~~~~~ ZERO SONIC VISIBLE CAVTO STD CAVTO CAVTO NOSE BACK ISt MK DEGREE OF CAVITATION (ARBITRARY:SCALE) Fig. 16 Mean volume loss per unit area of specimen exposed to fluid vs. degree of cavitation, for pairs of stainless steel specimens in mercury.

ater Bsho a maidmm velocity-damge ex onent of ihbot 4.9 xo a 12hour tesrt of stainless Qsteel for standard c6avitation, Other exponents obtained are i) 309 for a 305 hour test under the above conditions ii) 2A4 forP a 1l0 hour test of carbon steel iii) l 7 for a 5-minute test of age-hardened aluminum0o It is notesd tht all the presently available tests are for standard cavitationg wh erein the pressure in the vicinity of the test specimen is relatively modert:e for all veloities (Figo l4)o It may well be that greater vlity effects would be observed if a less-developed cavitation condition, with higher and more velocity-dependent test s cimen pressures hLad b@een us~edo The dage-eAponent is plotted in Figo 17 againt test durationo It is noted that a very smooth curve results even though the materials used for each duration (except the two ionger) are differento ethehr or not this result is purely coincidental is not presently knowno Further corroboration from additional tests must be awaitedo The preliminary data from the mercury tests do not a.icate even as great a velocity effect as that found with water, However9 more precise conclusions are not yet possibleO There is no indication in the present tests of the xsateence of a threshold velocity as sometims reported in the past (11, 13, l4) o owever, t he velocities have not been carried lo enough so that definite statements in this rega~d can be made0

-36z. 0X.2.. O ALUMINUM O STAINLESS STEEL A CARBON STEEL EXPONENT, n, FROM FIT TO CURVE DAMAGE It (VELOCITY)n 0 2 4 6 8 10 12 14 TEST DURATION - HOURS Fig. 17 Value of exponent, n, vs. test duration, for several specimen materials in water.

!i imP po'Cible to o~btgin fairly comprehensive results in a, iMnvestilgation similar to that described herein by testig a variety of materials with different fluids, and under different fluid-dynamic conditionsO However, some unifying hypothesis or correlating method is required to bring order out of the otherwise somewhat chaotic conditions which resulto A bubble energy quantum approach was previously suggested by the writer for this purpose. X,1 15)5 A hypothesized bubble energy quantum spectrum is shT n in Fkig. 18 in a shape suggested by the data and the existing knowledge of bubble dynamices The ordinate- n(E), is the nuber of bubbles from those "in the vicinity" of the d age specimes which deliver, upon their collapse, an energy quantum, E, to the surface of the specimeno The abscissa is E, the energy quantum3 The shape of any of the individual curves is' a function of the cavitation condition, which is thus the paraeeter disting u%_,,a ing the different curves Generally, the hypothesized cura es show a maximu anumber of bubbles of relatively small energys and increasingly smaller numbers of more energeti~c bubbles. The curve shape could be quantitatively investigated by using highsp ed motion pictures to delineate the bubble-size spectrumR and by using bubble-dynamics studies to compute the energy quantum delivered to the surface per bubble, under the kn n external pressure conditions (corresponding to the~ cavitation condition and the throat velocityLo As the cavitation condition bec, es less fully,~developds, mA el which is s9omewhat similar has also been previously proposed by ERao and ~hiruvengadam (l 6)o Hogwever, the con.ep~ tioins were independent,.

-38FIRST MARK CAVITATION STANDARD CAVITATION CAVITATION TO NOSE VISIBLE INITIATION RANGE OF DAMAGE - SONIC INITIATION HRESHHOLD NERGY FOR TEST MA.TERIAL E Fig. 18 Hypothesized bubble energy spectra for various cavitation conditions at a constant velocity, for a given material. Presumably, curves at higher velocity are generally similar, but at higher n(E) and E. The quantity n(E) = number of bubbles from those "in vicinity" of damage specimen which deliver an energy quantum E to the surface of the specimen, and E = energy delivered by an individual bubble to the surface of the specimen.

t il y high energy increases, at least until the cavitation condition is reduced to a minimur. This is illustrated by the curves (Fig, 18) the arrangeament of which was guided by the water damage results, showing a maximum of damage for cavi w tation to nose and visible inititiaion and a minimum for sonic initiation and first mark (the condition for zero cavitation would, of course, correspond to the abscissa) It ia issumed that at least a minimum energy quantum9 E, is required to produce any damage, (The surface stress must at least exceed the endurance limit.) Hence a threshold value for E. is shown:, which is obviously a function of the test material. {For harder and/or stronger materials this threshold would move to the right.) o second threshold9 further to the right9 is shown for bsingle-blow damage (corresponding to the crater-type pits))0 Th(^ total ~.mount of damage which occurs ia the form of labs or craters respectively is proportional to the integral under the applicable curve to the right of the applicable threshold. While a set of curves as described above would only be applicable to a given fluid6 and flow geometry9 itoe.o for example6 a venturi of a given design9 this approach appears to the writer to present a potentially extremely useful method for correlating cavita-ti'on damage resultso this general type of behavior is also ev:denced by tests with magnetostriction devices reported bly Nowotny (17) in which the fluid temperature is vtri.ed ove the range between solidification temperature-and *For ex p~e9 si ilar curves for ereacury (rather than water) would probably be deisplaced to the right9 i.e.9 to othe highaer enrgy range o

boiling temperature at a constant pressure. It is found tha': ~,.e dmamge reaches a maximum at some temperature well bel l,; boiling temprgature (as low as 50 C for atmospheric pressurc:~ in smoe of Nohiotny s work)? Presumably at temperatures mlea th o boiling temperature there are very many bubbles but the ol.lapse energy is small, so that the damage is smallo At la8er temperatures, there are fewer but more energetic bubbles with resultant increased damage, As the temperature is reduced still further, the energy of the individual bubbles continueo to increase, but the aumber of bubbles becomes extremely Small, so that the resultant dage decreases8 V Corclusionm A fairly comprehe sive set of cavitation damage data obtained in a flowing system (venturi) over a relatively wide range of applicable parameters has been presentedo Detailed co1clusios rregarding the effects of the variation of the differena parameters are pointed out and discussed in the body of the paper0 Finally, a correlating model to illustrate the effcts of these different parameters is suggested0 It is apparent frma all of the above that much'work yet remains to be accomplished in this field before it will be po sible to predict with any degree of certainty the cavitation damage to be sustained in various fluid flow components under apsplicable operating conditions0'However, in the writergs opinion, this does not ape ar to be an impossible obJectives, nd, in fact, considerable progress toward attaining it is

F o 0o ia mitt et e1o, a "Cavitation Damage Tests;$ wiith Va te; in a Cavitating Venturi", ORtA Technical Report 03424-4`T University of Michigan,, March, 19620 2e Fo Go Hamnitt, et al1, "Observations and Measurements of Flow in a Cavitating Venturi", ORA Technical Reort 03424-5-T, University of Michigan, May, 1962e 3o Vo Fo Cramer and Fo Go Ha itta "'Cavitation Pit DiamiterDIpth Observation for Stainless Steel in Water"' at nrnalReport No, o, 9 OA Project 03424, University of Jichigano 40 Wo J Walsh and P. Go ammitt, "Cavitation and Erosion Damage Measuremets with Bdioisotopes",, TransO ANS, Vlo 4, Noe 2, Nov 9161, po247; to be published, Nuelo 5, H0 N0, Botecher, "Failure:f Metals Due to tCavitation Under xperimental Condition", Tran0 ASNE, Vol 58, 1936, 1PPo 1t3 553604 J M. H1 oUsson, "Fitti Resistance of Metals Under C.vitation Conditions"g Tra~ ASME Vol, 59, 1937, pp. 39-408O 7. M. S. Plesset and A, T. T. Bllis, "On the Mechanisn of Cavitation Damage," Trans. ASMB, Yol 77, No. 7, Oct. 19S55., pp. 1055-1064. 80 A.0 Ellis, "'Production oi Accelerated Cavitaition DaEage by Acoustic Field in Cyliriea Cavity"t, Acoustical Society of America Journal, Volo 27, No0 5 $Spt0 1955 _~~~ g~P~e;~~~i~ 0~~~~~~r~t~

to Cavaiation ErosioY.e 9 ASME Paper No, 62-D39, 19 62, 10, Ro T, Knapp, "Recent Investigations of the Nechanics of Cavitation and Cavitation Damage" 9 Transo ASIE9 eol6 779 1955i pp o 1045-1054o 11,, J lo HRobbs, "Problems of Predicting Cavitation Erosioa From Accelerated Tests"09 ASE Paper N/oo 1-6lRD199 161,, 12? S. L, terr and L. Rosenaberg ""An index of C.%v itatio Erosion by Means of Radioisotopes". Transo As SE9 Vol0o.), 1958, pp. 1308-1314o 13o Jt Z, Lich'tmaEn D.o Bo allas9 C0 K. Chatt'en, "StudPy f Daging Eflfects of Cavitatiion Erosion to thips0 Underwater 8'tr ctures"t Transt ASNE, Vol0 80 19589, pp,, 213>41 3h9 1... J. if "Cavitation Dmage —A RXeview of Present Kniowled;e", Chemistry and XndustLry June 6, 1959, ppo 686-6916 I 5 Fo G. iam ittl "Proposed Model ~or Cavitation Pittii.~~ and Relat ed Thoughts". emo to Fiie Unive?3ioty o Michiga. ORA Project 03424, October 6, 19616 1&.: Raoq NSo and hiruvengadam9 A,,09 J HyEd, o y& iv.9 roct eA Soco CivO Eng,, $pteober 1961, pp, 37t-62 17~, B. lowot y9 Werkstof zerstorun dr. h KavitatioIn V. D, LX Verlag Gimbh, Bertin, IiW 7?, 1942, English?ransla tio published by Edwards Brothers, Ann Arbor. Wc -lihigaiO

Specification of Cavitation Conditions The cavitation conition for all tests is defin~d in'tersa of "degree of cavitation"". referring (except for initiatio)'to the extent of the cavitating region, Oi Sonic Initiation - First sonic nifestation beyond that of single phase flowo This was detected either by ear or electronically using a pie olectrien c~, ~L 4 Te results of these two methos were approx:imatey tbe s eo In all c ses, sonic initiation occurred at a higher throat pressure than visible intiatti o ii) Visible Initiation First appearance of a aore o@r less complete ring of cavitation, This always appares first at the throat exito iti) Cavitation to ose - The approxie ate locati. onf o;h termintion of the cavitatio regios is at th upsr~aaX) nose of t e test Specimen0 iv) Standard The app roiat location of the i;PRi"t.. of the cavitation region is at the center of tue;,,est v) First lark - The approxiate location of th@ te It~tion of the cavitation region is about 1 3/4 iiac 4s from the throat outleto VI) Sond ark 3 Sae a s first mark, bt at S3 ij 43

th trminazt'i@n of the Caviest aion sgi at 0th downstrea wada o f the test specimen, The location of these termination points is shown in Fi3E~ 2o Although the termination is not sharp, the cavital;ion conditions can be quite precisely reproduced0 -43a,