ML20126H417

Proposed Findings of Fact & Conclusions of Law in Form of Partial Initial Decision for Issue 8 Re Hydrogen Control.Ol Applications Should Be Denied.Certificate of Svc Encl

ML20126H417
Person / Time
Site:	Perry
Issue date:	06/13/1985
From:	Hiatt S OHIO CITIZENS FOR RESPONSIBLE ENERGY
To:
References
CON-#285-464 OL, NUDOCS 8506180336
Download: ML20126H417 (89)

Text

_ _ _ _

r u.

k

't O

June 13, 1985 i

UNITED STATES OF AMERICA NUCLEAR REGULATORY CONHI55I0t1 Before the Atomic Sofety and Licensing Board DONRC U

In the Matter or-

)

)

I THE CLEVELAND ELECTRIC

)

Docket Hos. 50-440 OL ILLUHINATING CO. ET AL.

)

50-44YALJJN 17 A10:15

)

(Perry Nuclear Power Pione,

)

0FFICE OF SECRETAb )

Unies 1 ond 2) 00CKETING & SERVICf BRANCH OCRE'S PROPOSED FINDIt4G5 0F FACT Af40 'CONCLUSI0f45 c

0F LAW IN THE FORM OF A PARTIAL It4ITIAL DECISION (ISSUE M8, HYDROGEf4 CONTROL)

Susan L.

Hiott CCRE Representative 8506180336 0506 ga m ow0,,40 3@9

4 O

TABLE OF C0tlTErlTS Page 1

I.

OPINION............................................

1 A.

HISTORY OF THE CASE.............................

3 8.

BACKGROUND OF ISSUE #8..........................

4 9

C.

APPLICABLE STANDARDS.............................

D.

COMPLIANCE WITH THE RULE.......................

14 14 1.

Operation or the System......................,

16 2.

Scenarios.....................................

19 3.

Containment Inte9rity.........................

c (a) Containment Vessel Capacity 19 (i) Positive Interno1 Pressure 19 (ii) tJegotive Internal Pressure 29 (c) DryWell Copacity 29 4

Contosnment Reiponse Analysif 30 5.

Equipment Survssotility 45 6.

Other Effects of 5) stem Operation 47 (a) DryWell Pool Leods 48 (b) Decoy Heat Removal 40 50 (c) Secondary Fires............................

7.

Scope of the Prelimsnory Analysts 50 II. F I rl D I rl G 5 0 F FACT 53 III. CONCLUSIONS OF LAU 78 77 IV. ORDER............................................

APPEllDIX A, URITTEN TESTItt0NY RECEIVED INTO EVIDEllCE APPErlDIX B.

EXHIBITS A PP E rlDI X C, COMPARIS0tl 0F C0 tit A ItittEt4T RESP 0 rise At1ALYSES l

L m

4 e

June 13, 1985 UNITED STATES OF AMERICA NUCLEAR REGULATORY CONNISSION Before the Atomic Sorety and Licensing Board 00LKETED In the Matter or

)

UShRC

)

THE CLEVELAND' ELECTRIC

)

Docket Nos. 3-440 OL ILLUMINATING CO. ET AL..

.m-Jal k2. 40 :15

)

(Perry Nuclear Power Pione,

)

0FFICE OF SECRETA 6 Y Units 1 and 2) 00CKETING & SERvlu.

BRANCH OCRE'S PROPO5ED FINDINGS OF FACT AND CONCLU5 IONS OF LAU IN THE FORM OF A PARTIAL INITIAL DECISION (ISSUE M8, HYDROGEN CONTROL)

Pursuant to 10 CFR 2. 754 (o) (2). Intervenor Ohio Citizens'for Responsible Energy ('OCRE*) hereby submits in the form or o portzol initial decision its Proposed findings or rock and conclusions of low relating to Issue NS, on hydrogen control.

The proposed findings or roct and conclustons or low rollow the form prescribed by-the Atomic Sorety and Licensing Board

('Ecord') in its April 18, 1985 Memorondum and. Order (Preposed Findtngs and conclusions).

I. OPINION A.

HISTORY OF THE CASE i

This is the thard portial initial decasion in this contested proceedsng on the, application for operating licenses for the Perry Nuclear Power Plant (PNPP").

The first portial initial s

decssion. LBP-93-77. 18 flRC 1365 (1983), deoit with quality l

assurances.ene second. LBP-85...

__NRC__

(1985), concerning emergency pionning and aiesel generator reliability, included o i

[g

r o o thorough ciscussion of the history of the cose which Will not be rencatee nere.

Briefly, the Applicones, The Cleveland Electric Illuminating Co.,

Duquesne Li9ht Co.,

Ohio Edison Co..

Pennsylvanio Power Co.

And The Toledo Edison Co.,

are seeking licenses to operate two coiling water reactors.('BWR5'), Units 1 and 2, at the Perry Eite an Loke Ceunty, Ohio, located opproximately 35 miles northeost of Cleveland on Lake Erie.

Their oppizcotton is opposed by the two intervenors in the case, Ohio Citizens for Responsible Energy ('0CRE') and Sunriower Alliance Inc. et al.,

on a number of grounds.

This Portial initial decision concerns the lost such issue, hydrogen control, to be considered in this proceeding.

This issue

(* Issue N8') was the subject or on evidentiary hearing held April 30 through May 3, 1985 in Perry, Ohio.

The decisional record of the proceeding on Issue No. I consists of the testimony and exhibits filed by the porttes, and the other evidence contained in the transcripts of the hearing.

Appendix A to this decision identifies the location of written testimony in the transcript.

Appendix B lists the exhibits identified, indicates the Board's ruling on any offer of on exnsbit into evidence, and identifies the location of admitted exhibits in the transcript.

In preporsng our decision, we reviewed and considered the entire record and the proposed findings of fact and conclustons of low submitted by the porties.

Those proposed findings and conclusions that are not incorporated directly or by inference l

1e

_3-

- s i n this portial initial decision are rejected as bezng unsupported by the. record or the cose or os being unnecessory to the rendering or this decision.

.This Board *s Jurtsdiction is limited to a determinoeion or rindings or roet and conclusions or low on matters put into controversy by the porties to the proceeding or round by the Board to' involve o serious sorety, environmental, or common

- defense and securtty question.

10 CFR 2.760o.

The Board hos mode no'such odditional~ determinations in this cose.

= --

' E.

BACKGROUND OFI55UE MS Issue No. 8 concerns the ability or the PNPP Mark III containment design (and equipment therein) to withstand the errects or-the ~ Combustion or hydrogen gas, produced from the reoetion.or ruel rod cladding with steam during a degraded core ocesdene.

Born OCRE and sunflowcr originally submitted hydrogen control contentsons.

The Boord denied admission or the contentions because they did not. meet'the'stondordifor the admission or hydrogen control contentions articul'oted in Metropoliton Edison Co. ~(Three Mile' Island, Unit 1), CLI-80-16, 11 HRC-674 (1980).

See Speciol'Preheoring. Conference hemorandum I rder'Concerning Party Sectus. Motions to. Dismiss and to Q

and

$ toy, the Admissibility or Contentions, and the Adoption or Special Discovery Procedures, LBP-81-24, 14LNRC 175, 207-08

.(1981).

CLI-80-16 required that on-intervenor wishing to litigote hydrogen control had to first specify a. credible L

1

. __ _ ~ ~...

.a 4

-q.

1 -

[

occident - scenario entailing hydrogen generation, combustion, contoinment breach or leakoge, onc offsite-doses exceeding 10 CFR Port 100 guidelines.

I Several months later, Sunflower moved to resubmit its hydrogen. control contention.

On Norch 3.

1982 we granted that i

oOtion, largely due to the publication in the Federal Re9ister of.o proposed rule wherein additional protection was to be J

required in Hork III BWRs (such as Perry) ato provide assurance that large amounts of hydrogen con be safely accomodated

[

i 46 Fed. Re9. 62281, December 23, 1981.

See LBP-92-15, 15 NRC 555. 560 (1982).

Applicants then moved for directed certification of that decision.

Their motion, opposed by the i

j NRC Staff and the intervenors, was denied by the Appeol Board, 1

ALAB-675, 15 NRC 1105 (1902),

i In LBP-82-15, we reworded the hydrogen control contention t

submitted by Sunflower to read:

I Applicant.hos,not demonstrated that the manual operation of two recombiners in each of its Perry units is adequate to assurs that large amounts of hydrogen con be safely accomodated without j

o rupture of the containment and

0. release of substantial j

quantities of radioactivity into the environment.

i i

(We subsequently designated OCRE os~the lead intervenor on this issue.-

See September 17, 1982 Memorondum and Order (Concerning j

Procedural Horions).)

l Discovery on Issue N8 closed September 30, 1982.

It I

I included one set of' interrogatories to Applicones by OC9E, one see or interrogatories to the NRC Staff by.OCRE (the responses to which we hod to compel), and a set of interrogatories and f

request for production of. documents to OCRE from'Appliconts.

?

i t

.--_,,_,__,,.-m...,.

-S-s Discovery marked the onset of a two-year dispute between staff and A p p l i c a n t s ', on one hand, and OCRE, on the other, concerning the scope of the issue and the need for the specification of a 1

credible accident scenario, pursuont to CLI-80-16.

The tortuous ihistory of Issue M8 is outlined more fully in our Horch 14, 1985 Memoroneum and Order (Motions on Hydrogen Control Contention).

OCRE riled o motion to reopen discovery on Issue N8 and several other issues on November 15, 1983.

We denied this j

motion, _but suggested that Applicants update their previou!

l aiscovery responses.

December 20, 1983 Memorandum and Order i

1 (OCRE Nation to Reopen Dascovery).

Applicants did 50, and the information supplied therewith prompted OCRE to move for the l

a reopening of discovery on July 30, 1984 and simultaneously to file onother set of interrogatories to Appliconts.

Appliconts a

voluntarily responded to some of the interrogatories and j

provided 50me of 'the documents requested.

1-i The Commission adopted the new hydrogen control rule for l

BWR5 with Mark III containments (and PURs with ice condenser i

containments) on January 17. 1985.

50 Fed. Re9. 3498.'

The rule i

requires the 2nsto110 tion of systems at such facilities capable of sorely handling the amount of. hydrogen generoted from o 75%

i metal-Water reaction.

On January 20, 1985 OCRE filed o motion seeking the rewording of Issue M8 to os to conform its lon9uo9e l

l With that of the new rule.

On January 28, 1985, the stoff moved l

for suomary disposition of Issue W8, based on its narrow' Wordings the contention's wording focused _on recombiner f

.aceauccy, while Applicants are using on igniter system for l

I --

degraded-core hydrogen control.

OCRE's motion was opposed by staff. and j

l

.,__.,...._...-______.__m._,_.

Applicants, wbtle the Stoff's motion was supported by Applicants and opposed by OCRE.

In our March 14, 1985 Memorandum and Order we granted OCRE's motion to reword Issue #8s Issue M8 now reads:

The Perry hydrogen control system is inadequate to assure that large amounts of hydrogen con be safely accommodated without a rupture of the containment and o release of susbstantial quantities of rodtocetivity to the environment.

In our view, this wording basically alleges that APPliconts' hydrogen control system does not conform to the new regulatory requirements of 10 CFR 50.44.

Our oceton rendered the Staff's motion moot.

On March.15, 1985 OCRE filed a motton for o continuonce of

/

WB / rom April 9 to June 3, alleging that the heartn9 on Issue T

additional Information was forthcoming on the issue, that there i

was insufficient time to prepare for a hearing on two complex technical issues [this issue and Issue #16, on diesel generneors], the unavailability of unnomed witness (es), and the lack of horm to other porties ensuing from the proposed extension.

In response to ourVDquest. OCRE indicated that it would not proffer Witnesses on Issue M83 Applicants and OCRE then agreed upon April 30 as the hearing date for Issue M8.

In light of this agreement, and the fact that the Stoff's SER on this issue would not be available untti mic-April, we continued th4 heartng until April 30.

See March 29, 1985 Memorandum and Order (Motions for Continuance of Hearing and to compel Appeoponce of f4RC Witness),

On March 18, 195* OCRE moved to compel the appearonce or Dr.

~

~

Norshall Berman of the Sandto National Laboratory.

OCRE alleged that there existed a genuine scientific disagreement between Dr.

Berman and the Staff on the basis of statements in the Grand Guir SER on the odequacy of the hydrogen igniter system.

AffidoVits filed by Dr. Berman and by Allen Notorrancesco, o Stoff witness, in response to OCRE's, motion indicated that o disagreement did not exist, and that the Stoff wsenesses were familior with-the Sandid research program on hydrogen control.

Thus, the exceptional circumstances necessory to compel the oppearance of a porticularly-named Storr witness were not present, and we I

denied OCRE's motion.

(See 10 CFR 2. 720 (h) (2) (i). )

Horch 29, 1985 Memorondum and Order.

Applicants' witnesses at the hearing were Eileen M.

Eu::elli, John D.

Richardson. Kevin W.

Hole clow, Roger W.

Alley, Dr. Bernard Lewis, Belo Karlovitz, and Dr.

G.

Martin Fuls.

Ms. Bu :elli is a senior licensing engineer with CEI. and is responsible for licensing issues relating to the PNPP hydrogen control system and for the preparation of the preliminary GVoluotion, submitted pursuant to the new provisions of 10 CFR 50.44.

Mr. Richordson as employed by Enercon Services and serves os a consultant to CEI on the hydrogen-control issue.

He also was employed by Mississippi Power and Light Co.,

the Grand Guir licensee, and handled all licensing and technical issues relating to the Grand Gulf hydrogen igniter system.

Mr. Holt: claw is a licensing engsneer with General Electric Co.

He hos managed the GE Severe Accident Progrom, which hos

.. s included

t. e the EWR/6 Standard Plant Probabilistic Risk Assessment and associated evoluotions, including hydrogen event risks.

Nr. A11ey'is employ 5d by Gilbert /Commonwoolth, Inc, the PNPP I

m,Ar' chi'tect-Engineer..where he is the manager or tne Structural Nuclear Engineering Section, Mr. Alley was involved in developing the origin $1 ASME contoinment design specification for PNPP and hos been responsible for finite element and stress analyses of the PNPP-containrent and drywell.

'D r. Lewis and ti r, Karlovie: are founders of Combustion and Explosives Research, Inc. L'C6mbex.'), o consu.Iting,ritm' providing services.to government, research institutes, and industry.on the'rundamentals or comeustion, flames, ignition, and explosions or gossi, liquids, and solids.

Dr. Lewis and Mr.

korlovite. hove extensive expertence in ene field or combustzon and have authored many papers on this subject.

Dr. Fuls has o PhD in mechanical engineering and developed the CLA5IX-3 computer program, used to evoluoteithe consequences of hydrogen burning in the PNPP containmene, The NRC Storr's witnesses were Dr. William Trevor Pratt, Allen Hotorrancesco, Li Yong, and Hukom Gorg, Dr. Pratt is the Principal Investigoror or the Accident Analysis Group of the Brookhaven Nationoi Laboratory and is romiliar wit'n'the MARCH

. computer code.

He has conducted-o number or studies on core meltdown 0ccidents and rev-iews or probabilistse risk ossessmenEs', including'thE-GESSAR-II PRA.

Dr..Pratt holds o PhD 9

o 3

- i n me choni' col e ngi nee r s ng and hos authored or co-outhored many

.)

I

9-t s

publications.

Mr. Notofrancesco is a Containment Systems Engineer in the NRC's Containmene Systems Branch.

He is the lead staff reviewer of the PNPP hydrogen igniter system, i

Mr. Yang is a Structural-Engineer in the Structural and J

l

.Geotechnical Engsneering Branch or the NRC.

He is responsible for review of. seismic and structural subJeets in the Perry' Final

Sofety Analysis Report and'hos contributed to the Perry SER.

Mr. Gorg is employed as on electrical engineer in the

.f Equipment Qualification Bronch of the NRC's Division of Engineering.

His responsibilities include review of licensee progroms for environmental qualification of safety-related 1

equipment and of equspment survivobility during and ofter a hydrogen burn event.

OCRE presented no direct testimony, but cross-examined both Stoff's and Applicants' witnesses.

In response.to this cross-J exomanation -Applicants-presentqq o rebuttal-witness,, James H.

Wilcox, o CEI weld'ing engineer, i

C.

APPLICABLE STAi4DARDS i

The regulation governing our disposition of Issue #8 is the Commission's new hydrogen control rule, 10 CFR 50. 44 (c) (3) (iv)-

j (vii).

This-final rule was published in the Federal Regiseer on Janu.'Y 25, 1985 (50 Fed. Reg. 3498).

This new rule requires that BWRs with Mark III containments and

'FWRs with ice condenser containments implement hydrogen control systems copoble of handling large amounts.cf hydrogen, as would I

result from o Three Mile Island-type occident.

I, i

At the outset it must be stressed that compliance with the l

rule connot be avoided by orguing that a degraded core occident

?

-foe resulting in the generation of Eubstantial quantitie5 of hydrogen is Unlikely to occur, os Applicants hove done (Appliconts' Testimony at 10-16).

Such on orsument is in errect on impermissible challenge to the rule.

The commission considered (and rejected) comments alleging that degraded core accidents are unlikely during the rulemoking process.

50 FR 3499.

10 CFR 50.44 (c) (3) (iv) ( A) requires Hork III piones to have a hydrogen control system capable of handling the amount of hydrogen generated from o 75% metal water reaction without loss of containment integrity.

This system must be supported by a suitable program of experiment and analysis.

10 CFR 50. 44 (c) (3) (vi) requires that the analysis (B) (1) evoluote the consequences of the generation of large amounts of hydro 9en (that equivalent to a 75% metal-water reaction),

including considerotson of the hydrogen control measuress (2) include the period of recovery from degraded core conditionss (3) use occident scenarios describing the behavior of the reactor system during and following a degraded core occident; (4) EuPPort the design of the selected hydrogen control systems and show that' (5) (i) containment' structural' integrity;is*

c maintained and (5) (ii) necessary equipment will survive the environment created by the burning of hydrogen.

The lateer ewo points are explotnea more fully in Sectsons (iv) (B) and (V) of the rule.

Section ( tv) (B) requires that containment structural integrity be demonstrated by on analytical technique which describes the containment response to the structural loods involved.

Two examples of occeptable I

-II -

onolytical techniques are given:

one is on undescribed method including the use of actual material properties wtth suitoble margins to account for uncertainties in modeling, material properties, construction tolerances, etc.

The other is a showing that steel containments meet the requirementE of the ASME Code, Service Level C Limies, considering pressure and dead load alone.

Section (V) requires that systems and components needed to establish and maintain safe shutdown and to maintain containment integrity must be copoble of performing their functions during and ofter exposure to the environmental conditions created by the burning of the amount of hydrogen resulting from o 75%

metal-water reaction, including local detonations, unless such detonations con be shown unlikely to occur.

While the "Supplementory Information' section of the Federal Register notice is not technically port of the regulation, it is port of its ' legislative history" which is useful in interpreting this regulation.

From that discussion we rind thoe the Commission's policy for preventin9 excessive radiation dose to the pubite is best assured by maintaining a leak tight contornment.

This is to be demonstrated by compliance with section (iv) (B), which requires a showing of ' structural integrity with margin."

50 FR 3500.

The methods specifically outlined in the rule are meant as examples onlys other methods may be acceptable, provided that

' convincing evidence" 15 present9d regarding their suitability.

50 FR 3501.

Similarly, the discussion 4t 50 FR 3501 illuminates the

m

-Q~

\\

i requirements of.$ection (V) on equipment survivability.

Equipmene survivability is esseneiolly equipment qualirication, l

except that the margins considered odequate ore less that those required for design-basis equipment qualification.

There must f

be odequate evidence, generated by the qualirication methods of 10 CFR 50.49(f), that components needed for mitigating the i

~ consequences of an occident, for maintaining integrity of the s

containment pressure boundary, for maintaining the core in o

}

tore condition, and for monitoring'the Course of on occident are

?

I Capable of functioning during and folloWing exposure to the environmental-conditions created by the postulated occident, including the burning of hydrogen.

The ciscussion cescribes on acceptoble method of analysisi first on acceptoble contotoment thermal i

analysts must ce performed.

Tnen, che enermal response of 1

ecutpment analy ec thereby should be compared to the thermal 1

responge of equipment curing qualificoricn testing.

Tne uoltfiCotton responEG Ihould enve1 Ope the hycroGer. burn thermal "esecnie.

521ectec reits ma/ Glio be necessary torossure survivacility.

The operabt;n of the hydr 09en Control E'Etem muEt not

'arther os;rr.vate the c urse of the occi=ent c.

enmar.9er :ne i

pione curins normal opernetens.

50 FR 3500.

t Further. guidance from the " Supplementary Informatton' section Will be usea in our determination of compliance with the rule, infro.

Section (vii) of the rule conhoins on unusuoi implementation Schedule.

Each applicone as required by section (vii) ( A) to l

r

--r 4..,

e

-l.5 -

submit by June'25, 1985 a schedule for meeting the requirements of the rule, Section (vs t) (B) requires OL opplicants to comply with"Seetion..(iv)(A) prior-to, exceeding-5% poweri.'hoVGVGr, complete rinal analyses are not required ir on acceptable preliminary onalysis has been submitted.

The rule gives no 4

criterio for determining what shall be addressed in the preliminary analysis and what may be deferred to the final 4

analysis, although it is the Commission's objective that the rule be complied with Without undue delay, Section (vii) (D),

We reserved our Judgement regarding this matter until the evidence was recetved, Our conclusions on the adequacy or the scope of the preliminary analysis submitted by Appliconts are presented below, A rinol word is needed.

Section (vii) (B) and other ports or the rule ((iv) (B) and (vi) (B) (3) ) appeor, if taken li terally, to stVe the Starr the sole power to determine whether compliance with the rule has been achieved.

The testimony presented by Starr and Applicants, not surprisingly, endorses this interpretation, I.e.,

they would have us summarily offirm the Stafr's rindings, we connot accept this interpretations were We to do so, We Would be abdicotang our responsibility as judges, i

The NRC Starr is not the trier or roct in this proceeding: We j

are.

The staff is just another party, the views of which Connot be giVen any more Weight than those of the other parties.

l

. Vermont vonkee Nuclear Power Corp, (Vermont Yonkee Nuclear Power Storion), ALAB-138, 6 AEC 520, 530 (1973): Consolidored Edison (Indian Point Nuclear Generating Station, Units 2& 3), ALAB-s m

~-

,v~,

_s...,

. 1 e

i 304, 3 NRC 1, 6 (1976)s southern california Edison (son Onorre Nuclear Generating Station, Units 2 & 3), ALAB-268, 1 NRC 383, L

~

389 (1975).

Thus, we muse subject the starr's position to the some scrutiny 4

as that of the other parties, and reject it if necessary.

D.

COMPLIANCE WITH THE RULE 1.

Operation or the System wi.,lijcont ine our o ttent, ion; to the, hydrogen con trol-sys tem 1

We used for degraded core events, the distributed igni,ter system.

Finding 4.

We nche that another hydrogen control system, the l

combustible gas control system, is used for design basis events.

Finding 3.

Because Issue M8 concerns the degree or compliance i

l with the Cct.nission's new rule on degraded core hydrogen control I

(Finding 2), we will not consider the design bosts system rurther.

The distributed igniter system consists of glow p)us 4

{

igniters distributed throughout the containment, Wetwell, and drywell.

Finding 4.

The system is powered by diesel generator-i backed AC class 1E power systems.

Id.

The system is designed j

to burn off the hydrogen resulting from o degraded core occident at low concentrations, so os to prevent occumulation of more dangerous, higher concentrations.

Id.

The system is to be manually actuated, in accordance.with i

PHPP emergency instructions, when the reactor water level

{

reaches the top or the active fuel.

Findin9 6.

However, these instructions, and the generic emergency procedure guidelines on i

i which they are based, are still under development.

Finding 7.

-v

-,,.,we n-n--y----m

.-y m

,u.--

n---i.-

---a n.+.--w w,-

eff*

The vent path for containment veneing, to be used for overpressure threats to the containment, has not even been established.

Finding e.

Although Appisconts clotm that the procedures will be available before PNPP exceeds 5% power (Tr. 3426), we cannot approve procedures Whach we have not seen and which are yet unfinished. 10 CFR 50. 44 (c) (3) (vii) (B ) requires the hydrogen control system to ce installeo and operable before exceeding 5%

power, (Compare 50 FR 3500.)

We interpret this section to mean that the Procedures for operating the system must also be finalized.

Applicants opparently shore this anterpretation.

The record indicates that these Procedures are of sufficient complexity to worront our thorough evoluotion.

For example, on analysis of containment hydroben concentrd, tion'is'necessory in those situations (such as o station blockout degraded core occident) where hydrogen concentration may be high.

Finding 13.

i Potential radiological releases are evoluoted before containment venting is done.

Tr. 3443 (Richardson).

We note that Applicants' design basis offsite dose calculations are in a state of flux (Tr. 3265-70, 3600, 3652-57 (Bu :elli)), and the flow path and rates for contoinment venting are undetermined.

Finding 8.

We therefore have no way of ossuring that the procedures used will be approPriore, that instruments relied upon by the operators are avoilable or reliable under occident conditions (compare OCRE Ex. 21 at 196), or that 10 CFR 100 guideltnes will i

i be met for containment venting.

We will require that the i

finalized procedures be prcvided for our scrutiny (and that of

?

16 -

the parties) and opproval before PNPP operates above 5% power.

2.

Scenarzos 10 CFR 50.44 (c) (3) (vi> (B) (3) requires Applicants to use in their onolysis occident scenarios that describe the behavior of the reactor system during and following a degraded core

accident, Applicants in their preliminary analysis evolunted two scenarios, a small break in the drywell with extended ECCS failure ('DWB"), and a transiene with a stuck-open relief Volve with GXtended ECCS foilure ('50RV').

Finding 9.

Recovery was

-ossumed to occur at the point of a 75% metal-water reaction.

i Finding 10.

Both scenarios used identical steam and hydrogen release histories calculated by the MARCH code.

Findsng 11.

The steam and hydrogen were assumed to be released to the drywell os Well 1

J cs to the suppression pool in the DUB cose.

Id.

The MARCH onalysis did not model recovery, but wos modified to yield a 75%

i

]

metal-water reaction.

Id.

I Since.the Commission considers the use of a single release history, with variation,0f key,porometers, to.be oc,ceptoble (50 FR 3502), we consider Applicants' approach with respect to the release histories for these two scenarios to be acceptable.

A l

75% metal-water reaction was assumed, os required by the rule.

We are aware of the uncertosnties ossociated with the NARCH code (OCRE Ex. 21 at 11, 18s Hoeorrancesco Testimony II at 4-5):

nowever, since the analysis did use on amount of hydrogen correspondtng to a 75% metol-water reaction, and st is not apparent that better onalytical tools exist (compare OCRE Ex. 21 ot 11), we find the use of-MARCH occeptoele.

-c

-,e.

m

,,7

,y-----

y

,7 r~

.-n-

-n.

Our findings on other o3pects of these 2 scenario 5, such as those used in the containment response analysis, are presented in that section of our opinion, infro.

we are concerned that the scenarios chosen do not represent the most severe challenge to containment from hydrogen combustion, as is required to be considered.

50 FR 3502.

The station blackout degraded core occidenk is such a scenario.

The igniters would not be available during a station blackout occident.

Finding 12.

The hydrogen would occumulates o 75%

metal-water reaction would result in a concentration of 28% in the containment.

Id.

A deflagration of this amount of hydrogen (os would result from igniter system actuotion when power is restored) would produce high overpressures.

Id.

Although Applicants' emergency procedures will have provisions for such situations (Finding 13), we cannot conclude that these provisions (measuring containment hydrC9en concentration and containment venting) are acceptable.

As discussed above, we Connot accept these unfinished procedures on faith.

Even if such procedures could be accepted. We are not convinced that they Will prevent deflagrations caused by rondom ignition sources.

We therefore find Applicants' spectrum of Scenarios to be deficients station blockout must be satisfactorily addressed before exceeding 5% power.

At',the*heofin9 APplicantt' objected to anf,dfscussion of.

~

station blackout, Clotming that it is on issue more appropriate for the final analysi! and implying that the new rule prohibits its consideration.

Tr. 3408-30.

We con easily reject their latter objection.

The Staff is requiring thor station blackout

's

-/E-I be considered in Applicants' final analysis, unless there is

)ustification for its exclusion.

Notorrancesco Testimony II ot t

6.

We don't see hoW this fact con be squared With on interpretation that the rule precludes o consideration of 0

m.

station blockout.

Moreover, os Applicants admitted (Tr. 3429),

the sentence at 50 FR 3502 on Which they rely pertains to the Starr's rindings at ice condenser plants.

The Commission goes on to state that I

applicants at different reactors may have to oddress other scenarios. 'Furthermore, we do not interpret the Commission's 1

statement that the rule does not require backup power for all types of hydrogen control systems at oli ofrected plants as a prohibition of a finding that backup power is necessary in o i

partscular proceeding, i

As f or Ap'plicon ts ' first objection, we would first note that a thorough discussion of the issue of preliminary vs. Final i

analyses is given below.

We think that this issue is of such i

importance that it must be oddressed now.

Indeed, we see no valid reason to defer it to the final analysis.

It may even be to Applicants' benefit to consider the issue now, as any design changes we may require might be mode.more eosily'now than after power operation.

Nor do we consider Appliconts' dtscussion (Tr. 3609-10) of f

the low probability of an extended station blockout leading to o i

degraded core occident (bosed on on onolysis supposedly showing that RCIC would keep.the core cooled for 9 hours1.041667e-4 days <br />0.0025 hours <br />1.488095e-5 weeks <br />3.4245e-6 months <br />) to be dispositive.

First. Applicants' witness Holk: claw was unsure i

f whether analysis ossumed that the RCIC system took suction first I

-l9 -

From,the suppression pool or the, condensate,storose tank...Tr.

3660.

On redirect he said that he believed that RCIC took

' suction first from the suppression pool.and then from the condensore storo9e tank.

Tr. 3664.

This method of operation is controry to that normally occurrin9.

Normally RCIC takes suction from the condensate storo9e tonk, and automatically switches to the suppression pool on low CST water level.

NUREG-0887, PNPP SER, Hoy 1982, at 7-25 to 7-27.

It thus appears that the method of operation claimed by Applicants would involve override of automatic functions.

This fact, along with the uncertainty expressed by Applicants' witness, and the unovailability of the onolysis for our scrutiny (and that of other parties) leads us to reject Applicants' claim.

We believe this is consistant with the Commission's finding that the new hydrogen rule is necessary to cope with unexpected events, regardless of the likelihood of hydrogen 9eneration therefrom.

50 FR 3499.

We therefore conclude that station blockout must be analyzed for degraded core hydec9en control, with our opproval, before PNPP exceeds 5% power.

3.

Contoinment Integrity (o) Containment Vessel capacity (t) Positive Internoi Pressure Applicants have performed analyses of the PNPP Mark III containment vessel to determine its ultimate internal pressure

~

~..

capacity.

Finding 14.

Their analyses included determination or static shell capacity, buckling analyses, and evoluotions of

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

• 4 k

penetrations according to service Limits c and D or the ASME I

code.

Findings 15, 16. 19s Applicants' Testsmony at 24-28.

k The Starr requested that Applicants perform on analysis or the PHPP containment demonstrating that the requirements or the i

t 4

ASME Code. Service Level C.

or subarticle NE-3000 ore met for on interna %* pressure of-45 Ps19 (Chd.7consideri ng,r dead lood)..

Finding 17.

This is essentially the some os the *occeptable method' of 10 CFR 50. 44 (c) (3) (iv) (B).

Appliconts performed this analysis; they believe that the.

most limiting penetration (P414) has o pressure copocity or 50 psig.

Appliconts* Testimony at 28.

However. We note that dead Icod, which the Starr and the rule require to be considered. Was neglected for penetrotions other than the personnel airlock and l

equipment hotch.

Finding 00.

Dead lood contributes.o maximum or 4.7% or the stress intensity contributed by ene internal i

pressure.

Id.

l l

Appliconts oiso neglected the errect or elevated l

temperatures due to hydrogen combustion on material properties.

i Finding 24.

Appliconts admit that this could decrease material I

i strength by 16,.'.

Id.

l Applicants also neglected the errects on stresses resulting

)

from the as-built, out-or-tolerance conditions or the PHPP

}

j contoinment vessels.

Finding 25.

Analyses at the design pressure. 15 psig. showed stress increases from 4 to 20 percent.

I I

Id.

l Appliconts have defended their neglect of these items by loaming that they are insignificant.

See, e.g..

Applicants'

(.

Ex. B-4 at 16 (dead lood): Alley. Tr. 3506-87 (elevated i

I '

temperature, claimed to be insignificant because there is l

l sufficient margin betueen the minimum properties used and the octual material properties of the limiting penetration. P414);

Alley, Tr. 3596 (out-or-tolerance condition, said to be insignificant os the Vessel shell capacity is greater than that i

of the limiting penetration),

4 We disagree.

We note'that containment capacity is on important factor in determining the adequacy of the hydrogen control system, 50 much so that on increase or decrease of 5-10 pst. con be significant.

OCRE Ex. 21 at 9-10.

If ue make the i

simplistic assumption,thot thp,oboye-no.ted effec,ts ore-odpitive, the containment copocity could decrease by 18.7 to 34.7 percent.

l I.e.,

instead or 50 psig, the capacity uould be 32.6 to 40.6 ps19.

We also disagree with Applicants' rotionali:Chions of these effects.

The hydrogen rule requires dead load to be considered, regardless of its taognitude.

Applicants' assertion that the i

I out-of-toleronce condition is insignificant is Predicated on the assumption that only the Ve$sel shell, and not the penetrations, j

Will be offected thereby.

This assumption is unsupported by 1

evidence, as no calculations uere performed to consider higher internal pressures.

Finding 25.

o Applicants' orgument that there is sufficient margin in the material properties used to account for the effects of elevated i

temperature also foils.

An examination of Applicants' Ex. B-4 i

shous that P414 is not the limiting penetrations P205 is.

For enis penetrotton, the stresses exceeded ASHE code allouables, j

assuming minimum specified yield strength.

Applicants' Ex. 8-4 l

f r

I vy-y y

y

---a-.,.--s e.

-gw

--m-,,e ae-

.----es-

i

-u-ot Toble 10.

Applicones therefore took credit for the octual

.I material properties, which allowed a pressure limit of 54.8

{

psig.

Id.

Findtng 20.

Had P205 been treated the some os all 4

other. penetrations, the limiting pressure would be 57000/63500 x i

j-45 or 40.4 psig, using the stresses and factoring method of I

Table 10 or Applicants' Ex.

8-4.

i' A decrease of 10% in strength due to hydrogen burn temperatures for P205 would result in a pressure capacity or 49.3 psig, using the octual material strength.

While this is only t

I 1119htly less than the pressure for P414, 4,9.7 psig (Finding 20), it illustrates that the use of actual material properties 1'

for P205 results in no margin to account for uncertainties in t~

l material properties, modeling techniques, or construction tolerances, os required by 10 CFR 50. 44 (c) (3) (iv) (B) (contrary to Applicants' ossertion at-Tr. 3589-90).

' 'Applicunts( o na l y s i s' *d i d, n c e 'c o n s i d e r. t h e

• e,f f e c t of' f

defective contoinment welds.

The PNPP containment vessels l

contain welds deviotsng from ASME code requirements.

Finding I

'd.

A' fracture and fatigue onolysis was performed, for design i

l Dosis loadings, by Aptech Engsneering services which indicated that the defects would not propogote for the loods considered, j

Finding 07, 33.

We believe thoi hydrogen combustion loads (as determined l

from on appropriate Containment response onolysis) should be l

evolunted for their effect on these defects.

We realice that l

the Aptech analysis convoins conservatisms, such as the i

temperature assumed for determining fracture toughness.

Finoing 33: Tr. 3590-91.

However, we note that the minimum rotic of fracture toughness to applied stress intensity factor is 1.34.

Finding 33.

I.e.

o 34% increase in stress would result in fracture initiation.

(According to the principles of linear elastic fracture mechanics, fracture results when the opplied stress intensity factor equois or exceeds the fracture t-ghness.

OCRE Ex. 13 at 2-4.)

We 'olso believe that conservatism is necessary, os the rule requires that there be structural integrity with margin.

We t

also are concerned that the technique used by Aptech to r

determine flod size moy not be sufficiently oCCurate.

We note the iRportance of flow si:e in the fracture mechanics 4

j evoluo: ion.

Finding 30.

The Staff considers the digital 1

radiographic enhancement technique used by Aptech to be 1

~

occeptable only when the radiographs used are of good quality, 4

and that it connot be considered a standard technique without l

further demonstrations of occuracy.

Finding 09.

The PNPP weld i

rodtographs sn question are of poor quality.

Finding 26.

The welds are now inaccessible for repair or re-rodtogrophy.

OCRE Ex. 18 at 1.

We would not chorocterire the ' result $ of calibration

• of the digital enhancement

,t e c h n i q u e given in.Toble.2 1 P. 11 Of,A,ppendix a of 4

OCRE Ex. 13, showing t h.1 t the technique con be from 0 to 90%

4 off, os o demonstration of occuracy.

Given these uncertainties, and the severe consequences of fracture initiotion (Findings 31, 30), we believe that on evoluotion of contoinment copocity must 3

include o thorough assessment of these defects for' hydrogen f

combustion loads, l

l At the hearin9 Applicants claimedithot they had evoluoted i

the stresses at these uelds for 50 psig and found that the

.- - -. _ _ - - - =

i 1

I stresses at the higher internal pressure Were less than at 15

~

psis, os analy ed by Aptech.

Alley, Tr. 3306, 3313.

We con gave tnis testimony no weight because the calculations are noe i,

f 4

available for the scrutiny or the 8oord and porties (Tr. 3315),

because the Cloimed stress reduction seems preposterous, and 5

beCouse We hove serious Problems With Mr. Alley *S Credibility.

For example, Mr. Alley stated thot the stresses in the i

containmene vessel are less than those onolyZed by Aptech, due to the addition of the annulus concrete.

Tr. 3313-15.

Mr.

l Alley later stated t

1 that the Aptech analysis did not credit for-the onnulus concrete (Tr. 3593). which is contradicted by OCRE Ex. 18, o Starr Eummary or o meeting on the subject which Mr. Alley attended.

1 Mr. Alley claimed to be unowore of the problem with the surroce roughness or the Welds (Tr. 3324), osain discused at the meeting summari:ed by OCRE Ex. 18.

Rother than admit that the digital enhancement technique could be inaccurate by a roctor or 10. Mr.

Alley questioned the occuracy or o table in the Aptech report i

(Tr. 3323), which he later Chorocterited 05-on excellent report

)

(Tr. 3341).

Mr. Alley's contradictory and evosiva responses do i

not meet our standard of reliable, probotive evidence.

(Nor do we consider the Starr's testimony on this matter to meet enis standard.

.Although Mr. yong stated that APPlicont's analyses demonstrate the adequacy of the at-built contosnment i

pressure capacity (Yong. Testimony at 2), his opinion.is based on i

a revseW of the onolyses contained in Applicants' Ex.

8-4.

l wnsen, os discussed previously, did not consider the as-built ecnditton of the containment.

And his statement that the I

(,

'derective welds would not impair the integrity of the l

l

.- _ __ _ _ _,_ -. _ _.._ _. _. -__ ___ _, _. _..--- ~,--- __.- _ -,_ _

.__m...

..__,.m

__..m

. _=. _.

-2 6 -

j containment (TP. 3733) was based on on opporently limited review l

t or the Aptech report, Which only addreffed de5ign bosis loads.

Finally, We must state that we find the cursory. ' rubber-stomp

• 7 noture or the-i

}

5tarr's testimony to roll short or our requirements.

The Starr j

should not support on opplicotton.

Carolino Power f,

Light j

(Sheoron Harris Nuclear Power Fione. Units 1-4),

LBP-77-19. 10 NRC 107 (1979), mediried and arrirmed, ALAB-577, 11 NRC 18 (1980).)

As-discussed above, the hydrogen rule basically requires e

that the contoinment serve os a leck-tight barrier to the i

i release or rodiotion.

Appliconts ogree that this is the' j

containment's function.

Finding 35.

We ore not convinced that j

meeting the ASME service level C limits ensures o leak-tight 1

containment.

I The PNPP equipment hatch is unseated by increasing internal i

pressures the only force opposing this action is the preload on the closure bolts, which is overcome at 21.6 psig.

Finding 37.

i j

Applicants take credit for the spring-back or the 0-ring seals l

to prevent teokoge.- Finding 38..However, Applicants neglect I

[

the phenomenon or compression set, Which increases os a result I

4 of thermal and radiation oging.

Finding 39.

We connot accept Applicants' rationali:otions or thiS c

problem at the hearing.

Mr. Alley claimed that maintenance 4

i i

procedures, Which he could not identify, would prevent

(

i l

compression set from becoming a problem (Tr. 3278).

Mr. Alley, who previously did not know ir the' hatch 0-rings Were made or 1

i i

f l

l

l

+

ethylene propylene, later testified that the 0-rings (now noming the specific compound) would not,be subject to, compression, set for the temperatures one pressures expected in o hydrogen burn event (Tr. 3582).

Hr. Alley mentioned the Arrhenius equation as a method by which this was determined.

But, later he didn't know What synergistic errects are (Tr. 3650).

We take orricial notice of the roct that the Arrhenius methodology 15 a method or addressing occelerored osing for the purposes or equipment qualificotton.

See NUREG-0538. Rev.

1, ' Interim Starr Position on Environmental Qualification of Safety-Reloted Electrical Equipment' or 16.

Page 15 or that document oedresses synergistic errects. Which i

should be-considered in acceleroded aging progroms.

(See also 10 CFR 50.49 (e) (7). )

We do not believe that a wseness who does not know what synergistic errects are con be considered I

compe te n t to testify on the ofrects or elovated temperature on 0-ring behavior and ogsng.

(We likewise reject Mr. Alley's testimony (Tr. 3593) on the i

smoothne55 or horch seatsng surroces 05 locking in credibility.)

We therefore conclude that equipment hatch leokoge (including the errects or ;ompression set) due to elevated temperatures and pressures must be r c ria a l l y considered.

We must also have reliable evidence concerning the errects of seal wear ona secesng surroce roughness on seal lookoge during degraded core occident conditions.

We likewise are concerned about the ability of the inflatable seals used in the personnel oirlocks to resise lookoge.

Finding 40.

They are considered to hove poor

27-resistance against severe occident conditions, and on inflatable seol useo n drywell oirlocks railed qualirication eesting.

Id.

We are not persuaded by Applicones orguments that hydrogen s

combustion conditions are expected to be less severe than design basis conditions, os this expectorion is based on the CLASIX 3 Code.

Tr. 3363, 3623.

Our dssopproval or the CLASIX 3 code is addPessed below.

~

~

~

• Finally, we must oddress the obility of Applicants' onolytical techniques to describe the containment response to the structural loods involvec.

At the hearing, Applicants orgued that questioning the validity and opplicability of their onalytical methods was in effect challenging the rule.

Tr.

3784-85.

We diso9ree.

The rule does not require the use of the 45t1E code, but permits other onolyses.

The evidence in our record indicates that the ASME code moy not be the most conservat1Ve method for oil ports and conditions encountered in the PriPP containment.

The ASME method does not consider leakage from the pressure-unseating equipment hatch.' A fracture mechanics onolysis is the oppropriate methodology for evoluoting defective welds.

And the ASME code would allow the dome knuckle to reach 82 psig (57000/31203 x 45, from Table 8 of Applicants' Ex.

8-4, based on minsmum specified yield stress), which is above the buckling pressure (determined from methods which may not be conserVotive) at which shell fracture may occur.

Finding 16.

I We therefore find no prohibition to questioning the Validity l

Of Appliconts' onolysis, We are porticularly concerned that Appisconts used fantre kloment techniques in thetr onolyses of

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

2g.

contonment vessel penetrations.

Finding 21.

Experiments t-conducted by the Sondio lectional Laboratory to qualiry such techniques have shown that they may be nonconservative.

Finding i

20.

Although Applicants claimed that their techniques are i

applicable for the lineor. elastic stress ranges under I

3 constderation (Tr. 3393), we are not sure that their analyses ore in race confined to these stress ranges.

The ASME code is r

based on redisersbution or stresses.

Finding 23.

Even the less 4

j.

severe service 1l 1evel A or B loodings may produce allowobis strosses exceeding i

1 j

yield and approachtng ultimate strength.

Id.

Applicants' early analyses,or penetrations showed that. local yielding.ogcur,s at i

j 20.3 psig for the main steam penetration.

Applicants' Ex. 8-4 i

f at Toble 6As Alley, Tr. 3359.

We thus cannot find that the j

rinite element techniques used by Applicants are qualirted for 4

]

the stress conges considered, s

i In conclusion, we do not oecept Appisconts' onolysis or i

containment Vessel capacity or their finding that the PNPP i

l containment vessel con withstand on internal Pressure of 50 psia.

We are not convinced that their analyses demonstrate structural integrity with margin at this pressure, let alone l

that they properly evoluote containment lookoge.

(We do not i

l

[

consider Appliconts' assertions that their service level D l

l calculations establish margin.

Applicants' Testimony at 28.

i This onotysis used octual material strengths, osain without margins, for the equipment hatch and personnel airlock and a j

plastic onalysis for P205.

Applicants' Ex. 8-4 ot 22.)

Until the uncertointies oddressed above are resolved, we i

t

._____=_..________..__._-..._~.__m, e Z9 -

must consider the PNPP containment copocity to be indeterminate.

We note that the ability of the containment to Withstand its 1

design basis pressure hos not yet been established.

Finding 36.

l This too must be rectified before we con opprove operation i

coove 57. Power.

4 1

(ti) Negottve Internal Pressure i

\\

Appliconts did not evoluote the negative pressure copocity

[

l 1

of the containment, but relied upon o conservative FSAR onolysts Of vacuum breaker oction for design basis events to th0W thok i

L i

i design negative pressure is not exceeded.

Finding 34.

We uould i

find this acceptoble, but for the uncertanty as to the ability 1

8 of the vacuum breaker to survive hydrogen burn pressures, as l

4 j

dtscussed below, i

(b) Drywell Capacity 1

Applicants did not perform on onolysis of the PNPP drywell, l

but relied on on analysis or Grand Gulf.

Finding 41.

The Grand

(

9 i

i j

Gulf'onolysis did no t consider",the ef f ects of ' voidt **in tHe 1

t' drywell walls voldt have been found in the PNPP drywell wall.

I Finding 40.

Although the voids have been repaired, and -

I Appliconts claim that they are reasonobly sure that there are no l

l ceners sn pioces not inspected for them (Tr. 3417), we connot i

occept this without further scrutiny.

We connot be sure that i

Applicants' evoluotion of the voids determined the cause of the i

t problem, which opparently narrowed the inspection process.

We are otso concerned that there may not be effective I

- nondestructive i

examination methods for detecting voids in the drywell woll, f,

Alley. Tr. 3416.

t I

w

{

I i

--,-w,------,,

-w~

n,.

,---,r-n,,.-,,.,-,-e.,--,

-,m.

- ---, --,,-., - - ----,, n ~-,,.

The Grand Gulf analysis is not ovoilable for our scrutiny or for that of other parties.

The Sandia Nationci Laboratory. in its revtew of the Grand Gulf igniter system, did not advocate taking GGNS calculations at face value, but recommended obtaining independent estimates.

OCRE Ex. 21 at 10.

Given the importance of drywell integrity (Finding 43). We cannoe accepe the GCNS onolysis on faith either.

We find Applicants' opproach on this matter to be unocceptable.

4. Containment ResponseAna1Ysis We examine herein the odequacy of the contoinment response code used by Applicants. CLASIX 3 (Finding 45), and assumptions and models used therein.

We Will concentroke first on the input porometers assumed for the scenarios analy:ed, then on the code seselr. and fino11y on variotions or these scenarios which mo, represent o more severe challenge to the containment than was analy:ed.

our ossessment relies heavily On the review of the Grand Gulf igniter system performed by the Sandia Nationoi Laboratory contained in NUREG/CR-0530, OCRE Ex. 21.

Finding 48.

We rind this document to be the most comprehensive and scholarly evoluotion of the igniter system in our record.

At the hearing Appliants and S t o f, f tried to, discredit this document by,clonmang that it was out of date and that the HECTR computer model therein was crude and did not model o Mark III containment.

Notofrancesco. Tr. 3724.

We reject this reasoning for several O

reasons.

=1)

Not the least of our reosons is the fact,that Mr.

Notorrancesco is obvtously not a combustion expert (Tr. 3672).

. = _ _.....

. r i

t We do not doubt that Sandia hos been Perinsng its HECTR 4

models: indeed, this as to be expected in any rield or sesentific endeavor.

Our record does not indicoke that these later versions predice lower pressures and temperocures than i

that in OCRE Ex. 01: in roct, o loter version or what Mr.

Notorrancesco termed o realistic representation of the Hork III t

containment (Tr. 3738) produced higher peak pressures than in i

OCRE Ex. 01.

Tr. 3741-40.

In any event, when decling wsth those fields or science that are ropidly changing, o line must be drown somewherei we cannot wait for the rinal version or f

HECTR berore we render our decision.

Appisconts and starr have j

both urged that that line i

ce drown at Grand Gulfs i.e.,

nothing more should be expected or Applicants than was requared of Grand Cult for operation above il Ex. 8-2 and 8-3.

$% power.

Notorroncesco, Tr. 3743s Applicants

• Since OCRE Ex. 21 (and no subsequent HECTR calculations) was 4

relied upon by the Starr in opproving power operation for Grand 1

Guit (Notorrancesco, Tr. 3744-46), we reel that it is oppropriate to rely on the document.for our decision.

Mr. Notorrancesco's arguments that HECTR overestimates flame l

]

speeds and comoustion completeness (Tr. 3737) are also without merit.

These are input parameters which were set to be identical to CLASIX 3 ossumptions in some or runs in OCRE Ex.

21.

Findings 80-64.

While other HECTR cases we have examined e

i use higher values for riome speed one combustion completeness, j

ehey are orrset by use or o smaller hydrogen source term and l

1ess conservative propagation porometers.

Findings 81, 85, 86, i

4

87.

Mr. Notorrancesco's comments abdue HECTR loc. king o.drfwell (Tr. 3738, 3742) are oiso unpersuosive.

Sondia evoluoted this enrough MARCH sensstivity studies, and round that presence or the drywell might lower peak pressures by less than 10 psi.

OCRE Ex. 21 or 148.

(We believe this is oiso verified by HECTR cose A-1, which models CLASIX 3 cose 5,

which is o one-compartment contosnment model.

Peak pressures from nECTR were only about 10 psi greater than CLASIX 3:

this suggestb that the low CLASIX 3 pressures result from other phenomeno, such as comportmentali:otson, rather than the presence or the drywell.

Findtngs 64, 88.)

Applicones rsnolly tried to use the orridovst or Dr.

Norshall Berman or Sandia riled in response to OCRE's motion ec compel his appeorance at the hearing, discussed above, to object to the use of OCRE Ex. 21.

Tr. 3725, 3745.

Dr. Berman, due to 5 torr and Applicant opposition to OCRE*s motion, was not available at the hearsng for Cross-examsnation.

We reel it is unrose and inconssstent for Applicants to object to Dr. Berman's GPPeorance at the hearing and then try to use his statements an on arridovit (which as not port or our hearing record) to cuerress these positson, As Judges with considerable technical expertsse, we are copoble or evoluottng the document ourselves.

We have done so, Al and find the document compe te n t,

sondio recogni:ed that none or jk/We rand this document to be for more credible than ene eestsmony or Dr. Fuls, who closmed romiliarity wsth the NARCH

ode (Tr. 3:55) but did not recogni=e that a itsting or inpue values was not for t1 ARCH but for CONTEMPT-LT26, and was aboue to cry to corretore the CONTEMPT listing with the MARCH manual (Tr.

3543

• 6),

and that of Mr. Notorrancesco, Who, as noted above, is not a combustion expert and Whose testimony We round too protective Or Applicants.

l

• 1 t

the ovoilable containment response models are sophisticated i

enough to predict burn pressures and temperceures with high t

occuracy.

OCRE Ex. 21 at 10.

Our purpose in relying upon this

' document is not to consider the HECTR results os obsolutes, but l

rather to compare the erreces or dirrerent input and modelans l

)

ossumptions with CLASIX 3.

We rind the qualitative insights i

I gained therefrom to be essential to our decision.

(For ease or i

J comparison, we have compiled in Appendix C to this decision the I

major for inpue.poromete,rs,and results, in uni (orm units or measure,,

the'CLASIX 3 ond HECTR runs for PHPP (from Applicants' Ex. 8-1) 9 j

and GGN5 (from CCRE Ex. 01).)

l.

The first input parameter we Will examine is ignition limit, i

l i

The PHPP CLASIX 3 onalysis used on ignition limit or 8%

1 hydrogen by Volume.

Finding 51.

Applicants claim that this is conservative.

Applicants' Testimony at 46.

However, we notw i

that the'the flampobility limits for hydrogen are geometry-j dependent.

Id.

8% hydrogen concentration corresponds to the downWord Propogotion limite olthough Dr. LGWis.later stated that thss limit is 8.5-10%.

Finding 60.

The lotter figures compare i

Well with that used by Sandio (9%).

OCRE Ex. 01 at 15.

i Sandia round that the placement of the igniter assemblies at 3

Grand Gulf close to ceilings (and the spray shield of the i

housing) would inhibit combustion errectiveness.

Finding 60.

4 I

Sondio rett that tenition would reliably occur at the downded propagation limit. but not at lower concentrations, Id.

Thts hos been Verified by hydrogen combustion tests at the HeVodo 1

Test Site.

Finding 61.

{

,mr-,.e%.__.----r--

-.,,,e

,,.m y

.m-

-,--9

---

s-,

r,-

,m-Sme.-,,-my.,

-- rwe

-,,,,.--,-~,-++,-w,.

, _ wy a,

~._.

4.

(We must reject Dr. Lewis' statements that the test which railed to ignite was due to incomplete mixins (Tr. 3520), and that proximity of the igniter ossembly to ceilings has no errect on unsuppop'ked by the NTS dato.

I ses errectiv'eness (Tr. 3523-14) as f'

It is obvious that Dr. Lewis has not keph'Lp with the latest

_LI research in this rield.

Tr. 3517-18, 3522, 3627.

We also rejece his ossertions that the Fenwoi tests proved the reliability of the ignition system.

Tr. 3627.

These eeses used centrol' ignition in a 3-foot diameter vessel (Korlovit. Tr.

3639, 3649), which we do not consider appli'coble to conditions IIIfcontatnment.)

in the Hork Sandia reported that, at Grand Gulf, all but a few igniters ore located within 2 feet or ceiling structures.

OCRE Ex. 21 or 195.

We note that the p1ocement of igniters at Perry is ssmilar to that'ot Grond Guir,' Applicants' Testimony"ot 34." OCRE Exi 16 corroborates this.

(We realite that'much discussion was held e

I at the hearing on whether this document Could be relied uPon to ossess igniter locations.

Although at hos been superceded by the list or igniter locations in Applicants' Ex. 8-1 (Tr. 3504),

the document provsdes valuable descrsprions or-the locations or

' igniters and or structures in close proximity, os.the Starr round important (OCRE Ex. 19 at 2).

Applicants' uitness Bu::elli rirst said that on teniter-oy-

.L/ Thss ss also apparent from the transcript or the McGuire OL proceeding, in which Dr. Lewss testified.

We take ortsetal notice-or'the rollowing portions or thoe transcript:

T r.- 3295-96 (February 26, 1981, where Dr. Lewis insisted that detonations j

ore imposstole in hydrogen concentrations less than 10%,

r contrary to the statements of other experts): Tr. 5055 (Horch

' 19, 1981,'where Dr. Lewis odmitted he was not familiar with the work or. Dr. John.H.5. Lee until leoring of it from other wsenesses of the hearin9)I Tr. 5062 (Horch 19, 1985, where Dr.

Lewis was not-romilior with the results or reses or glow plug tentters-conducted at Lawrence Livermore'Notionoi Laboratory).

f.

1 teniter-comparison could be perr'ormed to determine whether the exhibit locotions were representative or the current locottons.

l Tr. 3506.

Then, persons who were not witnesses inrcrmed Applicants' Counsel that locations with the some elevation.

octmuth..ond radi31 dimension may not be i dentical.

Tr. 3510.

Ms. Bu::elli subsequently, on redirect, changed her testimony to match counsel's sectements.

Tr. 3607-08.

Loter. she admitted

.that specifying the 3 coordinates identified a precise location.

i Tr. 3659.-

We ogree with the lotter poznt, and believe the document is occurate enough for our purposes.)

l Comparing OCRE Ex. 16 to Applicones' Ex. 8-1 (Toble 2.4-1).

L we find that the following i gniters have identical locottons ona L

ore located close to the i dentified structures:

1M56-054 through -060. 1M56-098. -100, and -101 (room ceiling):

1M56-067 through -075, 1M56-077 and -078 (underside or potor f

crane support ring)s 1H56-080 through -090 (containment vessel dome).-

The ro11owing igniters do not have identical locations but are very close to the identified structures:

t 1M56-0076 (some elevation and radial dimension, but 6 degrees f-dirrerence 0: 1muth01, underside or crane support ring)s 1M56-016 through -020 (all located within 1 root or location identified i

as i nside race or drywell top slob): 1M56-0:4 through -030 (011 i

located about i root below the 620' elevation, the HCU ricor i

i (see Finding 62))i-1H56-032 a.nd -033 (within a roo.t o( t h, e.,

underside or rioor slob for refueling pool): 1H56-040 (within o l

1 root or steone tunnel coaling): 1H56-040 through -045. -093 i

y e

,=-e

+--y-

+-9

-.v.syy y--wt-w e -yww am 4 w

,m-r*

y r-wg---v

---vm----ww

--e w

r----*egw-w

-*---r--*-~~-wem-e,m*

vv-ve------

J

/

~

through -096, and -099 (wsthin a rooe or room ceslings)s IM56-j 091 (within 2 f t.

cr room c'e i l t n g ).

i Tnus, we fina from tne information available choe 41 or the 100 PNPP igniters are ledated in closo-proxtmity to structures which could annibst upward flame propogotion.

(It is not possible to cetermine this for tne remosning igniters.)-

Sondte 4

concluded that~ such on oceangement w s.1,1 reliably tantre hydrogen I

in the 10% rong6. but; noe at lower concentrattons.

OCRE Ex. 01 or 195: Findings 59, 40.

We therefore believe thoe on j

oppropriate, conservative ignition limit for.contoanment response analyses-is ce the downward propogotion limit conge or 9.5'to 10%.

Finding 60.

This is somewhat greater than the 4

i value used in CLASIX 3.

.Finctng it.

The secona input porometer we exonine is propogoeton limies, i

Tnis refers to the concentration or hydrogen needed'ior flame r

propogot' ton into con odjocent comparement.

OCRE Ex. 21.ot 94.

The CLASIx 3 moddl ossumes a propogotton Itmit-or s% hydrogen.

Finatng 51.

Sandio consteered a more realistic input for n.

.a propogotion limits to be the observed geomeEEy-dependent 11msts (some os the ignition limits mentioned above).

Finding 87.

The

[

CLASIX 3 value should be conservoetve.

=Ihe third' input parameter is combustion Completeness.

As as expected, complete combustion results in higner pressures, as more nydrogen.is consumed..

Findings 79, 89.

CLASIX.3 uses a contustion completeness er 95,0.

Finding 51.

However, Nevodo

(

Test Site results snow. complete combustion ror byerogen concentrations above-7.7%.

Finding 50.

We therercre consider ene CLASIX~3 value nonconservative.

l

-,,-_.m._

, ~., _.

2 2.1--

._u-

--+--La--.-.

.,-wn.4m.-.~-.+--.+---a i

.A---.-

,.-.4 A

I 37-The fourth parameter we consider is flame speed.

Increasing the flame Speed will increase peak burn pressures, OE there is lets time ror ' heat 1~os s dur t ng ' the' burn. " Finding' 53.-

CLASIX~3

~

l uses o flame speed of 6 feet /second.

Id.

Based on their experiments in the VGES tank, the Sandio researchers considered this volut to be low by a factor of three or more.

Id.s OCRE Ex. 21 at 15.

AlthoJ9h Stoff witness Notorrancesco claimed that

~

the flame speed based on VGES data was too high (Tr. 3737), we l

believe that on important factor offecting flame speed has been neglected - the effect of ioni:ing radiation.

The action of ionizing radiation will increase flame speed.

Finding 54.

It even oppears that ioni:ing radiation, through the creation of radicals, con initiate o detonation, although ehe levels of. radiation in the contoinment ore too low to achieve this.

Lewis, Tr. 3526-27, 3615.

(Detonations result in supersonte flome speeds.

OCRE Ex. 01 or 187.)

Sandio also believed that flame acceleration (to speeds greater than cetermined in VGES) is likely.

OCRE Ex. 21 at 15, 17, 194.

Given these uncertatnties, we are convinced that a higher flame speed than that used in CLASIX 3 is necessary.

The fiftn porometer ss contoanment spray operation.

Coneoinment sprays are on important heat transfer mechanism that results in a significant reduction or pressure and temperature during a hydrogen burn.

Finding 56: OCRE Ex. 21 or 17.

CLASIX 4

3 assumes thot the containment sprays are cutomatically actuated e

ofter the first burn.

Finding 55.

i The PNPP containment sprays are o subsystem of the RHR j

system, which also functions'to cool the core.

Finding 57.

The I

1

sproys share common pumps, piping, and electrical systems with on ECCS subsystem.

Id.

S'ance o degraded core occident is premised on the assumption or o loss or core cooling, it may not ce conservative eo ossume sproys are avoilable.

Id.: OCRE Ex.

19 ot 4.

The Storr hos stated that the sprays should not be conssdered if their availability is questionable.

Findins 59.

Although Applicants stated that it may not be inconssstent to assume spray, operation in o degraded core occadent ir o-Volve railure as postulated (Tr. 3445), we believe that more than a volve 4

roilure is necessary to cause o degraded core occident, e rind spray availability to be questionable, and agree with the Starr that sprays should not be considered in Applicants' containment response analysis.

CLASIX 3 is nonconservative in this respect.

The final input parameter we will consider is the Wetwell spray carryover fraction.

This or course assumes spray availability.

CLA51X 3 sensitivty studies have shown that decreasing the wetWell sproy carryover fraction Will increase Wetwell temperature.

Findins 63.

(The weewell refers to the lower compartment in the containment and is bounded by the suppression pool and the HCU ricor.

Findin9 6C.)

The PNPP CLASIX 3 model assumes a werwell spray carryover rroction or 0.

9.

Findins 63.

The containment spray headers are located for above the w e t w e l l.'

Finding 65.

.The cross-sectional riow areas below the refueling-rloor in the Perry Mark III containment are quite low.

Findins 65.

The reactions or total contoinment cross-sectional dreo at the various elevations are (assuming total cross-

~_

sectional area to be 11,310 square feet, usin9 on inside contoinment radius of 60 feet, Applicants' Ex. 8-4 at Fi9ure 1).

689'-6

• eleva tion, 2778/11310 =.246: 664'-7' elevation,.224:

642'-0* elevation..271; 620'-6' elevoeion,

.168.

Thus, assumsns spray droplets are not impeded by structures and components at elevations above the HCU floor, the maximum amount of direct spray impin9ement that could reach the wetwell is l

obout 17%.

The remainder of the assumed carryover fraction is composed of sheet flow, which is assumed to have holf the heat transfer effectiveness os droplet flow.

Findin9 64.

However, this assumption has no experimental bosis.

Id.

We believe that it is improper to take so much credit for on unverified I

assumption.

'This toe is^noncenserv,ative.----

l Of oil the CLASIX 3 input parameters we have examined, we have found 011 but the propo9ation limits to be nonconservative.

We now turn to the odequacy of the CLASIX 3 code itself.

We fand that the oppropriore stondord by which to evolunte o computer model is that used by Applicants.

This standard requires that the Code either be available in the public domain j

one have o history of proven use or that it be validated by comparison wi'th results of other accepted codes or experimental i

dato.

Findin9 44.

Since CLASIX 3 is o proprietary code (Findins 45), the first test is inopplicable.

Applicants claim that CLASIX 3 has been validated by comparison with results of hydrogen combustion tests at Fenwol and the Nevada Test Site.

Finding 46.

CLASIX 3 9 ave conservative predictions for all but one test.

Id.

However, we l

co not believe that these tests are similar e'nough to the Hork III containment to consider the code validated.

'The Fenwol

eeses were conoucee: in a scoll (3-rece cionerer sossel).

Ic.

The Nevoco Test stre vessel contatnee ne comportsenes.

Id.

We do not rino it surpristng that CLASIx 3 con honele such sample configurcetons, as ene cause er its prediction or icw turn pressures as apparently its compartmeneoli:ocion.

Fineing SS.

We note that the CLA5IX 3 cose 5 reported in OCRE Ex. 21 save pressures only occut 10 psi less than the HECTR preetceton (one obout three times higher than those predicted by the usual CLA5IX 3 conrtsuroeion).

Finding S4.

This case used a one-compartment containment configuration.

Id.-

This indicates that avalidotion* with suen staple test configurocions as not ancaccetve of proper come performance for the nark III conectnnent.

Comportson with ceher oecepeee coces as the lose oleernoetve.

The appropriate ecces for comparison are HECTR one CONTEMPT.

Findings 49, 50, 73.

The NRC Stoff concurtee comporisons or CLASIX-3 with CONTEMPT-LTOS and round that CLASIX 3 uncerpredices cesospheric temperocures because se overpreciets neat transfer.

Fineang 73.

The S tarr rounc thts to ce noncenservoeive one seceed thor CLASIX 3 snouls not be usec.

Io.

(Applicones

.n enetr fanssngs urgec ence this assue te cererred to the final analysis.

We celseve, for reasons discussed melow, that it is oppropriote to consteer at now.)

Comparisons witn HECTR show that CLA51X 3 srecely i

unserprecters peak burn pressures.

Fincins 90.

HECTR cose 5-1 l

mocelec CLASIX 3 case 1.

Findina 81.

These cases useo a higher i

preburn pressure, os wcule occur sn the DUE occieent from ene i

1

Ql-Gdded dryWell oir mass.

Findins 79.

This was the~only HECTR come using the some hydrogen source term os CLASIX 3s all other HECTR runs assumed less nydrogen was released.

Finding 81.

19nition and propogotion limits were 10%, with 100% complete combustion, 6 ft/sec flame speed, and automatic sproy action.

CLASIX 3 Produced a peak pressure of 11.1 psisi HECTR produced a peak of 67.6 Psi 9.

Id.

CLASIX 3 cose 2 ond HECTR cases B-2 ond B-O' used input parameters similar to those in cose 1, except the HECTR hydrogen source term is less and the preburn pressure is not elevated.

Finding 90.

CLASIX 3 predicted a peak pressure of 7.4 psag, While HECTR predicted peaks of 56.5 (B-2) and 59.4 (B-2') psi 9.

Id.

CLASIX 3 cose 3 and HECTR cose B-3 assume the some input porometers as were used sn the PilPP onalysis.

Finding 82.

CLASIX 3 produced a peak pressure or 6.7 psi 9s HECTR's Peak Was 41.6 psi 9.

Id.

CLASIX 3 cose 6 and HECTR cose B-4 used the I

some assumptions, except the sprays were assumed off.

Finding 83.

CLASIX 3 produced a peak of 10.3 P5193 HECTR, 53.4 psi 9.

Id.

Other HECTR runs were conducted with input parameters i

j

' considered.more reoisstic by.Sandid,-i.e.,

rlomerspsed.ond

~

propo9ation limits.

Findins 85.

These runs (with the exception of-case B-6, the significance of which is discussed below) 9 ave l

peak pressures ranging from 57-to.69 psis.

Id.

The cause or CLASIX 3's predictioh of low burn Pressures is-its'compartmentolitotton model and. heat tronster assumptions.

Finding 88.

Wetwell burns will give small p'ressure excursions, l

I

'/2.-

While burns in the containment compartment cause hi9her pressures.

Id.

CLASIX 3 overpredicts oxygen transport back into the wetwell, thus ovosaing wetwell inerting.

Id.

In reality, hydrogen would also be pushed into the contain.nent compartment.

OCRE Ex. 21 at 93: Notofrancesco, Tr. 3749.

I.e.,

intercompartmental mixing is more rapid than assumed.

OCRE Ex.

21 at 93.

Better mixing makes global burns, with their ottendant hash peak pressures, more likely, Id. ot 17, 93.

We must conclude that the CLASIX 3 code is unocceptable, as it does not compare favorably with the accepted codes HECTR and CONTEMPT.

Indeed, comparison with these codes hos identified serious nonconservottsms in the CLASIX 3 approach.

CLASIX 3 fails to predict more than minimal pressure rises even assuming conservative input porometers.

We therefore Concur with the Stoff (0CRE Ex. 20) that CLASIX 3 should not be used.

We finally turn our ottention to conditions which could present a more severe threat to containment integrity than was previously examined.

Combustion in the drywell for the DWB occtdent only occurred when the transient was extended beyond the end of the hydrogen release.

Findin9 75.

More severe comoustion might occur if core reflood was assumed at on earlier time.

Findin9 77.

E.g.,

recovery at 5500 seconds into the transient would condense the large steam fraction in the drywell, raising the concentrations of hydrogen and oxygen.

Id.

Action of the drywell voeuum breakers would admit more oxygen and limit the degree of depressuricorion.

Id.

A simplifted l

l calculotion of-this event showed that pressures in the ronse of 50-55 psi could be t

I

{

t

attained, assuming preburn atmospherte pressure.

Id.

This is more severe than previously calculated for the DUB scenario (see Fsndang 47) and exceeds dessen basis p.r e s s u r e s.

We believe that the effects of early recovery on drywell combustion should be investigated using a reliable onolytical model, and.special attention should be paid to the resultant Wetwell hydrodynamic loads from suppression pool swell (05RE Ex. 19 at 5).

We are also concerned thot the base cases may have been terminated too early.

The 50RV transient was terminated wnen the hydrogen release wene to zero.

Finding, 74.

At the end of the transient, wetwell hydrogen concentration Was 26%i oxy 9en, 3%i seeon, 23%.

Id.

In the containment, there is 5% hydrogen, 2% oxygen, and steam is 31%.

Id.

(Similar conditions extse at ene ene of ene DWB eronsient, terminated ofter the drywell burn.

Finding 75.)

Eventually, the steam will condense, raising the nyorogen and cxygen concentrations.

OCRE Ex. 21 at 29.

The contosnmene vacuum breakers (not modeled by CLASIX 3: Finding open to edualize pressure and will admit oxygen, like

76) Will the drywell vacuum breakers in the cose Just discussed.

Finding t

34.

The effect of terminating these transients early may be significant.

HECTR cases B-6 and B-6' illustrate this.

Both i

cases used identical input porometers, except that cose B-6 Was terminated at the end of the hydrogen release, as was the CLASIX 3 SORV case.

Finding 86.

Case B-6' was allowed to contsnue l

post enot time.

Id.

Case B-6 produced a peak pressure of 21.6 psig, similar to the results obtained from CLASIX 3.

Id.

Case B-6*

produced a peak pressure of 69 psig.

Id.

The early termination of these transients may be another factor l

---.~

-4f y-.

contributing to CLASIX 3 nonconservatism.

~

~

l Sandio stoeed that a more severe threat to containment integrity Qou18 be posed 69,3ydcogen releases

• entering th,e_

contotnment or wetwell without rirst troversing the suppression 4

pool.

OCRE Ex. 21 at 197-198.

The PNPP onolyses assume that hydrogen flows through the pool and enters at the bottom of the wetwell for both 50RV and DWB cases.

Finding 66.

Drywell 4

byposs leakoge is not modeled in CLASIX 3.

Finding 68.

The maximum drywell leakage allowed by technical specifications is I

over 5 times the rated capacity of the drywell purge compressors.

Findin9s 69, 70: Richordson, Tr. 3498.

In some conoitions; drywell leokoge may be great enough to prevent flow through the suppression pool.

Finding 72.

Applicants admit that leakage of hydrogen through the drywell wall leak paths could occur.

Finding 71.

l They assumed that 14-19% of the total hydrogen could bypass the Pool.

Id.

They also claim that this effect would not offect I

the operation of the hydrogen control system or their analytical conclusions (Tr. 3500,.3649), but-no conrirming calculations have been performed.

Finding 71.

+

'We are'not convinced that the omo'unt of leakage Will be-

~

limited-to 14-19%,.os neither we nor the other ~ parties hove been

~

able to review the GE calculations.

We also.believe knot leokoge of hydrogen to the contoinment comportment will citer the mining and concentration ossumptsons of the containment response'onalysts, making global burns more likely.

The effects

-- of drywell leokoge on containment response to o degroceo core l

occicent must be-thoroughly ~ cnd formo11y evoluoted, -os the storf

.. ~

~

2-..%.

___.,,1 e

q5-nos required.

OCRE Ex. 19 de Si OCRE Ex. 20 or 2.

5.

Equipment survivabil ty Applicants' evoluotion of equipmene survivobslity is based on containment pressure and temperature profiles determined by use of the CLASIX 3 code.

Finding 90.

Since we found above that the CLASIX 3 ccee as inadequate and proeuces unreolistically low pressures and temperatures, we must find any analysis of equipment survivf.bility. relying on it,to be.likewsse ceficient.

We also have problems with the opproach used by Applicants, who woulo have us sumcorily offtem their analysis because it is cound by Grand Gulf's, which was accepted by the Stoff but is not part of our record.

Id.

We have no uoy to confirm that their heat transfer code, HEATING-6, is conservative, or that oppropriote modeling ossumptions have been used.

We must conclude that equipment survivability during and ofter hydrogen deflagrations has not been proven.

(Though not necessary for our decision, we would state that the apparent survival of electrical cables at the Nevado Test Site, despite burning and cracking of ccble insulation (Tr.

3716-17) may not be indacotive of survival in on octual opplication.

We have no information en whether the cables were aged to on end-of-life condition, as' required in equipment qualification. tests by 10 CFR 50. 49 (e) (5).

Equipment qualification one survivability are considerec by the Comm2ssion ce be essentially the scoe, and~10 CFR 50.49 is referenced my the rule.

50 FR 3501.)

Ue ore Concerned chat certain components have Very low

(

qualification pressures.

Finding ?3.

Applicants' justificotson

[.

for accepting these is apparently.not confirmed by any testing.

Id.

We find the use of enese components, not qualified to the hzgher pressures we believe wil occur from hydrogen burnzng, to I

be unocceptable, especially since the vacuum breakers are reized upon in determining containment negative pressure copocity.

4 Finding 34.

i -

We believe that a showing that essential equipment con survive detonations is unnecessory, os detonations are unlikely I~

to occur, even if detonable concentrotions were to form, Finding 94.

Diffusion flames pose another' form of thermal threo't to j-equipment. " Tests

• conducted bflthe Hydrogen control Ownerr Group i

t'HCOG*) in a 1/20 scale facility showed that diffusion flames exist when.the hydrogen release rotes exceed 0.4-0.5 lb/sec.

Finding 95.

Preliminary results from'these tests indicated that I

diffusion flames resulting from o 75% metol'-water reaction will result in unocceptable thermal loading to equipment.

Id.

The 1/4 scale t'ests planned by HCOG will deftne enermal.

envtronments resulting from difrusion flames.

However, HCOG will not consider diffusion-flames resulting a release history corresponding to a 75% metal-water reaction.- Finding 96.

The reason for this is that the hydrogen release rates needed for i

diffusion' flames are supposedly'not realistic for:0 75%.metol-Y water reaction..

Richardson, Tr. 3568, 3622-23.

Although we recognize thot-the 1/4 scale tests will be odoressed in the final analysise we connot defer this issue.

10 l

CFR 50. 44 (c) (3) (v) (B) requires-that evoluotions of equipment i

survivobility consider a 75% me tal-wa ter-reaction. - Applicants have announced.their defiance of ~ this regulation; we connot

-y3-ignore such a blatant violation of the Commission's regulations.

We are not convinced by Applicants' orguments about the hydrogen' release rates leading to diffusion flames being unrealistic for a 75% metal-water reoceson (as o sustained hydrogen release rote at that level would supposedly lead to on unrecoverable degraded core occident).

Richardson, Tr. 3602-03.

We believe that this logic challenges the Commission' rule.

The Commission specifically states that the 75% metal-water reaction is o limiting cose degraded core occident, and that oxiwotion beyond that point will be unrecoverable.

50 FR 3499.

In the rulemoking process the Commission rejected Comments which sought a reduction in the extent of metal Water reaction to be considered.

Id.

wa also rind the HCOG oPProcch illogical.

If the 1/00 scale tests indicated that equipment survival is a problem for diffusior., flames from o 75% metal-water.reoc, tion 3,then the logical opproach is to test these conditions in the 1/4 scale i

facility.

Rother'thon resolving the problem. HCOG oppears to be nicing from it.

We therefore rule that Applicants' opproach to this issue is I

illegal.

Applicants mubt commit to evoluoting the thermal environment, and equipment response thereto, resulting from diffusion flames using a hydrogen release history resulting from o 75% metal-woter reaction as a condition of licensing.

l c.

Other Effects of System Operation In this section we consider whether other effects of i

nydrogen control system operation will ogsrovate the course of l

on acc2 dent.

50 FR-3500.

~

I.

(o) DryWell Pool Loads sensitivsty studses ce Grand Gulf using the CLASIX 3 code have predicted violent overflow of the suppression pool into the drywell for some cases and assumptions.

Finding 97.

Sandia likeWise predteted pool surge into the drywell.

Id.

'I e is not clear from Mr. Richardson's evasive responses (Tr. 3485-96) whether the effects of these loods on equipment and components Within the dryWell have been evoluoted.

In any event, there is no evidence in our record for Perry on these loods (os calculated from on oppropriate containment response code) and their effects.

This problem must be evoluoted.

See OCRE Ex. 19 ot 5.

(b) Decoy Heot Removol The design basis for the Mark III containment assumes that cecoy heat is added to the-suppression-pool and is transfered to the ultimate heat sink by the suppression POOL Cooling mode of the Residual Heat Removal ('RHR*) system.

Applicants' Ex. B-1 ot 25-26.

.In a degraded core occident, hydrogen combustion adds heat to the containment otmosphere, in addition to the decoy heoc addition to the pool.

It is thus appropriate to examine

• ^ '

• ascoy heot"remo7ol d opob i } i't y *i fe ' thi s s i tuo t'i~oW; -~

The evidence in our record indicates that long term decoy heat removol is essential to maintoi,0 the core in a sofe conoition, and tnot operation of'o loop of the.RHR system in the 9 0'o 1 coolsng mode is necessory to ensure decay heat removal.

Finoing 98.

When containment oemosphere pressures exceed 9 psig, l

containment sprays, a subsystem of the RHR system, automoeicolly occuote.

Finding 99.

If pressure remains onove 9 psis, the l

_=

-V9 -

sproyk remoir. OperGtangi sproys take precedence over other RHR i

modes.

Id.

Operator action is necessory to align the systen to other modes.

Id.

Operation of both spray trains uill degroae pool cooling and l

mixtng.

Finding 100.

Apparently no calculoeions of pool r

temperature response With both RHR loops in the spray mode have be*n performed, at least not on our record.

Holecciow. Tr.

1 3470.

Although Applicones later claimed tnot the pool j

teroperoture peaks about 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> ofter o design bosts occident (uhile hydrogen burntng, os analy ed by Applicants, occurs within 1-0 hours) (Richardson, Tr. 3611). this may not be true J

for other-scenarios not considered by Applicants; nor is it clear that that the design basis colculotzon did not take credit for RHR pool cooling.

We find that the eff.ects of hydrogen combustion on cecoy heot removal capability must be evolunted.

Tho effects Predseted by Mr. John Humphrey, o former GE contoinment systems i

engineer, should receive sFecial. consideration.

(See Tr. 3479-85.).

The effects or pcol temperature on makeup systems, including the potential for pump covitation, Gust be evoluoted for.Pf4PP,- considering any loco 11y elevated pool temperature due co.the proximity of. kne succion strainers to.Sofety-relief bi disenorge quenchers.

(See OCRE Ex. 15; Richardson, Tr. 3469.)

- b Although Applicones claimed that calculations hove been performed which show that pump cavitation is not o problem for high pool temperatures (Tr. 3606), these calculations are not ovoilable for.our scruttny (or trot.cf tre parties, Tr. 3659-59), nor is it apparent that the effects of taking suction from l

. the Vicinity of'the safety-relief Vo1Ve quenchers Was considered in that calculotton.

We Connot 'occept on faith analyses Which we have not seen.

l l

t

,v..-

--.7

,,n+-,,,-

e

- 5Fo -

1 (c) Secondary Fires Applicants,have not e v o l p o t, e d.,; t h e p o t e, n t i, a l,, f o r sgg o n d a r y fires (burning of combustible materici in the containment, sgnited by hydrogen burning), although Applicants claimed that on analysis'for Grand Gulf showed no potential for seco'ndary r ires. - Richordson. Tr. 3580-81.

Thest logic is that ssnce the CLASIX 3 containment response results for PNPP are bounded by GGNS, secondary fires won't occur at Perry either.

Id.

We must reject this logic, os it relies upon the faulty 4

CLASIX 3 code.

Furthermore, we connot accept on faith analyses performed for Grand Gulf Which are not part of our record.

The evsdence in our record indicates that secondary fires are likely to occur, and will ofrect the containment pressure and temperature profiles.

Finding 101.

7.

Scope of the Preliminary Analysis As discussed previously, the new hydrogen rule permits the final onolysis of the hydrogen control system to be deferred.if i

an acceptable preliminary analysis is submitted.

But, there is no guidance os what constitutes on acceptoble preliminory analysts.

Applicants in their proposed findings urge us to interpret-the rule such that the Preliminary onolysis i n c l u d e s-4 only-the general requirements of 10 CFR 50. 44 (c) (3) (iv) ( A) and not the.more substantive requirements of sections (iv) (B),

(V),

and (Vi).. Claiming that the longuage of section (vii) indicates that.the Commission meant to exclude these sections from the 4

prelsminary analysis.

We connot agree.

The Federal Register notsce indicates khot the longuage of section (vii) underwent last-c.-

y

.m

-,.L

_y,.....m

.r a7 -

g.

w.w-

51 -

minute changes.

50 FR 3502.

We therefore connot infer from the wording of that the section what the Commission may have had in mind, as it is likely that its longuage was not chosen with

core, We believe thoc the Commission intended a Preliminary showing to be moca on all focets of the rules this is consistent

' wi th Qhat 'Qos ~requi~ red 'f br; ice. c'dndenser..~#1oM ts. S e qu o 7 thland ' -

e McGuire, Where equipment survivability Was demonstrated (50 FR 3501): the analyses for these Plants are apparently equivalent to the preliminary analysis under Sec tion '(Vii) of the rule, Id,

• x Applicants and Staff agreed on the scope of the. preliminary analysis before Applicants submitted the document.

Finding 102.

Not surprisingly, both Applicants and Staff objected to many of the matters raised by OCRE os beyond what they considered the Proper scope of the preliminary analysis, We find guidance for resolving this issue in section i

(vii) (D) of the rule.

This section lists the factors the Storf is to consider in setting the final schedule for compliance with the rule.

These factors are the skotus of efforts.to comply with the requirements, the smpact of the requirements on other plant mooirications, and the Commission's goal of compliance without undue delay.

Thus, the standard is that of I

i reasonableness, We find this to be the correct standard for our

purposes, It is obviously unreasonable to require Applicants to complete their final analysis (now scheduled for completion in I

mid-1986,. Finding 104), including results of experimental i

.-._=

_ t

~

i t

progross'new in pregress Ishe ene 1/4 sesle tests. cefere I

cperctt03 ctcve 5% power.

It is also unreasenc=le f:r 5t fr sn=

AFliccnts t On005e On12 tnOse GreG5 relatively frs? Of c0ntroversy (Or snelussen in tne prelisancry anslysis, deferrang the other, uncertcan ssues to tne f:nO! OnC;Ystn.

IP15 60ul:

' e unf air t0 the interven:rs, Gn: 25 what has esEentIC117 c

{

noppenee here.

We connet :Oncene su:M 3n stes n.

as it 12 i

tOntos0unt tc =eferr:ng c cOntestec as-sue t0 ene Staff fCr 3

r res 1ution, c prs tice pr n:=zte i

cy tee C0asissson an Censcisectec Edsson (Incton 30snt Unit 2).

+

g, -,i n ).

e,

--p Y.a,

.a,

n v.4 e,.

_. <. nr.

2 We _ce12 eve.tn.st ene 3 rec 5 c=cressed in ~PpisCGnts' t

prels=incry Onclysis sust ne sottsfacecrily res 1vec tefore 1

l 0;erati n etcve *% pc.er.

These inclusion Of these creas anci Otes that they COnsi=er tnen spaliccb!e Oca P O te.e r e r.c n i

s I

later.

Tnss sesns enst sur findings concerning centsznnent sne

=rywell integrsty, ecntcinnent response ecce, sn= defisgrarson d

-equt=nent survivshslity sust be cdeauctely.aceressed cs O i

I

On=tttOf f:r licensing.

5:nce Appliconts inten= to have pr0ce=ures (Or Operating the nycrcsen 0:ntrci systes cef0re ccerati0n sc0ve 8% P0ner (Tr; 3425), we fine this necessary as sell, os'ciscussed c=Ove.

r

-The effe:t of :gniter systes Operation in o cegrcred OOre sericent sn =eccy'nese resaval eccezility is Isc apprcpriste t:

oceress now, os aspisconts o= nit that contatnnent.nsat ressvs; 1sirelevsne to tne analysis Of tne hydr:sen centr:1_systen.

_AcFlicants' Ex.

9-l' o r 2C Tne Secrr's jussement on tne opproprsote scene er ene p

.+A h

,.-m

---w

-n,

---,-~m.,em, m.-

--, - - -, -,. ~ -, -. -

53 -

+-,s.

preliminary analysis is based on What was required at Grand Gulf i

-for. operation'obove 5% power.

Finding 103.

The opplicable GGNS SER supplement included o discussion of the 1/20 scale tests and comparisons or CLASIX 3 with CONTEMPT-LT28.

Id.

An earlier SER I

supplement.oddressed other matters, such os the Sandio evoluotion.-

Id.

We take official notice of the fact that Grand i

Gulf SSER 3 (NUREG-0831. Supplement 3 -July 1982) also covered secondary fires (pp. 22-20 and 22-21).

We therefore consider i t appropriate to consider these matters now, i

It is'opparent from OCRE Ex. 19 that Applicants have been GWore of most of the issues we have considered above (including dryWell pool' loads and dryWell bypass lenkoge) since ~1982.

We 1

- feel there has been enough time for Applicants to oddress these matters.

To defer them any longer would contradict the l

Commission's goal of prompt compliance with 10 CFR 50.44 i

(2) (3) (iv)-(vii).

i j.

Wi~th.the.exEsption.of diffusion flam'es'to,be'itudied in tre j

1/4 scale tests.-We! find no valid reason to defer any of the matters We'have addressed'herein.- We require o commitment from i

j Applaconts to address diffusion flames resulting from o 75%

metal-Water reaction in compliance With the rule, as discussed eorlier, but for.011 other mothers-the deficiencies We have; i

l noted must be corrected before operation above 5% power.

\\

II. FINDINGS OF FACT

=

1.

Issue MS Was admitted as c' contention in'this proceeding by Memorandum ond-Order-(concerning Lore-Filed contentions; Quality

{

- Assurance, Hydrogen Explosion, and Need for Increased Sorety o f.

4

_ Control System [ Equipment). LBP-82-15, 15 NRC'555 (1982).

Issue 4

4

(

e--s tvy y 9-

+

F T

w P-9

=

--'gert-r y e s'm* Pet-+'-+c*

• y v*P V

i

'-e-v'

+--v"*-'vv'-*ww-iew-qwWme'*v h-"H-e

-'-1f*W-'T-4' uth r

B-4 4

A--4 a+-J+

• nn-A

-+4o e,-

A

--~s.--A1-e.-a--

4

.-gy -.

t

• ~

5;

.,g.

NB.wos subsequently reworded by Memorandum and Order (Horions on Hydrogen Control Contention), LBP-85 __, 21 NRC ___ (March lw, 1985).

OCRE ts the lead intervenor on the issue and the only intervenor which participated in its litigation.

I 2.

Issue ~M8, as' litigated, reads os follows:

'The Perry hydrogen control system i s inadequate to ossure thoe large amounts of hydrogen con be safely accomodated without a rupture or the containment ond o release or substantial quantaties of radioactivity to the environment.

This wording basically clieges that the hydrogen control system does not comply with 10 CFR 50. 44 (c) (3) (iv)-(vt i).

March 14 f

Memorandum and Order at 7.

i 3.

The PNPP nydrogen control system used for design basis events is the combustible gas control system.

Applicants' Ex.

3-1 at 3: Applicants' Testimony at 35-37.

This system consists of the hydrogen mixing or drywell purge subsystem; two 100%

capacity. recombiners; the hydrogen analysis. subsystems and o l

backup containment Purge subsystem.

Id.

~

4.

The degraded core hydrogen control system is o distributea 1

igniter system. Applicants' Ex. 8-1.ot 5-12, 22-25; Storr Ex. B t

ot 6-2s Notorrancesco Testimony I or 3-4; Applicants' Testimony at 29-34.

This system consists of 102 thermal gl W plug ignaters placed throughout the drywell, wetwell, and upper i

coneoinment.

Id. ot 32-34.

The system is powered from'120 VAC, l

60 Hci Closs 1E_ power distribution panels copoble.of being powered _by the emergency diesel generators.

Id. o 32; Applicones' Ex. 8-1 at 9.

The. system is designed to burn

- hydrogen at low concGntrations (8%), thereby preventing high concentrations o'r hycrogen, which might ignite rondomly and i

i threaten containment integrity.

~56-

.am,~,-

from occumulating.

Id. at 31: Notorrancesco Testimony I or 3-4.

5.

The PHPP 19 niter system is of the some design, and uses the some igniter assembises and locations, as the Grand Guir Nuclear Station (*GGNS*).

Applicants' Ex. 8-1 or 23; Applicants' Testimony at 34.

The PNPP containment and engineered sorety system designs are also similar to those at GGNS.

Applicants' Ex. B-1 at Sectior. 5.

o.

The igniter system is manually initiated, in accordance with the emergency procedure guidelines, when the reactor water level reaches the top of the active fuel.

Applicants' Ex. 8-1 at 12 Applicants' testimony at 34; Bu::elli, Tr. 3424.

4 7.

The PHPP emergency instructions for operation or the igniter system, and the generic emergency procedure guidelines on which they will-be based, are currently under development.

Bu :elli, Tr. 3425-27.

3.

The PNPP emergency instructions will include provisions for venting the containment in the event or overpressure.

I Richardson, Bu::elli, Tr. 3441-43.

The vent path for PNPP (and resultant flow rates) has not been established.

Bu::elli, Tr.

3443.

9.

The degraded core occident scenarios considered by Applicants in assessing the igniter system ore (1) a small break in the drywell ('DWB') with extended ECCS railure, and (2) o transient with a stuck-open relier volve-(*SORV=) with extendeo l

ECCS railure.

Applicones' Testimony at 39: Notorrancesec 1

Testimony I ot 5.

'10.

For these scenarios, no ECCS riow is assumed until Just prior to reaching a metal-water reaction equivalent to 75% of the active fuel clodding, at which time recovery or coolant l

- dGG -

makeup systems is assumed to occur and the transient is terminated.

Applicants' Testimony at 38.

~

11.

The hydrogen and steam release rates used for these scenarios were calculated from the MARCH computer code.

Applicants' Testimony at 40s Applicants' Ex. 8-1 at 19.

Identical mass and energy releases were used for both scenariosi the steam and hydrogen were assumed released to the drywell as well as the suppression Pool (through the safety relief volves) for the DWB Case.

Id.

The MARCH onalysis was for on unmitigated occidents to model a degraded cdre accident, the i

release history was modified such that the maximum hydrogen I

release rate achieved prior to core slump was held constant until o 75% metal-water reaction was reached.

OCRE Ex. 21 at 11 Applicants' Testimony at 40-41; No.tofrancesco Testimony II at 6.

12.

The igniter-system would not be operable during a stoeion i

blackout degraded core occident.

Bu :elli, Tr. 3432.

The

~

I hydrogen would occumulate in containment.

Id.

The hydrogen concentration in containment resulting from o 75% metal-water reaction is about 08% by volume.

Bu::elli, Tr. 3438.

Although a detonable mixture, it would not detonate if power were

~

restored and the igniters actuated.

Lewis, Tr. 3439.

However, the deflagration of this amount of hydrogen would produce high

^

. ~

6.

overpressures'of obout 50-110 psi.

Lewis, Tr. 3440.

13.

In such a situation,-the emergency procedures would snstruct tne operator to determine the hydrogen concentration prior to oc;uoting.the-igniter system.

Bu :elli, Tr. 3441.

i

~

containmene venting might be used in such circumstonces'.

.52-

~,,n-

.w aen..

Bu::elli, Tr. 3442.

14.

Applicants performed on analysis of containment ultimate structural capacity for internal pressurication.

Applicants' Ex. 8-45 Appliconts' Testimony at 25.

15.

The analysis determined static containment vessel shell capacity.using actual material stren9ths.

Applicants' Ex. 8-4 at 4-7.

Values were computed for mean and lower bound values of material properties.

Id. o t-3.

Lower bound properties are computed by subtracting three standard deviations from the mean.

Alley, Tr. 3283.

Fifty percent of the cont'ainment vessel materials are expected to have strengths greater than the mean values, and 50% are expected to have strengths less than the meon, while 99% of the materials are expected to have strengths 9reater than the lower bound values.

Alley. Tr. 3284.

16.

The dome knuckle controls the capacity of the containment shell.

Applicants' Ex. 8-4 at 6.

At 78 Psi 9, hoop buckling occurs in the knuckle (ossumin9 lower bound material

~

properties).

Id.

At this pressure, W0ves form periodically around the circumference of the dome.

Id.

The shell may j

fracture Where the waves appeor.

Id.

The methodologies used for calculating the buckling pressure are based on perfect ellipsoidal shells (containing no residual stresses and I

' geometric imperfections) and have no factors of safety.

Alley, Tr. 3347-49.

t j

17.

The NRC requested tnot Applicants perform on onolysis of j

the PNPP containment demonstratin9 that the requirements of the ASME Boiler and Pressure Vessel code,Section III, Division 1,

Subarticle NE-3220, Service Level C Limits are satisified for o 7

L

-$8 -

Ioading combination or dead lood ond internal pressure.

The minimum pressure,was to be 45 psig.

Applicants' Ex. B-4 or 13.

18.

In response to this request Applicones analyzed the containment vessel cylinder and dome and penetrations.

Id.

The stress intensities at 45 psig for the containment vessel cylinder and dome were less than the allowable stress intensities.

Id. ot Table 8.

The analyses used minimum specificied yield strengths.

Id. at 13.

The pressure capacity, pursuant to ASHE Service Level C limits, for the cylindrical.

shell is 79 Psi.

Alley, Tr. 3596-97.

19.

Analyses or the personnel airlock and equipment hatch indicated that Service Level C limits would be reoched at 50.2 Psig and 52.6 psig, respectively.

Applicants' Ex. 8-4 at 15.

20.

Analyses or the other penetrations neglected dead load, as dead load contributes o maximum or 4.7% or the stress intensity contributed by internal pressure.

Id. at 16.

The limiting Penetration is P414', with a pressure capacies or 49.7 psig, when considering the errect or on adjacent penetration.

Id. 64 17.

Toble 10.

This calculation is-based on minimum specified material properties.

Id.

P205 has on allowable pressure copocity or 54.8 Psi 9, taking credit for material 1

certifications.

Id, at 16. Table 10.

The maximum internal i

pressures to meet Level C limits were calculated by roctoring up

.the stresses to reach the allowable stress.

Id. or 15.

21.

Applicants used rinite element techniques in their analyses l

or penetrations to meet ASME Level C limits.

Id. ot 14, Figures 7-14s Alley. Tr. 3406.

22.

-The Sandia National Laboratory has conducted experiments

- fPf -

involving internal pressurication of scale models or contoanment vessels for the purpose of qualifying onalytical methods.

Alley, Tr. 3383.

One of the models roiled cotostrophically o t' 195 psig internal pressure.

Alley, Tr. 3401.

Finite element analyses of that model predicted general membrone yielding at

~.

185 Psi 9, leoko9e at-205 psi 9, and cotostrophic failure at 226 psig.

Alley, Tr. 3404-06.

23.

The ASME code relies upon a redistribution or stresses a

caused by matersol yielding at discontinuities or penetrations.

Alley, Tr. 3386-87.

At service limit A or B loodings, allowable stresses may exceed yield strengths and may approach ultsmote strengths.

Id.

24.

The matarial strengths of the containment Vessel will be decreased about 10% due to increased temperatures from hydrogen burns.

Alley, Tr. 3286.

This errect is not included in

' Applicants' onalyses.

Id.

25.

The PNPP containment vessels, as built, are out-of-tolerance with respect to their specifications.

Alley, Tr.

3344.

An onolysis was performed, for-15 psig internal pressure, to determine the stresses resulting from these os-built dimensions.

Alley, Tr. 3345.

The onolysis showed that circumferential stresses at one point were 20% 9reater enon ideal conditions.

Alley, Tr. 3345, 3347.

The analysis concluded that the largest increase in circumf ert nttal stress was 4.37%, with the largest increase in vertical itress being 9%.

Alley, Tr. 3596.

The stresses at higher. Pressures were r.or evoluoted.

Alley, Tr. 3347.

Geometric imperfections will l

-decrease the shell's resistance to

./

-Go -

buckling.

Fd.

26.

The PNPP containment vessels contain welds deviating from ASME code requirements.

OCRE Ex. 13 at 1-1 to 1-3; OCRE Ex. 18 ot 1.

These defective welds are inaccessible for repair.

Id.

The deficiencies Were discovered upon a re-review or Weld rodtographs.

Starr Ex.

6.

SSER 4.

ot 3-2.

These radiographs t

are or poor quality, and the Weld surroces are rough.

OCRE Ex.

18 at 1.

27.

Applicants commissioned Aptech Engineering Services to perform on evoluotion or the rejectable Wel'ds.

Alley. Tr. 3304.

The Aptech.evoluotion included rrocture and ratigue analyses to determine crack growth and rrocture performance.

OCRE Ex.'13 at 2-1 to 2-9.

28.

Only design basis loadings were evoluoted by Aptech.

Id.

or 3-7 to 3-12; Starr Ex. 6 ot.3-3.

Credit was taken' for the annulus concrete.

OCRE Ex. 18 at 2.

Fatigue crack growth due to multiple hydrogen.derlogrations was not considered.

Alley, i

Tr. 3306, 3325.

29.

Aptech determined the rioW si e of the Weld derects by o digital radiographic enhancement technique.

OCRE Ex. 13 at 6-12.

Approximation of riaw depth by this methodology is o guide 1

or cid.in judging rioW depth at this time.

Storf Ex. 6 at 3-2.

Further-demonstrated occuracy or the technique is needed before 1

it con be considered a standard technique for rioW depth measurement.-

Id.

The technique is acceptable os on alternative to repeated radiography-provided that the original radiographs i

Cl-meet minimum specificoriens as ec tacge quality.

Id.

30.

Tne cepen or c surtces-ecnneceec crcet is an imperecne rcctor crrecezng its arcacsoticn; ene ceeper the crack, ene less stress needed ec initicee trccrure, cliey, Tr. 3326; OCRE Ek.

13 ct 25 31 '.

The acteracts used in ene FN?F cenecinnent vessels hcVe icwer resistance ec c prcposcting tracture enon ec tracture instiction.

Ie. or 2-3.

Under constant 1cces, centinues croca propcscrien as expected once a fracture nos instictec.

Ic.

E3 2.

The FNPF contain=ent weld ricws are locctec in heri:cntal er circunferentici welds and cre crientee such that longitu=inci stres.*es cPply.

OCRE Ex. 13 ct 3-4, 3-6.

If a prcPcscting circusterentici crac6 c urred in tne icwer contcin=ent vessel, ena intsrnci pressure would lift ene vessel up, to de restrainee cnly oy penetrations, cttcennents, and externci structures.

Alley, Tr. 3337-35.

33.

The Apre:n analysis, whien includes censervatiscs cs te

~

.-m.

s-

..Orc:k grcuth rctes'.cnd tracture toughness values, rcund that, for ene cesign 1cces consi=eres, very sac 11 crock grcwen wcu1=

cecur Over the 1tre cr tne picne, one ence the rinc1 rics sice woule not prepcscte cy c receture cecncnis=, given a scxinus L

cne-eine stress Iccdtns.

OCRE Ex. 13 ct 7-1; secrr Ex. e at 3-2 to 3-3.

The-rctic cr the frceture tcugnness, Kie, tc the opplied stress intensity fceror, Kc, rcnged Fren 1.34 tc 1.5e fcr the ricws cnd Icod concinctions cnclyzed.

OCRE Ex. 13 ct 7-3 4

i 34 The nescrive internci design pressure for the PHPP 4

i

- ecntainment is -0.9 psis.

Appitecnes' Ex. B-1 ct 14 Two 04-inch vacuum relier lines are included in the FNFF dessgn tc

-GL #

assure that negative pressure inside the containment does not exceed the design value.

Id.

Two additional 24-inch vacuum relier lines are provided for redundoney.

Id.

FSAR design basis calculations, toking credit for only two of the 4 lines, showed that for conservative assumptions, a negative Pressure or

-0,72 psig was reached.

Id.

Analyses or the ultimate negative pressure capacity of the PNPP containment were not performed.

Id. ot 15.

35.

The function or.the containment vessel is to serve os o low-leakage barrier to limit fission product escape into the c

environment in the event of on accident.

Alley, Bu :elli, Tr.

3264.

36.

The containment structural integrity test, conducted at 17.

Psig, has not yet been performed.

Bu elli, Tr. 3281.

The purpose of this test is to demonstrate the capability or the containment to meet specified pressures and to exceed design pressure.

Id.

The structural integrity test _is a one-time test.

Alley, Tr. 3282.

17.25 psig is'the highest internal pressure to which the containment will be tested.

Bu::elli, Tr.

3282.

An integrated leak rate test has not been conducted yet.

Bu::alli, Tr. 3281....Th e integrated Icok rate test is conducted or 11.31 psig.

Bu::elli, Tr. 3282.

37.

The PNPP equipment hatch is pressure unsecting.

Alley, Tr.

3272.

The Preload on the hatch closure bolts is the only force rasisting the internal Pressure.

Alley, Tr. 3273.

The original bolt preload was based on on internal pressure or 15 psis, and connot be increased.

Alley, Tr. 3274.

The internal pressure or

L f

whicn ene closure scit srelcoe is overcene zs 21.6 esas.

Alley, Tr. 3277.

L

25. -Applicants

• anaisses or ene eauspaent noten assunee enor rc t

Isakoge would occur or 45 cr 52.6 psig as the ceflections Of ene hacen cover were less enan ene preccapressica of ene 0-rings, whscn cre 'cssu=ed es nave sufficient sarsng-esch to prevent leakage.

Appisesnes* Ex. 5-4 at 14-15; Alley. Tr. 2275.

~

39.

CCEpression set is a pnencDenen wnerein the C-rsnss wCuls

{

- retaan scae er all cf tneir preccapression wnen ene icsc is

}

recovec;-i.e., enere is little er no spring-nsck.

Alley, Tr.

3277.

Ccapression see wall sncrease cue :: eneraal one f

ecdiscion oging.

Allef, Tr.'3273.

Rcugnness or arregularity of the nGtch seating surfcces wt11 4

cisc cause or increase leakoge.

Alley. Tr. 3275-79.

40.

The.PNFP personnel osricens use snrictomie seals to prevene f

leakGge.

Alley, Tr. 3362.

The Argonne Nationsi Lancratcry 4

censseers snflatable sec15-te have poor resistance cgsinst severe sccident consiesens.

Euccelli. Tr. 3370.

Inflat ble i

seals for the crywell perscnnel cirlocks faile: cualification 4

i testing at s te=perature of 465 cegrees Fchtenheit.

CCEE Ex. lo at 1.

Tne seals were recess;ne= cod requalifie= (t: icwer

' tenperatures, 330 cegrees-F).

Ic. ct 2.

The internci seal pressure nse to te cecrease= reca 70-103-psis to 62 psis, as increcses eenperatures would increase the pressure within ene seals eeycne enot'ec wntch eney were.qualifie=,

Id.

41.

The uitseste copocacy of the.orywell was noe onalyced fer i

?

i s

j

- Ferrys Applicones reited upon on analysis of the Crone Gulf

..c_.,,

crywell and upon'aesign ssnilcesties between the FhPF cno GGNE,.

- drywells.

Applicants

• Ex. E-1 ct 15-17.

.The ultinste pressure

• wfw sm-w,a,,,am e

e-

,e y

e og y,we-

,,-g

--e-

• ,gw,

,,9y-g

%--,,-g 4

e,

-9 Ky9p--g--w-y fv+--

-MC g

y7y-.,p-+

i I

-lo 4 ~

capacity of the Grand Gulf drywell was found to be +67 Psidi the negative pressure capability of the GGNS drywell head was found to be -89 psid.

Id. or 17.

4 2.. The GGNS onolysis ord not evoluote the effects or voids in the concrete drywell wall.

Richardson, Tr. 3414.

Voids have been found in the PNPP drywell wall.

Alley, Tr. 3415.

They hove beeA repaired, and creas of the drywell wall containing heavily congested reinforcement (conditions in which the voids had been found) were inspected for other voids.

Alley, Tr.

3416-17.

Inspection of the entire wall was not performed.

Alley, Tr. 3416.

43.

Loss of drywell integrity con cause containment failure by steam bypasss of the suppression pool in those occidnets involving a pipe break in the drywell.. Holt:clow. Tr. 3420.

44.

Applicants' standards for computer codes is that either they ore avoilable in the public domain and have o historyor use which hos proven their opplicability and validity, or they are used to onolyze sompie problems (within the range of applicability for the octual problem to be analyzed) Which h0Ve Known Solutions from other accepted programs, classical theory.

or experimental dato.

Alley, Tr. 3382.

45.

Applicones used the CLASIX 3 computer code to evoluote the temperatures and pressures resulting from hydrogen deflagrotions within the containment and drywell.

Applicants' Ex. 8-1 at 20-21, Appendix As Applicants'. Testimony ate 41-42.

.CLASIX 3 is not

~

available in.the public domain.

Fuls, Tr. 3547.

i 46.

CLASIX 3 results have been compared with hydrogen

'comnustton tests conducted at Fenwol and the' Nevada Test Sites except for one instance, CLASIX 3 predicted conservative l

.a

+

o

- C C,-

n.

"~"

. - ~ _,

pressures and temperatures.

Fuls. Tr. 3621.

The Fenwol tests were conducted in a 3 foot diameter vessel.

Korlevit:. Tr.

3649.

The Nevada Test Site vessel was o large open vessel with no compartments.

Fuls. Tr. 3662.

47.

The Perry CLASIX 3 onalysis for the 50RV occident scenario indicated that 32 wetwell burns and 2 containment burns would occur, with a peak pressure of 21.2 psig (containment) o n'd a peak temPeroture o'f 1762 degrees-F (wetwell).

Appisconts' Ex.

3-1 at Table 17 of Appendix A.

CLASIX 3 results for the OWB occident showed'38 wetwell burns and one drywell and one containment burn-Id.

Peak Pressure was 19.4 psis (wetwell and containment), and peak temperature Was 1763 degrees-F (wetwell).

Id.

48.

The Sandia Nationoi Laboratory conducted on e. valuation of the Grand Guir igniter system, including comparison or CLASIX 3 results with other computer codes, MARCH ond.HECTR.

OCRE Ex.

21: Notorrancesco, Tr. 3688.

49.

HECTR is a mor'e advanced containment response code than the MARCH subroutine MACE.

Pratt, Tr. 3700.

Version 2.0 or MARCH hos tncorporoted certain elements of the HECTR model.

Pratt, Tr. 3689.

50.

HECTR hos been evoluoted osoinst Nevado Test Site and was found to give conservative results.

Notorrancesco, Tr. 3737.

51.

The CLASIX 3 code used hydrogen burn porometers of 8 volume-% for ignitton and-propogotion with 85% combustion completeness.

Applicants' Ex. 8-1 at Appendix A.

Table 4.

52.

Hydrogen combustion tests at the Nevada Test Site showed combustion was virtually complete for hydrogen concentrations I

greater thon.7.7%.

Lewis. Tr. 3516.

53.

CLASIX 3 used a flame speed of 6 ft/sec.

Applicants' Ex.

8-1 at Appendix A, Table 4.

Sandio considered this flame speed to be significantly low, by a factor of three or more.

OCRE Ex.

21 or 16, 17, 93.

Higher flome speeds result in less time for heat transfer, and consequently, higher pressures and temperatures.

Id.

...r.

w 5+r Ioni=ing,radioeion has the errece of promoting occelerated combustion and increasing flame speed.

Lewis, Tr. 3528-29.

55.

The CLASIX 3 onalysis assumes that containment sprays are avoilable during the hydrogen burn, and cre'ossumed to be octuated ofter the first hydrogen burn.

Richardson, Tr. 3444s Appliconts' Ex. 8-1 at Appendix A, Table 9.

56.

Containment sprays are the dominant containment atmosphere

-heat transfer mechanism, and spray operation is very important to plant safety for hydrogen burn scenarios.

OCRE Ex. 21 at 12,

?.

57.

Containment sprays at PNPP ore o subsystem of the residual I

heat removal system; onother function of that system is low pressure coolant injection, on ECCS subsystem.

Richardson, Tr.

3444s OCRE Ex. 19 at 4.

A basic postulate of degraded core occidents is that coolant is unovailable.

Id.s Richardson, Tr.

3445; Notofrancesco Testimony II at 3.

The containment spray systems shore common pumps, pipang, and electrical power supplies.

Richardson, Tr. 3445.

SS.

.The stoff-has stated that sprays should not be considered in Applicants' containment response analysis if their l

ovatlobility is questionable.

OCRE Ex. 19 or 4.

59.

Numerous igniter ossemblies at PNPP. ore locored near

m ceilings or under other obstructions.

OCRE Ex. 16s Applicants' i

Ex. 8-1 at Table 2.4-1s Bu=elli, Tr. 3502.

60.

'Sondio round that the design of the igniter housing (the spray shield) and the placement of the igniters close to ceiling structures would inhibit combustion effectiveness, such that the system would reliably ignite hydrogen at the downward t

propogotion limit, but not at lower concentrations.

OCRE Ex. 21 ot'195-96.

The downward propo90 tion limit for hydrogen 15,8.5 to 10%.

Lewis, Tr. 3514.

61.

Hydrogen' combustion tests at the Nevada Test Site showed i

that ignition at lean concentrations (6%) could not be ochseved

.;-~

l for on igniter location at the top of the vessel.

Richardson, l

Tr. 3627.

62.

The CLASIX 3 model for the PNPP Mark III containment-uses 3 compartments, drywell, wetwell, and containment.

Applicants' Ex. 8-1 at 20.

The wetwell consists of the volume bounded by

~

the-suppression pool and the HCU floor (620' elevotion).

i Notofrancesco Testimony II or 3, 5.

The volume of.the 1

containment compartment is 959,388 cubic feet;_wetwell volume is 181,626 cubic' feet.

Appliconts* Ex. 8-1 at ~ Appendix A, Toble 5.

1 63.

CLASIX 3 sensitivity studies showed that decreasing wetwell spray carryover' fraction resulted in higher temperatures'from hydrogen burns. -Fuls, Tr. 3549, 3550-51.

The.PNPP'CLASIX 3 l

onolysis assumed o wetwell spray corrover fraction of.4669.

Fuls,~Tr. 3550s Applicants Ex. 8-icot Appendix A.

Table 9.

4 64.

-The wetwell sproy carryover fraction of..4669 includes sheet flow, which 15 assumed to have half the heat. transfer l

e

- O

-CB -

effectiveness os droplet flow.

Fuls, Tr. 3550s Applicants' Ex.

8-1 at Appendix A, pp.

3-4.

There is no experimental basis for this assumption.

Fuls, Tr. 3550.

65.

The cross-sectional riow area at containment elevation j

689*-6' is 2778 square feets at elevation 664'-7',

2534 square f ee r's at elevation 642'-O',

3070 square feets at elevoeion 620'-

6*,

1900 square feet.

Applicants' Ex. 8-1 at Fi9ures 2.-12 through-2.4-15.

The containmene spray headers are'obove these r

elevations.

Fuls, Tr. 3550.

66.

The CLASIX 3 onalyses assume ehot hydrogen is' released through the' suppression pool into the bottom of the wetwell t

compartment.

Richardson, Tr. 3496, OCRE Ex. 21 or 18~.

67.

iA drywell breox accident first results in the expulsion of the drywell air moss by steam.

Fuls, Richordson, Tr. 3533.

The dryWell atmosphere will then consist of steam and hydrogen.

2 Fuls, Richardson, Tr. 3534~.

All of the hydrogen is at first' assumed to be released to the drywell until 20 minutes into the transient, when half of the hydrogen is assumed,. f u.....be released directly to the drywell, with the other half going to the i

suppression Pool through the safety relief volves.-

Applicants' Ex. 8-1 at Appendix A, p.

is Richardson, Tr. 3496-97.

[

68.

The steam and hydrogen release into_the drywell and the 2

oction of the drywell purge compressors are assumed to pressurice the drywell such that the first row of drywell LOCA vents is uncovered and the hydrogen bubbles through-the l

suppression pool.

Applicants' Ex. 8-1 at 26s Richardson,'Tr.

3497.

The CLASIX 3 model ossumes bidirectional flow between wetwell and drywell (through'the suppression pool),

bidirectional. flow between wetwell and containment, and

unidirectional riow from containment to drywell.

Applicants' Ex. 8-1 at Appendix A, Figure 2.

.69.

There are two drywell purse compressors each rated at 546 serm.

Richardson, Tr. 3497: Applicones' ex. 8-1 at 26-27.

70.

The maximum allowable drywell byposs leokoge rate is 5843 scfm at 2.5 Psig and 32.645 serm at 30 psig.

Applicants' Ex.

8-1 ot Table 5.4-2: Richordson, Tr. 3497.

These values are technical specification limits for drywell leakage and are to 0.1'8 square feet teokage o'ea.

Bu::elli, Tr.

6 r

equivalent 3616.

No action is required to be taken ir'the results or c

periodic drywell leak testing (conducted at opproximate,1y 3 psi) sndicate bypass leakage values less than that allowed by the technical specirications.

Richardson, Tr. 3499.

71.

Ir drywell leakoge occurs in a DWB occident, hydrogen would leak out through the drywell wall.

Richardson, Tr. 3498.

General Electric hos done parametric calculations or bypass leakoge assuming 0.168 square feet leokoge area, and round thot i

14-19% of the total hydrogen generated could bypass the suppression pool.

Holt: claw, Tr. 3628-29.

No evoluotions have been performed or the errect or this leakage on hydrogen concentrations in the contoznment as o function of time.

c Richardson, Tr. 3500.

72.

The Brookhoven National Laborotory performed calculations or. bypass leakoge for,GESSAR which shoWed that, assuming a 2.7 l

l psid is needed to activate suppression pool riow and a low.

I f

internal drywell pressure, on opening os small os 4 inches in diometer (0.0873 square ree t) will eliminate rioW through the suppression pool.

Pratt, Tr. 3708-12.

The greater the pool l

s

-7o-height, the greater the dirrerential pressure needed to octivate pool riow, and the smaller the bypass leakoge area need to prevent pool riow.

Pratt, Tr. 3712-13.

The suppression Pool makeup system is automatically octuated at 30 minutes arter a-LOCA 519nol, and would raise the level or the pool.

Id.

73.

Comparisons or CLASIX 3 with CONTEMPT-LT28, an acceptable and conservative computer code, showed that CLASIX 3 uncerpredicted atmospheric temperatures because it overpredicted convective heat transfer.

Notorroncesco, Tr. 3686-87.

Because or this, the Starr round CLASIX 3 to be in nonconformance with the provisions of NUREG-05GG, and stGted that it should not be t

used for determining the most severe compartment temperatures.

Id.; OCRE Ex. 20 or 1.

74.

The CLASIX 3 onalysis or the 50RV transient was terminated when the hydrogen release ended (75% metal-water reaction was reached).

Applicants' Ex. 8-1 or Appendix A, p.

2.

At the end or the transient, the 90s Concentrations in the various compartments "are:i l

Orywell, CO23 = 17.45%s CN2] = 72%s CH23 = 0.94%

CH20]

9.5%.

=

49%s CH2] = 26%i CH203 = 23%i Wetwell, CO2] = 3%i CH2] =

l-containment, CO2]

5%: CH203 = 31%.

2%s CN2] = 62%s CH2]

=

=

(

Applicants

• Ex. 8-1 at Appendix A, Figures 10-21.

Final containmene and wetwell pressure at the end or the transient is 20 psio (5. 3' p sis).

Id. at Figures 6, 7.

75.

The DUB transient was continued pose the ene or hydrogen h

release to allow for o drywell burn, at which time the transient I

t was terminated.

Id, ce 2, Figures 22-28.

At the end or the transient, the sos concentrations in the various concentrations are:

I

i t

-7f-p-

P 1,

I c

as-v -, a.

.,;. un v,

~

.. un..-,. -

p..

r..,.,,.,-

n..

e.,. y e.

u -,.

.n-un.

u

, un a a,

un va.

+

z-uu-,,3 e.,.,.

J.,., a.

- oa.

,de..e..,

.< as Ccntoinsene, CO23 = 3:; CN2

= e7:; CH2: = 13;; CH202 = 251.

!=.

or Fssures 09-42.

Verwel: cne cenecznnene pressure ct ene P

t en: cf ene transtent as 21 : sic (s.3 : sis).

t Fasures Os, i.

76.

Tne centsinnent v :cus reiter vcives cre nee s :eles :n ene

... )... s.

=t a.

e-.y C, a- :. X

.~...

i

..ee very.cf O*45 crescent ter re ene e:ne ; stuiste= in 77.

i j

tne cnclyses, e.g.,

at $523 see nes'znt ene trcnstant.. ule result in concensstien Or ene Isrge steca trortsen (0: ut 75%)

sn ene crywell et ense t s ee.. Fuls. Tr. 3535-Iss Surrells. Tr.

1 3537.

stas:triee co ulariens er enis event sn vec ense ene 1

4 4

i

ncentrottens or owysen one nyerssen veule rise fren 2% an: II:

to 5 one-57%, resse:ezvely.

Fu15. Tr. 3534, 35el; *ewss, Tr.

.I 1

3 Ifel.

The a xtura, cr:;ansity n nficancele. rnen =ee r,es

stusticle.

.Fuls. Tr. 3535, 3540-41.

Tne pressure rase tr:n tne ::abus t:On Of ents ni.=ture. ul: te $3-55 pss, cssuntng i

inatt:1 ces E:ners: tressure, 22 Es ta :( ' ent crywell were l

.t..

5...ne.

w..

3.

pressure Of 6 Esic.

Le.ts, Tr. 3540.

ine stvture is Pten,

.....,.w.,

e-ts_-.

e..,

-a e, tsn. tea.

.v

.. r..,

.m

< j.,

4 i

cf tne crywell vacuun ereckers w0uld litat tne extent Of cecressurt:sta n onc a sit nere Oxysen ant the cry-ell.

Fuls,

.r.',w..

n 75.

Gas raxtures contos?.sts less enan 5% c=ygen Or nere enon l

4 i

$$1 steca (by volune) cre inert.

CCRE Ex. 21 ce 15.

it 7G.

The cressure rase.0:ases ty a nydrogen ceficgestion as 4

E t

v an

---e,-

y

+-

w e

-,s.

--w

-r

--,r, e

m.:es,

-m go

+,,,..-

r,

-7JL-mainly dependent un the snitial pressure (pre-burn) and the number or moles or hydrogen consumed.

OCRE Ex. 21 or 199.

~

-Small-chonges in initial pressure con produce signirscont changes in peak burn pressure.

Id. or 17.

Elevated initial pressures in contoinment are expected for the DWB scenario, in which the drywell oir moss is pushed.into the containment.

Id.

or 29. 94 80.

Sandio performed a number or calculations with the HECTR code os o comparison with the GGNS CLASIX 3 models.

OCRE Ex. 21 at 15-16, 28-29.

Some or these cases used, identical input parameters and compartmentolication models, while others used input values considered more realistic by Sandio.

Id.

In all cases HECTR predicted higher peak pressures than did CLASIX 3.

Id.

81.

HECTR cose B-1 modeled CLASIX 3 case 1.

Id. at 28.

Identical curn porometers (10% ignition and ProPogotion limits, 100% combustion completeness. 6 ft/see riome speed) were used.

1 Id. at 20-21.

Containment sprays were assumed to. initiate automatically arter the first burn.

Id.

These cases modeled on elevated initial, pre-burn pressure, os would result from.the DUB transient.- Id.: Id. at 29.

HECTR case B-1 used a hydrogen source term equoi to that in CLASIX-3; all other HECTR runs used a hydrogen source term less than thot in CLASIX 3.

Id. or 18, 20.

CLASIX 3 produced a peak pressure or 1.755 otm (11.1 psig);

HECTR, 5.6 Lem (67.6 psig).

Id. or 30.

82.

HECTR case B-3 models CLASIX 3 case 3, which used the some eurn and spray parameters os in the PNPP 50RV onalysis.

Id. at 20-21.

CLASIX 3 produced a peak pressure or'1.456 ohm (6.7 ps19).

HECTR's peak pressure was 3.830 otm (41.6 psig).

Id.

Ot-

73 --

30.

33.

HECTR cose B-4 models CLASIX 3 cose 6.

These cases are the some os the previously consicered case, except that sprays are assumea to be orr.

CLASIX 3 predicted a peak pressure of 1.701 orm (10.3 psis)s HECTR, 4.63 otm (53.4 psis).

Id. or 30.

84.

CLASIX 3 case 5 considered a comportmentali:oeion model consisting or one coneoinment comportment (i.e.,

the werwell was not constdered a separate compar tmen t) : HECTR case A-1 also modeled this condikion.

Id. at 20-21. 28.

CLASIX 3 Predteted a peak pressure or 3.361 ohm (34.7 psis)s HECTR, 4.09 otm (45.4 psi 9).

Id. at 30.

i 85.

HECTR cases B-5, B-6, B-6',

and B-7 used dirrerent burn parameters, but all modeled the CLASIX 3 compartmentalizoeion.

Id. ot 29.

All these cases assumed 100% combustion completeness, riame speeds which were o function of nyorogen concentration tusing a correlation developed from turbulent flame experiments in the Sandio VGES tank), and realistic propogotton limits (4.1% for upward propogotion, 6% for i

hOrl Ontal, and 9% for downward).

Id. ot 15, 18, 22, 29.

Case B-5 used on ignstion limit or 8%, with automatic spray action.

Id.

This cose gave o peak pressure or 4.877 otm (57 psis).

Id.

at 31.

Case B-7 was similar, but used o 10% ignition limit.

Io. at 20.

This case had a peck pres 5ure of 5.500.otm (66.2 psi 9).

Id. ot 31.

86.

Coses B-6 and B-6' were identical, except that B-6 was termanotec wnen the hydrogen release ceased, whereas B-6' was i

continued post that time.

Id. ot 29.

Both ossumed on ignition limit or 10%, and sproys were assumed on during the entire run.

.i Id. or 20.

Case B-6 produced a peak pressure or 2.467 cem '(01.6 l

~

_.74-b psts), while B-6' hos o peak er 5.693 otz (69 psts).

I.

or 31.

87.

Realistic-prcposorion liszts such as used in HECTR cases B-

$ throush 5-7 will cause upward propogotins turns to occur, limiting the buildup or hydrogen in upper compartments.

Ic. ce 15.

The CLASIX~3 propogotton medel is unreolistic.

Id.

SS.

In both CLASIX 3 and HECTR, wetwell burns have small pressure rises, wotle ceneosnaene burns leoa to hisn peo.

pressures.

Id. ot 16, 199s Applicones' Ex. 8-1 oe pppencix A, pp.

o-7.

This is mecause che small werwell roptely reocnes the necessory conditions for c nbustzon. and the resultant pressure rise ss small cue to the transport or gas from the verwell to the containment.

CCRE Ex. 21 at 199.

In CLASIX 3 inertsns or ene weewell (by exygen cepletion or by steon) is prevented mecause containment air is transported into tne wetwell at on orttrically high rote when ene post-combustion gases eccl.

Id.

s CLASIX 3 predtets roster ecolans than coes HECTR.

Io. or 16.

The result or ustns CLASIX i

)

3 wtth ~ its present ccaportsneneolication-is on unreolistic calculotson or ene mixtng and combustion process, leading to underesttmated hyaregen burn pressures.

Id. at 199, 9c.

Incomplete conbustion results in lower pressure esses enon complete combustson, wnien results in fewer burns with larger pressure rises accompanytns each burn.

Id. at 93.

5 90.

HECTR cases B-2 and s-O' modeled CLASIX 3 cose 2.

OCRE Ex.

21 or 25.

The input ossumptions were the some os for CLASIX 3 cose 1 (10% ignition and prcpogoeton linsts, automatic spray 4

cptten, and flome spees or i rt/sec), except that pre-curn pressure was not elevated.

Id. or 20.

The HECTR run useo o l

,,.,.2 m._...--

smaller hydrogen release than did CLASIX 3.

HECTR cose B-2 produced a peak pressure or 4.845 otm (56.5 psig), compared to the CLASIX 3 peak o f. 1.'504 otm (7.4 psig).

Id. at 30. (HECTR cose B-28 examined the treatment or wetwell sprays, and round that Wetwell inerting was delayed, but gave similar final results.

Id. ot 28.)

91.

Applicants have complied a list or equipment required to survive o hydrogen burn.

Appliconts' Testimony at 50s Applicants' Ex. 8-1 at Table 5.6-1.

1 92.

Applicants conducted a preliminary evoluotion or equipmene

(

survivooility based on a comparison of PNPP and GGN5 conecinment pressure and eemperature proriles calculated by the CLASIX 3 I

coce.

Applicants' Ex. 9-1 at 21Ai Gorg Testimony oc 3-6.

Applicones conducted on analysis, using the HEATING-6 heat econsrer computer code, or the igniter assembly, using the CLASIX 3 temperot.ure profile to confirm that the PNPP 1

cemperature profile results in lower equipment temperature than the GGNS Profile.

Applicants' Ex. 8-1 ot 21C.

Applicones base their conclusion or equipment survival On the similarity between i

GGNS and PNPP temperature profiles, the heat transfer analysis, l

the similarity of l

the PNPP and GGNS equipment lists, and the margins between the response calculated for GGN5 and the PNPP equipment qualification temperatures.

Applicants' Ex. 8-1 at 21D.

l 93.

The' containment vacuum breakers, 1N17F0010, 1M17F0020, i

1M17F0030, and 1N17F0040, are only. qualified to 15 psig.

L Applicants' Ex. 0-1 at 210. Toble 5.6-1.

The hydrogen mixing compressors are qualified to 14'.9 psig.

Id.

The compressor

check volves are qualified to 15.3 psig.

Id.

Applicants-Ossumed that,the check volves and vacuum breaker will not be exposed to' peak burn pressures.

Id.

The compressors at ggt 45

.Were evoluoted and shown to survive pressures of 24 psig.

Id.

These assumptions and evoluotions have not been confirmed by testing.

Bu::elli, Tr. 3570.

94.

The likelihood of a detonation occuring is very small.

OCRE Ex. 21 at 194.

95.

Tests conducted in a 1/20 scale model of the Mark III containment found that diffusion flames exist for hydrogen release rates exceeding 0.4-0.5' lb/sec.

Richardson, Tr. 3551.

Preliminary results of these tests indicate that thermal loading to equipment resulting from o 75% metal water reaction will be unocceptable.

Richardson. Tr. 3553-57.

96.

The Hydrogen Control Owners Group ("HCOG') is conducting rests an a 1/4 scale test facility to better define the thermal environments resulting from diffusion flames.

Richardson, Tr.

3552: Applicants' Testimony at 23-24.

HCOG does not intend to evoluote the thermal environments caused by diffusion flames using a release history equivalent to a 75% metal-water reactson in'the 1/4 Scale tests.

Richardson, Tr. 3568.

97.

CLASIX 3 sensitivity studies (us2ng input assumptions other chon those in Applicants' preliminary analysis) have predicted Violent overflow of the suppression pool into the dryWell.

4 Richordson, Tr. 3487, 3494.

Sondio also predicted that significant pool surge may occur in the drywell.

OCRE Ex. 21 at 195.

90.

Long term decay heat removal is neeeed to maintain the core

i sore condi. tion.

Richardson, Tr. 3478.

Successful decoy in a heat removal depends upon a noving on RHR loop operating in the suppression pool cooling mode.

Rtchardson, Tr. 3576-77.

99.

The PNPP RHR syseem includes suppression pool cooling and containment spray subsystems.

Applicants' Ex. 8-1 at 25.

Containment sprays are outomatically actuated at 9 psig.

Id. ot 26.

The Spray subsystem tomes precedence over other RHR runctions, except for LPCI during the first 10 minutes rollowing a LOCA signol.

Richordson, Tr. 3448.

If containment pressure remains above 9 ps19, both trains of RHR will operate in ene spray mode.

Richardson, Tr. 3448.

OPerotor action is necessary to align the RHR system from the spray to other modes, even

~

arter pressure drops below 9 psig.

Richardson, Tr. 3449, 4

1 100.

Hoving both RHR loops in the containment spray mode will cesrode suppression pool cooling and mixing.

Richardson, Tr.

3453-55, 3481-85.

101.

Equipment survivability tests conducted at the Nevado Test 4

' Site resulted in extensive burning of electrical cable insulation during tests involving 13% hydrogen; burning was for shorter durations at a hydrogen concentration of 10%.

Gorg, Tr.

i 3717, 3748.

Heat'loods imposed on electrical cable insulation by hydrogen combustion con result in pyrolysis or gosification of the insulation compounds.

OCRE Ex. 24 ot 1201, 1205.

The combustible gases thus liberated will offect the subsequent containment ' pressure and temperature nistories.

Id. at 1205-06.

.,....r 102.

Applicants submitted to the Staff on Horch 1, 1985 o-preltminary analysis of the hydrogen control system.

Applicants' Ex. 0-is Applicants' Teskimony at 19.

The onolysts

- included a description of the igniter system, on analysis or i

m m

-_ ~ -

_ 74 -

\\

containment capacity, o containment response onolysis, and a comporsson of the PNPP and GGNS design rectures.

Id.

Prior to submitting the analysis, Applicants and 5torf agreed on ies scope.

Id.; Applicants' Ex.

8-2, 0-3.

On March 21, 1985 Applicants submiteed on update to the preliminary analysis.

This submittol included on analysis of equipment survivability.

Applicants' Ex. 6-1 at 21A-210.

103.

The basis for the Stoff's Judgement of the scope of the preliminary analysis is what was required or Grand Gulf for operation above 5% power.

Notorrancesco, Tr. 3742-43.

The opplicable SER supplement. GGNS 55ER 5 (August 1984), includeo o discussion of the 1/20 scale tests, comparisons of CLASIX 3 with CONTEMPT-LT28, and referenced the findings or SSER 3, which relied upon HECTR calculatsons in NUREG/CR-2530.

Notofrancesco, Tr. 3740, 3744-46.

No HECTR calculations performed orter the issuance of that HUREG were relied upon by the Staff in its' evoluotion of Grond Gulf.

Id.

104.

The final analysis of the PNPP hydrogen control system is scheduled for completion in mid-1986, and is dependent on experimental and analytical work to be performed by HCOG.

Applicants' Testimony at 22.

III. CONCLUSIONS OF LAW Appliconts have failed to meet their burden of proof on Issue M8, as compliance with 10 CFR 50.44 (c) (3) (iv)-(vii) hos l

not been demonstrated.

i f

m mL.-.-~-

IV. ORDER UHEREFORE, IT IS ORDERED 1.

The opplication for licenses to operate the Perry Nuclear Power Plant. Units 1 and 2, at power levels grer.ter than 5%, is DENIED.

2.

Pursuont to 10 CFR 2.760(o), this is o partial initial decision that will constitute final action of the Commzsszon forty-five (45) days from the dote or issuonce unless a notice or appeci is filed pursuont to 10 CFR 2.762 or the Commission ozrects tnot the recora me certified to it.

3.

A notice or appeal of thzs decision or deszgnated portions thereor may be riled with the Commission, in the form required by 10 CFR 2.762(o), within ten (10) days creer service or this cecision.

4 To pursue on oppeal, briers in support or the oppellant's position must also be riled, within thirty (30) days arter rilsng the, notice or oppeal (forty days if the NRC Starr is the appellant).

The brser must comply with the requirements or sectson 2.762.

S.

Within t h'z r t y (30) days (Forty days for the NRC Starr) arter the time hos expired for the riling and service or all appellone f

orters, parties moy file opposing or supportins briers that i

comply with the requirements or Section 2.762.

i

.. _ ~,,

r.--.

.- e o.

~

4.

Filings that do not comply ut th the rule governing oppsols may be strakin.

Respectfully submitted, c5(

Susan L.

Hiott OCRE Representoeive 6075 Munson Rd.

Hencor, OH 440e0 (016) 255-3156 o

i

APPENDIX A WRITTEN TESTINONY RECEIVED INTO EVIDENCE APPLICANT 5' TESTIMONY Following Tr. 3041:

Applicants' Direct Testimony or Eileen H.

Bu::elli, John D.

Richordson, Kevin W.

Holt: claw, Roger W.

Alley, Bernard Lewis, Bela Karlovit, and G. Nortin Fuls on the Preliminary Evoluotion of the Perry Nuclear Power Plant Hydrogen Control System (Issue MB), with shotement of qualifications of said witnesses.

Following Tr. 3752:

Oral Testimony (without written testimony) of James H.

Wilcox i

STAFF'S TESTIMONY Following Tr. 3676:

Testimony of Allen Notorroncesco Regarding Issue M8 (Hydrogen j

Control), with statement or qualifications

("Notorrancesco Testimony I")

Testimony of Li Yang Regarding Issue #8 (Hydrogen Control), with statement of qualifications 4

Testimony of Allen Notorrancesco on the Hydrogen Control Issues CGntained in the Licensing Board Contention M 8

(*Notorroncesco Testimony II')

t Testimony of Hukom C.

Gorg Re9arding Issue M8 (Hydrogen Control), with statement or qualifications Statement of qualifications of William Trevor Pratt (oral testimony without written testimony)

l l

1 l

APPENDIX B i

Exhibits j

Admitted at Following Exhibit Identified at Transcript Transcript Number Description Transcript Page Page Page App. Ex. 8-1 The Cleveland Electric 3219 3243 3243 Illuminating Company Preliminary Evaluation of the Perry Nuclear Power Plant Hydrogen Control System, March 21, 1985 s

App. Ex. 8-2 Letter from M.

Edelman 3219 3243 3243 5

to B.J. Youngblood, dated February 5, 1985 re Perry Nuclear Power Plant i

Hydrogen Control Evaluation

[

App. Ex. 8-3 Letter from B.J.

3219 3243 3243 I

Youngblood to M.

i Edelman, dated February 20, 1985 re Acceptability j

of the Scope, of Hydrogen j

Control Design and i

Analytical Information to i

be Provided to Support Full l

Power Licensing.of Perry j

Nuclear Power Plant, Unit 1 y

i App. Ex. 8-4 Letter from M. Edelman 3219 3243 3243 l

-to.B.J. Youngblood,

~

dated February 11, 1985 re SER Confirmatory Issue (3) Containment Ultimate Capacity Analysis

~

l

Admitted at Eo1 iowl m Exhibit Identified at Transcript Tra n scri p t Number Description Transcript Page Page Page Staff Ex. 8

" Safety Evaluation 3675 3677 not bound Report, NUREG-0887, into record Supplement No.

6" OCRE Ex. 12 Letter from Mr.

C.O.

3261 3263 3263 Thomas to Mr.

G.G.

Sherwood, dated April 13, 1984 re Request for Additional Information Regarding the Severe Accident Review of Gessar II OCRE Ex. 13

" Analysis of 3305 3343 3343 Inaccessible and Potentially Rejectable Defects in Perry Nuclear Power Plant" authored by Warren P. McNaughton, Jeffrey R.

Egan and Jeffrey D. Byron of Aptech Engineering Services, dated July 1983 oCRE Ex.14 Letter h m. Edei.wan Yo 33ry 337e 3378 J l(eppler, d.ortc4. ritarc.b 7,

• g9 35 cc Dr well Air Isck 15sor sea 5 OCRE Ex. 15 Letter from M. Edelman 3965 3967 3%67 to Mr.

B.J. Youngblood, dated May 29, 1984 re Piping Design Review OCRE Ex. 16 Table 2.2-1, Igniter 3508 3508 3508 locations from the Perry Nuclear Power Plant Units l

1 & 2 Interim Report on l

the Hydrogen Control System i

6.2

I 1

Admitted at Following Exhibit Identified at Transcript Transcript Number Description Transcript Page Page Page

'OCRE Ex. 17 Attachment A,

" Experimental 3562 3562 3562

^ '

Study of H, Diffusion Flames Burning Above a Pool of Water", from the Combex Study'of Hydrogen Control at Grand Gulf Nuclear Station, dated 1981 OCRE Ex. 18 NRC Memorandum from John 3680 3681 3681 Stefano to B.J.

Youngblood, 1

dated May 4, 1983.re Summary Report of meeting with the Cleveland Electric Illuminating Company (CEI) on Perry Containment Weld Deficiencies OCRE Ex. 19

-Letter from A. Schwencer to 3682 3683 3683 D. Davidson, dated September 16, 1982 re Request for Additional Information-Regarding Degraded Core Hydrogen Control for the Perry Nuclear Power Plant (Units 1 and 2)

OCRE Ex. 20 Letter from B.J. Youngblood 3685

~~

3685 3685 to M. Edelman, dated August 30, 1984 re Request for Additional Information Regarding Hydrogen OCRE Ex. 21 NUREG/CR 2530 3691 3691 3691 Review of the Grand Gulf Igniter System OCRE Ex. 22 NRC Memorandum 3693

REJECTED, i

from Marc Wigdor, 3696 j

B-3

Admitted at Following Exhibit Identified at Transcript Transcript Number Description Transcript Page Page Page i

through Jack Rosenthal, to Brian Sheron, dated October 24, 1984 re Hydrogen Control Owners' Group and NRC Meeting, October 3 and 4,,

1984, discussing the use of BWR Heatup Code OCRE Ex. 23 Draft Report, "An 3701

REJECTED, Assessment of Postulated 3703 Degraded Core Accidents in the Grand Gulf Reactor Plant," by R.D. Gasser of Brookhaven National Laboratory, dated June 1982 OCRE Ex. 24 Paper presented at the 3714 3715 3715 Second International l

Conference on the Impact of Hydrogen on Water 1

Reactor Safety, " Electrical Cable Insulation Pyrolysis and Ignition Resulting from Potential Hydrogen Burn Scenarios for Nuclear s

Containment Buildings,"

by A.L.

Berlad, R. Jaung and W.T. Pratt l

1 l

e O

l B-4

APPENblX C C oM PA RI SOAl CF CON TA /^/M E'A/ 7-

/REsPousE AA/ ALY SE S CLAs/X 3 164 prop.

• /.

F'Amc spreb, Ajo, o F BOR AIS T m Ag,

9r%y, Pt.APPT CASE umtT um tf comkt rE F hrc.

5?RRls

. c. r-

.ww

• F (psie Noyff T

8 8

SS 6

AU TO 2-

1. A I 10 l' E -

pg pp S, gg UPP bwB B

B 8,6 6

A u'ro O

30 laOt

/3.S MAls SoRV B

B 85 6

Au ro

/

59

/o.2.o 23.9

$ Ass bwe 8

8 85 6

Avro o

2.6 I3 fir o it. g r

MNS

/

/0

/O

/00 6

t90To 2,4 o

,g

,y y,

GAIS 2

/o

/0

/00 6

Avro 2, 5 o

ty 3

, t; y ;

3,4

&AIS 3

6 8

85 6

Avro o

58 jos 2 67 2,G 3Al5 6

8 8

85 6

oFF q

t.(

ao2.3 34,g 2, 7

%NS 8

8 86 6

PF

• • O 0

68

/575

,to. 5 NO TES *

/. TA kfAl FROM TAot E /8 of APPEA/blX A of APPLtcAAtys' r.c, g ;,

.2.

YAKEAf FROM 7A BLE S 2.1 Aab -?./S o f OcRE Ex.2.1.

3.

Tcin P cRA rvat A a b p a rssv R e ina>u rn A 74KcM FRoM wc7w;u Adb Cos/TA/drwd7 VA LVEs, Lt.

HEC TR CASE 8-/

/nobEls CLAS/X 3 CASE l.

. S.

MEC TR CA5CS 6-2 A,0 b P -2.!

In c BEL Ct.hSt% 3 CA S c 2..

pfc1R CAsi 2..tl ALLowtb SPRAV' C A RR'{ C J E'R /^t To YM W 1"EL '- '

C-

i

// ECTR gu,

paop, PsAme' Siho N 3. or a u R.9 s 7,n,;w, pm No7ES PLAMT CASE UWT UM87 comm.cTF FT/src SPRMS c,7 ww

• F (ps,c

',9e M j

7,g jga7 47,4 GAIS B-l

./o

/c

/o o 6

Auto uns s-2.

so to.

/o o 6

AUTO

/

2.4

/922-SL5 qqns 6 -2.'

to

/o ja o G

Au-ro

/

At

/Bif 59.'l f

, GrMS B-3 8

8 85 G

AUTO l

31.

/877 t//. 6 GGNs A-I a

s a5 s

OFP Ll

• \\

jfjg jjg,ll i 56as a <4 g

a as off 7

3a

,g,y

$3,4 l %ns B-S g

g(xx) to o p(y,gy 2t,q Avro j

go jg9o gy.o l' ' ' ' i ' *

(6a 5 8-6

/o EO;.#d

/o o

@gd = 25.3 o4

/

/d

/89o al.6 1645 6-6 '

/o cdw;,)

/o FM#h M 3 0^/

1 IT do78 69.O

' 's h '

E645 B-7 fo 6p,/4.)

/oo p(.,,gp 25.3 Auro 1.

2I 41o32.

66.2.

i l

Atorts, coa riaveb U

+

4. NECTR CAsc g -5 sn tbELS C LAS$L 3 CMC 3,

i 9 HECTA Cast A-1

/no het.s csAsix.s cAss s,

8. # ECTR CASE B-9 setohet s CLAsix 3 CASK 6.

4.

R.D.Na). = 4. t '7. Fs it uPw Akb P Ao 44 6 AT/ 04/, 6 9. FoR. BoRI'EAA/744. PRoPv44A T'8^/,

1. */. Fok bew Moo 4R h M o.

8 t

/o 7A' KCAl FRom YABLES 2,2 AAlb 2./6 o F o C A E C x. p.l.

o

//.

TAKCAl Fhent 7M8LC4

2. s-l AMb

.3.18 OF GCkE EX* LI.

/2 FLArntE.TPccb CArLiuL Afeb Front E q uA 71 sal 2.1 o f O C R G" h* Al U^'k INN #'*# #'"

s

/ 3.

M*1'R CASK B-6 ' sbm7tcAL, -T'o 8.fo, KXCCPT FOR LoM4ER ROM 7/ntC.

j A.

& sousc.c riR sv FoR sec:rR case s g us enus. -ro cLASIX.3 s e u (4 c e~ 'fAM, ALL, OMR, Mc~c7R Cases MAvc A soo Rcc 1. Am s.ess rasa ctAssx 3 I

C-J-

o 1

I i.

i l

t i

CERTIFICATE OF SERVICE This is to certi'fy that copies of the foregoing were served by depo g in the U.S. Mail, first class, postage prepaid, this

/3 day of

• T~) a t 1986 to those on the service list below.

i, SNkC" W

y 7 4 0 :fj 1-Susan L. Hiatt C 0F SECRE7pg,

K N

BR Ncy

% - Y M b b D. W is '/oPAPP .C SERVICE LIST I JAMES P. GLEASON, CHRIRHAN / Terry Lodge,, Esq. ATOHIC SAFETY 1 LICEN5ING 80ARD 618 N. Michigan St. 513 GILHOURE DR. Suite 105 SILUER SPRING, MD 20901 Toledo, OH 43624 i Dr. Jerry R..Kline Atomic Safety.& Licensing Board. i U.S., Nuclear. Regulatory Commission Washington,'D.C. 20555 Mr..Glenn O. Bright Atomic Safety & Licensing Board U.S. N'uclear Regulatory Commission Washington, D.C. 20555 Colleen P. Woodhead, Esq. Office of the~ Executive Legal Director U.S. Nuclear Regulatory Commission ( Washington,.D.C. 20555 .M Jay.Silberg, Esq. Shaw, Pittman, Potts, & Trowbridge 1800 M Street, NW . Washington, D.C. 20036 Docketing & Service Branch i Office of'the Secretary U.S..Noclear Regulatory. Commission Washington, D.C. 20555 Atomic. Safety.&, Licensing. Appeal Bo'ard Panel U.S. Nuclear Regulatory Commission Washington, D.C. 20555 I I}}