ML20212M334

Responds to 851119 & s Re Adequacy of Emergency Alerting Siren Sys at Nuclear Power Reactors.Commission Continuing to Assess Issues Raised by Ltrs & Initial Decision

ML20212M334
Person / Time
Issue date:	08/07/1986
From:	Zech L NRC COMMISSION (OCM)
To:	Bright G, Carpenter J, Kelley J Atomic Safety and Licensing Board Panel
References
NUDOCS 8608250448
Download: ML20212M334 (1)

Text

,

,, UNITED STATES .

g NUCLEAR REGULATORY COMMISSION

[; WASHINGTON D.C.20555 y

~s.,...../

CHAIRMAN August 7, 1986 Judge James L. Kelley Judge James 11. Carpe.nter Judge Glenn O. Bright

Dear J1dges Kelly,

Carpenter, and Bright:

Thank you for,your letters of November 19, 1985 and May 16, 1986 concerning the adequacy of emergency alerting siren systems at nuclear power reactors. The' Commission is continuing'to assess the issues raised by your letters and your initial decision, and has been closely following the resolution of this issue in the licensing proceeding. The Commission will be asking staff for its views on the generic issue which you raise, with a view to deciding whether the relevant rules or guidance documents need to be revised Thank you'for bringing this matter to our specific attention..

Sincerely, M b. ,

Lando W. Zec , Jr i

l 8608250448 8!30007 PDR COMM3 NHCC

, CORFiESPONDENCE PCR ,

4

,. _ -n. - - --., , , , , . _ _ , - , - - - - , _ - - , _ .

,qg ,

..s- -

" '4 UNITED STATES f j NUCLEAR REGULATORY COMMISSION

- ATOMIC SAFETY AND LICENSING BOARD PANEL 0 k, WASHINGTON. D.C. 20555

. +...*

November 19, 1985

[

l Chairman Nunzio J. Palladino 4

Commissioner Thomas M. Roberts

Consiissioner James K. Asselstine 2

Conmiissioner Frederick M. Bernthal i Comunissioner Lando W. Zeck

Dear Mr. Chairman and Conunissioners:

We are writing this letter to bring to your attention evidence of possible generic safety problems involving nighttime emergency I

notification of residents in the plume emergency planning zones surrounding nuclear power plants. It appears that under the acceptance i

, criteria being generally applied by the Federal Emergency Management Agency to nuclear plant siren systems.-- criteria which are. based on 4 daytime conditions -- substantial numbers of.EPZ residents would not be aroused from sleep should an emergency siren notification be necessary between, say, midnight and 6:00 a.m., particularly if bedroom windows

were closed.

We are members of the Atomic Safety and Licensing Board presiding over the application of Carolina Power and Light Co., et al. for an operating

. license for the Shearon Harris facility. We recently conducted an 1

evidentiary hearing in that case on an Intervenor's contention that the

Applicants' sirens would not awaken sleeping residents in the EPZ between midnight and 6

00 a.m., particularly those who have closed their

windows and turned on their air conditioners. Both the Applicants and the Federal Emergency Managespent Agency presented extensive testimony '

, and responded to cross-examination on the contention, as reflected in i the Shearon Harris transcript.. Tr. 9356-9976. The parties will be i submitting proposed findings of fact on the siren and another contention during December, and the.8oard expects to decide those contentions in l early 1986.. In these circumstances, we as a Licensing Board have not drawn any conclusions about the particular siren contention before us in the Shearon Harris case. In any event, we would have no occasion to write to you about Shearon Harris at this time -- in advance of the normal review process -- because Harris is not, of course, an operating facility. Furthermore, should any deficiencies in nighttime notification emerge from the Shearon Harris record, we could fashion effective measures to deal with them on a site-specific basis. Rather, this letter is prompted by the possible generic implications of certain of the evidence in the Shearon Harris record, and our belief that such evidence should be called to your attention now. Most significantly, you should be aware of the following matters:

11/21...To EDO to Prepare Response for Signature of Chairman and Consn Review...Date due: Dec 3...Cpys to: RF..85-991

.. a ,

wh 9'/W

,, o' t Q, -

. t 5.-  % ~ _ _ _ . . _ . . . . _

Zip' ;'

./ - . .

.s

1. The FEMA design criteria provide for siren sound covera '

10 dB(a) above ambient noise levels or a minimum of 60 d8(c)ge of

2. The " FEMA 43" revi.ews of siren system designs for operating nuclear facilities are based on summer daytime conditions. Such reviews give no consideration to factors only applicable at night -- i.e., almost everyone is indoors and asleep, many with the windows closed. These factors are not offset by greater sound propagation at night.

3. The testimon air conditioning)y and isanthat, outdoorforsiren a house sound with the level of 60windows dB, the closed (central probability of arousa.1 from sleep is essentially zero. With the windows.

open, the probability is 7-8%. -(Tr. 9650).

4. The sound levels necessary for high probability of arousal are substantially above the FEMA minimum guidance. For example, a 50%

probability of arousal of an individual in a house with the windows closed requires outdoor sound levels of 90-99 decibels (Tr. 9927). The wide range in this estimate reflects the imperfect knowledge of siren arousal capability, since direct tests have not, been conducted, at least in the United States.

5. For the Harris assumed summer scenario -- including 36% of the.

houses in the EPZ with no air: conditioning'(windows open) -- aiousal is

~

estimated as approximately 70%. Assuming the accuracy and acceptability

of that estimate, lower probable arousal rates in other climatic areas of the United States nevertheless suggest possible generic safety concerns -- particularly in the winter season, when more of the population may sleep with the windows closed.

6. We further note that overall arousal percentages for an EPZ can l

mask the fact that, in areas'of moderate housing density but with a fair number of houses more than one mile from a siren, a substantial percentage of the houses would have an arousal probability of less than 50%. For example, at siren 70 in the Harris EPZ, about 100 houses of about 160 in that area; or 63%, are in the 70-80 dB area. The .

probability of arousing an individual in those houses with the windows closed is approximately 30%, and is 55% for household arousal with two residents. If these houses correspond to U.S. averages,18% would have one resident, 30% would have two residents, and the average. probability of arousing' the household would be 43%. ~ .Thus, although the risk.of not being aroused, averaged over the EPZ,.may be' roughly 30%, in some areas approximately 50% of the houses may- have a risk of non-arousal of roughly 57%. '

NRC regulations require "early notification" to EPZ residents (10 C.F.R.

l 50.47(b)(5)), and this has been further defined to mean that the desigr.

l objective of "the prompt public notification system shall be to have the capability to essentially complete the initial notification of the l

k e g

. - . . - =.. -- . . - = = . - - a.

l'... .

l 3-

}

public within the plume exposure pathway EPZ within about 15 minutes."

10 C.F.R, Part 50, Appendix E, IV D.3. The evidence we have cited indicates that in EPZs which rely primarily on siren notification (it is our impression that that is the case for most commercial reactors), such notification would not be " essentially complete" under some typical nighttime conditions within 15 minutes. To be sure, emergency notification typically includes "back-up" notification by police, fire and other emergency vehicles driving prearranged routes with sirens sounding. However- as was testified to in the Shearon Harris case, this back-up notification probably would not be accomplished in the 15-minute period. Tesgimony of Alvin H. Joyner at 42, ff. Tr. 9374. It seems doubtful tha; such back-up notification could be completed expeditiously if emergency workers had to first be reached and activated between midnight'and 6:00 a.m.

We also are aware that persons who are aroused from sleep might not only arouse their household, but that they would also tend to contact some

~

neighbors, friends and relatives in the EPZ. This process of " informal notification" was testified to at some length in the Shearon Harris hearing, and the Board does not doubt that such informal notification would occur to-some degree. However, the phenomeno'n cannot be controlled like a siren level or a police car route, and its likely effects are difficult to quantify. Therefore, we question whether l infonnal notification should be viewed as a. substitute for planned notification.

According to testimony in the Shearon Harris case, the siren systems around operating nuclear plants have not been field-tested to determine what percentage of EPZ residents would actually be aroused by their activation between midnight and six a.m. Tr. 9935-40. We believe that some such testing would be prudent and merits Commission consideration.

Respectfully submitted, Oe - = h l JysL.Kelley /

L h' DpJamesH. Carp $hter Glenn O. Bright

$, AY cc: H. Plaine, General Counsel l S. Chilk, Secretary Shearon Harris Service List l

, - . - -.-ww,o g - r-- ~---n- ---,,,-p ---v - ,n ---

3, .. .,

n, 8

• UNITED STATES N8N 4

[' j e

' NUCLEAR REGULATORY COMMISSION ATOMIC SAFETY AND LICENSING BOARD PANEL ,, ; 4 W ASHIN GTON. D.C. 20555 ,

sN

,/

May 16, 1986 N '

. Ny 1 s Chairman Nunzio J. Palladino I.

Commissioner Thomas M. Roberts \ j Comissioner James K. Asselstine / .

\' /

/

Comissioner Frederick M. Bernthal N

'l-Comissioner Lando W. Zech, Jr.

occ m FAC.. %..u._q 9 0 - g(_

Dear Mr. Chairman and Comissioners:

PR00.a ci. ._.

'On N vember 19. 1985, we wrote to.you to express our concerns,~

arising 'out of the Shearon Harris operating lic5nse proceeding, that existing siren systems at other nuclear plants might be deficient for nighttime alerting. On April 28, 1986, we issued our decision on the remaining contentions in the Shearon Harris case, including nighttime alerting, and authorized issuance of fuel loading and operating licenses for that facility. We found that the overall Shearon Harris alerting system, where sirens are to be supplemented by a tone alert radio in the bedroom of each residence in the first 5 miles of the 10-mile EPZ, meets-applicable requirements under sumer nighttime conditions.

Nevertheless, our findings in the Shearon Harris case underline our generic concerns about nighttime alerting at reactors which rely, entirely on sirens and so-called " informal alerting" (initiated by the sirens) in the first 15 minutes following a declared emergency.

We continue to be concerned about the technical adequacy of the FEMA standard for judging nuclear power plant siren systems for nighttime conditions. As discussed in more detail in our decision, the FEMA standard for a nuclear plant siren system in a sparsely-settled area like Shearon Harris is that it produce a minimum sound level of 60 dB (decibels) outdoors throughout the EPZ. This standard is applied to sumer daytime conditions, the time least favorable for sound propa-gation. However, as the record in the Shearon Harris case shows, the time least favorable for actually alerting people is midnight to 6 a.m.

when virtually everybody is indoors and asleep. Although nighttime conditions are more conducive to sound propagation, that advantage is much more than offset by closed bedroom windows and the state of sleep.

. 'Our second concern is with.the Staff's position, as expressed in the Executive Director for Operations' memorandum to you of February.24, 1986, which states in pertinent.part as follows:

The conclusion reached in FEMA's study, in response to the

ShearonHarris]hearingcontention,thatthissirensystem

. supplemented by informal alerting] can be expected to arouse and alert approximately 90% of the EPZ residents during a m m ,

h,w 'enAms'T4)"( V

~

, .., , -- , - _ . , - - - - - - , _ - , _ _ - - - . - - - . . - . , 9 i~

. a ,

O nighttime emergency serves to confinn FEMA's judgment that siren systems designed and evaluated in accordance with

- NUREG-0654/ FEMA-REP-1, Rev. I and FEMA-43 meet the NRC requirements for both daytime and nighttime alerting.

Unlike the Staff, we would not accept a 90% alerting level as meeting the Commission's requirement of " essentially 100%" alerting in the first 5 miles of an EPZ. We held that " essentially 100%" means a notification system capable of alerting greater than 95% of the EPZ residents in the first 5 miles. We therefore held that the 91% level expected for the Shearon Hasris sirens and informal alerting was insufficient. However, with the addition of: tone alert radios at Shearon Harris, an alerting level of 98% 'can be. expected, a clearly satisfactory level in our view.

Our third generic concern arises from a realistic scenario which the Shearon Harris case and the FEMA / Staff position do not address --

winter nighttime alerting. The Harris contention was focused on summer nighttime conditions apparently because the Intervenor believed that noise from air conditioners made that the " worst case." The evidence showed that air conditioner noise, typically in the lower frequency range,'had only slight effects on ale'rting. The key factor is whether the bedroom windows are open er closed because closing the window attenuates sound intensity by about 28 dB. In the Shearon Harris case, the study on which FEMA relied was based on the assumption that about 50% of the windows are open on a summer night in Eastern North Carolina.

In the winter, however, we would estimate that perhaps 75-90% of bedroom windows in that area would be closed. In an EPZ in a colder climate --

say, in Maine or Minnesota -- we would expect that most people would also have storm windows and that virtually all bedroom windows would be

closed at nicht. That expectation is borne out by a 1982 NRC Staff analysis titled " Evaluation of the Prompt Alerting Systems at Four Nuclear Power Stations," NUREG/CR-2655, PNI.-4226, which projected the following siren alerting rates under nighttime conditions

Plant Conditions Rural Area Trojan WinterNight(Raia) 60%

TMI 42%

Zion Winter Night(Windy Winter Night (Snow)) 51%

Indian Point Winter Night (Snow) 53%

~

Similarly-low alerting percentages could'be expected as a result of closed windows in hot summer climates.

To be sure, the large gaps between these alerting percentages and l the Commission's requirement of " essentially 100%" will be partly filled by informal alerting. If one assumes that 50% of the alerted households would engage in informal alerting, overall alerting for the above four l

E

~ ~

/.- ,

1 3-

, EPZ's would range from 54% to 72%. Using TMI to illustrate that computation, with 42% direct alerting the 50% assumption would result in

., 21% of the alerted households attempting to alert others. This 21%

might successfully alert a non-alerted household 58% of the time (42% of the time an already alerted household would be contacted). Twenty-one percent multiplied by 58% produces an alerting increment of 12%, so that the initial 42% would be increased to a 54% alerting level.

Furthermore, as the Shearon Harris record:shows, the Estimates in NUREG/CR-2655 were based on questionable data and are probably low by 10-20 percentage points. Nevertheless, even after taking account of informal alerting and factoring in ,the.most relevant data on the arousal

~

' capabilities of sirens, the nighttime alertinglercentages for the fou'r plants would probably ' range from about 70% to 85%, falling'well below even the Staff's lenient 90% standard. Such levels of alerting cannot reasonably be viewed as satisfying the " essentially 100%" regulatory requirement.

.We refrained from raising a wintertime alerting issue in the Shearon Harris case on ou.r own motion only. because the licensee is to install tone alert radios, supplemental to the sirens, in the bedrooms of residences in the first 5 miles of the EPZ. In our judgment, there is a serious question whethe,r othe'r existing siren systems, in the absence of an equivalent supplsmental system for nighttime alerting, can meet regulatory requirements, particularly under winter nighttime conditions. There are several supplemental alerting technologies (such

> as tone alert radios, automatic telephone dialers, activation of alerting devices with signals carried by regular household power lines, and mobile sound sources) that can be anticipated, in conjunction with siren systems already in place, to alert more than 95% of sleeping people in an EPZ in any season of the year. It appears that such systems can be cost effective. For example, in the Shearon Harris case, the Applicants testified that the approximate cost of purchasing and distributing tone alert radios to about 600 houses in the first 5 miles of the EPZ would be $80,000.

In summary'and conclusion, the practical effect of the FEMA and NRC Staff positions on nighttime alerting in 15 minutes based on sirens anc informal alerting are that (1) in the summer, about 10% of the people.

will not be alerted -(more in hot climates), and (2) in the winter in cold climates, 15-30% of th'e people will not.be alerted. If.the Connission agrees that such alerting levels are acceptable and correspond to an absence of undue risk in the first 5 miles of an EPZ, then the present regulatory standard of " essentially 100%" alerting should be changed. If not, we reconnend that timely action be taken to ensure that siren notification systems at existing nuclear plants are

1. e

- - - , . -- ym . _,,,3

, . .~ . . . . . .

. ..m t

i reviewed against appropriate technical and legal standards and supplemented, if necessary.

Respectfully submitted.

0- 9 JpsL.Kelley [

n new 1 NN ey s ow%

D JamesH.Carppter,

$'d Gle7n O. Bright #

cc: M. Malsch, Acting General Counsel S. Chilk, Secretary '

Shearon Harris Service List F

l i

l l

l I - . .

j _

9

% o e- e *w.. +w