ML20091R012

Forwards Info Requested by Jh Joyner in NRC Ltr Re Bases for Plant Protective Action Recommendation Logic Diagram

ML20091R012
Person / Time
Site:	Crane
Issue date:	07/06/1995
From:	Broughton T GENERAL PUBLIC UTILITIES CORP.
To:	Martin T NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION I)
Shared Package
ML20091R006	List: ML20091R012
References
C311-95-2284, NUDOCS 9509060182
Download: ML20091R012 (14)

Text

. _. _

f,:

GPU Nuclear Corporation i

L Nucle.;ar

= !r*

Middletown, Pennsylvania 17057-0480 (717) 944-7621 Writer's Direct Dial Number:

5 i

(717) 948-8005 -

.b July 6,1995 C311-95-2284 L

H

.]

i j

Mr. Thomas T. Martin J

Administrator, Region I i

U.S. Nuclear Regulatory Commission i

j.

475 Allendale Road 2 King of Prussia, PA 19406

[

Subject:

'Ihree Mile Island Nuclear Station, Unit 1 (TMI-1) l Operating License No; DPR-50 I

Docket No. 50-289 i

NRC Inspection No.' 50-289/95-05 j

' Request For Information About TMI PAR I.ogic Diagram' 4

(.

Dear Sir:

i i,

'?

This letter transniits the information requested by Mr. J. H. Joyner. in the NRC letter dated a

'e

' June 2,1995, about the bases for the Three Mile Island Protective A'ction Recommendation (PAR) I.ogic Diagram. As requested, Attachment I includes the assumptions and analyses

- utilized in the PAR logic diagram development, and the procedures by which operators f

estimate the release duration and the uncertainty associated with that estimate. Attachment I also provides sufficient. detail to demonstrate that this PAR logic will provide protective j

(actions equivalent to,the guidance in EPA-400, " Manual of Protective Action' Guides and

! Protective Actions for Nuclear Incidents."

a, i

'l Should you have any questions pertaining to Attachment I, please contact G. Simonetti at c

1 (717) 948-2081 for more information.

7

+

+

. Sincerely,

$. h

,9509060182 950706 "'

gDR ADOCK 05000289 T. G. Broughton PDR Vice hident and hr,'TMI

, ' " 5,kWM/mkk

,M l p; w ;:y:.'

~

SE

' %p :dj:ch'"iM.LGiEvans.TMI Senior Resident TOdr.

~

p l.

. E RJ.W 1Hernan - Senior Project Manager,

' ? nu ' syyOJ! H.' Joyner - Chief Facilities Radiological Safety'and Safeguards Branch 3'

3 pc vw

GPU Nuclear Corporation is$ subsidiary of General Public Utilities Corporation

[ i

, :n.

C311-95-2284 l

Page1 I

ATTACHMRNT I t

i PR'OTECTIVE ACTION RECOMMENDATION (PAR) LOGIC DIAGRAM BASIS l

Executive Summary 4

4 i

l The Environmental Protection Agency's " Manual of Protective Action Guides and Protective j

Actions for Nuclear Incidents" (EPA-400) requires evacuation at 1 Rem Total Effective Dose Equivalent (TEDE). However, it goes on to say, " Sheltering may be preferable 1

.to evacuation as a protective action in some situations." " Physical cbnstraints to evacuation-i e.g. inadequate roads" is an example it lists as one where sheltering would be preferable.

'Ihe area around the ' hree Mile Island Nuclear Generating Station (TMI) is highly populated with a road system that cannot always support an expeditious evacuation of the full 10 mile

- Plume Exposure Pathway Emergency Planning Zone (EPZ). ' Hence, for TMI there am times j

when sheltering is a preferable pmtective action.

If TMI's situation is such that there is an actual or potential release that could result in an unprotected population exceeding 1 Rem TEDE or 5 Rem Committed Dose Equivalent (CDE) to the Child Thymid, GPU Nuclear's preference is to evacuate the 5 mile radius and '

i 10 miles downwind. We will make such a recommendation unless sheltering is determined to be the preferable action. Exhibit 1 of TMI's " Emergency Operations Facility Procedure" (EPIP-TMI.27) provide the Emergency Support Directors with clear, definitive guidance on how to determine whether evacuation or sheltering is the prefermd protective action. The procedure will recommend evacuating the largest possible area that can complete the evacuation prior to one hour before the release is terminated. Such guidance provides a recommendation which will result in the least dose to the public and provide the maximum protection for those most at risk (closest to the site).

The following tables are provided to illustrate a comparison between ThH's Protective Action Recommendation (PAR) Logic Diagram and the Nuclear Regulatory. Commission's. diagram contained in Response Technical Manual (RTM)-92 Volume 1 Rev 2 and RTM-93 Volume 1 1

~

j

~ Rev 3. It also compares the dose received to the public using our recommendation against Pennsylvania's policy that only evacuations of the full EPZ will be considered. The general assumptions used to generate these tables are the same ai those used to initially generate the

- PAR I4gic Diagram. They are covered in detail under the " Assumptions" section of this response. The doses given mpresent the dose to the maximum exposed individual carrying out the corresponding protective action recommendation.

t b

y 6

i e

4

.________m_

C311-95-2284 Page 2 4

- Case 1: Long Estimated Release Termination Time (24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />)

Release Starts 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> after Generei Emergency declaration.

Release le terminated 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> efter declaration -

Recommendation Site Boundary Dose 2 Mile Does 5 Mie Dose J

GPUN Evecuate 5' mile radius end to miles downwind.

26.5 Rem TEDE

.530 Rom TEDE

.2EE Rom"cDE NRC Evecuate 2 mile 'redius and 10 rniles downwind. Shelter remeis 'no stee.

26.5 Rem TEDE

.530 Rom TEDE

.265 Rem TEDE l

g State Evocuate 10 mile radius.

26.5 Rem TEDE

.530 Rom TEDE

.265 Rein TEDE

- Case 11: Medium Estimated Release Termination Trne (5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />)

Release Starts 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> after General Emergency decieration.

Releas'e le terminated 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> after decle etion l

pocommendenen Site Bounde y Ume 2 Mise Dose 5 Mie Dose I

GPUN Kvecuate 2 mile radiue and shelter 10 rvules downwmd.

D Rom TEDE

.100 Rom TEDE

.050 Rom TEDE NRC Kvecuote 2 mile radius and 5 miles downwind. Shelter romaning eree.

10 Rom TEDE

.200 Rem TEDE

.050 Rem TEDE State pvecuate 10 mile radius 10 Rom TEDE

.200 Rom TEDE

.100 Rem TEDE l

[

~

~ Case 111: Short Estimated Release Termination Time (2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br />) f Release Starts 1 hours1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> after General Emergency declaration.

Release le terminated 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> after declaration f

I l

Recommendation Site Boundary Dose 2 Mile Dose 5 Mile Dose i

GPUN Shelter 2 mile radiue and shalter 10 miles downwind.

2.5 Rom TEDE

.050 Rom TEDE

.025 Rem TEDE b

NRC Evacuate 2 mile radius and 5 rrdes downwind. Shelter remaining eree.

5.0 Rem TEDE

.100 Rom TEDE

.025 Rom TEDE i

State Evocuate 10 mile radius.

5.0 Rom TEDE

.100 Rom TEDE

.050 Rem TEDE Case IV: Unknown Estimated Release Terminadon Tune I

Release Starte 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> after General Emergency decieration.

Re!eese is actually terminated 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> after declaration Recommendation Site Boundary Dose 2 Mie Dose 5 Mile Dose GPUN Evecuate 5 mile radius and shelter 10 miles downwind.

16.25 Rem TEDE

.325 Rom TEDE

.200 Rem TEDE l

NRC Evacuate 2 mile radius and 5 miles downwind. Shelter remaining erse.

16.25 Rom TEDE

.325 Rem TEDE

.200 Ram TEDE State.

Evacuate 10 mile radius 25.0 Rem TEDE

.500 Rom TEDE

.250 Rem TEDE Specific Assumptions for All.4 Cases:

Dose Retes in Rom /hr TEDE (efter release hee started)- Site Boundary = 5.00; 2 Mies =.100; 5 Miles =.050 i

Met.oorological coriditions "A" Stability, Southwest winde et 5 mph.

It is a winter weekday with normal weather.

e e

b

(

l-C311-95-2284 Page 3 i

PROTECTIVE ACTION RECOMMENDATION (PAR) LOGIC DIAGRAM BASIS

~

I Background-Prior to 1991, gu[ dance on how to develop a PAR was limited to pmviding evacuation times estimates for the full 10 mile Plume Exposure Pathwa'y Emergency Planning Zone (EPZ),

requiring that the user consider projected exposure to be received if sheltered vice evacuated, and a requirement to apply the " Keyhole" concept. This limited guidance caused considerable confusion and lack of consistency on the part of the users. One of the biggest problems dealt with determining the radius to evacuate if the General Emergency was based j

on potential failures. In such cases, the dose projections based on current conditions were l

often fairly low since core damage had not yet occurred, and the dose projections based on projected core damage were very speculative. As a result, in 1991'TMI Emergency Preparedness revised its PAR Imgic Diagram Procedure to provide a more user-friendly procedure that offered more definitive guidance. In developing this pmcedure, the department received considerable advice from the users themselves, namely the Emergency Directors, the Emergency Support Directors, and the Gmup Leaders Radiological and Environmental Controls.

Considerations-2 Pennsylvania Policy-Pennsylvania's policy regarding evacuation is that if any area of the EPZ must evacuate, then the entire EPZ must evacuate. GPU Nuclear, as well as the other 3 Pennsylvania utilities with nuclear plants, has voiced concern with this policy. The Pennsylvania utilities tried unsuccessfully on several occasions to convince the State to rethink its position. (We have not given up on the issue. We are presently discussing this issue with the new administration.) There can'certainly be times, particularly when the release duration is l

expected to be short and the release magnitude large, when such a policy imposes increased

. health risks to the public. On the other han,d if there is a longer, smaller release or

~

significant lead time before the release, this recommendation is not detrimental. In such situations, it would be beneficial if the plant's PAR was similar to ivhat the State would independently develop.

1 TMI Evacuation Time Estimates-The estimated evacuation times listed in both the 1981 and 1994 Evacuation Time Estimate 4

studies conducted for GPU Nuclear indicate there are dramatically.different evacuation times for a 2 mile evacuation versus a 10 mile evacuation. A 2 mile evacuation was estimated to i

take between 21/2 to 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> while a 10 mile evacuation could range from 6 to 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> depending upon the given condition.

0 The highly variable and lengthy evacuation times can make the decision process as to whether to evacuate or shelter more complicated than at a site with a smaller population density and thus shorter evacuation times. For example, at TMIjust extending the downwind sectors out to 5 miles from a two mile radius evacuation could lengthen the evacuation' time from 21/2 hours.to nearly 81/2 hours depending upon the conditions and sectors affected. O' the other hand, it might only lengthen it 10 minutes. '(Table 6.1, n

I h

(

C311-95-2284 Page 4 l

~

page 6-2 " Evacuation Travel Time Estimates for the Three Mile Island Nuclear Genenting Station Plume Exposure Pathway Emergency Planning Zone" February 1994)

~

Dose Projections,

A major concern prior to 1991 was what dose projection to use if the recommendation was being made based on deteriorating plant conditions. As was pointed out in NRC presentations, dose pmjections under such situations are highly suspect. It requires estimating variables such as time to containment failure, release duration, and extent of core damage. Since the damage has not yet occurrd it is not likely we could accurately estimate those variables. Therefore, limiting the recommended evacuation ra'dius based on a suspect dose projection does not seem very prudent.

I In all likelihood, our General Emergency declaration will be based on deteriorating plant conditions rather than dose projections. (This was the case in the 1995 TMI Biennial Exercise.) Once we recognize that we are moving towards the necessary core damage and i

potential containment failure that could result in exceeding the Environmental Protection Agency's Protective Action Guidelines (PAGs), we have approximately 15 minutes to j

determine an appropriate PAR. At this point, there is likely to be very limited data. We don't feel it is wise to delay our recommendation for a dose projection that would need to be l

based on anticipated conditions. Instead, we assume that all areas downwind of the release have the potential of exceeding the PAGs. Since we know protective actions are required, we can concentrate our efforts on the appropriate. recommendation, namely the option that would result in the least dose to the public. In other words, we are concentrating our efforts on trying to determine how long until the release might stop, something we feel we might be able to predict, rather than delaying our recommendation to try to estimate exactly when the release might start and what the extent of core damage might be. By concentrating on release termination time, we can better determine how much of the EPZ we should recommend to evacuate. In short, at a General Emergency, we will recommend evacuating i

as much of the EPZ as possible so long as it will not result in someone receiving more dose i

than he would get if he shelters.

l Even after damage has occurred, accurate dose projections can be difficult. Until samples L

are taken, assumptions need to be made on how much core damage has occurred and what is the exact isotopic makeup of the release. Again, in a situation where there would be

. sufficient time to complete the evacuation, limiting the evacuation radius based upon a dose projection that indicated areas outside 5 miles may not exceed the EPA's Protective Action j

Guidelines (PAGs) does riot seem prudent.

To summarize, our dose projection model will project doses as accurately a3 possible.

Certainly, the dose projections would be used to help us classify an event..' And certainly if the dose projections would indicate an area might exceed the PAGs and w'e hadn't already come to that conclusion based on plant conditions, then we would provide the appropriate recommendation based on the dose projection. However, we realize the limitations and the possible uncertainties.of this model. RTM-92 emphasizes the large uncertainties in dose

' projections on pages C-6, D-2, H-2, 3, 4, 6, and 7. Hence, once we have det. ermined that

. pmtective actions are required, either through the dose projection process or plant conditions, -

.then we will recommend evacuating all areas in which evacuation would result in less dose l

e

,.a e.

A

...4

.a-

..em*.e a-M 6 mal e_ _. - -

A a:

.-a Lm.-,

.ar.,am

.ms en u

P i

C311-95-2284 1

[

Page 5 1

i than sheltering thereby removing the reliance on potentially suspect. dose projections.

\\

l Guiding Principles-i In developing the diagram, we took into consideration State policy, the long evacuation times, and the. potential uncertainties in the dose projection process. We felt that the diagram.

should fulfill the following requirements:

The recommendation, as best we can determine, shou'ld result in the least dose to the j

public, and it should maximize the protection for those most al risk.

i It sinuld be consistent with the guidance provided by the Environmental Protection Agency. (Note: While this procedure was originally developed before the publication of EPA 400, it has been reviewed to verify consistency with that document.)

It should, as much as possible, be consistent with the generic guidance provided by the Nuclear Regulatory Commission. However, it should be GPU Nuclear's best -

recommendation based on all of the data available as opposed to strictly following l

generic guidance. (Note: Th'is procedure was originally developed prior to RTM-92.

The guidance used, which came from NRC presentations, was later published in RTM-92. When RTM-92 was_ published, a review of the diagram was conducted to

' verify that it still met the intent.of the principles of RTM-92.)

It should recognize that we can have long evacuation times. Therefore, it should take advantage of whatever lead time is available.

i It should be as consistent with the State as possible as long as it results in the least dose to the public and maximizes the protection for those most at risk.

C It should be simple and accurate, and it should be able to be determined quickly so that it can be provided in the 15 minutes available.

-[

It should provide consistent results from user to user.

Assumptions-Evacuation Time Estimate Studies (ETEs) -

It was assumed that the ETE conducted for GPU Nuclear gave reasonable estimations for the i

length of time to evacuate. A minimum of 2.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> would be required to complete the,

evacuation of.the 2 ' ile radius, and Wadng upon conditions, it could take as long as '

m approximately 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> to evacuate the full EPZ. _

s e

h m

,m

,~ -

a 4

C311-95-2284

^

s Page 6

)

3 I

Release Termination Time-d l

It was assumed that there can be situations when the release termination time can be' conservatively estimated.

7-i

.m j

It should be pointed out that we are actually dealing with release termination time and not release duration. Referring to EPIP-TMI.27, " Emergency Operations Facility," Revision 5, i

NRC Inspection Report No. 50-289/95-05 states under section 7.2 that "The discriminator between the recommendation of evacuation of an area vs3 ' heltering is the release duration."

s j

Technically, this is not true. The procedure actually uses the time f, rom General Emergency i

declaration until the time the release is estimated to stop."iThe difference is significant if the General Emergency declaration is based on deteriorating plant conditions rather than the 1

release at that time (i.e., plant conditions rather than presentidose projections).

j

mpf, Our method bases its release termination time on param$ters'w'e feel can be reasonably and

{

accurately determined. We did not use release duration 15ec5ise one must know both the l

starting time and the stopping time. Determining the'sta'rftime of a release would'be l

extremely difficult because to make that determination"ok'would have to accurately predict accident progression and containment performance, nei.ther!of which we Would have any j

control over. However, there could be situations when%do have the ability to control when the release will stop. ' For example, if the release (is. through a leaking steam generator, we know that we can emergency isolate that generator; ati300' degrees. ? A conservative estimate using cooldown rates can predict the terminatio $Itidae."..

Oigy 9Wf The inspection report asked that we identify procedures by%shich.

j release duration and the uncertainty associated with'that'duritidri.E As. stated earlier,'we do

'not use release duration in the PAR process; we use rel55't5rmhstion time. There are no

)

formal procedures used to determine that time. Release".Erinination time is based 'on

, experienced operator estimates of when the release can ly(iminat'ed t

' plant manipulation. TMI-l Abnormal Transient Procedure 1210-5, "OTSG Tube Ieakage")

i.

I provides formal guidance used to determine when a gekf56Ecari be isolated or when we can go on decay heat removal. From there, the operaidrsis5 reactoi shutdown rates and.

l cooldown rates to estimate that release termination tiin'ei$,'IM unchrtainty associated with this l

estimate is generally plus or minus one half hour. ~ This sanie estimate is used for the dose '

projections as well. The projections assume the release has"aheady started resulting in its

}

release duration equalling the PAR Diagram's release teimination time. ' Hence, if we are unable to arrive at an appropriate release termination tim' dhen~we 'would also have e

inappropriate dose projections.

TMl$@ ?. '

4Q9y$

We recognize the critical role esthnating the release,terminatid@n' time l

s

- importance in training. We consider a non-conservative [feldase termination time to be a l

serious error similar to a non-conservative Emergency'Actibri Ievel determination., We expect our operators to be properly trained and exercisiFiljood judgement. Certainly, any f changes to conditions that would cause a change in'the%WnMrelease termination time i

would be substantial enough to require going to the b$EiT of thfdiagram (see step #4 of 1

l GTables 1-4 in Note 7.).

dR.

l

.f

$$[

I 4

1

'A

~

Q Ch4 -

J

]g l

2 nu:..n

3 C311-95-2284 Page 7 l

~

]

We stress to our F.znergency Directors and Emergency Support Directors that they must be confident in the release termination time. If they are not, then they are to assume a release termination time of 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br />. Eight hours was chosen as the typical time it would take to cool 4

down to the point.,where a release could be terminated. An 8 hour9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> release termination time.

would msult in a minimum of a 5 mile evacuation except in l adverse weather situation t

where it would yield a 2 mile evacuation. In 75% of the possible situations, it would yield a 10 mile evacuation downwind.

j Sheltering Protection Factors-

)

The procedure only credits sheltering with saving I hour worth of d'ose. 'Ihe figure was j

originally arrived at through sheltering guidance that stated the types of homes in the TMI i

l area would provide a 50% dose reduction for up to 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br />. When EPA 400 was issued this sheltering assumption was examined and determined to still be appropriate. The typical a

house in the TMI area is a wood frame house with a basement. For external dose, EPA 400 l

suggests a Dose Reduction Factor of.6 for this type of house. Our minimum evacuation s

time is 2.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />. Two and a half hours of dose times the.6 factor equal one and a half j

hours of dose. 'Ihis implies that over a 2.5 hour5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> period the house will reduce the dose by 1 j

. hour worth over what would have'been received outside the house. Approximately the same j

figure would be arrived at for the inhalation dose if there is no cmdit for special preparations j

being taken. Theoretically, the protection offered by sheltering will be some number in

• excess of I hour dose savings, for exposures in excess of 2.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />. However, as time goes.

on the protection sheltering offers becomes less and less significant. Our assumption of I hour wo'rth of dose savings over a 2.5 hour5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> period also appears consistent with the assumption RTM-92 Volume 2 uses on page 110 that a frame house provides a 1/2 reduction i

in total dose.

1 Plume Travel Time-l

. It was assumed that plume travel time would not have a significant effect in determining whether to evacuate or shelter. Plume travel time is highly variable and unpredictable.

i Furthermore, examination of wind speed characteristics around TMI have shown that winds are typically 5 miles per hour or greater. When winds do become very light (less than 3 mph) they rarely stay that light for more than an hour.,

l tt r

i Ground Shine-In determining whether to evacuate or shelter, there'was no additional dose added to.the.

sheltering case to cover the dose received if evacuated after the release had terminated. For the situations where sheltering would be considered (i.e., release termination time period of a l

few hours), the dose from ground shine will typically represent only a small fraction of the.

l total dose received. It is true that there would still be dose delivered from ground shine after the release had stopped..However, if that dose were significant,'those that were sheltered could then be evacuated. 'Ihe additional ground shine dose received'in that evacuation

, process would be more than compensated for by the additional dose savings, above the 1

' hour credit already taken that sheltering would actually offer.

a 4

k e

' 1 %=

4 o

e T

l

\\

f n l

+

8

hb 4:5 J, a

, *'Qir l

. =

C311-95-2284

~;.

Page 8

~

i Dose Comparisons-Dose comparisons between the evacuation alternative and the sheltering alternative were conducted for the later evacuees who would be the most at risk. With the data we hhd available, we could not say how far from the plant the earlier evacuees would get before j

getting tied up in traffic.

4 For those closest to the plant, it was assumed that the evacuees would first get caught in traffic close to home. 'This assumption seems reasonable based on the ETE's conclusion that one of the places where queuing is expected to be most prevalent is along Route 441 (River Road which is the closest main road to the plant) (page 6-1 ETE). Donceptually, those that have the furthest to travel (those closest to the plant) would be the most likely to get caught in traffic, would get caught in that traffic closest to the site, and would be caught in traffic 2

j for the longest time. An analogous situation would be the exiting of a crowd from a football stadium. Those furthest from the exits are the most likely to be delayed, they would be delayed furthest from the exit, and they would be delayed for the longest time.

i i

Qualitative Analysis-4 Long Release Termination Times-Long release termination times can result from either a release with a.long release duration or an extended period of time before the release starts followed by a shorter release duration.

l In either case, at a General Emergency as a result of actual or projected doses exceeding the PAGs or because actual or expected plant conditions could result in a release that would exceed PAGs, the PAR Logic Diagram recommends evacuating a 5 mile radius and 10 miles downwind. Our definition of downwind sectors are those that we expect to be affected and 4

the adjacent sectors (outside the 2 mile ring) during the course of the release. With very long release termination times, we are likely to conclude that all sectors are expected to be

{

affected during the course of the release. Thus, our PAR will then match up with the State's. Such a match enables preplanned protective actions to be used. As RTM-92, Volume 2 page 196 points out when using preplanned protective actions, "The areas to be evacuated, the message to be used, the routes to be followed, are in place for fast t

implementation."

~

l For this case, dose projections for some of the recommended evacuation area may not necessarily exceed EPA PAGs. Our reason for recommending that the area be evacuated i

anyway is that we have potentially long evacuation times. We know we have damaged fuel; 4

so there is the potential for significant exposures. We need to take advantage of whatever time is available to us in case the dose projections turn out to underestimate actual dose.

Very Short Release Termination Times-It is extremely difficult to foresee a situation where plant control has been lost to such an t

extent that core damage has occurred, and yet the plant has been cooled down to the point where any releases can be stopped in a very short period of time. However, good emergency preparedness calls for a plan to be developed for such a circumstance.

[

g a

r h

I

_ _ ~

C311-95-2284

[

Page 9 For evacuation of only the 2 mile mdius, the primary delay in evacuation is the preparation and mobilization time. Most people (85 %) would not be able to begin evacuating for the first hour and nearly half of the people (45 %) will not begin until an hour and a half has passed (page 5-6,," Evacuation Travel Time Estimates for the Three Mile Island Nuclear Generating Station Plume Exposure Pathway Emergency Planning Zone" February,1994).

The inspection report justifies evacuating the public in short release durations simations based on the conclusion that " dose savings to other residents will be significant in that some portion of the population will have evacuated the affected area prior to plume passage." Based on i

our ETE for a 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> release termination time, our conclusion is that only 15% of the public

}

might receive less dose while 85% would receive more. To us, suc'h a situation would j

justify sheltering as an option.

{

An obvious concern is choosing a sheltering option when evacuation could be implemented.

That is why the release termination time must be based on something the operators feel they 1

have control over. There should be very little error when a very short termination time is I

chosen. Furthermore, in the short termination time situation, there should be fairly i

immediate feedback if the termination time turns out to have been estimated in the non-conservative direction.

For example, suppose that we lose all makeup capability during a primary to secondary leak in excess of 50 gallons per minutes while steaming through the condenser. There:is no core i

damage at the time of the declaration, but incore thermocouples have indicated a slight i

superheat condition. Based on our EALs, we would declare a General Emergency. Now lets also suppose that the operators estimate,' based upon their cooldo'wn progress, that they I

will be able to isolate the leaking steam generator within one hour. Our PAR Iagic Diagram I

procedure would suggest that they recommend sheltering th'e 2 mile radius and 10 miles downwind. If after this recommendation something then happened to change their ability to i

meet the one hour termination time, the same PAR Imgic Diagram procedure would require i

them to return to the beginning of the diagram because there has been a substantial change in l

conditions (see step #4 in Tables 1-4 of Note 7 ). The time between the initial and subsequent recommendations was not completely lost. Since the sheltering recommendation was given, some of the preparation and mobilization would have been accomplished. -

Therefore, the subsequent evacuation shouldn't take quite as long as it would have if there were no sheltering prior to the evacuation. Admittedly, this would not be the way we would prefer the sequence of events to occur, but because of the rather immediate feedback that would be available on the estimate for short termination times, the possible consequences are I

limited.

4 Now, take the case where we did properly estimate the release termination time of I hour, but our procedures required an automatic evacuation at the General Emergency. Lets also

~

j

'~us that'85 % of the people wil1~not even have left their house in that 1 bour period. Instead, suppose that fuel damage and with a significant release occurs for that hour. Our ETE tells they are going in and out of their house, completely unprotected, preparing to evacuate because that was the recommendation they were given. We see the consequences of going with an automatic evacuation when it cannot b' practically implemented as more of a risk e

' than inaccurately estir'nating a conservative release termination time that was based on something the operators had control over.

- =. -

(

~

i.

.C311-95-2284 Page 10 Release Termination Times of Several Hours-It is in this termination time frame that our diagram provides its most benefit. It provides the operator with meaningful information about what can practically be carried out. The i

population inside TMI's EPZ is approximately 175,000 people with at times an additional 30,000 transient population. This population is not evenly distributed around TMI. A large 2

portion can be found in the northeast quadrant of the 2-10 mile ring. As was pointed out earlier, including this area in the evacuation recommendation significantly lengthens evacuation times. Just including the 2-5 mile area of this quadrant can sometimes significantly alter the evacuation time. Adding this area based on a default recommendation l

of 2 mile radius and 5 miles downwind or because suspect dose projections indicate a possibility of exceeding the EPA PAGs can have serious consequences on those most at risk, i

namely those residing in the 2 mile ring. Therefore, before we recommend evacuating out to 5 or 10 miles, we would like to believe that such an action will result in those people receiving less dose then if they had sheltered.

4 If we believe that an area cannot be completely evacuated before the time we have conservatively estimated that the release will stop, then we will recommend sheltering areas I

further out in order to expedite the evacuation of those areas nearest the site. With longer release termination times, larger evacuation areas can be recommended.

f This system provides a well defined method to maximize the protection for those most at

[

risk. Since the evacuation times are relatively short for the 2 mile radius, it is highly likely that our PAR will recommend that~they evacuate. If more time is available, we take.

advantage of that time by evacuating a larger area. And, if sufficient time is available, our i

recommendation would closely match the State's policy. The significant advantage of this system is that it provides the Emergency Directors and Emergency Support Directors with clear guidance on what to do in the limited release termination time situation.

An example can help illustrate the advantage of our system. Assume that there is a primary to secondary leak with insufficient makeup capability. Secondary side cooling capabilities j

' somewhat degraded. However, the emergency response organization has determined that

.y can get the plant on the Low Pressure Injection (LPI) System in 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />. The wind is blowing into Sector 1 (the north-northeast sector). It is a weekday with normal weather.

Assume that core damage occurs 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> after declaration and lasts for one hour until the plant is placed on LPI. Assume that the dose near the site is 100 Rem TEDh for that 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />. Depending upon release pathway and meteorological conditions, it is certainly possible that the doses could be a factor of 100 less at 2 miles. Our recommendation would be to evacuate the 2 mile radius and shelter 10. miles downwind. This would result in no dose to those inside 2 ~ miles because they would have been able to complete their evacuation prior; to j

plume arrival. The dose received by those outside 2 miles would be about one half of the dose t. hey could have received if they had evacuated. In this example, it would be something..

less. than 500 mrem. On the other hand, suppose we had tried to evacuate out to 5 miles,

^ ither becan:e it was an' automatic requirement or because dose projections indicated that i

e those outside 2 miles could exceed-1 Rem. Those near the site could then receive up to 100 '

Rem since now they wouki be caught in traffic on Route 441.while those outside 2 miles could receive up to 1 Rem'. We see our procedure having the benefit of providing substantial

. dose savings and the avoidance of early health effects. If there were 10 Rem at 2 miles, the _

[

+

[

C311-95-2284 c

l Page 11

~..

potential doses near the site could be 1000 Rem making our methodology even more 4

l appropriate.

Unknown Release. Termination Time-4 In a relatively fast-breaking event, the operato'rs may not know exactly how long it will be before they can stop the release. They may be in.the process of ~ working on that information j

when the General Emergency is declared, but they have insufficient data to make an accurate estimate. In such a situation, the procedure has them use a default value of 8 hoiars. Eight j

hours was chosen because it represents the typical times it has taken'to get the plant fully i

cooled down. It also provides a reasonable initial protective action recommendation. An 8 hour9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> release tennination time will msult'in a minimum of a 5 mile evacuation except in 2

one adverse weather situation where it yields.a 2 mile evacuation. In 75% of the possible situations, it will result in a 5 mile radius and 10 miles downwind evacuation.

4 1

MMA

• The inspection report felt that this recommendation, appears in conflict with EPA-400 in that 1

j it would be possible for areas that werci sheltered to receive in excess of 10 Rem and not.

even be considered for evacuation. 'Ihis is uhirue.'sBy definition, there'must be a substantial l

change in conditions after 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br />. Either/tlie; release has been terminated or the evacuation has been completed. In either case, it requires.the operator to go back to the beginning of the diagram. Since our basic premise that.is taught to all of.the' Emergency Directors and l

Emergency Support Directors is to evacuate all of the potentially affected areas that can i

complete their evacuation prior to the releascitermination, the release termination time should

)

be adjusted to the known time or by an additi8nal.8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br /> if unknown. In either case, a new l

incommendation would be made. Again,'tlie' additional area would be made based upon the l

new total termination time. Hence, if sufficichitime is available, the previously sheltered area would then be evacuated.

l It should be pointed out that in developing gresponse to the inspection report it was found Tables 1-4 of Note 7, it says, "Determinet"XQ;would improve the d that there is a place where some word changes hich equals the number of hours from i

NOW until the current or projected release'isl estimated to STOP. If it is unknown when the -

release will stop, assume "X" =8." The word,?.NOW", was chosen to ensure that the opemtors took advantage of any lead time. prior to the commencement of the release. It is obvious to us that for the tables to work properly lwhen they are revisited the user must count l

the time that has already passed since the first recommendation. For clarity, the wording should be changed to ensure that the total release; termination time is used. The procedure l

will be revised to change the word, "NOW"[td "ths time of the General Emergency t

{

declaration.

4: Mig,

' mA.,

\\s?f%g ' Y h

Deviations from Procedure-f e

j The procedure tries to anticipate all of the pssible situations. However, we recognize that it l

i is possible that, given a totally unique and unasticipated scenario, the procedure may give an.

l answer the Emergency Director or Emergendj[S6pport Director feels is inappropriate. To l

. cover such cases, there is a note'in the prti5dukth'at gives'them the latitude to exercise good judgement and deviate from the proc &lEN%niwer if mmry.

f x

u,~

ff.,

m k

=-

i.

C311-95-2284 Page 12

. EPA-400-The inspection report questions whether this diagram conforms to the guidance set forth in EPA-400, "Maewl of Protective Action Guides and Protective Actions for Nuclear Incidents." While the original diagram was developed under the guidance issued by the EPA prior to EPA-400, it was carefully reviewed against EPA-400 to ensure it was consistent with i

the new guidance. Some modifications were made at that time. However, our basic philosophy was already consistent and did not need to be changed.

EPA-400 says, " Sheltering may be the prefermd protective action when it will provide j

protection' equal to or greater than evacuation, based on consideration of factors such as i

source term characteristics, and temporal or other site-specific conditions." A significant

. portion of our PAR Logic Diagram procedure is devoted to conducting the evaluation of which, sheltering or evacuation, offers the greatest protection. Section 5.5 of the manual goes into a more detailed discussion of the merits of sheltering versus evacuation. Page 5-18 l

of the manual provides guidance on how to make a determination on whether sheltering or.

evacuation is the preferred protective action. In developing our diagram, we took into i

consideration the items listed such as, " population distribution, the sheltering effectiveness of residences and other structures, institutions containing population groups that require special consideration, evacuation routes, logical boundaries for evacuation zones, transportation systems, communication' systems, and special problem areas." Long evacuation times result in the same prol,lems as long mobilization times as described in 5.5.2 #4 on page 5-19.

i I

EPA 400 recognizes this when they say, "it (sheltering) may be the temporary protective l

action of choice if rapid evacuation is impeded by physical constraints to evacuation (e.g.

inadequate roads)" (5.5.3 #7, bottom of page 5-21 and top of page 5,22..)'

\\

l We also noted the comment in section 2.3.1 of the manual where it says, "However, reliance on large dose reduction factors for sheltering should be accompanied by cautious examination of possible failure mechanisms,' and, except in very unusual circumstances, should never be relied upon at projected doses greater than 10 Rem." Unlike the inspection report, we do not read that statement to say that. sheltering cannot be used for situations where there are dose projections greater than 10 Rem. Instead, one should not assume large dose reduction i

3

' factors to justify sheltering when dose projections are greater than 10 Rem. Our methodology only assumes very conservative dose reduction factors.

(

We see the situations where we could potentially recommend ' sheltering an area where dose projections. exceed 10 Rem as very rare. The first would be an extremely high dose rate puff release. For very shon time periods, an evacuation recommendation could result in the majority of people receiving significantly more dose than if they,had sheltered. We-interpreted such a situation as "an unusual circumstance".

, 'The second situation would occur if we limited the evacuation area due to the short release

' termination time. Foi example, assume that our release termination time caused us,to only recommend evacuating the 5 mile radius and shelter 10 miles downwind even though our dose projections indicate areas outside 5 miles may exceed 10 Rem. In such a situation, the

. area inside the 5 mile radius has the potential for doses much greater than 10 Rem.

Furthermore, attempting to evacuate the full 10 miles could result in those living in the

.o Ih

' auw &

.~

[,

C311-95-2284

(

Page 13 j

5-10 mile ring getting more dose than they would get if they sheltered since they may not be l

able to complete their evacuation prior to release termination time. We do not believe it is J

the intent of EPA-400 to require us to evacuate an area that we are estimating might receive 10 Rem based on a dose projection with potentially large uncertainties if such a recommendation would result in others getting lethal doses, especially if such a I-recommendation could result in all areas receiving more dose. Again; we see such a i

situation as "an unusual circumstance."

4 Conclusion-3 g

)

The inspection report also incorrectly assesses the diagram as being insensitive to specific j

dose pro'ection results. Clearly, any dose projection in excess of EPA guidance will. result in a protective action recommendation if one has not already been taken based upon plant conditions. The difference is that once we know we have the potential to exceed the PAGs l

we do not limit our recommendation based upon a potentially suspect dose projection.

Iristead, we evacuate unless sheltering results in being the preferred protective action because it provides " protection equal to or greater than evacuation, based on consideration of f.tetors 4

such as source term characteristics, and temporal or other site-specific conditicas." We will j

always evacuate an area of projected dose of 10 Rem unless there are unusual circumstances, J

j-The difference is that we do not necessarily wait for that dose projection. Ifit made sense to 1

evacuate, we would have taken that action. We realize that our procedure is sensitive to l

evacuation times and estimated release termination times. However, because of the uniqueness of our evacuation times, we felt our recommendation should be sensitive' to them.

j And all recommendations based on dose projections are already sensitive to being able to estimate that release termination time. We feel there are circumstances when we can make l

an accurate determination of that time, and if not, we have e reasonable default value. By j

concentrating our efforts on release termination time, we are focusing on a parameter that we are more likely to have some control over and some experience in determining.

.We cannot say the same about many of the other parameters that go into the dose projection

]

. process suc.h as estimating, core damage or isotopic mix in a severe core damage situation.

We believe that our procedure not only meets EPA-400 guidaace, but more importantly, provides a recommendation that gives the least dose to the public and maximize the 3

J i

protection for those most at risk around TMI.

i i

l i

)

i s

e b

4 1-j

~

}

t i'

,'Q g s, 4

!$bML idu.

, s

,