ML20052A980

Summary of the Zion/Indian Point Study

ML20052A980
Person / Time
Site:	Indian Point, Zion, 05000000
Issue date:	03/31/1982
From:	Murfin W SANDIA NATIONAL LABORATORIES
To:	NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES)
References
CON-FIN-A-1019 NUREG-CR-1409, SAN80-617, SAND80-0617, SAND80-617, NUDOCS 8204290505
Download: ML20052A980 (70)

Text

. _ _.

NUREG/CR-1409 SAND 80-0617 R1, R 7 P

e Summary of the Zion / Indian Point Study i

i Walter B. Murfin s i 1

p g

.s s.

~

'%4 J' +

~. <

+

t Q

A

- i,%.

+

l,b ~ c s

.,C o.,.

3 s

~

Bandia National Laboratories F 2 9 0 0 - Q( 3-001 Prepared for U. S. NUCLEAR REGULATORY COMMISSION i

820429 0 605

e f

t 4

r I

s e

i

.5 y,L a.

5

(

P ry

)

t

, 't

.i s

NOTICE This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the

./,

/..

United States Government nor any agency thereof, or any of their employees, makes any warranty, expressed or implied,'

or assumes any legal liability or responsibility for any third r

1*

t party's use, or the results of such use, of any information,

.h

- apparatus, product or proc %s disclosed in this report, or

{

represents that its use by such third party would not infringe

./ '

privately owned rights.

i i

t' g

.3.,

1 l

4 i

t.

s' f'

1 fl.

?l

}

/ ~ rf~

-g q

x.

n,f or l

'c Available from s

E GPO Sales Program r

Division of Technical Information and Document Control i

U.S. Nuclear Regulatory Commission Washington, D. C.

20555 4

and i

. ~ -

National Technical Information Service Springfield, Virginia 22161 I

g 4 - &

1 t

c

,'s..

,-aw-

--e

,,mn.

,-p

1 NUREG/CR-1409 SAND 80-0617 R1, R7 a

SUMMARY

OF THE ZION / INDIAN POINT STUDY W. B. Murfin Printed:

April 1980 Sandia Netional Laboratories Albuquerque, NM 87185 operated by Sandia Corporation for the U.

S.

Department of Energy Prepared for Division of Reactor Safety Research Office of Nuclear Regulatory Research U.

S. Nuclear Regulatory Commission Under Memorandum of Understanding DOE 40-550-75 NRC FIN No. A1019

d Abstract Results of a study by Sandia National Laboratorie's (SNL), Los Alamos National Scientific Laboratory (LANSL),

and Battelle Columbus Laboratories (BCL) for the identification of reactor core-melt accident mitigation measures at the Zion and Indian Point plants are summa-rized.

Mitigation strategies have been identified that show promise of providing large reductions in consequences for specific accident sequences.

However, without an overall risk analysis, it is not clear to what extent a given mitigation scheme reduces overall risk.

The study evaluated filtered-vented containment systems, steam explosions, hydrogen burning, hydrogen control measures, melt / concrete and melt /MgO interactions, and meltdown phenomenology.

Steam explosions have been determined to be unlikely to present a threat to contaimment at Zion and Indian Point.

It has been determined that a one-time burn of the quantities of hydrogen expected in a meltdown accident with modest initial overpressures could threaten containment integ rity.

Penetration of the basemat in a meltdown accident could not be confidently established; if penetration occurs, it will probably be after 3 to 4 days.

A core retention device at these plants might give an additional delay of a few hours to a day.

O Y

l

Foreward This report _is a summary of a study conducted jointly by Sandia National Laboratories, Los Alamos Scientific Laboratory, and Battelle Columbus Laboratories.

The work of the many investigators at all three laboratories should be recognized.

Those who devoted much of their time to this project are listed in Appendix G.

Many others, both at the laboratories and at the U.S. Nuclear Regulatory Commission, made valuable contribu-tions as consultants and reviewers.

Secretarial staffs of the laboratories are also commended for their high performance in the face of a very demanding schedule.

k 4

b 2

Contents

1.. Background and Study Objectives 2.

Study Basis 3.

Study Organization 4.

Summary and Conclusions Appendices - Summary of Individual Tasks Appendix A - Task I:

Filtered Vented Containment System Design Studies Appendix B - Task II:

Steam Explosions Appendix C - Task III:

Hydrogen Burning Appendix D - Task IV:

Hydrogen Control Appendix E - Task V:

Melt Interactions with Concrete /Mgo and Basemat Melt-through Appendix F - Task VI:

Meltdown Phenomenology Appendix G - Study Personnel References Figures l

Figure 1 Containment Failure Modes Figure Al Design Option 1 Figure A2 Design Option 2 Figure A3 Design Option 3 Figure A4 Design Option 4 Figure AS Design Option 5 1

s 1.

Background and Study Objectives

1.1 Background

~

The Reactor Safety Study (RSS)1 evaluated the risk to the public f rom potential accidents at light water reactor (LWR) plants.

The study evaluated the combined probability of events leading to releases of radioactivity to the public, quantified the magnitude of the releases, and estimated the public consequences.

The RSS determined that the maximum consequences and risk arise from core meltdown acci-dents in which the containment fails.

Containment failure modes identified in the RSS were slow overpressurization, overpressurization due to hydrogen burning, penetration by missiles generated by steam explosion, failure to isolate containment, basemat penetration, and interface-system LOCA (valve failures leading to releases outside of containment).

Consequences resulting from basemat penetration--in the absence of other failure modes--were determined to be relatively low.

In response to direction from Congress, the U.S. Nuclear Regula-2 tory Commission (NRC) prepared a plan for development of improved safety systems in early 1978.

One of the facets of that plan--which was also accorded high priority by the Advisory Committee on Reactor 3

Safeguards --was the development of filtered vented containment systems (PVCS).

In the same time frame, several studies (e.g., Ref-erence 4) pointed out the potential of filter venting concepts.

The major issues involved in conceptual designs of an FVCS were explored 5

in a program plan submitted in October 1979.

In addition, the NRC l

4

7 Lessons Learned 6 and Special Inquiry Group recommended investiga-tion of the possible role of an FVCS in accident mitigation.

In late 1979 NRC analyzed the risk from the Zion and Indian Point plants, using accident probabilities calculated for the pressurized water plant used in the RSS combined with actual popu-lations and weather for Zion and Indian Point.

As a consequence of the high population density and questionable effectiveness of evacuation, the hazards for these plants appeared to be higher than for other similar plants.

Several potential methods were identified for reducing the risk at these plants, including reducing the probability of valve failures that could bypass containment, reducing the probability of an unisolated containment, improving the containment cooling system, and the installation of mitigation systems.

In late December 1979, NRC requested SNL, LANSL, and BCL to perform a joint study of accident mitigation measures at the Zion and Indian Point plants.

The study began on January 7 with an organization meeting.

This report summarizes the joint laboratory effort.

i l.2 Objectives The objective of the study is to identify some of the methods for significantly reducing the likelihood of large airborne releases

~

of radioactivity that could result from core melt accidents in light water reactors (LWRs) in which the containment ruptures aboveground.

I Of particular concern are plants located in high population densities that could make the effectiveness of evacuation questionable, such as 5

tihe Zion and Indian Point facilities.

The laboratories were also requested to evaluate the effectiveness and to identify the design criteria appropriate to each mitigation scheme, in order to provide NRC with a technical basis for decision making.

A decision whether or not to require mitigating measures at these plants, or an outright recommendation for or against such a decision was not an objective of this study.

NRC also requested consideration of some of the events in the meltdown sequence--such as steam explosions and hydrogen burning--

that could negate the ef fectiveness of mitigation measures.

Studies of unisolated containments, valve reliability in interfacing systems, end improved containment cooling were separately conducted.

2.

Study Basis The RSS pointed out that specific combinations of system failures and human errors would produce more-or-less predictable

" accident sequences."

Probability of occurrence and consequences to the public can be estimated for each specific accident sequence.

The, combination of probabilities and consequences for all possible sequences defines the overall risk.

It was determined that a relatively few accident sequences dominate the overall risk.

That is, their contribution to the risk is so great that it overshadows the combined contributions of all other accident sequences.

The RSS also determined that maximum consequences from reactor accidents occur if the containment barrier fails or is bypassed in

~

meltdown accidents.

In these very low probability accidents l

6

l l

)

o radioactive material sould be released from the core debris into I

the containment atmosphere or into cooling water.

Failure of the i

containment could release a part of the radioactive material into the atmosphere.

Figure 1 shows a logic tree or fault tree of events leading to containment failure.

Previously identified modes of containment failure are:

steam explosion generated missiles penetrating contain-ment, failure by hydrogen burning, failure by static overpressure, failure by basemat penetration, loss of containment isolation, and interfacing-system LOCA.

This study has identified an additional event, massive steam generator failure which could bypass contain-ment; this can be considered a special case of loss of containment isolation.

Although any specific mitigation scheme may prevent containment i

failure for some accident sequences, it may be ineffective against others.

For example, a relief valve with a low flow capacity could prevent containment failure caused by slow pressurization but would be ineffective if faced by the release and detonation of massive l

quantities of hydrogen.

There are accidents which cannot be materially affected by any mitigation scheme.

These are accidents in which the containment is bypassed, for example, the interf acing system LOCA.

Other methods l

are available to reduce the risk from these accidents.

~

1 l

Each mitigation system has a certain risk of spurious opera-tion.

Some strategies also have the potential of worsening some accident sequences.

A mitigation system will offer a certain I

reduction of consequences when faced with some accident sequences, 7

i

y 4

i t

p -

!=

1-

-ill-ill igeO

's;

ai, i

"ATI o

,!j!

,i.

!s A lili.o

'li,

is4f-i j;;

11 a

in

=!'

11

=

I!!

O JlE!.Y 0

~~

m 8

,5

~

1 l

f,!-..1 E

I!

9.-

iii 5

ini'!

z a

., px

!i i

ei ;V 4

E v

,v,. a' =

&s T 4-:

5

( u

';y

/i i' [h1 il!

p x!! _'ih l g l

E h

i,,b

[

eg 3

3 gh.

~ji w

r-j !!! (__,f

_ il f3. !'. -!y rs r

O ds!

i 2

A, s i; 33 8

if it works as designed.

There will also be an increase in conse -

quences if the system is operated when it is not needed, or if it exacerbates an accident sequence.

Most accidents do not require additional mitigation measures to avoid containment failure.

Operation of the system in these more benign sequences leads to unnecessary release.

The releases would be small, but the probability is expected to be higher than the probability of accidents for which the system gives large consequence decrements.

An overall risk assessment is required to determine if combination of low probability and high risk reduction when the system works as designed competes favorably with a combination of higher probability and low risk increment when the system operates unnecessarily or improperly.

This study has focused on the identification of design concepts for filtered-vented containment systems (FVCS).

The qualitative advantages, disadvantages, and hazards of several strategies are pointed out.

However, competing risks and benefits could not be quantified within the time frame of this study.

Related issues in this study included an evaluation of contain-ment disruptive events--steam explosions and hydrogen burning--and of phenomena that might impact the design of an FVCS--core / concrete interactions and meltdown phenomenology.

3. Study Organization The study was broken down into six major tasks:

Task I

- Design of FVCS 9

Task II

- Steam Explosions Task III - Hydrogen Burning Task IV

- Hydrogen Control Task V

- Melt Interactions Task VI

- Meltdown Phenomenology The study was conducted from January 2, 1980, until March 31, 1980.

At the time of reporting nost tasks were essentially completed, with only a few details to be filled in.

It is not expected that accomplishment of any remaining work will alter the conclusions of this report.

Detailed reports of the study are contained in two companion volumes:

" Report of the Zion / Indian Point Study:

Volume I,"

SAND 80-0617/1, NUREG/CR-1410, Sandia National Laboratories, April 1980 and " Report of the Zion / Indian Point ~ Study:

Volume II,"

LA-8306-MS, Los Alamos National Scientific Laboratory, April 1980.

These reports should be consulted for more detailed information.

i 4.

Summary and Conclusions This study has identified potential methods for mitigating the consequences of postulated meltdown accidents at Zion and Indian Po in t,' in which the containment might rupture aboveground.

Failure to isolate containment and interfacing systems LOCAs were not e

included.

The study concentrated on FVCSs.

Phenomena that might l

destroy containment integrity, such as steam explosions, hydrogen

~

burning, and rapid steam release were also investigated.

Other l

10

phenomena investigated--core / concrete interaction and meltdown phenomenology--impact the design conditions for an FVCS.

Accident sequences were selected to cover the envelope of expected design conditions:

slow and rapid pressurization, hydrogen burning, and large amounts of steam generation.

Future plant-specific system analyses might determine that the particular sequences used are not dominant sequences at Zion and Indian Point.

The most challenging design condition is a pressure " spike" caused by rapid steam release or hydrogen burning.

Rise times of the spikes range from a few seconds to a few minutes.

The peak pressures in some sequences would be likely to destroy containment integrity.

The rise time is so short that extremely high venting rates would be required for attenuation.

Mitigation strategies investigated include containment pressure relief or depressurization with high and low venting rates, primary system venting, containment flooding, and combinations of these.

Some strategies that show promise of handling the widest range of accident sequences also appear to have a high potential for harmful system interactions, or require considerable operator intervention.

It is not clear whether complex strategies covering a wide range of acc.ident sequences, but with considerable adverse potential, are preferable to simple strategies effective against a more restricted set of accident sequences, but with minimum potential for harm.

For any given venting strategy several levels of filtering effectiveness are possible.

The simplest option considered is a 11 J

gravel-filled suppression pool.

Each added level of sophistication reduces the release of radioactive material but adds to the cost.

Some combinations of venting strategy and filtering options show promise of great reductions in potential consequences for a single cccident sequence under restricted, optimistic conditions.

However, these reduced consequences have not been balanced against added risk from faulty operation.

The probability of steam explosions has not been established.

However, detailed analyses have shown that if steam explosions do occur, they would be unlikely to threaten containment integrity, and hence would not negate the effectiveness of filtered venting.

The quantity of hydrogen expected from metal / water reactions would be unlikely to form a detonable mixture if uniformly distri-buted throughout containment.

However, deflagration of this quantity of hydrogen would present a threat to the containment if there is a modest initial overpressure.

The burning of half the hydrogen might cause damage but would be unlikely to burst contain-ment., Current methods of hydrogen monitoring and control are inadequate to deal with the quantities of hydrogen expected in melt-down accidents.

Control methods have been identified that show consi.derable promise, but large-scale development efforts would be required.

Penetration of the basemats at Zion and Indian Point could not be either definitely established nor positively ruled out.

If pene-tration occurs, the best estimate is that it will be delayed 3 to 4

~

days, and the penetrating material will be solidified.

A core t

l l

12 l

t

retention device at those plans could give an additional delay of a few hours to a day.

The understanding of meltdown is dominated by phenomenological uncertainties.

Present methods of predicting the course of meltdown rely on simple bounding calculations.

The predicted timing of events is believed to be reasonable, but the motion and conditions of the molten core are uncertain.

It is not certain that the core debris is coolable either in the vessel or in the reactor cavity.

Produc-tion of a steam " spike" when debris contacts water in the cavity cannot be ruled out for any sequence.

Restoration of ECCS can halt meltdown under some circumstances.

Ilowever, the possibility of massive steam generator tube rupture caused by thermal shock or sudden steam generation should be inves-L tigated.

The most important uncertainty is the ef fect of competing risks on the overall risk reduction of mitigation schemes.

Another 1

important uncertainty is the rise time and magnitude of pressure

" spikes."

If the calculation of these is excessively conservative, l

simpler venting strategies could appear relatively more attractive.

l b

13

Appendices - Summary of Individual Tasks Appendix A Task I:

Filtered Vented Containment System Design Studies A.1 Objectives The objectives of this study are:

1)

Development of conceptual FVCS designs, 2)

Development of preliminary design specifications, 3)

Evaluation of design options and strategies.

It has become apparent that quantitative evaluation of design strategies could only be performed subsequent to a detailed risk analysis.

Each design strategy has advantages and disadvantages that are not yet quantified.

Because the design strategies can only be qualitatively evaluated, the design specifications cannot yet be completed in detail.

Subject to this reservation, the objectives of this task have been met.

A.2. Design Conditions It is not practical to design an FVCS to be successful against every. conceivable accident.

Any given design will be successful when conf ronted with some accident sequences and unsuccessful for others.

It also appears that some design schemes that are successful against certain accident sequences have a concomitant high likelihood of spurious operation.

14

In order to determine the environmental conditions and con-g sequences and evaluate the design options, we have followed the I

following stepwise process:

1)

Selection of a set of accident sequences, 2)

Calculation of the containment environment, 3)

Design of a system for the environment, 4)

Calculation of consequences, with and without the FVCS. -

We have investigated the response to a set of accident sequences.

The sequences were chosen on the basis of their significance at the PWR analyzed in the RSS with consideration for design differences between the RSS plant and the Zion and Indian Point plants.

Also each of the accidents offers a severe challenge to FVCS design.

Because of plant differences not yet analyzed, it might turn out that some of the accident sequences chosen for analysis are not dominant sequences for Zion or Indian Point.

However, the sequences do exhibit the conditions expected to 1.imit the behavior of an FVCS:

both slow and rapid pressuriz-tion, with and without hydrogen burning.

It is believed that these, sequences fairly define the envelope of conditions with which an FVCS might be conf ronted.

The sequences chosen for study and the reasons for the choice are as follows (symbology follows the RSS):

TMLB':

Loss of of fsite and onsite AC power for 16 hours1.851852e-4 days <br />0.00444 hours <br />2.645503e-5 weeks <br />6.088e-6 months <br />, resulting in loss of secondary heat l

15

removal,. followed by the return of AC power and restart of the containment coolers.

This sequence has a high (approximately 120 psia) steam spike at about 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />, and a slow pressurization af ter about 7 hours8.101852e-5 days <br />0.00194 hours <br />1.157407e-5 weeks <br />2.6635e-6 months <br />.

I AB-Burn:

Large loss of coolant accident (LOCA) plus loss of of fsite and onsite AC power for 16 hours1.851852e-4 days <br />0.00444 hours <br />2.645503e-5 weeks <br />6.088e-6 months <br />, followed by the return of AC power and restart of the containment coolers.

If the containment atmosphere is flammable, the hydrogen is assumed to ignite when the molten core drops into the cavity.- This sequence has a high (approximately 90 psia) spike due to hydrogen deflagration at about 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> and a slow pressurization starting.at about 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />.

S HF-Burn:

Small LOCA plus loss of ECCS recirculation 2

capability, resulting in the loss of con-tainment spray recirculation capability.

If the containment atmosphere is flammable, the hydrogen is assumed to ignite when the molten core drops into the cavity.

This sequence has a modest (approximately 70 psia)

.apike due to hydrogen deflagration at about 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />, and a slow pressurization due to melt / concrete interaction.

16

S G:

Small LOCA plus loss of heat sink for con-2 tainment coolers and containment sprays.

This accident results in containment over-pressurization before meltdown.

The sequence has a slow pressurization beginning early in the accident and a high steam condensing load.

TMLB":

A variation of TMLB', in which AC power is assumed to return after about 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br />, leading to restart of containment coolers, containment sprays, and ECCS injection.

This sequence has the same pressure spike as TMLB', plus a very high steam condensing load when the ECCS and containment sprays start.

A-Vent:

Large LOCA causing erroneous actuation of containment venting.

All engineered safety features are assumed to operate on demand.

This sequence has a high potential for system interaction, because the engineered safeguards are all in operation.

The computer codes used in the RSS for determining the containment environment were used in this study.

These i

l codes have been extensively modified and improved since the RSS, and the latest versions were used.

l 17 l

The containment environment--pressure, temperature, composi-tion, and fission products--were calculated by the MARCH / CORRAL O

codes by Battelle Columbus Laboratories.

MARCH treats the thermal-hydraulic processes during and following meltdown, and calculates containment pressure, temperature, and composition.

9 CORRAL calculates fission product transport and deposition in the containment system.

10 Accident consequences were calculated using the CRAC code.

This code calculates statistically averaged health and economic consequences, using site specific meteorological and demographic data.

For this study the evacuation model used in the RSS was modified to reflect the idea that evacuation could be ineffective in areas of high population density.

The consequence calculations of this study were limited by time constraints.

Evaluation of the design consequences should include determination of dominant sequences for each plant, and consequences for all the dominant sequences, as well as consequences of faulty operation.

This should then be combined in an overall risk analysis.

The most severe design condition appears to be a rapidly risi.ng (of the order of a few seconds to a few minutes) pressure spike.

The pressure spikes occur in several ways:

rapid release of steam from the primary safety valves, flashing of primary water when the reactor vessel f ails at high pressure, rapid generation of steam when the molten core falls into water in the reactor cavity, rapid steam generation when the accumulators dump onto 18 f

the molten core, and hydrogen burning.

In some sequences many of these phenomena occur in rapid succession, and the pressure increments are ef fectively superimposed.

These rapidly rising pressure spikes (which will be responded to fully by the containment structure) impose a severe burden on any practical venting system.

Although the calculation of these pressure spikes may be somewhat conservative, we do not at this time see any means of excluding them from consideration.

A.3 Mitigation Strategies Each mitigation strategy can be thought of as reducing conse-quences for a class of accidents for which it obviates containment failure, and increasing the consequence for another class of acci-dents.

The relative contributions to overall risk can only be addressed by a detailed quantitative risk analysis, which is beyond the scope of this study.

In the absence of a risk analysis, it has not been possible to evaluate design strategies quantitatively.

Obvious advantages, disadvantages, and hazards are pointed out, however.

The design strategies are listed in the order of increas-ing operator involvement.

The venting strategies are synopsized in Table I, with a qualitative rating of advantages and disadvantages.

Vent Strategy 1 - Low Volume Containment Pressure Relief In this strategy containment is vented at a low rate (about 15,000 cfm) by a valve which opens when containment pressure exceeds a preselected setpoint, and recloses when the pressure falls below the set point.

The advantages of this strategy are 19

M O

Table 1.

Synopsie of the vent Strategy Options

• Not Likely to Exacerbate Independent Ahle to Mitigate High System Consequence Operator Free from Lower of Able to Mitigate Steam Spike Rydrogen Burn Reliability Interections Accidents Judgment i

1.

Low volume containment

+

+

+

+

pressure rel'4ef la.

Low volume containment

~

0

+

+

+

+

depressurtsation 2.

Migh volume containment 0

0

+

0 pressure rellet 2a.

High volume containment 0

0 0

+.

0

+

pressure relief plua low volume containment depressurisating 3.

Passive containment

+

+

0 0

0 flooding plus low volume I

containment pressure relief 3a.

Passive containment

+

0

+

0 0

0 flooding plus low volume containment depressurination 4.

Diversion of sump and

+

+

+

accumulator water plus low containment pressure relief ea. Diversion of sump and

+

0

+

+

volume containment pressure relief depressurisation 5.

Anticipatory primary

+

=

+

+

system depressurisation in-containment plus low i

volume containment depressurmatton Sa.

Anticipatory primary

+

0

+

+

system depressurisation in-contelnment plus low volume containment depressurination 6.

Anticipatory containment 0

+

0 depressurization 7.

Anticipatory primary system

+

+

0 depressurization es-contain-ment plus low volume containment depressurization "DetInitton et Symbols

+ denotes a highly favorable aspect, O denotes a moderately favorable aspect. - denotes an unfavorable aspect.

e a

e e

l its simplicity and the minimum potential for adverse ef fects.

A major disadvantage is that it cannot handle accidents in which containment fails because of a pressure spike.

~

Vent Stategy la - Low Volume Containment Depressurization This strategy widens the set of accidents which can be handled to those in which the containment is threatened by deflagration of the hydrogen produced in melt / concrete interaction and may' mitigate releases from delayed isolation failure.

It is similar to the previous strategy, except that venting continues even after the containment pressue falls below the set point. (Appropriate hardware might include a rupture disk. )

Disadvantages are that unnecessary venting often occurs, and ECCS recirculation failure can be initiated by reducing containment pressure below the saturation pressure of water in the sump.

(The latter hazard can be avoided if depressuri-zation is slow enough.)

This strategy still cannot handle many accidents in which the containment is threatened by a pressure spike.

Vent Strategy 2 - High Volume Pressure Relief The pressure spikes in accidents such as TMLB' and S HF-Burn 2

can be successfully mitigated by pressure relief at a rate of about 10 million cfm.

This would require using the 18 foot equip-l ment hatch as a vent.

It is doubtful that reliability comparable to that of other strategies can be achieved with such a large vent.

Venting could also be carried out at the more manageable--but still high--rate of 300,000 cfm using the normal 3 foot diameter purge penetrations.

This could handle steam generation caused by inad-vertent or delayed pumping of ECCS water onto the molten core.

It 21

would not be adequate to mitigate the major pressure spikes, unless the generation of steam is significantly slower than has been I

assumed.

Failure of the relief valves to reseat could present a major hazard.

Vent Strategy 2a - High Volume Containment Pressure Relief with Low Volume Depressurization This strategy could be accomplished by installing high volume rescatable relief valves in parallel with a low volume rupture disk.

This strategy combines the advantages of vent strategies la and 2.

The hazard of recirculation water flashing is also present with this strategy.

Vent Strategies 3 and 3a - Passive Containment Flooding Plus Low Volume Pressure Relief or Depressurization About 1 million gallons in the lower part of the containment could reduce the pressure spike caused by molten fuel falling into the reactor cavity.

It could also quench the core in.the cavity and retard basemat penetration, assist in cooling by its large heat sink, capture many of the fission products, and retard reactor vessel failure by cooling the outside of the vessel.

Conta inment flooding could be passively intitiated from a large elevated tank, and could possibly even be usable in containment sprays.

This would be.particularly attractive at Indian Point, which is surrounded by hilly terrain.

A major disadvantage is the effect on wiring and electrical equipment in the flooded part of the reactor building.

Care would have to be taken to ensure that contaminated material coul.d not back up into tne water line.

22

Operator judgment might be required in the decision to flood con-tainment.

A combination of containment flooding with low volume

~

venting might be used to remove noncondensables, although it is L

not clear that venting would be necessary in order to handle almost all accidents.

Vent Strategies 4 and 4a - Diversion of Sump and Accumulator Water Plus Low Volume Pressure Relief or Depressurization This strategy requires operator judgment that the core is no longer coolable under any circumstances.

Misjudgment could cause an otherwise preventable meltdown or facilitate basemat penetration.

Because of these hazards, this strategy does not appear attractive.

Vent Strategies 5 and Sa - Anticipatory Primary System Depressuri-zation into Containment Plus Low Volume Pressure Relief or Depressurization Deliberate depressurization of the primary system after most of the water has boiled off could be helpful in TMLB' or small LOCAs.

This would allow discharge of the accumulators into an_ intact primary system, retarding meltdown.

The delay in melting has been calculated to be 90 minutes for TMLB'.

The pressure spike at reactor vessel fail'ure would also be significantly reduced.

Deliberate depressuri-zation would require an operator judgment.

This has th-obvious disadvantage that an error in judgment could cause a LOCA that would not otherwise have happened.

This reliance on human judgment may pre-sent so many problems as to put a premium on alternative strategies.

Vent Strategy 6 - Anticipatory Containment Depressuriza tion The strategy would counter containment overpressurization by

" foreseeing" a core melt and venting containment in advance.

23

During the time between the initial core melting and reactor vessel failure, there is ample time to reduce containment pressure to near atmospheric, using a 2 foot diameter vent.

Venting could be initiated by pressure relief at a low volume rate, as in Vent Strategy 1.

High volume venting could be automatically initiated at any time that there is a strong presumption that core melt is imminent; operator override of automatic venting would be possible.

The operator could also initiate anticipatory venting.

The criteria used to make the decision could include high radiation levels in containment, high hydrogen concentrations, high temperatures in the reactor vessel or primary loops, prolonged low water level, and high containment pressure.

Valving arrangements can be designed for this strategy with high reliability and low unavailability.

Anticipatory containment venting introduces a larger window for unnecessary releases; it can be anticipated that most accidents would not lead to containment failure even without anticipatory venting.

However, the magnitude of the releases prevented if containment failure were averted would be much larger than any of the more numerous unneces-sary, releases.

In the absence of a complete risk analysis, it is ambiguous whether anticipatory containment venting does more overall good than harm.

There is also a possibility of deleterious system interactions when the containment is rapidly depressurized; spurious l

venting while the ECCS was operating could precipitate a meltdown.

l i

Against these major disadvantages can be balanced the possibility of

~

j handling most accidents, and of getting rid of hydrogen early in the accident.

24 l

ent Strategy 7 - Anticipatory Primary Depressurization Ex-Containment Plus Low Volume Containment Depressurization This strategy can prevent containment threatening hydrogen burning during small LOCAs and transients leading to meltdown and would also prevent a significant portion of the steam spike.

How-ever, this strategy may be unacceptable because of the enhanced possibility of interface system LOCAs, the possibility of operator error causing an ex-containment LOCA, and the possibility of caus-ing ECCS recirculation failure.

The external venting would take place through the filter system, but any spurious operation would necessarily release some radioactive material.

A.4 Vent Filter Design Options Four options for filtered atmosoheric venting and an option for contained venting have been considered.

Each system can operate passively for 16 hours1.851852e-4 days <br />0.00444 hours <br />2.645503e-5 weeks <br />6.088e-6 months <br />, af ter which time it is assumed that AC power '

is returned in TMLB'.

The systems will also operate successfully for any other accident, provided a vent strategy appropriate to the accident has been selected.

The primary condensing / cooling component is a vapor suppression water pool.

Suppression pools are a well proven method of cooling and condensing, require less volume than crushed rock or gravel for the same heat sink capacity, provide a long-term solution to heat removal with the addition of a heat exchanger, will trap most of the solid or liquid particles, and can be treated with sodium thiosulfate to trap most of the inorganic iodine and part of the organic iodine.

m 25 L

\\

\\\\

Y t

~

['^

\\

1 3

A suppression pool with a water space of 150,000 ft and an 3

air space of'150,000 ft would be adeq} ate for, condensing.

Cooling s

i h would be required after 16 hours1.851852e-4 days <br />0.00444 hours <br />2.645503e-5 weeks <br />6.088e-6 months <br />. g Pressure drop can be maintained N

y 4v below 2 psi.

u Placement of the suppression pool inside containment would be s

preferred.

However, this has not been considered'because of the s

inordinate difficulty of backfitting.

(

The sTzing of other components depends on the flow rate.

The x

sizes given in the following descriptions are those required for Vent Strategy 6.,

The normal 3' foot diameter purge lines allow a s

3 f

choked flow of more than 250,000 cfm into the suppression pool.

The flow of noncendensables out of'the suppressibn pool is limited N

oabout 12 inches.

to a maximum flow' of 50,000 cfm' b-fifice f

The nominal flow is 40,000 cfm.

Naturally,'a low volume vent sbrategy would require smaller filters.

Design Option 1 - Capture of particulates and elemental iodine

'4 This design has a gravel bed submerged in a suppression pool.

The filter efficiency of this device requires verification in a large scale test.

'(Figure Al.)

,1 Design Option 2 - Enhanced capture, of particulates and elemental iodine N-This. option has a suppression pool followed by a gravel / sand filter with a total frontal area of 100 ft x 120 ft and'a depth of

~

N 16 ft.

(Figure?A2.)

)

'*3 i

26 a

!s n

/

\\

\\

l 1

d (A)WITHOUTACPOWER O

~

~

, VACUUM B,REAKER LINE h

VENTLIfE a

__:_~

MMT

\\ - v or)

/

I, GRAVEL SUBMERGED IN ALKALINE WATER JL (B)WITHACPOWER O

-~

BLOWER h

HEAT EXCHANGER FrouRE A1, DESIGN OPTION 1 27

L (A) WITHOUT AC POWER

~

VACUUM BREAKER LINE h

VENT LINE I

NNbY m

/

VM ALVALINE WATER

/I

\\

A CONDENSATE

/

GRAVEL-SAND-GRAVEL (B) WITH AC POWER O

.L BLOWER n

8 2_

4 iT+Ii9##M m

1. G4. 6#M@

l HEAT EXCHNiGER FIGURE A2.

DESIGN OPTION 2 28

7 Design Option 3 - Capture of particulates, elemental iodine, and inorganic iodine This design has the suppression pool and gravel / sand filter

(

as in Design option 2, followed by an ignition source for burning

(

hydrogen, a gravel bed to condense water from hydrogen burning, a bed of impregnated charcoal (8 x 16 mesh, 1000 ft2 frontal area, 2 inch depth) and a set of HEPA or roughing filters.

The charcoal container is surrounded by a 20,000 gallon water tank for cooling.

(Figure A3).

Design Option 4 - Capture of particles, elemental iodine, organic iodine, and holdup of xenon In addition to the components described in Design Option 3, this design includes 100 tons of activated unimpregnated charcoal.

The bed would have 1000 ft2 frontal area and a 5.5 foot depth.

Dividers would maintain the correct L/D ratio.

This and the charcoal filter would be in a container submerged in a 100,000 gallon water tank for cooling.

The pressure drop across the xenon bed would be 0.4 psi.

Holdup of krypton would be impracti-cal--about 10 times as much charcoal would be needed.

(Figure A4.)

Design Option 5 - Contained Venting In this option, the containment would be vented through a suppression pool to a second containment.

An ignition source ignites hydrogen before the second containment, with gravel upstream and downstream for heat sink and condensing.

Because there are no filters, a full 250,000 cfm vent rate is usable.

3 The second building needs a volume of about 1,000,000 ft with a design pressure of about 62 psia.

(Figure AS.)

29

(A) WITHOUT AC POWER g

~

.= VACUUM BREAKER LINE h

VENTLifE n

g _ BYPASS i

ti,T!??M16'

__=--

^ :.; ; ;

B%R'~ M w

/

n

/

WWNW ZEOLITE - IMPREG. CHARC0AL -

ALKALINE WATER f

ROUGHING OR HEPA FILTERS

/

GRAVEL-SAND-GRAVEL-SPARK IGNITORS-GRAVEL (B) WITH AC POWER D

L BLOWER l

8 n

V i

4N

=

p______q.

k BLOWER l Wh%'WS3 yQ l

1 2;;- rt,m ;

u W4W%iWN4tW I

W=N'im i

\\

/

l HEAT EXCHANGER FIGURE A3.

DESIGr4 OPTION 3 l

30

a (A) WITHOUT AC POWER D

~

, VACUUM B,REAKER LINE h

VENTL!fE a

n I

g-_BLP,A_5S_

N l

'a^m ^mr,"uc m

l B 9'*2 A m

/

/

WM4;99Ptt,

/'

ALKALINE WATER

/'

/

y ZEOLITE - IMPREG CHARCOAL.

/

GRAVEL-SAND-PLAIN CHARCOAL - ROUGHING GRAVEL-SPARK OR HEPA FILTERS IGNITORS-GRAiEL

^

(B) WITH AC POWER O

BLOWER n

g v

BLOWER g;MWetwaf; g

~

_ _ ~

&+W%,sm WW W#

\\

/

HEAT EXCHANGER FIGURE A4.

DESIGN OPTION 4 31

(A) WITHOUT AC POWER

, VACUUM BREAKER LINE O

h VENT LINE n

I sp,; m;;

DPl!RLRGR!iWE e

/

/i ALKALINE WATER f

GRAVEL-SPARK IGNITORS -

GRAVEL (B) WITH AC POWER O

BLOWER A

8Y V

9

=

n I

k EBEU3REE 1

+;;w z,n; s

@@iSWPJBd!Ni2

+

x

~

~

HEAT EXCHANGER FIGURE A5.

DESIGN OPTION 5 32

T A.5 Evaluation of Design Options MARCil code calculations for Vent Strategies 6 and 7 have been carried out both for filtered atmospheric venting and for venting to another containment.

The MARCli logic for venting to a suppression pool is in some respects unrealistic, although the results have been hand corrected.

The corrected results indicate that peak pressures for TMLB' can be significantly-reduced--by 37 psi for atmospheric venting or 27 psi for contained venting.

This reduces the pressure spike below the containment failure pressure, although design pressure is still exceeded.

Time has not permitted detailed calculation of the ef fects of all the vent strategies with the entire spectrum of design accidents.

Comparative consequences for sequence TMLB' have been 10 calculated using the CRAC code for each of the design options listed in Section 5.1.4.

The calculations have assumed that a vent strategy had been selected that averted containment failure, that the vent system worked as designed, and that i

the containment would have failed catastrophically without venting.

The risk increment due to spurious or inadvertent operation and lower risk reductions due to possible imperfect operation need to be considered.

The magnitudes of risk reduction are thus optimistic, and only indicate the ideal target toward which a design might be directed.

33

i

~

For these calculations site specific meteorology and populations were used.

Evacuation was limited to a 5 mile radius from the site, reflecting the idea that evacuation may be ineffective in regions of high population density.

The calculations show that Design Option 1 eliminates all early fatalities, reduces latent cancer fatalities by an order of magnitude, and reduces land interdiction by two orders of magnitude.

Design Option 2 reduces latent cancer fatalities by an additional 1-1/2 orders of magnitude.

Design Options 3 and 4 provide only small amounts of consequence reduction beyond Option 2, and thus appear less cost-effective.

These calculations only indicate the relative merits of each design option--under the assumptions used for the analysis--and by no means indicate the overall value of filtered atmospheric-vonting.

Some of the most aerious possible failure modes of a vent filter system are:

sand-filter blockage, suppression pool water loss, and hydrogen explosion in the wetwell.

These failures coul,d result in unfiltered releases.

Improper operation of filters could result in reduced mitigation.

Spurious operation of valves could result in unplanned releases.

Failure of valves to operate could result in the system responding as if the FVCS wore not p' resent.

A.6 Cost Estimates The estimated costs at Indian Point, using the existing 3 foot diameter purge penetrations, with a 1 foot filter orifice, 34

including labor, overhead,and profit and 25 percent contingency, but not including automatic control logic or costs for licensing and shutdown time, are approximately $16 million for Design Option 1 and $32 million for Design Option 4.

These costs correspond to a seismic Category I structure.

The estimated costs at Zion for the same conditions are very similar.

The use of a low volume rate venting strategy could reduce these costs.

• A.7 Design Specifications If the decision is made to backfit the Zion and Indian Point plants with an FVCS, detailed design specifications will have to account for a wide variety of conditions under which the system will be implemented.

A checklist of system specifications that need to be established is given below.

Detailed quantification of these items cannot profitably be addressed without a detailed quantitative assessment of the competing risks of various vent strategies.

I.

Functional Requirements A.

Containment pressure reduction B.

Containment temperature reduction, if necessary C.

Mitigation of radioactive release II.

Operatic 731 Requirements A.

Decontamination factors for important isotopes B.

Quality assurance criteria (especially for sand filters)

C.

Maximum filter loading capacity and re-entrainment characteristics D.

Maximum and minimum flow rates and pressure drops i

35

E.

Heat removal and condensate drainage requirements F.

Capability to withstand operating environment G.

Instrumentation requirements III.

Resistance to Hazards A.

Resistance to earthquakes, tornadoes, and missiles B.

Resistance to fire and hydrogen explosions within filter system C.

Resistance to steam explosion from within containment IV.

Reliability A.

Valve actuation reliability B.

Reliability of mechanical components (air coolers, hydrogen recombiners, heat exchangers)

C.

Likelihood of spurious operation D.

Likelihood and impact of human error E.

Filter failure or bypass modes and likelihood of -

occurrence F.

Emergency power requirements G.

Redundancy V.

Control A.

Actuation logic B.

Flow rate control C.

"Zero-release" options VI.

Sabotage Protection A.

Passive operation versus operator control B.

Protection of piping, valves, and filters from unwanted access l,

VII.

Inspection' Considerations 1*

A.

Ease of access B.

Frequency of inspection 36 l

C.

Inspection objectives:

1.

Evidence of structural damage or degradation 2.

Water infiltration, weathering 3.

Contamination with foreign matter D.

Impact on plant operating procedures VIII.

Testing Considerations A.

Frequency of testing B.

Testing objectives 1.

Ef ficiency degradation versus time 2.

Flow resistance versus time 3.

Component availability C.

Testing methods IX.

Maintenance Considerations A.

Ease of access B.

Periodic replacement of filter materials (especially charcoal)

C.

Grooming of filters (especially sand bed)

X.

Postaccident Safety and Repair-Restoration Considerations A.

Shielding criteria B.

Access to plant after accident C.

Difficulty of restoring reactor to service 4.8 Implications of Unresolved Issues To show that a vent-filter system is beneficial, it is necessary to show that the risk reduction potential for acci-dents which would have led to containment failure without an 37

FVCS is greater than the risk increment for accidents which would not have led to containment failure and in which the FVCS results in increased releases.

Thus, the most important remaining task is an overall risk analysis for each venting strategy and design option.

In addition, several unresolved phenomenological uncer-tainties need resolution.

Some of these have application to the rise time of pressure spikes; e.g., vessel f ailure mode, molten core-water interactions in the cavity, rate of steam production from accumulator dump, mode of deposition of the core in the cavity, and whether hydrogen is ignited in one massive burn or a series of smaller burns.

If the calculation of pressure spikes is excessively conservative and if the spikes are not as threatening as they now appear, some of the simpler venting options would look more attractive.

Other unresolved phenomenological issues impact the officiency of saitigation, e.g., decontamination factors for very large suppression pools, loading limits for sand filters, cnd' efficiencies of gravel and crushed rock for steam conden-cation.

A.9 Summary and Conclusions The major conclusion is that venting strategies and design options are available that appear to offer major consequence reductions for a specific accident under specific, optimistic 38

conditions.

The relative merits of the several venting strategies and design options cannot be assessed without considering the overall risk reduction of each strategy and each option.

The most challenging design issue appears to be a pressure spike from rapid steam generation and/or hydrogen burning.

The calculation of these spikes is believed to be conservative, but we do not believe they can be ruled out of consideration.

Some venting strategies can handle even the most severe spikes and some cannot.

In the absence of a complete analysis of all com-peting risks, it is not clear whether it is preferable to design a system for the most challenging pressure spikes or to concen-trate on a more limited accident spectrum.

The resolution of the most important uncertainty--the ef fect of competing risks--is underway.

Appendix B Task II:

Steam Explosions B.1 Objectives The objectives of this task are to determine best estimates of the work potential from steam explosions and the loads on the primary system, to determine the realistic potential for generat-ing missiles capable of penetrating containment, and to evaluate the major-uncertainties.

B.2 Discussion of the Problem There are six basic, interrelated components in the determina-tion of threats to the containment integrity from steam explosions:

39

1) the extent of premixinig of molten material and water, 2) the I

dynamics of the thermal interaction, including molten material frag-mentation, 3) the configuration of the system at the time of the steam explosion, 4) the delivery of dynamic loads to the primary system, 5) the response and f ailure modes of the primsry system, and

6) the penetration of structures by missiles, if any are generated.

The first three components have the greatest uncertainty.

The extent of premixing in a pouring mode is necessarily limited.

Experimental evidence strongly suggests that an explosive event is triggered no later than when a premixed region of molten core material and water contacts a solid structure.

The present judgment is that the pouring mode would be very unlikely to involve more than 10 to 20 percent of the core material in a coherent interaction.

The detailed interaction physics subsequent to premixing is not well known, but the general characteristics can be established from the available experimental data.

Experimental pressure tran-sients indicate a rapid heat transfer period of perhaps a millisecond or less.

Based on the observed initial rapid pressure increases, it cppears reasonable to assume a high degree of fragmentation of both the core material and water.

This is followed by a rapid pressure decay.possibly due either to preferential venting through a mixed fuel-water-vapor zone above the interaction region or entrainment of surrounding cold water.

The system configuration influences the mechanical energy generated by the steam explosion, particularly through the degree 40

of inertial confinement.

The amount of entrained material influ-ences the energetics of the explosion through heating due to fuel entrainment, cooling due to massive water entrainment, or addition of working fluid due to limited water entrainment, and hence influences the dynamic loads to the structure.

Currently, a single most probable configuration cannot be justified.

Furthermore, there is no one-to-one relationship between the work potential of the explosion and the final impulse applied to the system, nor between impulse and the damage potential.

Because of these uncertainties, a unique best estimate is neither possible nor meaningful, although a reasonable range can be established.

B.3 Accomplishments This study has investigated a range of best estimate explosive energies and configurations and evaluated the damage potential of steam explosions within this range.

Because of the phenomenological uncertainties and the finite number of cases that could be consid-ered, additional points lying outside the likely range were also investigated.

'The range of steam explosion work potentials and initial config-urations was independently investigated at SNL and LANSL.

A wide variet of postulated f uel/ water ratios, expansion characteristics and slug types were investigated.

In the LANSL work, the SIMMER II two-dimensional fuel / coolant interaction codell was first calibrated against Sandia steam explosion experiments.

After calibration, two-dimensional calculations were 41

performed with varying f ractions of the core in the premixed zone, differing high pressure equations of state, and mixing configura-tions.

Some of the calculations showed lower plenum failure early in the explosion.

Venting through the lower plenum in these cases would ef fectively limit any threat to the upper head.

For those explosions that could credibly deliver a slug to the upper head, 1000-1500 MJ appears a reasonable limit for the slug kinetic energy.

The dynamic loads were in every case centrally biased.

The effects of slug shape were also investigated; it was found that a centrally biased slug is the most likely.

The ef fect of the fluid kinetic energy on the primary system structural response depends on the coherency with which the momentum of the fluid is delivered to the structure.

It appears that the fluid dynamics of the upward directed steam expansion is strongly af fected by the behavior of the pool of water or molten core material above the interaction region.

Two-dimensional analyses as well as the experimental data strongly suggest the breakup of the pool into a more dispersed configuration.

The result is an impact with the reactor head which has less damage potential than a coherent slug.

The Sandia estimate considered mixing 10 percent of the core with 10 MT of water; a one-dimensional isentropic eryansion gave a slug energy of 300 MJ.

It should be noted that isentropic expansion does not give maximum slug energy.

The SIMMER calculations more nearly approach the maximum because of continued heat transfer during

~

expansion.

Mixing of 40 percent of the core with 20 MT of water was also concidered.

This configuration--which gives a slug energy of 3000 MJ--is out of the range of best estimate configurations.

42 e

The range of estimated slug energies runs from 300 to 1500.MJ.

Tne high, out-of-range estimate was also considered--even though it is physically unlikely--in order to gain insight into the ultimate strength of the reactor vessel head.

The most energetic slugs were calculated for highly tamped explosions.

In these the fuel not participating in premixing formed an initially coherent cover over the mixing zone except for relief up the downcomer.

The configuration and extent of mixing at the time of fuel / water contact is subject to great uncertainty, however.

Low power zones at the core periphery would probably be only partially molten at the time of support plate failure and could alter the course of the expansion, and the assumed configuration of a uniform thickness pool above the mixing zone is highly idealized.

Furthermore, limited intermediate scale experiments with corium A+R and water indicate a trend toward low explosive efficiency.

For these reasons, the proba-bility distribution of explosive energies is believed to be skewed toward the lower end of the range.

Loading conditions on the head were approximated as follows:

'l )

A hemispherical water slug with a kinetic energy of 300 MJ.

2)

A hollow, hemispherical water slug with a kinetic energy of 300 MJ.

3)

The spatially and time varying pressure from the highest SIMMER calculation which would not have i

failed the lower head.

i I

43 i

4)

The peak pressure, as in 3, uniformly distributed over the head.

5)

A hemispherical water slug with a kinetic energy of 3000 MJ.

6)

Impact of the upper core support plate with a kinetic energy of 300 MJ.

Head failure was predicted in some calculations; however, the failure mode involved meridional (North-South) fractures commencing ct the pole.

Such failures would not produce large mass missiles.

Previous estimates of large missile generation have been based on excessively conservative analyses and failure criteria.

Steam explosions could generate small mass missiles, corresponding to control rod drive assemblies.

Conventional empirical penetration formulas, which are known to be conservative, indicated that some of these missiles could penetrate the overlying missile shield.

Detailed calculations with the CS012 code which more realistically represented the expected situation showed that the missile shield would be penetrated only at the upper end of the velocity range and'that any missiles penetrating the shield would be destroyed in the process and hence would present no threat to containment integ-rity." For the high out-of-range estimate, failures leading to separation of the head, or a portion of the head, could not be confidently excluded.

However, dynamic analyses showed that the head velocity was too low to constitute a threat to containment even if separation occurred.

44

Quasistatic overpressures--as distinguished f rom explosions--

could be generated in the interaction of fuel and water.

These overpressures exceed the static strength of the vessel which has been determined to be 40-50 MPa.

Static vessel failure is not expected to present problems other than those already dealt with in Task I, namely, release of steam into the containment.

However, further investigation of vessel f ailure at ultra high pressure is needed to understand all the ramifications of this failure mode.

Steam generator tube failure due to quasistatic overpressuri-zation has also been investigated and found to be possible.

Steam generator tube failure could present a path for relcase of fission products outside of containment unless the main steam isolation valves were closed.

It is therefore recommended that positive means of isolating the steam generators be investigated.

As a consequence of the particular geometry at Indian Point and Zion, ex-vessel steam explosions are not capable of generating threatening missiles; f urthermore, the shock pressures due to ex-vessel steam explosions are not significant for the containment structure.

However, ex-vessel reactions Letween fuel and water could generate large volumes of steam, with attendant quasistatic pressure surges in containment.

B.4 Implications of Unresolved Uncertainties Physical effects exist that mitigate the explosive damage potential.

There is some evidence that the likelihood of explo-sions is diminished at high pressure; however, the experimental data are in conflict and as yet insufficient to conclude that 45

explosions are absolutely precluded.

The solid structure remain-ing in the vessel could inhibit fuel / coolant mixing and reduce explosive yields.

The water not vaporized in the interaction could alter the explosion pressure by further heat transfer.

Lower head f ailure early in the expansion has been mentioned as a limiting factor.

Based on simple bounding calculations, this limit appears to be approximately 1500 MJ.

Time constraints N.

have prohibited the extensive combination of structural and inter-action calculations that would be required to determine this limit more precisely.

It is unlikely that knowing the effects of head failure better would alter the conclusions, because a slug with a kinetic energy of twice the apparent limit has been determined unlikely to tnreaten containment.

A fundamental and controlling uncertainty is that associated with the degree of participation of core material.

For the extent of premixing to become larger than some fraction of the total core, a high degree of uniformity in the pouring of the core into the water and a precisely timed trigger are required.

For greater involvement, the mixing process must change to one in which water percolates up through the molten core to establish premixing.

This phenomenon is highly unlikely as a mechanism for preiaixing.

I There would be a long time frame during which premixing in other regions would disappear by gravity separation.

Coupled with this argument is the strong tendency for dynamic loads on the head to be reduced with a greater degree of core involvement.

This occurs because the pool above the interaction region becomes more shallow, 46

thereby having a greater tendency to disperse and impact the reactor i

head as a spray.

Thus, although the degree of core material partici-pation is uncertain, the potential for obtaining higher estimates of containment damage through steam explosion is still very low.

Another important question is the degree of inertial confinement.

Phenomenological uncertainties make a unique realistic estimate of the system configuration impossible (see Sectic n B.2).

Although the idealized confinement used in this study cannot be excluded, the actual situation can be expected to depart from the ideal.

Further calculations could help to quantify the mitigating effects of depar-ture from ideality.

The effects of tamping could perhaps also be experimentally verified, but it is not possible to specify the degree of tamping to be expected in a core meltdown.

Because this study has been largely based on ideal confinement, and because slug energies appear limited by lower head f ailure anyway, better knowledge of the effects of confinement, although intensely interesting, would probably not change the conclusions.

One uncertainty that might change the conclusions about the damage potential of steam explosions is related to the failure modes of the reactor head if loading distributions change from centrally biased.to more uniform.

The uncertainty is mainly associated with higher strecses away from the pole in the Indian Point reactor due to head reinforcement around control rod penetrations.

This tendency could lead to the generation of larger missiles.

However, it is expected that the velocity of the missiles would be too low to pene-trate the missile shield and containment.

47 I

4 B'. 5 Conclusions The estimate of the range of work potential in steam explosions in the reactor vessel is 300-1500 MJ.

The lower end of the range is believed to be more probable.

It has been determined that explosions in this range would be unlikely to pose a threat to the containment at Zion and Indian Point and would therefore not negate the effective-ness of an FVCS.

This conclusion is specific to Zion and Indian Point and does not apply to other plants.

It is expected that resolution of the phenomenolog'ical uncertainties would tend to strengthen this conclusion.

Quasistatic steam pressure developed in the interaction could fail steam generator tubes, which could allow radioactive material to bypass the containment barrier.

Quasistatic steam pressures could also fail the reactor vessel.

Ex-vessel steam explosions do not pose a threat to containment at Zion and Indian Point.

Appendix C Task III:

Hydrogen Burning C.1 Objectives The objectives of this task are to define the loads delivered l

to the containment from hydrogen burning, to evaluate the contain-ment response to those loads, and to determine a best estimate of the threshold level for containment failure from hydrogen burning.

C.2 Work Accomplished The threat to containment from deflagration or detonation of hydrogen was investigated by determining pressure pulses generated 48

by burning with a variety of hydrogen concentrations, initial water vapor content, and initial temperatures.

The hydrogen was assumed to be homogeneously distributed throughout this volume.

The initial conditions were representative of those to be expected in core melt-down accidents but are not specifically tied to particular sequences.

They can therefore be applied to a wide range of sequences.

Loadings on the containment were bounded by considering both isochoric burns and detonations.

The latter are very conservative; detonations are impossible for the concentrations considered.

How-ever, very rapid turbulent deflagrations can produce dynamic loads in excess of quasistatic burns.

A " pseudo-detonation" would be an extremely conservative upper limit on turbulent deflagration loads.

The kinetics of the detonation process was investigated parametri-cally by varying threshold pressures and reaction rates.

The CSQ code 12 was used for calculating the " pseudo-detonations."

At hydrogen concentrations comparable to uniform mixing of the total quantity of hydrogen produced by metal / water reactions an isochoric burn from initially dtmospheric pressure would not produce a clear cut threat to the containment.

However, burn from a modest initial pressure appears dangerous to the containment integrity.

At concentrations comparable to uniform mixing of half this quantity, detonation pressures--without consideration of structural response--

appear to exceed containment failure pressure; however, detonations would not be sustainable at these concentrations.

Shock-wall and shock-shock interactions cause locally higher pressures in the deto-nation calculations, but the durations are very short.

49 b

Analyses of containments at Zion and Indian Point (the Zion-containment was analyzed at LANSL and the Indian Point containments ware analyzed at SNL) showed the structures at both plants to have similar ultimate static load capability in spite of great differences in design (8.6 x 105 Pa at Indian Point and 1.05 x 106 Pa at Zion).

Dynamic calculations showed that both containments would respond to an isochoric burn as if it were a static load.

It thus appears that any isochoric burn giving pressures higher than 8.6 x 105 Pa (125 psia) at Indian Point or 1.05 x 106 Pa (153 psia) at Zion would fail containment.

Lower pressures cause damage such as concrete cracking, which might allow escape of some fission products if the liner leaked.

6.9 x 105 Pa (100 psia) is recom-mended as a reasonable estimate for the threshold of failure at either plant.

This corresponds to burning of approximately the quantity of hydrogen expected from metal / water reactions with a modest initial pressure; the precise amount depends on the initial conditions.

When the internal pressure at Zion exceeds 5.6 x 105 Pa (82 psia), a moment reversal occurs for which the building was not designed; the ef fect of this has not been determined.

Although the two containments have similar static pressure capability, their dynamic behavior is somewhat different.

A dynamic analysis of a detonation pulse having a peak pressure of 6

1.7 x 10 Pa (250 psia) at Indian Point showed considerable damage, but not catastrophic failure.

A detonation pulse with a peak pres-sure of 1.4 x 106 Pa (200 psia) applied to the Zion containment showed damage that could be regarded as complete failure.

It should 50

be pointed out that detonations would have been impossible at the hydrogen concentrations considered.

However, locally high concen-trations might allow detonations in a restricted region, even though the average concentration is not detonable.

Asymmetry could not be investigated because of the short time constraints of this study.

Asymmetric ignition is quite possible; however, our current belief is that asymmetry would not cause gross effects in an isochoric burn.

In a detonation, much of the loading comes from shock interactions and reflections, and asymmetry could play a significant role in detonation pulses.

At the present time, we are not able to estimate what the ef fects would be.

Containment penetrations were also investigated.

It has been determined that even the largest penetrations are so heav!1y reinforced that no major weakening of containment is expected.

The thermal ef fects of hydrogen burning on the containment structure have been evaluated.

There does not appear to be a structural problem due to the elevated temperatures.

However, we have not investigated whether or not elevated temperatures and pres.sures would cause problems with safety related components.

C.3 Summary and Conclusions An isochoric burn of about the amount of hydrogen expected from metal / water reactions could seriously endanger containment at either Zion or Indian Point if there is a modest initial overpres-sure.

The average containment composition with this amount of hydrogen would not be detonable.

Therefore, whether detonation 51

would also fail containment may be moot.

Locally richer concen -

trations could conceivably allow detonations in restricted regions.

This might cause local f ailures at somewhat lower average concen-trations.

For the purpose of FVCS design an isochoric burn 5

pressure of 6.9 x 10 Pa (100 psia) is recommended as a reasonable threshold failure pressure.

Some permanent damage is expected at lower pressures.

Pressures of 8.6 x 105 Pa (125 psia) at Indian Point or 1.05 x 106 Pa (153 psia) at Zion are the best available estimates for catastrophic static failure pressures.

The effect of hydrogen burning on safety related components has not been evaluated.

It has been determined that the thermal effects of hydrogen burning would not cause structural problems.

The principal unresolved questions are:

a)

The effects of asymmetry b)

The effects of other gases (CO, CO, CH4) 2 c)

The effects of nonuniform concentrations d)

The level at which damage should be regarded as catastrophic Resolution of these uncertainties might change the estimates of failure pressure either upward or downward.

Present analytical metho's are not capable of quantitative resolution of these d

uncertainties.

m 52

Appendix D Task IV:

Hydrogen Control D.1 Objectives This task encompasses a study of hydrogen generation, solu-it bility, combustion, control, monitoring, radiolysis and detonations in primary systems, and development of a hydrogen accident compendium.

D. 2 Results of Survey By far the major sources of hydrogen in meltdown accidents are metal / water reactions and melt / concrete interaction.

Metal /

water reactions produce about 1000 Kg of hydrogen ani melt / concrete interactions could produce a comparable amount.

Burning of 1000 Kg represents a threat to containment integrity (see Section C.3).

Current control methods--recombiners--are flow-rate limited and are of no use in serious accidents involving large amounts of hydro-

\\

gen.

Chemical gettering schemes are also flow-rate limited.',

\\

Transport of hydrogen is important both in the primary systep and in containment.

Kinetics of solubility in boric acid solutions is almost unknown (however, a recent Japanese paper has appeared on this subject).

Detection and monitoring of hydrogen is extremely important.

Knowledge of hydrogen concentration is imperative if operator action is to play a role in accident mitigation.

An instrumentation devel-opment program would be required for adequate gathering and display of hydrogen concentration information.

53

b s

s

%s

?

s Calculations of pressure due to hydrogen deflagration or det.o-nation can be carried out for simple situatiots.

Variances from the simple calculable cases--asymmetry, nonuriiformity, ef fects, of contaminants--require experimentation at large s'cales. ?The results of laboratory scale experiments may not be~ translatable to contain-(

mont sizes.

s One effective method.of control is deliberate buNning at, low

~

concentration =

Industrial experience with this method' exists.

Spark igniters could be placed around containment and automatically ectuated.

The reasoning is that several successive burns--each at low concentration--has much lower damage potential than a single large burn.

~

3 Injection of water fog into the containment atmosphere could be a useful method of suppression.

Volume fractions as low as 0.05 oercent could be useful.

Completely passive injection methods x

c::n be visualized.

A significant experimental progran--including l

large scale tests--would be required for developmente Halocarbon fire suppressants'could be used to s prAss the j

flammability limits.

Total inerting is also p$ssiblE, but very i

4 i l

large volumes of the inerting gas would be required.

Accidental suf focation risk should be considered for any inerting policy.

Venting the containment could remove hydrogen as it was gen-erated.

In some accident sequences, some of the venting stategies

listed in Section 5.1.3 would maintain hydrogen below the danger level.

Primary system venting outside containment could be partic-l ularly useful for hydrogen control, althot2gh the many disadvantages and hazards of this strategy need to be carefully weighed.

54

1 Oxygen could be controlled by hydra:ine, although the process liberates large amounts of heat.

Oxygen control also involves operational problems such as difficulty of inspection in nonbreath-able atmospheres that could conceivably increase overall risk.

~

D.3 Implications of Unresolved Uncertair. ties None of the control methods described can be considered current

" state of the art" at containment scales.

No matter what system of monitoring and control is chosen, a considerable development effort will be required.

The first priority should be a more quantitative ranking and evaluation of control methods.

Based on our present knowledge and inclination, we believe that multiple ignition sources and water fog appear to be the most attractive near-term alternatives.

D.4 Summarized Conclusions Hydrogen produced in metal / water reactions and melt / concrete interaction presents a threat to containment integrity.

Current monitoring is inadequate as a basis for operator judgment.

Currently installed control methods are not suitable for the quantities of hydrogen expected.

An PVCS, with proper venting strategy, would provide control 1

in some accident sequences.

Several other control methods are l

pcssible, but all would require extensive development.

Multiple o

ignition sources and water fog appear attractive for relatively l

near term development.

l l

l 55

Appendix E Task V:

Melt Interactions with Concrete /MgO and Basemat Melt-through E.1 Objectives The objectives of this task are to determine gas and aerosol production from interaction with concrete or Mgo, to predict the time of basemat melt-through, and to evaluate the ef fectiveness of a core retention device at Zion and Indian Point.

E.2 Results Calculations of melt / concrete interaction for representative sequences were performed with the WECHSL13 code.

In every case, the melt solidified before penetrating more than a fraction of the basemat.

The time to penetrate the remainder of the basemat was estimated, using the admittedly inadequate experimental data base.

Penetration times were calculated by this method to range from three to four days for sequences TMLB' and AB.

Aerosol production was found to vary by more than an order of magnitude, depending on the initial conditions.

The latter are poorly known, and a conser-vative best estimate of 500-1000 kg is suggested.

The gas production and penetration predictions of WECHSL were compared with those calculated by INTER 14, which were used for con-tainment pressure determination.

1NTER predicts approximately twice as much gas as WECHSL.

Both codes are conservative in that the concrete type used in the analysis produces more gas than the

~

concrete expected at those plants.

On the other hand, neither code accounts for gas production from radiatively heated concrete, which could be significant.

Pending resolution of these uncertainties, 56

the INTER prediction (which has been used in MARCH calculations) is believed to be reasonable.

Neither WECHSL nor INTER properly accounts for all gas species.

A preliminary calculation, using more accurate equilibrium data,

+

showed high concentrations of methane.

Bounding calculations of the minimum floor area required to prevent penetration were performed.

The methods used are necessarily simplistic and the assumptions may not be realistic.

However, the calculations suggest that penetration arrest is unlikely for the configurations at Zion and Indian Point.

The available experimental data on interactions of core melts with candidate retention materials has been reviewed, with particu-lar attention to Mgo.

The interaction has been found to be primarily chemical rather than thermal.

The nature of the reaction could be eutectic formation, grain boundary dissolution, or a combination of these.

The threshold temperature and heat of reaction are not known.

A parametric study of melt /MgO interaction was performed, using a model developed for this study.

It apppears that the parameters under the designer's control (the dimensions of the device) or those that can now be estimated (melt temperature and composition) have less ef fect on device performance than do parameters which are not known (such as threshold temperature and heat of reaction).

1 The results of the parametric study were combined with an admittedly scanty experimental data base to estimate the eff!cacy of a backfitted core retention device at Zion and Indian Point.

The best estimate is that such a device would delay melt / concrete 57

intoractions for a few hours, or at most a day.

During this delay tima gases and aerosols would be reduced but not eliminated.

Af ter failure of the device, a typically violent interaction with concrete is expected, with gas and aerosol evolution proceeding as if no device had been present.

The best estimate is that conse-quences will not be increased by the installation of a core retention device, and benefits could be realized for some accident sequences.

E.3 Implications of Unresolved Uncertainties Uncertainties in heat transfer from molten core debris are important for gas release and aerosol production.

These uncertain-I ties (including the fraction of radiative heat loss) are relatively less important in estimating melt-through time.

Melt-through time is dominated by uncertainty in the behavior of hot solidified debris.

This uncertainty also makes it impossible to state with confidence whether penetration arrest will occur.

The effect of water in the reactor cavity is open to question.

The effects could range from a modest increase in penetration time to total prevention of penetration.

Uncertainties concerning melt /MgO interaction are of paramount importance in evaluating the efficacy of a core retention device in the limited space available.

These uncertainties would be less I

controlling if a backfitting method could be developed allowing a thicker layer of MgO.

However, in view of the present lack of knowledge, holdup of the melt for several days cannot be confidently predicted.

I 58

~

E.4 Summary and Conclusions Basemat penet' ration cannot be positively confirmed or excluded.

The best estimate for penetration time, if it occurs, will be 3 to 4 days.

If penetration occurs, the material entering the subsoil l

will be solidified.

Positive arrest of penetrations is thought to be unlikely at

(

Indian Point and Zion, unless the core debris is water cooled and fragmented.

Current data on melt /MgO interactions are inadequate for con-fident evaluation of a specific design of core retention device at those plants.

The best estimate is that a core retention device at these plants would delay melt / concrete interactions for a period ranging from a few hours to one day.

An increase in hazards due to core retention devices is not anticipated, and some benefit in reduced gas and aerosol penetration is possible.

Appendix F Task VI - Meltdown Phenomenology F.1 Objectives The objectives of this task are to study current models of the meltdown sequence, to determine the predictive capability and uncertainties of current models, to investigate debris coolability, and to determine the impact of restoration of power.

F.2 Work Accomplished O

A survey of the MARCH code has been performed and several areas of uncertainty have been identified which might dominate the melting 59

sequence.

Major uncertain, ties which bear on FVCS' design include:

fuel motion, time of vessel breach, mode of vessel failure, state of materials exiting vessel, occurrence and scale of interactions, and mode of entry into lower plenum.

These phenomenological uncertainties dominate our understanding of the configurations at the times of fuel / water contact, pressure vessel failure, and entry into the reactor cavity.

These uncertainties cannot be well bounded at this time.

MARCH models are simple, fast, and straightforward.

Studies of the interaction of melts and water (ignoring the energetics of steam explosion) have been carried out.

It-is not possible either to assure coolability or to definitely rule out the possibility of coolability either in-vessel or ex-vessel.

The production of a rapid steam spike of the order of 50 psi (almost independent of accident sequence) when a large quantity of melt is introduced into ample quantities of water in the reactor cavity cannot be precluded from consideration without better knowledge and further analysis.

Pressure transients within the vessel could fail steam generator tubes.

Less than catastrophic vessel failure might reduce pressure spikes.

Restoration of AC power within 3-1/2 hours would be likely to prevent core damage in T!4LB' (although steam. generator tubes might be ruptured if cold feedwater were suddenly supplied).

Restoration before 4-1/2 hours could probably I

save the primary vessel.

Restoration thereaf ter might allow cooling i

of the debris ex-vessel, although sudden dumping of ECCS water onto core debris could cause a pressure spike.

60

~

A parametric study of fission product behavior shows wide uncertainty bounds.

CORRAL 9 appears to fall within the band of likely behavior.

F.3 Implications of Uncertainties Because of time constraints, it has not been possible to quantify the uncertainties which have been identified.

Design specifications for an FVCS should reflect that these uncertainties exist, however.

If current methods of calculation of pressure spikes are in fact too conservative, simpler FVCS designs might have a high probability of success.

The possibility of massive steam generator tube rupture requires investigation.

This failure could allow radioactive material to bypass the containment barrier, making an FVCS irrelevant.

Uncertainties in fission product behavior in the containment impact strongly on the uncertainties of risk assessment, although present calculational tools do not appear unreasonable.

F.4 Summarized Conclusions The uncertainties inherent in calculational models translate to uncertainties in results that may dominate the predicted design conditions for an FVCS.

The steam pressure spikes, identified in Task I as the limiting design condition, cannot be ruled out at this time for some core melt sequences.

An additional concern is the possible bypassing of containment through steam generator tube failure either in high pressure meltdown or from thermal shock on restoration of feedwater.

61

Appendix G:

Study Personnel In addition to the personnel shown (who have devoted a major share of time to the project), many others have contributed valu-able ef forts as consultants, reviewers, and assistants.

NRC Project Manager:

C.

N.

Kelber Project Leader:

W.

B.

Murfin LANSL Coordinator:

M.

G. Stevenson Task No.

Task Title Principal Investigators I

Design of FVCS A.

S.

Benjamin (Lead)

H.

C. Walling G.

J.

Boyd R.

A. Haarman (SAI)

D.

T. Pence (SAI)

P.

Cybulskis (BCL Lead)

R. Wooton R.

G. Jung II Steam Explosions M. L. Corradini (SNL Lead)

D. V. Swenson M.

Berman C.

R. Bell (LANSL Lead)

W.

R.

Bohl R.

Alcouffe R.

J.

Henninger III Hydrogen Burning A.

Loads M. Berman (Lead)

R.

K.

Byers S.

L. Thompso n B.

Containm nt Response W. A. Von Riesemann (SNL Lead)

M. Huerta E-P Chen C.

A. Anderson (LANSL Lead)

J.

G.

Bennett F.

Biehl s

T.

Butler L.

Carruthers l

K. Cooper IV Hydrogen Control M.

P. Sherman (Lead)

M. Beaman 62

V Interaction with W. B. Murfin (Lead)

Concrete /MgO and RD. - A. Powers Basemat Melt-through R.

K. Cole VI Meltdown Phenomenology R. L.. Coats (Lead)-

J.

B. Rivard R.

J. Lipinski P. S.

Pickard F.

Gelbard J.

P. Odom M.

S. Senglaub P. B. Bleiweis (SAI)

J.. Tills (CSI)

O e

63

References 1.

Reactor Safety Study, WASH-1400, NUREG-75/014, U.S.

Nuclear Regulatory Commission, October 1975.

2.

Plan for Research to Improve the Safety of Light Water Nuclear Power Plants, NUREG-0438, U.S.

Nuclear Regula-tory Commission, March 1978.

3.

" Interim Report No. 3 on Three Mile Isla7d Nuclear Station Unit 2,"

letter from Chairman M.

W.

Carbon, Advisory Committee on Reactor Safeguards to Chairman J.

M. Hendrie, U.S.

Nuclear Regulatory Commission, May 16, 1979.

4.

Underground Siting of Nuclear Power Reactors:

An Option for California.

A Summary of the Technical and Economic Implications with Recommendations, Nuclear Assessments Office, California Energy Commission, April 1978.

5.

A.

S.

Benjamin, Program Plan for the Investigation of Vent-Filtered Containment Conceptual Designs for Light Water Reactors, NUREG/CR-10 29, S AND79-10 88, Sandia Laboratories, October 1979.

6.

TMI-2 Lessons Learned Task Force Final Report, NUREG-0585, U.S.

Nuclear Regulatory Commission, October 1979.

7.

M.

Rogovin and G. T.

Frampton, Three Mile Island:

A Report to the Commissioners and to the Public:

Volume I, Nuclear Regulatory Commission Special Inquiry Group, January 1980.

8.

R. O.

Wooton and H.

I. Avci, A Users Manual for MARCH l

(Draft), Battelle Columbus Laboratories, 1979.

9.

A.

K.

Postma, P.

C. Owzarski, and D.

L. Lessor, " Transport and Deposition of Airborne Fission Products in Containment Systems of Water Cooled Reactors Following Postulated Acci-dents," Appendix J of Appendix VII, Reactor Safety Study, WASH-1400, NUREG-75/014, U.S.

Nuclear Regulatory Commission, 1975.

I 10.

I. M. Wall, S.

S.

Yaniv, R.

M.

Blond, P.

E.

McGrath, H. W.

Church, and J.

R. Wayland, " Overview of the Reactor Safety Study Consequence Model," International Conference of Nuclear Systems Reliability Engineering and Risk Assess-ment, Gatlinburg, Tennessee, June 19-25, 1977.

11.

L.

L.

Smith, SIMMER-II, A Computer Program for LMPBR Dis-rupted Core Analysis, NUREG/CR-0453, LA-7515-M, Los Alamos Scientific Laboratory, 1978.

l 64 i

l

12.

S. L. Thompson,'CSO-II, A Eulerian Finite Difference Program for Two-Dimensional Material Res ponse - Part 1.

Material Sections, SAND 77-1339, Sandia La aoratories, 1979.

13.

M. Reimann and W.

B. Murfin, " Calculations for the Decom-position of Concrete Structures by a Molten Pool," PAHR Information Exchange Meeting, Ispra,. Italy, October 10-12, 1978.

14.

W.

B.

Murfin, A Preliminary Model for Core / Concrete Inter-actions, SAND 77-0370, Sandia Laboratories, August 1977.

I O

w 65

Distribution:

U.S. Nuclear Regulatory Commission (630 copies for Rl'and R7)

Division of Document Control Distribution Services Branch 7920 Norfolk Branch Bethesda, Maryland 20014 J.

F. Meyer (30)

Office of Nuclear Reactor Regulation U.S. Nuclear Regulatory Commission Washington, D.C.

20555 C.

N. Kelber (30)

Office of Nuclear Regulatory Research U.S.

Nuclear Regulatory Commission Washington, D.C.

20555 P.

Cybulskis (5)

Battelle Columbus Laboratories 505 King Avenue Columbus, Ohio 43201 LANSL C. A. Anderson, Q-Div., MS576 LANSL C.

R.

Bell, Q-7, MS553 LANSL J. Jackson, DAD /NRC, MS146 LANSL M.

G.

Stevenson, Q-DO/RS, MS555 (5) 1767 H. C. Walling 4400 A.

W.

Snyder 4410 D.

J. !!cCloskey 4412 J. W. Hickman 4412 G. J.

Boyd 4413 N.

R.

Ortiz 4413 W.

B.

Murfin (5) 4414 G.

B.

Varnado 4414 A.

S. Benjamin (5) 4420 J.

V. Walker 4421 R.

L.

Coats 4422 D.

A.

Powers 4422 J.

B.

Rivard 4423 P.

S.

Pickard 4425 R. J.

Lipinski 4440 G.

R.

Otey 4441 M. Berman 4441 M.

L.

Corradini 4441 M. Sherman 4442 W. A.

Von Riesemann I

4442 E-P Chen 4443 D.

A.

Dahlgren 4444 R.

K.

Byers 5522 D.

V.

Swenson 5524 M.

Huerta 3141 T.

L. Werner (5) 8266 E. A. Aas 3151 W. L. Garner (3)

For DOE / TIC (Unlimited Release) 66

t' Org.

Oldg.

Name Rec'd by' Org.

Bldg.

Name Rec"d by*

• Recipient must initial on classified documents.

o 0

>