THE UNIVERSITY OF MICHIGAN College of Engineering Mechanical Engineering Department Cavitation and Multiphase Flow Laboratory Report No. UMICHM 01357-32-I DETAILED CAVITATING FLOW REGIMES FOR CENTRIFUGAL PUMPS, AND HEAD VS. NPSH CURVES (to be presented, 1975 ASME Polyphase Flow Forum) by F'.G. Hammitt Financial Support Provided by: National Science Foundation Grant No. GK-1889 and the Argonne National Laboratory December 1974

ABSTRACT Available published data for detailed cavitation performance of pumps, showing ratios between NPSH values for first inception and head fall-off are presented. For.23 test points, it is found that these ratios vary between -1.5 and ~18.5, so that no meaningful prediction based on this past experience is as yet possible. Comparative cavitation sigma data in the same "hardware" between water and liquid metals is reviewed. No substantial difference due to fluid change per se is found. However, large cavitation "scale effects" within any given fluid exist. ACKNOWLEDGMENTS - Financial support for this investigation was provided both by the Argonne National Laboratory and by the National Science Foundation (Grant No. GK-1889).

-1I. INTRODUCTION A major problem in the design of pump impellers in cases where cavitation is probabl-e, and where knowledge of the exact cavitation condition existing is important, is relating the precise cavitation condition to the external pump performance as measured in terms of the conventional parameters resulting in the usual AH vs. NPSH curve (Fig. 1, eg.). Details of the existing cavitation regime may be important either because of cavitation damage, or for reasons particular to various applications. For example, it appears that cavitation must be avoided completely in the sodium circulating pumps for the LMFBR in one current design philosophy. It is there desired to use the acoustic signal from collapsing sub-cooled boiling bubbles in the core as a portion of the safety circuit. Distinguishing of such a signal from cavitation is difficult, so that zero cavitation may be required for such a system. On the other hand, if cavitation noise is not prohibitive per se, then cavitation damage becomes the probable limitation, and limited cavitation is permissible. However, without special provisions only the conventional Head vs. NPSH curve can be measured, so that it is desirable to relate this curve to the detailed cavitation regime to adequately design the sodium pumps for this application. Since present analytical techniques are inadequate to provide the necessary detailed information on the structure of the cavitation regime in pumps of this type, it has been necessary in many cases to make detailed model experimental studies using transparent components and sophisticated acoustic instrumentation. Early studies of this sort for cavitating liquid metal pumps (1-3) were made by NASA and Pratt and Whitney - CANEL. While various tests so far performed (1-8, eg.) indicate no significant difference in cavitation performance between liquid metals and water in the same hardware, all detailed tests to date show that cavitation commences at a much higher NPSH than that corresponding to the first fall-off of head or other externally measurable conventional pump parameter.

-2II. Compilation and Presentation of Available Data A. Comparison of Cavitation Inception between Water and Liquid Metals Various comparative cavitation sigma tests between water and liquid metals in the same "hardware" including particularly pumps and venturis, have been made over the last decade and reported in the literature (1-8, eg.). While most of the work concerns sodium, a mercury pump was tested in this laboratory (5). The present author has summarized the comparative test data so far available elsewhere (4). While large cavitation sigma "scale effects" exist for all liquids due to various causes, the presently available data, which is quite varied and considerable (as well as theoretical considerations) indicate no significant difference in cavitation sigmas due to different liquids per se. The effects of gas content are also not likely to be substantial as long as they are within the usual ranges found in engineering equipment. The presently available gas content data has been summarized elsewhere (9) by the present author and others. B. Detailed Pump Cavitating Flow Regimes - Head vs. NPSH Detailed pump cavitation tests wherein the earliest initiation of cavitation is detected either acoustically or visually* are reported over the past.15 years from various laboratories in the U.S. and abroad, such as Pratt and Whitney - CANAL (1,2), NASA-Lewis (3), NEL (10-14), Neratoom (7, 8, 15),UKAEC-Risley (16), and our own laboratory (5). From these sources it is possible to compute ratios between the NPSH corresponding to the first visual or acoustic appearance of cavitation and NPSHcrit, corresponding to conventionally defined cavitation inception, i.e., either 2 or 3% head drop-off. Twenty-three such tests points have been found from this reported data. These are listed in Table 1 along with the conventional pump parameters as far as possible. They are plotted * For first inception, the methods appear to be equivalent.

-3on Fig. 1, which is a generalized Head vs. NPSH curve, normalized to noncavitating head and NPSH.it While the scatter of points is considerable, I have divided the axis very roughly into regions of possible damage and noise, possible noise but probably no damage, and zero noise or damage. Of course these divisions assume "average" pump designs according to the present authors opinions, and are not in any case directly backed by experimental evidence at this time. The data is further grouped in Fig. 2, showing "frequency of event" vs. ratio of NPSH for inception (NPSH nc) to NPSH crit. In this way it is seen that the most probable value for this ratio is -2.5, but in some cases it is in the range 15 - 20. It appears that for "good" pumps (high S = suction specific speed), the ratio tends to be large, whereas it is moderate for moderate S pumps (12), but this is not always the case. At this point I know of no good explanation for the wide variation of this NPSH ratio. It thus still appears that no prediction of the detailed cavitation behavior of a new pump design, based on published past experience along, is meaningful. Hence, either detailed transparent (or acoustically instrumented) model tests, or the development of more realistic computing models than are presently available, seems necessary to assure required pump performance for such cavitation —sensitive applications as the sodium pumps for the LMFBR. The rough division of the overall performance regime shown in Fig. 1 seems to provide as accurate a general prediction as is. presently possible. This indicates that an NPSH "safety factor" of ~4 is necessary to avoid the probability of damage, and ~10 to avoid bubble noise. Separate detailed tests in water and sodium are probably not necessary, if full-scale water tests are made.

-4III. CONCLUSIONS The following general conclusions seem possible at this time. 1. No significant differences engineering-wise in cavitation performance of pumps between different liquids exist due to the change in liquid per se, but large cavitation "scale effects" exist for all liquids. This statement does not include damage considerations. Thus separate detailed cavitation performance tests between water and sodium, eg., are probably not necessary, if damage considerations are not involved. However, the only sure way to avoid damage (and bubble noise) is to assure this complete absence of cavitation in the pump. This will in general require an NPSH many times that for head fall-off (perhaps ~10 x). However, limited cavitation, requiring an NPSH "safety factor" of ~4,may be sufficient if only damage (and not noise) is the consideration. 2. Examination of 23 separate test points shows that the "most probable" ratio between NPSH for first inception and head fall-off is ~2.5. However, several pumps, particularly those with high suction specific speed, show ratios in the range 15 - 20. 3. Prediction of detailed cavitation behavior of a new pump design is not possible from the published past experience. Hence it appears that either detailed visually or acoustically instrumented model tests, or the development of improved computing models, are necessary for cavitation-sensitive applications such as the sodium circulating pumps for fast breeder reactors (LMFBR).

References 1. G.M. Wood, "Visual Cavitation Studies of Mixed Flow Pump Impellers," Trans. ASME, J. Basic Engr., 85, 1963, p. 17-28. 2. R.S. Kulp and J. V. Altieri, "Ca;vitation Damage of Mechanical Pump Impellers Operating in Liquid Metal Space Power Loops," NASA CR-165, July 1965 (Pratt and Whitney Aircraft - CANEL contractor report). 3. M.J. Hartmann and R.F. Soltis, "Observation of Cavitation in a Low Hub-Tip Ratio: Axial Flow Pump," ASME Paper No. 60-HYD-14, 1960. 4. F.G. Hammitt, A. Keller, O. Ahmed, J. Pyun, E. Yilmaz, "Cavitation Threshold and Superheat in Various Fluids," Proc. Conf. on Cavitation, Fluid Mach. Group, I. Mech. E., Edinburgh, Scotland, September 1974, p. 341-354. 5. F.G. Hammitt, "Observation of Cavitation Scale and Thermodynamic Effects in Stationary and Rotating Components," Trans. ASME, J. Basic Engr., D, 85, March 1963, p. 1-16. 6. J. Bonnin, "Thermodynamic Effects in Cavitation," Edinburgh Conf. (see no. 4), p. 355-362. 7. R.H. Fakkel, etal., "Comparison of Cavitation Test on the SNR 300 Prototype Sodium Pump, Carried Out Using Water at Room Temperature and Liquid Sodium at 5800 C," Edinburgh Conf. (see no. 4), p. 193-202. 8. R.H. Fakkel, C.J. Hoornweg, etal., "Development, Design, Construction and Full Scale Sodium Testing of Prototype Sodium Pump for LMFBR Power Plant," Proc. Bath Conf. on Nuclear Reactor Pumps, I. Mech. E., Bath, England, 1973, p. 197-205. 9. F.G. Hammitt, Effects of Gas Content upon Cavitation Inception, Performance, and Damage," J. Hyd. Research (IAHR), 10, 3, 1972, p. 259-290. 10.I.S. Pearsall and J. Forsyth, "Cavitation Speed Scale Effects in Pumps," Proc. IAHR Section for Hyd. Mach., 6th Symp., Rome, 1972, Paper I-7. 11. I.S. Pearsall, "Design of Pump Impellers for Optimum Cavitation Performance," Proc. Fluid Mach. Group, I. Mech. E., 187, 55/73, 1973, p. 667-678. 12. W.M. Deeprose, N.W. King, P.J. McNulty, I.S. Pearsall, "Cavitation Noise, Flow Noise and Erosion," Edinburgh Conf. (see no. 4), p. 373-381. 13. I.S. Pearsall, "A Review of Cavitation Scale Effects in Hydraulic Machines," preliminary draft for Cavitation Scale Effects, Working Group No. 1, Section for Hyd. Mach., IAHR, 1974. 14. I.S. Pearsall, "Acoustic Detection of Cavitation," yait, on Vibrations in Hyd. Pumps and Turbines, Paper No. 14, Proc. Inst. Mech. Engr., 1-A66-67, 181, Part 3A. 15. R.H. Fakkel, "Sodium Pump Development," Sodium Cooled Fast Reactor Engr., International Atomic Energy Agency, IAEA-SM-130/23,:Vienna, 1970, p. 343-374. 16. G. Seed, L.F. Bowles, I.D. Macleod, "Design, Testing, and Commissioning of Sodium Pumps for 600 MW(t) Prototype Fast Reactor," Bath Conf. (see no. 8), p. 173-185.

- fi"0 |ZG 22.. | li 13.. (~ Iz.'7 | I'''-. I, 5 "-' /;"' "":-:-, t -.- i ~~ ~ ~ ~t - ~ - ~ —-- - i: - ~ — ~ -- ~~ ~~ ~ ~ i- i r<. < - f l-7. I.... t.., i __,_.. _tI -~ _ ___ 3 LO'O....I - 2Z"....I (C _______ ____ /4 8Z1 a1 44ojdo~2 I,.3 _._',37 a|, 7t,_. | 1 6 1 MpH *_ I I I U I ____ I~ic3,3-:'I F0 2I.o E, 1io, o IW le (,0 &'7i:: _____:, K,' 21O —-.1T ~' t3 ]... _l [ t'NI =EL.!I - i I7Z i' X c6$~t'1 SS II i8 t|~~ ~t - _ 5w-_t Cz 8_ - ___. ooI7C) 22__ ___ ___ -'7 | "-''~, ~'j.~9!,,,,,V4 O~q,:, I,, z7XLor Ao-"7I 5S c3 s J, A i. C I_______ I-I p....... ",,.. ---- - -,,. Ct —--— t i- -, I w i F I.____ _ _ I. ( Ir- A 2l. " vO lAO! I; I J N 1 _-__ __ _. - I' ( f$ iDL2;':{R.. Io 1 ----— ^1 1. S 11 ~;'1 (OJE ( ____ t IqJWS II 1;'1 1 12 Pt~~~~~~~~~ a ~~ ~ (

R.Mic,'to>9\- n l l Q ~t r tJ lle, i, l.) I - -- -:() 4 W_'t- i.,, 1*~L~t5Gol - I - I - jotg Itz-<! t S z.Lcl-..i....... ct"I/,..'~ - 40 I cq1 F_-{' lo -- --- --- lA —1 l, I _. 1tN4- I | I i.I..L. I; 18~ I Jo. zVi/ — -I- YIII2 2., r II~,!1i-L5 I I 7.'. 11._1 it__ L_ _I _ _ I shs -I I I I I -- -. _-_ t 2.. /EL~ —-— ~ —------— 7'.... |......... ___,_'__':,I Atu~F-' L., ]..'-1- 1,,..... ____,I I i I L

(i (( (.) (t) 1.0............___ Probable Cavitation Noise Probable Zero Cavitation Effects I Possible Damage but Probable Zero and Noise Damage Fig. 1 Generalized Head vs. NPSH Pump Curve <: | ik- NPSHCRIT I I I I 0.01 1.I I II I I 1.0 2.0 4.0 -6.0 8.0 10 1 2 14 16 18 NPSH /NPSH 621a I. ~ ~ ~ ~CI I N PSH INC./NPSH CRIT. 621a