ML20041D261
| ML20041D261 | |
| Person / Time | |
|---|---|
| Issue date: | 02/01/1982 |
| From: | Quinn L NRC OFFICE OF THE EXECUTIVE DIRECTOR FOR OPERATIONS (EDO) |
| To: | Murley T NRC OFFICE OF THE EXECUTIVE DIRECTOR FOR OPERATIONS (EDO) |
| References | |
| NUDOCS 8203050033 | |
| Download: ML20041D261 (5) | |
Text
"
fi n.
k. EB2 D
3 FEB 1 1982 1982u 7 ne 2:w&'L* nim ni E
/
Tid H
13 l/
N MEMORANDUM FOR: Thomas E. Murley Director Regional Operations and Generic Requirements Staff FRON:
Liz Quinn Regional Operations and Generic Requirements Staff
SUBJECT:
PRIORITIZING CliGR AGENDA ITEMS During the past couple of months I have had opportunity to observe this office address its newly mandated responsibilities of reviewing generic requirements to be imposed by NRC on one or more classes of reactors.
I was astonished by the momoth volume of items initially identified as potential candidates and pondered how the Committee to Review i
Generic Requirements could identify and focus on those requirements which would contribute most significantly to the health and safety of the public while optimally using both NRC and licensee resources.
I also questioned how safeguards-related issues could be prioritized with safety-related issues. How can the randomness of possible accident sequences be weighted against the malevolent act of sabotage.
I support the CRGR's position that each Office should prioritize their proposed requirements but by what technique do you rank the overall agency initiatives received by the Comittee. Even though each requirement will have had a regulatory analysis performed, that just justified the individual requirement not how it relates with other requirements competing for possible licensee attention.
Therefore, I believe that there is a need for a ranking approach that'
/
can be applied by the DEDR0GR staff for the CRGR that is a less
/
burdensome analytical technique than probabilistic risk analysis while retaining the basic concept of risk and incorporating in a natural manner, cost considerations. A ranking scheme that has these attributes has been developed by Dr. Vernon L. Grose (published in " System Safety in Rapid Rail Transit", ASSE Journal, Volume 17, November 8,197a pp. is-1 l
M; " Controlling Terrorism and Other Security Threats Against -
l Multinational Corporations via Systems Methodology," IACP monograph l
1975; and " Converting Health Hazard Control from Moralism to Manage-l
_/
ment," Professional Safety, March 1980). The scheme presented here is
/
based upon his wort
/
/
It is proposed that Table 1 detailing the categories of concern be disseminated to the NRC staff. As part of the regulatory analysis package submitted to the CRGR tney would identify and support which l
level of the three categories was appropriate for their requirement.
l Thus they would be identifying the consequence of failing to correct i
OFFICE)
.... ~..
....... ~.... ~....
........ ~..........
..-.--m
.~.
- m m.~ ~ ~. m
~ ~. - ~ ~ ~ ~..
SulNAME) 8203050033 820201
~~.
.. ~ ~ ~.- ~ ~..
.. ~.. -. - ~ ~ ~
~ ~~ ~.~.....~
-. ~. ~. - ~... -
PDR REVGP NRCCRGR onek PDR NRC FORM 318 (10-80) NRCM Om OFFICIAL RECORD COPY usam mi-m..eo
1 l
FEB 1 sa2 4
i :
the concern, roughly estimating the frequency at which the consequence may be expected to occur, and the total cost of implementing the 4
requirement.
t In order to rank regulatory requirements, it is necessary to obtain a clear unde standing of the concerns they address and the consequences r
of failing to address those concerns. This understanding is obtained through construction of a " hazard scenario," prepared by the proposing Office.
In the teminology of probabilistic risk assessment, hazard i
I scenarios are simply event trees. This terminology is not employed here because this approach, as will be seen, can be employed effectively with less rigorous descriptions of the sequence of events leading to a hazard state. Every accident, or for that matter successful sabotage, is composed of many contributory elements. Hazard scenarios are_a brief description of combined causative factors which could lead to an accident. They could be described as "an anatomy of an accident."
l A given regulatory requirement may address many hazard scenarios.
In
,such a case it may be possible to combine a large msnber at those scenarios based upon their similarity in tems of the ranking factors l
given below. Nevertheless, after this combination has been achieved, there may still remain several hazard scenarios that are addressed by the proposed rule change. These should be ranked separately and the regulatory requirement should be ranked according to the highest priority scenario.
In categorizing the scenario, it is important to ensure that the frequency category corresponds to the expected frequency for the entire scenario, not just the fault addressed by the regulatory initiative.
For example, suppose the concern addressed by the regulatory initiative can only have catastrophic consequences in combination with-failure of two other mitigating systems. Then the frequency of the hazard scenario must be the expected frequency of the fault addressed by the regulatory initiative occurring in combination with failure of F
the other mitigating system from other random causes. Needless to say, this frequency is likely to be much lower than the expected l
frequency of the fault addressed by the regulatory issue.
Once the appropriate level of the proposed requirement has been identified by the proposing Office, for each category, the DEDR0GR staff could combine the level codes and locate the alphabetical code on the Priority Ranking List, Table 2 (which would not be circulated) to determine a numerical ranking for this requirement.
i The philosophy employed in ranking these categories is similar to that l
presented in NUREG-0739 An Approach to Quantitative Safety Goals for 4
Nuclear Power Plants, and may be sumarized as follows. Some levels of risk are intolerable and must be reduced no matter what the cost. -
Categories involving these risk levels are ranked primarily by risk.-
omcep suRuaMe>
oare >
NRC FORM 318 (10-80) NRCM 0240 0FFICIAL RECORD COPY uso,o i_m.m
FEB 1 1982
. In cases of approx mately equsi risk, categories are ranked by cost (to obtain the greatest risk reduction within resource limits).
Finally, in cases of approximate equal risk and cost, categories are ranked by severity of consequence to reflect society's apparent risk aversion when comparing a single infrequent large accident with a number of small accidents presenting equivalent risk.
For those categories involving risks below the " intolerable" risk levels, the ranking process changes to one driven primarily by risk reduction per dollar expenditure. Categories involving approximately equal risk reduction per dollar expended were then ranked in accordance with total risk reduction. Finally, in cases of approximately equal risk reduction per dollar cost and equivalent total risk reduction, the categories were ranked by severity of consequence to again reflect societal risk aversion to infrequent large accidents.
The details of the ranking approach can be modified without effecting the applicability of the general method. I:owever, this method does provide a consistent and logical approach to ranking proposed regula-tory requirements from all the Offices of NRC.
b[
Liz Quinn Regional Operations and Generic
.f
~
/
Requirements Staff p
Attachment Table 1 Table 2 l
l Distribution l
Vstello l
LQuinn l
DEDR0GR cf l
Central File l
PDR (NRG/CRGR) i l
I omce p.D.E.D..R..O.G..Rr...................
e SURNAME).b.d.00I
.h...
6t /...)../. 82.......
con >
OFF1ClAL RECORD COPY usam mi-m-m nac ronu sia 00% sncu cao
TABLE 1 CONSE0VENCE Level Criteria A
Severe core damage.
Essentially involves loss of all installed safety features. Severe, direct breach of containment.
B Severe core damage.
Containment fails to isolate or leak integrity lost.
Fission product release mitigating systems operate to reduce release.
C Severe core damage. Containment has degraded leak integrity then fails by basemat melt-through. All i
other release mitigation systems have functioned as designed.
D Limited to modest core damage.
Containment systems operate but containment leaks in somewhat degraded mode. No failures of engineered safety features beyond those postulated by the various design basis accidents are assumed.
EXPECTED FREQUENCY OF OCCURRENCE Level Criteria F
Once per 10 reactor years.
2 G
Once per 10 reactor years.
3 H
Once per 10 reactor years.
4 I
Once per 10 reactor years.
5 J
Once per 10 reactor years.
6 K
Once per 10 reactor years or more.
ESTIMATED COST OF REQUIREMENT i
Level Criteria L
Less than $ 10K per reactor unit.
M Less than $100K per reactor unit.
N Less than $1M per reactor unit.
0 Less_than $10M per reactor unit.
P Less than $100M per reactor unit.
I
.~..
TABLE 2 PRIORITY RANKING 1.
AFL
- 26. AIL 51.
CHL
- 76. DHL 1 01. BK0 2.
AFM
- 27. BGL
- 52. CHM
- 77. AKO 102. CJP 3.
AFN
- 28. AIM 53.
CHN
- 78. BIO 103. CK0 4.
AF0
- 29. BGM
- 54. CHO 79.
BJN 104. DGP 5.
AFP
- 30. AIN 55.
CHP 80.
BKM 105. DHO 6.
AGL
- 31. BGN
- 56. AKL 81.
CIO 106. DIN 7.
AGM
- 32. AIO
- 57. BIL 82.
CJN 107. DJM
.8.
AGN 33.
BG0
- 58. CIL 83.
CKM 108.
DKL 9.
AGO
- 34. AIP 59.
DFL 84.
DF0 109. BKP 10.
AGP 35.
BGP
- 60. AKM 85.
DGN 110. CKP 1
11.
AHL 36.
CGL
- 61. BIM 86.
DHM 111.
DHP 12.
BFL
- 37. CGM 62.
BJL 87.
DIL 112.
DIO 13.
AHM
- 38. CGN
- 63. CIM 88.
AKP 113.
DJN 14.
BFM 39.
CG0 64.
CJL 89.
BIP 114.
DKM 15.
AHN
- 40. CGP 65.
DFM 90.
BJO 115.
DIP 16.
- 41. AJL 66.
DGL
- 91. BKN 116. DJO e
17.
AHO
- 42. BHL
- 67. AKN 92.
CIP 117.
DKN 18.
BF0
- 43. AJM 68.
BIN 93.
CJO 118.
DJP
- 19. AHP 44.
BHM
- 69. BJM 94.
CKN 119. DK0 20.
BFP
- 45. AJN
- 70. BKL 95.
DFP 120.
DKP 21.
CFL 46.
BHN 71.
CIN 96.
DG0 22.
- 47. AJO 72.
CJM 97.
DHN l
23.
CFN
- 48. BHO 73.
CKL 98.
DIM 24.
CF0
- 49. AJi 74.
DFN 99.
DJL 25.
CFP 50.
BHP 75.
DGM 100. BJP i