ML20043F087
| ML20043F087 | |
| Person / Time | |
|---|---|
| Issue date: | 05/14/1990 |
| From: | Taylor J NRC OFFICE OF THE EXECUTIVE DIRECTOR FOR OPERATIONS (EDO) |
| To: | |
| References | |
| TASK-PII, TASK-SE SECY-90-173, NUDOCS 9006140153 | |
| Download: ML20043F087 (8) | |
Text
o0Oe00000e000000e0000000 8
RELEASED TO THE PDR
- p"%,g i
un/9o G
1:
6 t. -
inite 5
/
eeeeeeeeeeeeeeeeeeeeeeee
.e...
y POLICY ISSUE May 14, 1990 SECy_,0_1,3 (Information) f_gr:
The Commissioners
[
I f.nqm:
James M. Taylor Executive Director for Operations p
Sub.iect:
USE OF PRA-BASED SOURCE TERM METHODS Puroose:
To inform the Commission of staff plans to evaluate the use of PRA-based mechanistic source term methods in areas, other than siting, where deterministic methods (TID-14844) are i
currently used.
i Backaround:
Following the January 9, 1990, briefing by the staff on the status of development of updated source terms, the i
Commission directed the staff to submit a paper to the Commission " discussing the pros. cons, and feasibility of employing PRA-based mechanistic source term methodology (i.e., NUREG-1150 methodology) in all areas, other than siting, where deterministic (TID 14844) source term methodology is currently employed."
i Discussion:
A PRA-based mechanistic source term would consist of a i
-mechanistic estimate of fission product quantities and types based on our best understanding of the progression of a given accident sequence or sequences and the relevant physical phenomena involved.
As such, there could be many source terms depending on the choice of accident sequences and plant design. The source term would be based on I
probabilistic analysis of the likely sequences.
It could be highly plant specific, or could envelop a number of potentially generically important sequences.
Further division of the source terms would depend on whether the release is a release within the containment or a release to the environment.
A source-term used in the context of a replacement for TID-14844 would generally be defined as a release within the containment, t
Contact:
NOTE:
TO BE MADE PUBLICLY AVAILABLE
-W. Beckner, RES 1N 10 WORKING DAYS FROM THE DATE 492-3975 OF THIS PAPER M. Cunningham, RES 492-3965
'( 0*
l 9006140153 900514 PDR SECY 90-173-PDC l
W.
(
w
'f
-The Commissioners x E
This paper provides a description of the general pros and cons of using PRA-based mechanistic source term information, the types of source term information which are currently.
available (from the second draft of NUREG-1150), and the types of information which are required to address the issue of TID-14844 replacement.
Because PRA-based data have not yet been developed in an appropriate form to address this issue, a staff plan to generate and review the additional information is also provided.
General Pros and Cons The TID-14844 source term was originally developed to serve as a conservative basis for siting considerations. Over time, it has come to be used for a variety of other regulatory purposes.
The positive aspect of the use of TID-14844 is that it provides a single, well-understood, source term for many purposes. However, it does not fully represent the modern, more realistic, understanding of severe accident source terms. Use of TID 14844 may be optimistic in some applications and overly restrictive in others.
PRA-based source term methods provide a more accurate prediction of the expected course of a severe accident. The principal negative aspects are that these methods are much
- more complex and the results are still very uncertain. As discussed in the enclosure, PRAs such as NUREG-1150 provide source term information for many accident sequences in both probabilistic and non-probabilistic forms. To apply this-information in the context of a licensing review one would have to decide which sequences are to be considered and at what confidence level the source term information should be applied. To be most realistic,this would have to be done for each area-on a plant specific' basis. Thus, the application to specific issues in a plant or set of plants becomes much more complex and would require additional' applicant and staff resources to generate and review the information.
Areas amenable to use of pRA-based mechanistic source term information The areas (other than siting) where it may be appropriate to use PRA-based mechanistic source term information to replace the traditional TID-14844 source term are the following:
= _ = _
/
t
'The Commisaioners 3-a.
Fission product cleanup systems (sprays and filters) c.
Containment leak rate d.
Isolation valve closure time e.
Equipment qualification f.
Post-accident sampling systems g.
Shielding requirements for some emergency response i
facilities Source term information affects the design of equipment and associated technical specifications for equipment operation in various ways.
For example, typical information needed would'be the timing, location, and duration of a radioactive release, and the type and quantity of radioactive material, j
However, only a portion of this information may be needed, depending on the application involved.
For example, for proper design of some equipment,; only the time of initial release to determine start times may be needed.- Some equipment may only need a total integrated release to obtain radiation doses. Other equipment design may require detailed time-dependent release data for proper design and.
operation.
Staff olan for data development
'A highly. complex PRA-based mechanistic source term, which could be different for each particular plant and sequence involved, is not viewed as practical and the expected benefit is not believed to warrant such complexity.
- Rather, the staff is working toward a replacement for TID-14844 which would be an envelope of expected source terms from the most likely core melt scenarios as analyzed in state of the art PRAs.
Such an enveloping source term could be developed for each plant; however, the complexity and benefits of this approach are not favorable.
Rather, the staff believes that development of a single or limited set of source terms (for example, BWR vs PWR) would be sufficient for incorporating-up to date information and providing for more realistic treatment'-of certain engineered safety features without increasing the complexity of application.
Based upon the work done so far, the quantity of radioactive material in such a new source term is unlikely to be significantly different from TID-14844 in its effect on offsite consequence; however, differences in isotope species, chemical form or timing of release could affect the design of certain engineered safety features.
s:
x i
'i.
i
-The Commissioners The Office of Nuclear Regulatory Research will continue to work to develop the needed information, to review this information in the context of the regulatory areas identified above, and to make. recommendations on changes to present regulatory practice.- This work is being done as part of the effort already underway by the staff assessing longer term changes to the source term as discussed in SECY-89-341 and Staff Requirements Memorandum #M900109.
,/'
J or xecutive irector for Operations
Enclosure:
-Source Term Information in:NUREG-ll50 DISTRIBUTION:
Conunissioners OCC=
OIG LSS GPAJ REGIONAL OFFICES EDO
SECY
t Enclosure L
Source Term Information in NUREG-ll50 l-PRAs such as NUREG ll50 include both a set of detailed mechanistic (non-probabilistic) source term methods as well as less detailed probabilistic methods for extending the mechanistic source terms to large sets of accident sequences for uncertainty analyses. The results of the mechanistic analyses-generally are applicable only for a very specific accident sequence in one plant. They can be grouped together, however, to generate surrogate source terms for groups of accident sequences in a plant, or for an entire plant.
Such analyses may also vary significantly among plants and can be influenced l
by the calculational models used (e.g., the industry's MAAP code versus the staff's Source Term Code Package and its replacement, the MELCOR code).
Currently, source terms (and the unreleased quantities of radioactive material of interest in some issues such as equipment qualification) used in PRAs can be calculated and displayed in a number of ways (e.g., single-valued I
estimates, probability distributions). The Source Term Code Package (STCP) and the MELCOR code can generate point estimates of source terms based on mechanistic models, yielding simulations of fission product behavior in the reactor and containment systems for severe accident sequences (see, for example, NUREG/CR-4624, " Report on Radionuclide Release Calculations for Selected Severe Accident Sequences"). However, the STCP and MELCOR codes are large complex codes and are not now, by themselves, completely suitable for the uncertainty or probabilistic analysis of the many source terms required for PRAs.
In NUREG-ll50, a simple and parametric source term code was developed to
. simulate the detailed mechanistic codes and permit source term uncertainty analysis. Source term parameter probability distributions were used as input.
Such distributions were provided by a panel of source term experts after i
review and consideration of the physical processes, calculation results from the detailed mechanistic codes, and results from experimental studies.
For each accident sequence, or group of sequences having similar accident l-progression and phenomenological characteristics, the model generated an extensive data base of probability distributions for such source term characteristics as the magnitude of radionuclide release (by chemical group).
Because of their simplified nature, however, some source term parameters (e.g., timing) were not well characterized.
The results from the parametric code can be presented in the form of probability distributions of the fractions of radionuclides released (by chemical group) for individual accident sequences or sequence groups.
L Presentation methods include the probability distributions shown in Figure 1, l
or complementary cumulative distribution functions, CCDFs, shown in Figure 2, j
as absolute frequency distributions (per reactor year) summed over all L
accident sequences (as shown in Figure 3), or as sets of individual measures l
of such distributions (e.g., medians, means). These results can be cast into a variety of other forms, depending on the intended use.
In NVREG-1150 (second draft), source term results were provided in two ways, like those shown in Figures 1 and 3.
I
t l
l-z.
[
l
~
1.
l l
Release Fraction f
1.0E+00 1
l 1
1.0E- 01 g
median 2
~~
1.0E-02 g 3,
~
1.0 E - 03 g 5
1.O E- 04 g 1.0E- 05 NG I
Cs Te Sr Ru La Ba Ce Elemental Group Figure 1 Probability Distribution of Radioactive Release Magnitude for Specific Accident Sequence
l...
~ _
CONDITION AL EXCEEDANCE PROB.
1.0E+00 -
E V SEQUENCE 1.0 E - 01 g STATION BLACKOUT 1.0 E - 0 2 :
SGTR N
~
N 1.0 E - 03
' ' ' ' '.0E-01 1.0 E - 04 1
g 1.0E-04 1.0 E-03 1.0E-02 IODINE RELEASE FRACTIONS FIGURE 2. CONDITIONAL EXCEEDANCE PROBABILITY FOR SPECIFIC ACCIDENT SEQUENCES (EXAMPLES ONLY)
$l
l
-J
~ d
't' EXCEEDANCE FREQ. (PER R-Y).
1.OE-04 1
-l N
j 1.O E- 0 5 5
~
'l E
1.OE- 0 6 1
I
~_
1.OE- 07 5 1.OE- 08 1.OE-04 1.OE-03 1.OE-02 1.OE-01 1.OE+00 IODINE RELEASE FR ACTION FIGURE 3. EXCEEDANCE FREQUENCY OF IODINE RELEASE FRACTION (INCLUDING ALL ACCIDENT SEQUENCES) (EXAMPLE ONLY)
.-_