ML20087P090

Reply Brief on Commission Review of ALAB-729 & ALAB-744. Plant Operation Should Not Be Approved Until Safety Issue Resolved.Certificate of Svc Encl

ML20087P090
Person / Time
Site:	Three Mile Island
Issue date:	04/03/1984
From:	Weiss E HARMON & WEISS, UNION OF CONCERNED SCIENTISTS
To:
References
ALAB-729, ALAB-744, NUDOCS 8404060034
Download: ML20087P090 (32)

Text

~ , S S. l .

~

S iRC LCS 04/03/84

.gA I?R ~5 R:) np' UNITED STATES T #iERICA NUCIEAR REGJINIORY CONISSION BEFORE 'I1IE COtEISSION In the Matter of )

)

MLTROIOLITAN EDISON C04PANY ) Docket tb. 50-289

) (Ibstart)

('Ihree Mile Island Nuclear )

Station, Unit No.1) )

UNION OF CCtCERNED SCIENTISTS' REPLY BRIEF ON REVIEW OF AIAB-729 Arm AIAB-744

1. Environmental Qualification of Safety Equipment

'Ibe core of both the . Staf f _ and GPU's argunents is that the Canmission

,, need -not 'either by certification or litigation resolve the question of whether

'IMI-l can be safely operated despite having equipnent unqualified to survive an accident environment because the -entire matter is being dealt with

" generically." E.g., ' NIC Staff's Brief Concerning the Ccmnission's Review of Specific Design Issues in AIAB-729, March 19,1984, p.12 (hereinafter " Staff Brief"); Licensee's Brief ~on Review of AIAB-729 and AIAB-744, March 19,1984, pp.17,320 (hereinafter "GPU Brief") .

As LCS argues. in our main brief, this disregards the fact that the

!- environnental qualification issue has two dimensions, one generic and one plant [-specific. . Ihis distinction hus always been recognized by the Commisston and also by the O)urt of Appeals in UCS v. NRC, 711 F.2d 370 (D.C. Cir.1983) .

'Ihe !EC is not free to license plants on the grounds that their safety problens are not unique. See UCS Brief on the Ctanission's Review of AIAB-729; l March 19,1984, p. 4-6 (hereinaf ter "UCS _Brief") . ,

-8404060034 840403 PDR ADOCK 05000289 0 PDR _ , _

7. ,,

Indeed, the CPU and Staf f position is in all essential elements the same L ~as the . Staff's . long-discredited position on the so-called " unresolved safety problems" described L and overturned by the Appeal Board first in Gulf States

Utilities Co. (River Bend, Units 1 and 2), AIAB-444, 6 NIC 760 (1977). @e NRC used to routinely grant construction permits and operating licenses to new plants without even disclosing in its safety evaluations that the plant in

question was subject to an unresolved generic safety problem. (In fact, the list of unresolved safety problens was itself kept secret by NRC for years.)

In the River Bend case, supra, the Appeal Board categorically rejected the renarkable argunent that simply because a safety problem applies to more than one.. plant and is therefore part of a generic program, that problem can be

. disregarded in individual construction permit cases.

Wis doctrine was applied to operating license cases in Virginia Electric and Power Co. ' (North Anna Nuclear Power ' Station, Units 1 and 2) , AIAB-491, 8

~

q. NRC 245' (1978) .- 'Ihe Appeal Board held therein:

'of course, these ' unresolved' issues cannot be disregarded

' in individual-licensing proceedings simply because they.also -

i - -

have generic . applicability; rather, for an appl'icant to succeed, there must be some explanation why construction or-1 operation can proceed even tnougn an overall solution nas not

'been found.-

.In River Bend,1we'said that such explanations should appear 'in the Safety Evaluation Report for the facility. We also described generally the t'ype of reason which would be

-sufficient toilet construction to [ sic] go on in.the face of an

, unresolved generic question. Where operation of. a facility is -

J? involved, similar analysis ^ is necessary; but, as to certain-  !

issues, the justification for giving an applicant the green ilight can obviously be more difficult- to come by. Ebr example,

.tne reason orten given ror anowing construction-activity is

that there is still: time to find'a solution and build:it into the plant's design. At the ' operating ' license stage, that reason.is'not available. But there may be one-or more other

. justifications for permitting tne plant to operate. The most coupon are that a solution satisfactory for the particular

. facility has been inplemented; a restriction on the level or nature or operation acequate to eliminate the problen has been zimposed;-or the safety issue does not arise until the-later-

. years of plant operation. '

Id.(at.248, emphasis added.

We same principle applies here. We NRC ray not use the label " generic" as a shield to avoid the plant-specific dimensions of a safety problem.

-Indeed, the circunstances here are more compelling than in the River Bend and North Anna cases. Here, the factual evidence of record dmonstrates that much safety equipnebt in WI-l is unqualified and the record patently contains no evidence supporting a finding of " interim" safety nonetheless.

Nor does the fact that the Cburt in ICS v. NRC permitted NRC to determine in the first instance. whether the plant-specific safety determinations may be made by rulemaking or adjudication alter these principles when NRC has done neither, %e question- here is whether, in the face of a safety issue fairly raised in a hearing, a plant can be permitted to go into operation when there has been no lawful determination in any context that it is sufficiently safe.

Were are several other pints requiring some rebuttal. First, both GPU and. the Staff attenpt to make much of the fact that (CS withdrew its

-sponsorship of Contention 12; to what effect is unclear. UCS successfully mcued the ASIB to accept the Cbntention, did extensive cross exanination and filed detailed and exhaustive findings of fact on this ' issue. GPU claims that the Board, when it adopted the (bntention, limited the scope of the issue.

GPU Brief, p. 18-19. We only limitation placed in this issue by the Ibard, as GPU well knows, . was its limitation to safety equipnent in the containnent and auxiliary buildings. In addition, the Board's second question in connection with ~ UCS Contention 12 covered virtually all of the ground contained in the contention:

2.' Which items of Ibgulatory wide 1.89 [the Ibg. Wide implenenting GDC 4 at the time the question was formulated]

have been grandfathered with respect to MI-l? Explain any justification for allowing restart without ccmpliance with the granatatnered items.

anphasis added._

_4 _

mis . central question -- the safety justification for operating with mqualified equipnent -- was never : answered by GPU or the Staff. It is

. disingenuous of GPU to imply that some " limitation" of the issues by the Board

~ redounds to their benefit.

'Ihe only significant . limitation was the one unilaterally imposed by the '

- Staff and unequivocally rejected. by the ASIB. %e Staff presented evidence only on the qualification of equipment to survive a design basis small break

.ICCA,' with 1% fuel failure. We ASIB had the following to say on that subject:

1154. We have not been able to discern why the Staff approached BQ/tCS-12 with an analysis of a design basis small-break IOCA with its assumption of cne percent failed fuel. When this very narrow testimony was presented, we

. questioned counsel about the Staff's rationale for its change of position.since the earlier testimony, but our efforts were not productive. Tr. 21,885-92. 'the ' Staff's proposed findings are no more helpful. Nowhere has the Staff explained how it uses the Board's superfluous agreement that the Staff may limit its analysis and testimony to accidents that are clear and close analogs to the TMI-2 accident (Staff proposed finding

= Paragraph 6) to justify an analysis of a design basis LOCA with an assunption of only one percent failed fuel.

~1155.~ 'Ihe analysis by the Staf f's witness of a design basis small-break IDCA with .its assumption of one percent failed fuel is not one of the " clear and close analogs to the 'IMI-2 event.. ." pranised by the Staff (Tr.19,487) . We Staff's .

claim that it presented such-an' analysis (proposed findings

. Paragraphs 6 and 7) is mfounded. '1he design basis small-break

-IOCA does not envelope-the TMI-2 accident. The Staf f's -

analysis has been useless to the Board in deciding our questions on UCS Contention 12. It added nothing to our . .

understanding of the ability of. the safety equipment in 'IMI-1.

~ at restart to' withstand the harsh environment of a .'IMI-2 type accident or in accord with the criteria of Ibgulatory 02ide.'

A. 89.-

.1156. In our view the Staff has defaulted and the decision -

, : must rely chiefly on Licensee's testimony and argument.

14 NBC 1211, 1402, footnotes omitted, emphasis added.

We ' are amazed by l GPU's ' argunent that there - is evidence of record ' to L support ~a , finding that the plant can be safely operated and that the Licensing Board made " substantive ~ findings" on that evidence. GPU Brief at 16. ..As we 4

. noted in our main brief, the Staff was fomd to have " defaulted" on the issue and GPU's evidence was found by the Board insufficient even for a " qualitative judgment of the risk of allowing interim operation..." 14 NRC at 1402, 1403.

- In addition, GPU's testimony that 95% of equipent was qualified was demonstrably false. See UCS Brief, n.1 at 3.

In fact, We Board found:

1181. We have not addressed each of the UCS proposed findings because we believe that they have prevailed to the extent that UCS has demonstrated that all of the safety eculpment at TMI-1 will not meet all the criteria of Regulatory Guide 1.89 at the time of restart.

Id. at 1409, m phasis added.

nat is the pertinent " substantive finding" that the Board made.

It is regrettable that the Board's failure to address UCS's proposed findings in detail in favor of a general finding that we " prevailed" has had

. the consequence that for both the Appeal Board and the Canission, this issue

-has been presented in largely abstract terms. (ES's Proposed Fir. dings of Fact and Rulirgs of law on UCS Contention 12, July 13,1981, analyze the record in

-detail .and we urge the Gmmission to read them. 'Ihey demonstrate, inter alia, that not only did the Staff make no attempt whatever to determine whether the equipment - in MI-1 can withstand the -accident environment which occurred during the MI-2 accident (Tr. 21,913-21,916) , . it had not even evaluated the environmental qualification of equipment installed or modified as a direct result of the MI-2 accident. Wis includes such ccrnponents as position indicatkon for ' the IORV. and safety valves and instrinnentation for automatic EEW initiation. See IES Proposed Findings of Fact and Rulings of Iaw on (ES Contention -12, para. 669, p. 269-270. Nor does the Staff have any specific plans to review such equipent.

Id. at para. 670,.p. 270.

Were are many other' specific deficiencies treated in IES's proposed findings. %is is .

scarcely an abstract issue.

w

In this connection, we draw to the Ctxnmission's attention a serious misrepresentation by GPU, which states in its brief:

Ch a more general level the Licensing Board observed that Licensee, fran the evidence it presented, was making good progress in complying with lE Bulletin 79-OlB and Commission Order CLI-80-21. Id. at 1400.

~

GPU Brlef, p. 12.

te ASIB never made any " observation" that GPU had made " good progress."

Ch the contrary, at the cited page, 14 tE 1400, the Board simply described GPU witness Braulke's (inaccurate) evidence that 95% of the equipment was qualified and stated:

It appeared that the Licensee was making good progress in conplying with IE Bulletin 79-01B and Ccanission Order CLI-80-21. We turned to the Staff witnesses for verification.

Daphasis added.

We Board then proceeded to find that the Staff " defaulted" in supplying any " verification" (14 . NRC at 1402) and concluded that Braulke's testimony itself was insufficient to support even a " qualitative judgement of the risk of allowing interim operation.. ." Id. at 1403. GPU does not cite this crucial section of the decision. We failure of the evidentiary record to support a safety judgment is precisely what made it necessary for the ASIB to seek refuge in the Cbmmission's " generic" treatment of the issue, as a substitute for a factual finding favorable to GPU. See Id. at 1403. 'Ib suggest to the (bmmission that the ASIB found " reasonable progress" is a gross misrepre-sentation of the truth.

Bodh the Staff and GPU arg ue that 'IM I-l should not be treated

" differently" than other reactors, citing CLI-81-3,13 IK 291 (1981). E.g.,

Staff Brief, p. 4. Of course , that decision dealt only with NUREG-0737

" lessons . learned" coupliance deadlines, and is, therefore, not directly on point. N rever, the (bmmission "enphasized" that it " expects the Doard to find to the contrary when the record so dictates." ,Id at 295-6. Ibrhap's most

~

imp)rtantly, (CS is not asking the 0:xnmission to treat this plant dif ferently than other reactors. tb plant should be permittted to operate without either a) otrnplying with environmental qualification requirements, or b) providing plant-specific data justifying a conclusion that interim operation pending qualification would not pose undue risk. Se Cbmmission has always accepted this principle. It is the underlying principle of CLI-80-21. GPU cannot be heard to complain that it is being treated " differently" wnen it has never rebutted the factual record in this case that substantial safety equipnent in

'IMI-l is not qualified.

Another misstatement by GPU and the Staff requires correction. Both state that the generic rulenaking and environmental qualification "subsuned" the 'IMI-2 lesson learned. The Staff states that the rulemaking " considered the implications, of that accident in its determinations regarding the radiation source term appropriate for equipnent qualification purposes."

Staff Brief at 11. See also GPU Brief, n.12 at 11. In fact, in developing the rule the (bmmission expressly excitried the 'IM I-2 " lessons learned" and stated that they would be considered separately.1! -

Finally, the Staff notes that "at least one" piece of equipnent has not been shown to be qualified - the Bailey E/P (bnverters for the EEW flow control valves. 'Ihe Staff claims that it is still " reviewing" the licensee's justification for these conponents. Staff Brief at 13.

Ebr one thing, we find it extremely difficult to believe that only one piece of equipnent remains mqualified. As noted in our main brief, as of the last published review, the majority of the equipnent items requiring qualifi-

-1/ "2hese p3sitions were developed prior to the 'Ihree Mile Island thit 2 event. Any reccanendations resulting fran the staf f's review of that event will be provided later." NUREG-0588, at lii, Decenber 1979. See also, CLI-80-21,11 NBC 707, 716 (1980) .

c

cation had not been demonstrated to be qualified. LCS Brief, n.1 at 3. Since then, meetirgs between the Staf f and GPU have taken place and information has been exchanged, but as yet, no evaluation has been published. See also Staff Brief, n.8 at 13. Ibr, of course, has this material been subject to questioning. Based on past history, we would be astonished if, in fact, only one piece of unqualified equipnent rmained.

Even more imprtant, the Staf f is at best disingenuous when it states that GRI's justification for these cmponents is still "under review." We have just learned that the Staf f unequivocally rejected GPU's specious

" justification" before filirn its brief, for the same reasons we pointed out in LCS Brlef, pp. 8-9.

01 March 8, 1984, Mr. James van Vliet, NRC's Project Manager for T4I-1, noted that GPU's proferred justification for continued operation with unqualified empnents in the ITW systems relied "upon a probability argment based on low probability of events, which is made even lower by the ISI program and then as a third justification, I guess, you discuss feed and bleed." Transcript of Meeting with GPU on T4I-1 Enviz'onmental Qualification, March 8, 1984, p. 5. Mr. van Vliet rejected GPU's claim of justification for continued operation (JCO) as follows:

- We have not in the course of environmental qualification progrms accepted probability [as] an argment for JCO's.

_Id. at 5.

y And I think what we would have to say is simply what I

. said and what Walt [Jensen] said, we can't give you credit, design basis credit for feed and bleed at this point and we have to reject it.

_Id. at 12 Just to samarize where we are right now with respect to the ((CS Show Cause] petition, ard also we need - I think I want to point out at the present time for reasons we explained earlier, we cannot accept the JCOs on your E/P converters.

Id. at 177, emphasis added.

'Ibe Staff also seriously misleisds the (bmmission by implying that the

- "Staf f's review to date" forms the basis for the renainder of its statement that "it appears that test data is not available to demonstrate the qualifi-cation of at least one cmponent, namely the Bailey E/P Oanverters for the EN flow control valves." Staff Brief at 13.

A more truthful statement would be that GPU has identified at least one cmponent for which qualification has not been denonstrated and GPU's justifi-cation for continued operation for that cmponent is unacceptable. (In LCS's view, the fact that GPU advanced a low probability argunent and reliance on feed and ' bleed as justification for continued operation demonstrates its intransigence or ignorance of the Omnission's criteria for acceptable IOs and the Appeal Board's ruling on the viability of feed and bleed in this proceeding. See AIAB 729,17 NIC 814, at 852.)

In reality, the Staff's - review of the status of environmental qualifica-tion at MI-l is in an embryonic stage, as a perusal of the transcript of the March 8,1984 meeting on 'IMI-l environmental qualification confirms.

2. Emergency Feedwater Reliability GPU's r>rief ' attempts to mw confusion concerning the origin - of the Board's ' analysis of EEW reliability and ~ of the - system's faikure rate and fundamentally hinges 'its appeal on the familiar; refrain that it should not be treated " differently" than other plants.

Se, ASIB's concern about the reliability of the 'IMI-l EEW system stems

- directly 'fra : the 'IMI-2 Jaccident. 'Ibe unavailability on demand of -the EEW system was a potential contributor to the' severity of the accident. 14 NRC at 1355. 'Ihe' Board noted that "[t]he NIC Staff (IE) considered that improvunents

,in plant . procedures. and ~ technical specifications were an immedlate need in

' order: to limit the possibility of 'a similar ' occurrence.. ." Id.~ at 1356.

9 4

Ee Ibard went on to note that the MI-l EEW system was not safety-grade, the overall system did not meet th$ single failure criterion, "and a high degree of reliability could not be . anticipated." _Id . De Board observed correctly that, while it made no judgment about the adequacy of the original design, "our task is to decide whether the requirements of the Q:mmission's orders have been met and whether the' improved EEW reliability is adequate to protect the health and safety of the public." Id. It noted that the adequacy of the EEW design was challenged by (CS and is the " thrust" of Ibard Question

6. Id.

GPU is extremely critical of the Board's use of' the Staff's historic data for failure of safety-grade EEW systes of 1 in 25 reactor years. Id. at 1356-1357. GPU Brief p. 32ff. GPU's criticism is based in a fund 6 mental misconstruction of the use to which the Ibard put this data. Since, as noted above, the orginal 2I-1 EEW system was not safety-grade, but will be in the "long-term" ' (i .e . the first refueling outage af ter restar t) , the ASIB felt that. the adequacy of the future system shotild ' fairly be jtriged by cmparison with the historic failure ratos,of safety-grade EEW systems. Id. at 1356-1357. Eus, using safety-grade systems for cmparison (when MI-1 will J,~~ .

not be safety-grade at restart) is generous to MI-1.

De high historic failure rate for safety-grade EFW systems -- 1 failure in 25 rea5 tor-years - and the high rate of challenge to EEW systes (Id. at 1357), sled the ASIB to lock further for assurance that. the MI-1 system will

^

be sufficiently reliable to . perform its vital safety functions. Berefore ,

w the Ibard specifically sought a plant-specific quantitative failure analysis

, . e .- '

for'WI-1: +

,7 A

. We knew that the licensee proposed upgradin3 the EFW system to safety-grade but in view of the past record with

' safety-grade' EEW systems at other plahts, % felt cmpelled to examine the reliability of the system.

Id. at 1358.

/ *y.

Eus, the ASIB used this data in a correct and prtdent manner, not as a basis for a conclusion on the WI-l EEW system, but as a reason to inquire in greater depth about the reliability of the MI-l system. GPU deliberately chose not to present a failure rate analysis of the WI-l EEW systs, despite being given more than ample opportunity by the ASIB. Licensee performed no evaluation of the probability of loss of main feedwater at MI-1. We generic data for five B&W plants over a two year period (i.e.10 unit-years) showed a loss of main - feedwater frequency of 0.3 per plant-year. GPU estimated that the uncertainty attached to this frequency is less than a factor of 10. Tr.

16,618-20, Keaten. Wis is a high rate of dmand for EEW.

GPU also made no attempt. to estimate the probability of failure of the EEW systs or the probability of failure of all decay heat rmoval systems at M I-1. Tr.16,629, Keaten. Instead, GPU's strategy was, and continues - to be, to put forward the most generalizcd and unsupported assertions of reliability while sniping at hard data which indicate to the contrary.

GPU claims , for example, that if success criteria other than those employed in the Staff quantitative analysis were used, the results would be better because operator action could be credited. Ibw do we know that accounting. for operator action and misaction would not make the results worse?

Adding another variable to the analysis does not by itself aid GPU, when no analysis of its effect has been done. Sheer speculation cannot substitute for d

facts. ,,

2f . GPU claims that the historic failure rate data are not applicable to MI-l because half of the eight failures were during start-up operations and the WI-l EEW is not "normally" used for startup or shutdown. GPU Brief p. 32 2e implication that the WI-1 failure rate would be much lower for this reason is unwarranted. Use of an EFW system for normal startup can allow discovery of probles that might be undetected at MI-1. Tr. 16,663-6, Keaten. See also Tr. 16,654-5, 16,659-61, regarding the significance of

- failure during -tests.

GPU suggests alm that we should rely on greatly improved operator t' raining. GPU Brief, p. 35 and n. 24. (bnsidering the docunented ineptness

. of GPU's post 'IMI training program as remarked by both the Special Master and

~

the ASIB, reliance on this to significantly improve the quantitative failure rate of IFW would. be absurd. 'Ihe Special Master found, inter alia, "[ f] rom this pattern one must conclude that the training department did not take seriously the Licensee's obligation to teach the subjects required by the Ccmnission's Order and that the operators did not take seriously their obligation to learn it." 15 NRC 918, 1017 (1982). 'Ibe Licensing Board, in its decision after the cheating hearirgs, stated ".. . the Board is forced to conclude that we did not see what we thought we were seeing, and that the Licensee's training and testing progran was best described as the opposite of esse quam videri (to be, rather than to seem) ." 16 NRC 281, 357 (1982) .

In the sane vein, GPU argues that in drawing conclusions about the comparative risks of plants, " consideration needs to be given to the integrated response of all plant systems.. ." GPU Brief p. 37. Ibwever, no integrated reliability study has been done for 'IMI-1, 'by either the Staff or GPU,inor is there any ' reason whatever to believe that such a study would show-

'IMI-l to be , safer than current data indicate either on a ccuparative or absolute basis. It is always possible to argue for more gendral study; that is no reason in itself to disbelieve the results of the EEW failure analysis.

Itreover, ' studies done in other plants. have identified potential ccmnon mode

. failures for B&W plants that' could constrain the ability 'of the plant to deal

^

twith 1oss' of feedwater. See (ES Proposed Findings of Fact, paras. 435-440, j pp. 180-1'11. 'Ihus, -consideration of " integrated response" is not likely to benefit:GPU's case.

4

' One of the lowest points in this proceeding occurred when the Staff tried to discredit its own MI-1 EFW eliability analysis. See LCS Brief at 16-17.

tbw GPU suggests that the hard data on EFW reliability should be disregarded

~

in favor of the Staff witness's "judgnent" that the probability of core danage

at MI-l is less -or the same as for "other plants." GPU Brief at 37-38. 'Ibe

Staff witness was not remotely qualified to offer such a sweeping judgment.

See UCS Reply Findings on Board Questions 2 and 6, July 27, 1981, paras.

I 105-108.

GPU.also claims'that the failure analysis placed "some bias" against the

' B&W design. GPU Brief , p. 36. We problem is not with the analysis, but with

' the B&W design. We fact is that B&W steam generators have a much smaller

. inventory than other IWRs and thus dry out much more quickly, in about five

. minutes, as conpared with twenty. 'Ihis is not a " bias"; it is a recognition

. of a physical dif ference between the plants.

GPU also mounts an attack on the ASIB's use of quantitative analysis, claimirg ' a distinction between this and the St. Lucie case and arguing that some "tnique" circunstance is required to justify quantitative analysis. GPU i .

Brief,;pp. 30-31. . %at is, GPU argues that unless the EI-l EEW can be shown in advance to have a higher rate of challenge or failure than other EFW systems in other' plants, no quantitative assessnent should be done, nor should

- loss of HM.be considered a design-basis event for MI-1.

2is argunent - fatally. misconstrues AIAB-603 and the Ocnsnission's decision in the St. Lucie matter , CLI-81-12,- June 15, 1981. .

GPU is simply wrong in 4 asser' ting 3 that fit 'is' necessary to find, af priori, sone *cnique circunstance"

. . present at 21-1 in order to review the reliability. of the emergency feedwater -

system. It confuses the. evidence on the record which led to a finding in St.

' Lucie .of the - tnreliability of off-site ' and on-site - power with a "sp :cial situation" justifying the review in the. ff rst place.

c

In fact, the Appeal Board in AIAB-603 found that total loss of power at

~4 St. Iucie had a probability of 10 to 10- per year, considering the probability of . loss of offsite power and the probability of failure of both diesel generators. 12 NIC 30 at 45. It is, of course, true that this finding was based ' on the St. Incie factual situation, as any record finding is, but this is no different than the 'IMI-l case, where the reliability assessment prodtced by the Staff was based on the particular design of the EFW system.

'Ibe Appeal Board fully recognized that its St. Lucie finding might have implications beyond that plant. It stated:

Our finding that station blackout should be considered as a design basis event for St. Incie Unit 2 manifestly could be applied equally to thit 1, already in operation at that site.

By a parity of reasoning, this result may well also obtain at other nuclear plants on applicant's system it not at most power reactors. Our- ?lrisdiction, however, is limited to the matter before us .- licensing construction of St. Incie 2. Beyond that, we can only alert the Commission to our concerns.

AIAB-603,12 tac 30, 32, enphasis added.

'Ibus, the ' finding that compelled consideration of station blackout as a

' design basis event was simply the finding based upon the record of a relatively high probability of occurrence of the event, irregardless of whether such a high probability was " unique" to St. Incie. As the Cmmission

-noted in reviewing AIAB-603 and ' leaving it undisturbed:

'Ibe Appeal Board finding relevant to this review was that the probability of total loss of on-site and off-site AC power

- station blackout - was sufficiently high that protecting the plant against such an occurrence was warranted.

y CLI 12, June 15,1981, Sl. op. - at 2.

By contrast, GPU would have the Cbmmission rule that loss of EEW need not be considereed at SI-1, even though it has a higher probability than station blackout at St. Incie, because no showing has been made that 'IMI-l is " unique" Such a standard would supplant. the imC's duty to ensure the

~

in this respect.

safetyL of the plants it licenses and permits to operate with a general maxim i

that a plant shall be licensed unless it can be shown that it suffers fram

" unique" ' safety. problems. 'Ihis proposition is empletely at odds with NIC's

- responsibility and finds no support in either AIAB-603 or CLI-81-12.

GPU claims likewise that sme special circmstance need be shown in order to justify reliance on quantitative estimates of the reliability of the 'IMI-l EEW system. 'Ihe argment apparently is that, while the reassuring qualitative judgments of IIW reliability contained in C E and Staff testimony can be relied upon, the quantitative reliability assesment cannot be used unless B&W plants are worse than other.1MRs.

We see no logical support whatever for such b curious positon. Both the qualitative judgments arxl the quantitative assesment are directed toward the same question: Is the ']M I-1 . EEW system sufficiently reliable or should failure of EEW be considered in deciding whether to permit ']MI-l to operate?

'Ihere is no inherent reason that we can perceive why the former should be considered but not the latter. tbr is there any precedent or logical reason why an additonal threshold of " uniqueness" should be established in order for the Commission to consider and rely upon the quantitative reliability assessment.

Finally, GPU states that the Board cited "with approval" Staff testimony

'that the 'lMI-li EEW - system at restart will be cmpa~able with sme other operating plants and about equal to industry average. GPU Brief, p. 27. 'Ihe Board indicated no such "apprwal," ~ it simply described the witness's opinion.

14 NRC at- 1372. Moreover, at the time of the hearing, the Staff witnesses believed 'that the EEW system was essentially seimically qualified, based on GPU's representations. See App. Tr . 325, Lic. Ex. 15. It was learned only later . 'that the system is not seimically qualified. App. Tr. 345. . In addition, in the years since the ' testimony was given, all other operating 9

5

- ~. .

-reactors; have had. at least one refueling outage and thus, most EFW systems

- should ! be fully safety-grade by now. 'Ihese developnents adversely af fect

'IMI-l in cmparison with other plants.

'Ibe Staff's position on the issue of W reliability can be briefly stenarized: 1) at restart the EEW system will be safety-grade for small break IDCA and loss of ' main feedwater accidents . (but not other design basis accidents) and thus is suf ficiently reliable, and 2) feed and bleed provides "further, thotgh not required assurance of the protection of public health and safety until restoration of feedwater" Staff Brief, p. 17.

'Ihe. former is nothing more than a syllogism. In the face of data showing B&W .EEW systens are challerged at a relatively high rate, failure data for safety-grade systems "not.. . indicative of high reliability (14 NRC at 1356-7),

and a plant-specific ' failure mode analysis showing that after it is fully safety-grade, the 'lMI-l EFW is expected to ' have a ' failure probability of 1.5X10

-4 . per year '(Id. 'at 1370; , it is not ' sufficient to fall back on the bald

~

assertion that the' system is adequately reliable.

~

As to i the Staf f's attunpt to resuscitate feed and bleed, albeit coyly

~

couched as "tnnecessary," we are frankly astonished. First of all, when the Staff argues that it "does not. believe that it is necessary.. .to rely on feed and bleed as a backup" to EFW (Staff Brief, p.17, emphasis in original) , the Staff deliberately ignores 'its own testimony and the Appeal Ibard's findings-based m that testimony.

'Ihe{ Staff-- position during- this proceeding was that feed and bleed p neces'sary ' to meet NRC rules pending full upgrade of the EEW system to safety

- grade'. >

Until- EFW system-upgrading at 'IMI is completed, the Staff is relying on the feed and bleed mode of core cooling to protect against events tor which the erw system is not tully satety

. grade.

NRC Staff Proposed Findings, para. 435, emstasis added.

Based on our consideration of the evidence on the record of this proceeding,-we find that althotxjh the EEW system at 'IMI-l

- will~ not be safety-grade .at the planned - time of restart, it will have been upgraded to significantly improve its reliability, that operator action within about 20 minutes to actuate the safety-grade HPI pumps and initiate feed and bleed cooling can protect against failures of both the main and mergency feedwater systems, that feed and bleed cooling can be

~

. continued until feedwater is restored and thus there is

-reasonable assurance that the public health and safety will be adequately protected against feedwater transients if TMI-l is allowed to restart prior to full upgrading of the EFW system to

-safety-grade.

NRC Statt Proposed Findings, para. 441, emphasis added.

Se ' Appeal Board, relying on the Staf f's testimony and ICS Proposed Findings on % opened Hearing (April 10, 1983), found as follows:

We consider the EEW systs sufficiently reliable for events within the limited scope of this proceeding. However, the staff has indicated that feed and bleed is relied upon for those events for which the EFW system is not fully safety-grade, such as main steam line break. Furthermore, the staff testified that the EFW system function following a safe shutdown earthquake has not been demonstrated since portions of the -system piping and controls are not-Seismic Category 1.

While these events (such as a main steam line break and a severe earthquake)_ are outside our purview, it is necessary to

note our concerns over the possible reliance upon feed and bleed.. If the staff wishes to rely on feed and bleed, regardless of whether the event postulated is within the scope of the restart proceeding, then it should prortptly complete its analysis of the feed and bleed process to assure its viability.

17 NBC 855, emEilasis added, footnotes onitted.

~

Se Staff, in its zeal to convince a majority of the Omnission to vote m

for ' restart without :considering the risk : to . public health and safety, apparently hopes the Carmission will. overlook the facts that:

' l. y te 'IMI-1 ; HW ' system is not safety-grade for all design basis accidents, such as' main . steam line break and a severe earthquake;

' - 2. - Eleed and bleed -is' relied tpon by the Staff to protect public health and safety against those design basis accidents- for which EEW is not 1 s safety-grade; and

, ~3. Feed and bleed :has not been shown 'to be a viable. method of pro'fiding E

adequate core cooling for 'IMI-1. ,

i

. . . . - - - - ~ - - - - - - - .

+ 4

-l8-

'Ihese ' facts contributed . to the Appeal Board's finding that the eviden-

~

~

tiary - record 1 in , this ; prceeeding is inadequate to determine "whether there is

. reasonable assurance that 'Ihree' Mile Island thit No.1 can be operated without endangering the health and safety of'.the~ public." 17 NRC at 823, 895. .It was

' left to the Commission to make that determination "after examining all systems

~

.and considering information within and outside this record... ." Id., mphasis.

added.

'In contrast, the Omnmission is improperly focussing primarily on the F . question of whether information and issues were properly excluded fron. this L

record while. apparently ignoring the excitxled information which shows that Tt4I-1, in its present condition, cannot be operated without undue risk to public ^ health and safety.

We J return now to the Staff's positions that, for -small break loss of

. coolant rccidents . and main feedwater transients, " feed and bleed should be

' considered to provide further,

though not required, assurance of the

protection 4of the public health and safety until restoration of feedwater" and that "the feed and bleed -core cooling system has been .shown analytically to' be' .

. effective l for a. period of 'several hours following a postulated loss of all .

.feedwater." L Staff Brief, p.17.

Here Lagain, the Staff ' apparently; hopes the (bmmission' will werlook the fact that the ' Appeal . Board ~ reviewed the Staff's assertion that feed and bleed -

. has beg -shown analytically to be effective" and found that assertion . to be without merit.

, 'Ihe: conclusions of these analyses lend some support for -

the position that feed and bleed can prwide adequate core

. cooling 'at 'JMI-1.~ However, because of the uncertainties l involved in the ' analyses and the ' failure of the. staff witnesses

~

-to' adequately address those~ uncertainties in their testimony, we are unprepared to state conclusively that feed and bleed .

will successfully provide core cooling at 'ITII-1. As noted by

- LES, .staf t. witness Sheron testified at the reopened hearing h.'? .

n

'that the adequacy of feed and bleed is within the range of experimental uncertainty. Additional investigation at the uncertainties inherent in the analyses would be needed before a definitive statement on the viability of feed and bleed cooling could be made.

17 NIC 852, enphasis added.

Furthermore, less than two weeks before filing its brief, the Staff sunnarized , for GPU, the reasons why the Staff presently can give no credit for feed and bleed. James van Vliet, NRC's project manager for ']MI-1, and Walton Jensen, one of the principal Staff witnesses in the reopened hearings

-who tried (and failed) to convince the - Appeal Board .that feed and bleed coolirg has been analytically demonstrated for 'IMI-1, told GPU the following:

MR %N VLIET: As far as feed and bleed goes, I think that our position has been, and of course you can check the record on this, it has been that we think we have analytically demon-strated that there is a high probability that feed and bleed would work but as I understand it we have never looked at feed and bleed as it relates to a soecific plant, to the kind of depth that we would have to to give it credit for design basis accidents.

Now as far as what might be required for us to consider design basis _ credit for feed and bleed, I will ask Walt Jensen to expaund on it a little bit and Walt can explain some of the kirds of thirns we need to look at.

MR. ' JENSEN: As Jim said, we have seen analysis for our feed and bleed for several hours after a loss of feedwater event and it seems as though it will work for the period of time that we have looked at it. But before we can give credit for it as _ a system that could mitigate design basis events, we would have to have a good bit of additional informaticn such7s

.we would have for any such system, ICCS system or any kind of ergineered safety feature.

4 I have got a list of some things that we have thought about if I can read it.

. Che thing that needs to be defined is the time period we

, . are talking about. How long would you need to be in feed and bleed? We need to develop a design basis, in other words, for 0 -the system. What is the design basis for it in terms of time?

You would go on loss of feedwater, I guess, and how long would-pu have to stay on feed and bleed before you could switch to.another system such as DIR [ Decay lieat Iemoval] ? Or ,

w)uld you assune that other equipnent would be available after

some period of' time, so you would have to be -- so you would not be'on bleed and feed then.

E

And we would like to see some couplete analysis over this period of time for the reactor system and the containment ,

system such as the system pressure and the tenperatures throughout the primary loop and the void fractions that might occur in the head and other places, pressurizer, hot leg.

b _

MR. MA[U]S: . Are We going to get a transcript of this so j we don't have to take detailed ' notes?

' MR. VAN VLIET: It'is up to you.

MR. MAIE: All right. .

MR. JENSEN

'1 hat would probably be a help.

_ 'Ihe flow is going into the system fran the safety injection pumps and the flow is going - through the IORV and safety valves.-

4 We would also like to see analysis of the containment temperature and pressure and I guess the one- problem that comes to mind is the brittle fracture limits fran Appendix G on the supporting thermal stress evaluation for-the reactor vessel, the types of operator action that might be required to keep the limits from being exceeded and does he have to control the PCRV safety injection plunps .or what kind of instrumentation would he rely on to do_this?

Can this instrument'ation really tell what he really needs to know about temperatures within the reactor vessel?

-And then we need to look at other equipment that would be

- needed to tunction during the feed and bleed process and we need to have justification that this equipment can operate for whatever period of time it is supposed to operate.

l. _

And the hich pressure injection punps, can'they operate p - for- the requirec number of hours? The safety valves,- the L PORVs, would they be N W in feed and bleed in this time?

And then we need to-be sure that this equipment is L

designed to appropriate standards for safety-related equipment

, including redundancies, seismic qualification,-environmental

'~

yqualification and emergency power requirements?

.Next after the period of feed and bleed, the reactor core

still'needs to be cooled and I do not know what you would do, whether you would go on -RHR or DHR.

In any case, you would t need uto look at that equipnent and be sure that equipnent would still be available and we would look on that as part of this feed ^and bleed system which would also meet safety related standards.

~

l'Ihat ;would incitzle redundancy, seismic, emergency power.

'One thought we had 'about the DHR system is the' fact that it has

a single drop line with a valve set which might be proce to

. single ta11ure.- .'

"~

And then we need to look at the operator, to list all of the operator actions that are necessary during the teed and bleed process and post-feed and bleed recovery, justify that ,

the plant procedures exist for feed and bleed operation and post-teed and bleed operation, Justify that operators have been adequately trained in these procedures.

Next we need to look at experimental verification that -

for the feed and bleed cooling process, provide experimental data justifying the discharge model for the safety valves and/or the PORV and then verify to do a cmputer code adecuately for the experimental data.

• Ihat is the list that we came up with.

MR. VAN VLIET: I might add that the list was really put together in about a 24-tour period and the final review may or may not have more things, listings, we just don' t know but certainly I hope that gives pu a feeling for the things we would need to look at.

MR. VAN VLIET: I believe if we had to write an evaluation today, and of course one of the aspects of that is the [LCS Show Cause] petition [Concerning EIM], that we would have to only respond to what you have on paper here and that of course is your reliance on feed and bleed.

And I think what we would have to say is simply what I said and Wiat Walt said, we can't give you credit, design basis credit for feed and bleed at this point and we have to re]ect it.

Transcript of Staff Meeting with GPU on 'IMI-l Enviranental Qualification, March 8,1984, pp. 6-10,12, emphasis added.3/

-3/ UCS wishes to bring ' to .the Cbmmission's attention the fact that Walton Jensen was one of the Staff's principal witnesses before the Appeal Ibard in the reopened hearings and testified, under oath, to the effect that much of the information he sotght from GPU in the March 8,1984, meeting was already known or was mimprtant in assessing the viability of feed and bleed. See, for exanple, .Sheron and Jensen, ff. App. Tr. 83, at 40, regarding the alleged mimportance of exprimental discharge data for the safety. valves, and App. Tr.:194-199, 210-211 (Jensen) regarding whether the HPI pumps would be capable of performing as required during feed and

' bleed -- Question by counsel for GPU: "Mr. Jensen, just to get to the bottom line of Mr. Pollard's questioning about the HPI pump capability, does the use of the HPI ptnps for cystem makeup for extended periods of time during normal operation of the plant give pu confidence in the'HPI (Ebotnote Cbntinued) w e- >4 n- - ~-- - e - --

l Ihus, contempraneously with its assurances to the (bmmission that feed l and bleed can be relied upon as a backup to EEW, the Staff is telling GPU that feed and bleed cannot be relied upon because the Staff:

1. Has "never looked at feed and bleed" for a specific plant,

2. Does not know "how long you need to be in feed and bleed,"

3. Needs an analysis specific ta T4I-l of containment and primary system temperature and pressure, void fractions in the reactor vessel head, pressurizer, hot leg and "other places,"

4. Is concerned about the ptential for exceeding the brittle fracture limits of the reactor pressure vessel during feed and bleed and whether the existing instrmentation is adequate to tell the operator "what he really needs to know,"

5. Ibes not know whether the high pressure injection pumps can operate for the required number of hours and. whether the PCRV and pressurizer safety valves will be damaged by feed and bleed operations,

6. Does not know whether the equipnent needed for feed and bleed meets NRC regulations concerning redundancy, seismic and enviromental qualification, and mergency power supplies, (Ebotnote (bntinued) pump l capability for feed and bleed conditions?" Answer. (WITNESS JENSEN):

~"Yes, it would. - 'It would show the pmp oculd operate for' a long period of time with ' flows not much different from what it would .have in feed and s bleed operations." (bunsel for GPU" " Fine, that is all I have."' App. Tr.

210-211. Other' exmples of Mr. Jensen's propensity to say whatever he

- t! bought was necessary to extract a decision favorable to restart fran the Licensing and Appeal Board abound. Ebrtunately, LCS cross-exmination was often able - to establish that Mr. Jensen had no factual basis for his testimony. In the case of the above aspects of his testimony on the viability of feed and bleed, we had ~ to await the transcipt of the March 8, 1984, meeting for him to acknowledge the lack of a factual basis for his testimony. _ See also, Staff Cbunsel Cutchin's successful objection to ICS's 'attemp to establish that feed and bleed could threaten the reactor vessel brittle fracture limits. He claimed the issue was not relevant, in contrast to the Staff's current position. App. Tr.~300. '

k c

7. Is concerned that the Decay Heat Ibmwal System is prone to a single failure,

.8. Ibes not know whether plant procedures and operator training are adequate for. feed and bleed operations and post-feed and bleed recovery, and

9. Has neither experimental data for the discharge characteristics of the MI-l PORV and safety valves nor, of course, a canputer analysis of feed and bleed utilizing experimental data for the 'IMI valves.

In sum, while the Staff Brief appears designed to lead the Cmmission to overlook the fact that although feed and bleed is necessary for design basis accidents, the Staff is in no position to vouch for its viability as a backup to IFW, not even as an allegedly " unnecessary" backup to EEW for small-break IICAs and main feedwater transients.

3. IORV With regard to the use - of the IORV for depressurization during low temperature operations, the argunent is reasserted that' the PORV is not needed because the operator would have more than ten (10) minutes to terminate the cause of the overpressurization. 'Ihe Staff asserts that this 'is true whether or lnot a steam bubble -is maintained in the pressurizer. Staff Brief, p. 23.

'Ihis assertion is false and not in accord with the record. It appears to be ,

based solely on the Appeal Board's sanguine, but utterly unsupported observa-tion that "[il fI an ' overpressurization event were to occur during a cold shutd'own condition with no bubble in the pressurizer, the IORV also should only serve 'a secorrlary safety function as a backup to operator actions (e.g.,

shutoff HPI > and increase letdown flow) to terminate the event." 17 NRC at 864, enphasis added, footnote anitted.

l l

'Ihe problem is that the evidence in the record is to the contrary and the  ;

l Appeal Ibard made no attempt whatever to reconcile its cbservations with the record.

Ebr one thing,- even GPU's witness conceded that it is probable that when there is not a bubble in the pressurizer, the operator does not have ten (10) minutes to act to prevent overpressurizing the reactor vessel. Tr. 8976, Jones. 'Ibere was no evidence contradicting this. As Mr. Pollard testified:

Another good illustration was the technical specification which Mr. Jones {GE's witness] just finished reading, which says, in effect, that the NRV may not be taken out of service unless specified corditions are met. 'Ihat is that pu either had a bubble in the pressurizer or the high pressure injection pump breakers were locked out or the injection valves were locked closed.

I think this quite well illustrates that the protection ajinst overpressure when you are cold is the PORV. It is true the operator should try to make sure that nothing happens in the plant that would cause 'an overpressure condition, but if-something did happen, it is the ERV that is going to limit that pressure rise.

Tr. 9032-9033, Ibllard.

'Ihe operator is instructed in the 'IMI-l procedures to use the ERV for

~ depre'ssurization under normal and transient conditions. Tr. 9033 ff, Pollard.

'Ib refer- to - the ERV as a '" backup" is incorrect. It is the sole protection

[ against overpressurization. It is only a backup in the sane sense that the emergency core cooling system (ECCS) is a backup to a loss ~ of coolant

. accident. 'Ihat is, if the primary system, despite its high quality, has a rupture, the ECCS must deal with it. ' So too, the ERV "is only a backup in

?

the sense that if the operator makes a mistake ' and causes an' event which resul,ts 'in increasing. pressure, the ERV is there to terminate it." Tr.

r .9031-9032,.- Pollard.

Both GPU and the Staff also ~ assert that during inadequate ' core cooling conditions, the depressurization functions of the PORV are less significant

- than depressurizing with' .the operative steam generator. E.g. GPU Brief, p.

.' 4 5. ' As we have pointed out repeatedly, the 'IMI-l procedures require use of

-2 5-

~ooth the steam generator and ~ the IORV; the operator is directed to use the PORV to maintain ICS pressure within 50 psi of - steam generator pressure. Lic.

Ex. 48 . at 26.0-27.0. Moreover, after depressurization, the operator is directed to use the PCRV to control ICS pressure below 150 psig. Lic. Ex. 48 at 28.0.

In addition, the Staff and GPU totally overlook the fact that there is no safety-grade means - of depressurization -even using the operative steam generator. Neither the atmospheric dunp valves nor the turbine by-pass valves

-are safety-grade as this record clearly establishes, e .g . Tr. 16,557-59, Keaten. 'Ihus, there is no way to depressurize through the steam generators without the use of non-safety grade equipnent. Finally, as even the Staf f testified, one function of the IORV is to give the operator a means of

~ depressurizing the y aary systs that is independent of the steam generators.

Jensen , f f . Tr . 8821, at 3.

In belatedly claiming that the use of the IORV during low temperature operations is not within the scope of this proceediri g, the Staff raises an objection which it never . raised during the hearing. Sta f f Br ie f , pp. 22-2 3.

~

'Ihe Staff did not object to the UCS testimony nor attempt to have the

pertinent sections stricken. It cannot be heard ' now to raise a last minute objection after four years of litigation, particularly when the result would be .to prevent consideration. of evidence clearly establishing a vital safety function of the PCRV. 'Ib ignore this issue after the evidence is in would confirm ' .that the scope of this proceeding was ' manipulated to deliberately

. exclude 'the " bad news" while including the good.

~Moreover, the Staff's " nexus" argunent distorts the thrust of the tCS contention and the basis upon which it was argued to and accepted by the ASIB.

As ;tCS has em@asized from the outset of this proceeding, one of the primary

~

. lessons -learned .ftcm the 'IMI-2 accident is that previous safety analyses did .

not properly recognize - the manner in which systems then considered to be unrelated to safety may in fact be called upon to mitigate accidents. TS d

. believes that systems used in accident prevention and mitigation are important to sa fety, within the meaning of tGC's r ules , should be classified as safety-grade, and upg raded to meet the pertinent Commission ' rules for safety-grade systens (redundancy, diversity, quality assurance, testability ,

etc.) so that they_ are sufficiently reliable to perform their accident-mitigation ancVor prevention functions. 'Ihe PCRV is one of these.

While there was vigorous dispute on the merits among the parties as to whether! the PCRV does per form functions impor tant to safety, there was no dispute that ' there is a clear connection between this issue and the E I-2 accident. Ib party objected to the adnission of TS (bntention 5. IBP-79-34, 10 tGC - 828, 83 6. It was not claimed by any party that the lessons learned fran MI-2 could be limited mechanistically to snall-break IOCAs or loss of feedwater events.

In support of the view that the lessons learned from the EI-2 accident

- enconpass this' issue, TS 's testimony . cited the IGC's original " lessons learned" evaluation, NUREG-0578:

Several nonsafety systems were used at'various times in the mitigation of the accident in ways not considered in the safety analysis; for example, long-term maintenance of _ core flow and cooling with the-steam generators and the reactor coolant pumps. 'Ihe present classification system does not adequately recognize either of these kinds of effects that ynon-safety systems can have on the safety of the plant. Thus requirements for nonsafety systems may be needed to reduce the frequency of occurrence of events that initiate or adversely affect transients and accidents, and other requriements may be-

'needed to improve the current capability for use of nonsafety

- systems during transient or accident situatiom.

Id., p. 18, em@ asis added.-

The relatively high frequency of A00s (Abnormal Operational (bcurrences] places a reliability demand on the

' o;eration of the PORVs and associated equipment that is higher taan originally envisioned. Also, the operation of sone _

- canponents and. systems provided for emergency core cooling have been'. challenged more times than was previously expected as a -

I result of A00s. Therefore, there is a need to consider the upgrading of the PORVs, block valves, and the associated control and power equipment to a safety-grade classification to achieve greater valve reliability and to minimize the number of challenges to the operation of the emergency core cooling conponents and systens. Ibwever, the merits and degree of upgrading of all pressure-relief equipnent associated with the pressurizer requires further evaluation, which should be accomplished on a' longer term basis.

Id., pp. A-3 and A-4, emphasis added. Pollard, ff. Tr. 9027 at 5-15, 5-16.

The 'IMI-l accident, thus, directly raised the question of whether systems previously condidered unrelated to safety do, in fact, perform safety functions. That is the " nexus" for this contention. IES argued and presented evidence to demonstrate that the PCRV is a conponent which does perform vital safety functions. No party objected to any of this evidence. Belated efforts after the record has been made, to look at only sane of these safety functions while ignoring others, are inexcusable. If the next accident involves an overpressurization event, one can only imagine the recriminations that would follow.

There is a good reason why the Staff urges the Cbmmission to . rule this issue to be beyond the scope; the Staff's substantive ' position on the use of

, the IORV-for depressurization during low pressure operation is indecipherable.-

y.

Pages 23 and 24 of the Staff's brief can only be read as aduitting that the Appeal Board was wrong in finding that the IORV is only a backup to operator action:

-7 Tneir conclusion may' be' correct depending on the particular circumstances involved (e.g., depending on the.

reliability of the alarms) . The record, however, contains no "

such details. .

Having admitted that the record does not . support a finding that the IORV is only a " backup", . the Staff then seeks . to ' have . the . Ccanission accept other

" reasons" W1y the -IORV need not be safety grade, while 'in the same breath ,

conceding that' these new " reasons" are also not part of the record. Staff Brlef, p. 24 and: n. 17.- An. astonishing footnote - follows, containing. a

soliloquy supported neither- by record citation nor citation of any kind whatever, nor even any attribution. We claim is made that "the Staff believes that vessel - failure will not result" from unmitig ated vessel

. overpressurization. Wis ~ unsupported " belie f" on the part of the Staf f's lawyers (no technical Staff docunent nor published technical opinion of any kind is cited) has no status in this proceeding and is entitled to no-

. consideration. %e Staff is well aware of this. Rus , its new " reasons" are apparently intended as a signal to the Ccmnissioners that they need not really worry about this- issue and can seek some way to avoid the merits.

LES is aware of no docunentation supporting the Staf f's new belief and we consider this transparent ploy to be highly unethical. Indeed, we believe that - the - Staff. has ' prwided no citations to support its belief precisely to

. prevent ICS frcun. mounting a rebuttal to anything cited. Bis is particularly egregious considering that LES attenpted during the reopened proceeding to explore the question of potential vessel overpressurization and were prevented

. fran doing so because the Appeal Board sustained Staff attorney Cutchin's objection that the issue "has nothing whatsoever to do'with the issues within this procceding."

App. Tr. 300. - Se Ccmnission may not lawfully rely on the Staff's " belief" either- implicitly or explicitly. See SAPL v. Costle, 572' F.2d 872 ~ (1st 'Cir.1978) . -

J In sunenary, the evidence of record conpels a finding that the IORV should besafeyy-grade. No anount of legalistic pettifoggery can obscure this fact.

m

~4. J Systems Interaction -

In -essence, both GEU and the Staff argue that, on this issue, 'no progress is reasonable progress. .'tCS sees'no need to reply.

~

wmgf s m y y: g y ,a w.

c ~

=

,) l$ f- h *'$, #&

s

5. : ~ Main Steam Line Rupture Detection System RSLRDS)

. Wis [ issue first arose. because ' it becane apparent that failure of the

- non-safety grade MSIRIS could shut off all feedwater (both main and emergency) tio both steam generators. De ASIB " resolved" the issue by directing GPU to propose' a solution. Rather than upgrzde the system to safety-grade prior to restart, GPU proposed the short-term expediency of- disconnecting the signal between the EW system and the non-safety grade MSIRIS. GPU apparently - hops -

thereby to consign the remaining safety hazards associated with the non-safety grade MSIRIE to that vast' universe "beyond the scop of the proceeding."

Pr ime : among = those hazards is the potential for overpressurizing the.

containment 1 following .a main steam line break inside containment if the non-safety grade MSIRDS fails to shut off main feedwater to the affected steam generator.- Neither GPU nor the Staff denies the existence or the seriousness of -. this problem. . Mr do they deny that making the system safety-grade would prevent it. ney simply wish the Commission would disregard the issue.

Frankly,- it never occurred to LCS that GPU would hit upon a short term

" solution" so manifestly inadequate. 'Ihe Ippeal Board recognized the~ " concern for.~ overpress' u rization. of the contairinent"' and stated: " Prior to acceptance of

[the; GPU] proposal, we recommend that ' the potential ~ for containment L overpressurization' as aT result of MSIRDS failure be evaluated." 17 NRC at

[834. -.It has'not(been.

Itc should , be: enshasized ' that the Oamm'ission did not take review of

. whether this tissue is within the scope of the proceeding. - We therefore assune t

tihaE itr ' wishes' to - address the merits. On the merits it is ' not subject to w l serious: dispute that the 'R4I-l MSIRDS poses a serious safety hazard.

4 4

1. ~

A l'. .. .,

3. J.

s-

+

- CONCIUSION-

' For the ~ above-stated ' reasons as ' well as those treated in LCS's main

brief, the Catmission stould . find that the operation of 'IMI-l would pose undue

. risk to the public ' health and safety.'

Ibspectfully Submitted, u

Ellyn R. Weiss General Counsel Union of (bncerned Scientists Harmon, Weiss & Jordan 2001 S Street, N.W., Suite 430 Washing ton, - D.C. 20009 (202) 328-3500

- Dated: April 3, 1984

? .

• s 4-Y t

-}~

i t ,+.4' d

f ,

^ N u.

(::

00CMETED 35NHC UNITED STATES OF AMERICA NUCLEAR REGULATORY COM4ISSION '84 APR -5 A10 :18 0FFILE-OF SLCat.1Ah 00CnETING & SERVir.l.

BRANCH In the Matter of- )

)

METROPOLITAN EDISON COMPANY ) Docket No. 50-289

) (Restart)

(Three Mile Island-Nuclear )

Station, Unit No. 1) )

. CERTIFICATE OF SERVICE y .'I hereby. certify that copies of " UNION OF CONCERNED SCIENTISTS'-REPLY BRIEF ON REVIEW OF ALAB-729 and ALAB-744" have beeniserved on the following-Epersons by deposit-in the United States mail, first. class' postage prepaid, '

this-3rd day of. April'1984, except as indicated by an asterisk.

Nunzio Palladino, Chairman- _ Gary J. Edles, Chairman U.S.~ ~ Nuclear . Regulatory Comission Atomic. Safety and Licensing Appeal Board Washington,-'D.C.=20555 U.S. Nuclear Regulatory Comission Washington, D.C. 20555 Victor Gilinsky , - Commissioner -

.U.S. ' Nuclear Regulatory Comission~ Dr. John H. Buck Washington,'D.C. 20555- Atomic Safety and Licensing Appeal Board U.S.~ Nuclear Regulatory Comission

, James Asselstine,~ Comissioner' . Washington'D.C. 20555 U.S. Nuclear Regulatory Comission ' . . _ _

-. Washington,:D.C.;20555 Dr. Reginald.L. Gotchy Atomic-Safety and Licensing Appeal Board

, > Frederick Bernthall,Comissioner _U.S. Nuclear -Regulatory Comission

?U.S. Nuclear Regulatory Comission Washington D.C. 20555..

Washington , - D.C. '20555 Judge Christine N. Kohl

Thomas ~ Roberts,! Comissioner. -Atomic Safety and Licensing Ap' peal Board-U;S. NuclearnRegulatory Comission' U.S. Nuclear Regulatory Comission Washingtoni D.C.l20555

Washington, D.C. 20555

. Docketing and Service Section _Ivan W.~ Smith, Chairman

~ Office'ofthe-Secretary ( .

Atomic Safety and'LicensingLBoard l~. .. U.S. Nuclear Regulatory _ Comission - U.S.SNuclear Regulatory Comission JWashington,.D.C. 20555 Washington,-'D.C. 20555 1 <

.~E s x ,

4 s -r 4 l

c -

m.

-2

. Sheldon J. Wolfe, Alternate Chairman.

• William S. Jordan, III Atomic Safety and Licensing Board Ha. mon, Weiss &' Jordan

- U.S.JNuclear Regulatory Commission 2001 S Street, N.W.

~ Washington,'D.C. 20555 . Suite 430 Washington, D.C. 20009 Mrs. Marjorle'Aamodt R.D. #5- .

John A. Levin, Assistant Counsel

'Coatsville,-PA 119320 Pennsylvania Public Utility Corrinission P.O. Box 3265 Maxine Woelfling, Esquire Harrisburg, Pennsylvania 17120

Office of Chief Counsell Department'of Environmental Resources ANGRY /iMI.PIRC

505. Executive House 1037 Maclay Street

. P.O. Box 2357. Harrisburg, PA -17103

- . Harrisburg, PA~ 17.120

• Steven C. Sholly.

Ms. Louise Bradford Union of Concerned Scientists

' Three Mile Island ~ Alert 1346 Connecticut Ave. , N.W. , Suite 1101 1011 Green: Street- Washington,'D.C. 20036

" Harrisburg,'PA 17102 Richard J. . Rawson Dr . - Judith H. - Johnsrud Office of. Executive Legal Director Dr. Chauncey Kepford _ .U.S. Nuclear Regulatory Corrrnission-Environmental' Coalition on Washington, D.C. 20555 Nuclear Power 1433 orlando: Avenue. Thomas A. Baxter,-Esq.

State College, PA;16801 Shaw, Pittman, Potts & Trowbridge 1800 M. Street, N.W.

- Washington, D.C. 20036

/

1. Hand delivered.

P

)

i rf l:

4

,6 .' .- .