ML20083B494

Georgians Against Nuclear Energy Amend to Petition for Leave to Intervene.* Suppls 910614 Petition W/Listed Petitions

ML20083B494
Person / Time
Site:	Vogtle
Issue date:	08/09/1991
From:	Carroll G GEORGIANS AGAINST NUCLEAR ENERGY
To:	Atomic Safety and Licensing Board Panel
References
CON-#391-12184 91-647-OLA-2, OLA-2, NUDOCS 9109250102
Download: ML20083B494 (9)

Text

- - _ _ _ - _ _ _ _ _ _ _ - _ _ _ _ - _ - _ - _ _ _ - _ _ _ _ _ - _ _ _.

/ 3 / f//-

s

( 6!,f D 5

' W i.

UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION 91

.n 15 I? :B ATOMIC SAFETY AND LICENSING BOARD Defore Administrativo Judges:

Sheldon J. Wolfe, Chairman Dr. James H. Carpenter Dr. Thomas S.

Elleman In the mattor of GEORGIA POWER COMPANY, ET AL.

(Vogtle Electric Generating Plant, Units 1 and 2)

Docket Nos. 50-424 & $0-425 - U # ' ",

ASLBP No. 91-647-OLA-2 August 9, 1991 GANE AMENDMENT TO PETITION FOR LEAVE TO INTERVENE GANE appreciates the consideration of the Atomic Safety and Licensing Board in schedt' ling a prehearing conf erence. GANE also appreciates your granting of standing by accepting our recently established standing. The rules you forwarded have been read and honored in this filing. If thero is a need to reestablish our spokesperson's right to represent GANE, an affidavit is attached from our treasurer, Dennis Hoffarth.

GANE supplements its June 14 Petition with the following contentions:

1. The NRC Branch Technical Position EICSD 2 " Diesel-Generator Reliability Qualification Testing" dated 11/24/75, of tho standard Review Plan (Appendix 7-A of NUREG-75/087) established a reliability goal of 0.99 (at a nominal 50%

confidence level) for a plant's diesel generator system. GANE contends that Georgia Power's request to delete the 5:100 test failures standard will unacceptably lower the confidence level in the statistical reliability of the diesels at Plant Vogtle to prevent a aangerous station blackout. (The NRC's basis of stating diesel reliability in terms of a statistical confidence level of 50% is used in every calculation of reliability included in this petition to assure consistency with the NRC's basis. Note that at higher statistical confidence levels the reliability would be lower.)

2. Georgia Power's request to delete the 5: 100 criteria runs counter to NRC Regulatory Guido 1.108 which sets forth periodic testing provisions "to

-)Y) 9109230102 910809 PDR ADOCK 05000424

}

G PDR

_ - _... - _ - - ~ - - -

GANE Cont entions Page Two provide a basis for taking those corrective actions needed to maintain high inservice ro11 ability of installed diesel generator units." The reliability required in Regulatory Guide 1.108 pc sition C.2,a (9) is to " demonstrate the required reliability by noans of any 69 consecutive valid tests (per plant) with no failures." Completing 69 tests without a failure yields a.99 reliability at a confidence level (B) of 50% and is therefore consistent with the Branch Technical Fooition shown in Contention 1. This is shown by statistical calculationi B=1-(p)n 69 B = 1

.99 B =.5 or 50%

(B = confidence coef ficient, p = probability of starting (reliability),

n - number-of consecutive-tests without failures)

Ref erence: Handbook of Probability and Statistics with Tables, 2nd Edition, Burlington/May After the reliability demonstration, Reg. Guido 1.108 C.2.d presents a logical series of corrective actions to increase the testing frequency should a diesel generator system fall below the.99 reliability basis at a statistical confidence level of 50%.

1) If the number of failures in the last 100 valid tests is one or zero, the test interval should not be more than 31 days (p =.983 or.993 0 B =

50%),

2) If the number of failures in the last'100 valid tests is two, the test interval should be r.ot more than 14 days (p -.973 9 B = $0t).

3) If the number of failures-in the last 100 valid tests is three, the test interval should be not more than 7 days (p'=

.963 0 0 - 500,

i

4) If the number.of failures in the last 100 valid tests is four or more, the test interval should be not more than 3 days (p =.953 0 A = 50%).

This standard is consistent with the 99% reliability goal _at a 50%

confidence level-of the branch technical position since increasing costing is required when there is more than 1 f ailure per 100 tests.

GANE contends that-Georgia Power's request is very distant f rom this 7

standard in allowing a 31-day interval for.95 or' greater rollability and a 7-day interval for.90 or greater reliability, rurther, the reduced' sample size undermines'the statistical ability to ascertain when the diesel has fallen below:the marginal level of reliability (90%), a significant warning.

flag of declining diesel reliability.

?

3. Georgia PoweE's request to delete tho 5:100 criteria for establishing diesel generator reliability should not be allowed because it is inconsistent i

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

__.,_..rm

..___._-_.-,..._,m.._._--....-

..-,-..,_.-,,mr,..

.i

GANE Contentions Page Three with the basis and recommendations of Generic Letter 84-15. Enclosure 3

' Example of Diesel Generator Performance Technical Specification" clearly maintains the "failutes in the last 100 valid demands" as the basin for t he

• Reliability Levels Remedial Actions"

(" Reliability Program" item 2).

4. Georgia Power's request is inconsistent with the requirenente of Regulatory

-Guide 1.103, regulatory position C.2.d.1-4 which Georgia Powet congnitted to in its FSAR and represents a significant reduction in the level of safety J

provided. Eliminating the 100 start basis from the survelliance testing schedult effectively reduces the sample size supporting current diesel reliabi,ity by a factor of five. Georgia Powet presents no reliability calculations in their 50.92 evaluation to show that their proposed test plan neets-NRL's requirements of demonstrating 99% reliability per plant or 954

-j

- reliabilit,' per diesel at a 50% confidence level GANE contends that Georgia j

Power's teei nical specification change request must be denied as it conflicts with the vital safety analysis upon which its operating license is based.

5. Georgia Power's request selectively includes those portions of the Generic Letter it desires and deletes those criteria it does not want. Georgia Power includes no criteria from the GL 84-15 that would declare the diesel inoperable due to excessive failures and low reliability (regardless of how bad the. diesel performance may be) such as that of." Table 4.8-2 of Attachment to Enclosure 1" or " Items 2 and 5 in Attachment to Enclosure 3" of the Genetic Letter._ GANE contends that the methodology Georgia Power intends to use to establish operability of the diesels will not provide adequate indication of declining. generator reliability.

j l

6. If the 100 valid tests basis is eliminated and replaced 'only by a "rero out l

of 7" and a 'one out of 20" valid tests basis then the fundamental diesel

-reliability assumptions of the NRC's Standard Rev4ew Plan are being abandoned-(i.e. 0.99 reliability at 50% confidence level) as well as the requirements of-

~

95% reliability per diesel generator stated in. Generic Letter 84-15. Thus,

- with this change the safety assumptions of Georgia Power's and the NRC's

~

- accident analysis are no longer supported by the surveillance testing i

-~m-un

'e.,

,v

,en-_n, ween.e,,--,

n, w-s wavn.

4 t

(

GANE Contentions Page Four provisions of the Technical Specifications, L;io failurgs out of 7 valid ten u yielda_caly_at l

8 = 1 - (p)n 8a 1 - (p)7 f

.50 = 1 - (p) 7

(.$)1/7 =.906 p=

reliability of 90.6% at a $0% confidence level.

(B = confidence coefficient, p = probability of starting (teliability),

n = nurrber of consecutive tests without f ailures)

Allowing one failure in 20 and using the F-distribution tables to calculate the sum of the terms on one end of the binomial expansion establishes only at (mg/ma) F p/(1-p)

=

(38/4)* 1.171 p/(1-p)

=

p

.917 reliability of 91.7% at a 50% confidence level.

(p a probability of starting, 3 = number of successful tests, n = total number i

2(n-S+1) degrees of freedom, of tests, mi - 2S degtecs of freedom, my a

F = F-distribution table value for S=.5)

References:

Handbook of Probability and Statistics with Tables, 2nd edition, Burlington

'y, Probability and Statistics in Engineering and Management Science, nes/ Montgomery GANE contends neither 90.6% or 91,7% reliability meet the NRC's reliability goal or would warrant the decreased t w frequency per the Genoric Lettor, rather these probabilities would requ

.ccelerated testing that borders on the disqualification limit of 90%. The NRC's intent of requiring a 99%

reliability per plant or a 95% reliability per diesel generator as the minimum desired level cannot be assured (even at the f airly low statistical confidence level of $0%).' without the 100 testa criteria basis included in Technical Specifications. GANE contends that 0:7 or 1:20 tests is just too few tests to establish the high diesel reliability (95%) with any reasonabic statistical confidence.

~

. - a

___,i._.-

GANE Contentions Page Five

7. GANE contends that allowing "2 failures in 20" as proposed by Georgia Power f or the threshold to initiate corrective actions only establishes a t p/ (1-p) = (mg/mg) F p/ (1-p) - 36/6

• 1.1006 p

.868 reliability of 86.8% at a confidence level of 50%.

GANE contends that Georgia Power proposes to change Technical Specifications to allow a diesel that fails to meet NRC criteria for operability (90M to continue service with only increased surveillance testing. Georgia Power's request represents a significant reduction in thu assurance of diesel reliability below that previously required and below that established by the NRC for safe operation and as such must be denied. (Reg.

Guide 1.108) in ' orte

8. GANE contends that while decreasing the testing frequency may s

circumstances be warranted where high reliability has been established and considerations of wear and increased 'end of life" f ailures derninate, such is not the case with the Vogtle diesel generators.

The failure of the Vogtle diesel generator during the Site / Area emergency of 3-20-90 as described in Nureg 1410 (Sections 2.4,3.2,3,3,6.5, Appendices 41&J) and the f ailures experienced by the Vogtle generators both bef ore and after the Site / Area Emergency exhibit a characteristic of the " shake down" or

" burn in" period of typical failure curves below:

l l ""W W

.Q$ "

P FA Lunf u

l l

l 2AM1 l

t f/IILUElb e

7741E ~+

4

'r'/ PILA L l'AttuXC LUEVE n,.

_=-=-

In circumstances such as Vogtle's, increasing the test frequency would actually be beneficial to safety to further ' shake down" any other hidden, neglected, or otherwise unrecognized problerns (like CALCON switches, air start valvo clearances and binding, voltage regulator problems) and reduce operator error ' events (such as inadvertent starting and improper duplex filter operation) by increased operating experience.

I i

f i

Gane Contentions Page Six

9. GANE contends that the true reliability of the Vogtle dieseln may not have

.been accurately evaluated or reported to the NRC as evidenced by several repeat past cited violations including 50-424/87-57 as well as NRC investigations that are still pending (Augusta Chronicle 9/17/ 90). In this light, the significant threat posed by a station blackout and the fallute of the Vegtle diesel to perform during-the Site / Area Energency of 3/20/90 (Nureg 1410 (sections 2. 4, 3. 2, 3. 3, 5.1, $. 4, 5. 5, 6. 5, 6. 6, 6. 7 ] ), approving this change proposed by Georgia Power at this time would be adverse to safety.

10. GANE contends that lack of operator familiarity was shown to be a significant f actor in the 3/20/90 Site / Area Energency and increased, rather

[

than decreased, testing is recommended to offset the hazard posed by human error. Section 5.5.2 of Nureg 1410 discusses the lack of familiarity among the i

operators with the annunciator load sequencer and diesel pneumatic control j

systems, Section 5.4.1 of Nureg 1410 cites lacl. of operatot familjarity with the diesels and the disabling effect of the c(nfusing array of instructions, Section 10.3.2 (Nureg 1410) investigates diesel generator lockout following i

shutdown and finds that the control circuits are complex and not well understood, while 10.4.3 shows that operators have had difficulty diagnosing causes. The diesel is so crucial a factor.in preventing station blackout and so complicated, a rigorous testing schedule is warranted. Enclosure 3 of the ACRS Briefing on Status of Staff pollowup Actions Resulting from the Invertigation of the March 20, 1990 Incident at Vogtle Unit i dated 3/25/91 (3-b) reiterates this in its status report of the ongoing evaluation of diesel generator operation saying, "Because of the time between these tests, the operators may not be as alert and prepared to diagnose EDG problems as they might be if EDG operating procedures were used more frequently " Section 3-1

'for.the ACES Briefing also notes the importance of operating experience to f acilitate diagnosis. Thu 4/11/91 ACRS Meeting Transcript contains an enlightening discussion of the importance of the EDGa-to prevent station blackout but more significantly on p. 61 Mr. Rubin's discussion highlights tho infrequ-ncy of operator experience.with toe diesels and points out that testing is nometimes the only experience en operator gains with a diesel

' generator prior-to_ emergency-need. if increased testing cannot be required

~ under the jurisdiction of the 2.714 rules, certainly decreased testing cannot be allowed and Georgia Power's request must be denied.

11. GANE has-filed a Freedom of Information Act Request for Licensee Event-Reports and-Special Reports about the Vogtle Diesel Generators after not finding most of these documents in the Public Document Room.and being denied a l

l I-

- _. _ _. - ~ _.. -. ~ - _ - _.

GANE Contentions Page Seven request by Georgia Power. We ask that these docunents be allowed into the c

__ discussion as they becone available. For now, we have reviewed an 8/7/90 Special Repert on Valid Diesel railures and a 2/28/91 Licensee Event Report which raise grave concerns as to the current statistical reliability of the

~

generators in service. The 8/7/90 report shows 6 valid failurea in the last 79 tests (projects to 7.6 valid f ailures in 100 tests) or generator 18 and $

valid failures in 43 valid tests for the 2A generatot which projects to 11.6 failures per 100 tests. The later report we have been able to see, 2/28/91, shows the 2A generator record a bit improved at 6 failurus per 76 tests for a 7.8 out of_100. Reg. Guide 1.108 requires special evaluation of the generator units if the nunter of f ailures in the last 100 valid tests is eeven or more and requires the licensee to describe corrective measures to increase the reliability of the generator, provide an assessment of the reliability of electric pnwer and provide the basis for continued plant operation. GANE boldly contends that these diesels may not NOW be qualified for nuclear service until reestablished. If the 2.714 rules don't allow this party rollet l

on that. level, we officially contend that this poor record of diesel reliability at the very least warrants the rigor of the original testing

~

criteria of 5:100 and 1:20.

12. GANE contends that the TDI generators are rated for 25,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> of

-service and expected by the industry to operate for 50,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br />, that this is common knowledge, end that excessive wear of the diesels by testing as f requently, as atl3 day intervals is not a legitinate concern and does not provide a rationale for decreased testing of the problematic diesels at Plant Vogtle.

134 GANE contends that Georgia Power has a history of reluctance to face the problems with its diesel generators as well as a history of-not availing.

themselves of INPO information and NRC notices that would help them address chronic and' serious problems with their generators. Section 6.2 of Nureg 1410 notes this lack of use of available information. GANE's experience in a previous petition concerning the high jacket water temperature. switch bypass which is now pending in appeal-before the Commission le that_ Georgia Power has avoided the issue and spent its energy on attacking the legality of GANE(a

-method of expression rather than the technicalities of why their diesels don't work properly. Section 5a of the.3/25/91 ACRS Briefing-implies that testing the diesels offsets Vogtle's lack of use of information notices, etc. and supports our contention that Georgia Power should retain the 5:100 and 1:20 criteria for their'Vogtle diesels.

[ --, _

- -.. -. - - ~ ~

~. - - - - - - - - -

l i

i l

'r!

GANE Contentions ha'dM Eight

14. GANE centends that EDG reliability is acknowledged by the !}9h tbfpiWi.112 :15 most significant role in reducing the threat of the serious situation of 4

station blackout (ACRS Transcript of 4/11/91 Meeting) and to relax:the. testing

requirements for the Vogtle Plant which had a Site / Area Emergendh"bn".3/20/90

{

would send a confusir.g messago to other power plants which should be able to i

learn from the mistakes at Vogtle. The request should be denied.

15. GANE contends that neintenance of the current testing requirenents will l

help Vogtle operators identify the problems with their diesels. If they are t

having trouble proving the diesels at the level of reliability required by the NRC perhaps they will have more incentive to take corrective measures, such as

~

finding a way to repair the CALCON sw' itches which have shown chronic fallutes throughout their operating history, probably due to contamination of the pneumatic logic -(Nureg 1410, Appendix I and ACRS Brjefing 3/25/91, Section 3).

GANE believes this damaged logic should be replaced with a more rugged electronic logic. Perhaps then their diesels would prove reliable.

E I

Respectfully submitted, h'7/ hN Glenn Carroll Representative for GANE

,e

= Dated %n'd signed in Decaturi Geor'gla

' " + * - ' ~ - ~ ~ - ~ ~

t-August 9,- 1991 i

e e

d h

P

(--

-