THE UNIVERSITY OF MICHIGAN INDUSTRY PROGRAM OF THE COLLEGE OF ENGINEERING INITIAL PHASES OF DAMAGE TO TEST SPECIMENS IN A CAVITATING VENTURI AS AFFECTED BY FLUID AND MATERIAL PROPERTIES AND CAVITATION REGIME F' Go H nmitt Lo Lo Barinka Mo J. Robinson R. D. Pehlke C. A. Siebert December, 1963 IP-644

ACKNOWLEDGMENTS The authors wish to acknowledge the financial support of NASA for the majority of the work herein discussed. In addition, financial support for some of the tests was provided by Pratt and Whitney Division of United Aircraft Corp. The authors would also like to thank the following research personnel of The University of Michigan for their assistance: K. Aksus, Vo Biss, Ho Ring, E. Rupke, J. Brigham, Y. Sozusen, E. Shippey, W. Smith, R. Ivany. ii

ABSTRACT The detailed characteristics of pitting in the early phases of cavitation damage incurred by test specimens inserted into the diffusing portion of a cavitating venturi are shown pictorially and discussed in terms of degree of cavitation, fluid, material, duration and velocity effects. Quantitative damage results are presented from these tests for mercury and water as test fluids, and for a variety of test materials. Various possible damage correlating parameters are discussed and examined. iii

TABLE OF CONTENTS Page ACKNOWLEDGMENTS............................................... ii ABSTRACT...................................................... iii LIST OF TABLES.......................................... v LIST OF FIGURES................................................ vi I. INTRODUCTION........................................... 1 IIo PITTING CHARACTERISTICS AS AFFECTED BY FLOW, FLUID AND MATERIAL PARAMETERS................................ 8 A. Degree of Cavitation Effects..................... 8 B. Fluid Effects.................................... 18 C. Test Material Effects........................... 24 III. QUANTITATIVE DAMAGE DATA............................... 38 A. Degree of Cavitation Effects..................... 38 Bo Throat Velocity.................................. 40 C. Duration Effects................................. 42 Do Fluid and Material Effects....................... 43 IVo CORRELATION WITH MATERIAL MECHANICAL PROPERTIES....... 50 A. Theoretical Considerations....................... 50 Bo Experimental Results............................. 52 V. CONCLUSIONS............................................ 56 VI. REFERENCES............................................ 58 VIIo APPENDIX - SPECIFICATION OF CAVITATION CONDITIONS...... 59 iv

LIST OF TABLES Table Page I Typical Mechanical Properties of Materials Tested.... 9 II Comparison of Sizes and Percent Circular Pits on Mercury Samples.................................... 21 III Comparison of Sizes and Percent Circular Pits on Water Samples........................................ 22 IV Comparative Size Distribution of Pits in Mercury and Water......................................... 23 V Data on Surface Finish as a Cavitation Damage Parameter............................................ 25 v

LIST OF FIGURES Figure Page 1 Drawing of the Damage Test Venturi Showing Location of Specimens, Specimen Holders, and Cavitation Termination Points................................. 2 2 Schematic Drawing of Overall Mercury Loop Layout....... 3 3 Schematic Drawing of Overall Water Loop Layout......... 4 4 (a) Drawing of Test Specimen, (b) Photograph of Test Specimen, Showing Polished Surface.................... 5 5 Normalized Axial Pressure Profiles for Several Cavitation Conditions with Water (1/2 inch venturi test section No. II)....................................... 7 6 Cavitated Surface of Stainless Steel Specimen Number 84-3 for "Visible Initiation" Cavitation Condition in Mercury at a Throat Velocity of 34 feet/second. Surface was Etched Before Testing. Mag. 37-1/2 X. (a) 6 hours duration, (b) 15 hours duration............ 10 7 Cavitated Surface of Stainless Steel Specimen Number 107-3 for "Nose" Cavitation Condition in Mercury at a Throat Velocity of 34 feet/second. Surface was Etched Before Testingo Mag. 37-1/2 X. (a) 6 hours duration and, (b) 10 hours duration............................ 11 8 Cavitated Surface of Stainless Steel Sample Number 71-3 for "Standard" Cavitation Condition in Mercury at a Throat Velocity of 34 feet/second. Surface was Etched Before Testing. Mago 37-1/2 Xo (a) 6 hours duration and, (b) 10 hours duration................... 12 9 Cavitated Surface of Stainless Steel Sample Number 116-3 for "Back" Cavitation Condition in Mercury at a Throat Velocity of 34 feet/second. Surface was Etched Before Testing. Mag. 37-1/2 X. (a) 6 hours duration and, (b) 10 hours duration................... 13 10 Cavitated Surface of Stainless Steel Specimen Number 84-3 for "Visible Initiation" Cavitation Condition in Mercury at a Throat Velocity of 34 feet/second. Mag. 37-1/2 X. (a) Duration 25 hours, (b) same as (a) but Specimen Number 107-3 and "Nose" Cavitation Condition.............................................. 14 vi

LIST OF FIGURES CONT'D Figure Page 11 (a) Cavitated Surface of Specimen Number 84-3 for "Visible Initiation" in Mercury at a Throat Velocity of 34 feet/second. Mag. 37-1/2. 10 hour duration. (b) Cavitated Surface of Stainless Steel Specimen Number 99-3 for Same Conditions as (a).................. 15 12 Cavitated Surface of Stainless Steel Specimens for "Standard" Cavitation Condition in Water at a Throat Velocity of 64.7 feet/secondo Mag. 37-1/2 X. (a) Sample 4-3 at 191 hours duration and, (b) Sample 5-3 at 191 hours duration.............................. 19 13 Cavitated Surface of Sample Number 55-3 Showing Elongated Pits in the Direction of Flow, Duration 1 Hour, Mag. 120 X. Actual Width of View Shown is Approximately 46.5 mils............................................... 27 14 Cavitation Pit in Columbium-l Zirconium After 40 Hours Exposure Showing Wake Formationo Actual Width of View Shown in Approximately 46.5 mils....................... 27 15 Macrograph of Cavitated Surface of Carbon Steel Samples Numbers 64-1 and 65-1 for "Nose" Cavitation Condition in Mercury at a Throat Velocity of 34 feet/second. Outside Samples are Undamaged New Samples for Comparison. Mag. 4 Xo (a) 4 hours duration and, (b) 7 hours duration................................................... 30 16 Same as Figure 15 Except, (a) 17 hours duration and, (b) 37 hours duration................................ 31 17 Same as Figure 15 Except, (a) 77 hours duration and, (b) 100 hours duration................................. 32 18 Magnified Picture of Damage Cut-Off Region in Figure 15(a), (a) Sample 64-1 and, (b) Sample 65-1o Mag. 25X.. 34 19 Macrograph of Cavitated Surface of Stainless Steel Samples Numbers 47-3 and 48-3 for "Standard " Cavitation Condition in Mercury at a Throat Velocity of 34 feet/ second. Undamaged Samples Included for Comparison. Mag. 4 Xo (a) 17 hours duration and, (b) 47 hours duration... 35 vii

LIST OF FIGURES CONT'D Figure Page 20 Macrograph of Numbered Side of Sample Number Cb-l Zr 20 After 50 hours Exposure to "Nose" Cavitation Condition in Mercury at a Throat Velocity of 48 feet/secondo Mag. 4 X................................ 36 21 Microsection Through Gross Damage Area on Numbered Side of Sample Number Cb-l Zr 20 as Shown in Figure 20. Etched Mago 100 X. Actual Width of View Shown in Approximately 46.5 mils........................... 36 22 Microsection Through Cavitation Pit on Sample Number Cb-Ar 20, After 50 Hours Exposure to Cavitation Showing Material Microstructureo Etched, Mag. 1000 X Actual Width of View Shown in About 5 mils............ 37 23 Actual Specific Volume Loss vs Cavitation Condition for Constant Times for Stainless Steel in Mercury.... 39 24 Actual Specific Volume Loss vs Throat Velocity for Constant Time and Various Cavitation Conditions for Stainless Steel in Mercury........................... 41 25 Mean Depth of Penetration vs Time for Stainless Steel Cavitated in Water................................... 44 26 Mean Depth of Penetration vs Time for Various Materials and Velocities in Mercury.................. 45 27 Actual Specific Volume Loss vs Time for Different Sample Materials in Mercury......................... 47 28 Actual Specific Volume Loss vs Time for Different Sample Materials in Mercury.......................... 48 29 Normalized Cavitation Damage vs Tensile Strength Comparison of Mercury and Water Data................ 53 30 Normalized Cavitation Damage vs Strain Energy to Failure, Comparison of Mercury and Water Data,........ 54 viii

INITIAL PHASES OF DAMAGE TO TEST SPECIMENS IN A CAVITATING VENTURI AS AFFECTED BY FLUID AND MATERIAL PROPERTIES AND CAVITATION REGIME Io INTRODUCTION The cavitation damage investigation being conducted in the authorss laboratory, and some of the preliminary results, have been 1,2 5,4 described in previous papers2 and project reports 4 In brief, cavitation is producedin aventuri test section (Figure 1) located in either of two closed-loop tunnel facilities (Figures 2 and 3). Damage is observed on two (or three) test specimens (Figures 4 and 1), which project into the diffuser section. One surface (indicated in Figure 4) is metallographically polished and monitored for pittingo The damage produced so far in these facilities is generally in its relatively initial phases in that it consists largely of individual small (order of mils diameter and from hundreths to tenths of mils depth) pits or craters which have not yet overlapped. While in some cases somewhat more severe damage has been observed, it is still not of the extensive type sometimes observed in the field after long serviceo Nevertheless, it is felt that a study of these initial phases of damage will lead to an improved understanding of the later phases of damage0 The test specimens utilized are thin, tapered sections oriented parallel to the flow, and projecting away from the wall about 2/3 of the venturi radius at that axial location (Figure 1). *On some of the longer run mercury specimens damage has been obtained which appears very similar to that noted on the interior passages of the centrifugal pump used in the loop after about 600 hours of cavitation4'1 -1

4L25!<^- B | x3 STUD SPECIMEN HOLDER 4 WEAR SPECIMEN AV --- STND TANDARD / CAV TO BACK / CAVTO IT MK ---- 3.015 VISIBLE INITIATION- -- \ rCAV TO 2NDMK SECN B.___5.361" ---— B- B -5.773" | *||l- SECTION A-A 2 - 554 ~ -- ----- 6.522"-,, -- -46.52 ------- — |~~ -|~~~ 1~-14 578".412"- -Z AXIS -786- 1001 --— 1.75" Figure 1. Drawing of the Damage Test Venturi Showing Location of Specimens, Specimen Holders and Cavitation Termination Points.

3_ 176 8- - DRIVE PULLEY BEARING HOUSING STUFFING BOX THROTTLING VALVE COOLING WATER IN THROTTLING VALVE SUMPTO STORAGE TANKTANK STORAGE TANK (OUT OF LOCATION) TO MAIN LOOPTI DRAIN Figure 2. Schematic Drawing of Overall Mercury Loop Layout. Figure 2. Schematic Drawing of Overall Mercury Loop Layout.

/ HIGH PRESSURE TANK CAVITATION TEST VENTURI \ / Y FLOW STRAIGHTENER FLOW DIRECTION \ \-FLOW DIVERTOR GAS ------------------ Uj ---' j \' -\ f^ —- - * --- UJ ----— PRESSURE FLOW MEASURING ORIFICEy / --- _LOW PRESSURE TANK SURGE TANK/' VARIABLE SPEED SSURGE TAENK MOTOR AND MAGNETIC CLUTCH DRIVE UNIT / — SCREEN CNESSUREO "\ ____________ I I y L~~~~~~~~~~ONTKUL; ^ 1256 PUMP Figure 3. Schematic Drawing of Overall Water Loop Layout.

-51r _____ _-I_.... 0.20.0" — S 0.6t00". —0.500"-0.120" o.o" 1___ 0.740"0 0.745" 28~ 40.060" 1 1004 Figure 4(A). Drawing of Test Specimen. Polished Surface 1005 Figure 4(B). Photograph of Test Specimen.

-6By suitable adjustment of the flow-rate and pressures the cavitating region can be caused to terminate visually upstream, upon, or downstream of the test specimenso An arbitrary selection of "degrees of cavitation" has been made according to the apparent termination point. These conditions are listed in the appendix. As the degree of cavitation is adjusted from "Visible Initiation" toward more fully developed conditions, the positive pressure gradient, associated with the cavitation termination region, shifts downstream as indicated in Figure 5, so that the static pressures in the vicinity of the test specimens depend on degree of cavitation and flow-rate (velocity), with the higher pressures being associated with the lesser degrees of cavitation. Previous papers1'2 concerning this investigation have discussed some of the characteristics of the pits which are formed and the probable mechanism involved, and hve presented some preliminary quantitative data on the volume removed by cavitation from the test specimenso The present paper examines and discusses the effects upon the pitting characteristics of cavitating flow regime, fluid properties (mercury and water), and material properties. In addition further quantitative damage data from tests wherein the above parameters were varied over substantial ranges is presented. From this data normalized mean depth of penetration is plotted against the various material mechanical properties in an attempt to show possible correlationso *"Mean depth of penetration" can also be thought of as "specific volume loss", ie0 volume removed per unit area exposed0

1.4 _ _ _ _ —- Reynolds No. = 306x /05 -Sonic _ - - i 1l = 306 x 05-Visible CI.2 -- _ = 3.05x /05-Cav to Nose 1'*<- "\1 -,1;= 3.02 x /05-Standard Z ~S1 Throat Outlet~ ] --- - - ~IV j I IL IC 0~ tO 6 Throatr ltnlet/ -r 2..End Entrance Diffs I 2 3 4 5 6 9 I0 II 12 13 14 15 16 7. —-- PRESSURE TAPS - 7 1225 AXIAL DISTANCE - INCHES Figure 5. Normalized Axial Pressure Profiles for Several Cavitation Conditions with Water (1/2 inch venturi test section No. II).

II. PITTING CHARACTERISTICS AS AFFECTED BY FLOW, FLUID AND MATERIAL PARAMETERS A. Degree of Cavitation Effects Figures 6 through 11 are composite photomicrographs of the initially polished surfaces of five stainless steel* test specimens after exposure to cavitation in mercury at 34 feet/second throat velocity under the cavitation conditions (as defined in the Appendix) of: i) Visible Initiation ii) Cavitation to Nose iii) Standard Cavitation iv) Cavitation to Back for various durationso An examination of these figures shows the following. 1. Pit Size and Number Density Distribution As the degree of cavitation is increased from "Visible Initiation" (lowest degree producing substantial damage in mercury) through the intermediate conditions to "Cavitation to Back", the location of the damaged region tends to shift toward the rear of the specimen, as would be expected. Also the type of pitting varies throughout the damaged region. Toward the upstream end of the region the pits are very small and numerous; toward the downstream end they *The applicable mechanical properties of all materials involved are listed in Table I, -8

TABLE I Typical Mechanical Properties of Materials Tested Tensile 0.2% Yield Hardness Strength Strength % Elong. % Reduction Measured Bending Fatigue Elastic Material Condition psi psi in 2 in, in Area Rockwell B BHN Strength psi Modulus Stainless Steel Annealed 98,000 37,000 55 65 76 140' 35,0006 (300 Series) @ 107 cycles 28 x 10 1010 Carbon Annealed 50,000 30,000 40 71 48 85* 2,000 Steel @ 10 cycles 28 x 10 Aluminum Annealed 13,000 5,000 35 - - 23- 5,500 1100-0 (2SO) @ 10 cycles 10 x 10 Aluminum Age 45,000 40,000 12 - 57 90- 13,5006 6061-T6(61-ST6) Hardened @57 cycles 1x Plexiglass - 10, 445* - 2-7 - - M90-M100 - 6 Rockwell 0.4 x 10 Cb-lZr 20% Cold-Worked 56,000 52,500 6 84 -- Stainless Steel - 95,000 78,000 22 - - 185 - 29 x 10 (410) Aluminum AMS- - 30,500 23,000 2.0 - - 75^- - 10 x 10 4220 T-77 Stellite - -- -- 31 - Measured value or converted from measured value. XX Typical value.

-16become considerably larger and not as numerouso This is particularly true of the small crater-type pits of which there are very many on these specimens, and which are presumably caused by the implosion of single bubbles.* In the middle region of the damaged area the oratertype pits reach a maximum number density, and cover the complete size spectrum. The distributions of pit size and number density is believed to result from the fact that bubbles are most numerous in the upstream portions of the damaged area but, because of the low static pressure in these regions (Figure 5), have a minimum driving force for collapse. In the downstream regions of the damaged area the static pressure is considerably greater because of the action of the diffusing portion of the venturi downstream of substantial two-phase flow (Figure 5), Hence the driving force for collapse in this region is much greater. There is a more violent collapse of these few bubbles which penetrate into the region of higher pressure, so that generally fewer but larger pits of the single-bubble crater type are produced. For "Cavitation to Back" (Figure 9) there is no region of large pits, ioe. the region of higher static pressure is downstream of the specimens entirely. 2, Detailed Flow Effects Although the previous comments have been concerned with the axial distribution of pitting, an examination of any of the composite photomicrographs (Figures 6-11) shows that the distribution also is not uniform in the circumferential direction. The pattern is in the *This presumption is based on strong circumstantial evidence which indicates that all pits, symmetrical or otherwise, are single event failures, but, because of their symmetry, the crater-type are also single blow eventso The local distribution of the larger, irregularly-shaped pits presumed due to many blows is believed more a function of inclusions or other surface defects than of fluid-dynamic parameters,

-17form of a wake, originating along one side of the specimens. This can be seen particularly well in Figure 11, which shows the two specimens tested together in the venturi at "Visible Initiation" after 10 hours duration, As installed in the venturi (Figure 1), the test specimens are separated by an angle of 90, leaving a complimentary angle of 270 o In Figure 11, the specimens are oriented as they would appear to an observer opposite the smaller (900) opening. It is seen that the wakes originate from the flow through the narrower opening, indicating that the relative flow to the test specimens has a cross-wise component in the direction of the wakes, away from the more restricted region of the 90 separation, and generating bubbles as it encounters the sharp lengthwise edge of the test specimens. 3o Quantity of Pitting The quantity of pitting of course increases with test duration as is particularly evident from an examination of Figures 6, 7, 10, and 11, where the development of pitting for "Visible Initiation" and "Cavitation to Nose" can be compared for various durations up to 25 hours, Precise quantitative results from weight measurements, showing the effects of duration and degree of cavitation, will be discussed in a later sectiono However, for the moment it is noted that the damage resulting from these two degrees of cavitation is quite similar4 It is believed that in both cases the damage is caused mainly by bubbles generated in local cavitation from the test specimens themselves, *Venturis with three equally spaced specimens have been used in the more recent water tests,

-18and that the main influence of the increase of degree of cavitation in this range is simply a reduction in the local static pressure affecting the local cavitationo Bo Fluid Effects lo Type of Pitting Some information on the effects of fluid parameters, when they are varied over a very wide range, on pitting characteristics is afforded by a comparison of the water and mercury resultso Since, for the test materials used, corrosion was not significant in either fluid, it is believed that the differences noted are mainly due to the large density and vapor pressure differences between the fluids (all tests are at room temperature)o Figure 12 is a composite photomicrograph of two stainless steel specimens exposed to "Standard Cavitation" in water, at a throat velocity of 65 feet/second, for a duration of 191 hourso The mean depth of penetration is estimated to be about equal that of the specimen tested in mercury under the same cavitation condition for a duration of 10 hours (Figure 8)0 As will be noted, the damage on the water specimen is composed predominently of relatively large pits of irregular outer contours, presumably the result of fatigue failures due to repeated (relatively weak) blows, rather than the smaller, apparently single-blow, crater-type pits encountered on the mercury specimens0 This apparent predominence of more powerful blows in the mercury test would of course be expected both from the density and vapor pressure (less cushioning effect) differenceso

-202o Shape, Sizee, and Spatial Distribution. of Pits Detailed pit tabulations have been made on a pair of stainless steel specimens exposed to 100 hours of "Visible Initiation" cavitation in mercury at 54 feet/second throat velocityo These are presented in Table II and can be compared with similar observations from stainless steel specimens cavitated (at various degrees of cavitation and durations) in water at 65 feet/second throat velocity presented in Table III (reproduced from reference 1 for convience)o Comparison shows that in all sizes the proportion of approximately symmetrical (ioeo crater-type) pits in the mercury tests is greater (order of 25-30% ) than that in the water tests The tabulated proportion of craters is of the order of 15% but it is believed that the actual proportion is greater because it is likely that many of the somewhat irregular pits, especially in the smaller sizes, are also the result of single blowso Thus the trend noted in the previously discussed photomicrographs is somewhat substantiated by these observations The proportion of the total number of pits found in each size category is shown in Table IV (exclusive of the "small" category which was not counted in the mercury tests)o It is noted that the water specimens have a larger proportion than do the mercury specimens in the larger size categorieso Assuming that the larger pits are generally of the multi-blow fatigue type, this is consistent with the previous observationso

TABLE II Comparison of Sizes and Percent Circular Pits On Mercury Samples TABLE II-A % Circular Pits of Total Pit Size Categories (mils) VVL VL L S Sample Sample Cavitation Velocity Duration Observed 5.52 Dp<11.04 2.76<-Dp<5.52 1.15 < Dp<2.76 0.46<Dp<1.15 No. Position Condition (fps) (hrs.) Area % Cir. Total % Cir. Total % Cir. Total % Cir. Total 64-3 Front Nose 34.8 100 Polished 7.7 26 13.8 29 12.8 94 -- Surface Numbered 16.1 31 18.75 32 25 200 - Side Opposite 8.9 45 25.7 39 16.9 130 - Side Sub-Total 11.8 76 22.6 71 21.8 330 - (Both Sides) Sub-Total (All Surfaces) 10.8 102 20 100 19.8 24 -- 63-3 Back Nose 34.8 100 Polished 7.4 27 16.1 31 19.2 73 - Surface Numbered Side 13.33 15 15.0 40 15.3 150 -- Opposite Side 10.0 20 22.2 45 18.2 110 - Sub-Total (Both Sides) 1.4 (Both Sides) 11 35 18.8 85 16.5 260 -- Sub-Total (AllSurfaces) 9.7 62 18.1 116 17.1 333 - Total (64-3 & 63-3) 10.4 164 19.0 216 18.6 757 - % of Total (exclusive of 14.4 19.0 66.6 small) TABLE II-B Total Pits Per In2 HR. Sample Sample Cavitation Velocity Duration Observed VVL VL L S No. Position Condition (fps) (hrs.) Area Pol. Sides Pol. Sides Pol. Sides Pol. Sides 64-3 Front Nose 34.8 100 --- 7.0 2.54 7.8 2.38 25.3 11.05 -- 63-3 Back Nose 34.8 100 --- 7.3 1.17 8.3 2.84 19.6 8.7 - Where: Pol = Polished Surface and Sides = Total for Both Sides

-22TABLE III PIT COUNT TABULATIONS* Pit Count and Various Sizes Pit Size (mils) WL 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 lo.05 81 15.1 325 Standard Cavitation Opposite Side 18.75 lo 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 lo.l 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 hrs. Total All 0 53 5.4 74 7.0 158 14.0 050 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 1226

-23TABLE IV Comparative Size Distribution of Pits Percent of Total Pits (Exclusive of "Small") in Category VVL VL L Mercury 14o4 19o0 6606 Water 17o0 2305 59o4

-24Co Test Material Effects 1. Effects of Grain Boundaries on Pitting Location The stainless steel specimen shown in Figure 8, cavitated in mercury, was etched prior to testing to determine the existence of any possible relation between the grain boundaries and the location of pits and their shapes0 As expected the crater-type pits seem to fall completely at random, their location not influenced by the grain configuration~ Sometimes, however, the profiles of the irregular-shaped pits follow the grain boundaries for part of their contour, indicating that the location and shape of this type of pit is a function of the local surface material properties in addition to the flow parameterso 20 Effect of Surface Finish To investigate the effect of surface finish, two sets of stainless steel specimens were prepared for mercury tests; one set was polished on all wetted surfaces (I 2.0 microinch rms), and the other remained in the machined condition (20 to 25 microinch rms) on all wetted surfaceso Each set was tested for 50 hours duration at "Standard Cavitation" with 54 feet/second throat velocityo The results, listed in Table V, are somewhat surprising in that the completely unpolished specimens showed the smallest weight loss, the completely polished specimens the highest, and a set of regularly prepared specimens (polished only on the usual "polished surface", which is about 1/9 the wetted surface) tested under the same conditions, was intermediate. While the spread in volume loss is substantial (about 50% in the extreme case), it is not necessarily beyond the possible scatter due to inhomogeneity of material, dimensional differences between specimens, minor flow differences, etco

-25TABLE V Data On Surface Finish As A Cavitation Damage Parameter Test Conditions: Standard Cavitation 2,5" Flow Rate (34 fpso) Stainless Steel Samples Total Weight Loss Ratio To Sample Condition For 2 Samples (mg.) Unpolished Regular (47 & 48) 4o27 lo20: 1 Polished Sides (78 & 112) 5046 1o53: 1 Unpolished (81 & 82) 3o56 1: 1 Rolled iL To Flow (11-120, 12-120) 4o85 1o36: 1 Note: Ao Surface finish of sides of regular samples (47 & 48) is between 20 - 25A in (rms)o Bo Surface finish of polished surface of regular samples is about 2 / ino (rms)o C. Surface finish on 78 & 112 was about 2 / in. (rms) on both sides and polished surfaceo Do Surface finish on 81 & 82 was about 35 / ino (rms) on "unpolished" surface and about 20 - 25 M in. (rms) on sideso Eo Surface finish on 11-120 & 12-120 was similar to regular samples (ioeo, 47 & 48)o

-26It thus appears that the metallographic finish given the monitored surfaces of the test specimens in these tests does not significantly protect them from cavitation damage; in fact the available evidence shows the oppositeo However, the machine finish used is extremely smooth compared e.go with an ordinary cast pump impeller, so that no conclusions on the effect of rough surfaces upon cavitation damage can be drawn from the present tests. 3. Effect of Direction of Rolling In some of the tests, it was noted that the damage often took the form of pits elongated in the direction of flow, ioeo parallel to the long direction of the test specimens (Figure 13) which, in the ordinary specimens, was also the direction of rolling, Hence it was suspected that these pits might represent the removal of inclusions, which had been squeezed by rolling to the elongated profile of the pits. To check this supposition, a set of stainless steel specimens was fabricated in such a manner that the direction of rolling was normal to the polished surface, ioe. normal to the direction of flow. These were tested for 50 hours duration under the same conditions used for the polished and unpolished specimen tests described in the last section, The results are also listed in Table V. The volume loss was somewhat increased (13o6%) over the standard testo This relatively small change is not necessarily beyond data scatter so that it cannot be concluded that the direction of rolling has any significant effect upon volume loss, No profusion of elongated pits was produced,

H A J 0 U H Os C ~ P5* a3 Cl )- -- 0 ~,,......:.:.:.:.:....... U 0 9; 000; 0 s > 4V >,z4 jEjf 000 c~.......i;';~:"::': ":::::::':'";::::::::..... C~&* iiiiili-i-iiii#iliii':i:~i~iiff~~Z:...... CH ~ ~ ~ ~ ~ ~~~~ 0 x: -~i::: 00 0~~~~~~~~~~~~~~~~~~~~~~~~~~~3C 0) 4

-284. Specific Material Effects ao Carbon Steel Tests in Mercury It was shown in a previous paper that the general form of pitting encountered with water on carbon steel was similar to that observed on stainless steel, providing the test was short enough so that corrosive effects did not become significant. The present tests show that the same is true with mercury, although corrosion is not so significantly involved. However, the damage rate of carbon steel with either fluid is considerably greater than that of stainless steel, so that in the case of the mercury tests, where damage obtained in a given duration is of the order of 100 times that with water, effects of a new kind, as described below, have been observed with carbon steel. In most tests conducted during the present investigation, it has been observed that the location of pits within the damaged region is essentially random (within certain limitations as, for example, that larger pits form in regions of higher static pressure, etco)o However, in one extended carbon steel test it has been very clearly demonstrated that the previously incurred damage sometimes influences the location of subsequent attack0 It is believed that in this case this is due almost entirely to a shifting of the cavitation region due to the change in surface * A mercury venturi provides an intensified cavitation damage test in a flowing system, perhaps intermediate in degree of intensification between a water venturi and a water magnetostriction unitO **Very large pits (Figure 14) sometimes form a wake influencing the distribution of subsequent pitso (Reprinted from refso 1 and 4 for convenience )

-29condition, rather than to any required incubation or preconditioning period for portions of the surface, since the demarcation between damaged and non-damaged regions is at all times extremely sharp, and even the single-blow crater-type pits are generally absent from an undamaged portion until the demarcation line has passed. This behavior is illustrated in a series of low-magnification photographs (Figures 15 - 17), showing two damaged carbon steel specimens, held between two unused specimens for comparison, after exposure to "Cavitation to Nose" in mercury at 34 feet/second for various durations, starting with 4 hours and proceeding to 77 hourso In the first photograph (Figure 15), the orientation of the specimens has been inadvertently reversed, so that the adjacent sides correspond to the 270~ separation in the venturio It was previously shown (Figure 7) that the damage for this particular cavitation condition follows a wake pattern originating in the 900 gap, and angling across the specimens from this region, It is believed that this flow pattern results in the initial damage distribution, covering an area similar to that' shown in Figure 7o When the specimens were exposed to further cavitation the damaged area continued to spread in the upstream direction as indicated in the succeeding photographs (for which the orientation is correct)o The wake pattern is still apparent, but the metal on the upstream side of the damaged region has

-30(a) (b) Figure 15. Macrograph of Cavitated Surface of Carbon Steel Samples Numbers 64-1 and 65-1 for "Nose" Cavitation Condition in Mercury at Throat Velocity of 34 feet/second. Outside Samples are Undamaged New Samples for Comparison. Mag. 4X. (a) 4 Hours (b) 7 Hours.

-31(a) (b) Figure 16. Same as Figure 52 except, (a) 17 hours duration, and (b) 37 hours duration.

-32i~~~~i- ~i _::S:.t:E,.::......................... R R @_ S _~~~~~~~~~~~~~~~~~~~~~~~........... (a)............... iiil il~ ~B j?!~ z ii itiV0000-.000-<=i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.-L:_......... iii~isiiu.i~~i im-0!~~~~~~~~~~~~~~~~~~~~~. | ~~~~~~~~~~~~~~~~~~~~~~~~:i:S::', R:7.' < ~ —':~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...:.... ili - l -~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~:: ii/~'~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~i &. -'li lE.... tt; L.R'0': 0 a R,RR.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...... (b) Figure 17. Same as Figure 52 except, (a) 77 hours duration, and (b) i00 hours duration.

-33now been attacked. However, the demarcation between damaged and undamaged areas is still extremely sharp, with the damaged portion having incurred several layers of pitting, while the undamaged region is essentially untouched. The region of the demarcation is shown with higher magnification and more detail in Figure 18o This occurrance of a sharp damage cut-off, which advances into the undamaged material, is not unusual for heavily damaged specimens, having been previously observed eogo in tests with a rotating disc apparatus,5,6 Also somewhat similar behavior has been observed to a much lesser extent in the present tests with stainless steel (Figure 19)o bo Columbium-Zirconium Alloy Tests in Mercury Damage of a sort not observed in the tests on the other materials has occurred in tests with mercury on Cb-1 Zr alloy, in which there was an eventual** macroscopic fatigue-type failure, in which a thin surface layer covering an area of the order of Oo015 square inches peeled-off to a depth of about 2 to 3 mils (Figure 20)> Figure 21 is a section through this region showing the continued propagation of cracks at the ends of the regiono The microstructure of the material is shown in Figure 22. *This type of failure is believed due to the cold-worked condition of the material, which would not be typical of high temperature applicationso **Prior to this the pitting was generally similar to that observed on the steels,

-34*Iiiiiiiiiiiiiiiiii^iii (a) (b) Figure 18. Magnified Picture of Damage Cut-off Region in Figure 52 (a). (a) Sample 64-1, and (b) Sample 65-1. Mag. 25X.:R~~z'~~~~~~~~e 1,Ka~et?~u- s'Zsa&eCto~Elo r ~~'

-35-~~~~~~~~~~~~~~~~~~~~7 0 (a) i i | z _l~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~...... _;~~~~~~~~~~~~~~.....-....... l _ *E~~~~~~~~~~~~~~~~~~~~~~~~~~~.........................11.....I I. 1~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~........_..._..!...s s| _ X I E _ |, 11 - _ _ — I _ _~~~~~~~~~~~~~~~~~~~~~~~~~~~............ _ l i l - E g i B B llg l_ 1 I~~~~~~~~~~~~~~~~~~~~~~~~~~........... l | 1|E i! 11~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~....... _ 1 11 - -.0'''z' _ z _ I l~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~................. _ * | | _:-: iii-Eiigii00:i~~~~~~~iii~~ii-:igigi.:.igiiigii:: i i 0::- I......I..I.*.._...I _ | | | _. E E EEE~~~~ikEVEEEEEE7000T'-':.X.-gEEE' i:::j::000020000000 7il:.................. - I 11 - __1 I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~............. * I 11- -I I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~............ - 111 - _ -I I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~........... * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.................................... __~~~~~~~~~~~~~~~~~~~~~~~~............. V -—::"::::::::SS —::.:::::'-::SSS::S Ei::".............................................:.: ": h.;;;-;-~~~~~.....;-.X...........'l00000000000000000000000000.0-0 —.g...gg:'..,.......'.-............-.'......-..............' - X: 000000000000000000................. I xxxSx: ]:...000...........00000...................... - ~~~~~~~~~~~................... -D~~~~~~~~~~~..... C B ] %.................. 1 S e e........................ S r 3 s t. / e c th~~awmaged:eza:-eples >1:>nt^1*t1(ied f~o................................................... ~ z d (} ) ~t oz~ s d r~ t~ o

-36Figure 20. Macrograph of Numbered Side of Sample Number Cb-lZr 20 After 50 Hours Exposure to "Nose" Cavitation Condition in Mercury at a Throat Velocity of 48 feet/second. Mag. 4X. Figure 21. Microsection Through Gross Damage Area on Numbered Side of Sample Number Cb-lZr 20 as Shown in Figure 58. Etched. Magnification 100X,

-37lvB Figure 22. Microsection Through Cavitation Pit Labeled in Figure 68 After 50 Hours Exposure to Cavitation. Etched. Mag. lOQOX.!??!iz~i;iiiil iiiiiiiii~~~~~~~~~i~~~iii ~ ~ ~-iii-iiiill~ ii:i~~~~~~~~~~~~~~~~ iiiii- iiiiiiii. iiiiiiiiiiiii!iiiiiiii~ Figure 22~~~~~~~~~~~~~~~~~~~~~~~~~~~~~. M i rsi o T hrug Cavi t:ation::~ PitLabele Cavitation. Etched. Mag. 1000X.~~~i

III. QUANTITATIVE DAMAGE DATA A great deal of quantitative damage data has been accumulated with both water and mercury (room temperature in both cases) as test fluids, and stainless steel, carbon steel, aluminum, various brasses, plexiglass, Cb-l Zr, and other refractory alloys as test materials during the course of the present investigation;34 and the work is still in progresso While it is not possible to report these results in detail in the present paper, the highlights will be summarized. Ao Degree of Cavitation Effects The change of cavitation condition from initiation towards increased cavitation results in an increase in the number of bubbles in the vicinity of the test specimens, but a decrease in the collapsing pressure and hence energy of collapseo As might be expected under these conditions the damage goes through a maximum as the degree of cavitation is increased. With water this maximum occurs for "Visible Initiation", whereas, for mercury it occurs for the increased condition of "Standard Cavitation"o The preliminary results indicating the latter, previously published, have since been confirmed by much more comprehensive data (Figure 25). The reason for the differences between water and mercury in this regard is not presently known, but the situation is indicative of the fact that the cavitation phenomenon does not in general scale according to the simple similarity lawso *While the volume losses for water are calculated from pit counts and those for mercury from weight measurements, so that the absolute magnitudes are not entirely consistent, the data sets for each fluid should be internally consistent and hence can be used as indicated~ -38

-39240 MATERIAL: STAINLESS STEEL- 302 220 FLUID: MERCURY VELOCITY:0 34 FT/SEC _ 64 FT/ SEC 200 100 HRS. 180 160 -0 "= 140 U) U) 0 -J 120 w > 100 u J~ ~ )50 HRS. a. 80 60 40 20 ---.. -- -- ^ —_ —,, -- -- -- ioo H R S. VIS. NOSE STND. BACK CAVITATION CONDITION 1015 Figure 23. Actual Specific Volume Loss vs. Cavitation Condition for Constant Times for Stainless Steel in Mercury.

-4oB, Throat Velocity Previous investigations of velocity effects upon cavitation damage have generally shown that damage increases very rapidly with velocity, approximately proportional to the 6th power of the velocity.7 Although no theoretical justification of such a relation is available, it has been roughly confirmed in several tests: 5,6 i) Rotating disc with through-holes as cavitation inducers.o ii) Liquid jet impinging upon rotating test specimen8 iii) Cavitation upon an ogive in a water tunnel. However, the relations between the force-time loading on the surface and the volume removed, and again between the main stream velocity and the surface loading generated by collapsing bubbles is far from clear for any of the above devices, or for the venturi of the present tests. However, there is a basic difference between the present tests and those referenced above, in that in all the previous cases (but not in the present venturi) the geometry is such that the size of the damaged region changes with velocity. With the relatively small test specimens used in the present venturi this is not the caseo Velocity effects in the present tests have been generally much less* than previously reportedo5'6'7'8 In fact for well-developed cavitation the damage exhibits a maximum at an intermediate velocity and tends to decrease at higher velocity. See Figure 24 for mercury tests on stainless steel wherein the velocity is varied over The velocity exponents for the present tests, positive or negative, generally lie between 1 and 2. **An example of a decrease of damage with increased velocity is also reported for a rotating disc facility,9

-41240 - 220 200 100 HRS.180 100 HRS. 160 -:,, 140 o 120 >J - ___ __ _ _ o 100 IL u 80 ven A 50 HRS. \ Q3 50 HRS. 60 M — ATERIAL: STAINLESS STEEL-302"FLUID: MERCURY r VISIBLE INITIATION - CAVITATION 0 CAVITATION TO NOSE 100 HRS. 40 - CONDITION: 0 STANDARD CAVITATION - | - l ^ CAVITATION TO BACK. 20 _______ ^___________________<^\.^A50 HRS.lI -__ ___ _ — == \ _|.=0 50 HRS. __ -—' —---- - - 30 HRS.0 10 20 30 40 50 60 70 80 THROAT VELOCITY, FT/SEC 1022 Figure 24. Actual Specific Volume Loss vs. Throat Velocity for Constant Time and Various Cavitation Conditions for Stainless Steel in Mercury.

-42about a 2:1 range for three degrees of cavitation. Damage decreases uniformly with velocity for the most developed cavitation ("Cavitation to Back"), shows a maximum at an intermediate velocity for "Standard Cavitation" and increases uniformly with velocity for "Cavitation to Noseo" Similar behavior has also been observed with the other materials tested in mercuryo So far the water tests are not sufficiently comprehensive for definite conclusions. The dependence of velocity effect on degree of cavitation can be explained by the fact that the static pressure in the vicinity of the test specimens is a strong function of throat velocity for the initiation conditions (approaching the squared relation of single-phase flow), but does not vary substantially with throat velocity in the area of the test specimens for well-developed cavitation (io.e two phase flow), remaining close to vapor pressure for all throat velocities. Thus the violence of collapse on the test specimens is not substantially a function of throat velocity for well-developed cavitation. It appears that the above effect is coupled with another which somehow reduces either the number of bubbles in the critical region, ioe,, adjacent to the test specimens, as the throat velocity is increased, or the static pressure in that region. In. either case, further evidence of the lack of simple scaling of cavitating flow regimes is afforded, Co Duration Effects* Virtually all the curves of volume loss vs. test duration from the present investigation show, after an initial high damage rate, *Op Citp footnote p1 12

-43a leveling-off period, followed eventually by a second period of high damage rate. The longest duration water-test curves presently available (Figure 25) are of this type~ In a sense, the mercury tests might be considered as accelerated (or intensified) water tests, in that for the same duration a much greater quantity of damage is incurred, so that it might be presumed that the behavior observed in mercury over a relatively moderate test time may illustrate the behavior to be expected in a much longer water test. If this is the case,** it is apparent that the relatively simple pattern so far observed in water will become more complex as duration is further increased. Figure 26 shows the presently available longduration effects with mercury, at two throat velocities for stainless steel, and a single velocity for carbon steelo In all cases there was heavy pitting during the initial observation period, although in some cases the weight loss during this period was not large, Since the mercury curves become considerably more complex than those for water, it is not clear how best to extrapolate to longer durations. Do Fluid and Material Effects lo Fluid Effects A direct comparison between the damage quantities attributable to mercury and water "other things being equal" is difficult because of the ambiguity of that condition, and the difficulty of achieving true similarity, For example, as previously discussed, the maximum **It is obvious that it is not entirely the case, since these are differences in type of pitting between mercury and water, as previously discussed0

ZOO zoo --------------------- CAVITATING FLUID: WATER lllX CAVITATION CONDITION: STANDARD 6.00 VELOCITY: 64.7 FT/SEC - en SPECIMEN MATERIAL' STAINLESS STEEL 5 SPECIMEN NUMBERS: 4-3 8 5-3 --- z- / _oo ~~z~~~ ~/ 4.00 I- - DURATION. HOURS o 3.00 1.00 0 20 40 60 80 100 120 140 160 180 200 DURATION, HOURS Figure 25. Mean Depth of Penetration vs. Time for Stainless Steel Cavitated in Water.

11JJt1175tt111 700 PLEXIGLAS AT 34 FT/SEC 600 - 0 I I I /-STAINLESS STEEL AT 64 FT/SEC z 500 -- z /-Cb-Zr AT 34 FT/SEC w - / P 400 nzLL / \-STAINLESS STEEL AT 34 FT/SEC 1 300 -- I z / w 200 -_ — - - - -- - — ___ w /-Zr / ^J^^^ CARBON STEEL AT 34 FT/SEC 1200 CAVITATING FLUID:MERCURY CAVITATION CONDITION: STANDARD 0 00 0 50 100 150 200 250 300 350 400 450 500 DURATION, HOURS Figure 26. Mean Depth of Penetration vs. Time for Various Materials and Velocities in Mercury.

-46damage is obtained for different degrees of cavitation for the two fluidso However, rough comparisons have been made4 between the maximum damage conditionso The making of such comparisons is further complicated by the fact that most of the quantitative damage results for water are based on calculation* from pit tabulations4'10 (weight losses are usually insufficient for significant measurement), whereas those from mercury tests are based on direct weight measurement. The details of the estimating procedure are given in reference 4. For two sets of tests for which comparisons were possible the ratios of mercury to water damage were 4705 and 135, giving a numerical average of 91. Hence a ratio of approximately 100 seems indicated. 20 Material Effects In a previous paper comparative data in water between aluminum, carbon steel, stainless steel, and plexiglas were given. The volume loss of the aluminum was very great compared to the steels; carbon steel showed about twice the volume loss of stainless steel in short tests before corrosion became significant; and plexiglas appeared to show less damage than stainless steel,** Comparative results are shown for the mercury tests (Figures 27 and 28) between carbon steel, stainless steel, Cb-l Zr, and plexiglas for two degrees of cavitationo It is believed that corrosion is not *In spite of considerable care in measurements10, the volume loss computed from pit counts was always considerably less than that derived from direct weight loss measuremento It is felt that a significant volume removal in the form of a thin uniform layer, by another mechanism other than the formation of discreet pits is involvedo **Based on pit-count datao Since pit observation on the transparant plexiglas was difficult, the quantitative data on plexiglas is uncertain. However, the over-all conclusion is confirmed by lack of pitting in the venturi walls themselves after very long exposureo ***Based on direct weight measuremento

-47600 CAVITATION CONDITION: CAVITATION TO NOSE 540 FLUID: MERCURY / VELOCITY:34 FT/SEC X O STAINLESS STEEL -302 _ MATERIAL X CARBON STEEL - 1010 0O COLUMBIUM - 1% ZIRCONIUM 480 Lr SEL -30 420 z J.____..... 360 - --- - --- w ) 300 O 240 w 180 0 10 20 30 40 50 60 7 80 90 100 TEST DURATION, HR 1025 Figure 27. Actual Specific Volume Loss vs. Time for Different Sample Materials in Mercury.

-481300 1200 - CAVITATION CONDITION: STANDARD FLUID: MERCURY VELOCITY: 34 FT/SEC/ (0 STAINLESS STEEL -302 MATERIAL X CARBON STEEL- 1010 1100 0 COLUMBIUM - I% ZIRCONIUM Il PLEXIGLAS 1000 2200 900 I x Nz0 NZ /5^~ 8~ ~2100 — ~ - I / )Z 800 /) -'2000 Cj 0 w -J 700 w _. o j600~~~~~~~~~~~ i c~1900 o0 600 2' L) L LLI _ __ __3__ ____ _0 500 400 ^^T -— Al^^ —,00 -'-' "- - 0 10 20 30 40 50 60 70 80 90 100 TEST DURATION., HR 1026 Figure 28. Actual Specific Volume Loss vs. Time for Different Sample Materials in Mercury.

-49important for any of these combinations. However, there is some dependence of relative resistance of the different materials on degree of cavitation, but for both cavitation conditions, stainless steel shows least damage and plexiglas most.

IVo CORRELATION WITH MATERIAL MECHANICAL PROPERTIES A. Theoretical Considerations An over-all objective of cavitation damage research is the determination of a grouping of material, fluid, and flow parameters which could be used to correlate cavitation damage in general, and thus allow the prediction of damage in advance, For the present this objective appears unattainable in the completely general caseo However, under certain restricting assumptions, it may be possible to determine suitable correlating parameterso In this connection it is usually assumed that: i) consideration is to be limited to a single type of cavitating flow regime, and ii) only mechanical damage effects are significant (or, alternatively, only chemical effects)o These assumptions may be applied to the present venturi tests, where.h4 the fluid-material combinations are such that corrosive effects are generally insignificant (except perhaps' for the carbon steel in water combination) It has been found by many previous investigators that increased tensile strength, surface hardness, yield strength, endurance limit, etc. will usually result in reduced cavitation damage, under the above assumptions, although there are usually numerous exceptions. More recently it has been emphasized that strain energyl may be of paramomnt importance. -50

-51It has been previously stated in connection with the present investigation that the pits appear to be either craters resulting from single-bubble implosions or fatigue-failures resulting from numerous blows, that the crater-type of pit becomes relatively more numerous for a high density fluid such as mercury, and that neither mechanism appears negligible. It is believed that the parameter groupings which would correlate these two types of damage are different. It seems likely that strain energy to failure may be of predominent importance in determining resistance to the formation of crater-type pits: however, resistance to conventional fatigue failure would certainly also involve the endurance limit** A very large number of blows which did not create stresses in excess of the endurance limit would not cause failure or any permanent effect, even for a very brittle material. In addition to the above, there are numerous other parameters and related effects which may require consideration. For example, there is the question of coupling between the relatively short range forces produced by a bubble implosion and the material surface with its ability to deform elastically out of effective rangeo It is believed that this mechanism may be important in the behavior of plexiglas, which was almost immune from damage in water, but was very susceptible under the more intense forces created in cavitating mercuryo *These statements are meant to apply only to cavitation damage in relatively initial phases as observed in this investigation. As already mentioned, macroscopic fatigue failures have also been observed eventually (Cb - 1% Zr alloy). **Or the ultimate tensile strength, to which it is approximately proportional,

-52Some other considerations which may be important are: i) recovery time required by material between blows, since these may be imposed at a relatively high frequency, ii) ratio of dynamic to static mechanical properties of material, and iii) corrosivity of material in fluid in question, which, although negligible under static conditions may become important under cavitating conditionso Bo Experimental Results All the applicable quantitative experimental data, available to date from the present investigation, has been normalized to stainless 4 steel under the same duration and flow conditions, These individual data points for each material and fluid have then been averaged in such a manner that the averages were weighted according to quantity of damage in terms of mean depth. of penetrationo The normalized values were then plotted against ultimate strength, yield strength, Brinell hardness, and strain energy to failure, respectively. These plots are very preliminary, since much data is yet to be gathered0 However, the plots show the large magnitude of data scatter, and indicate that, of the parameters used, the most successful correlation is achieved with strain energy to failureo Since hardness and yield strength are more or less proportional to ultimate strength for the materials used, only the plots for ultimate strength and strain energy to failure are presented (Figures 29 and 30)o The correlation of averaged points to strain energy is quite good

-532.0 TEST FLUID: MERCURY a WATER 1.8 ~ AVERAGED AVERAGE POINT FOR WATER | AVERAGED AVERAGE POINT FOR MERCURY 1.6 VERY HIGH =-ALUMINUM- 6061-T6:. I I AT 15 HOURS,4.42 1.4 uI' |_ 1.2 0 L -PLEXIGLAS wI CARBON STEEL- I NORMALIZING POINT < 1.0 - --- - - - \ - - - FOR STAINLESS STEEL Cb-I Zr(ANNEALED) I II I I I I I1 r-Cb-IZr(20%COLD WORKED).8 — %- W t __ _ ___ _I______ _ >Q ~4 E__ _ liii __ -= __ __ N.6 J.4.2 _ — ALUMINUM-6061-T6 AT 50 HOURS Itj|0~~~~~~~~ LL III IIT T 1035 0 I 0 20 40 60 80 100 120 140 -3 TENSILE STRENGTH, PSI xO 10 Figure 29. Normalized Cavitation Damage vs. Tensile Strength Comparison of Mercury and Water Data.

-542.0 TEST FLUID: MERCURY 8 WATER 1.8 - AVERAGED AVERAGE POINT FOR WATER [ AVERAGED AVERAGE POINT FOR MERCURY -VERY HIGH 1.6 -/ ALUMINUM-6061-T6 E/ Ar AT 15 HOURS,4.42 1.4 - _ ___ -PLEXIGLAS 02 1.2. Cb-IZr(20/ COLD WORKED) E, -Cb-I Zr(ANNEALED) _ lJ. CARBON STEEL -NORMALIZING POINT < 1.0 - — FOR STAINLESS STEEL 0 L.6 N.4 /ALUMINUM 6061-T6.2 - 6 7 AT 50 HOURS 1036 0 6I 0 1000 2000 3000 4000 5000 6000 7000 STRAIN ENERGY TO FAILURE, FT- LBF/IN Figure 30. Normalized Cavitation Damage vs. Strain Energy to Failure, Comparison of Mercury and Water Data.

-55except for plexiglas in water, where as previously mentioned, mechanisms not considered in the choice of correlating parameter are believed involved, However, the scatter of individual points is very large, so that a prediction of damage a priori from such a correlation might not be meaningfulo It is believed that a good correlation can only be achieved in terms of more complex groupings of material properties which consider more completely the mechanisms involved,

Vo CONCLUSIONS The following conclusions appear to be the most significant: i) The detailed characteristics of the pitting observed on test specimens inserted into the diffusing portion of a cavitating venturi, depend significantly upon degree of cavitation, fluid, and materialo The observed trends are generally explainable in terms of the hypothesized mechanismso ii) While the distribution of crater-type pits is essentially locally random in the early phases of cavitation damage, it becomes a function of the previous damage as this becomes sufficient to cause perturbations in the flow pattern. The local distribution of irregularly shaped pits, on the other hand, is often a function of surface imperfections, grain boundaries, etco iii) Although the early damage is composed of individual, generally non-overlapping pits, gross macroscopic fatigue-failure has eventually been observed for at least one material, iv) The effects of throat velocity upon quantity of damage in the cavitating venturi is much less than previously supposedo In some cases damage decreases with an increase of velocityo v) The damage with mercury in the cavitating venturi is the order of 100 times greater than that with water for the same range of velocity, so that mercury tests provide an accelerated cavitation test, perhaps midway between a conventional flowing system and an ultrasonic deviceo -56

-57vi) Of the conventional mechanical material properties, damage from the present tests is best correlated by strain energy to failureo However, the scatter of individual data points about the best curve is very large.

VIo REFERENCES 1. Hammitt, F. Go "Observations on Cavitation Damage in a Flowing System." Trans. ASME, Jo Basic Eng,, Series D, V. 85, No. 3, September 1963, pp. 347-359o 2. Hammitt, F. Go, et al. "On Transient Loading Effects in Cavitation Pitting." ASME Paper to be presented at 1963 Annual Winter Meeting of ASME, November, 1963, 3o Hammitt, Fo G., et alo "Cavitation Damage Tests with Water in a Cavitating Venturi," ORA Report No. 03424-4-T, Nuclear Engineering Department, The University of Michigan, (April, 1962). 4. Hammitt, Fo G., et al, "Cavitation Damage Tests with Mercury and Water in a Cavitating Venturi and Other Components." ORA Report No. 03424-9-T, Nuclear Engineering Department, The University of Michigan, (September 1963), 5. Lichtman, Jo Zo, "Possible Contributions of Reentrant Flow to Cavitation Erosiono" ASME Paper Noo 62-HYD-3. 6. Kelly, Ro W,, Wood, Go Mo, Marman, Ho V., Milich, J. J., "Rotating Disk Approach for Cavitation Damage Studies in High Temperature Liquid Metal." ASME paper presented at Hydraulics Division Meeting, (May, 1963). 7. Knapp, Ro To, "Accelerated Field Tests of Cavitation Intensity." Trans. ASME, Volo 80, No 1, (Jan. 1958) 91-102; and "Recent Investigations of the Mechanics of Cavitation and Cavitation Damageo" Trans. ASME, (Octo 1955), 1045-1054. 80 Hobbs, Jo M,, "Problems of Predicting Cavitation Erosion from Accelerated Testso" ASME Paper No. 61-HYD-19. 9. Personal communication from Ao Thiruvengadam to Fo Go Hammitto 10o Barinka, L Lo,, Hammitt, Fo Go, "Detailed Investigation of Cavitation Pitting Characteristics from Cavitating Venturi Tests." ORA Report No 03424-8-T, Nuclear Engineering Department, The University of Michigan, (April, 1963)0 11o Thiruvengadam, A., "A Unified Theory of Cavitation Damage," ASME Paper No. 62-WA-118. -58

VIIo APPENDIX SPECIFICATION OF CAVITATION CONDITIONS The cavitation condition for all tests is defined in terms of "degree of cavitation", referring (except for initiation) to the extent of the cavitating region in the venturi. i) Sonic Initiation - First sonic manifestation beyond that of single phase flow. This was detected either by ear or, in the earlier water tests, by using a piezoelectric crystal. The results of these two methods were approximately the same. In all cases sonic initiation occurred at a higher throat pressure than visible initiation. ii) Visible Initiation - First appearance of a more or less complete ring of cavitation. This always appeared first at the throat exit. iii) Cavitation to Nose - The approximate location of the termination of the cavitation region is at the upstream nose of the test specimen. iv) Standard - The approximate location of the termination of the cavitation region is at the center of the test specimen, v) Cavitation to Back - The approximate location of the termination of the cavitation region is at the downstream end of the test specimen. vi) First Mark - The approximate location of the termination of the cavitation region is about 1-3/4 inches downstream from the throat outlet. -59

vii) Second Mark - The approximate location of the termination of the cavitation region is about 3-1/2 inches downstream from the throat outlet, The location of these termination points is shown to scale in Figure,1 Although the termination is not sharp, the cavitation conditions can be quite precisely reproduced.