ML18033B561
| ML18033B561 | |
| Person / Time | |
|---|---|
| Site: | Browns Ferry  |
| Issue date: | 11/20/1990 |
| From: | Office of Nuclear Reactor Regulation |
| To: | |
| Shared Package | |
| ML18033B560 | List: |
| References | |
| NUDOCS 9011300116 | |
| Download: ML18033B561 (3) | |
Text
~pS REGS~
O~
Cy i~
O
/p ++*+~
UNITED STATES NUCLEAR REGULATORY COMMISSION WASHINGTON, D. C. 20555 EtICLOSURE SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION TENNESSEE VALLEY AUTHORITY BROWNS FERRY NUCLEAR PLANT UNITS 1
2 AtID 3 DOCKET NOS. 50-259 50-260 AND 50-296 By letter dated tlay 31, 1990, Tennessee Valley Authority (TVA) (the licensee) submitted an analysis supporting their claim that the toxic chemicals emitted during a barge accident on the Tennessee River would not affect the operators in the control room of the Browns Ferry plant.
In this submittal, out of the 20 different chemicals identified to be transported by barges, TVA eliminated I4 because these chemicals were either not toxic or TVA was able to prove that they would never reach toxic concentrations in the control room.
The remaining six chemicals (acrylonitrile, benzene, chlorine, ethyl benzene, toluene and vinyl acetate) could, under certain circumstances, produce toxic concentrations in the control room.
- However, TVA demonstrated by analysis that, except for chlorine, toxic concentrations for these chemicals would not be reached until more than 2
minutes after their presence was detected by odor.
This time period and means of detection are in conformance with the guidance provided in RG 1.78.
As for chlorine, which would reach toxic concentrations in about 15 seconds, TVA conducted a probabi listic analysis demonstrating that the probability of chlorine invading the control room is less than 1E-6.
This probability level meets the criteria of SRP Section 2.2.3 which states that "the expected rate of occurrence of potential exposures in excess of the 10 CFR Part 100 guidelines of approximately 1E-6 per year is acceptable if, when combined with reasonable qualitative arguments, the realistic probability can be shown to be lower."
The NRC staff reviewed and accepted TVA's analysis (SER dated September 14, 1990) provided that all control room personnel would undergo training in recognition of threshold odors for the five hazardous chemicals whose presence in the control room is based on their odor detection.
In the letter 'dated October 15, 1990, TVA disagreed with the staff's provision and claimed backfit arguing that in all previous licensing actions the NRC interpreted the RG 1.78 definition of frequent shipment as applying to each individual chemical and not to the aggregate of chemicals shipped by barges.
Since, in the case of Browns Ferry, none of the toxic chemicals were shipped more frequently than 50 tiIIIes per year, there is no need to consider them in the safety analysis.
Although this interpretation of the RG 1.78 guidance by TVA and previously by the NRC was incorrect, reversal of the staff 's past inconsistent application of this position would require imposing a backfit (see cover letter),
and in the case of Browns Ferry would necessitate proving that either:
(1) a concern for adequate public health and safety exists, or (2) a substantial improvement in overall safety could be realized.
- However, as detailed below, the staff has concluded that such proof is neither generically nor plant specifically substantiated.
The staff extended TVA's evaluation of Hay 31, 1990, from a probabilistic analysis for release of chlorine gas alone to an aggregate of all six toxic chemical gases.
TVA's analysis was previously reviewed and accepted by the 90ll300116 901120 PDR ADOCK 05000259 P
PDC NRC (SER dated September 14, 1990), including their methodolgy.
In the chlorine probabilistic analysis TYA considered probabilities of individual factors which could cause release of chlorine gas to the control room.
These factors are:
Probability of accident per barge-mile-Shipments of chlorine per year-Length of river within 5 mile radius from the plant-Probability of toxic gas release per accident-Probability of transport of toxic gas cloud toward the plant 1.8E-6 30
",.1 miles 2.3E-2 9.1E-2 Nultiplying together all these factors, total probability of chlorine ingress into the control room was determined to, be 8E-7.
Since al 1 the probabi 1 isti c parameters in this analysis are applicable to other toxic gases, substituting frequency of transport of all toxic gases of 118 shipments per year, for the frequency of chlorine shipment, total probability of the ingress of all toxic gases to the control room is determined to be approximately 3E-6. It should be remembered, however, that before the 10 CFR Part 100 guidelines could be
- exceeded, all control room operators would have to be incapacitated at the same time that an accident occurs which, without the operators intervention, would cause a core melt and result in release of radioactivity.
For a normal operating plant, the probability that a core melt would occur within two hours after the operators became incapacitated is also estimated to be very low, approximately 2E-6.
Therefore, the total probability that the 10 CFR Part 100 guidelines will be exceeded are definitively well below the SRP Section 2.2.3 criterion of 1E-6.
It can be concluded, therefore, that even if all toxic gases transported by barges past Browns Ferry are included, the predicted probability that a toxic gas release would result in a severe accident condition exceeding 10 CFR 100
,guidelines is sufficiently small and meets the staff's established regulatory position.
Principal contributor:
K. Parczewski Dated:
November 20, 1990
Distribution (Doeg~et File NRC PDR Local PDR S.
Yarga G. Lainas J. Partlow
'F. Miraglia N. Russell EDO F.
Hebdcn S. Black B. Hilson P. Kellogg C. Patterson M. Krebs T. Poss OGC E. Jordan K. Parczewski ACRS (10)
BFN Reading File cc:
Plant Service list 14-E-4 14-H-3 12-G-18 12-G-18 12-G-18 16-H-15 RI I RII RII 15-8-18 MNBB-3701 7-D-4