Text

Performance of Safety Injection Pumps W/Air Ingestion, Final Rept
	ML20246L962
Person / Time
Site:	South Texas
Issue date:	03/13/1989
From:	Swift W; CREARE, INC.
To:	;
Shared Package
ML20246L942	List: ML20246L937; ML20246L962; ST-HL-AE-3035, Evaluation of Possible Air Ingestion Due to Free Surface Vortices;
References
	TM-1332, NUDOCS 8903240333
	Download: ML20246L962 (22)
	v • d • e

n ,

ATTACHMENT 2 r.; .

ST-HL-AE .sas i

~

L PAGE _ / _ _ op __.t f im- 1552 reGO .

Performance of Safety Injection Pumps With Air Ingestion W. Swift Creare Inc.

I Final Report 13 Mamh 1989 TM-1332 CREARE INC. 6800-HANOVER, NEW HAMPSHIRE 03755 MARCH 13,1989 8903240333 890317 F~

hDR ADOCK 05000498 PNU

... e

. ,- s t s TM-1332

.' @rOOre. c ATTACHMENT !

ST-HL-AE .5 e 55

1.

SUMMARY

L PAGE d._.OF A This report describes the result of an evaluation of dw: performance of safety injection pumpt at Umt 1 of the South Texas Project Electric Generating Station. The evaluation was performed by Creare Inc. for Houston Lighting & Power Company.' Walter Swift performed the evaluation at Creare with support from Victor Iannello;- Albert McIntyre was the project manager for HL&P.

a The evaluation consisted of an independent assessment of the behavior of the core spray (CS), high pressure safety injection (HPSI) and low pressure safety injection (LPSI) pumps i during a postulated loss of coolant accident (LOCA - 1) assuming a double ended pump ,

suction guillotine break with minimum safety injection (two of three pump trains in operation).

The esaluation included a review of assumptions and the calculation methodology performed by UL&P (MC-6126) with respect to pump operation. There are two major aspects to the operation of the pump system:

1) at the calculated flow rates for the safdty system, what quantity of air (if any) is ,

entrained in the sump pipe, and !

)

2) if air entrainment occurs in the sump, how will this air affect pump performance? l Otr evaluation does not address the first question. ' Die methodology followed in the study calculation (MC-6126), which is based on guidelines set forth in NUREG - 0897, Rev.

1 [1], predicts an' air ingestion rate in the sump of 8.94 % under the LOCA - I conditions. !

This prediction is based on a c . ilated Froude number of 0.759 at the sump, which depends on a sump pipe velocity U, W m in turn depends on calculated flow rates and piping losses.

We have reviewed the approach ced to calculate the aggregate flow rate and thus amve at the sump pipe velocity U. Our independent calculation would predict a marginally smaller value of U than that given in MC - 6126. However, the difference is small and well withir the uncertainties of any similar calculation of system losses.

We then used the predicted sump air ingestion rate of 8.94 % and reviewed the calculations used to predict the air ingestion quantity at the pum ps. We concur with the MC -

6126 result that the rate of air ingestion at the CS, LHSI and 3HSI pumps predicted by the i NUREG 0897 methodology is 7.86 % - 7.87 %. Furthermore, we have found that the l calculations in MC - 6126 follow the NUREG 0897 methodology to predict an NPSHR l

penalty as a result of the predicted air ingestion rate. (However, NUREG 0897 is not to be applied to situations where the air ingestion rate at the pump cxceeds 2 %).

Finally, we reviewed the statements about the safety significance of the pump performance given on page 13 of MC - 6126. An analysis is presented which relies on the results of tests on a single stage, low specific speed pump under various combinations of air ,

ingestion rates and NPSH. The analysis concludes that at the NPSH available (30 ft.) and the 8 % air ingestion rate predicted by the NUREG analysis methodology, the pumps (CS, LHSI and HHSI) would be att to sustain at least 80 % cf the normal (liquid) head. The ar..dysis concludes that under the. calculated conditions, there was no significant hazani to safety.

This conclusion depends on a prediction of the amount of head degradation for the CS, LHSI and HHSI pumps under air ingesting conditions from laboratory test data on other pumps. NUREG/ CR - 2792 [2] addresses this issue. Figure 3-2 from this reference (Figure 9 of this report) is a composite plot illustrating the variability in two - phase performance of pumps which have been documented in the literatum. An important conc.lusion given in 2

4

_ __________._mm_-___-__m.-. _ _ _ m

ATTACOT[EtiT ST-HL-AE- pg TM-1332 h*

x

[@[6 , PAGE 3 - 0F _ 21

'I n section 4.1 of the NUREG is that for ingestion rates above 2 %, pum? behavior is highly

? variable, depending on the specific pump design, the flow rate, anc other, unidentified variables. It is because of this variability that it is not possible to predict with absolute certainty how one pump will behave at some operating condition based on tests of another pump at similar conditions, or even from tests of the same pump at diffemnt conditions. ;

It is not appropriate to apply the data of Merry directly to predict the performance of the CS, LHSI and HHSI pumps. The pump used by Merry m the laboratory tests was a low specific speed (810 - U.S. units)I, single stage machine, quite unlike the mu.It i - stage safety ,

pumps with mixed flow impellers (specific speed = 3110 per stage). Based on our review of the technical literature, there is no smgle set of data or a proven analytical method which can ;

be directly applied to these pumps to exactly predict their behavior at an 8% level of air :

ingestion. However, one can make judgments about the likely behavior of the safety pumps based on the existing data base.

For example, it is known that the presence of an inducer substantially improves the cavitating and air ingesting performance of centrifugal pumps. It has also been demonstrated by test (as noted in MC - 6126 and NUREG/ CR -2792) that muldstage pumps perform with less head degradation than single stage pumps under similar air ingestion rates. Arguments can also be made that based on existing data, mixed flow impeliers will operate with less head degradation than low specific speed designs at similar air ingestion quantities. However, it is also known that operation with a sufficient amount of air ingestion over a period of time can result in the accumulation of air at pump impeller inlets, radically changmg their ability to perform and occasionally causing mechanical problems. The combination of all these factors still cannot be used to quantify the amount of head degradation which would occur under the conditions studied in MC - 6126. One can only conclude that there is a reasonable probability that the head degradation under these condiuons would be comparable to or less than the values given in the literature - which are comparable to values found in Merry's tests.

2. AIR ENTRAINMENT LEVELS IN THE PUMPS ,

The calculation methodolog: n:;ed in MC - 6126 was checked for the LOCA - 1 analysis. Our independent analysis predicted the same levels of air ingestion at the pump inlets as those identified in MC - 6126.

3. HEAD DEGRADATION / SAFETY SIGNIFICANCE The analysis in MC - 6126 predicts an air ingestion rate of about 8 % at the pump inlets using the NUREG 0897 methodology for the LOCA - 1 set of assumptions. At air ingestion rates in excess of 2 %, NUREG - 0897 gives no guidance as to how to evaluate the behavior of centrifugal pumps. Section 3.2.2.2 of NUREG 0897 and Section 4.1 of NUREG/ CR -

2792 address the issue. Basically, there is so much variability in pump behavior (as indicated by tests) at ingestion rates above 2 % that the specification of generic technical guidelines for the prediction of behavior over a wide variety of pump types is not appropriate. However, there is also a reasonable amount of similarity in pump behavior at moderately low air rates 1 The specific speed - rpm (gpm)0.5 j (f,)0.75 - of the Merry pump is incorrectly listed in I Table 3 - 1 of NUREG/ CR - 2792 as 1,074.

3

ATTACHMENT ST-HL-AE- 3 e a s' TM-1332

.' ( & (& . PAGE f 0F _ .s I from widely differing tests and types of pumps. A review of the more important literature on air effects m pumps allows one to draw some conclusions about the likely behavior of these safety pumps based on results from other tests. The following sections discuss important features of the safety pumps, review existing test data with respect to air ingestion performance at similar operating conditions and discuss these results with respect to the probable behavior of the safety pumps.

i 3.1 KEY FEATURES OF THE SAFETY PUMPS Number of Staees The CS and LHSI pumps are each five stage mixed flow machines. THe HHSI pumps are sixteen stage machines with mixed flow impellers. In general, multistage pumps show less head degradation than single stage pumps which use the same impeller.

Specific Soeed An important hydraulic feature of a centrifugal pump is its specific speed.

l N = N(rpm) * (Q(gpm))0.5 /(H(ft.))0.75 s

This parameter characterizes design dimensional and speed relationships for a pump. Pumps of comparable specific speed will generally have similar performance characteristics, although there is also vanability because of individual design decisions for each pump. The CS and LHSI pumps are each five stage machines with a best efficiency flow rate of about 2700 gpm, and a corresponding best efficiency head of 460' (or 92' per stage) at a speed of 1780 rpm.

This corresponds to a design specific speed of 3114 (per impeller stage). The impellers would be classified as mixed-flow designs.

Incocer The CS and LHSI pumps are equipped with integral inducers at the inlet to the lowest stage impeller. In general these screw - like impellers provide improved performance with l respect to cavitaticv and air or vapor entrainment.

3.2 EXISTING AIR / WATER TEST DATA Test data from a number of pumps over various operating conditions are reviewed in this i section. These data illustrate the variety in results of air water performance for pumps of different specific speed and number of stages for changing air / water flows and at various ;

levels of NPSH.

Floriancic I31 Tests were performed on a single stage pump and a three stage pump of comparable design to evaluate the behavior of the pump unc er mixed air / water and argon / water flow. The test results show two important trends:

4

o.

ATTACHMENT ST-HL-AE- 3* 35 TM-1332 Ckgg , PAGE 5 0F Al

1) the performance of a multistage pump under two phase flow situations is superior to a single stage pump of the same design, at the same ' flow conditions, and

2) the performance of a single stage, pump under two phase flow conditions deteriorates as the upstream pressure is lowered.

The test pump is listed as an N, = 95 (Mks system) design. However, no definition for N, is given in the paper. The U.S. equivalent value is ambiguous becausc Sulzer (the company at which the tests were performed) has used differing equations from time to time to define N '

s The value lies between 5163 and 1397 in the comparable U.S. units, depending which of two definitions of N,is selected.

The tests were limited to volumetric air fractions below about 6%. Figures 1,2,3, and 4 (Figures 3-7, 3-8 and 3-9 from NUREG/ CR - 2792, and an additional figure from the original paper) illustrate features of the effects of air on performance.

Figum I shows head, efficiency and power variations with flow for various air flows on a single stage p, ump. For air fractions of 2% or less, them is no serious degradation in head -

even over a wide range of liquid flows This characteristic is typical of the behavior of all centrifugal pumps which wem evaluated during the writing of Rev.1 of NUREG 0897. As the air fraction increases to 6.4 %, head degradation becomes significant, and at high flow rates -

severe. Figure 2 shows similar performance relationships for three stages af the same pump of Figum 1. It is clear that there is a significant reduction in head degradation in the multistage machine, particularly at flows lower tien best efficiency point. 'Ihis effect is summarized for rated flow in Figure 4. Figum 3 shows another important characteristic for centrifugal pumps when air is entrained. As the suction pressure is reduced, the effects of air tend to increase head degradation. This is generally coupled in some way with the NPSH requirements of the pump. But, there is insufficient information given in this reference about the cavitation characteristics of the pumps tested to arrive at any quantitative conclusions.

Creare Air / Water and B&W Air / Water I41 Tests were performed at Creare on a single stage, mixed flow pump (N, = 4200) with a 1.94" diameter impeller in air water mixtures. Comparable testing with a similar pump (N, =

4317 - nondimensional, single phase performance curves of the two pumps agreed within 5 %)

having an impeller diameter of 12.3" was carried out by Babcock and Wilcox. Single phase and two phase performance were documented for all four quadrants of operation. Figure 5 characterizes the two phase test results near design flow coefficient for each pump. The i Figure (from reference 4) shows how head degradation varies with void fraction ap. '

l The test results are presented in terms of homologous parameters h/aN and v/aN, and void fraction a p. h is the ratio of test head to single phase head, aNis the ratio of test speed to rated speed, y is the ratio of volumetric flow to rated flow, and pa is the ratio of volumetric air flow to total volumetric flow.

5 w_--__-___-_ ._ ._ -__. _

j

.~ -

(

ATTACHMENT

[- Ch[@[$,, ,

ST HL-AE- J'M PAGE 6 0F .2j se ._ . ,. .. - . . - - _ _ . . . . . _.. .. ...

-e i

a.J- - ,l [

i

._._T .

l '

Air Volume.Fract1 4 i - I 120 !

i - - -

l . l - - - .m l '

j 8 e l l l l

? Ps 0%

w 100 -

--t---- - - - - - - - - - .- -. ~ - " -

, I

= \. p 2,05%

30 . - . . . ..

7 . 3 _ _ _

5 e ae 60 q

I I{ . . .j ..

i e -..-. . - .e .

. a

,e . _. _. _ . ._

2 2 '

.... . .L .-l. . . . . ...

_j ._

4o 4 .. ... ... . . .. . .. _ _ _ , ..

l .

p ,08%

4

..- . . 1. '

j. .

[6,40%

20 l l l1 l l-N 5,07%

/ 0%

3to . . .. .. . . - . .

W..&

j 8 300 . .. _. . . _

1

.. . . .. . .i . . I_.__l.. i

. ). 3. ,.,

x 90 .. . __.. . . . . . .

. ... . .!' . j .

'_ . _ , j 2,05 %

u ' ' 4,08 %

! l ! l I 30 . . 4 ,.

j. ,- . ~ . . . . .

70 .. .

i.

l 3

4..

1 I

d. . . .d ,

I 5,07%, '

1 k

j 60< l -- # -- -

... 4 l ' l i

,I- - - --A- 6,40 %

l

y-i -.;.-----..-

5 M, 50 . ,

A* 5 . ,i _ } _ 1. . - . .I. .

.i . _ . .!. .._. . . ._ _ . i j ~

l , !

l 40 '

'l 100- , , . .

2 m- _ -

o '

) i lI l I c 90,. _L._. . .. ,. _, . . .' ' _ , !

l'

. I _.. '.1 . 2._

M 80- ' - - - .-- -- -- . -- -.

t

.j. 3 70 _.1 .. . . . _ .

.p 7 q __ .p. 1 _ . .. . .

=w_

___. 'q _ .. .L . !_ . .'_ _ '

. ,. _ . . __q _ . . _ . 2,05 %

fa i i

! i I I t---

e/--- i L d 50- -- { 1;y

I

- - --1 i j -

I ' '

D U 40 .-_2.. . . _. I 4,08 %

e i e g 30 . .. l . ....2._._._4 O E

..d i i

. 5,07%.

6,40 %

g .y 20- -

-4 - -- - - - -

p- -. 4 - - - * - - ----.. ---- -- -- J w e . , i W I: 10- I I

- - - - + - - -* A--

- 4 i ' n.

O I ' ' ! I I ""

0 , 10 20 30 40 50 60 70 80 90 100 110 120 130 Volumetric Flow Rate of Water x 100 '

Rated Volumetric Flow Rate i Figure 1. NORMALIZED HEAD, POWER AND EFFICIENCY IN AIR / WATER FLOW FROM FLORJANCIC [3] - SINGLE STAGE PUMP 6

. o' ATTACHMENT ST-HL-AE- 305

- PAGE 2 0F A l g ,

1M 7 %

.m o _

110 Vt /Q., 3,9% '

\ [' leas"#'

Vt /O.,I 4,9%- V'/Ea,- 2,1%

$00--

lt c , ;- l 'NI i t 90 l

, ! $. jNTDg a

V/Q.,

t 6,1% MIk !

o 80 k k s TL M N 70 110-Vt /Q.,- 0, 2,1%

@ 00 I ;

$90 .

O,,,*52 N l\

g s \ Vt /0,,- 3,9%

o N A'go .

Y g, .[s '~

N Vt IO.,[- 4f9%'

' Va lQ.,- 6,1%.

o Y/Q.,-0%

a b 100 ft/G.,- 2,1%

y& T N70 4 \ k Vt /0,,- 3,9%

O60 '

\9/., 4,9%

t

~Y a /d.,

6,l'%""

g. / [

a p /

'u e /

n n,0 -f u e20 -

14 lE.

10

- [

0 .

M's 0 10 20 30 40 50 60 70 80 90 100 110 120 l

J Volumetric Flow Rate of Water x 100 1 Rated Volumetric Flow Rate Figure 2. NORMALIZED HEAD, POWER AND EFFICIENCY IN AIR / WATER FLOW FROM FLORJANCIC [3] - THREE STAGE PUMP 7

1 j,_

d-

~. .

q

'f ATTACHMENT 1 ST-HL-AE- 3 e.sr !

.' [N[h .

PAGE F - 0F J/ --

l q

AIR VOLUME' f TRACTION I

100 Y3 $ 3w { -

l w

2% l 90-

\ \ -

\ \ \ "

j c, 80-

\ 3 w 4%

x k k. "

K' 5% )

e 70-

$ \! l l

li & 60- -

. \h A

50 . .I "

6%

0 0,1 0,2 0,3 0,4 1/ Inlet Pressure (atm'1)

Figure 3. EFFECT OF INLET PRESSURE ON AIR / WATER PERFORMANCE FROM FLORJANCIC [3J i

i l l l

8

~

.x

C -

ATTACHMENT AGE OF __e l _

AIR VOLUME FRACTION.

100 --- - - - - -

4 3 2%

3 w 90-

_ A-3 N 3 w\ ,

A 3

\ %

s s 4%

i

'8~

80 \, \ g

\ ,

\ --

5%

h 70 g , g z N \

f 7 60

\

3 i

\

50 -

\ 6%

0 0,5 1,0 1/ NUMBER OF STAGES Figure 4. EFFECT OF NUMBER OF STAGES ON AIR / WATER PERFORMANCE FROM FLORJANCIC (3J I

i 4

9

ATTACHMENT

( & (Q .

ST-HL-AE- so3s PAGE /4 0F J /

1.s --

G CREARE - 1.94" DIAMETER ,

. a l I" Q BABCOCK & WILCOX - 12.3" DIAMETER i.s e se 8p e

'I s' ' I 1 ee !

I r '.5. "e =

,em f.s ..

5

,* t ,

a ,

8 e j

se 8% ,

1 j... . .

. .%.L

. # a. _ _ _

e:S m e:r s;s nipse Es e.6 og Ree ~ ee;, i;.

Vold froctlon OF g

E45 e

-8.5 --

S

-1.0 -

f Figure 5. COMPARISON OF CREARE AIR / WATER AND B&W DATA -

l TWO PHASE FOR v/aNBETWEEN 0.80 AND 1.20 [4]

1 10 l

..e"

4 ATTACHMENT TM-1332

.' @Q. ,

GE - JL- 0F Q i

Three important conclusions can be drawn fmm these test results:

1) them 'is a fair amount of scatter in the results showing the. variability in behavior (which includes uncertainties in measurements) between two nearly identical hydraulic designs at comparable operating conditions,

2) below 10 % air fraction the head remains above 60 % of the single phase value (h/otN = 1), and

3) thert is a fair amount of agreement between the two pumps in the tand of degradation with increased void fraction.

Unfortunately, there is no information provided in the reference (or other references relating to these tests) regarding variable suction pressures or the interaction between cavitation performance and air / water behavior.

Murakami et al. [5L I61. I71 This group of Japanese researchers have tested several series of pumps - axial and ,

centrifugal - with variations in speed, air volume, numbers of blades, etc. Among the tests am 1 data on the combined effects of air and NPSH. Some of the results of their tests are summarized in Figum 6.

In tests on a low specific speed pump to evaluate air entrainment effects, they found virtually no change in performance below air fractions of about 4 % by volume [5]. Between 5 % and 10 % volumetric air fraction, they noted " discontinuities" in pump. performance characteristics as a result of the introduction of air. The test pump was a radial flow centrifugal design with an 8.8" diameter impeller having 5 blades. There is evidence from their flow visualization during the tests that the discontinuities in the performance characteristics result from abrupt changes in the flow of air bubbles through the pump.

Basically, because the machine is highly radial flow, the blade to blade pressure gradients are large. They tend to promote the accumulation of air bubbles in low pressure areas through the blade passages which severely distorts the liquid flow patterns. The attached bubbles grow in volume, then are swept out of the impeller - accompamed by flow and pressum fluctuations.

Comparative tests [6] were performed on three similar radial flow impellers of approximately the same design discussed in [5]. The effect of number of impeller blades was evaluated. A minor shift in peak efficiency occurred as the blade number varied (3,5 and 7 blade impellers wem tested). But the performance curves were comparable - specific speed at peak efficiency was within +/- 10 % for the three machines. The tests showed that for the three blade impeller, the flow discontinuities occurred at lower air fractions (about 3 %) - !

probably because of the stronger blade to blade pressure gradients associated with the fewer number of blades. In general, as the number of blades was increased, the quantity of air which the pump tolerated befom evidence of " discontinuities" in flow also increased.

Later tests were also conducted on the three pumps to evaluate the combined effects of cavitation and air entrainment on performance [7]. The liguid cavitation performance of the 5 and 7 blade pumps were quite similar, that of the 3 blade impeller differed significantly from the other two. Figure 6 shows the combined effects of air and varying cavitation on the pump 11

c.-

1 ATTACHMENT Greare . .

%"V^%'# u O7 \

Ci U [s4 = 0 I s:- .

.4 %..

O6 <

3Ag; , , , ,, - - -

, j -

. oos c.og 05-f

\ . Ci 0 07

^

sa9adurcumas=w

, . . e . . _

O4 ,

'h' .,..,,, . . tv.. . . . _- _

. \

r - ,- x '

+ _<

\

. y' 03 - i Pump Ps

' (pr = O 2 85)

C4 AC1,

. . =- %. ,, ..

O 06 -

' ' d $"*_"'Q~,Tj~ ~_~_ p t

4

** X _ . J'.g. -

_ /,

g\ \

0o4 ' ' '

O5 lO 15 7 20 Figure 6. PERFORMANCE OF PUMP PS DER MR MMG CONDITIONS [7]

12 i

,a ATTACHMENT

. ST-HL-AE F35 TM-1332 h[ @ [$ . PAGE /L_0F 21 performance. The abscissa in the figure is a = NPSH / (ul /2g). The shading in the figure shows the band of "discon:inuous" behavior. As airflow is increased or as NPSH is decreased, the flow in the impeller becomes unstable once the discontinuity band is reached.

Neumann & Lualdi I81 Tests were performed on a mixed flow impeller in highly aerated conditions to determine the air handling behavior under differing blade settings and under differing means of air induction. Details of the pump design are not given in the paper - except that a Worthington model 16 MS 5 diffuser pump was used, with an equivalent specific speed of about 5100 (U.S.

units). The pump produced a head rise of about 29 ft. at 7000 gpm,740 rpm.

Figure 7 illustrates some of the results of these tests. In the region of best efficiency, two different methods were used to inject air. The dashed curves show behavior using the first device which did not control uniformity of the air entering the loop upstream of the pump.

The authors describe flow conditions with a large degree of scatter in the data. The second device assumd uniform bubble distribution of air. It is evident from the figure that upstream air flow distribution uniformity will strongly affect head degradation of the pump. It is also 4 obvious that for relatively low levels of air (6 %) the effect is less pronounced.

Merry I91 Tests by Merry were performed on a radial flow centrifugal pump to demonstrate the behavior of the pump under combined conditions of air ingestion and cavitation. These test results were used in NUREG/ CR - 2792 to illustrate these effects. While these results are useful for characterizing generic behavior of centrifugal pumps under these conditions, the curves cannot be arbitrarily applied to other machines. Figure 8 shows how the requirements for NPSH are increased as air quantities increase. The specific speed of the pump at best efficiency is 812 in U.S. units.

3.3 DISCUSSION Figure 9 from [2] presents a large array of two phase pump test results which includes most of the data discussed in the previous section. The figure shows that for air fractions less than about 5 %, most of the existing two phase pump data shows head degradation less than 20

%. I.e., the collective set of data, which incluces a large amount of variability in pumps, test conditions, etc., tends to indicate that for operation near design conditions, head degradation will be less than 20 % for an inlet air fraction of 5 % or less. The specific data of Florjancic indicates that multistage pumps will perform with even less degradation than a comparable single stage unit for the same air ingestion rate.

Figures 10 and 11 show the data of [6] and [9] in a reduced format to provide some guidance on the relative effect of NPSH on two phase head degradation. The vertical axis in each figure is test head divided by single phase rated head at the test flow and speed. The i horizontal axis for each figure is o/o . o is defined here as the ratio of NPSH to rated head.

cr o cr is the value of a at the 3% head degradation point for single phase flow. By replotting the data in this format, one can evaluate differences between radically different machines on the l_

same relative basis. On each figure, a dashed line shows the approximate locus of 3% head degradation for different air ingestion rates.

13

t-

.s p '

1

ATTACHMENT

- A i

ST HL-AE- 3ess

_PAGE_ 19 op___p_7 L.-

@ege.

J

. , . . .T .. i .

l. : i . .i .

i. t l -(. ' y 6 . l . ;- -l.

l i.. ..., .... . . - . . , .

..'_. ..,._. ._J........ . .:.. '

I 4 1 i !

. l

y . .

a

. . . .o . .I; , l i

i, . . . . ..i..

, , q.. ;

i : ;, i / i , .

.}

l [' d :

/

..'... ..l. _ . . . . . ... . , ..

g . . .

I s i i . '} :. Z i

o i: : o l . r 1

. . . .. .W .. . . . i . . . . . ... . . . .

I ...I ,

n 1 t I .;

r ,

,f ;, . . : . i ,-

i- .

i.

. a U

l .. ;. . . _ . . . . ,i .

i I . . .. . , _ . _. , . . ...'

.. 1. .

m g (. e/. <

8 s s

a Z"

i i l ,

.. ...J. ... o- ; !..<_ ...,.. . ,,,e i ..i. .

.. ., ..j , . , . , . . . . . .... _ . . . . .. . . . . .

l . / . +

, , , e M

', 3 - ,

i ' !/

,s ,,.

l

... l

,. I ,; .

f

_. 4_ . . . . .: .. .

j ...li Q no t ,.

j.,l ; .

'=,

. . . .e # i.......... . . , . . ... ...

._..).

j i : i g .

l..!< i. , s

.s s

9 8;

, s. !

E- p t k . . . . . . ..

g . <h -6 . . . . . . . , . ..

, s. .

[ ' '

db !

...,, , . . .", .6 .

' J. ..,

i ,7 . .. .I... . .t i , ! i -

i 3:

l

...t ..

l l.. . N. .I j

.. ...s /. l

. i.. . h.,i, v3

' 'W a i

. . r,/)

[ O

. p .. .

i !

j ..... p g . ,

i . t

! ,. . . j..

l ,

.I.;

, .;y

. t-

,,,,a I 'W ; , i.

i' ' :> Q i P ,I * . .

O l l,

. . . . . . ..t. . a. . . i.

.l W J .,' f. .

1

. <J W i /

i , . et : - ' , ; U.

l, I l l l{ . 2,. ,.. I. a ! . a.he

. . w.e ...l'.<...g.

. . . .l .... , .- go. . . . . . ..... ..y... .. 4... . e .. . .l I !

i

.i i i !.

k Q d l

.. ld g

i i .

i s' ,

.i <_ p is

.s

i. jl.'.

I i

U'

r. W go l2

.. . .. .J . s ..

q :

i j

) ._ l . t s

(o w Iw> , . I . . . .. -[ .. ., i- ..... l.. . :. -

e

' ,i

. i .

t , . i V

4 l' ' i g yU I

i

, I 1

.! i l'

i

!I M

'. 3 m 4

..! i I i l ! '

i d -;

e l

i .

. l 4

. . .r ' , l

W .

i l 1

6 , I i, ,

a. .

g .. . !... . . . . . .. . . . . .

.f. .,

Wl ! 4 -

!/- ! '

.a :

1 l

,. , ,. .i . . . .

. . . . . . . . . . . . 'i. ..: i. .

' ' i

2 l '

O i ! , I ! ! i h! I/ i i

. V. . : a ; I ;

. 4.. .; I, !

J l . l 3

. . / ' '

f' l ! ' j- l 1

l

.j' /' i !

o . r'.

2 ., - l . . .

, i- I!

l--. j..l. -

t.

q. l .d.r

..L.

I!

. S, g

m

, i ! ,.' i V .1

].

, i (l. 4 l co

. ;t ,' ..; .j..3

7. .

.g . . . + ,. .._ .

, . , g

's7 .

.. . I r j 'l .! ,

,i l. .

id'g :

ei

[..

d ,.

. "C . ,,,s . .. .1 j. i ... .j . ;.. ~,'I" l, ' @'

n2 i - i i . , - . d_ ' ~ ' -

i l l

...l...'.g M '

- , j i ( ' i l1 '

. .a . ,. 3 .

i

. i e I i t b.. '

l '! - -' )

4 i I j ! .

L 2 _! . ... l' .. . ,

.L.,.

' O

/

d.

sn O

av i-o i

f i

14

> ,y' -'*

(&(0 ' . . ATTACHMENT ST-HL-AE- 3635 '

PAGE >< OF Al I I I g g 3 0

m o

- g o

4 W , ,

e e

~

\s a:\\\ :;; -

e e e

m a

i I""

~

g _

g w

b.

s ;

d Z l

- w ~

q WE W

= E >>

" h i B om o

~

N O

h2 -l

2 O

Od wm e

s

_s-

==

- - as EE $5 a

o-w o

z

_ h* 5 2

=gu

!!!=

Ndd

~

==- L'% - o ,

E et-* +-

E ic I I i 1 1 9 9 9 M

q o

- - o o o o ;

Y H/ H i

i 15

o ATTACHMENT u I 1.2 , . h. ~n"cbAq,3g37 ,, ,

~, i

{f%fQ .$

.' 0 A

l 6 0 l $ 0 A A l 1.0 0

g g et

_ , e e. , e

@ Y M 6 m

i gM 6 6 O

w x

08 -

e(+ %

E D %

w O-i a

- Y e Y# +

E3 4 o Cy DE!

S 0.6 .

+ g 66 ,

+

~

W

. s%. 3 ,

W b 0*4 -

g M m@ 6 .

U c

a e C+ h c ,

W 0*2

do C

1 SYMBOL DEFINITION k ~

XX a

W O

creare Asv & s/w Data (12)

B&W A/W Data (12]

x6

o O c-E s/w Data (12) z3 C -t- G-E S/w Data (14]

0.0 . A semiscale s/w (10,111 ,

g Y Stepanoff (14] (Hr = 54 Ft) j a Stepanoff (} 4] (Hr = 188 Ft) l Herry (31) {

Murakani & Minemura (30)

A Floriancic I32)

-0.2 e i i i 0.00 0.04 0.08 0.12 6.16 0.20 )

PUMP INLET VOID FRACTION Figure 9. PUMP PRESSURE RISE IN AIR / WATER AND STEAM WATER FLOWS [2]

]

16

.2 - '

I .. ' . i L .

h(@(Q --

pkgglA5

_y

%d TWNs-s -

Fy v vX O w N ,

l RMM l

g Q QM4 % _

Qn 9h 1

El--

9 g

k ** -m 5 hN 4 1 h w 5 z

o knA.

I c. A<

Q j %' _

I h' I i n o W

l a+ o Y -e

!B

~

[]+ 0 z.

t o o'

W c

[]..

\ U f ti b Z u

[1 -

o M

El

/ b m

m a

6 e

e A

I I I I l l l l l l l O

9 - -

9 m h 9 q y n q - o ,

" - o o o o o o o o d !

s 17 1, 1

.-._m.-. _ _ _ . ._

!la t ..

53

~

$ = gh

,d!z

- h> rp yN E" g O b' ,

l E s " 4 S -

A 2 1 7, .

R 51 0 O T 3 i bl .

A A i ]

9 C 3, . . fy [

C; o af R o I q f A g' _ p l i

y :=

z T I

g a f.

C; ,3 W

4. g, 5; N O

I T) T

< A CI s. R D _

I M R A

?

a G

- l r E -

C s f D -

8 (_- D A -

E

{.s l l _

C : c N ,2 h

N O

C l H

S P

G y

N N F _.

O G . i S

T C

G .

E F

F m E I

,1

[ o -

1 1

e r

u g

i F

O 2 1 1 9 8 7 6 5 4 3 2 1 0 1 1 O O 0 O O 0 0 0 0 C

/

5

!!l l! .!

. o ATTACHMENT.u ST-HL-AE. 3

'. PAGE N OF M TM-1332

$ TOGO -

Figure 12 then shows the combined set of data in terms of a ratio at which there is a three pr cent head drop for varying air ingestion rates. The result shows that for these two sets of pump data, the degradation in head (fmm NPSH effects) is expected to be less than 3%

of the nommal two phase value for o ratios above the dashed curve. One could conclude themfore, that for an expected air ingestion rate of 4%, a sigma ratio of more than 1.4 is required to assum no appreciable head degradation beyond that predicted by just air ingestion considerations alone. This is equivalent to a required increase in NPSH of 1.4 times the NPSH required for all liquid flow. At 8% air fraction, the NPSH requirement would be about twice the all liquid value.

The calculations in MC - 6126 indicate that for the LOCA - 1 analysis, the net positive suction head available at the pumps is about 30'. The pump performance curve shows an NPSH requirement of about 11' at a corresponding flow rate for operation with 100% liquid.

This imphes that the sigma ratio for the LOCA - 1 analysis is 30/11 = 2.7. Using the results of the Merry tests and the work of Murakami & Minemura, this would imply that the additional head degradation due to NPSH would be less than 3 % even up to 10 % air fraction.

The test results given in the previous section cannot be strictly applied to predict exactly how the safety pumps would behave under air ingesting conditions. As noted earlier, there is a great deal of variability in the pumps and test conditions. However, based on the collective results in the literature and the design featums of the safety pumps (the fact that they are multistage and have inlet inducers), it seems highly probable that these machines would have been capable of sustaining at least 80 % of the nominal rated head for air fractions of up to 5

% at the pump inlet under the LOCA - 1 conditions and assumptions. It is also likely based on the data and margins of uncertainty that the pumps would have been able to sustain operation with less than 50% head degradation for air fractions up to 8% for at least short periods of time.

References

1. Serkiz, A.W.; Containment Emergency Sump Performance; NUREG-0897 Revision 1, !

U.S. NRC, October 1985.

l

2. Kamath, P.S., Tantillo, T.J. and Swift, W.L.; An Assessment of Residual Heat Removal and Containment Spray Pump Performance Under Air and Debris Ingesting Conditions; Creare TM-825, NUREG/CR-2792, September 1982.

3. Florjancic, D.; Influence of Gas and Air Admission on the Behaviour of Single and Multi-Stage Pumps; Sulzer Research Number 1970, pp. 35-44.

4. Kamath, P.S. et.al.; Two Phase Performance of Scale Models of a Primary Coolant Pump; EPRI-2578, Electric Power Research Institute, September 1982.

5. Murakami, M. and Minemura, K.; Effects of Entrained Air on the Performance of a ,

Centnfugal Pump; Bulletin of the JSME,17(110), August 1974, pp.1047-1055. 1

6. Murakami, M. and Minemura, K.; Effects of Entrained Air on the Performance of i Centrifugal Pumps; Bulletin of the JSME,17(112), October 1974, pp.1286-1295.

19

- _ _ _ _ _ _ _ _ _ _ - _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _______-______m

... . .e.

~

. p.

ATTACHMENT i (N(0 *. j- AGE o of_p; ;

1 L.

2.5 .

i . , ,

i , , ,

l l

2.0 -

z O

~

i 8 O

<l.5 -

a -

E V m 8 -

o Q b a.

N O

4 l .O(- -

b y o MERRY - FIGURE II ge Q MURAKAMI- FIGURE 10 ro 0.5 -

O i I i I i l i l i O O.02 0.04 0.06 0.08 0.10 g ,

VOLUMETRIC AIR FLOW F

TOTA L VOLUMETRIC FLOW Figure 12. COMBINED DATA FROM [6] & [9] SIIOWING LOCUS OF 3%

HEAD DEGRADATION 20

O

. > f TM-1332

[ @ [g . ,

1.. 3 g i PAGE a i _ op _,2,

7. Murakami, M., Minemura, K. and Takimoto, M.; F/fects of Entrained Air on the Performance of Centrifugal Pumps Under Cavitating Conditions; Bulletin of the JSME, t .

23(183), September 1980, pp.1435-1442.

8. Neumann, B. and Lualdi, R.; Pumping of 3 resented at the :;

-Worthington Int. Pump Symposium, Kishm,October ev, U.S.S.R., Highly21-25, Aerated

,974. Liquids; Paper

}

9. Merry, H.; Effects of Two-Phase Liquid / Gas Flow on the Performance of Centnfugal Pumps; National Engineering Laboratory, East Kilbride, Glasgow, C130/16, IME '

Conference on Pumps and Compressors for Offshore Oil and Gas, June 29-July 1,1976,' '

pp.61-67.

i 1

i l

21

- __ - _- _____-_

ML20246L962

Text

SUMMARY

Navigation menu

Search