ML19347E292
| ML19347E292 | |
| Person / Time | |
|---|---|
| Site: | McGuire, Mcguire  |
| Issue date: | 04/14/1981 |
| From: | Blum S, Jeffrey Riley BLUM, S., CAROLINA ENVIRONMENTAL STUDY GROUP |
| To: | |
| Shared Package | |
| ML19347E289 | List: |
| References | |
| NUDOCS 8104240532 | |
| Download: ML19347E292 (52) | |
Text
e L.f
. /m \\ r r
4
)
lY
':-m
/7 t '.1 161981 UtIITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION e
BEFORE THE ATOMIC SAF2TY AND LICENSING BOARD
'g s
In the Matter of
)
)
DUKE PC' DER COMPANY
)
Docket lios.
50-369
)
50-370
('dilliam B. McGuire Nuclear
)
Station, Units 1 and 2
)
INTERVENOH'S FRUFOSED Fillu11*GS OF FACT AND CONCLUSIONS OF LA'd IN THE FCRM OF A SUPPLEMENTAL INITIAL DECISION In accordance with 10 CFR 52.754, Intervenor, Carolina Environmental Study Group, hereby submits Proposed Findings of Fact and Conclusions of Law in the form of a Supplemental Initial Decision.
Respectfully submitted, Shelley Blu..
1402 Vickers Avenue Durham, ii.C. 27707
)
(917) h93-223d F ls A l
f JeMeL.Riley d54 Henley Place l
Charlotte,
'i.C.
20207 (704) 375-k3h2 l
April
/f 1961 l
l 5194240532
a y
~
9 t
s
<W
- cnin g-lN[ AER 1 O1981
- UNITED STATES OF AMERICA i
NUCLEAR REGULATORY COMIIISSION
\\ '.{h
- nnt:r:
L -s v
BEFORE THE ATOMIC SAFETY A!TD LICENSING BOARD
/
/wte Robert M. Lazo, Chairman Dr. Emmeth A. Luebke, Member Dr. Richard F. Cole, Member In the Matter of
)
DUKE PCWER COMPANY
~
)
)
Docket Nos. 50-369
)
50-370 (William B. McGuire Nuclear j
Station, Units 1 and 2 SUPPLEMENTAL INITIAL DECISION (REOPENED OPERATING LICENSE PROCEEDING)
- 1. EACKGROUND.
On April 16, 1979, the Atomic Safety and Licensing Board (" Licensing Board") issued an Initial Decision in the captioned proceeding.
Duke Power Comeany (William 3. McGuire Nuclear Station, Units 1 and 2), LEP-79-13, 9 NRC 469 (L979).
The Initial Decision authorized the Director of the Office of Nuclear Reactor Regulation, upon making requisite findin6s with respect to uncontested matters not embodied in the Initial Decision, to issue operating licenses for the two units.
9 NRC at pp. 547-d.
The Licensing Board stayed the effectiveness of the Initial Decision "until further order by the Board following the issuance of a F
supple..ient to the NRC Staff s Safety Evaluation Report ("SAR")
addressing the significance of any unresolved safety issues."
'li.
On May 23, 1960, the NEC Staff issued the aforementioned supplement to the SER.
2.
Subsequently, Applicant moved the Licensing Board to
- lift the stay.
In response, Intervenor, Carolina Environmental Study Group (CESG), noted'ts opposition and moved the Board to reopen the proceeding.
By Memorandum and Order of November 25, 19d0, the Licensing Board denied Applicant's motion to lift the stay, reopened the McGuire operating license hearing, and admitted new contentions regarding the combustion of a sig--
nificant hydrogen release.in a Wetinghouse pressure suppression containment.
3 The reopened hearing began on Feb. 2h, 1981 and con-cluded, except for filing of a Staff affidavit and response of the parties thereto on March 19, 1901.
Proposed findings of fact were to be filed by Applicant on April 6, 1981.
Staff and Intervenor were to file April 13, 1961. Intervenor requested, and horeby receives additional time to file.
4 PARTIES.
The parties to the proceeding were those to the earlier Operating License proceeding, Applicant, Duke Power Company, NHC Staff, and intervenor, Carolina Environmental Study Group (CESG).
3.
CONTENTIONS AND ISSUES.
In a revised matinto reopen the operating license proceeding, Aug. 13, 1960, CECG put forth the following contentions:
Contention 1:
The licensee has not demonstrated that, in tne event of a loss-of-coolant accident at McGuire:
1.
substantial quantities of hydro 6en (in excess of the design basis of 10 CFR 550.hh) will not be generated; and 2.
that, in the event of such generation, the hydro-gen will not combust; and 3
that, in the event of such generation and com-bustion, the containment has the ability to with-stand pressure below or above the containment design pressure, thereby preventing releases of off-site radiation in excess of Part 100 guideline values.
Centention 2:
Neither licensee nor NRC staff has demonstrated that a McGuire ice containment will not breach as the result of the rapid combustion of quan-tities of hydrogen which a dry containment would withstand.
Contention 3:
Neither licensee nor NRC staff has demonstrated that the emergency planning radius of 10 miles is sufficient for protecting the public from the radioactive releases of a low pressure, ice condenser containment ruptured by a hydrogen explosion.
Contention u:
Licensee and UHC planning do not pro-vice for crisis relocation which would be required as a result cf containment breach and radioactive particle re.3ase.
6.
These were admitted by the Board in its memorandum and order regarding CESG8s Motion to Reopen Record of Nov. 25, 1900.
7.
CESG filed supplemental contentions 5 and 6 on Nov. 7, 19eo, with its reply to Applicant's motion for summary disposition.
The Board in its memorandum and order of Feb. 13 denied CE30's motion to add contentions 5 and 6.
They are:
Contention 5:
Under current practice the NHC is required to issue an environmental impact statement as to the consequences of Class 9 accidents.
Such an environmental impact statement is required for McGuire.
l Contention 6:
The eg.ergency plan for McGuire must, due to the special circumstance of close proximity to a large populatinn center, be revised to provice an emergency response for the city of Charlotte in i
the event of a Class 9 accident.
8.'
The Board concluded that the cperating licence proceeding must l
.g-be reopened'to hear evidence on the hydrogen generation control matter, stating that CSSG's contention "may well shed significant light upon the key safety findings which are required to be made before operation of McGuire Units 1 and 2 could be authorized,"
noting that "the commission and others recognize that the question of hydrogen-generation control is a significant one."
LlMITED AFFEARANOES 9.
Limited appearance statements were made by 49 members of the public.
Additionally a letter was filed documenting one of the statements.
10.
A majority of the persons appearin8 expressed concern about the safety of nuclear generation of power.
The great individual variety in expressions made it apparant that them views were not the result of an orchestrated campaign.
Those advocating the operation of the plant were primarily concerned with economic matters and the availability of electrical power.
MOTION FOR SUPllAR'I DISPOSITION: FOR DISMISSAL 11.
Applicant moved for summary disposition for a license authorizing fuel loading, initial criticality, sero power NRC physics testing and low power testing for McGuire Unit 1.
Staff and CESG filed answers Nov. 7, 1981.
The Eoard granted those parts of Applicant's motion not concerned with low power testing.
The Board denied summary disposition in the matter of low pofer testing.
"McGuire is a thin shelled reactor and the hydrogen issue is'related to the pressure capability of the containment i
l I
i
5 structure."
12.
Applicant moted dismissal of CESG from the case (Tr. 5236).
The Board denied the motion (Tr.5253) before concluding the hearing Intervenor has made/morethan a prima facie case.
March 19, 1961.
e O
-,,. - - - ~ _,,. -
-,,,---.,,,.we,,
,..,yo-
.y,,-
g.~,,,.,,.
,-_,-g..,,., <,..,,,.
,<y-,
,.,,--,-n..,
,e,
-rm-,
FINDINGS OF FACT ON CONTxOVERTED ISSUES ACCIDENT CREDIBILITY 13.
Applicant has repeatedly asserted that it is CESG's burden to prove a credible accident in which excessive hydrogen release We note tha t Applicant did not designate the S D te-occurs.
2 quence as the one and only basis on which the case could be lit-igated until January 6, 1981, at which time it served DPC Exhibit 5a, b, and e on Intervenor.
14.
Staff has argued that "the legitimate area of inquiry here would be whether or.not the proposed hydrogen mitigation system by Duke has been tested against, for purposes of this licensing action, appropriately representative accident (s), and we think that CESG may properly inquire with respect to whether or not probabilities of some other type of scenario have been established such that maybe the other scenarios should be the one we are look-ing at."
(Tr. 4404) 15.
Intervenor did not, in filing the concentions which this Board admitted, and which are the basis for this proceeding, in any way commit itself to the burden of proving a credible accident scenario.
16.
Contention 1.
The licensee has not demonstrated that in the event of a loss-of-coolant accident at McGuire:
1.
Substantial quantities of hydrogen in excess of the de-sign basis of 10 CFR 550.44.i'l not be generated.
17.
In accepting this contantion this Board did not premise what
~
the loss-of-coolant accident would be nor that CESG would have tc prove the credibility of any ocher accidents this Applicant would,
~7-subsequently, choose to deal with.
18.
As to the credibility of an accident, we do not find the defin-ition advanced by Applicant's witnesses of "very, very remote" acceptable (Tr. 2993).
In the practice of the AEC and NRC the use of " credible" accords with the dictionary definition " capable of being believed; worthy of belief".
In judging w'. tat accidents are to be considered in PSAR's, FSAR's, and SER's, the Commission is concerned with the physical possibilities of largo reactor coolant system pipe breaks; and experience has shown us that large pipes can and do occasionally break.
A loss of coolant results.
In this sense the various accidents considered in the MARCH code com-puter runs by Sandia and Brookhaven National Laboratory are cred-ible.
The probability of occurences is a separrte matter.
"Very, very remote" addresses likelihood, not possibility or credibility.
We find the accidents discussed in CESG Exh. 40, 40A, and 59 to be credible.
This is not to say that they will happen in the life of the McGuire plant, or that they will, if they do happen, neces-sarily result in steam generation.
But they may.
l 19*
Applicant has intimated that the Commission order of Septembeg 26, 1980 on Docket No. 50-289 (restart) required an intervenor to demonstrate a credible accident.
There is no language in the order of this nature.
There is a discussion which relates "cred-ibility" to likelihood:
20.
"There is ac limitation in Part 100 regarding the types of l
accident that may be analyzed." (p. 3) 21.
The capacity to assess the likelihood of acciG its by Staff has proved remarkably deficient.
The multiple failures at TMI have been viewed so unlikely as not to be credible (NASH-1400).
l l
, ~ _. _ _. _, _,.. -_
r 8
The fire mechanism at Brown's Ferry had been given no considera-tion; i.e., it was taken as completely unlikely.
The Commission is aware of the problems of defending against possible acciden:
consequences, although in the absence of actual experience certain accidents seem highly unlikely.
Hydrogen release in excess of 10 CFR 550.44 is a case.in point.
For :he past 18 years the NRC and AEC have implemented Part 100 by postulating an accident that entails release of sub-stantial quantities of fission products from the reactor The quantities postulated could only occur if there core.
was a substantial ECCS failure.
This arguable inconsistency between Part 100 practice and other NRC and AEC rules intended to assure adequate ECCS design has long been recognized.
(p. 3)
The Commission tolerates this arguable inconsistency because 1:
recognizes its responsibility for anticipating and averting pos-sible accidents of severe consequences
- despite the administrative intent, effectut td through a program of engineering design, reg-ulations and I&E, to make such accidents extremely unlikely.
In In the instant matter this Board recognizes that the physical pos-sibilities must, as CESG maintains, be examined.
22*
The Staff received burdens both in regard to reviewing Appli-can:'s response to Contention 1, and directly in Contention 2, in which the ulnerability to hydrogen combustion pressure of an i
ice condenser containment is compared to that of a dry containment.
l
- I: is because an accident can be conceived and credenced that design standards are arrived at.
It is the object make :he probability of credible accidents acceptably low.of design :o is only through experience that it can be established what the I:
actual likelihood is.
Given the short existence of nuclear cpe' ration and the long odds against the accidents embodied in design objectives, and the uncertainty as to achieving these l
design objectives, all accident probabili:y estimates mus: be viewed as having a high level of uncertainty.
23.
It is-the Board's burden, given a record of the possible consequences of physically credible accidents, for which it is not presently possible to assess an actuarial probability, and in which small differences of estimated probability are of doubt-ful utility in this single instance, to determine whether the health and safety of the public are subjected to an unacceptable risk.
24.
This Board finds that the burden of demonstrating accident credibility does not fall on CESG.
CESG has done no more than bring to the attention of the Board and the parties the possible accidents already deemed credible by the Commission which would have the potential for causing hydrogen releases of a significant quantity in terms of Part 100 considerations.
25.
Accident probabilities have been calculated for the Sequoyah I
P ant (CESG's exhibit 61 and discussion in transcript) which are l
l both more likely and more serious in consequences than S D.
In 2
this example and in others, we find the work done on Sequoyah safety to be the best evidence available for !!cGuire, a similar ice condenser plant.
e e.
-w
-e
..iYDROGEN GENERATION 26.
The Applicant's model of a base case accident, as mitigated by an interim deliberate ignition system (IDIS), based on MARCH and CLASIX codes, would develop a maximum pressure of 16 psig, Exh. 5b, Table 6, Fig. 1-6, and burn 350 lbs. of hydrogen of the 1550 lbs. released over a period of 3600 seconds (60 minutes).
97*
The Applicant has a two-pronged solution for dealing with this small loss of coolant' accident (LOCA).
(1) There will not be a repeat of TMI because:
- a. Operators will respond appropriately;
- b. New instrumentation will provide adequate information to prevent the errors in judgement made at TMI.
(2) There will not be a repeat of TMI because even if a PORV (pilot-operated relief valve) sticks open and the ECCS does not operate :
- a. The McGuire steam generators hold more water than TMI's, and more time will be available to act;
- b. Both recirculating fans will operate;
- c. Both containment spray systems will operate;
- d. The IDIS will operate to mitigate the possible effects of the hydrogen release;
- e. The CLASIX base case, JVD12., will be played out;
- f. Containment pressure will at no time exceed 16 psig; 1
l
- g. The ECCS will come on in time to prevent core slump, i
which is estimated at 110 to 120 minutes.
Ecuicment Asoects l
l
~3*
l In Applicant's base case it is assumed that the ECCS does not initially operate.
If the operators have been trained to not turn
. off the ECCS prematurely this failure to operate means:
(1) The injection pumps are out of service or (2) Onsite power has been lost.
(3) If onsite power loss has made the injection pumps inoper-able, then the atmosphere recirculation fans, the contain-spray systems, and the igniters will not be operable.
ment (4) If onsite power is restored, the time of restoration will be critical.
Rel' eased hydrogen will not be circulated and will be far from uniformly distributed.
Detonable con-centrations may have developed in the ice condenser (IC).
The baskets from which ice will have melted form a grid system.
They are formed from sheet perforated by square holes.
(5) The large SG capacity can affect the time scale of the accident, but not necessarily in a significant way.
Boil-dry will take longer and reactor coolant heat and pres-sure increase, causing PORV operation, will come later if, as in the case at TMI, feedwater loss initiates the accident sequence.
If the initiating event is a pipe break the SG capacity beccmes immaterial.
(6) Core and reactor cooling, if restored in the later stages of the accident, may result in a low enough pressure in i
the reactor coolant system (RCS) to such in containment l
l air and result in hydrogen ignition by an uncooled fuel l
rod.
l (7) A power outage of two or more hours will result in core slump and the rapid discharge of hydrogen, steam and fission products into the containment (Exh. 5a, Table 3).
fi -- -
-. ~. -
. 29.
The Applicant has chosen to defend against one specific LOCA, the S D.
There is a large universe of LOCA's involving small, med-2 Lum, and large pipe breaks which do not involve stuck PORV's.
~
This makes Some of these,are viewed as more probable than S D.
3 them equally worth considering in that even an actuarial probab-ility, as distinguished from an estimated probability,,provides
' no information as to when a specific event will occur.
It only speaks to what the timewise incidence will be in a sufficiently large sample -- a sample which is not realizable in the matter of nuclear power plants.
30.
In a number of small LOCA's hydrogen evolution is more rapid than in S D.
(The Board has, since this hearing, decided to take 2
official notice of the BNL report as it is simply reporting ad-ditional computer runs using the same MARCH code as relied on in l
DPC Exh. Sa and 5b, and in CESG Exh. 40 and 40A.)
This would make the situation more sensitive to outage time.
This is not l
I to say that the Board accepts as fact the projections of the MARCH and CLASIX computer codes.
31.
In selecting S,D as a model accident and using the MARCH l
I and CLASIX codes to compute the consequences for various cases the Applicant set itself a problem it could solve.
There is noth-ing in CESG's contention 1-1 that can be construed as specifying that the Applicant need only defend against an approximation of the TMI accident of its choosing.
Nor, in the Board's mind is it sufficient, particularly in the case of a McGuire thin shell con-tainment, designed on the basis of a 15 psig accident, to consider l
only one very specific circumstance in which hydrogen is released l
I and its effect mitigated by deliberate burning.
It is the history
. of the industry that few accidents have been repeated.
There has been but one Fermi core melt, one Browns Ferry fire.
It is likely that some of the lessons of TMI have been adequately learned; that the fear of " going solid" has been placed in per-spective so as to not cause injection flow to be improperly curtailed; that the thermohydraulic state of the RCS with re-gard to subcooling vill be clearly indicated.
However, the most basic lesson which TMI had to teach is that neither licensee nor NRC staff anticipated what did, in fact, happen.
The Board's obligation in this proceeding is to consider what can happen and to determine if an adequate defense has been made against such occurrences.
32.
The Board questions certain of the assumptions made in the computer programs on which Applicant and Staff rely.
The burn-ing behavior of hydrogen, as the record amply shows, depends on a number of variables:
concentration, oxygen concentration, ad-mixture with air and steam, ignition source, container morphol-i ogy (CESG Exh. 59, NUREG/CR-1762, pp. 7-240).
The base case assumes homogeneous compositions in the LC, IC, and UC at the mcments of ignition.
Ignition occurs simultaneously for a combustible composition taken to be distributed in a regular l
way in a particular compartment.
l 33.
The first ignition in base case JVD12 (DPC Exh. 5b, Fig. 6),
at 10 v/o hydrogen, occurs at 1420 seconds (23.7 minutes) after
' hydrogen release begins (id. Fig. 1).
The most rapid rate of j
hydrogen release is at about 1272 seconds (21.2 minutes).
The i
circulation rate of 60,000 cfm (id., Table 4), would require h
6 20 minutes to move 1. 2 X 10 cf, the volume of the containment.
It would clearly require a number of cycles to fully mix a hetero-geneous containment atmosphere.
Yet, just 2.5 minutes after the peak rate of hydrogen release a concentration of 10 v/o has been reached at the igniters and results in total burn of the hydrogen in the LC.
We do not find this unrealistic assumption acceptable.
Yet the entire detailed projection, the individual peak temperatures, pressures, and times, is based on it.
34 It is established by the record that above certain concen-trations steam inerts hydrogen / air mixtures (CESG Exh. 59, Fig.
10), although there is some difference among expert witnesses as to the precise concentration (Tr.
).
It is no more realistic to expect that the steam distribution in the LC will be uniform than the hydrogen distribution at the oresumed time of first ignition.
Whichever igniter at the 763' level first is contacted by the combustible mixture (DPC Exh. Sb, Fig. 3.3.1, 3.3.7)* it is altogether unlikely that combustion will propagate throughout the lower compartment effectuating an 85 lb. burn (id. Table 6) resulting in the temperatures and pressures shown (id. Fig. 1 and 5).
- The igniters at the 763' level are those closest to the lower IC doors and in the mainstream of recirculating con-tainment atmosphere.
15-35.
It is Applicant's position that only the base case is credible and that the maximum differential pressure to which the containment may be exposed is-16 psig (Tr.
).
Staff accepts Applicant's position (Staff Ex. K). CESG's proffer of testimony in this area was not admitted by the Board at hearing.
The failure of CESG to offer a witness the Board required as qualified does not, however, relieve this Board of the obligation of making its own determination.
As in the matter of the need for power before the Catawba ASL3, it is not the role of the Board to be an umpire, calling balls and strikes, but rather to be a determiner of fact so as to make an informed and reliable decision based on a complete record, (Initial Decision, Docket Nos. 50-143, 50-414, June 30, 1975, pp. 53-55, 1 103, and fn 39).
This Board has undertaken to add to the factual record by appropriately noticing documents, Appendix 3.
It has further undertaken to reason through some of the problems insofar as present knowledge and the record make it possible.
O e
~
,&-w,-
-,,.,,.,..e.,---,--+-.,-%,---..--,,,_,_,-%-._,%,,..,__#,,,.,.,
,,,-vwy.
,.,., - -. -,, - -. -, _.,., -ey---.
,, - -, -, -g,
,w,
,,--,-,--.-r*,,m
,,,w,.,.--,y,yw w w -, -
. 36.
The mixing of hyd. ogen, steam, and air in the LC, whether the recirculation. as are on or off, is a matter of critical importance in regard to the when and the where of deliberate ignition.
37.
Questions in regard to the mixing of the containment atmosphere were neither addressed nor raised by Applicant nor Staff.
The combustion of uniform mixtures was assumed.
The uninvestigated status of the mixing question is reccgnized by investigators (The Behavior of Hydrogen During %:cidents in Light Water Reactors, NUREG/CR-1561, Mixing of Hydrogen in Containment, pp.
183-5).*
38.
In Applicant's base case, hydrogen release from a stuck PORV would vent from the pressurizer reliefs tank rupture discs.
T he pressurizen relief tank is located between.70 and 110 of are.
(FSAR Eig. 1.2.2-10, PSAR Fig. 1-16).
Mecirculated air enters the LC through 6 hatches, 4 of which issue into the reactor c ompartment at positions of about 335, 25 155. and 205 l
There will be an interface between the steam and hydrogen 0
entering the region of 70-100 and the air entering at the four hatches.
The locadon of this interface will depend on the relative volumes of flow of these two components.
T he recirculation flow volume is 60,000 cfm.
At 4752 sec onds ( 21.2 min. ) prim-to first postulated ignition in tne base case, steam release is c omputed as 48.4 lb./ min. : hydrogen release as 1.07 lb./ min. (DFC axn. 5a, Tables 1 and 21 At t b2 The board takes official notice of Sandia Laboratory's :!UREG/CR-1561.
The Board is of the opinion that the Staff shculd, early in this proceeding, have brought this basic report to the attention of the Board and the parties.
There was ample time for so doing.
The date of the report is August, 1000.
computed. temperature of 771 F, the corresponding volumes of flow are 27,800 cfm hydrogen, 139,600 cfm steam, for a total of 167,400 cfm, about 2.6 times the flow of recirculated air.
It is obvious that the ice condenser doors in the are 25-155 will be venting steam and hydrogen to the ice condenser.
Due to the condensation of the steam, essentially pure hydrogen will vent fro the top of the IC and, due to its very low density, as well as upward momentum (DPC Exh. bb, Fig. 3 3.1 (mis-numbered] Containment--
Plan View [actually an elevation]), rise to the top of the containment.
The IC doors near the ends of the IC, at about 0
230 and 310, will rcaalve essentially unmixed recirculated air.
Somewhere between, based on the relative volumes of recirc-ulated air and RCS steam and hydrogen discharge, in the vicinities 205 and 335, there will be a mixing zone, in no way defined in the record, which will be in port combustible and may be ignited by the igniters at 774'0" elevation (DPC Exh. Sb, " Fig."
lactually Table] 3 3 1, Glow Plug Locations) IIHEIB-33,34, 35,
& 36 The steam inerted portion of the transition zone will be steam stripped in the IC and may be ignited by IEHMTB-52, 53, 54
& 55.
Combustion will not propagate to the hydrogen rich zone in the IC as 90 of arc is approached because the oxygen level necessary to support combustion will not be present.
- 39. This detailed consideration of a moment 2 5 minutes before hydrogen combustion, according to the base case, is supposed to
' start makes clear the deficiencies in the homogeneous composi!Aon model.
As the accident progresses, the absolute and relative volumes of steam, hydrogen, and containment atmosphere will change, the mixing zone will change is size and location, and the patterns and intensities of turbulence will change.
- 40. As 1) the entire period of hydrogen release in the base case is only about 60 minutes; 2) as the upper containment, supplied with the gaseous effluent from the IC will be most hydrogen rich in the vicinity of 90 of arc, and most air rich between 230 and 310 in the region where the air recirculation fan intakes are located (DPC Exh. 6); and 3) as both hydrogen and steam emissions undergo great changes during this time (DPC Exh. Sa, Tables 2 & 3, CESG Exh. 40, Fig. 17, 18, & 20), in which 4) only 3 times the containment volume has recirculated, it is apparent 5) that the containment atmosphere will be least homogeneous during and i
subsequent to the rapid release from the RCS and will only slowly move away from this extreme heterogeneity as the steam and hydrogen release rates decline.
The contribution to mixing by the upper containment spray sys, tem will, of course, depend on whether there is onsite power.
RCS release rates do not decline until 6960 sec-onds into the accident (116 min.).
The dynamics of the accident process are as conducive to the heterogeneity of the containment atmospherecompositionasthegeo[metryofthestructure.
4, Once it is comprehended that uniform burns will not occur in l
1 the LC or IC, the detailed computer projections must be rejected as invalid.
This Board finds the MARCH /CLASIX projections invalid.
This applies to the sensitivity studies as well as the base case.
- de find that the record is incomplete as none of the parties hac
' providad a credible and realistic study or, which to base findings as to the temperatures and pressures which will develop in a small LCCA in the absence of ECCS, whether with or withouc an igniter system.
- 42. A limited, but not literal use can be made of Applicant's sensitivity studies (DFC Exh. Sb, 2.2 4).
Witn nalf the atmospheric recirculation rate, LC hydrogen concentrations will be higher than for corresponding times in the base case.
The more hydrogen released before ignition, the higher the peak temperature and pressure will be.
When ignition occurs later than in the case case, 1420 seconds (23 7 min.) after initial hydrogen release, the time is 2220 seconds (J/ min. ) for case JVDC'/, a single peak as high as 117 psia with burning occurs simultaneously in LC, 10, and UC.
This kind of result can occur if,due to power outage and
-restoration,the igniters were not activated until late in hydrogen release.
Neither Applicant nor Staff has addressed such a credible
~accident scenario.
- 43. In the absence of onsite power, the most likely cause for an inoperative ECCS, recirculation fans, hydrogen skimmer fans, spray train pumps, and igniters would not operate.
Dealing witn tnis circumstance has not been addressed by Applicant or Staff.
- 44. Confining consideration to steam and hydrogen release, a matter for which the MARCH code is qualified, the Brookhaven study (CESG Exh. 59) indicates a serious alteration of consequenecs in terms of steam release. depending on the time of initiating (or reinitiating) coolant injection.
In case no. 3, an 3 D, burn 10 v/o to 0 v/o, 2
power restored at 89 minutes, a pressure peak of 62 psia is computed at 110 minutes.
Sandia MARCH computations indicate that dangerously high hydrogen concentrations would develop in the LC and UC by 100 minutes (CESG Exh. 40 A, Fig. 17 & ld).
- 45. Core slump has not been considered by Applicant for what appear 9d to be inadequate reasons, the events associated with core slump are somewhat uncertain (Tr.
).
Sandia computes core slump as occuring in a little over 90 minutes for S D 2
(CESG Exh. 40 A, Fig. 17, 10, 19, & 20).
For the large break, AB, sequence, core slump begins at 22 5 minutes (id. Fig. 22, 23, 24, & 25).
For core alump and 10% burn in the AB sequence a peak pressure greater than 150 paia is computed for both LC and UC at 25 minutes.
In BNL's studies the lowest pressure generated by a hydrogen burn at core slump is for an S H with 2
steam break, case no. 7, 39 psia.
The highest pressure, calculated for case no. 15 for a TMLB is 115 psia.
The average for the 14 cases considered is 71.5 psia.
- 46. The airconium/ water reaction although a major and rapid source of hydrogen generation is not the sole source.
Nadiolysis, water reaction with other metals, iron sinc and aluminum, may contribute significant mmounts of hydro 6en over a period of time.
[
l However, this Board has decided to not consider hydrogen release mechanisms which would become significant only after core slump.
Such matters reach beyond the thrust of contention 1.
- 47. Applicant's witness testified to the efficacy of jets as miXi"8 devices (Tr.
).
The necessary conditions for jet mixing are not present in the base case.
The venting from 1
the pressure relief tank rupture discs will be in an atmosphere df steam and hydrogen.
The steam and hydrogen jets will lose force and directionality in the confined and irregular structure of the LC (13 Ali, Fig. 1.2.2-10 and ?3AR, Fig. 1-16).
Jet mixing will, during periods of more than 25-305 of the maximum discharge are rate, when releases /at the same rate as the air recirculation rate, not occur.
The basic condition for jet mixing is that the jet operate in a plenum containing the second fluid, a condition which will only be realized at periods of minimal steam and hydrogen release rates.
- 48. There are three FORV's set for 2335 psig opening and three safety valves set for 2465 psig opening (FSAR Table 5 2.2-2).
There is also a pressurizer spray nozzle.
This requires 7 small pipe connections.
The failure of any one of these would lead to a LOCA which would be approximately equivalent to the base case, except that it could not be corrected by closing a block valve.
As the fundamental assumption in the base case is the loss of the ECCS, any small break accident is at least equally a candidate for consideration.
Depending upon the location of the break, a new set of specifics in regard to mixing and ignition would arise.
49.
The CLASIX model for the combustion of hydrogen generated in a LOCA is invalid, lt assumes a uniform mixture of hydrogen and oxygen in the igniter regions which will not be realized.
50.
The values of pressures, temperatures and time put in evidence as the Applicant's base case are invalid.
For purposes of comparison the accident sequence and assumptions of the base case result in the lowest pressure increase in a total of 30 cases considered by Applicant, Sandia Laboratory, and Brookhaven National Laboratory.
The Applicant has unduly limited the con-sideration of accident sequences in holding that only its base case is credible.
51.
The most rapid evolution of hydrogen durin6 a LUCA without ECCS occurs, according to the MARCH code, at core slump.
Applicant has not considered any instances of core slump nor provided a credibie reason for not so doing.
52.
Applicant has sought to apply good methods of operator selection and training.
Uperator availability is limited. A full complement of operators for five d hours shifts is not available.
Twelve hour rotating snirts cause more adjustment problems than d hour shifts.
First year operation is accompanied oy two to three times the error incidence in operation as later operation.
Undermanning and 12 hour1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> shifts offset the benefits of improved training.
Operators will have only had about 3 months actual operating experience.
The shift technical advisors will not meet the educational requirements of NUH5G-0737 App. G.
Applicant is not ready to operate with the unresolved hydrogen l
l control and mitigation questions that exist.
53.
The loss of onsite power is not susceptible to minimization by operator action.
34 Applicant he failed to demonstrate that substantial q'uantities of hydrogen in excess of the design basis, 10 CFR 250.hu, I
will not be generated.
55.
Applicant has failed to demonstrate that the.'cGuire containnent has sufficient strength to withstand pressures that L
can be generated by hydrogen release and combustion.
, Ooerational Aspects 56.
The studies made of the TMI-2 accident indicate that, what-ever the contributing causes initiating the accident, once under way the cutting back of co oling injection made possible core uncovery and hydrogen generation.
Instrumentation which assures i
that the primary system ts subcooled and the level of coolant in the reactor is sufficient.to cover the core, in combination with l
l operator performance which assures the maintenance of subcooling f
by avoiding premature termination of coolant injection, should materially reduce the chance of a repeat of the TMI-2 accident.
57.
McGuire Unit 1 subcooling instrumentation has been in place
[
for a year and has functioned satisfactorily over a three week test period (Tr. 3009).
The reactor level-indicating equipment is to be installed by January 1, 1982 (Tr.
).
It is intended i
to provide further indication of the approach to an inadequate core cooling situation (Tr. 3016).
58.
The reactor operators (RO's,' and the senior reactor operators (SRO's) will receive about 2% years of training (Tr.
).
Oper-ators are required to have high school diplomas, though higher I
degrees are not a basis for exclusion (Tr. 2875).
An RO may l
l advance to SRO by 12 weeks of training, 3 months of observation, and passing an NRC examination (Tr. 2883).
An SRO may advance to shift technical advisor (STA) by four weeks of simulator and classroom training (Tr. 2921).
This training program includes instruction in "STA responsibilities and accountability; manage-ment and supervisory skills; transient and accident analysis; plant chemistry; PWR heat transfer; PWR thermal and hydraulic
. transient response; small break LOCA analysis; and mitigating reactor core damage."
(Tr. 2921, Testmy. 15A).
- 59. The role of the STA is primarily that of accident assessment and analysis.
The training program for STA appears to contribute little if anything beyond what might be expected for an RO or an SRO.
Certainly the four-week program does not begin to approach the 520 hours0.00602 days <br />0.144 hours <br />8.597884e-4 weeks <br />1.9786e-4 months <br /> of College Level Fundamental Education prescribed in NUREG-0737, App. C, 6.1.2.
The record gives no assurance that this or other requirements will be met by Applicant's STA's, supra 6.2, 6.3, 6.4, 6.5, 6.6, 6.7, 6.8.
There is no assurance that Applicant's STA's will be qualified to carry out the objective of reducing the probability of abnormal or emergency condition occurrences and mitigating the consequences of these conditions if they do occur.
- 60. Presently there are four operating shifts.
Each shift is manned by unit coordinators, a shift supervisor, an assistant shift supervisor (SRO), 2 licensed operators, and 2 unlicensed operators.
The plan is to have five eight-hour shifts (Tr. 2948).
Shifts of up to 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> in 24 may be worked for a total of 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> per week.
No testing has been done to determine the long-period an operator may work safely and efficiently (Tr. 2951).
est Perhaps the best measure of Applicant's witnesses' confidence that premature termination of ECCS operation is not credible is to be found in their choice of language in defining " credible".
Mr.
Canady uses "believe" three times in 15 lines (Tr. 2993, 11. 14-18).
Mr. Barron (Tr. 2994, 1. 4) uses "believe" once and Mr. Reed uses it twice (Tr. 2994, il 12-15).
No one was able to testify "I know".
61.
CESG raised a number of questions in regard to human fac-tors and the expectations in regard to the performance of oper-ators in an accident situation.
The importance of "over-learn-ing" (Tr. 3630) as a means of avoiding stress-induced amnesia (Tr. 3636) was testified to.
Distributed learning experiences were seen as superior to concentrated training (Tr. 3627-28).
Applicant appears to use a mix of repeated classroom, simulator, and on-the-job training, a'lthough the record does not indicate that Applicant has developed criteria for over-learning or, other than examinations and job performance ratings, means for assessing the effectiveness of its program.
62.
In regard to the 80% average passing grade for NRC R0 and SRO examinations, over-learning slows the loss by forgetting of test performance (Tr. 3631-33).
63.
Conceptual learning was differentiated from rote learning and distinguishes the expert from the non-expert (Tr. 3638).
Operator training, and particularly STA training, should be di-l rected to conceptual learning.
64.
In addition to the short term effect of inducing amnesia, continued stress causes deterioration in the capacity to perform (Tr. 3981).
Boredom also is an enemy of emergency performance (Tr. 3847-48), the longer the period in which there is no stim-ulus, the more likely a signal will be overlooked.
Again, over-training provides a compensation.
To determine Applicant's sit-i uation in regard to these matters, a job survey by qualified human factors experts would need to be performed (Tr. 3850-51).
The record does not show that the data has been acquired and the study made.
~. -
. 65.
Applicant called reouttal human factors witnesses (Tr.
4715-5841).
It was the opinion of these witnesses that neither boredom, information overload, fatigue, or stress would have any adverse effect on the safety of operation (Tr. 4745-48, 4754, 4764).
As this Board was deprived of the opportunity of exam-ining CESG's psychological witnesses, it sought a second opin-ion and found one in NUREG/CR-1764, a study of human factors in operators.
This study bro.ught to the Board's attention the role of the Circadian rhythms in job performance and the adverse effects on the performance of persons working rotating shif ts, particularly 12-hour shifts.
The Board thinks it is essential to supplement the record in this regard.
66.
The number of operator errors in a plant changes with the age of the plant, NUREG/ CR-1764, Taole 7, a document which the Board officially notices.
Operation errors and total errors, including maintenance, installation, and administrative are about 3 times as high in the first year of operation as in the r
l third year.
Although Staff argues, for example, in permitting it operation under an interim hydrogen rule for a year or so, is precisely the initial period of operation which should be most carefully safeguarded to compensate for new-plant, new-team errors.
We note that most operators will have only had about three months actual operating experience when McGuire is scheduled
(
co start (Tr. 3025).
67.
The academic psychologists, although having no familiarity of with Applicant's operation (Tr. 4672) also have no conflict interest in that they are neither employed by the nuclear in-dustry nor the environmental movement.
By calling the attention l
of the Board to group-think, cognitive dissonance, reverence for l
. authority and risky shift matters (Tr. 3854-58, 3837-42), all of which address the mindset problems reported by the investiga-tors of the TMI-2 accident, the question of mindset in the instant matter is raised.
The behavior of the parties is illustrative of these observations.
Applicant's witnesses and counsel are in total agreement as to confining their analysis and mitigating measures to a single accident scenario.
Applicant's witnesses provide unanimous assurance that common psychological conditions, boredom, fatigue, stress, cannot affect the safety of operatt~on.
Staff also speaks with a coherent and consistent voice -- we have identified problems, called these to the attention of the Applicant and the Applicant has come up with acceptable solutions.
The one faintly dissonant note was provided by the witness from Sandia, but he appeared at the Board's request, not at the Staff's initiative.
The Board is persuaded that the testimony submitted by Applicant and Staff witnesses has been influenced by group-think, deference to authority, and has resulted from cognitive dissonance, i.e., an ignoring of items of information and inter-pretation which do not support their position and a shift toward accepting a greater level of risk than would have been arrived at had the individual parties been acting with complete independence.
68.
The Board is not reassured that Applicant's unit coordinator l
does not know, even approximately, the number of control panel indicators to which an operator must respond and the controls which he must operate (Tr. 3027).
The Board is concerned about an operator selection and training program in which 10 to 12 cand-l idates take an SRO examination but only 7 pass.
In view of the
fact that STA's are to be college-level material and are drawn from SRO's and RO's, it concerns the Board that Applicant has only a high school diploma requirement for operator candidates and not even a quota for college-trained candidates.
It is a concern that none of the managers testifying were certified safety engineers or fire protection engineers (Tr. 2846, 2847, 2850).
It is a concern that the lag time for response to an emergency situation was a matter of peripheral knowledge rather than active investigation (Tr. 2869).
The Board is concerned that Applicant, having considered the matter of dealing with station blackout (Tr. 2888 1. 22, 2889 1. 3) opposes testimony dealing with it on the grounds credibility has not been demon-strated.
Applicant's procedures for dealing with blackouts are prima facie evidence of credibility.
Training for responding to large LOCA's is sufficient to attest to the credibility of such accidents (Tr. 2887, 11 22-25).
The 3oard is concerned that there has been no training for use of the mitigative system although it is in place (Tr. 2959, 11. 8-11).
69.
The Board agrees with CESG's witnesses that psychological factors enter into operator selection, training, and on-the-job performance and finds that such factors have not been actively taken into account in the Applicant's training program, and in ignoring the requirements of NUREG-0737, App. C for the qualif-ication of the critical new person in the operating roster, the shift technical advisor.
In regarding these factors as relevant in an individual's performance (Tr. 4765) we find that a study of Applicant's selection and training procedures and operational environment by a qualified independent third party is essential to messr_e_the_ safe ogeration of McGuire.
HYDROGEN COMBUSTION 70.
CESG's Contention 1-2, that Applicant has not demonstrated that hydrogen, if generated, will not combust, has been in part mooted by Applicant's choice of deliberate ignition as a means of mitigating hydrogen release.
71.
In the event that power is restored after an outage during a LOCA, the igniters could intensify the hazard by touching off an accumulated, high conce'ntration of hydrogen.
In such circum-stance-the use of igniters should be conditioned by firm informa-tion as to the hydrogen concentrations of various regions of the containment (Tr.
).
72.
Until the Applicant has demonstrated a program and instrumen-tacion to avoid this possible circumstance, operation of McGuire will not be acceptable.
I i
i l
~
j l
t CONTAINMENT STRENGTH AND MITIGATING ALTERNATIVES 73.
The matter of containment strength is given special signif-icance by the uncertainties in the release of hydrogen, and the role of the mitigation scheme.
In the worst credible case, the metal-water reaction would go to completion, the containment at-mosphere would contain about 30% hydrogen, which on deflagration could develop a pressure of about 13 atmospheres, or 190 psia (R & DA, H.W. dubbard to Victor Gilinsky, 4 August, 1980, offic-Lally noted).
None of the estimates of McGuire containment strength indicate recention of integrity at such pressure.
Ap-plicant calculates a 1 in 2 chance, zero standard deviations, of survival at 67.5 psig.
Staff calculates a 1 in 769 chance, plus three standard deviations, of surviving 120 psia, and a corresponding i in 769 chance of not surviving 48 psia.
74 It has long been the practice of the Commission and of its predecessor to consider the worst credible case in establish-ing design criteria.
A Condition IV Limiting Fault for which McGuire by regulation was designed, the " Double-Ended Rupture l
l of the Largest ripe in the xeactor Coolant System" ts defended against (xESAR 14.4-1).
~1he threat agatnst the containment sys-l tem of a substantial hydrogen release is,.like a double-ended break of a large pipe, not anticipated to De an event of a nign probacility.
Just as the chance of a large pipe break ts very
- In DPU Exn. 50, Table 8, the ourning of 1200 pounds of hydro-i gen of a possible 1900-2000 pound total, gives a calculated pressure of 117 psia.
Scaling up to the full possible burn I
results in a calculated pressure of 185-195 psia.
remote, but in the interest or puolic health and safety must be designed against, the possibility of containment breach by a hydrogen explosion must, in strict parallel, be designed against.
75.
There are a number of options open to the Applicant:
(1).
The strength of the containment may be increased so that in the event of the most severe hydrogen combustion event, the requirements of 10 CFR 550 App. A-V Criteria 50 and 51 are met:
Criterion 50 -- Containment Design Basis.
The reactor con-tainment structure, including access openings, penetrations, and the containment heat removat system shall be designed so that the containment structure and it internal compart-ments can accomodate, without exceeding the design leakage rate and, with sufficient margin, the calculated pressure and temperature conditions resulcing from any loss-of-cool-i 1
ant accident.
This margin shall reflect consideration of (1) the effects of potential energy sources which have not been included in the determination of the peak conditions, such as energy in steam generators and enerav from metal-from de-water and other chemical reactions that may(result
- 2) the limited graded emergency core cooling functioning, experience and experimental data availab?e for defining accident phenomena and containment responses, and (3) the conservatism of the calculational model and input parameters.
Lemphasis suppliedj Criterion 51 -- Fracture prevention or containment pressure boundary.
The reactor containment boundary shall be designed with sufficient margin to assure that under operating, main-tenance, testing, and postulated accident conditions (1) the ferritic materials behave in a nonbrittle manner and (2) the probability of rapidly propagating fracture is minimized.
The design shall reflect consideration of service temperatures and other conditions of the containment boundarv material during operation, maintenance, testing, and postulated accident conditions, and the uncertainties in determintne (1) ma-terial procerties, (2) residual, steadv state, and transient stresses, and (3) size of flaws. Lemphasis supplied]
(2) The atmosphere may be inerted.
This measure has been re-These reac-quired of operators of Mark I and Mark II reactors.
tors have about 257. of the volume of an ice condenser containment and are more readily inerted.
Unlike ice condenser containments l
they do not require semi-weekly door frost removal.
An inert atmosphere poses problems in worker protection.
(3) A deliberate burnoff system of demonstrated reliability and effectiveness may be used.
It is clear that such a demonstra-tion is presently lacking.
The testimony of Sandia's witness indicates that under certain postulated conditions the nonselec-tive use, i.e., just turning on of the glow plug igniters, may be
" fraught with danger". (Staff Exh. E)
(4) Other ignition suppression systems including halon and fogging have been proposed.
The use of halon seems doubtful on a number of counts.
A fogging system does not, in fact, appear yet to exist.
Other systems include filtered vent and rapid inertion by operation of a jet engine. (CESG Exh. 40; RDA-TR-178700-0C2 76.
c.e state-of-the-art for mitigating the effects of hydro-gen release is :learly in an early stage of developmer.
The effectiveness, cost, and reliability of reasonable mitigation alternatives are not yet known.
The choice before the Board is whether to permit the use of the mitigation system which Applicant has installed and Staff approves, but about which sertous sarety questions have been raised, or to withold authorization to operate.
Given the established standards of safeguarding against accidents of presumed low but unquantified probability, but of severe con-sequences, the Board must, based on the substantial uncertainties developed in this record, deny authorizatien for McGuire operation.
- A draft of Sandia Laboratory's " Analysis of Hydrogen Mit-igation for Degraded Core Accidents in the Sequoyah Nuclear Power Plant" was provided to the Board and the parti.es just orior to commencing the hearing.
R & D A's Report (Staff Exh. M) was provided the Board and the parties March 13, 1981.
33-77.
With regard to containment strength, Intervenor offered testimony that tanded to show that there was a possibility that a faulty weld grinding job remained uncorrected.
We find the view of quality control offered by Mr. Lanford to be eye opening, and find that the fact tha't his report was not responded to when made, nor read by those personnel at Applice,c responsible for quality control until the week before they testified, appalling.
78.
Neither Staff nor Applicant were able to respond to the specific situation pointed out by Mr. Lanford.
Both made general references to quality control.
79.
We must find that there is an uncorrected grind, more than 2 la inches from a weld, scoring the steel 1/8 inch deep, which has a small, but real effect on the strength of the containment.
O e
~. - - - -
---.4-
--_,,.,-,--------y,__,n.
_,.,,.._.,,,,.,.c-,--...-yw,,.
.,,,w--,,
...y-.
p.--rm-,-
y w--,--
.----,,y-,,...,%.p.---,--,-,r-e,---
, POLYURETHAUd IUSULATION 80.
Polyurethane insulation is used between the ice condenser chilled air ducts and the containment wall (Staff Exh. O, P).
Polyurethane totally volatili:ec at a lower temperature than the hydrogen flame temperature characteristic of burning 8.5 v/o hydrogen (Tr. 5110, CESG Exh. 40 A, Fig. 2) of 1600 F.
The gases formed are combustible and will leak into the containment (Tr. 5127, 5130, 5131).
There are 27,000 lb. of polyurethane in the containment (Tr. 5136).
81, The combustion of 27,000 lb. of polyurethane would result in a volume of 900,000 cubic feet of gas measured at standard condi-tions (O C, 14.7 psia)(Affidavit of Jesse L. Riley, Mar. 27, 1961)?
The amount of oxygen initially present in the containment 13 only sufficient to burn about 9,000 lb. of polyurethane.
The nat volume of the combustion gas would be 100,000 cf at standard conditions.
If the heat of combustion were removed, the presence of this gas would cause a pressure increase of about d5 in the i
l containment.
If, instead of burning, the polyurethane were to 1
completely pyrolyze, i.e. gasify, it would result in 250,000 cf l
at standard conditions and an increase in containment pressure of about 205, about 3 psi.
l 82.
Applicant and Staff recognize that polyurethane is combustible and can pyrolyze.
The matter to be resolved is how much pyrolysis ic 2oard admite the dilcy Af fidavit of On. 27, l'JUI, to the record.
Applicant, cpposen the admission, April J, 19"1.
ib o affidavit is, in effect, a response to Staff's filing on the same subject.
. pyrolysis and/or combustion can occur if there is a steady burn of hydrogen / air in the IC.
The 250 lb. estimated as combustible
- 7 Applicant's witness would add 3x106 Btu's to the containment, small compared to the hydrogen burn contribution of 80x106 Stu's (Tr. 5116, 5119).
However the potential contribution of polyureth-6 ane ccmbustion is one third of 324x10 Etu's. greater than the hydrogen contribution.
83.
Hydrogen can burn at richer concentrations than 8.5 v/o.
Sandia's Dr. Eerman is of this opinion (Tr.
).
As the concen-tration oh hydrogen burned increases, so does the flame temperature.
It reaches a maximum of 4150
? at 32 v/o (CESG Ixh. 40 A, Fig. 2).
The melting point of iron is 2795 F.
The structural strength of steel diminishes rapidly at temperatures over 700 ? (Riley Affidavit, Attach. 1).
34 The parties agree that the detail of the construction of the enclosure of the polyurethane foam initially shown by Staff (Staff Exh. 0) does not, in part, apply to McGuire (Riley Affidavit, Maren 27, 19ci, Otarr Arfidavit, Maren 27, 1901, Rasin Affidavit, April 3, 19d1).
85.
If no higher flame temperatures than 1600 F sre encountered, Applicant's view that only a small amount of pyrolysis and no direct combustion will occur has some plausibility.
86.
If flame temperatures are higher, and the physical possibility exists, the galvanised steel sheet form shown as "Y" in Applicant's Affidavit (Hasin, April 3, 19cl) will soften.
The Glastrate insulation sealing strip, 2710956, ficer glass with an crganic binder, will deteriorate and the seal cetueen the polyurethane roan and the reactor side of the the keystone form (::RC Affidavit,
, March 27, 1901, Fig. 1 & 3) will fail.
Tne gases formed by pyrolysis will be released at this defective seal.
The pressure of a large volume of gases will be sufficient to df flect part "Y".
The gases will readily escape into the 10.
87.
The gases escaping into the IC'will ignite anc further extend the region of the "Y" form which is heated.
Up to a hydrogen concentration of 29 v/o there exists a residue of ox76en after combustion (CESG Exh. 40 A,' Fig. 2).
88.
With the deterioration of the Glastrate seal, part 2710956, the opening of a gap between the edges of heat sof tened part "Y" and the downward movement of the hydrogen and the chilled air ducts, flame as the ice melts (Tr.
),
air will be able, by convection, to gain access to the foam, entering the gap as pyrolysis slows.
A furnace fueled by polyurethane will result.
The quantity cf poly-urethane contributing heat and combustion gases to the containment will be, in all likelihood, substantially greater than 250 lb.
89.
The Board finds neither Applicant nor Staff gave consideration to the possible pyrolysis and combustion of the polyurethane insulation as a consequence of the deliberate ignition of hydrogen.
The board, recognizing that ignition temperatures can reach upward from 1600 F and that steel sheet is softened at this and higher temperatures, finds that substantial uncertainty exists as to the magnitude of temperature and pressuce increases which develop in the containment as a result of polyurethane pyrolysis and conbustion.
This adds a further factor of uncertainty to the development of pressure in an ice condeneer containment durint a hydrogen release accident.
37 ACCIDENT IMPACT 90.
The impact of a breach or containment after tne expulsion or fission products by the core is unacceptable.
The MARCH code indicates that a significant release would be underway by 4115 seconds (69 minutes)
(DPC Exh. 5a, Table 3) in the S D scenario 2
considered by Applicant.
The Reactor Safety Study lists nine release categories for hypothetical PW< accidents (NUREG-0534 Supplement, Table 6.1.4-2).
These consider noble gas releases of as much as 90% of the core inventory and iodine releases as high as 70%.
91.
The uncertainties in regard to hydrogen release and mitiga-tion and the fact that it is possible for the peak pressure of burning hydrogen to greatly exceed even the most optimistic estimate of the ultimate strength of the containment require that McGuire not operate until these matters have been satisfactorily resolved.
l I
l l
- SAFETY FACTOR; CONSERVATISM AND PROBABILITY 92.
The initial design basis for the McGuire containment was 15 psig.
At the time of the construction permit proceeding the pressure calculated for the design basis accident was 12.6 psig (PSAR).
It has since been increased'but is still ler;s than 15 psig (FSAR).
93.
The accident at TMI-2 demonstrated that the release and combustion of hydrogen cou'Id result in substantial peak pres-sures.
The peak at TMI-2 was 28 psig.
The combustion of a sim-ilar quantity of hydrogen in a McGuire containment is readily calculated by means of the gas laws:
pvt 1=PY22 0
For TMI, P7 = 28 psig + 14.7, V1 = 2.05 X 10 cf.
6 ror McGuire, V2 = 1.2 X 10 cf., P2 - 72.95 psia = 58.25 psig + 14.7 94, A safety factor is conventionally calculated as the ratio of the average ultimate load to the maximum anticipated load in normal use.
Assuming that the design basis accident of 15 psig is within the scope of normal use, Applicant's ultimate pressure calculation of 67.5 psig leads to a safety factor of 4.5.
Staff's average ultimate pressure of 84 psig leads to a safety factor of 5.6.
These safety factors are evidence of the conservatism the NRC and the AEC have required in the construction of nuclear stations.
When a peak pressure of 58 psig is considered, safety factors of 1.16 and 1.44 respectively result.
For the more severe pressure peaks calculated by the MARCH code for other credible accidents there is no factor of safety.
95.
It is not within the discretion of this Board to abandon the policy of conservatism put in place by the Commission.
It is the burden of the Applicant and the Staff to demonstrate j
conservative safety tactors for the Meuuire containment struc-ture in reference to the pressures which can develop in credible hydrogen release accidents.
It will not.be enough to assert that the probability of an accident sequence is low, or that the likelihood is very, very remote -- this is true of LOCA's gen-erally, for which the Commission requires a defense.
The op-l tions remaining to the Applicant include structural upgrading for the spectrum of accidents which can lead to the generation of significant quantities of hydrogen; the demonstration that the combustion of released hydrogen cannot occur and that the pres-sure of the unburned hydrogen does not constitute a severe chal-t i
lenge to the containment; a precisely controlled deliberate l
l burnoff of hydrogen such that no substantial peak pressure can i
result; or a demonstration that accidents resulting in signif-icant release of hydrogen are a physical impossibility.
O e
l l
l l
. RIPENESS 96.
The Applicant has urged the Board before, during (Tr. 5239),
and after the hearing, and most recently the Commission, to author-ize McGuire operation expeditiously, asserting that further authorization is required by May 15, 1981.
Applicant also sought dismissal of CESG from the proceeding (Tr. 5238, 5235).
It seeks a 35% power level license to avoid what it sees as constraints on the Board in the matter of authorizing full-power operations (Tr. 5240).
97*
The Staff, on the other hand, has been hard put to keep current with the developments.
It has provided the Board and the parties with a series of relevant documents either just before the hearing, during the course of the hearing, or after the hear-ing, including reports from Sandia Laboratory, Brookhaven Na-tional Laboratory, R & D A, Staff's Environmental Qualification of Safety-Related Electrical Equipment, Feb. 24, 1981; Reactor and Containment Systems Performance Assessment, testimony re-l ceived Feb. 26, 1981; and Affidavit of Vincent S. Noonan, Harold E. Polk, Krysztof I. Parezewski, and Charles G. Tinkler, concerning the polyurethane problem, March 27, 1981.
99*
Testimony by Dr. Berman indicates that Sandia Laboratory I
l studies on hydrogen combustion in ice condenser containments are still under way.
The cover letter to Dr. Meyer from Robert A.
Bari, for the Brookhaven National Laboratory study (Jan. 23, I
1981) indicates the short-term nature of the study and the haste i
with which it was undertaken.
Although this report was trans-mitted to NRC under a cover letter dated Jan. 15, 1981, it was not made available to the Board and the parties until Mar. 3,
-~,,,m,-
- m. -- -
4
. 1981.
It is the responsibility of the Staff to provide the Board and the parties with material information and, additionally, to provide evaluative comment (Tr. 3646).
The providing of the foregoing materials has been inexcusably slow.
The draft Sandia report is dated Dec. 1, 1981.
The first complete and legible copy of a letter intimating to the Board and the parties that hydrogen studies were undertaken at Sandia and that a report existed was mailed under a cover dated Feb. 20, 1981.
The report itself was not presented until the hearing opened.
The Staff has given "no indication of the significance of these (Sandia and BNL] reports to this proceeding..." (Tr. 3648).
99.
The disadvantage to the Board and the parties of late transmittal of available relevant materials has been, to some extent, overcome by the efforts of those involved in the hearing.
However, the disadvantage that has not been overcome is the in-complete status of the investigation of credible hydrogen re-lease scenarios and the study of effective hydrogen mitigation techniques.
The record is palpably incomplete.
The Fenwal i
l study offered by Applicant (DPC Exh. Sc) is obviously based on an over-simplification of containment conditions.
The
\\
Livermore ignition tests raised new questions about steam sup-l pression of ignition.
Clearly the consideration in Sandia and BNL studies of alternative sequences in resuming injection dur-
.ng an accident as these affect pressure development are
. germane.
100.
The information presently available is not sufficient to
~
I reach a decision.
The Board affirms CESG's motion that the Staff i
advise the Board and the parties of a complete list of the projects underway, that it regularly update the Board and the parties on the status of projects, and that it distribute the findings on these projects to the Board and the parties as soon as they 'are received by the NRC.
At such time as there appears to the Board to be sufficiently conclusive material to make a determination, the hearing shall reconvene.
i e
f I
r
- _. -.... - _,, -...,.,. ~. -.
,_,7.,,, _. _, _., _ _,,,
. Conclusions of Law 1.
The Board is cognizant of its high responnibility in this matter, stated in the Regulations to be to decide if there is " reasonable assurance that the activities can be conducted without endangering the health and safety of the public." [ App. A, VIII,(b)(3)] and to consider the environmental factors ((b)(7)].
In every case, therefore, a Board must. perform a balancing of the risk consequences versus the benefits, and the probability of the event versus the effect of the event.
When a relatively unlikely event has catastrophic consequences, the decision making process is difficult.
2.
In the case of the McGuire reactors, it is necessary to determine if there are sufficient safety features under
$100.10(a), so that the consequences of a breach are sufficiently minimized.
The failure of Applicant and Staff to put on evidence with regard to Contentions 3 and 4 of CESG require us to conclude that there is insufficient emergency planning in the event of a breach, that the ten mile radius is not a sufficient barrier and that crisis / relocation would be required and has not been planned for.
l 3-The legal effect of that evidentiary decision by the Applicant and Staff is to give full credibility to the postulation of severe consequences of failure.
Those l
consequences, the appreciation of which marks the limited appearance. statements and probably dictated the appearance l
of both the City of Charlotte and County of Mecklenburg, are to be figured in to the " reasonable assurance" standard.
l
- q.. -
That is, in carrying its burden of proof (see 52.732), Applicant l
must provide evidence sufficient to remove doubts created by the serious residual dangers of the situation.
4.
The severity of the consequences, which we must assume, lacking l
evidence to the contrary, makes us take the position that the liklihood of a breach from a hydrogen producing accident must be l
demonstrated to be vanishingly small.
5.
We also note that the decision making scheme of the Commission requires a separation of funcItions (52.719) and that the Staff is satisfied with what the Applicant has proposed is not sufficient.
It remains the Board's decision to find whether or not a " serious safety [or] environmental matter exists," (52.760a).
I 6.
The separation of functions of the Board and Staff is similar to the separation of powers creating a system of checks and balances in our Constitution.
Although the l
Board is a creature of the Administrative branch of government, it is judicial in nature and must remain aloof from the pressures placed on administrative and legislative bodies by the public and by private interests.
- Thus, in deciding the safety inherent in the design of this plant, the Board must not consider the communications sent to the Commission by public and private figures.
While there is no l
doubt of their sincerity, they cannot be familiar with the l
real questions of safety raised herein.
Similarly, this Board may not consider the profits to be gained by Applicant from the immediate operation of the McGuire planc, or the nature of its investment.
Although it is proper for Applicant to be motivated by profic, it is not proper for ultimate safety decisions to be made in light of potential profits.
.. ~.
- 4.: -
7.
After our evaluation of the evidence, then, and after taking judicial notice of the documents noted above in the findings of facts, we find that chare are unresolved questions remaining in regard to the safety of the plant.
The questions are as follows:
(1) the actual value to be given the containment strength, con-sidering the R&D Associates calculations, the possibility of a flaw in the welding process, and our general unease at the ex post facto recalcuation of the strength of an existing structure.
In this light, we conclude that Applicant's 67.5 psig figure overstates the strength, and that Staff's 48 psig figure does not provide the level of confidence that we consider appropriate.
(2) in consideration of worker performance and psychological phenomena, we conclude that Applicant has not trained its on-the-job personnel to be aware of, identify and compensate for human factors.
Further, no site specific human factors analysis has been made of the McGuire plant.
In that this area has been stated to present the greatest risk to safety (af ter the review of TMI), Applicant must provide assurances h
l that, for example, information overload will not occur w en the newly revised procedures are followed, and thaw risky-shif t will not occur within the on -site decision making group.
This has not been done.
Rather, Applicant has stated that these factors are not applicable to nuclear reactor decision making, a response we find insufficient.
1
--.,w-9
,--.---..-,..s,.
,.,--..---..,m
__,,ry--.-y-y
- s-(3) in regard to problems _anet ing the confined burning of hydrogen, we were informed that 2periments are on going.
It is signficant that two such experimenters were sufficiently cautioned by their findings and those of others to note the possibility of danger (Dr. Berman relying on Dr. Lee and Dr. Strehlow).
Much more experimentation needs to be done before the a priori conclusions of Applicant's witnesses can be confirmed.
13 l
(4) steam erting remains a mystery as to its probability of l
occurring and effects if it does occur during a LOCA.
In this connection, experiments have been proposed which would include l
l those concerning the parameters of the creation of the condition, its continued existence thereafter, adn also the MARCH analysis of operator intervention to resume ECCS and the creation of steam and hydrogen.
MARCH was only used to postulate operator inter v-ention in one accident scenario.
MARCH and CLASIX are under development and neither treats mixing adequately.
(5) the possibility of pyrolysis of foam insulation by burning in the ice condenser region is a phenocenon which has net received the usual Staff review.
If Applicant is correct about this burning location, then we would like to see some ' independent calculation about the ability of the wall panels to contain the gases produced, heat transfer, and the incremental pressures so produced.
8.
In regard to these questions, it is note worthy that we have reaped the harvest of Sequoyah studies, but that they do not settle the issue.
Rather, the results obtained raise
. new and more significant health and safety issues.
There is no doubt in the minds of the Board that conducting a public hearing, with active Intervenors, is an aid to deciding these issues, and that while eliminating the hearing would mean l
faster licensing, it would not mean more safety, or even, ultimately, more electric power.
i 9.
There is not a reasonable assurance that the activities l
authorized by a license permitting operation from 5% to full power can be conducted without an accident which would endanger the health and safety of the public.
(
10.
The issuance of the license could be inimical to the health and safety of the public, 11-The issuance of the license could significantly affect the t
l quality of the human environment and therefore should require l
the prepartion of a supplemental environ = ental impact statement.
l 12.
The appropriate course of action from an environmental and safety view point is the denial of the requested license.
ORDER Based on the foregoing findings and conclus.n 1s, we direct the Staff to devise a program of research to resolve the questions raised by this hearing.
Said program will be submitted to Applicant and Intervenor for conment within a reasonable period of time.
If objections are raised, this Board may then direct further issues to be studisd or otherwise alter the program of research.
Pending the completion of said program, issuance of an operating license
48 IS DENIED.
't THE ATOMIC SAFETY AND LICENSING BOARD Emmeth A.
Luebke, Mernber Richard F. Cole, Member Robe rt M. Lazo, Chairman Issued at Washington, D.C.,
this day of May, 1981.
I i
i l
l l
~
, APPENDD( A In APPLICANT EXHIBITS For ID Evidenc 5 A-D Anal 7 sis Hydrogen Control - 4 vol.
311e 311e 5s Errata to 5 A-D 3290 3290 6
Diagrams, air flou path in containment 3336 3337 7
Lattice Frame 3490 3616 d
Stipulation re CESG's psychologist witnesses 4674 4674 CESG EXH181TS 40 Sandia Report-Sequoyah, SAND 00-2714, NUREG/CR-1762, draft 3215 3215 4654 40 A A Final version 4670 4670 41 Stephan West affidavit 3402 42-57 Various NRC, NSAC papers 3702-3016 50 NRC Staff answer to CESG interogatories 301d 59 BNL Report, Pratt to Bari, Jan. 15 'dl 3022 4654 59 4663" 60 Britannica p. on solders 3c23 4654 61 Table 9-1, Dominant Accident Sequence Sequoyan, CH-16by 4523 4526 62 Chapter Vill Sandia CR-1459 4d02 NRC STAFF EXHIBITS l
E Berman to Murley, Feb. 9 '01
" App. B" 3160 l
F Hanrahan report: OPE Review Hydrogen Control i
for Sequoyah 3550 G
SER Suppl.
2 to McGuire 4006 H
SER Suppl.
3 to McGuire aOO6 1
SER Suppl.
4 to McGuire 4006 J
FE1S, UUREG-0063, Jan. 1961 4006 h
Staff Analysis of Hydrogen Control Measures for McGuire, App. A 4007 43S3 L
Fig. 7, John Lee paper given Jan. 27, 19d1 h297 l
M H & D Assoc. Report 4656 4663" l
N toard Notification, Feb. 24, 1901 4656 0
Fig 22.2-25 Supplement 4, Sequoyah San 5125 5126 P
Fig. 22.2-2A Supplement 4, Sequoyah SEH 5174 5175
"" Limited purpose", no casis for findings
--m
r
b Appendix B Board's Exhibits Ex 59.
Full consideration Ex 61 Full Consideration The Behavior of Hydrogen During Accidents in Light Water Reactors,NUREG/CR-1561, August, 1980.
Clarification of TMI Action Plan Requirements, NUREG-0737, November, 1980.
Review of Staffing Requirements for Near-Term Operating License l
Facilities, NUREG/CR-1764, Oct. 1980.
TMI Related Requirements for New Operating Licenses, NUREG-0694, l
June,1980.
R&D Associates, H.W. Hubbard to V. Gilinsky, Hydrogen Problems in Sequoyah Containment, August 4, 1980.
R&D Associates, H.W. Hubbard to V. Gilinsky, Sequoyah Containment l
l l
Analysis, July 15, 1980.
Some Hydrogen Control Considerations for Ice-Condenser Plants, I
R&DA-TR-17800-002, Feb., 1981.
l l
l
l Certificate of Service I hereby affirm that copies of Intervenor's Proposed Findings of Fact and Conclusions of Law.
. were served by placing copies for all persons at the U.S. Nuclear Regulatory Commission by Federal Express, addressed to Robert M. Lazo, Esq., as below, i
and by placing all other copies in the U.S. Mails, postage l
prepaid, addressed as stated below:
both on the If day of April, 1981.
Robert M. Lazo Esq.
William L. Porter Esq.
Chairman, Atomic Safety and Associate General Counsel Licensing Board Duke Power Co.
l U.S. Nuclear Regulatory Commission PO Box 2178 Washington D.C. 20555 Charlotta NC 28242 (alse enclosed:
copies for J. Michael McGarry, Esq.
Dr. Emmeth A. Luebke Debevoise and Liberman Dr. Richard Cole 1200 17th St. NW Edward G. Ketchen Washingecn DC 20036 Chairman, Atomic Safety and Dr. John M Barry Licensing Appeal Board Dept of Environmental Health 1200 Blythe Blvd.
Chairman, Atomic Safety and Charlotte NC 2S203 Licensing Board Panel Diane B. Cohn Chase R. Stephens, Public Citizen Litigation Group Docketing and Service Section Suite 700 l
Office or the Secreatary) 2000 P. St NW Washington DC 20036 Mayor Eddie Knox City Hall, 600 E. Trade St Charlotte NC 28202 s --
th
%, dL Shelley Blum-
- _ -