ML20003C163

Testimony Re Hydrogen Generation,Combustion & Containment Response,To Be Presented Re Contentions 1 & 2.Prof Qualifications & Certificate of Svc Encl.Related Correspondence

ML20003C163
Person / Time
Site:	Mcguire, McGuire
Issue date:	02/17/1981
From:	Jeffrey Riley CAROLINA ENVIRONMENTAL STUDY GROUP
To:
Shared Package
ML20003C158	List: ML20003C157 ML20003C163
References
ISSUANCES-OL, NUDOCS 8102260732
Download: ML20003C163 (15)

Text

-

Q-- - -. -__ _ _. __

W CORRESPONDENCE S

+

UNITED STATES OF AMERICA NUCLEAR REGULATORY COMMISSION 4

- #ecug 2 FEB 191981

• i 9"

BEFORE THE ATOMIC SAFETY AND LICENSING BOARD Omes of the 5808%M

, p & Sarace /

4 In the Matter of ) os

)

DUKE POWER COMPANY ) Docket Nos. 50-369-OL

) 50-370 OL (Wi'lliam B. McGuire Nuclear )

Station, Units 1 and 2) )

TESTIMONY OF JESSE L. RILEY REGARDING HYDROGEN GENERATION, COMBUSTION, AND CONTAINMENT RESPONSE

1. Q. What is the purpose of your testimony in regard to hydrogen generation?

A. It is to show, on the basis of verifiable fact, that the physical possibility exists for hydrogen to be gener-ated in substantially greater amounts than have been considered by Duke in its testimony. Such hydrogen, mixed with containment air, can be ignited. The combin-ation of hydrogen with oxygen can produce pressures sufficient to breach and burst the McGuire thin shell containment, with attendant release of radioactive fission products.

s

2. Q. Why is it relevant that the physical possibility exists for generating hydrogen in greater amounts than Duke has considered?

A. On ignition of larger amounts of hydrogen greater pressures l and temperatures will result. The amount of hydrogen released at TMI would be sufficient to breach the much smaller, thin shell McGuire containment. Duke has made i

the 8 to 10 v/o (volume percent) hydrogen concentration i

i present at the time of the hydrogen explosion the basis

! for its studies. Because the McGuire containment is only 59% the volume of the TMI containment, the same amount of hydrogen would result in concentrations of 13 5 to 17%

and a pressure of about 58 psig (pounds per square inch gauge) rather than the 20 psig peak at TMI. Containment Dreach would result in severe radiological consequences due to the failure of the last line of defence and the unimpeded release of radioactive gases and particulates.

3 Q. What actual quantity of hydrogen was generated at TMI?

l How does it relate to the amount of hydrogen that it is physically possible to generate?

81022607M

A. The generation of hydregen in the TMI accident was about 45% of the amount which potentially could have been gener-I ated. In the McGuire FSAR, in accord with 10 CFR 550.kk, it had only been required to consider dealing with a slow release of up to 5% of potential hydrogen, although it was asserted that this was a conservatism, that a maximum release of 2% was more realistic in a loss of coolant accident (LOCA). The release at TMI was about 10 times greater than the amount considered, far beyond the engineer-

. ing expectations of the NRC staff and, presumably, the utilities licensed under 10 CFR 550 44 Duke's testimony, A. D. Miller'(Appendix HYD to Report NSAC-1, " Hydrogen Planning", pp. 7-9), summarizes the material balance infor-mation which leads to the value of about 45% metal-water (i.e. hydrogen generation) at the time of the hydregen explosion at TMI.

4 Q. Did all of the hydrogen generated at TMI combust?

A. About two thirds of the hydrogea generated at about 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br /> after the accident initiated. This combustion caused the 20 psig peak pressure.

5 Q. Did this place the TMI containment in jeopardy?

A. No. It was designed on the basis of a 60 psig steam release accident. The ultimate capability has been estimated to be in the vicinity of 170 psig.

6. Q. 'What is the corresponding design basis for a McGuire containment?

A. The design basis for a McGuire containment is 15 psis.

Before the TMI accident the sum of air and steam pressures received primary consideration in a LOCA. In'a dry containment like that at TMI this combined p assure was estimated to not exceed 60 psig. This corresponds to a mean containment atmosphere temperature of about 290 F.

and does not depend on the condensation of steam. In an ice condenser containment like McGuire the steam released in a LOCA is condensed by passage over ice. Lower con-i tainment air is driven into the ice condenser and upper containment by the steam resulting in a calculated pressure of less than 15 psis.

7 Q. .Was the pressure due to hydrogen releases in excess of a 5% metal water reaction taken into considera. tion?

A.- No.

~

d. Q. -WereLthe pressures that could result as a consequence of.

combustion takan.into account?

A. No. ,

/

1 .

e ., *

]

9. Q. Are there any undisputed facts in regard to hydrogen generation, the pressures resulting from the combustion of hydrogen, and the capacity of a McGuire containment to withstand the pressures that can develop?

i A. Yes. Based on discovery there area number of critical facts about which there is no dispute. These reflect i established areas in chemistry, physics (the gas laws),

and engineering:

i i i A. At temperatures above 1700-1800 F zirconium, which  ;

i constitutes 96% of the Zircalloy-4 fuel pin sheath alloy, reacts with steam to form hydrogen.

! Zr e 2 H 2O 2H2 + Zr02 The reactignJoules) gives off a great deal of heat, 246 BTU (2.62 x 10 per gram mol of hydrogen. 1

!* B. The amount of zirconium in a McGuire core is about

45,232 lb. Completely reacted it will form 1983 lb of i, - hydrogen (900 kg), a volume of about 353,080 sof istandard cubic feet measured at 1 atmosphere and 32 F).

' C. Hydrogen combusts with oxygen in the ratio of 2 volumes of hydrogen gas to 1 of oxygen at the same pressure.

The-weight ratio is 2.016 to 16.000. ,

i At ngraal temperatures in an air filled containment, circa 120 F and 100% R.H. (relative humidity), hydrogen at r concentrations greater than 4v/o can be ignited. Combust-ion is not complete and flame propagation is upward only.

i At concentrations of 8v/o and higher combustion becomes

increasingly complete and flame propagation more rapid

[- and.in any-direction.' At a' concentration of 19v/o hydrogen

' detonation can occur. The detonation range extends to about 550/o. -0paese relation were published by Shapiro and Mofette, WAPD-sC-545 and,is reproduced in WASH-1400.)

D.. The' combination of stoichiometric. mixtures of hydrogen and containment air wor 1.d lead'to pressures greater than 100 psia, assuming an maitial pressure of;only 1 atmosphere. ,

E E.. Combustible hyd5o8'n 'ir m&xtures ignite at temperatures

! uns low as.1080-1090 F (560-590 C).

L l F. Containment yield' values calcu1Gted for Sequoyah range from'27 psig (R&DA) to 36 psig.(Ames). For the 50% thicker

.McGuire shell the.valuss of 40 5 and 54 psig correspond.

r . Duke reports.(Hydrogen. control, Vol. 2,_p. 4-11) 45-50 psis .

.for initial yield at the base.. Duke does not report a yield value for the shell. -Based on yield: stress tests of.the steel ~a value.of 45-50 psig would be obtained. Duke reports an ultimate value of 67.5 psig. .The ultimate value of h3;5 psig calculated by TVA for Sequoyah. translates to

~

65 3 psig for.the thicker McGuire shell.-

.. - - , - . . . . _ . .- . - . - _ _ . . . - - - - . . - - = = = . . .-

• ey e -use- ** * * * * * * -

_ . _ e ** m e

4-

, G. The most severe consequences of hydrogen combustion occur with the ECCS (energency core cooling system) inoperative. To expose the core the ECCS must be inoper-ative. The spray system is also inoperative and the two

, circulating fans are off.

Duke calculates an upper containment pressure of 117 psia for the co g lete burning of only 14v/o hydrogen with 1 fan '

i and 1 (of 2) spray train operating. A 100% metal-water reaction will provide a hydrogen concentration of about 25v/o. (Hydrogen Control, Vol. 2, Table 8)

H. The maximum temperature and pressure for hydrogen burning in air results when the hydrogen and oxygen are in the stoichiometric (precisely combining) ratio of 2:1.

This 4000,F.maximum temperature rise at constant volume is to about

10. Q. Are these all the facts of the matter?

A. There are some additional facts not addressed by Duke or NRC consultants which are verifiable.

I. Steam release in a LOCA raises thg lower containment

' temperature. At a temperature of 200 F, as a result of the

increased proportion of steam in the atmosphere, hydrogen l at the stoichiometric relationship would be at s concentra-

tion of about 16v/o, still in the, totally combastible range, although outside the detonatable range. Ice selt will

initially hold down the temperature of the upper containment.

The pressure in the containment, without cogbustion, if the atmosphere reached a temperature of 200 F would be about 37 psig. A lack of. containment cooling leads to

. higher temperatures and higher pressures.

J. To prevent containment burst by hydrogen combustion

, a number of means have been considered. 1hese include attempts to aasure effective ECCS performance so as to preclude hydrogen release; and a variety of mitigation measures including inerting and a Halon release system.

!- .11. Q. Given these facts, what is the basis for controversy?

A. ' The occurrence ol* a hydrogen release accident as the result of a LOCA is~not a foregone conclusion. .However the means proposed for avoiding hydrogen release and for mitigating

- the consequences if it occurs, as well as the magnitude of i - possible releases are subject to question.

A. Absent a LOCA there is no potential for the accident.

' All. parties accept. the' premise that LOCA's can occur and 4

should be planned and designed for. ;The LOCA's planned-

~ for exclude'only' reactor. vessel rupture and reactor. head separation, either'of which would produce extreme'conse-quences and which,'if'the. containment-did not immediately y- --

87 F- -i+ y f :.egw

rupture, would result in hydrogen generation and po;sible combustion and containment breach.

B. Given a LOCA short of reactor breach, the ECCS is designed to flood the core and to keep it covered with water by means of both high pressure injection pumps and low pressure injection pumps. If the ECCS functions until criteria testified to by Duke are met, Canady et al, (1) the coolant pressure exceeds a specified minimum and is increasing, (2) thepressurizerlevelexceedsaspecifieg minimum, (3) the coolant system is subcooled at least 50 F, and (4) adequate auxilliary feedwater is being injected into at least one non-faulted steam generator, if these criteria are satisfied core cxposure, if any, will be brief and the extant of the metal-water reaction under 5%.

C. If TMI operators had observed the foregoing criteria, core exposure and excessive hydrogen generation would not have occurred. ,

D. There is no assurance that Duke will be able over the planned operating life of the plant to maintain those patterns of behavior in all who operate the plant so as to assure avoidance of a TMI type accident.

E. There is no assurance that the additional instrument-ation to determine steam saturation inside the coolant system and to indicate water level in the reactor will be sctively and accurately operational at all times in the planned operating life of McGuire. Lapses would make difficult the determination of just those conditions which proved critical at TMI.

I F. Each of the exercises performed by Duke, the NRC, l and their consultants assumes an inoperative ECCS. All that is necessary to realize this assumption is the loss of AC power to the ECCES equipment. There are many multiple failure paths that can lead to the loss of AC power: 1) an electrical fire in control room switching gear, 2) burnout of specific transformers, 3) loss of trunk power and a coincident failure of the diesel electric generators to start, 4) failure of both low pressure injection pumps, 5) a seismic event causing aberrant performance of instrumentation and/or controls.

G. There are three PORV's (air operated pressurizer relief valves) in series with block valves. They are set at 2335 psig. There are also three mechanical safety valves, spring loaded, set at 2485 psig. The status of the PORV's is indicate'd on the control panel, that of the safety valves is not. .There are no block valves in series with the safety valves. This provides three possibilities for LOCA's which are not monitored in the control room.'

H. An MIT study of nuclear reactor operators working on simulaters showed that, after a simulated accident, it took the operator 20-30 minutes to recover competence.

Given a large LOCA in which automatic ECCS response is defactive, due to a failure of the controls, possibly as a result of mis-wiring, a fairly common error, blowdown would have occurred and hydrogen generation been either complete or well underway.

I. The possibility of an anticipated transient without scram (ATWS).has not been eliminated. Although upper there head is injection and other mitigants have been applied, still only one scram system. The consequences of a persist-ing ATWS have apparently not received consideration.

Indefinite flooding is not feasible, as the events at Chalk River proved. A cessation of flooding will result in core exposure and hydrogen generation.

The totality of kinds of serious reactor accident has probably not yet been experienced. Some of theseThis will result in core exposure and hydrogen generation.

conclusion is supported by the fact that the AEC did not anticipate either the Fermi or Browns Ferry accidents.

The NRC did not anticipate Three Mile Island. The Kemeny and Rogovin studies, as well as the Lessons Learned task force reports make clear a mindset which is perhaps best characterized by the expression, " Damn the torpedos. Pull Speed ahead." It may be that the host important lesson to be learned--that there are still lessons to be learned, has i not been learned. Murphy's Law still seems to hold. If something can go wrong, it will.

K. Duke, TVA, and NRC staff in carrying out calculations r

l on hydrogen release appear toR&DA havemore set arealis'tically 50% metal-water l reaction as an upper limit.

considers a range of cases including 100% metal-water reaction. There is no basis for the extent of the reaction i

to 50%. It creates the illusion that the problem is less l

i serious than it is. Slightly different circumstances rather than the inherent properties of the system would have resulted in a TMI release approaching 100%. All that would have been required (see NUREG-0600) would have been l

i slower rates of coolant injection, a longer period of j opened PORV.

These considerations apply to hydrogen generation. What

12. Q.

about ignition?

A. The NRC has not identified the ignition mechanism. Some of the consultants have assumed that the large number of electrical devices in the containment provide an assured random source of ignition when combustible levels of hydrogen are reached. The proposal of the distributed ignition system is a way of expressing disbelief in the

efficacy of. random source ignition.

Known means of ignition are a flame, an arc or statig spark of sufficient energy, a solid at dull red heat, 1100 F or Higher. The glow plug and the recombiner are instances of the latter.

lt is my opinion that the ignition at TMI occurred by contact of a combustible hydrogen-air mixture gin the reactor by means of a fuel pin exceeding 1100 F. The supporting basis for this view was provided by the I and E investigation. of the accident (NUREG-0600, part I-4) . At time of ignition, cooling flow was on. The outer fuel pins in the core were at a low temperature, the more central pins at a high temperature (Fig. 1 4-11). The g inlet temperatures to the coolant pumps had fallen below 200 F (Fig. 1 4-12C).

The EMOV (PORV) had just been opened when the pressure spike occurred (p. I-h-49). Two thumps were heard in rapid requence (p. I-4-47). The EMOV had been closed for

. the preceding 26 minutes (Table I.4-4) . During this time, due to cooling of the reactor toinlow the atmospheric boiling point, the pressure in the reactor became sub-atmospheric. When the EMOV opened, air rushed through the pressurizer and into the reactor providing the oxygen necessary for combustion. Hot fuel pins ignited the mixture. The burning gas blew back through the pressurizer and ignited the approximately 8v/o hydrogen in the con-tainment. Hence the " double thump,".

This explains why the combustible mixture had not ignited earlier. There was no effective random ignition source.

As motors and relays are enclosed to protect them from the humid atmosphere and spray, and have been known to operate completely submerged, it is not surprising that they did not serve as ignition sources. -

I am of the opinion that it would be premature to identify ignition sources in a possible McGuire hydrogen generation accident.

.13 Q. What controversy do you see in regard to containment strength?'

A. There are uncertainties that have been given a more credible treatment by R&DA than other consultants. The R&DA value for containment yield at Sequoyah of 27 psig translates to 40.5 psig for McGuire. The most significant fact is that a number of scenarios for McGuire exceed the highest ultimate value proposed, that of 64 psig in an NRC study. Only one study expressed a probabilistic confidence limit for ultimate. That was at the 955 level.

Considering the consequences of containment rupture and the usual uncertainty associated with ultimate values, this does not appear to be a conservative approach.

With only 1 spray and 1 fan Duke reported a pressure peak (foregoing) of 102 psig. Only by the arbitrary holding down to a 25% level of the netal-water reaction can the case be made that the containment weuld not leak or breach. Such an arbitrary choice appears to have no other justification -

than the desired result of not failing the containment.

14 Q. How does an NRC finding that operation is acceptable at an assumed 25% metal-water reaction fit in with NRC concepts of conservatism?

A. It does not fit. The McGuire containment shell was designed and constructed for a maximum 15 psig pressure.

With yield at 40-45 psig and burst at 65-70 psig there is genuine conservatism with a safety factor at yield of at least 2 7 Even with distributed ignition scenarios assuming two fans, two spray trains, and a 100/50% burn or 10/8 v/o hydrogen, pressures of 20 psig are estimated. (Hydrogen Control, Vol. 2, Table 6, JVD14) Safety factor 2. As soon as 1 fan is dropped, pressure is up to 42 psig and the safety factor becomes 1, which is zero safety margin.

Loss of AC power, the most probable cause of an inoperable ECCS, would disable the distributive ignition system.

Considering the rapidity of hydrogen generation in a large LOCA without ECCS, a peak hydrogen explosion pressure of about 200 psig-is credible (R&DA,'4 August, 1980, Table 2, 9 a), 836 kg, 13 3 atm = 195 pria, 160 psig). The safety factor in this scenario is about 0.2. Conservatism has vaniched. Chancing better describes the approach.

15 Q. Do you think the distributive ignition system as installed by Duke will be an effective means of avoiding serious pressure releases due to the generation of hydrogen?

A. The answer is yes and no, depending on the circumstances.

If the hydrogen release is slow, corresponding to gradual j core exposure as coolant boils out of the roactor, the consequence of the relatively low decay heat rate

! associated with low power operation, no ECCS operation, and the fans and spray trains fully operational, the hydrogen in the containment will approach a uniform mixture with air. When the concentration of some pockets exceeds 4v/o, these will burn off with relatively small pressure consequences. Assuming no equipment railures, this process can be expected to ecitinue as lo'ng as slow

hydrogen evolution occurs. For exanple, at 15 metal-water reaction the rate of hydrogen production is about 3,000 l

.sef/ minute. The f ans circulate containment atmosphere i at a rate of 60,000~ ef/ minute. The coolant system l is ejecting steam along with the hydrogen. The upper

.. ___- . - ~ _.-

_9 limit of hydrogen concentration is about Sv/o. Under these conditions distributed ignition appears to keep the containment atmosphere from greatly exceeding the lower combustible limit of 4v/o and to distribute the burnoffs over a long enough period of time or the heat removal proceeses, conduction to spray droplets and conduction to equipment and structure, so as to mitigate temperature and pressure rises to structurally acceptable limits.

Larger LOOA's at high power levels are another matter, particularly when combined with an electrical failure.

Duke has already had an electrical failure at McGuire (minutes of 240th ACRS meeting, December 5,1900 p.300) in which loss of the venting cooling in the pressurizing compartment led to extensive equipment burnoit and the loss of the ambient cooling system during het functional testing.

In a large LOCA at full power without off-site or on-site AC there will first be a blowdown of the reactor content, refilling of the reactor by the core flood tank, heating and boiling off ~of the core flood tank water. It will take 30-40 minutes for the core to become sufficiently exposed and hot for the metal-water reaction to begin.

Total core exposure under these conditions would take several hours and be a limiting factor in the rate of core exposure and the extent of the metal water reaction.

However a failure to scram as an e'lement in the accident would result in complete core exposure in about 6 minutes.

Under these conditions there would, during the blowdown, be an intense heating of the lower containment. This wouldbethecaseforaLOCAwithscramaswellasanATWy.

The minimum heat injected would be of the order of 4 x 10 BTU. The ice in the ice condenser would largely have melted.

Hydrogen would be issu.ing from the opening in the reactor coolant system at a rate of up to 20,000-30,000 cf/ minute.

l The hydrogen flow would be turbulent. A volume of essentiel3y pure hydrogen would drive the initially emitted and uncon-densed steam before it. If igniters were on, no ignition would occur in 'either the hydrogen-steam transition band or in the pure hydrogen. Oxygen is a requisite for ignition.

If ice remained in the ice condenser, the hydrogen would be stripped of steam. There would form a narrow, moving tran-sition band of hydrogen and air. This combustible band could be ignited by the igniters in the upper plenum of the ice condenser. However the pure hydrogen following the transition band would not be ignitable and would flow into the upper containment. Due to its buoyancy (hydrogen has only 7% the density of air) it would rise to the dome.

Depending on the coherence of this hydrogen cloud, and the degree of progress in mixing with the upper containment air, ~

only a minor amount of hydrogen-air combustion would occur on contact with the dome igniters. The remaining hydrogen L

cloud would-gradually diffusively mix with the approximately equal volume of air in the lower part of the upper con-tainment (Hydrogen Contr31, Vol. 2, Fig. 3 3 1 (2 pp.',

{ig. 3 3 5). The result would be to make the entire remaining amount of hydrogen not only combustible, but detonatable. When enough oxygen had diffused to the vicinity of a dome igniter the entire combustible volume  !

would explode. The igniter system would have caused the event which it was intended to forestall.

What appear to have been similar concerns have been '

expressed by Sandia National Laboratory (Marshall Berman to the NRC's Thomas E. Murley and Durwood F. Ross, Feb. 9, 1981, p. 4). Quoting:

In summary, we think that: 1. The potential for hydrogen generation may be larger than previously estimated. 2. The codes being employed for accident analysis may be very inadequate for licensing decisions and should not be used to replace physical intuition and engineering judgment.- 3 The probability of lower compartment inerting during an accident may be higher because of computational uncertainties and because of the possible existence of an additional physical mechanism for inerting (condensation fog).

4 If a detonation were to occur in the IC upper plenum, calculations have shown that containment failure could not be conservatively precluded.

Deliberate ignition in the upper compartment is a difficult, quantitative problem. Igniters placed high in the upper compartment will probably be bene-ficial most of the time, but circumstances can be postulated where they might be dangerous. Igniters in the IC upper plenum appear to be justified only if one abandons the entire concept of lean mixture combustion, and replaces it with early, rapid, rich mixture, local combustion. We think that such a concept is fraught with danger.

16. Q. The Fenwal study (Hydrogen Control, Vol. 3) was undertaken ta examine the utility of the distributed igniter system.

Will you comment on it?

A. The Fenwal study was performed with a small, 133 7 cf spherical tank with the igniter.placed at the center of its~3 17 foot radius. The McGuire containment is 115 feet in diameter, 171 feet high, with a free volume of 1,200,000 cf. It is compartmentalized and of complex geometry (id.

Fig. 3 3 1). The J1 igniter locations result in one location per 39,000 cf. The volume ratio per igniter is 300:1, McGuire:Fenwal._As a model, the Fenwal equipment

~

is not appropriate.

In the Fenwal test either a well mixed atmosphere of hydrogen, air, and steam was subjected to ignition or

i l .

a well premixed stream of combustibles was let flow into the sphere. Nowhere did the test deal with heterogeneities in composition, fronts, transition regions, accumulations, the slow diffusive mixing process, or rapid turbulent mixing. The tests were confined to deflagratable compo-sitions and the low end of the temperature-pressure scale.

The detonation region was not approached.

I do not question the findings reported. I do question the applicability of the findings to the actual problem.

~

17 Q. Can you propose any remedies for the problems you have mentioned?

A. Yes. Of the mitigation measures that have been proposed I consider inerting the most reliable. It is relatively simple and inexpensive to remove oxygen fromair. There are several processes. In one, air is passed over finely divided copper at an elevated temperature. The proposed Halon system is expensive. It is cumbersome and depends

, on a sensitive, sophisticated triggering system which may be adversely effected by long exposure to a containment atmosphere. It has the capability of quenching a deflagra-tion. It operates by detecting pressure rise. It would not operate rapidly enough to suppress a detonation and would only further contribute to the development of

containment pressure.

Containment inerting poses problehs for the personnel who are required to pc2 form the frequent ms.intenance necessary'for operation of the ice condensor system. The atmosphere ~would not have to be deinerted during maintenance. '

There is a well developed air tank, air hooe technology for working in oxygen deficient environments. The buddy system would provide additional assurance for personnel. However some added hazard would have been introduced.

More basic than mitigation measures or administrative efforts to preclude the occurrence of hydrogen generation, and in accord with,the initial premises of the AEC in regard to safeguarding, would be providing a reasonable safety factor for the maximum credible hydrogen explosion accident by structural modification. This would call for a yield value of the structure of- 250-300 psig and an ultimate value in the vicinity of 400 psig. This structural upgrad-ing could be attained by a variety of means.: The-present circumferential stiffeners make no contribution to overall yield or ultimate. The present spacing cf about 10 feet is too great. Closer-spacing of stiffeners, at perhaps

- 2 5 foot intervals, would make a contribution which is

. susceptible to calculation. Additional vertical stiffeners might also be required. The 5 foot annulus gap provides~

adequate work space.

Another means of strengthening would involve the outer shell, if.it were found to have sufficient structural q .- ,,.ec. , ., , -. - ,y m,vr- ,, ._,vw, ,y --r. , .-.r,1 ,-.,w,. ,yv.~-,.-.e-.-=.y r,.

i l

... l capability. Bracing could be devised such that the outer containment could provide support for the inner shell if it were under sufficient stress to initiate yield. The contact of the bracing with the inner shell could either be jointed or sliding contact to accomodate differences in thermal expansion.

18. Q. Have you had any professional experience in the matter of explosions?

'A. Yes. I was called on to investigate a totally unexpected chemical explosion which hospitalized the five people in the laboratory where it occurred. I was also called on to investigate an industrial explosion of a presumably safe process which killed two people, totally incapacitated a third person, and did extensive material damage.

19. Q. Did you reach any conclusions as a result of these investigations?

A. Yes. Our deficient knowledge is a factor. We frequently plan on the basis of thinking that we know all that it is necessary to know. In both the explosions I investigated there were significant knowledge gaps which were not recognized. This is a recurring pattern in chemical explosions.

Another c'ommon factor was operator judgment prior to the explosion. 12 one case color' changes in the material would have served a. a warning to a person very highly knowledgable in that ffeld. The operator was not required to have this qualificatiun. In the second explosion I.

investigated the operation had been carried on for many years by another operator without adverse event, although there had been a prior event that could have served as a warning but was not approprir.tely construed. The second accident happened when t substitute operator carried out what was taken to be a ecnpletely safe operation. He was perhaps not aware that a backup safety system was not functioning due to a mechanical failure.

l Professional Qualifications of JESSE RILEY

)

My name is Jesse Riley. I am currently a consultant to the R and D department of CelaneseFibers Company. My home address is 654 Henley Place, Charlotte, NC.

.I have participated, as a member of the Carolina Environmental Study Group in the McGuire CP and OL proceedings, the Catawba CP proceeding, and the application to transship spent fuel from Oconee to McGuire.

I mm a member of the Conservation Council of North Carolina and have served on the Board. Presently I am a me=ber of the National Energy Committe9 of the Sierra Club and chair the Nuclear Legislation Steering Committee and the Nuclear Subcommittee.

I enrolled in the School of Engineering at Northwestern University but subsequently transferred to Arts, majoring in chemistry, physics, and minoring in math. I graduated cum laude with honors in chemistry and physics. I am a member of Sigma Xi and Phi Beta Kappa. My masters degree is from the University of Chicago where my thesis studies involved reaction kinetics and the isotopic tracing of reaction mechanisms.

I have been employed by Universal Oil, Products Co., Arundel

. and Consolidated Engineering, the U.S. Navy, Shell Oil Co., and Celanese Corp. At UOP I did research on gas analytical methods and low energy gas separation technology. At Shell my work in part was concerned with the phase relations between air and volatile hydrocarbon liquids. At Celanese I have been concerned with a wide variety of problems including heat transfer, rheology, materials testing, design of safe X-ray equipment, morphological, studies by l X-ray diffraction, reaction kinetics, vibration in jets, the structure of cellulosic materials, and the characterization of visco-elastic particles. I have invented several processes and hold 151 patents. During the last two decades 1 nave been called on to diagnose and correct manufacturing problems.

Mypublications are in the fields of rheology, materials testing, reaction kinetics, and the morphology of particles formed in the acetylation of cellulosic fibers.

I have found that physical intuition, which is closely related to engineering judgment, is a more powerful tool in problem solving than reliance on the literature or computer codes which usually have turned out to be simplistic.

i i

, e n. w w - , , . , g , , , . , . , s,n - ,w--

ENCf d) 4 UNITED STATCS OT AMERICA NUCLEAR REGULATORY COM;iISSION DoCKETEP BEFORE THE ATOMIC SAFETY AND LICENSING BOARE 198f > '-L Office of in, In the Matter of I Cs Wat & Service Branch /

)

DUKE POWER COMPANY ) Docket Nos. 50-369 OL g

) 50-370 Cw (William B. McGuire Nuclear )

Station, Units 1 and 2) )

AFFIRMATION OF SERVICE I hereby affirm that copies of " Testimony of Jesse L. Riley . .

and a letter to the ASLB concerning prefiled testimony in the captioned matter, dated February 17, have been served on the following by deposit in the U.S. Mail this 17th day of February, 1981:

Robert M. Lazo, Esq. OEdward G. Ketchen, Esq.

Chairman, Atcmic Safety and Counsel for NRC Regulatory Licensing Board Staff U.S. Nuclear Regulatory Office of the Executive Commission Legal Director Washington, D.C. 20555 U.S. Nuclear Regulatory Commission Dr. Emmeth A. Luebke Washington, D.C. 20555 Atcmic Safety and Licensing Board William L. Porter, Esq.

U.S. Nuclear. Regulatory Associate General Counsel Commission Duke' Power Company Washington, D.C. 20555 Post Office Box 2178 Charlotte, North Carolina 28'42 4 Dr. Richard F. Cole Atomic Safety and Licensing Chairman Board Atomic Safety and Licensing U.S. Nuclear Regulatory Board Panel Commission U.S. Nuclear Regulatory Washington, D.C. 20555 Commission Washington, D.C. 20555 J. Michael McCarry, 111, Esq. Chase R. Stephens Debevoisc and Liberman Docketing and Service 1200 Seventeenth Street, N.W. Section Washington, D.C. 20036 Office of the Secretary U.S. Nuclear Regulatory Chairman, Atomic Safety Commission and Licensing Appeal Board Washington, D.C. 20555 U.S. Nuclear Regulatory Commission 3,ayor Eddie Knox Washington, D.C. 2055s,,

600 East Trade St.

Charlotte, NC 20202 Dr. John Barry 1200 Blythe Blvd. "All NHC copies were sent c/o Charlotte, NC 20207 hr. Ketchen by Federal Express t

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

- 2-Diane B. Cohn, Esq. Shelley Blum, Esq.

William B. Schults, Esq. 1402 Vickers Ave.

Public Citizen Litigation Group Durham Suite 700) N.C. 27707 2000 P Street, N.W.

Washington, D.C. 20036 George Daly, Esq.

Room 215 301'S. McDowell Street Charlotte, N.C. 28204 34.sJ A JfsseL.Rileyfor ESG D

ap .- - w t' -m --. -a w- eawee- .- + . + m,e---rm--w- -

n-w 4+-e-e -.w --w--e -W----- w