ML16341B216
| ML16341B216 | |
| Person / Time | |
|---|---|
| Site: | Salem, Susquehanna  |
| Issue date: | 08/24/1979 |
| From: | Anderson F NRC OFFICE OF STANDARDS DEVELOPMENT |
| To: | Mattson R NRC - TMI-2 LESSONS LEARNED TASK FORCE |
| Shared Package | |
| ML16341B218 | List: |
| References | |
| NUDOCS 7910240699 | |
| Download: ML16341B216 (6) | |
Text
gt.A RfgI I((l,"~(f(l o
+a*++
UNITEO STATES NUCLEAR REGULATORY COMMISSION WASHINGTON, D. C. 20555 August "4, 1979 EhliL 2 r
I p
I
~
'OTE TO:
Roger J. Mattson, Director, TMI-2 LessoIIs'Learned Task Force FROM:
Fredric D. Anderson, Site Designation Standards Branch
'Division of Siting, Health 8 Safeguards Standards 4
SVBJECT:
COMMENTS ON RESPONSE TO BOARD QVESTION CONCERNING CLASS 9 ACCIDENTS (NOTE:
MATTSOH TO CVNHINGHAM, 8/16/79)
I reviewed your response to the Salem spent fuel board question on Class 9
accidents and the TMII-2 accident.
-I disagree with the position taken by the Task Force regarding the TMI-2 accident being classified as a Class 9
accident as defined in the proposed Annex to Appendix D of 10 CFR 'Part 50 (stated in Appendix I to R.G. 4.2, Revision 2, July 1976).
I have reviewed this position as a concerned NRC staff member and one of the staff listed in the cc list for your reply.
In truth, the definitions for accident classes as given in the Annex for environmental reports are not applicable to the determination of an actual accident.
Actual accident scenarios should not be forced to match the perceived analytical models for accidents used to postulate potential consequences.
Because of the many differences expected in accident scenarios depending upon the purpose.for. the evaluation, the NRC staff uses different accident scenarios (classes) for analyzing system failures as given in SRP and R.G. 1.70, for analyzing consequences from design basis accidents as given in SRP, R.G. 1.70 and specific guides such as R.G. 1.3, 1.4, 1.24, 1;25 etc.,
and for analyzing environmental impact from accidents as given in R.G. 4.2.
Therefore, the correct answer to the Board question would be that the Class 9 accident nomenclature was not applicable to assess, any real.
accident scenario.
If NRC feels compelled to respond to the Class 9 accident question for TMI-2 in a manner consistent with the conceived public understanding of a Class 9
- accident, the answer should be that the TMI-'2 accident divas not a Class 9
'accident.
The reason is that the commonly held definition for a Class 9
accident fs an accident resulting in fuel melt with failure 'of engineered
. safety features designed to mitigate the consequences of the accident--
the extreme case is containment failure.
Since TMI-2 did not experience fuel melt or ESF failure of a mitigative system, the accident cannot be classified as a Class 9 accident by common definition.
If NRC feels'hat the staff should force-fit the TMI-2 accident scenario with operator error and bad judgments into the R.G.
4.2 class accident scenario
0
Roger J. Hattson 2
which evaluates how radioactive material could be released to the environment and the consequences, the THI-2 accident auld not be deemed a Class 9
accident since the releases did occur by evaluated pathways.
Rather than defining whether conditions of the THI-2 accideqt. corresponded to the defined classes given in R.G. 4.2, I would rather'describe what accident classes did not occur at TMI-2.
l.
Class 9 accidents invo')ve sequences of postulated successive failures more severe than those postulated 0 'esia s ino 'D.ote tive systems and engineered safety features.
The THI-2 accident conditions were not more severe than postulated for any of the engineered safety features though the protective systems did take some severe conditions.
The consequences were not severe as defined for Class 9 accidents they were rather trivial.
2.
All Class 8 accidents are defined as events initiated by pipe breaks and reactivity additions and are relatable to design basis accidents
'sed in the safety aqalysis report.
The THI-2 accident was not initiated by either a pipe break or a reactivity addition 'and was not relatable to such accidents evaluated in the SAR.
3.
All Class 6 and 7 accidents involve fuel handling conditions.
THI-2 accident did not involve'fuel handling.
4.
Class 4 accidents are defined for BHR facilities.
THI-2 is a
PHR facility.
5.
Class 1 accidents are defined as trival incidents and TMI-2 accident was not a trivial.incident.
The accident classes that remain as contributo'rs to the THI-2 accident are:
Class 2, small release outside containment; Class 3, radwaste system failure; and Class 5, fission products to primary and secondary system (Pl(R).
The reported consequences from the THI-2 accident are consistent with the predicted conseouences from.such class accidents as the Clas's 2, Class 3
and Class 5 defined accidents.
If the NRC wishes to discuss the THI-2 accident in respect to system analysis categories as given in R.G. 1.70, then a further review of the B)I-2 scenario would be required.
In conclusion, the THI-2 accident was not a Class 9 accident by any rational analysis and comparison of the definitson for a Class 9 accident with the THI-2 accident scenario.
I would recommend that the NRC staff (maybe your Task Force) prepare an analysis and evaluation of the TMI-2 accident to show the accident development scenario in terms of the response of various systems especially in respect to fission product releases from the fuel; where, what and how much of the fission product inventory was inside containment, inside auxiliary building and released to the environment; and
Ik
r ~ ~ r g~
~
~
~
~
Roger J. Hattson 3
what the effect of the hydrogen exploding iozidj; the reactor vessel (regardless of the impossibility) would have been.
There are probably other unresolved questions in addition to the above that need investigation for comparison with measured or predicted results from the THI-2 accident.
cc:
Guy H. Cunningham Barry Smith Janice ftoore Gary Zeck Malt Pasedag Robert Tedesco John Yogi eric<de John Guibert.
Fredric D. Anderson Site Designation Standards Branch Office of Standards Development