ML20211A657

Discusses Considerations of Societal Risk in plant-specific, site-specific Application of Safety Goals & Definition of Subsidiary Criteria

ML20211A657
Person / Time
Issue date:	06/11/1997
From:	Sherry R Advisory Committee on Reactor Safeguards
To:	Advisory Committee on Reactor Safeguards
Shared Package
ML20211A627	List: ML20211A623 ML20211A641 ML20211A657
References
ACRS-R-1723, NUDOCS 9709240321
Download: ML20211A657 (8)

Text

[

UNITED STATES o,?

NUCLEAR REGULATORY COMMISSION

'{

f ADVISORY COMMITTEE ON REACTOR SAFEGUARDS g

c ASHIN' ton, C. C. 20s'.5

....+

June 11,1997 MEhiORANDUM TO:

ACRS Members FROM:

Rick Sherry, Senior ACRS Fellow d

SUBJECT:

CONSIDERATIONS OF SOCIETAL RISK IN PLANT.

SPECIFIC, SITE SPECIFIC APPLICATION OF SAFETY GOALS AND DEFINITION OF SUBSIDIARY CRITERIA Backcround in its November 18,1996 report (ACRS 1996), the ACRS stated that "the safety goals and subsidiary objectives can and should be used to derive guidelines for plant specific applications." In its April 11,1997 report (ACRS 1997), the ACRS stated further that "Quantification of the LERF [large, early release frequency) at each site is needed to ensure the appropriateness of the choice of the LERF acceptance criterion proposed in draft Regulatory Guide DG 1061 and draft Standard Review Plan sections that support risk-informed, performance based regulation."

In a Staff Requirements Memorandum dated April 15,1997 (SRM 1997), the Commission stated (referring to Direction Setting Issue 12), "The staff should develop objective standard (s) for the application of risk informed, performance based and risk informed less prescriptive approaches to regulatiors on an expedited basis. Such standard (s) could be in the form of individual plant safety goals and subsidiary objective performance criteria as discussed in the issue paper."

11EE An important issue that arises when considering application of the safety goals on a plant-specific, site specific basis is whether individual risk alone is adequate for characterizing risk or should other risk metrics, such as societal risk, be considered in assessing risk informed regulatory applications?

ATTACHMENT 2 9709240321 970919 PDR ACRS R-1723 PDR

o

.DJscussiott The two quantitative health objectives (QHOs) delineated in the NRC Safety Goal policy statement (NRC 1986) are both stated in terms of individual risk. Because of this formulation of the QHOs, and as pointed out many years ago by the ACRS (ACRS 1983), " Larger societal risks are permitted for the nuclear power plant which has the larger surrounding population...." The ACRS has proposed in the past that a societal risk goal be applied to early fatalities (ACRS 1980). Recently, other countries have developed no. lear power plant quantitative afety criteria that explicitly include societal as well as individual risk (Versteeg-1992).

Contained in the NRC Safety Goal Policy Statement are separate views of Commissioner Bernthal, which include the following:

The absence of such explicit population density considerations in the Commission's 0.1 percent goals for offsite consequences deserves careful thought. Is it reasonable that Zion and Palo Verde, for example, be assigned the same ' standard person' risk, even though they pose considerably different risks for the U.S. population as a whole? As they stand these 0.1 percent goals do not explicitly include population density considerations.

Although it may have been acceptable to neglect explicit considerations of societal risk in the safety goals when the goals were to be used in a generic fashion to assess the risk of the population of plants as a whole, it is less clear that this is acceptable when the goals are used t

to assess the acceptability of proposed changes on a plant specific basis.

The Sandia siting study (NRC-1981) indicates that total (as opposed to individual) early fatality risk is very sensitive to site population distribution and that the mean number of early fatalities is determined by the average density of the exposed population. In the Sandia siting study, there was a difference of three orders of magnitude in the calculated mean number of early fatalities among 91 sites using the same source term, same emergency response, same wind rose, and same rneteorological record. The only difference in this analysis was the use of a site specific population distribution. On the bases of 1970 census data, the population density within five miles of the plant among the 91 sites examined had the following characteristics:

Median

- 40 people per square mile 90th Percentile - 190 people per square mile Maximum

- 790 people per square mile Figure 1 shows the calculated number of early fatality results from tids study.

Benchmark calculations were performed for the " Task Force on Interim Operation ofIndian Point" (NRC 1980) to assess the impact of population density on offsite consequences. For these calculations, four high-density sites (Zion, Indian Point, Limerick, and Fermi), one average density site (Palisades) and one low-density site (Diablo Canyon) were selected. A j

j 3

3 l

standard plant was then located at each site, identical weather sequences and emergency responses were arsumed, and wind rose weighted 1970 popui tion distr,'butions for each site were utillud. This analyses allowed the comparison of the calculated total early fatality risk i

for each site. Figure 2 summarius the results of this comparison in the form of l

complementary cumulative distribution functions (CCDFs) for total early fatality risk. This figure shows that the three sites with the highest population density have similar risk profiles and are substantially above the risk curves for the " average" site and for low population-i density sites. On Table 1 these curves have been reduced to single valua nowing the j

expected consequences (number of early fatalities per year). From this table, it is obsened i

that the total early fatality risk is an order of magnitude greater for ti ' three highest

]

population density plants than for the " average" plant and more than two orders of magnitude l*

greater than for a low population density plant.

Table 1 -

a l

Expected Annual Consequences l

Site Early Fatality Risk i

i e

Indian Point 6.1x10'8 4

1 l

Zion 4.7x10

Limerick 3.5x10'8 j

Fermi 9.2x10d

^

d Palisades 2.9x10 4

l Diablo Canyon 1.6x10 s i

Figure 3 shows the distribution of population densities within one mile wide annular rings within five miles of each miclear plant (population densities projected for the year 2000) j (NRC-1979). This figure indicates that the population densities for the top 10 percent of j

nuclear plant sites exceed the median site population by approximately one order of magnitude. These results indicate that if all plants were to exactly meet the individual early 1

fatality risk QHO then the top 10 percent of the plants would contribute the h of the c

.l societal risk.

M These results suggest that a number of the high population-density sites may have an order of j

magnitude or larger societal risk than the median site. - For plants at these sites, it is not clear that a risk met:ic that only considers individual risk is adequate. Consequently, the Committee may wish to consider whether societal risk should be considered in applying the safety goals on a plant specific, site specific basis.

T

g.

4 Strawman Annrneh One possible approach for addressing societal risk impacts for high population-density sites that would not result in the need for revisiting the definition of the safety goals and could be implemented using the approach delineated in draft Regulatory Guide DG 1061 (NRC-1997) is discussed below.

For the very high population-density sites, the QHOs for individual early fatality risk and individual latent cancer fatality risk could be reduced from their nominal values by a i

" population density reduction factor" to compensate for the increased societal risks associated with the high-population-density sites. On the basis of the results presented above, a i

reduction of approximately a factor of 10 in the individual early fatality QHO would be required for the highest population density sites to assure that the societal early fatality risk for these sites would not exceed the societal risk for an " average" site. The reduced QHO could then also be used to calculate a site specific LERF. However, for these sites, it seems prudent to _ require a full Level 2/ Level 3 probabilistic risk assessment and direct comparison with the (reduced) QHOs because of the large societal risk potential. - The use of the simplified contairanent analyses procedures contained in Appendix B to draft Regulatory j-Guide DO-1061 and use of a LERF as surrogate for the QHos may not be appropriate.. The 4

core damage frequency (CDF) could probably remain at lx10d since this value for the CDF has been demonstrated in NUREG-1150 to result in an individual early fatality risk that is substantially below the safety goal QHOs.

Esferences (ACRS 1980)-

"An Approach to Quantitative Safety Goals for Nuclear Power Plants,"

NUREG 739,- Advisory Committee on Reactor Safeguards, U.S. Nuclear Regulatory Commission, October,1980.

(ACRS-1983)

"ACRS Comments on Proposed Safety Goal Policy Statement," Report to Nunzio J. Palladino, Chairman, U.S. Nuclear Regulatory Commission -

from J. J. Ray, Chairman, Advisory Committee on Reactor Safeguards, January 10,1983.

(ACRS-1996)

" Plant Specific Application of Safety Goals," Report to Shirley A.

Jackson, Chairman, U.S. Nuclear Regulatory Commission from T. S.

Kress, Chairman, Advisory Committee on Reactor Safegu -6, November 18,1996.

(ACRS-1997)

" Risk-based Reguintory Acceptance Criteria for Plant-Specific Application of Safety Geals," Report to Shirley A. Jackson, Chairman,.

U.S. Nuclear Regulatory Commission from R. L. Scale, Chairman, Advivry Committee on Reactor Safeguards, April 11, 1997.-

,~r,

-e.,--.

.e~,-

n, a

9 5

l (NRC 1979).

" Demographic Statistics Pertaining to Nuclear Power Reactor Sites,"

NUREG 0348, U.S. Nuclear Regulatory Commission, October,1979.

- (NRC-1980)

.Bemero R. M., et al., " Task Force on *nterim Operation of Indian Point," U.S. Nuclear Regulatory Commission, August,1980.

(NRC-1986)

" Safety Goals for the Operation of Nuclear Power Plants; Policy Statement Publication," U.S. Nuclear Regulatory Commission, August 4, 1986.

- (NRC 1981)

Aldrich, D.C., et al, " Technical Guidance for Siting Criteria Development," NUREG/CR-2239, December,1982.

o' (NRC-1997)

"An Approach to Using Probabilistic Risk Assessment in Risk Informed Decisions on Plant Specific Changes to the Current Licensing Basis,"

Draft Regulatory Guide DG-1061, March 28,1997.

l (SRM 1997)

" Staff Requirement - COMSECY-96 Risk Informed, Performance-Based Regulation (DSI-12)," Memorandum to L. Joseph Callan, l

Executive Director for Operations, and Karen D. Cyr, General Counsel, from Annette L. Vietti-Cook, Acting Secretary, U.S. Nuclear Regulatory Commission, dated April 15,1997.

1 (Versteeg 1992)

" Showing Compliance with Probabilistic Safety Criteria and Objectives," Reliability Engineering and System Safety, 35 (1992) 39.R, cc:

J. Larkins, ACRS i

R. Savio, ACRS S. Duraiswamy, ACRS 4

A. Cronenberg, ACRS ACRS Senior Engineers i

l 5

i i

i

(

i R

1

/

' '. 60 50 40 m

4 b, 30 W

I

~

l 20 10 I'

I' I ""

0 0

200 400 600 800 1000 MEAN EARLY FATALITIES CONDITONACON SST1 i

Figure 1 Histogram of Mean Early Fatalities l'

for 91 Sites, Conditional on an SSTI release.

Assumptions:

1120

,MWe reactor, a representative meteorological record, and Summary Evacuation.

I

FIGURE 2 EARLY FATALITY RISK FOR DIFFERENT 5!TES 1

,,s.

10'4 i iiion i iiion i

,iiio

,, iini, i i i ii a

_~

1. 1.P.

2.

ZION

~

3. LIMERICR 4

FEIMI

! 5 E.

>Alf ? Aft _t 6.

>!ASLO CAmvii' r

i l10-6 P

m

{

l

~

6

~

C l3 10'I 1

\\

4 l

10'I l

~

~

4 2

gs

/' i ' ' ' ' ' ' '

10'I 0

I I

3 4

5 10 10 10 10 10 10 4

X.EARLYFATALIT!ES(SUPPORTIVETREATENT) i i

j NOTE: THE"E ARE LARGE UNCERTAINTIES WITH THE ABSOLUTE VALUES PRESENTED IN 4

ASSUNPTIONS:

2)i1.P. UNIT 3POWERLEVEL(3025fefT).

1 SURRT DESIGN.

i 3h WITHIN 10 MILES. ENTIRE CLOUD EXPOSURE + 4 N00R5 4ROUN j

N0 SHIELDING BEYOND 10 MILES.. ENTIRE CLOUD EXPOSURE + 7 DAY GROUND EXPOSURE SHIELDING BASED ON NORMAL ACTIVITY.

4

4) WIND ROSE WEIGHTED 1970 CENSUS POPULATION DISTRIBUTION.

i 5J 3DENTICAL 91 WEATHER SEQUENCES FOR ALL SITES.

2 l

4 l

~

. - - -. = -

l POPULATION DENSITIES WITHIN p

ANNULAR RINGS TO 6 MILES

,,, g PROJECTED FOR THE YEAR 2000 j

(X Percent of the Plants Hove Population

{

Denshies Lees Then or Equal to That Shown) 500 M

=

90TH PERCENTILE -

d I

wa<D0ma 300 Em 1

w 9m WE I

80TH PERCENTILE 200

}

100 50TH PERCENTILE.

I 10TH PERCENTILE J

i i

i 0

1 2

3 4

5 MILES FIGURE 3 i

9 j

1

-, - - -