ML20213F990
| ML20213F990 | |
| Person / Time | |
|---|---|
| Issue date: | 09/09/1986 |
| From: | Read J, Soffer L Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML20213F903 | List: |
| References | |
| FOIA-86-776 NUDOCS 8611170273 | |
| Download: ML20213F990 (9) | |
Text
. \
REGULATORY SIMPLIFICATION OF FISSION PRODUCT CliEMISTRY J. B. Read L. Soffer ,
Regulatory Improvements Branch, Division of Safety Review & Oversight U.S. Nuclear Regulatory Commission, Washington, D.C. 20555 ABSTRACT The requirements for design provisions intended to limit fission product escape during reactor accidents have been based since 1962 upon a small number of simply-stated assumptions. These assumptions permeate current reactor regulation, and, while providing a high level of plant safety, are too simple to deal with the complex processes that can reasonably be expected to occur during real accidents. Potential chemical processes of fission products in severe accidents are compared with existing plant safety features designed to limit off-site consequences, and the possibility of a new set of simply-stated assumptions to replace the 1962 set discussed.
INTRODUCTION The defense-in-depth policy of nuclear power plant regulation leads not only to redundant provisions for accident prevention, but also to provisions to mitigate the off-site radiological consequences of any accidents that might occur despite preventative efforts. The design of dose mitigating safety features is chiefly determined by the most cost-effective means of meeting regulatory requirements imposed by the Nuclear Regulatory Commission (NRC) upon the plant design.
8611170273 861113 PDR FOIA PDR SHOLIYO6-776
. 2 1
Commission policy is to cast regulatory requirements in the least prescriptive form possible. Rather than specifying required equipment or design details, the NRC states the capabilities that mitigative provisions should have in terms of postulated accidents and the maximum off-site doses permitted. This leaves the means of achieving accident mitigation in the .
hands of the plant designer, while setting the minimum acceptable mitigation in terms of the ultimate goal of reduced off-site radiological consequences.
The postulated accidents used for this purpose are called design basis accidents (DBA), since they determine the overall acceptability of the plant by the NRC.
the most severe current DBA is that postulated in TID-14844,2 commonly called the " TID source term." This DBA consists of an assumed instantaneous dispersal into the containment free volume of the entire core inventory of noble gas fission products and half the core inventory of iodine fission products. Dose mitigation features are acceptable if it can be shown that the plant can retain this dispersed inventory sufficiently well to prevent off-site doses from becoming larger than those specified in 10 CFR Part 100 of the regulations. In the dose calculations, the containment is assumed to be pressurized to its designed capacity, leaking at its maximum permissible value but intact.
Implicit in the TID source term is the use of a single postulated release of energy and of the three fission product elements, Kr, Xe and I, to represent all aspects'of all accidents that should be mitigated. This is a gross .
simplification, suitable to its intended temporary uses in 1962, but no longer .
sufficiently descriptive of real demands that might be made of engineered safety features.
3 DISCUSSION A detailed model that included all the chemical processes affecting the transp1rt of fission product elements following an accident would be a monumental undertaking. Such a model would have to describe such diverse processes as the inhomogenous reactions by which most volatilizations would .
occur, Szillard-Chalmers processes, radioloysis and gas phase free radical reactions. Many processes that might be described as " laboratory curiosities,"
such as the hot atom reactions of alkali metal ions recoiling from the decay of noble gases, would be capable of significant effects. The current generation of computer models fall for short of such completeness; none the less, they are already too complex to be practical for use in the licensing review of nuclear power plants.
A licensing review is the means by which the NRC staff reaches consensus as to whether it will suppart or oppose an application for either a new reactor operating license or a modification to an existing license. For equipment and structures designed to reduce off-site doses in the event of an accident, the staff must decide if more mitigative features are needed, if the capacities of the proposed equipment is sufficient, and what surveillance requirements are needed to assure proper maintenance. Accident descriptions useful for such purpose are quite different from those used in probabilistic risk assessment, since accidents within a plant's design basis necessarily contribute little to that plant's risk. Risk is dominated by very low probability accident sequences in which the engineered safety features fail, and it is obviously not possible to' establish the capacities and surveillance requirements for ,
equipment based upon sequences which assume that equipment to fail. ,
- . 4 Mitigative features can be active, such as containment sprays, or passive, such as the containment itself. Passive systems have the advantage of not requiring power, ard therefore they are more likely to be available if needed, but, since they must rely upon inherent processes, passive systems are usually slower-acting and less efficient than powered safety features. The present .
codification of staff review practices, the Standard Review Plan 3 (SRP),
disallows many naturally occurring chemical and physical processes as being ineffective in reducing off-site doses.
Revision of NRC use of severe accident assumptions in licensing of future plants is to be implemented by changing the SRP and the Regulatory Guides and contractor reports referenced in the SRP. These revisions will attempt to contain the following improvements:
1.) Both active and passive dose mitigation features will be considered, 2.) All, except for the least likely severe accident sequences for particular plant designs will be used to assess dose mitigation features, 3.) Any margin-of-safety involved in a dose calculation will be explicitly stated and explained, without reliance on undue overestimation, and 4.) The review will stress design properties that decrease off-site risks.
It is not at present feasible to automate an engineering review, and the current generation of accident-modeling computer codes is useful only as a tool when applied to SRP revisions and not as a direct replacement for T10-14844. What is needed for licensing reviews is the time-dependent rates at which fission products enter the containment, the accompanying energy and gas release rate into the containment atmosphere, and a general description of the physical and chemical states of the fission products in the release, for each of a set of representative accidents.
5 Figure 1 is the calculated rate of iodine release to the containment of a PWR undergoing a TMLU accident, a transient in which neither the steam generators nor " feed-and-bleed" is able to remove decay heat due to equipment failures.
The public risk from this sequence is dominated by the possibility of containment failure. The licensing review under the SRP, however, must examine .
the plant design for its ability not only to maintain containment integrity, 5
at a but to recapture airborne fission products in the containment atmosphere rate sufficient to prevent their leaking into the environment in amount capable 6
of exceedir.g allowable dose guidelines In Figure 1, the released iodine is modeled as a condensation aerosol obtained from vapor produced in the core by both coherent vaporization o)
Cs I tc) -+ CsI (Q an heterogeneous reactions, as, for example.
(2) 2 Cs*(c)
- o"(c) + Mt o (9 --* 2 cs ou (9 Such an aerosol would be rapidly depleted by 0.2M sodium borate or boric acid solution sprays, and would have no significant diffusive loss through the containment walls and isolation valves. Even more rapid depletion of HI or 12 would be predicted. If containment failure is prevented, the most likely means of fission product iodine escape into the environment would be as iodoalkane vapor produced in the containment sump solution.7 The licensing review, therefore, assures off-site safety by requiring prompt concentration of fission product airborne releases into the sump, combined with steps to prevent the radiolytic evolution of iodoalkanes from the sump solution.
in the TMLU sequence of Figure 1, about 30% of the core inventory of iodine is ,
estimated to be released between 180 and 200 minutes after scram, with much
6 smaller release continuing until 1000 minutes. This release history could be combined with depletion and transport models of any level of complexity and integrated numerically. The existing SRP, however, is structured to combine diverse engineering disciplines' reviews by the use of first-order linear differential equations,
- ijb k (3) dI E m where x; is the amount of isotope i in location j. Solutions to (3) are the Bateman function (8), and the coefficients,h;g, can be chosen to represent growth, decay, deposition, filtration or transport, provided only that the processes represented are suitably first-order. A conservative release history, chosen to include a margin for error in the calculations, can be ootained easily by simplifying Figure 1 as a first order process. For example, the peak release rate of 3% per minute can be assumed to apply to the full period of 180 to 200 minutes, and the rate of 10-6 per minute to the period 200 to 1000 minutes. These assumptions double the integrated iodine release to 60%.
In Figure 2, containment concentration histories have been plotted for the simplified model above and for the TID-14844 release currently described in Section 6.5.2 of SRP. In both cases, a typical PWR containment spray mitigative safety feature with a molecular iodine and particulate depletion coefficient of 0.33 per minute has been assumed. Under the current SRP, 49%
of the iodine core inventory is instantaneously airborne as molecular or particulate' forms, and 1% as " organic iodine." As modeled by the current generation of NRC computer codes, all of the iodine is particulate, there being .
no algorithm for conversium by radioloysis or other reaction to any other form.
y From Figure 2, it is seen that the present SRP places risk dominance at the site boundary during the early hours of the accident upon molecular iodine escaping during the first few minutes, while over the 30 days following the accident, the 1% of the iodine assumed to persist as organic vapor becomes more important. As typified by the TMLU model, however, for most accidents .
beginning as transients, fission product release does not become significant until a few hours or more have passed. Indeed, where hydrazine is used as spray additive to reduce iodine, all of this material would have been expended and oxidized by the containment atmosphere prior to the TMLU iodine release.
CONCLUSIONS Both off-site risk and core-melt frequency of a nuclear power plant design are usually dominated by a comparatively small number of accident sequences.
Among these sequences will be found the accident conditions most likely to lead to consequences which, were they not mitigated, would exceed allowable dose guidelines. The need for, capacity of, and surveillance requirements upon dose mitigating design features can be established by using simplified first-order generation and depletion assumptions to calculate fission product concentrations in the containment atmosphere and other internal plant volumes.
Such assumptions are capable of approximating, but with a reasonable margin-of-safety, the internal fission product concentrations predicted by accident simulation codes, yet are simple enough to be convenient for use in dose assessment modeling.
REFERENCES ,
- 1. "U.S. Nuclear Regulatory Commission Policy and Planning Guidance 1986",
NUREG-0885, Issue 5, February, 1986.
- - . 8
- 2. DiNur,no, J.J., Anderson, F.D., Baker, R.E., and Waterfield, R.L.,
" Calculation of Distance Factors for Power and Test Reactor Sites,"
Technical Information Document 14844, U.S. Atomic Energy Commission, March 23, 1962.
- 3. " Standard Review Plan for The Review of Safety Analysis Reports for Nuclear Power Nuclear Power Plants," LWR Edition, USNRC, NUREG-0800, July, 1981.
- 4. Silberberg, M., Mitchell, J.A., Meyer, R.0., and Ryder, C.P.,
" Reassessment of The Technical Bases for Estimating Source Terms,"
NUREG-0956, USNRC, May 23, 1986.
- 5. Code of Federal Regulations, Title 10, Part 50, Appendix A, General Design Criteria for Nuclear Power Plants, Criterion 41. U.S. National Achives and Records Administration, Washington, D.C.
- 6. ibid, Part 100, paragraph 11.
- 7. Beahm, E.C., Shockley, W.E., and Culberson, 0.L., " Organic Iodide Formation Following Nuclear Reactor Accidents," NUREG/CR-4327, December 1985.
- 8. Bateman, H. " Solution of a System of Differential Equations Occuring in the Theofy of Radioactive Transformations," Proc. Cambridge Phil. Soc. 15, 423(1910).
.. 9
. . FIGURE 1. Csl release rate to containment atmosphere for the PWR TMLU sequence, as fraction of iodine inventory per minute.
(private communication, E.G.Cazzoli, Brookhaven Nat'l. Lab.)
l 59i Fej
?~ -
d1 2 5$ }
5,~ l esi
}
h21 b
z,.-1.
E 3]
re -
U3k feI 7: ,
S ,,, g,, go,a sie eso is.o is o 1O 0 10 8 80 8 Tir1C titlNF FIGURE 2. Containment atmosphere concentration, as percent of iodine inventory, computed for current assumption releaserateinFig.1, assuming 0.33mingandfortwicethe depletion rate.
\
\
blo%
=
2 " 1{ '
! ! SRP G.5 2. 2mm a As I AND 2 i PARTitutATE g }%
m I i
2 i
2 1 1 '
ys% ...........[...........--..--....--..
l 9
i 1
,. %- l I
e L ________ _
I' u
CJ4 * > - - '
O 30 40 (,0 83 (60 IO l't0 Ib' 0 14 0 200 0 M enE5 AFTER SCRAN\