ML20248C966
| ML20248C966 | |
| Person / Time | |
|---|---|
| Site: | Oyster Creek |
| Issue date: | 05/21/1998 |
| From: | Roche M GENERAL PUBLIC UTILITIES CORP. |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| Shared Package | |
| ML20248C969 | List: |
| References | |
| REF-GTECI-A-46, REF-GTECI-SC, TASK-A-46, TASK-OR 1940-98-20188, GL-88-20, GL-89-22, TAC-M83652, NUDOCS 9806020287 | |
| Download: ML20248C966 (22) | |
Text
\\
.L
{
GPU Nucleer. Inc.
U.S. Route #9 South
(
Post Office Box 388 NUCLEAR Forked River, NJ 087310388 Tel 609-9714000 (609) 971-4814 May 21, 1998 1940-98-20188 U.S. Nuclear Regulatory Commission Attention: Document Control Desk Wr.shington, DC 20555 Gentlemen:
Subject:
Oyster Creek Nuclear Generating Station (OCNGS)
- Docket No. 50-219 Facility Operating License No. DPR-16 Individual Plant Examination for External Events -
Response to Request for Additional Information NRC letter dated December 10,1997 (6730-97-3287) requested additional information regarding the OCNGS Individual Plant Examination for External Events (IPEEE) submitted to the NRC for review on December 29,1995 in response to NRC Generic Letter 88-20, Supplement 4.
The attachment provides an itemized response to each of the NRC questions. If any additional information is required, please contact Mr. David J. Distel, GPU Nuclear Safety & Licensing, at (973) 316-7955.
'ncerely, e_
9806020287 990521 y Michael B. Roche i
f AN Oyy9 Vice President and Director Oyster Creek
./DJD
Attachment:
- 1. Itemized Response to NRC Request for Additional Information.
- 2. Fragility Analysis of Combustion Turbine Fuel Oil Tank.
- 3. Assessment of Potential for Liquefaction and Pennanent Ground Displacements at Designated Facilities.
- 4. Recirculation Pump Suport Structure Drawings.
,vv2q-
- 5. Seismic Fragilities of Civil Structures at Oyster Creek Nuclear Generating o,
Station.
L
t k
- 7.. Information Notice 93-$3 Evaluation, b
cc:.' Administrator, Region I OCNGS Senior Project Manager
' OCNGS Senior Resident Inspector l
c 3
1940-98-20188
- kg -
Page 1 o f 21 ATTACHMENT I GPU NUCLEAR RESPONSE TO THE REQUEST FOR ADDITIONAL INFORM ATION REGARDING OYSTER CREEK NUCLEAR CENERATING STATION IPEEE SUBMITTAL (TAC NO. M83652)
SEISMIC 1.
NRC OUESTION Provide a discussion of how operator actions were incorporated into the analysis, including a table of the most important actions identified Also, describe how the human error probabilitws were derived and the method used and the basisfor quantification.
RESPONSE
The quantification methods and values of the human actions used in the Seismic PRA are the same as those in the Level i OCPRA. The six volume OCPRA Level I report was provided as pan of the IPE Submittal in December 1991. The Human Action Analysis, Section 6, includes a description of the quantification methods utilized in the Level 1 OCPRA.
The base OCPR.t model was adjusted to incorporate the impacts of seismic events (i.e., fragility analysis). Other modifications to the OCPRA model for the seismic analysis include the removal of containment heat removal recoveries, loss of offsite power grid recovery and balance of plant systems (with the exception of ofTsite power systems). Due to the nature of seismic events, long term containment heat removal recovery and recovery of ofTsite power via the grid, were not considered probable and were set to guaranteed failure. In addition, preliminary fragility analysis indicated that many balance of plant systems were not seismically rugged and were removed from the model (i.e., guaranteed failure). This treatment is considered conservative since in very low acceleration seismic events many balance of plant systems have a reasonable likelihood of success. The above modifications to the base OCPRA result in the removal of many of the human actions modeled in the base OCPRA.
As part of potential system interaction walkdowns, access to areas of the plant which require lecal manual actions following a seismic event were reviewed. In no case was operator access to perform local manual action impeded by failure of systems, structures or components following a seismic event. It should be noted that the failure (i.e., fragility) of block walls and other structures is defined as the onset of cracking, not structural co!bpse.
Due to the importance of station blackout at Oyster Creek, particular attention was given to the controls for the near site combustion turbines. The controls for the combustion turbines are located in the 4160 VAC A and B Switchgear Room in the southwest corner of the turtJne building. Normal access to the 4160 VAC A and B Switchgear Room is via the turbine building truck loadmg bay. Additional access via the 4160 VAC D Switchgear Room is also available as well as access via the 4160 VAC C Switchgear Room or 4160 VAC D Switchgear Room via the
)
rollup doors. Multiple pathways exist through the turt ine building as well as access from the truck loading bay on the southwest side of the turbine building.
Given that operator access to plant areas where local manual action was required was not restricted due to the seismic event, operator action values from the Level i OCPRA were considered appropriate for use. In addition, most of the human actions modeled in the base OCPRA, and therefore the Seismic PRA, are based on guidance contained in the symptom base emergency operating procedures (EOPs). The EOPs are the primary guidance used to respond to emergency conditions and are the guidance which would be used following a seismic event.
IPEEERAI DOC -
04/10/98 I
i 1940-98-20188 k;'
Page 2 o f 21 Table 1 -Top Ten lluman Actions for the Seismic PRA (Sarted by risk achievement worth)
I Human Action Description
- Mi.A -
c Error -
Fussel-4
' - Risk ;
' Risk
. Idiatiner Rate ~
Vesely ?
' Achievenient L Reduction.
Operator vents containment on primary llAOVi I?.E-03 2.22E-03 2.9280 0.9978 pressure control using hardened vent.
l Operator manually starts containment ilACCS 1.3 E-02 1.55E-02 2.1801 0.9844 spray /ESW - drywell spray.
Operator manuelly starts containtnent ilACC3 5.0E-03 4.33E-03 1.8611 0.9956 spray /ESW - Torus Cooling following IC makeup failure.
Operator manually depressurizes reactor llAAD4 9.0E-04 7.35E-04 1.8156 0.9993 l
on lowering water level, IC Makeup Failure.
Operator manually depressurizes reactor ilAAD3 9.0E-04 5.76E-04 1.6394 0.9994 on lowering water level, stuck open EMRV.
Isolation Condenser Makeup from fire ilAMU2 4.0E-03 1.92E-03 1.4778 0.9981 protection, condensate transfer failed.
Manual actuation of core spray following IIACS4 8.0E-03 3.20E-03 1.3964 0.9968 logic failure.
Operators align Sombustion turbine LOPCTS 1.0E-01 1.23E-02 1.1115 0.9868 within 60 minutes.
Operator opens CRD test bypass line.
IIACDI 5.0E-03 3.78E-04 1.0752 0.9996 Isolation Condenser makeup via llAMUI 4.0E-03 2.98E-04 1.0743 0.9997 condensate transfer.
Operator actuates containment spray in liACC4 7.0E-03 1.33E-04 1.0189 0.9999 torus cooling following IC failure.
Operator restores power to one vital AC llARE2 2.7E-02
.i.75 E-04 1.0131 0.9993 bus following station blackout and successful combustion turbine alignment.
Operators align combustion turbines LCT30M 4.0E-01 9.49E-03 1.0131 0.9900 within 30 minutes.
Operators bypass MSIV closure and ilAOL3
Operators initiate standby lig"U control llAI3I3 1.6E-02 4.22E-05 1.0026 1.0000 following an.ATW'.;.
IPLEERAI DOC 04/10N8
I l
1940-98-20188 l
[gpu '
Page 3 o f 21
.-m Table 1, provides the top 15 important human actions in the Seismic PRA. The human actions reported on this table are sorted by risk achievement worth. In addition to the human action description: the error probability, fraction importance, risk achievement and risk reduction are also provided.
It should be noted that the OCPRA and Seizmic PRA human actions are developed considering various conditions or scenario specific considerations. For example, several containment spray initiation actions (IIACCS, HACC4 and IIAr.C3) appear in the top fif teen human actions. These actions are related to various incoming con 6tions, such as previously successful or failed systems, which can affect the performance shaping ' actors for the action. The methodology employed in the development and quantification of the O?PRA and the Seismic PRA ensures that only one of the human action.s related to a system or function is applicable to a given scenario. In the case of the containment spray example, only one of the actions would be applicable to a given scenario. See Section 6 of the OCPRA for further details.
2.
NRC OUESTION Provide a discussion of offsite power recovery via the combustion turbines modeled in the analysis, including:
operator actions required and needed instrumentation fragility analysis calculationsfor the Combustion Turbine Fuel Oil Tank, as well as soil-related structural analysis.
routing of combustion turbine power (if routed via the 4kV emergency switchgear, additional discussion of how the switchgearfragility was incorporated into the modelfor off-site power recovery is needed).
RESPONSE
Loss of offsite power recovery is modeled in the Level i OCPRA. In the case of the seismic model, loss of offsite power recovery is modeled from the near-site combustion turbines located on the Forked River site. Power from either of the two combustion turbines is provided via an underground cable to the on site station blackout transformer. Controls and instrumentation for the combustion turbine are k>cated in the 4160 VAC A and B Switchgear Room in the southwest corner of the turbine building. Normal access to the 4160 VAC A and B Switchgear Room is via the turbine building truck loading bay. Additional access via the 4160 VAC D Switchgear Room is also available as well as access to the 4160 VAC C Switchgear Room or 4160 VAC D Switchgear Room via the rollup doors.
The combustion turbines are started by the load dispatcher upon the request of Oyster Creek via microwave relay. Should the lead dispatcher be unavailable local action at the combustion turbines located on the Forked River Site is required to start the turbines. Starting of the combustion turbines is modeled in the Seismic PRA. Failure to start the combustion turbines within 30 minutes is assigned a failure probability of 0.4 (basic event: LCT30M) and failure within one hour is assigned a failure probability of 0.1 (basic event: LOPCTS).
Fragility analysis of the combustion turbine fuel oil tank is provided as Attachment 2. The
" Assessments of Potential for Liquefaction and Permanent Ground Displacements at Designated Facilities" is provided in Attachment 3.
Power from the combustion turbines is routed to the 4160 VAC B Switchgear. Power from the combustion turbines can be used to power the "B" AC electrical loads including non-essential loads such as a feedwater train and condensate transfer as well as the safety related loads associated with the 4160 VAC D Switchgear. (The capability of the combustion turbines to supply the non-safety loads on the 4160 VAC B Switchgear is not modeled in the Level 1 OCPRA and the Seismic PRA.)
IPEEERA DOC 04/W98
1940-98-20188 Page 4 o f 21 Th current fragili y of the top event LX, Seist'uc Failure of the Combustion Turbines, represents t
the loss of offsite, swer recovery and should contain fragilities associated with the combustion turbine fuel oil tank, associated SBO transformer and the combustion turbines themselves.
Inadvertently omitted from the Oyster Creek Seismic PRA was the fragility associated with the com!>ustion turbine fuel oil tank. In addition, corabustion turbine power is routed via the SBO transformer through the 4160 VAC B Switchgear and the 4160 VAC D Switchgear. The routing of power through 4160 VAC Switchgear B and D was not reflected in the Seismic PRA. A sensitivity study which models the impact of the combustion fuel oil tank and the requirement for 4160 VAC B and D Switchgear is developed.
Table 2 Combustion Turbine Seismic Split Fraction Values (with Combustion Turbine Fuel Oil Tank)
Combustion Turbine Power Seismic PRA Sensitivity (Top Event LX)
Value Case Value LX1 - SEIS1. g Levels 0.007 to 0.26 4.66x 10" 7.16x 10" LX2 - SEIS2. g Levels 0.26 to 0.46 2.16x 10" 2.92x 10" LX3 - SEIS3. g Levels 0.46 to 0.62 1.20x10" 7.92x 10" LX4 - SEIS4. g Levels 0.62 to 0.815 9.12x 10" 9.60x 10" l
la the current Seismic PRA, the 4160 VAC B Switchgear is not questioned following the independent or seismic failure of either offsite power or the failure of the startup transformer. That j
is, if power is not available at the switchgear then independent failure is not questioned. To ensure that all cases contain the appropriate failure causes, the independent failure of 4160 VAC B Switchgear is added to top event LX, thereby requiring the success of the 4160 VAC B for success of combustion turbine power.
The above term prevents the assignment of offsite power recovery where the loss of power is the result of the independent failure of the 4160 VAC B Switchgear.
The independent failure of 4160 VAC D Switchgear is questioned at top event ED. Failure of power from the 4160 VAC D Switchgear is due to either the independent failure of the switchgear or failure of power supply from the "B" switchgear and the diesel generators. In this situation, a failure at top event ED could be the result of the independent failure of the switchgear.
To appropriately reflect the dependence of the combustion turbine power recovery on the loss of the 4160 VAC Switchgear a new top event is added to the event tree. This top event represents the j
failure of the diesel generator number 2 following a loss of power from the 4160 VAC B Switchgear. Top event ED is adjusted to remove the failure of the diesel generator and now represents the independent failure of the 4160 VAC D Switchgear. The split fraction assignment rules are adjusted to reflect the event tree changes as described in Table 3.
This sensitivity case produces a core damage frequency of 3.62x10+ per year versus 3.61x10+ per year for the base case. This represents an insignificant increase in the total core damage frequency due to seismic events. This result is expected due to the following:
Combustion Turbine Fuel Oil Tank fragility is commensurate with the other fragilities associated with providing combustion turbine power. Therefore, the effect on the probability of failure of combustion turbine power is small (see Table 2). The large change is in the lowest seismic acceleration initiating event (SEISI) which contributes very little to the total core damage frequency.
3 1
l H' lit 1R Al IXX' 04/1ON8
1940-98-20188 4*
Page 5 o f 21 The relatively low independent failure of the 4160 VAC Switchgear B and D when compared with the failure of the diesel generators and the diesel generator building.
The low relative contributions of the independent failure of the electrical power system (including the dt.:sel generators and combustion turbines) when compared with the seismic failure contributions.
Table 3-Sensitivity Case Split Fraction Assignment Rules Split Split Fraction Assignment Rule Assignment Rule Description Fraction EIS EB=S Diesel Generators not required since offsite power available.
E12 DW=S'EC=S+ DC= F DG2 without common cause since DGl is successful (EC=S) or not questioned (DC=F).
EID D W=S
- DB = S
- DC=S' EC= F
- E B = F DG2 with common cause failure following failure of DG l.
E!F 1
Default set to Guaranteed Failure EDI DB =S* EB-S Failure of 4160 VAC ID (swgr only), normal all support available (DB=S*EB=S).
ED2 El=S*DB=S*EB=F Failure of 4160 VAC ID (swgr only), when offsite power via 4160 VAC IB is unavailable (EB=F) and when bus IC is successful or failed due to loss of 125 VDC control power (includes breaker transfer failures from 4160 VAC 1B to EDG.)
EDF i
1)efault Set to Guaranteed Failure LYS (EC=S + ED=S)+ S BO'lC=S* M U=S Recovery of offsite power not required since either
- VR=S*SR=S 4160 VAC Bus successful or isolation cond:nsers successful.
LYl (ED=F*E/=F) *LX=S'SBO*1C=S' Failure of recovery of offsite power (30 min) given (VR= F+ SR= F) that 4160 VAC buses B and D not independently failed with offsite power and diesels failed.
LY2 (ED= F*El=f) *LX =S'SBO
- 1C= S
- Failure of recovery of offsite power (1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />) given (M U= F+ FP= F) that 4160 VAC buses B and D not independently failed with offsite power and diesels failed.
LYF I
Default Set to Guaranteed Failure 3.
NRC OUESTION Provide a discussion of the pracedure usedfor the evaluation ofLOCA pathways in the analysis and drawings ofthe recirculation pump support structures.
BESPONSE l
All applicable piping systems were walked down on an area basis to evaluate the potential for a i
l seismically induced LOCA and to evaluate the impact of any potential leakage or rupture. The walkdown was conducted in accordance with EPRI NP-6041 SL, "A Methodology for Assessment of Nuclear Power Plant Seismic Margin," Revision 1, and SQUG " Generic implementation Procedures (GIP) for Seismic Verification of Nuclear Plant Equipment," Revision 2. Two (2) seismic capability engineers and the engineer responsible for the development of the seismic PRA risk model performed the walkdown. A summary of the results of the walkdown is provided in EQE report " Development of Equipment Seismic Fragilities for OCNGS IPEEE" issued July 1996. A copy of this report is provided in response to Question 5.
l IPE01:RAI DOC 04/10/98 i
l I
L-
1940-98-20188 fq',
- Page 6 o f 21 l
A walkdown was performed by a Seismic Review Team (SRT) to assess the potential for seismic interaction concerns and to determine components and subsystems (including distribution systems such as piping) with I'3PFs below the specified value. The SRT walked through all areas containing components and subsystems within the scope of the seismic PRA models. The SRT inspected components and subsystems soAiently to obtain confidence that all potential failure-modes and seismic interaction issues are addressed. In particular, the issues identified on page 5-15 to page 5-17 of EPRI NP-6041-SL were considered during the walkdown of piping, valves and other components that farm part of the pressure boundary. As stated in EPRI NP-6041 SL, piping systems typically have a llCLPF greater than 0.5g unless conditions similar to those described on pages 5-15 to 5-17 exist.
During the walkdown, the SRT identifies conditions that might impact the ability of the piping systems to main'.ain pressure boundary. For example, as stated in EPRI NP-6041-SL, only primary stresses are considered in the faulted condition. Seismic anchor motions are not included with the l
SSE loads. Ilowever, seismic anchor motions have been found during actual seismic events to l
cause damage to piping systems. Therefore, such anchor motions are included in the evaluation l
and assessment of piping systems if they are judged during the walkdown to be significant, in addition, the SRT looks for support conditions that may have a brittle failure mode that would prevent the supports from deflecting sufficiently to relieve thermal expansion.
EPRI NP-6041-SL provides a procedure and criteria for assessing seismic adequacy of structures, supports and components including the potential for pipe leak or rupture. Th'.s methodology was followed in the evaluation of LOCA pathways in performing the IPEEE at Oyster Creek as described in EQE report " Development of Equipment Seismic Fragilities for OCNGS IPEEE",
Drawings of the recirculation pump support structures are provided in Attachment 4.
4.
NRC OUESTION l'rovide the cutoffthreshold values offragility curves; and the results ofsensitivity studies if any, on Ihe e))ects ofthe cutoffoflower tails qffiagilin curves.
RESPONSE
The EQE International report, " Development of Equipment Seismic Fragilities for OCNGS IPEEE,"( Attachment 6) provides a discussion of the threshold fragility value. The following quote is from page 1-4 of that reference.
"The overall impression gained from the walkdowns and review of A-46 Screening Evaluation Work Sheets (SEWS) packages is that the plant is adequately constructed for this low seismicity site. The A 46 packages demonstrate a sufficient degree of conservatism in almost all cases to satisfy the screening criteria. Where this degree of conservatism does not exist, detailed estimates of median capacities are developed for the seismic PRA."
The screening threshold fragility descriptions are:
Au Sa Se llCLPF Equipment structural integrity 1.0g 0.40 0.32 0.30g IP111RAI DOC 04/10/98
1940-98-20188 l
Pa p p
- Page 7 o f 21
.aca l
The threshold fragility, when convolved with the mean seismic hazard curves results in an annual pro sability of failure of an individual component of approximately 8.8E-6. A specific fragility has been completed for equipment not satisfying this threshold fragility. The results of the equipment fragility are provided in Section 5.
This threshold fragility value was, in part, chosen based on the results of exercise of the Level 1 OCPRA risk model. The exercise estimated the impact of the guaranteed failure of the balance of plant systems on the total core damage frequency. These systems included the feedwater, condensate, TBCCW, circulating water, service water and instrument air systems. Since it was known that the balance of plant systems were unlikely to survive higher acceleration seismic events, this sensitivity case was chosen to best represent a risk model to base the threshold fragility.
4 The resultant core damage frequency was 8x10 per year. The thresheld fragility value was chosen, based on experience, as 10% of this core damage frequency. No other sensitivity cases were perfonned.
5.
NRC OUESTION Provide the references 3-2, 3-3 and 3-8 which contain the results ofthefragility analyses referred to in the IPEEE.
RESPONSE
The reference 3-2 is provided as requested as Attachment 5. References 3-3 and 3-8 were used ii.
the development ofinitial models however, these references were superceded by later revisions of J
reference 3-8. The latest revision of reference 3-8 is provided, in place of referraces 3-3 and 3-8
{
i requested (see Attachment 6). This revised reference reflects the fragilities used in the Oyster Creek Seismic PRA.
3-2 EQE International, " Seismic Fragilities of Civil Structures at Oyster Cneck Nuclear Generating S:ation in Suppon of the IPEEE Program", Report Number 50124-R-003, July 1994.
3-8 (revised) EQE International, " Development of Equipment Seismic Fragilities for OCNGS IPEEE," July 1996.
l FIRE I.
NRC OUESTION in the detaded evaluation, use of afire severstyfactor was appliedforfive of the eight :ones that had not been screened up to this point in the analysis. Theformulation and application of thefire severityfactor is considered to have technicalflaws. Theformulation ofthefactor is based onfire events that have been recorJed in the EPRI Fire Events Database. Consideration of the extent of automatic or manualfire suppression on mitigation of these events was not addressed in the i
formulation of the severin factor. (11 is anticipated that these t) pes < suppression uwe employed in some of the events, thereby limiting severity.) Because thefrequemy associated with thefire can not be tctaly independentfrom thefire severityfactor, use of thefactor artificially decreases thefirefrequency. II' hen thefire severityfactor was applied in an area wherefire suppression was credacd, the fire severityfactor was only applied to the scenario uhere fire suppression was unavailable; this is inconsistent with theformulation of thefactor. The engulfing fire assumed when using thefire severityfactor is not always the limiting case; ie., a smallerfire of higher frequemy couldpose as much risk ofnot more risk. The use offire severityfactor is considered technically unsubstantiated, therefore, assessment offire damage is warranted for thefive :ones mentioncJ above, in which fire severity factars were used, please modelfire suppression and propagation to determine the probabdity that thefire will damage critical targets before it is suppressed, andprovide the results ofthe analysis.
IPIIERAI tXX' 04/10/98 I
I i
l 1940-98-20188
[pp Page 8 o f 21 o.
l l
RESPONSE
1 l
There are many issues associated with the request for additional inforr. tion in this question. The issues seem to be associated with the concerns about the use of " fire severity factor" There is l
some confusion with regards to this term which rr.ay be better represented by the term " Initiating Event Fire Severity Factor". Consider, the overall fire frequency provided in Table 4.1-11. This 4
overall plant fire frequency is 4.02x10 per year. The indication of this frequency is that approximately once every 2.5 years a fire event occurs which if not automatically or manually suppressed would result in an "all engulfing fire". Over the approximately 30 years of operation at Oyster Creek this would equate to 12 fire events which if nat suppressed would have resulted in "all engulfing fires". This is not consistent with experience at Oyster Creek. In overview, the fire seserity factor is attempt to adjust the fire event initiating event frequency to bener represent Oyster Creek experience without eliminating the potential for the "all engulfing fire" scenarios while preserving the frequency of those small fire events in the database.
The EPRI Fire Events Database was reviewed to more adequately address those items which were not "all engulfing fire" potential events. Items included in the database range from smoking relays or electrical equipment which were de-energized (no fire) to a car fire which occurred in the station parking lot. Since the EPRI Fire Events Database was not complete, i.e., not all fields were entered for each fire event, specific fields which provide a representation of the severity of the fire were reviewed. Two fields of the database, when taken together, provided a reasonable representation of the efTect of the fire events. These fields were the "COMEFFECT" and "DIRECTLOSS". The suppression database field was not complete and was not considered as part of the evaluation of the initiating event severity. Following a review of these database fields,,i was determined that less than 1% of the fire events in the database had significant equipment tamage (greater than 3 components damaged).
The intent of the application of the fire sever ty factor was to better represent the frequency of fire events which could, in the absence of fire suppression, result in an all engulfing fire event while preserving the frequency of the smaller fire events. la all cases, the fire initiating event frequency is preserved. That is, in no case is the fire initiating event frequency reduced using the fire severity factor. Where a fire severity factor is applied within a fire zone evaluation, the all engulfing fire frequency is reduced by the fire severity factor while the less severe fire events are addressed with the remaining frequency. Consider the example of a fire severity factor of 0.01 is applied in a given fire zone. The frequency of the all engulfing fire is represented as 0.01 times the initial frequency provided on Table 4.1-11. The remaining 0.99 of the fire frequency on Table 4.1-11 is assigned to the small (but still bounding) fire. The impacts of the smaller fire are based on fire rone configuration and most significant impacts on the OCPRA.
The fire severity factor is only applied in fire zones which have fire detection and automatic suppression. In this way, the concern that some fire events in the database may have been extinguished by manual or automatic means was addressed. Since the availability of fire automatic suppression systems generally varies between 0.02 and 0.05, the magnitudes of severity factor is not in question. The fire severity factor reflects the fraction of the Five Events Database which contains events which are either not fires (smcking electrical components) or do not affect plant equipment (car fire in station parking lot).
A fire severity factor was applied in five fire zones. These are the Cable Spreading Room. A and B Battery Room, Turbine Building Basement, Reactor Building 23 and 51 foot Elevatio s. The combination of fire severity factor and automatic suppression was not initially intended but did occur in three of these five fire zones.
These three fire zones used severity factor in favor of additional analysis which was judged would provide similar results, in the case of the Cable Spreading Room (OB-FZ-04), fire severity factor is used in favor of credit for manual suppression and the use of the remote shutdown panel located in the 480 VAC switchgear room. This analysis simplification may have been a misuse of the severity factor. Ilowever, credit for manual suppression and use of the remote shutdown panel would most likely result in similar if not lower core damage frequency values. The fire severity IPElIIMI IXX' 04/10Mt
f l
1940-98-20188
[Q,,
Page 9 o f 21 I
factor used in the Cable Spreading Room case is 0.1. Similar potential misuses of the severity factor occurred in the OB-FZ-08A/B and TB-FZ-ilD fire zones. Ilowever, credit for manual suppression and use of the remote shutdown panel is judged to result in similar core damage frequency values.
in conclusion, the use of severity factor to address weaknesses in the fire events database is not i
considered technically flawed. Careful application of the severity factor and preservation of initiating event frequency (to capture smaller fire events potentially risk significant with higher frequency) was the original intent. For the areas in which fire severity factor was applied in l
conjunction with automatic fire suppression, additic.nal sensitivity studies and analysis should be l
performed to demonstrate that overall conclusions regarding the core damage frequency is not significantly impacted.
GPU Nuclear will perform a re-evaluation of all five fire zones which utilized fire severity factor and provide the results of the analysis to the NRC no later than October 1998.
2.
NRC OUESTION In the detailed evaluation.for seven of the eight remaining :ones, credit was givenfor automatic fire suppression systems in mitigating thefire after some components are assumed to be damaged it is not apparent that this method addresses all thefactors that affect the likelihood offadure to suppress thefire bcfore damage of more critical components has occurred. In order to assume that the suppression systems will effectively control or extinguish afire, the analysis must provide reasonable assurance that the systems are designed. installed and maintained in accordance with nationally recogni:ed standards and codes. l'ertf, that the automatic suppression systems were designed and installed to meet applicable NFPA codes.
1
RESPONSE
Seven of the eight remaining fire zones credit either severity factor or suppression. As illustrated on Table 4.6-3," Oyster Creek Plant fire Areas and Zones Quantitative Evaluation Results", only five of the eight fire zones credit fire suppression. Of the five fire zones which credit fire suppression, only three of these zones use a severity factor, Fire zones, which credit automatic suppression of fires, were reviewed in the performance of the fire analysis to determine if adequate fire suppressant coverage was available to provide a reasonable likelihood of successful fire suppression. In the cases where automatic fire suppression was modeled the major combustion sources were considered to have adequate fire suppression coverage to provide a reasonable likelihood of fire suppression success.
The Table 4 provides the r.pplicable NFPA codes to which the automatic suppression systems were designed and installed.
1 I
l i
)
WM IRAI IXX:
04/10/98 l
1
)
_______.)
1940-98-20188 fsper Page 10 o f 21 4
uw i
Table 4-Selected Fire Zone Fire Protection Systems Fire Area /
Description Type Applicable Comments 7mne Code OB-FZ-4 Cable Deluge NFPA 15 Designed and installed to protect Spreading Automatic concentrations of cable trays which are Room the principle fire hazard.
OB-FZ-5 Control Room 11aloa NFPA12A Designed and installed to protect cabinets Automatic IF-4F and 1R-6R, SF-9F and 7R 11R, Local 10F and 11F. See note 1.
Application OB-FZ-6A "A" 480 VAC llalon NFPA 12A 7,ne wide suppression.
Switchgear Automatic Room Total Flooding OB-FZ-8C A & B Battery llalon NFPA 12A Zone wide suppression.
Rm, Tunnel, and Automatic l
Elect. Tray Rm Total Flooding TB-FZ-I I D Turbine Wet Pipe NFPA 13 Zone wide suppression.
Building Sprinkler Basement Automatic l
Closed llead NFPA 15 Ilydrogen seal oil unit. NFPA 15 used for Spray (guidance) guidance.
Automatic RB-FZ-1 D Reactor Deluge NFPA 15 Designed and installed to protect Building-51' el Automatic concentrations of cable trays which are the principle fire hazard.
Deluge See note 2 Water Curtain suppression at ceiling over Automatic southeast equipment hatch.
Deluge See note 2 Spray nozzles protect northwest stairwell.
Autm.
RB-FZ-l E Reactor Deluge NFPA 15 Designed and installed to protect Building - 23' el Automatic concentrations of cable trays which are the principle fire hazard.
Deluge See note 2 Water Curtain suppression at ceiling over Automatic southeast equipment hatch.
Note 1: Local application of safety related cabinets does not maintain llalon concentration of 5% for 10 minutes as the panels are not totally enclosed. Intent of the system (s) is to provide a suppressant to allow control room operators to bring manual suppression to bear.
Note 2: These systems use the NFPA codes for guidance. They are unique designs installed to comply with Appendix R commitments to prevent spread of fires across fire zone boundaries. Each one of these designs is an extension of the cable tray deluge systems which are designed and installed to NFPA codes.
3, NRC OUESTION lluman recovery actions are credited in the estimates of upper bound CDF where conservative assumptions were related No details are provided concerning the methodology employed to calculate the likelihood that the recovery action is unsuccessful. There are issues unique tofire situations that relate to psychological and environmental stressors (impact of smoke and suppression agents, reduced visibility, etcJ. No indication is supplied in the Oyster Creek IPEEE submittal as to whether thesefactors were considered Please provide the details as to how the human errorprobabilities were estimated IPEEIRA!IXX' '
04/10/98
l 1940-98-20188 kj Page 11 o f 21 I
RESPONSE
in the " Revised Estimate of Upper Bound Core Damage Frequency" (Section 4.6.2), some of the more conservative modeling assumptions associated with the initial fire evaluation of the OCPRA plant model are relaxed. Some of these assumptions are associated with non-physical conditions which can occur when using the OCPRA to model fire events. Other assumptions are the result of l
operator action to perform actions directed by the Resprmse to Fire Procedure (2000-ABN-3200.29). The following is a summary of the fire zones and the associated assumptions-1 OH-FZr10A (Monitoring and Change Room - 46 Foot Elevation). The impact on the 4160 VAC Switchgear ID was due to an assumed failure of the ventilation supply and exhaust fans for the "B" 460V switchgear room. As stated in the Response to Fire procedure, control of these fans remains available from the associated remote shutdown panel. Operation of this pai.el is directed by plant operating procedure 346 Normal operation of the ventilation system is directed by procedure 331. Several hours were determined to be available before ventilation of the area ia required. The operator action of the recovery of ventilation was not explicitly numerically evaluated, but the guaranteed failure of the 4160 VAC lD Bus was changed to a 0.066 failure rate, as discussed below.
OH-FZ-22A (New Cable Spreading Room / Mechanical Equipment Room - 63' 3"). The saw recovery described above for OB-FZ-10A (recovery of ventilation for the "B" 460 VAC Switchgear Room)is applied in this fire zone. This recovery action is described below.
TH-FLilH (Lube Oil Storage, Pumping and Purification Area). The original assumption was I
the guaranteed failure of 4160 VAC Bus IC. This was caused by possible damage to control circuits for the associated diesel generator and loss of control room operation of supply breakers for the 480 VAC switchgear. As noted in the revised evaluation, this only impacts circuit breaker control from the control room. The protective capability of these c!rcuit breakers remains unaffected, such that circuit protection (i.e. ability to trip on individual circui* faults) remains available. No operator recovery for these scenarios was evaluated, since the primary circuit i
protection capability remains available and operator action would only be required following circuit breaker failure to isolate an independent circuit failure.
TB-FZ-IIC (Switchgear Room, West end of Turbine Buil%g Mezzanine Level). The original fire evaluation modeled the guaranteed failure of 125 VI J Jus C. This assumption was made based on an assumed hot short failure of the power supp' or isolation condenser *B" which is also failed as a result of the fire event. The assumption is relaxed to reflect the failure ofisolation condenser "B" only (i.e., removing the guaranteed failure condition for 125 VDC Bus C). As noted for TB-FZ-llB, above, primary circuit protection remains available, such that operation of fuses in any failed power circuits would prevent the failure of 125 VDC Bus C itself. No operator recovery was considered for the revised evaluation of this area.
TB-FZ-IIE (Main Condenser Bay Area). As in the case with the fire zone TB-FZ llB, the original assumption was the guaranteed failure of 4160 VAC Bus IC. This was assumed to be caused by possible damage to control circuits for the diesel generater and loss of control room operation of the supply breakers from the 480 VAC switchgear. While this could fail operation of the associated diesel generator (but not the normal supply to bus IC), the supply breaker protective functions remain available. Therefore, the revised evaluation considered DG failure for all fires in this area, but removed the guaranteed failure term for bus IC. No operator recovery was considered for the revised evaluation of this area.
I' II*III'.RAI tK1C 04/10/98 f
p L
J
1940-98-20188 l
spes Page 12 o f 21 TB-FZr11F (Main Feedwater Pump Area.). The original controlling assumption was the failure i
of 4160 VAC Bus ID due to an assumed hot short failure of the associated core spray pumps.
)
Since the core spray pumps are not initially expected to be required and are isolated (circuit l
breaker open) from dTision 2 electrical puner, this assumption is removed. In addition, the circuit i
breaker would pro ide protecticn to the bus (i.e. trip) should core spray system 2 be initiated during a shorted condition. No operator recovery actions were considered for the resised evaluation of this area.
TB-FZrilH (Basement Southeast End). The original controlling assumption was the failure of 4160 VAC Bus IC. A review of this fire zone confinned that the cables associated with 4160 VAC Bus 1C are located in a cable pu'l pit. The pit is separated from the rest of the fire zone by a concrete wall and a layer of pea gravel and sand below the pit cover plate. Therefore, this assumed failure was removed in the rnised evaluation. No operator recovery actions were considered for the rnised evaluation of this area.
RB-FZelF (Reactor Building-19 Foot Elevation). The original controlling assumption in this fire zone is the failure of all equipment in the four corner rooms as well as the torus area. Resiews of the fire zone indicate that spread of a fire event to all corner rooms is unlikely due to concrete walls associated with the corner rooms and low combustible loading. The fire scenazio of highest consequence (northwest corner room) was chosen as the representative fire event. No operator recovery actions were considered for the rnised evahiation of this area.
MT-FA-12 (Main Transformer and CST Area). The controlling assumption for this fire area was the initial modeling of general transient initiating event. This event was determined to most closely represented by a loss of ofTsite power event. No operator recovery actions were considered for the revised evaluation of this area.
CW-FA-14 (Circulating Water / Intake Arra). The controlling assumption was the failure of all equipment in the intake stmeture area. The failure of all ESW pumps results in the inability to use the intake as the ultimate heat sink. The afTect is that core damage is dominated by long term failure of the containment due to overpressurization. This scenario develops over a period of approximately 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />. The restoration of a single ESW pump within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> is therefore modeled.
Of the ten areas evaluated in the Resised Estimate of Upper Bound Core Damage Frequency, seven areas are resolved by modeling adjustments that represent phenomenological issues. The remaining three fire zones utilize operator recovery actions. The values used in the evaluation are screening values. Actions were reviewed for the Oc available to perform the action as well as ensuring that the action could be accomplished outside the zone of influence of the fire. Both recovery actions are proceduralized. In both cases, significant time is available for the actior.s to be performed. Two operator recovery actions were estimated: Restoration of 4160 VAC Room Ventilation and Recovery of an ESW or Circulating Water pump to prevent containment overpressurization.
For fire zones OB-FZ-10A and OB-FZ-22A, the resised evaluation includes the restoration of 4160 VAC room ventilation. A value of 0.066 was used for the rnised evaluation. This value is judged to be extremely conservative, given the length of time available to perform the action and the availability of normal and recovery procedures for restoration of room cooling. Also, this action is performed at the remote shutdown panel located in the B 460VAC Switchgear Room, such that the operator does not have to enter the area impacted by the fire to perform this action.
l l-l l
l
1940-98-20188 GPu,
Page 13 o f 21 These components are normally controlled and their status is indicated in the control room, such that the operator has indication as to their status on a continuing basis, inclucting the potential for failure due to fire-related impacts. As directed by plant operating procedure 346, the Group Shill Supervisor (GSS) will send an operator to the B 460VAC Switchgear Room. Tids operator uilt then take the Train B control switch to the " ALTERNATE" position. This action automatically trips, thcn (within approximately 2 seconds) restarts the B 460VAC Switchgcar Room and A/B Br.ttery Room Fans (SF1-21, EF1-21 and EF1-20). Plant operating procedure 346 then directs the operator to operate these components as required by the normal operating procedure for the system (331).
As noted in Table 12-3 of NUREG/CR-1273, an initial screening value for procedurally directed actions, without considering operator recovery, is 0.05. Due to the length of time available (estimated to be at least 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />), this value is judged to apply to the recovery of switchgear room ventilation. His lower failure rate (compared to 0.066) wasjudged to adequately beund the value used in the revised evaluation for these areas. Table 12-3 gives an initial screening error rate of 0.025 for this form of proceduralized action when recovery by the operator is considered, as would be the case for this action, u hich would be performed well aner the fire has been extinguished.
In the case of recovery of an ESW pump to prevent contaimnent overpressurization, the recovery action was judged to be similar to the currently analyzed recovery of containment spray following system failure (top event RC). This recovery action was incorporated into the reiised evaluation of CW-FA-14, which is an open area, with no intervening combustibles between the various components of concern. The discussion prorided in the IPEEE report details the separation between the control and power circuits for the various ESW pumps, such that a single credibic fire does not have the potential to d:miage all four pumps.
In the case where an engulfing fire did occur in this arca, repair procedure A100-APR-3531.01 directs the plant mainten:mcc staff to establish temporary power to ESW pump 1-3 (IC) from 4160VAC switchgcar ID and to recover from damage to ESW pump 1-4 control cable 63-326.
For pump 1-3, this procedure specifics the mcgger, continuity and pump rotation tests required to assure successful recovery using a pre-staged cable, which includes high-voltage cable tenninations that have been pre-installed specifically for this purpose.
Given this information, the currently analyzed recovery (i.e. removing the guaranteed failure condition for top event RC) was judged to adequately bound the recovery of at least one ESW pump following a fire in the intake structure. Based on the CDF change from the initial to the resised evaluation, this resulted in an efTective recovery factor of (5.57E-7 /1.41E-6 =) 0.395, or recovery from two in every 5 fires in this area. Due to the procedural guidance and length of time available (i.e. approximately 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />), this was judged adequate to provide a screening bound for operator failure to perform this action before loss of NPSH to the containment spray pumps.
IIFO 1.
NRC OUESTION l
1he IPEEE submittal estimates thep equency ofhigh winds with speeds exceedmg 168 mph as SE-l 7 peryear, and estimates the contribution of straight winds to thisfrequency as neghgtble (See l
Figure 7 ofthe submittal). However, the staff in a letter dated February 26,1990 (from Alexander l
Dromerick, Senior Project Afanager, NRR, to Mr. F. E Fit: patrick, Vice President and Director Oyster Creek Nuclear Generating Station) has estimated that the frequency of tornado wind speeds exceedmg 168 mph is 7E-6 per year. Even the frequency of straight winds with speeds greater than 168 mph may be greater than IE-6 per year. At Seabrook, in a memorandum from Intham P Gamill to Leon Reiter, dated Aug 16,19M (Public Document Accession No.
M08240398) it is noted that windspeeds of up to 150 mph may be more likely due to non-tornado phenomena. From afgure in this memorandum, one would estimate, at Seabrook, afrequency of about 2E-6 peryearpr the probability that straight wind would exceed 168 mph. One can obtam similar results at Oyster Creek site, by using the approach given by Batts, Cordes, Ruasel, Shaver
{
l
1940-98-20188 Pgpes Page 14 o f 21 l
==ca and Simiu in NBS Building Science Series 124, " Hurricane Wind Speeds in the United States",
May 1980. This report finds that a Weibull distribution fits the hurricane windspeeds. From Figure 6 of the report onefinds that at milepost 2450 (near Atlantic City, NJ), thefrequency of exceedance of125 mph is about SE-4 per year, for hurricane winds. A goodfit to the data in the p=exp(-u ), where u=(v+ 669p660,,.s windspeed, and y =11.
report is given by r
This yields about 1.2E-6 per year for hurricane wind speeds greater than 168 mph. One notes further that Changery, in NUREG/CR-2639, gives the 1000 year retu.n period straight wind as
!I7 mph. In contrast. Table 3 on page 5.1-22 of the IPEEE submittalgives a 100,000 year return 1
per:odjor a straight wind ofI02 mph, a much more optimistic value.
1 Since thefrequency of wind speeds greate~ than 168 mph can be reasonably estimated at about 1E-3 per year, one cannot screen out high winds on the basis <fwind speeds exceeding 168 mph being less than IE-6 per year. Ilowever, a reasonable approach would be to show that at speeds somewhat greater than 168 mph the probability of core damage is low. 7he structures of concern are the diesel generator vaults and the oil tank compartment. Therefore:
Estimate as afsmetion of wind speed, for winds speeds greater than 168 mph, the probability of failure of the diesel generator vaults and oil tank compartment: include the e))ects of both wind pressure and tornado missiles.
RESPONSE
GPU Nuclear concurs with the Staff on the under-estimation of the frequency of high winds with speeds in excess of 168 mph. In addition, GPU Nuclear agrees that the frequency of high wind 4
speeds of greater than 168 mph can be estimated in the 1x10 per year range.
In the " Evaluation of Diesel Generator Building Subjected to Tornado-Wind Generated Loading -
Oyster Creek (TAC. NO. 49392 and 49394)", February 26, 1990, the staff concluded that the critical walls of diesel generator vaults and the oil tank compartment can withstand the impact of the most damaging missile, (hard missile with large kinetic energy) i.e., utility pole, together with t'ie wind loadings associated with a 168 mph wind. The Pickard, Lowe and Garrick report, " Oyster Creek Tomado Missile Risk Analysis",(June 1983), indicates that "No scabbing damages occur in the TORMIS (tornado-missile) simulations on the 18-inch reinforced concrete walls. llence, the 18-inch walls and 12-inch roof slabs are not vulnerable to penetrator-type tornado missile damage." In order to develop a rough estimate of the fragility of the diesel generatar vaults with respect to high wind speed the following needs to be considered:
Missile impact loadings to
- Diesel Generator Building Walls
- Diesel Generator Building Rc,of Wind loadings
+
Without additional detailed structural analysis, several assumptions are required to develop a rough estimate of the fragility of the diesel generator vaults and oil tank compartment. First, the safety mar <; ins incorporated in well engineered structures are assu.ned to result in mean failure probability of approximately 2 to 3 times design failure. Second, the kinetic energy of a missile is related to the square of velocity. Third, the high confidence low probability of failure (llCLPF) is assumed to be equal to 168 mph. The estimation of the llCLPF is based on the fact that the no scabbing occurred in TORMIS simulations at 168 mph. Therefore the 168 mph wind speed is used as the llCLPF value and is judged to be conservative.
IPl I f.RAI tXX' 04/10/98
1940-98-20188
@ g '"*
Page 15 o f 21 13ased on these assumptions, a general statement of the mean wind speed for failure of the diesel generator building vaults can be expressed as the square root of 2.5 times the high confidence low probability of failure (11CLPF) wind speed. This is can be expressed as approximately 1.6 times 168 mph which is equal to 266 mph. Figure 1 provides the fragility of the diesel generator vaults with respect to wind generated missiles. Three curves are presented in cumulative probability distribution format for assumed mean failure wind speeds of the square root of 2,2.5 and 3 times the llCLPF. Table 5 provides the full distributions for these cumulative probability distributions.
The NRC staffs letter (" Evaluation of Diesel Generator Building Subjected to Tornado-Wind Generated Loading - Oyster Creek (TAC. NO. 49392 and 49394)", February 26,1990) indicates that the diesel generator building vaults are capable of withstanding 300 mile per hour winds and 2.0 psi depressurization (with no evaluation of load combinations). The 168 mph wind speed evaluation of the diesel generator vaults includes the combination of wind and missile loads. It is therefore assumed that the fragility curves presented in Figure I represent the combination of the wind and missile loads. Also, the fragility is assumed to be dominated by the missile loads.
It is assurned that the diesel generator building fragility is dominated by missile loads and therefore the contribution of the diesel generator building roof to the fragility is not considered.
Direct impact of a missile on the diesel generator building roofis not considered likely. The force associated with a roofimpact is estimated to be much lower since the trajectory is more likely the result of a drop rather than a forced impact. The resulting diesel generator building fragility to high winds is estimated as:
O Mean 5th Median 95th Diesel Generator Building 266 163 258 394 liigh Wind Fragility
)
This results in an estimated cumulative failure probability of 0.05 at a wind speed of 168 mph.
1 i
r i
l IPCIIRAl. DOC 04/)05/g
1940-98-20188
- Qps, Page 16 o f 21 Figurei Emergency Diesel Generator Building Walls Wind Generated Missile Fragility (Mean Va!ues of 238,266 and 291 mph) 1.00 1
,p
- ya
>f.
ar 0.00 -
' -?
f f
4 1
}..,
0.80 -
g
~
l
~
... - 0 M
}
0.70 -
0.60 -
> lh 2
/
a 0.50 -
2
, I e
m
,9 0.40 =
- - -~-
3 0.30 =
p
.,I j
0.20:
0.10 --
, s i
O.00 100 150 200 250 300 350 400 Wind Speed l-+-Mean 3 238 mph -e-Mean = 266 mph -a-Mean = 291 mph l IPEl:I:RAIIXK' 04/10/98
1940-98-20188 alImp Page 17 o f 21
~
Table s Emergency Diesel Generator Building Wall Fragility to High Wind Generated Missiles Mean = 238 mph Mean = 266 mph Nisan = 291 n:ph Mean 5th Median 95th Mean 5th Median 95th Mean 5th Median 95th 238 168 233 322 266 168 258 394 291 168 278 458 l
Bin Value Probability Cumulative Bin Value Probability Cumulative Bin Value Probability Cumulative 1
119 1.00E-03 1.00E-03 1
107 1.00E-03 1.00E-03 1
99 1.00E-03 1.00E-03 2
129 1.00E-03 2.00E-03 2
119 1.00E-03 2.00E-03 2
112 1.00E-03 2.00E-03 3
133 1.00E-03 3.00E-03 3
124 1.00E-03 3.00E-03 3
118 1.00E-03 3.00E-03 4
136 1.00E-03 4.00E-03 4
128 1.00E-03 4.00E-03 4
122 1.00E-03 4.00E-03 5
139 1.00E-03 5.00E-03 5
131 1.00E-03 5.00E-03 5
125 1.00E-03 5.00E-03 6
141 1.00E-03 6.00E-03 6
133 1.00E-03 6.00E-03 6
128 1.00E-03 6.00E-03 7
142 1.00E-03 7.00E-03 7
135 1.00E-03 7.00E-03 7
130 1.00E-03 7.00E-03 8
144 1.00E-03 8.00E-03 8
137 1.00E-03 8.00E-03 8
132 1.00E-03 8.00E-03 9
145 1.00E-03 9.00E-03 9
139 1.00E-03 9.00E-03 9
134 1.00E-03 9.00E-03 10 146 1.00E-03 1.00E-02 10 140 1.00E-03 1.00E-02 10 136 1.00E-03 1.00E-02 11 148 2.00E-03 1.20E-02 11 142 2.00E-03 1.20E-02 11 138 2.00E-03 1.20E-02 12 150 2.00E-03 1.40E-02 12 144 2.00E-03 1.40E-02 12 141 2.00E-03 1.40E-02 13 151 2.00E-03 1.60E-02 13 147 2.00E-03 1.60E-02 13 143 2.00E-03 1.60E-02 14 153 2.00E-03 1.80E-02 14 148 2.00E-03 1.80E-02 14 145 2.00E-03 1.80E-02 15 154 2.00E-03 2.00E-02 15 150 2.00E-03 2.00E-02 15 147 2.00E-03 2.00E-02 16 156 2.00E-03 2.20E-02 16 152 2.00E-03 2.20E-02 16 149 2.00E-03 2.20E-02 17 157 2.00E-03 2.40E-02 17 153 2.00E-03 2.40E-02 17 151 2.00E-03 2.40E-02 18 158 2.00E-03 2.60E-02 18 155 2.00E-03 2.60E-02 18 153 2.00E-03 2.60E-02 19 159 2.00E-03 2.80E-02 19 156 2.00E-03 2.80E-02 19 154 2.00E-03 2.80E-02 20 160 2.00E-03 3.00E-02 20 157 2.00E-03 3 00E-02 20 156 2.00E-03 3.00E-02 21 161 4.00E-03 3.40E-02 21 159 4.00E-03 3.40E-02 21 158 4.00E-03 3.40E-02 22 163 4.00E-03 3.80E-02 22 161 4.00E-03 3.80E-02 22 160 4.00E-03 3.80E-02 23 165 4.00E-03 4.20E-02 23 163 4.00E-03 4.20E-02 23 163 4.00E-03 4.20E-02 24 166 4.00E-03 4.60E-02 24 165 4.00E-03 4.60E-02 24 165 4.00E-03 4.60E-02 25 __
167 4.00E-03 3.00E-02 25_
167 4.00E-03 5.00 E_-02...
25 167 4.00E-03 5.00E-02,.
26 170 1.00E-02 6.00E-02 20 170 1.00E-02 6.00E-02 26 170 1.00E-02 6.00E-02 27 172 1.00E-02 7.00E-02 27 174 1.00E-02 7.00E-02 27 175 1.00E-02 7.00E-02 28 175 1.00E-02 8.00E-02 28 177 1.00E-02 8.00E-02 28 179 1.00E-02 8.00E-02 29 177 1.00E-02 9.00E-02 29 180 1.00E-02 9.00E-02 29 183 1.00E-02 9.00E-02 30 180 1.00E-02 1.00E-01 30 183 1.00E-02 1.00E-01 30 186 1.00E-02 1.00E-01 31 183 2.00E-02 1.20E-01 31 187 2.00E-02 1.20E-01 31 191 2.00E-02 1.20E-01 32 186 2.00E-02 1.40E-01 32 192 2.00E-02 1.40E-01 32 197 2.00E-02 1.40E-01 33 190 2.00E-02 1.60E-01 33 197 2.00E-02 1.60E-01 33 202 2.00E-02 1.60E-01 34 193 2.00E-02 1.80E-01 34 201 2.00E-02 1.80E-01 34 208 2.00E-02 1.80E-01 35 196 2.00E-02 2.00E-01
_35 205 2.00E-02 2.00E-01 35 212 2.00E-02
_ 2.00E-01 36 198 2.00E-02 2.20E-01 36 209 2.00E-02 2.20E-01 36 217 2.00E-02 2.20E-01 37 201 2.00E-02 2.40E-01 37 213 2.00E-02 2.40E-01 37 222 2.00E-02 2.40E-01 38 204 2.00E-02 2.60E-01 38 216 2.00E-02 2.60E-01 38 226 2.00E-02 2.60E-01 39 206 2.00E-02 2.80E-01 39 220 2.00E-02 2.80E-01 39 230 2.00E-02 2.80E-01 40 209 2.00E-02 3.00E-01 40 223 2.00E-02 3.00E-01 40 235 2.00E-02 3.00E-01 IPEIIRAl.1X)C 04/10/98
1940-98-20188 Qisu Page 18 o f 21 T'M 5 Emergency Diesel Generator Building Wall Fragility to High Wind Generated Missiles Mean = 238 mph Mean = 266 mph Mean = 291 mph j
l Mean 5th Median 95th Mean 5th Median 95th Mean 5th Median 95th 238 168 233 322 266 168 258 394 291 168 278 458 l
Bin Value Probabikty Cumulative Bin Value Probability Cumulative Bin Value Probabikty Cumulative l 41 211 2.00E-02 3.20E-01 41 226 2.00E-02 3.20E-01 41 239 2.00E-02 3.20E-01 42 213 2.00E-02 3.40E-01 42 230 2.00E-02 3.40E-01 42 243 2.00E-02 3.40E-01 43 216 2.00E-02 3.60E-01 43 233 2.00E-02 3,60E-01 43 247 2.00E-02 3.60E-01 1
44 218 2.00E-02 3.80E-01 44 236 2.00E-02 3.80E-01 44 251 2.00E-02 3.80E-01 45
,,,, 22 0_,2.00E-02.
4.00E-01 45
_240, _
2.00 E-02...
4.00E-01 45 255
. 2.00E-02 4.00E-01,j 46 223 2.00E-02 4.20E-01 46 243 2.00E-02 4.20E-01 46 259 2.00E-02 4.20E-01 '
47 225 2.00E-02 4.40E-01 47 246 2.00E-02 4.40E-01 47 263 2.00E-02 4.40E-01 48 227 2.00E-02 4.60E-01 48 249 2.00E-02 4.60E-01 48 268 2.00E-02 4.60E-01 49 229 2.00E-02 4.80E-01 49 253 2.00E-02 4.80E-01 49 272 2.00E-02 4.83E-01 i
. 50..
232
_2 00E-02.
_5.;00E-01
_ 50._
256
_2.00 E-02_
5.00E-01 50 276
. 2.00E-02_
5.00E-01 j 51 234 2.00E-02 5.20E-01 51 259 2.00E-02 5.20E-01 51 280 2.00E-02 5.20E-01 l 52 236 2.00E-02 5.40E-01 52 263 2.00E-02 5.40E-01 52 284 2.00E-02 5.40E-01 '
53 239 2.00E-02 5.60E-01 53 266 2.00E-02 5.60E-01 53 289 2.00E-02 5.60E-01 '
54 241 2.00E-02 5.80E-01 54 270 2.00E-02 5.80E-01 54 293 2.00E-02 5.80E-01 l
_ 55...
244 2.00E-02 6.;'10 E-01_
_5 5._, _273_
2.00E-02_ _ 6.00E-01 55 298 2.00E-02 6.00E-01..,)
56 246 2.00E-02 6.2CE-01 56 277 2.00E-02 6.20E-01 56 303 2.00E-02 6.20E-01 !
57 249 2.00E-02 6.40E-01 57 281 2.00E-02 6.40E-01 57 308 2.00E-02 6.40E-01 58 251 2.00E-02 6.60E-01 58 285 2.00E-02 6.60E-01 58 313 2.00E-02 6.60E-01 59 254 2.00E-02 6.80E-01 59 289 2.00E-02 6.80E-01 59 318 2.00E-02 6.80E-01
_.60.
257
. 2 00E-02_
7.00E 01 60 293 2.00E-02
._7 20E-01_
60 324 2.00E-02 7.00E-01 61 260 2.00E-02 7.20E-01 61 297 2.00E-02 7..'0E-01 61 329 2.00E-02 7.20E-01 62 263 2.00E-02 7.40E-01 62 302 2.00E-02 7.40E-01 62 335 2.00E-02 7.40E-01 63 266 2.00E-02 7.60E-01 63 307 2.00E-02 7.60E-01 63 342 2.00E-02 7.60E-01 64 270 2.00E-02 7.80E-01 64 312 2.00E-02 7.80E-01 64 349 2.00E-02 7.80E-01 65.
.,, 273 _
.. 2.;00 E.02 _
8.00E-01 65 318
. 2.00E:02 8.00E-01 65 356 2.00E-02 8.00E-01 66 277 2.00E-02 8.20E-01 66 324 2.00E-02 8.20E-01 66 364 2.00E-02 8.20E-01 67 282 2.00E-02 8.40E-01 67 330 2.00E-02 8.40E-01 G7 372 2.00E-02 8.40E-01 68 286 2.00E-02 8.60E-01 68 337 2.00E-02 8.60E-01 68 382 2.00E-02 8.60E-01 69 291 2.00E-02 8.80E-01 69 345 2.00E-02 8.80E-01 69 393 2.00E-02 8.80E-01
_.70 297 2 00E-02.
9.00E-01_
. 70_
354 2.00E-02.
9.00E-01 70 405.,,.
2.00E-02,_
9.00E-01 71 302 1.00E-02 9.10E-01 71 362 1.00E-02 9.10E-01 71 415 1.00E-02 9.10E-01 72 306 1.00E-02 9.20E-01 72 368 1.00E-02 9.20E-01 72 423 1.00E-02 9.20E-01 73 310 1.00E-02 9.30E-01 73 374 1.00E-02 9.30E-01 73 432 1.00E-02 9.30E-01 74 315 1.00E-02 9.40E-01 74 382 1.00E-02 9.40E-01 74 442 1.00E-02 9.40E-01 75 320
- 1. 00E-02_
9.50E-01
_7 5_,_
390 1.00E-02 9.50E-01 75_ _ _
454 1.00E-02 9.50E-01 76 324 4.00E-03 9.54 E-01 76 397 4.00E-03 9.54E-01 76 463 4.00E-03 9.54E-01 77 327 4.00E-03 9.58E-01
'77 401 4.00E-03 9.58E-01 77 469 4.00E-03 9.58E-01 78 330 4.00E-03 9.62 E-01 78 406 4.00E-03 9.62E-01 78 475 4.00E-03 9.62E-01 79 333 4.00E-03 9 66E-01 79 411 4.00E-03 9.66E-01 79 482 4.00E-03 9.66E-01 80 337 4.00E-03 9.70E-01 80 417 4.00E-03 9.70E-01 80 490 4.00E-03 9.70E-01 IPI:I I RAI DOC 04/10/98
1940-98-20188 ammy Page 19 o f 21 I!Lble5 Emergency Diesel Generator Building Wall Fragility to High Wind Generated Missiles l
Mean = 238 mph Mean = 266 mph Mean = 291 mph Mean 5th Median 95th Mean 5th Median 95th Mean 5th Median 95th 238 168 233 322 266 168 258 394 291 168 278 458 Bin Value Probability Cumulative Bin Value Probability Cumulative Bin Value Probability Cumulative 81 340 2.00E-03 9.72E-01 81 422 2.00E-03 9.72E-01 81 497 2.00E-03 9.72E-01 82 342 2.00E-03 9.74 E-01 82 425 2.00E-03 9.74 E-01 82 501 2.00E-03 9.74E-01 83 344 2.00E-03 9.76E-01 83 429 2.00E-03 9.76E-01 83 507 2.00E-03 9.76E-01 84 346 2.00E-03 9.78E-01 84 433 2.00E-03 9.78E-01 84 512 2.00E-03 9.78E-01 85_
_349_
2.00E-03 9.80E-01 85 437 2.00E-03 9.80E-01 85 518 2.00E-03 9.80E-01 36 352 2.00E-03 9.82E-01 86 442 2.00E-03 9.82E-01 86 525 2.00E-03 9.82E-01 87 355 2.00E-03 9.84 E-01 87 447 2.00E 03 9.84E-01 87 532 2.00E-03 9.84 E-01 88 359 2.00E-03 9.86E-01 88 453 2.00E-03 9.86E-01 88 540 2.00E-03 9.86E-01 89 363 2.00E-03 9.88E-01 89 459 2.00E-03 9.88E-01 89 550 2.00E-03 9.88E-01 90_
367 2.00E-03 9,.90E-01_
90 467 2.00E-03
_ 9.90E-01 90 561
_2.00E-03 9.90E-01 91 371 1.00E-03 9.91 E-01 91 474 1.00E-03 9.91 E-01 91 570 1.00E-03 9.91 E-01 92 374 1.00E-03 9.92E-01 92 479 1.00E-03 9.92E-01 92 577 1.00E-03 9.92E-01 93 378 1.00E-03 9.93E-01 93 485 1.00E-03 9.93E-01 93 585 1.00E-03 9.93E-01 94 382 1.00E-03 9.94 E-01 94 491 1.00E-03 9.94 E-01 94 595 1.00E-03 9.94 E-01 95 386 1.00E-03 9.95E-01 95 499 1.00E-03 9.95E-01 95 606 1.00E-03 9.95E-01 96 392 1.00E-03 9.96E-01 96 508 1.00E-03 9.96E-01 96 619 1.00E-03 9.96E-01 97 398 1.00E-03 9.97E-01 97 520 1.00E-03 9.97E-01 97 636 1.00E-03 9.97E-01 98 407 1.00E-03 9.98E-01 98 535 1.00E-03 9.98E-01 98 658 1.00E-03 9.98E-01 99 421 1.00E-03 9.99E-01 99 559 1.00E-03 9.99E-01 99 692 1.00E-03 9.99E-01 100 457 1.00E-03 1.00E-00 100 622 1.00E-03 1.00E-00 100 785 1.00E-03 1.00E-00 liTI:FRAI DOC OU10/98
f 1940-98-20188 k *'
Page 20 o f 21 2.
NRC OUESTION '
NUREG-1407 requests that a licensee " perform a confirmatory walkdown of the plant. The walkdown would concentrate on outdoorfacilities that could be affected by high winds, onsite storage of ha:ardous materials, and ojfsite developments. " Please provide a concise summary of l-the findings of this walkdown and rssolution of any identified potential vulnerabilities. In particular, pleaseprovide an assessment ofthe efects offailurefrom high winds andtornadoes of non-safety related structures and equipment on the functioning of safety related structures, systems and components. NRC Information Notice 93-53, Supplement 1, discusses further the concern withfailure ofnon-safety related structures afecting safety related structures.
RESPONSE
Walldowns were performed at various stages of the development of the Oyster Creek Individual Plant Examination for External Events. A summary of the walkdowns and fmdings are
' summarized below.
Structural and Site Walkdowns. These walkdowns were performed in support of the seismic evaluations of structures and outdoor facilities as well as a general site walkdown. Plant areas walked-down include:
Nearsite Combustion Turbines Area (including fuel oil tank and gas supply lines) l Intake and Discharge Canals Areas
+
Intake Structure Area
+
Fire Protection Pump House Area (including fire pond and fire pond dam)
+
Condensate Storage Tank Area Reactor Building Areas (external and internal)
+
Turbine Building Areas (external and internal)
Main Office Building (external and internal)
Diesei Generator Building (external and internal)
Stack area Seismie Fragility Walkdowns. This walkdown encompassed all areas of the reactor, turbine and main office buildings as well as the diesel generator building, condensate transfer building and switchyard areas. Walkdowns of the drywell were also performed. The purpose of the walkdown was to gather information to support the development of equipment fragilities at Oyster Creek.
Seismic-Fire Interaction Walkdown. A separate walkdown was performed with the specific goal of gathering information to support the development of the seismic fire interaction evaluation.
This walkdown centered on the potential for seismically induced impact of the various fire suppression systems at Oyster Creek including the issues identified in the Fire Risk Scoping Study. A full list of the plant areas included in the walkdown are contained on page 4.8-1 of the Oyster Creek IPEEE.
Fire Hazard Walkdowns. Several walkdowns were performed to support the evaluation of the fire analysis of the Oyster Creek IPEEE. A full plant walkdown was performed to assess the general state of the fire protection programs at Oyster Creek. Specific attention was devoted to the assessment of housekeeping, fire analysis targets, fire ignition sources, and fire propagation paths.
Following completion of the initial screening phase of the fire IPEEE, another walkdown was performed in support of the detailed analysis. This walkdown included the ability of the fire suppression systems to mitigate fire events. Specific attention was devoted to the important scenarios identified in the initial screening phase of the fire IPEEE.
A final walkdown of the fire areas which are analyzed as part of the detailed phase of the fire IPEEE was performed. This walkdown focused on the confirmation of the analysis results
= contained in the detailed phase of the fire IPEEE.
It' eel RAIIX)C 04/10NB
1940-98-20188 b'
Page 21 o f 21
--a=
I l
WALKDOWN FINDINGS Two improvements were identified in the Seismic-Fire interaction walkdowns which are also applicable to high wind scenarios. These plant improvements were the Anchorage of the Arrowhead Demineralized Trailer and the anchorage of the high pressure CO rack outside the 2
turbine building. The anchorage of the Arrowhead trailer was no longer anchored and the embedded eyehook was missing. The anchorage was re-installed and prevents the trailer from becoming a missile in high wind scenarios The high prestcre CO rack located outside the turbine 2
building has the potential to overturn in seismic or high wind scenarios resulting in the CO bott'es 2
becoming missiles. The high pressure CO rack and anchorage has been replaced.
2 In the case of tornado events, Oyster Creek utilizes the isolation condensers with long term makeup from the torus via the core spray system. All systems and structures utilized in this method of decay heat removal have been demonstrated by calculation as capable of withstanding wind speeds of 168 mph. A more detailed evaluation of the issues associated with information Notice 93-53 is provided in Attachment 7.
I NRC OUESTION As noted in NUREG-1407, section 2A, the latest probable maximum precipitation criteria published by the National Weather Service callfor higher rainfall intensities over shorter time mtervals and smaller areas than have previously been considered; this could result in higher site flooding levels, and greater roofponding levels. please assess the effects of applying these new criteria to Oyster Creek. Additionalinformation is given in Generic Letter 89-22.
RESPONSE
Oyster Creek has not assessed the effect of higher rainfall intensities over shorter time intervals and smaller areas than have previously been considered (GL 89-22). GPU Nuclear plans to assess the affects and will provide the result of the analysis to the staff no later than October 1998.
The design basis flooding criteria for Oyster Creek are described in the UFSAR Section 2.4.
Oyster Creek flood design considerations were addressed in the Oyster Creek Integrated Safety Assessment Systematic Evaluation Program, NUREG-0822, and the Oyster Creek NRC Safety Evaluation Report for the Full Term Operating License (NUREG-1382). Plant modifications to address reactor building and turbine building roof ponding levels have been implemented.
IPEEIRAIIXX' 04/10/98
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ -