ML20215C796

Forwards Evaluation Constituting Reasons Why Mod to diesel-driven Containment Spray Pumps to Become Totally Independent of Ac Power Remains Unjustifiable

ML20215C796
Person / Time
Site:	Zion  File:ZionSolutions icon.png
Issue date:	12/10/1986
From:	Leblond P COMMONWEALTH EDISON CO.
To:	Harold Denton Office of Nuclear Reactor Regulation
References
2400K, NUDOCS 8612150418
Download: ML20215C796 (9)

Text

__

y a One Fwst Nemonal - Rese, Chcago, anois

- 7 ssess

~

A Reph to: Post 0500 Bat 757 Chee00.lenois80000 0767 December 10, 1986 l

Mr. Harold R. Denton, Director office of Nuclear Reactor Regulation U.S. Nuclear Regulatory Commission Washington, DC 20555

Subject:

Zion Nuclear Power Station Units 1 and 2 Diesel-Driven Containment Spray Pump NRC Docket Mos. 50-295 and 50-304 References (a): August 26, 1985 letter from S. A. Varga to Cordell Reed. I (b): November 19, 1985 letter from cordell Reed to H. R. Denton.

(c): July 7, 1986 letter from S. A. Varga to D. L. Farrar.

Dear Mr. Denton:

Reference (a) transmitted the NRC Staff's review of the Zion Probabilistic Safety Study and recommended two modifications to further reduce risk from Zion Station. One of these two recommendations was a request to modify Zion's diesel-driven containment spray pumps to become totally independent of A-C power.

Reference (b) provided Commonwealth Edison Company's comments on the staff's review. Included in these comments was a discussion of the reasons behind Commonwealth Edison's decision to not modify the contairunent spray pump.

Reference (c) provided additional justification for implementing the modification and requested that comunonwealth Edison reconsider its position. This submittal is transmitting the results of this reevaluation.

Reference (b) has been reviewed in detail and it does provide new insights. However, Commonwealth Edison Company has also carefully analyzed the appropriate accident sequences and has concluded that the proposed modification remains unjustifiable.

8612150418 861210 PDR ADOCK 05000295 P PDR 001

'\ \

r Mr. H. R. Denton December 10. 1986 The details of our evaluation are contained in the enclosure to this letter. We are available to discuss this issue at your convenience.

If any further questions arise regarding this matter, please direct them to this office.

Very truly yours, P. C. LeBlond Nuclear Licensing Administrator im Enclosure cc: Zion Resident Inspector J. A. Norris - NRR 2400K

EfCLO6URE In reference (c) the NRC staff requested us to reconsider our position concerning modification of the Zion diesel-driven containment spray pump, to make it a.c. - independent. We have examined this letter in detail and it provides new insights. However, it is still our viewpoint that the modification is not justified.

The NRC calculations show a population dose reduction benefit equivalent to $1.5 million for the modification. Commonwealth Edison responded, on November 19, 1985, that we felt that the calculations were based on overly conservative estimates for RCP seal LOCA's, loss-of-offsite-power frequency, and source term magnitude. Reference (c) explained that the dose reduction benefit would come from reduction in the consequences of the following:

a. seismically induced failure of offsite power and all service water pumps, worth $1.0 million;

b. loss of offsite power exceeding eight hours, with failure of the steam-driven aux. feedwater pump and all diesel-generators, worth

$0.3 million;

c. RCP seal-LOCA sequences, worth $0.2 million.

Reference (c) responds to our objections by explaining that:

1. the first two sequences are not dependent on RCP seal-LOCA assumptions;

2. A-44 work supports the loss-of-offsite power numbers used in the NRC analysis; and

3. use of new source term work would affect the results very little.

This response discusses each of the above sequences in turn, and then briefly discusses the source term.

The first sequence is a seismic sequence involving loss of offsite power and loss of all service water pumps. Because ceramic insulators in the switchyard are more fragile than any other important plant component, any earthquake large enough to damage other plant equipment will probably cause a loss of offsite power. The next-most-fragile important plant components are the service water pumps, and the service water pumps are needed to cool the emergency diesel generators. According to the original ZPSS analysis, there is sufficient probability of an earthquake of that size to make this an important sequence. We now believe that the service water pumps are not as sensitive to seismic damage as first estimated.

The NRC staff acknowledges this possibility on page 3-7 of the attachment to reference (a) which evaluated the ZPSS. Sandia's own seismic consultants, Jack Benjamin & Associates, felt that the ZPSS seismic analysis was conservative and that the seismic core melt frequency could be high by a factor of 2 or 3. This means that the $1.0 million benefit of the spray pump modification for the seismic sequence could be only $300,000 to

$500,000. Recent seismic work indicates that the benefit is even less.

Attachment 1 is a reproduction of Page 7.2-19 of the ZPSS, and it shows seismic frag 111 ties for several components, including the service water pumps. The horizontal axis represents earthquake peak ground acceleration.

The vertical axis represents probability of equipment failure. The solid curves represent the median values, including randomness, and the dashed curves represent the uncertainty at the 10% and 90% confidence levels.

These curves for the service water pumps suggest that at least some failures should be observed at 0.39 ground acceleration, and that for earthquakes as large as 0.69, a sizeable fraction of such pumps should fail. The Seismic Qualification Utility Group (SQUG) seismic experience data base

• includes 121 pumps similar to Zion's service water pumps, that have experienced earthquakes between 0.3 g and 0.69 Of these 121 pumps, only one failure occurred. This occurred on a deep-well turbine pump at a pumping station in chile, and it occurred on a corroded casing next to an old weld repair of an earlier crack. Because nuclear-grade pumps are minufactured, inspected, and maintained to higher standards, we believe this failure should be ignored.

Therefore, the data shows zero failures in 121 examples. These include earthquakes and pumps as follows:

peak Ground Acceleration # of Pumps 0.39 44 0.35g 56 0.49 4 0.42g 4 0.559 3 0.69 _.lq 121 The absence of failures in this data implies that the fragility curves for service water pumps should be shifted to higher g-levels. The tail of the lower bound should be no lower than 0.59, and the median curve should be much to the right of that. This shifts the service water pump curve considerably to the right of the Refueling Water Storage Tank (RWST) curve shown in Attachment 2, especially for lower g-values. This makes NWST failure the dominant seismic failure mode. If RWST failure occurs, the containment spray pump is of no value. This significantly reduces the benefit of the modification. We estimate this reduction to be a factor of ten or more. The $1.0 million benefit for the seismic sequence is reduced to $100,0uu sr less.

• EQE Incorporated, Eummyy_oLttle SeismiqJderntacy of_ Twenty C1annes of Plu_tcmenLEeqt u tred Ior the Cafe,_SlltLtyjow11 of_guglgar_P_laitn, July 1986 (Draft)

The second sequence described in reference (c) involves non-seismic loss of offsite power. The analysis assumes a loss-of-offsite-power frequency of .006/yr. for 8-hr. losses. In the attachment to reference (c),

the validity of the data is affirmed by reference to the A-44 work and NURBG-1032. Attachment 3 is a reproduction of Figure 2-1 of NSAC-85, Losses of Off-Site Power at U.S. Nuclear Power Plants, All Years through 1984.

This figure shows an occurrence rate, for losses lasting 8 hr., of .001/yr.

This graph reflects data for all nuclear plants for their total lifetimes to date. We believe that the EPRI data is correct for the average plant. Use of this 1984 EPRI data reduces the initiating event frequency, for loss of offsite power lasting 8 hr. or more, by at least a factor of six. This reduces the modification benefit for the second sequence from $0.3 million to $50,000. Because of Zion's six transmission lines, which are on two separate rights-of-way, and because of Commonwealth Edison's grid design, we believe that Zion is above average for offsite power reliability.

Therefore, the $50,000 benefit based on generic EPRI numbers is conservatively high.

The attachment to reference (c) acknowledges that the RCP LOCA o sequences would be affected by new seal LOCA assumptions. The ZPSS assumed leakage of 300 gpm per pump, 30 minutes after loss of seal cooling. Current work bupports leakage values of 25 gpm or less. We do not know the details of the sequences mentioned in reference (c), but we judge that this dramatic reduction in leakage nearly eliminates those sequences. This eliminates the

$0.2 million benefit of the spray pump modification for those sequences.

It is our view, therefore, that the total dollar-equivalent dose-reduction benefit of the containment spray pump modification is

$150,000 or less. This is without consideration of the new source term.

The attachment to reference (c) acknowledges that use of the recent source term work would reduce the dose estimates by a factor of 2 or 3. With the new source term, the modification's benefit is only $50,000 to $75,000 per unit.

The maximum $7b,000 benefit compares to a 1984 cost per unit of

$220,000. Therefore, the modification is not justified.

2400K

s- ,a 0.75 -

/

! F /

/ OFFSITE POWER a50 -

! CERAMIC INSULATORS

/

i / /

! a2s

-/ /

i /

! O I/ I I I I I  ! I I I I I I

0.1 0.2 0.3 0.4 0.5 0.6 0.s 1.0 2.0 3.0 4.0 5.0 s.0 to 10 I

i

'A e <

/ /

0.75 -  ! /

j F / SERVICE WATER PUMPS l /

am f f

/ /

.  ? a2s -

/

G 7 /

o I I I I di l I I i i 1 I

( at 0.2 a3 0.4 0.5 0.s as 1.0 2.0 3.0 4.0 5.0 s.0 to 10 1.0 l

/

/" /"

/

175 - /

,/ '$d

, s

, F f / AUXILIARY sLDG FAILURE OF SHEAR WALL

! am -

/

\ ,7 /

l 0.2s -

/ l 1

/ / a o I L# 1 1 1/ l l l l l 1 1 I k i at a2 0.3 0.4 0.5 0.8 as i

1.0 2.0 3.0 4.0 5.0 s.o to 10 $

~

i DAMAGE-EFFECTIVE GROUND ACCELERATION (g's)

Figure 7.2-2. Fragility Families for Key Components

,I f

4

I l

'O r ,e

/ l I 0.15 -

/ /

/ / @

0.50 -

/ / REFOELING WATER STOR AGE TANK

/

l 025 -

/ ,/

O I U I I U I I I I I I I 0.1 0.2 0.3 0.4 0.5 0.6 0.8 1.0 2.0 3.0 4.0 5.0 6.0 8.0 10

'O

,e ,e

/ /

'O 0.15 -

[ g' SOIL FAILURE BENEATH REACTOR BLDG I

p l l l 0.50 -

l l

N / /

l 7 0.25 - / /

8 / ,/

/ -

o i I/ I i _ P' s I a I I I I I I l 0.1 0.2 0.3 0.4 0.5 0.6 0.8 1.0 . 2.0 3.0 4.0 5.0 6.0 8.0 to i

1.0 ,

i /* p , /

0.75 -

/

I / DIESEL GENER ATORS 0.50 -

^

/ / t 0.25

/ / G Q.

7 j

/ / R i 0 I # I I I Is I I I I I I E l 0.1 0.2 0.3 0.4 0.5 0.6 0.8 1.0 2.0 3.0 4.0 5.0 6.0 8.0 10 w I

DAMAGE. EFFECTIVE GROUND ACCELERATION (g's) i Figure 7.2-2. Fragility Families for Key Components (continued) i

{

J l

. . . yNN:fP, r w 0

1 4

8 ,

9 1

3, 5 8 4 I 9

9 r 0 s

1 e.

v e 0 2

,ot u s 8 pn a ,

w s 9 eitim 1 s r u

- s0 - n o sf 3f o I 8 H r it a o n a r X ef y oh5 a u d n t st 4 r a ner : l a h ese2lae csgs y T

t eono ao e r l 7l wf it r

7 r

e g

o eret s sgr es n o

me w eea r

wgrep n

e L a vs 3 r er yI e 2dat n s a

h h elnhhv ee e r e t

u Y rTOTTe o i 6 o e F * * *

• H t y i n S e ,

a t

o r N M e s P i

I 5h T r e

n w a o P

, h T e r t e

g i

S I

4 n -  :

o f f

L O

i l l

4 A 8 f 9

1 I 3 o r

h e s g h e u t a s o

r e , s h o t W L s

r a

e y

g/ I 2 1

/ 2

" l la la t , e r

T o r o u f g e

g i

F I 1 a

r e

v A ,

0 0 8 6 4 2 0 1 0 0 0 0 0 0 0 0 0 MS3.uBa!gy E

7C

l l

i I

j 0.10 0.08 -

M S 0.06 Average for all years through 1984 1 E

! us

g -

a l

!! Total 7 g 0.04 -

Note: For the 3 most recent years-1982,1983,1984 C s

• There were 7 losses of off-site power

• Only 2 were longer than 30 minutes

• The lo'ngest was 2:45 0.02 -

/-Weather I /

• The average for ell durations was 0.045

/ events per site year I i , l I , I ,

0 , I i I i I i I i i 0 1 2 3 4 5 6 7 8 9 10 l

Longer Than This Many Hours y O

I E Figure 2-1. Losses of All Off-Site Power Per Site Year Lc.sr Than X Hours i 3

, u

! _