ML20134P252
| ML20134P252 | |
| Person / Time | |
|---|---|
| Site: | Nine Mile Point  |
| Issue date: | 02/14/1997 |
| From: | Mccormick M NIAGARA MOHAWK POWER CORP. |
| To: | Mary Drouin NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES) |
| References | |
| RTR-NUREG-1560 NMP1L-1182, NUDOCS 9702250233 | |
| Download: ML20134P252 (7) | |
Text
1 i
1 NIAGARA MOHAWK G E N E R ATIO N NME ME POWT NUCLuM SMTOUWE ROAD. P o. BOX 63. LYCOMING NEW YORK 130R3 BUSINESS GROUP MARTIN J. MCCORMICK JR.
%c2 President Nuclear Engineering February 14,1997 NMPIL 1182
. Mary Drouin Office of Nuclear Regulatory Research Mail Stop T-10-E50 US Nuclear Regulatory Commission l
Washington, DC 20555-0001 RE:
Nine Mile Point Unit i Nine Mile Point Unit 2 Docket No. 50-220 Docket No. 50-410 DPR-63 NPF-60
Subject:
Comments on Draft NUREG-1560
Dear Ms. Drouin:
l A December 11,1996 letter from F. J. Miraglia (NRC) to B. R. SyMa requested
(
comments on draft NUREG-1560 by February 14,1997. The purpose of this letter is to l
provide Niagara Mohawk's comments, which are attached.
l l
i Since the NUREG discusses issues of broad industry interest as well as issues specific to j
Nine Mile Point Units 1 and 2, our comments, attached, are grouped accordingly. The first group ofcomments deals with generalindustry issues. The second group of j
comments deals with issues specific to Nine Mile Point Unit 1 The final group of j
comments deals with issues specific to Nine Mile Point Unit 2.
j It is requested that my office be kept informed of the resolution of our comments. Thank you for your cooperation in this regard.
l i
i l
A00 0
250076 M822gggggggo 9
b Page 2 -
i 1
Questions can be referred to Robert F. Kirchner at 315-349-7323. If we can be of further assistance, please call.
Very truly yours, m
Martin J. McCormick Jr.
Vice President Nuclear Engineering Attachments i
cc:
(Dammmagt Control Desk i
Mr. H. J. Miller, NRC Regional Administrator Mr. S. S. Bajwa, Acting Director, Project Directorate I-1, NRR Mr. B. S. Norris, Senior Resident Inspector D. S. Hood, Senior Project Manager, NRR Records Management 4
s J
Draft NUREG-1560 General Comments G-1 The definition of a quality PRA outlined in the NUREG is subjective and should not necessarily be derived from comparison of IPE results. While the information on PRA quality is useful, the PRA quality section of this NUREG should be characterized as specifically applicable only to the NUREG compcison effort.
Any applicability to the creation of an industry standard should be classified as developmental in nature. An industry-wide standard for PRA quality should be based on a broader and more deliberative development effort that involves practitioners from various organizations. This effort should include, among other things, an assessment of the benefit versus cost of additional analytical detail in specific areas of PRA.
G-2 (Page xix) The NUREG indicates that MAAP is not a comprehensive computer code. MAAP models the complete progression of a severe accident and is comprehensive in its treatment. MAAP can be characterized as limited by the uncertainty in severe accident phenomena and simplification due to the need for relatively fast computation in PC and Vax environments, but we believe it should not be called non-comprehensive.
G-3 (Page xix) The NUREG indicates that invalid HRA assumptions have been used in' IPEs. However,in section 13 of the NUREG such HRA issues are classified as inconsistencies with potential explanations. The characterization of appent HRA anomalies is more appropriately described in section 13. In fact, on Page 1-3 the authors indicate that they did not consider the accuracy of the information in the IPEs. It would appear that references to invalid HRA assumptions cannot be defended by the body of the NUREG and it is premature to label such issues cs i.
invalid.
G-4 (Page 3-1) Statement "However, BWRs only remove heat directlyfrom the primary system..." ignores the shutdown cooling function and, for a few older BWRs, emergency condensers.
r G-5 (Page 12-13) The NUREG indicates that' isolation is a' negligible contributor to because the containment is inerted and isolation failure can be early release partly;riu'mber oNaises~ario"[Efduritik normal operationii easily detectedrA j
response. Foi containmenfisolation suce'ess' Eportion of these valves must transfer closed on demand. An inerted containment has no bearing on this failure mode. The reliability of such containment penetrations is more tied to design issues (i.e.,%e of fall'close'd'AOVs and check valves). It should be pointed out 4
that containment isolation, while reliable, is not a negligible contributor to risk.
The risk achievement worth of such valves can be significant in terms oflarge-l early release.
I G-6 (Page 6-16) The use of 1/3 failure in place of zero failures is called inappropriate.
The use of a fraction of a failure is a common technique for dealing with censored i
4 i
M data. While not always appropriate, it should not be characterized as inappropriate in all circumstances.
I G-7 Section 18 presents a comparison of NUREG-1150 results with IPE results as a whole. A more interesting comparison would be between the individual NUREG-1150 results and the corresponding IPEs. This would provide more detailed information on specific modeling issues.
l.
2 4
9 9
f l
_ _ _ _ _ _.. __ _ _ _ ~__ _
~e Draft NUREG-1560 Comments Directly Relating to NMP1 1
NMP1-1 (Page 3-13) "New pump seal materials have been implemented at Nine Mile Point 1... " should read "New rgactor recirculatiQn Pump seal..."
NMPI-2 (Page 1012,10-13, Il-16) The report indicates that NMP1 feedwater coolant j<
injection function is emergency diesel backed. NMP1 is not designed to use diesel generators to power feedwater pumps. NMP1 CRD can be powered l
from EDGs. Combined with the emergency condensers, this provides NMP1 l
with a reliable system functiwally similar to HPCI at other plants. In addition, j
NMP1 has a dedicated hydro pot er plant available along with procedures to align it as necessary.
NMPI-3 (Page Il-19) The NUREG indicates that no details are available for NMP1 l
LOSP DHR sequences. LOSP DHR sequences are similar to other DHR i
sequences except LOSP is the initiator. While low frequency, the dominant LOSP DHR sequences include failure to align heat removal systems, failure to align emergency condenser make-up, failure of an EDG, and failure of an emergency service water pump. The equipment failures result in a loss of 2
j service water which in tum fails RBCLC, instrument air (which fails containment venting), and shutdown cooling.
NMPl-4 (Page 12-16) The statement relating to footnote 12.4 is based on an incorrect interpretation of the NMP1 IPE. The treatment of the timing of shell melt i
through at NMP1 is similar to that of other plants. The footnote indicates that NMP1 is an exception in this regard and it is not. The incorrect interpretation appears to result from a different definition of early containment f tilure. NMPC defines an early failure as one that occurs within 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br /> of the initiator. Thus, the timing for Level I, II, and III considerations are based on the same reference point. NMPC believes this definition provides the best input to issues relating to accident management and emergency planning (i.e., timeframe for evacuation). The NUREG defines early containment failure as within a short timeframe around RPV failure. When RPV failure occurs later in a sequence, NMPC bins any associated shell failure correspondingly. Thus, such a late containment failure at NMP1 would actually be an early failure using the i
NUREG definition.
i "
NMPI-5 (Page 13-18,13-19) The range of SLC initiation HEPs attributed to NMP1 is incorrect. The NUREG indicates that NMP1 SLC initiation failure probability ranges from 2E-3 per demand to IE-4 per demand. The range described in the NMP1 IPE is from 0.4 per demand to IE-4 per demand. In addition, the NUREG indicates that the IE-4 value is an exception and no obvious seasons for 1
such a low value could be found (other BWR 1/2 licensees have SLC initiation in the 0.1 to 0.01 range). The IE-4 value applies in specific ATWS sequences where the operator has previously successfully prevented MSIV closure and lowered vessel level to control power. These actions give the operators an expanded time window for SLC injection. Per our operator action dependency approach, if the operator is not successful in the preceding actions, the IE-4 v-m l
b
.1 value is not used. By considering the need for prerequisite successful operator actions in the model and the long timeframe available, the NUREG authors should find a measure of comfort in the IE-4 value. If the operator is unsuccessful at maintaining the MSIVs open, the reduced timeframe and operator dependency issues lead to a 0.4 per demand HEP for SLC initiation. Other values (e.g.,2E-3) are used under a separate top event which considers the mitigative effects of SLC injection later in the scenario. In terms of comparison with regard to the IE-4 value, it is likely that the corresponding case (i.e., MSIVs open) is binned to success or a longer term sequence by other licensees. In terms of comparison with regard to the 0.4 value, while higher than other older BWRs, the value is in the range of BWRs as a whole.
l t
a i
s i
i i
1
i
)
Draft NUREG-1560 Comments Directly Relating to NMP2 NMP2-1 (Page 4-20) The NUREG claims that nearly all ti; venting probability is associated with drywell venting and presumes that this relates primarily to containment flooding. This is true only in the level II portion of the model. The NMP2 model includes all containment venting modes. Containment pressure control is the primary mode relative to level I logic. For level II, flood related venting is more noteworthy because either the sequence is successful in the level I model or equipment failures occur such that the pressure control mode is of little benefit in level II. Combustible gas venting is used for a very small number of sequences. Thus, the statement on Page 4-20 should be revised to make it clear that drywell venting is only predominate when focusing solely on the level 11 portion of the NMP2 IPE model.
NMP2-2 (Page 9-34) The NUREG makes reference to a hardened containment vent to be installed in 1993. NMP2 has had a hardened vent since original design. The modification discussed in the IPE dealt with making the system more reliable. It was subsequently determined that the modification was not necessary.
NMP2-3 (Page Il-41) The NUREG refers to RCIC failure due to failure to over-ride a steam-tunnel temperature trip and indicates that only River Bend and Perry identify this failure mode. NMP2 includes this along with other RCIC isolation signals in Top Event OA. This top event requires operators to over-ride RCIC isolation within 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> of the initiator. In the NMP2 linked event tree approach, failure modes related to a given system are not always directly modeled in the system's " main" top event. In this case there are several top events that relate to the RCIC system.
NMP2-4 (Page 11-49) "RCIC is assumed tofail because ofhigh temperature trips since no DHR is available" should read "RCIC is assumed tofail because of high temperature trips since no room cooling is available" i
i
+
i l
l l
I 4
]