ML20117K963

Forwards Detailed Operability Evaluation of Svc Water Sys W/Respect to Potential for Boiling in Containment Fan Coolers

ML20117K963
Person / Time
Site:	Point Beach
Issue date:	09/09/1996
From:	Link B WISCONSIN ELECTRIC POWER CO.
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
Shared Package
ML20117K965	List: ML20117K963 ML20117L005 ML20117L014
References
VPNPD-96-065, VPNPD-96-65, NUDOCS 9609120268
Download: ML20117K963 (9)

Text

Wisconsin Electnc POWER COMPANY 231 W MicNgort PO Box 2046. Milwaukee WI 53201-2046 (414)221 2345 VPNPD-96-065 September 9,1996 Document Control Desk US NUCLEAR REGULATORY COMMISSION Mail Station PI-137 Washington, DC 20555 Gentlemen:

DOCKET 50-266 AND 50-301 DETAILED OPERABILITY EVALUATION OF THE SERVICE WATER SYSTEM WITH RESPECT TO POST-ACCIDENT BOILING IN CONTAINMENT FAN COOLERS EOINT BEACH NUCLEAR PLANT. UNITS 1 AND 2 As reported in Licensee Event Report (LER) 266/30196-005-00 on August 30,1996, we have determined that I the containment fan coolers (CFCs) in both units are potentially susceptible to senice water (SW) boiling in the cooling coils during a design basis loss of coolant accident (LOCA) with a concurrent loss of offsite power l

(LOOP) and the normal sequencing of safety-related equipment. The transient in question was identified in Westinghouse Nuclear Safety Advisory Letter NSAL 96-003.

ta discuned in the LER and in previous meetings with Region 3 and others of the NRC staff, we hereby submit oui detailed operability evaluation of the Senice Water System with respect to the potential for boiling in the containment fan coolers. This evaluation confirms the operability determinations of previous evaluations of the transient, in that the CFCs and SW System are operable under these conditions. The evaluation also includes a schedule for final corrective action.

If you require additional information, please contact us.

1 Sincerely, 5

ll Bob Link Vice President gJy Nuclear Power GDA 9609120268 960909 PDR ADOCK 05000266 P PDR

~ ~ ~ ' ' '

Attachments cc: NRC Resident Inspector, NRC Regional Administrator AsubsMhn ofliluvnshLkudwitathn l J

VPNPD-96-065 Attachment A Summary Evaluation Page 1 of 8  !

4 CONDITION DESCRIPTION l Our research has indicated that original design and licensing reviews demonstrated that SW System boiling would not occur in the containment fan coolers under steady-state conditions. 1 Specifically, the PBNP FSAR has described the capability of containment fan coolers to perform their function without raising the exit temperature of service water to the boiling point.

Therefore, the potential effects of service water boiling in this region were not reviewed during the original design and licensing of the Point Beach Nuclear Plant. Currently, a transient condition has been postulated whereby boiling may occur in the containment fan coolers.

l Westinghouse Nuclear Safety Advisory Letter NSAL 96-003, " Containment Fan Cooler Operation During a Design Basis Accident" identifies the potential for steam flashing in the containment fan coolers (CFCs) during a design basis accident. The issue was presented in context of a typical CFC that was supplied by a closed loop component cooling water (CCW) l system. Wisconsin Electric reviewed the applicability of this issue to the Point Beach l configuration, and en August 1,1996, determined that the transient described in NSAL 96-003 did apply (Refer to LER 266/30196-005). It was determined that the containment fan coolers (CFCs) in both units were potentially susceptible to service water flashing in the cooling coils during a design basis loss of coolant accident (LOCA) with a concurrent loss of offsite power l (LOOP) and the delayed sequencing of safety-related equipment.

The transient condition occurs when the SW pump flow coasts down much quicker than the CFC air f'ow. Until SW flow is re-established, hot and moist air is drawn over the cooler coils and raises the temperature of the resident service water. Simultaneously, the resident liquid will be l removed by the downstream vacuum caused by the static hydraulic head (elevation) of the coolers. This latter phenomenon is called " column separation" and is characteristic of the Point Beach SW System design. The service water remaining in the coils will be quickly heated into water vapor. The resulting vapor pressure will help evacuate the downstream piping ofliquid.

At this point, restoration of service water flow could result in a waterhammer and potentially challenge the operability of the CFCs and the Service Water System.

The scenario may be summarized as follows:

1) LOCA and LOOP occur concurrently.

2) SW Pumps trip due to LOOP and coastdown quickly.

3) Containment Fans trip due to LOOP and coastdown to 1/4 full speed in about 3 minutes and to a full stop in 33 minutes under normal density conditions. (Undec the post-accident density conditions, one would expect this coastdown time to be much less).

4) Emergency Diesels energize the bus at 10 seconds into event.

5) First " block" of service water pumps start at 25 seconds into event.

VPNPD-96-065 Attachment A Summary Evaluation Page 2 of 8

6) Second " block" of service water pumps start at 30 seconds into event.

Based on the above sequence, the containment coils could be exposed to post-accident environment conditions for a period of 25-30 seconds without service water flow. If one assumes a completely homogenous environment in containment for this initial period of time, the coils would condense a saturated steam mixture with a temperature ranging from 220

• F at 4 secs to 280

• F at 14 seconds into the event. Using a simplified transient heat transfer analysis, we determined a coil full of 75

• F service water would quickly (within 10 seconds) reach a saturated state when exposed to these conditions.

Summarv - Evaluation of the Condition We contracted the services of two consultants, Fauske & Associates, Inc. and Sargent & Lundy, to evaluate this potential transient with respect to the Point Beach containment fan cooler configuration. Our initial operability assessment used the qualitative engineering judgment of these consultants as a basis for determining the system to be operable. This information was communicated to the NRC during previous conferences. Subsequent to the initial operability determination, both of our consultants and our own staff have done a substantial amount of additional evaluation to bound the waterhammer analysis and to demonstrate through theoretical and experimental analysis that the bounding analysis is conservative.

Based on the work of these consultants, we have concluded that the Point Beach containment fan cooler system will withstand the postulated transient and continue to be operable, as our original operability assessment had determined. The important functions that ensure SW system operability under this condition include containment heat removal, service water pressure boundary, and containment pressure boundanj. Both consultant reports are included for review with appropriate supporting attachments.

Summary of the Bounding Analysis Sargent & Lundy (S & L) focused their work on determining the bounding waterhammer load and piping stress analysis. Using conservative assumptions and input, their analysis demonstrates that the bounding waterhammer loading would result in pipe stresses that exceed design allowables, but do not exceed interim operability criteria. The S & L report includes the thermal-hydraulic analysis, pipe stress analysis, and a time-response analysis. These analyses are attached to this document, and summarized below:

1. The Thermal-Ilydraulic Analysis (THA) used conservative, bounding assumptions to determine the magnitude of boiling and the resultant waterhammer loads imposed by the transient. The thermal-hydraulic analysis also provided the basis for determining the potential SW flow restrictions caused by two-phase flow. These restrictions may delay the refill of containment fan coolers and challenge their capability to remove their design basis heat load within the time constraints of the design basis accident.

VPNPD-96-065 Attachment A Summary Evaluation Page 3 of 8

, Conservatively, the THA assumed a homogeneous containment atmosphere and the containment temperature profile in FSAR Section 14.3.4 to derive the external temperature imposed on the CFC. Many other conservative assumptions are made, and include consideration of voiding and steam condensation to accelerate any potential water slug.

No credit was taken for SW flow to the two lower CFCs in a unit, and no credit was taken for cushioning of the water slug by the presence of air released from the fluid during vaporization (i.e., column separation).

The THA considered several possible waterhammer scenarios occurring at several different locations in the SW System. Detailed description of these scenarios is provided in the attachment. For calculation of waterhammer pressure, NUREG-5220 was used. The analysis determined that the waterhammer peak overpressures for the steam generation portion of the transient are bounded by those for the refill portian of the transient. The

calculated maximum fill velocity for the refill portion of the transient was 6 ff/sec. Refer to

Attachment B.

2. Stress Analyses. Hydraulic transient time-histories were developed for evaluation of the piping and supports. A peak overpressure was also conservatively calculated (1,050 psig for the 8-inch lines and 1,370 psig for the 2-1/2-inch lines) for the purposes of hoop stress evaluation. Conservatively, the stress analyses assumed a Safe Shutdown Earthquake (SSE) to occur coincidentally with the waterhammer event. The calculated stresses were related to i

the code allowable values and the interim operability criteria described by our commitments

to IE Bulletin 79-14. Interim operability criteria are based on ASME Section III Appendix F limits. l Preliminary analyses indicated that pipe support stresses and piping stresses at several locations would exceed code allowable values. Therefore, iterative analyses were conducted to ensure that the interim operability criteria were met. In accordance with the interim operability guidelines, the supports in the system that exceeded interim operability criteria as a result of water-hammer loads were assumed to fail. The piping analyses were simplified and bounded by running the analyses without supports in the model to determine whether the pipe stresses would meet interim operability criteria.

4

As described in the attached analysis, the maximum combined pipe stresses calculated for this event (the refill transient) do not exceed interim operability criteria. In addition, the attached analysis determines that the hydraulic loads on the CFC cooling coils were found to be acceptable. Refer to Attachment B.

2

3. A Time-Response Analysis of the transient was conducted to determine the capability of the

containment fan coolers and SW System to remove their design basis heat load within the time constraints of the design basis accident. As described in the FSAR, the containment integrity analysis assumes the CFC system heat removal function starts at 60 seconds into the accident.

I l VPNPD-96-065 l

Summary Evaluation Attachment A )

{. Page 4 of 8 j The time-response analysis concluded that adequate liquid flow to the containment fan j coolers would be established within about 36 seconds, which satisfies the 60-second

)

i assumption. At this time, the transient is ended. Refer to Attachment B. >

4 i Summary of the Theoretical /Exnerimental Analysis l Fauske and Associates, Inc. (FAI) focused their efforts on demonstrating, through theoretical

! evaluation and experimentation, a more realistic estimate of the waterhammer loads. Using j industry literature from expens in the field of thermal-hydraulic analysis, FAI has demonstrated that the bounding analysis discussed above is conservative with respect to what theoretical and j experimental data would predict. In addition to their theoretical work, FAI conducted an i experimental scaled model of the thermal-hydraulic configuration similar to the Point Beach

! containment fan cooler configuration. The results of the experiment indicate that waterhammer l loading would be minimal compared to the bounding analysis results.

Refer to Attachment C for detailed discussion of this analysis.

POTENTIAL ADVERSE EFFECTS I

As discussed previously, there are two adverse effects created by the postulated CFC transient; l

(1) voiding and two-phase conditions in the SW System could delay containment heat removal l functions, and (2) waterhammer could stress piping and piping supports beyond their design allowable limits.

i l

! 1. Potential Delav in Performino Containment Heat Removal Functions i a

i j Voiding in the SW lines could delay the heat removal function of the containment fan cooler.

The containment fan cooler and SW pipe would have to refill with liquid after pump-start, and

! initial flowrates may be limited by the passage of steam through the downstream throttle valves.

! These flow restrictions could impede restoration of SW flow to the containment coolers and I delay their capability to remove the necessary design basis heat load within the time i' requirements of the FSAR (60 seconds), i I l

The time response analysis determined expected flow rates to the CFCs based on the calculated CFC downstream back pressure and found that once three pumps are running, the flow rate to the i CFCs will exceed the minimum required flow for design heat removal performance. Since a i minimum of three service water pumps are expected to be operating within 36 seconds, the design heat removal performance will meet the FSAR criteria.

1

2. Potential for Pinino and Suonort Stresses to Exceed Code Allowables 4

) i

< Analysis of the PBNP configuration during this transient has resulted in the conclusion that the ]

j hydraulic forces generated in the service water circuits of the containment fan coolers are not i sufficient to damage the pressure boundary integrity of the containment fan coolers. Therefore, I i waterhammer-induced pressure boundary failure and its consequential effects are not considered i

i l

. _ . . - - . - . - ~ . - - . - - - - - . - . - - - - - - - --- - --- =-

l

( t i VPNPD-96-065 Attachment A l Summary Evaluation Page 5 of 8

(

j credible. As discussed previously, analyses show that pipe stresses are within the Point Beach l interim operability stress limits.

COMPENSATORY MEASURES i

I l 'Although evaluations conclude that the design basis accident will not result in a break of the i SW pressure boundary, PBNP procedures, guidelines, and training provide additional assurance

[ that the unlikely failure of the service water pressure boundary will be identified and remedied.

1 Existing alarms and monitors will prompt the appropriate operator action to isolate any such SW

) leak. Radiation release monitoring and alarm is provided in the service water effluent, and j procedures prompt containment fan cooler isolation if necessary (Reference Alarm Response f

Book (ARB) IC20-C2-1,2C20-C2-9, RMSASRB CI-lRE-216, and RMSASRB CI-2RE-216).

i In addition, a SW containment fan cooler low-flow alarm would prompt appropriate corrective i action if waterhammer caused a significant leak in the containment cooler (Reference ARB col-l B2-3 and C2-9). No additional procedural guidance for this contingency is necessary.

a The potential safety risk of this event scenario will be minimized by limiting the time period until final corrective action is taken. A period of 14 raonths has been determined to pose an insignificant increase in safety risk, yet should allow sufficient time to prepare for safe and  ;

orderly design changes, if necessary.

1 I

Pursuant to GL 91-18, we have reviewed the operability of the existing condition. Operability evaluations have determined that the stresses on the safety-related piping and supports may exceed code allowables, but operability will be assured based on the interim operability criteria (based on ASME Section III Appendix F values). These criteria permit operation for an interim l period only.

Even though analyses show that the SW System pressure boundary remains functional during the transient, we have conducted a probabilistic safety assessment (PSA) analysis to evaluate the effects and risk impact ifit was assumed that this waterhammer scenario caused a SW pipe

~ failure l

VPNPD-96-065 Attachment A Summary Evaluation Page 6 of 8 inside containment. This PSA analysis used industry methods documented in EPRI TR-105396, "PSA Applications Guide", to determine the risk impact on the plant if this condition were to exist for 14 months (through Fall 1997). Refer to Attachment D for details.

The PSA analysis conservatively included evalua* ion of any event-initiator that could result in elevated containment temperature conditions. It was assumed that combining these initiators with a LOOP would result in a waterhammer. These initiating events included steamline breaks inside containment, LOCAs of all sizes, and Anticipated Transients Without Scram (ATWS).

Many of these initiators have very mild containment conditions during the first minute of the transient, and any postulated waterhammer effect would be limited.

The PSA analysis calculated that, following a postulated waterhammer transient and SW ,

failure, if the operators failed to isolate the SW leak inside containment, the overall PBNP Damage Frequency (CDF) and Large Early Release Frequency (LERF) would increase by 6.5E-8/yr. When this change in frequency is extended over the proposed 14-month interim period, the resulting change in Core Damage Probability (CDP) and Large Early Release Probability (LERP) were calculated to be 7.6E-8. Because the CDP is less than the PSA Applications Guide value of 1E-6, the acceptance of the waterhammer scenario my be considered "non-risk significant" and the change may be justified without the need for additional mitigating actions or analysis. Similarly, the LERP value is below the 1E-7 threshold and confirms that the scenario may be considered "non-risk significant".

Based on the nature of this transient phenomenon, the extremely low probability of the postulated sequence of events, and the expected complexity of modifications to minimize the potential effects of boiling, we plan to implement permanent corrective measures to both units prior to completion of the Unit 2 refueling outage scheduled for Fall 1997. Based on the proximity to the upcoming Unit 2 outage (scheduled shutdown October 5,1996), insufficient time exists to design and review permanent repairs in Unit 2 this fall. We believe this plan meets the intent of our commitment to interim operability criteria for safety-related piping.

All parameters assumed in the bounding analyses envelop currently observed operating and design limits. Therefore, no additional admini trative restrictions need to be created to ensure the integrity of the bounding analysis during the interim period of this condition.

PROPOSED RESOLUTION Engineering has been studying potential solutions to these transient issues during the develop-ment of the attached analyses. Several possible approaches have been identified, including the following:

. Modifications to the existing supports and piping systems to accept the analyzed loads from the postulated waterhammer event.

• Installation of an uninterruptable power supply on two or more existing service water pumps.

This approach would keep a minimum number of service water pumps running during the time-frame of concern.

VPNPD-%065 Attachment A Summary Evaluation Page 7 of 8 i l

e Installation of two additional service water pumps with fast starting diesel drivers. The l

timing of the diesels would be such that they provide water pressure to the fan coolers before i boiling begins. I e

Installation of flywheels on two or more service water pumps. This concept would keep the pumps rotating for sufficient time to prevent boiling in the fan coolers, until the service water pumps are loaded onto the emergency diesel generators.

Modification of the containment fan cooler service water discharge motor operated valves logic, such that they stay closed until the service water pumps are loaded onto the emergency diesel generators. We would also need to install an additional valve in the MOV bypass line l

which would close under these conditions as well. Additionally, a source of pressurized water could be installed to keep the fan cooler service water piping pressurized and filled with water during the identified transient conditions.

Each of these solutions presents risks to normal operation of the service water system and to other postulated accident scenarios which must be accurately identified and thoroughly reviewed to assure a responsible resolution. For example, installation of fly wheels could be designed to assure adequate flow is delivered during the first minute of the transient, but in addition, the inertia of restarting them after a minute of coastdown would need to be carefully considered in diesel loading and in service water flow after they restart. In addition to engineering analysis, sufficient time needs to be available to make necessary operating and emergency operating procedure changes, conduct thorough operator training, and to plan and implement all necessary documentation updates for a modification of this significance.

The above risks associated with pursuing these modifications in a short time frame are considered significant. As summarized previously, our PSA analysis has determined that accepting this situation for 14 months is not risk significant. Therefore, we are proposing that appropriate modifications be designed and implemented prior to the completion of the Unit 2 refueling outage in Fall,1997. Appropriate changes to Point Beach Technical Specifications and FSAR will be implemented at that time.

SUMMARY

/ CONCLUSIONS Our analyses conclude that service water boiling may occur in the vicinity of the containment fan coolers oc!ng the postulated design basis accident, however the effects of that boiling would not  !

compromise any safety-related function. The magnitude of the potential waterhammer would not cause a loss of service water system pressure boundary integrity, and the effects of cooler voiding would not impair the capability of the containment heat removal system to perform its function within the time limits of the analysis. Theoretical and experimental analyses have confirmed that the bounding analysis is conservative in calculating the magnitude of waterhammer loads. ,

Although the calculated waterhammer loads resulted in piping and support stresses exceeding code allowable values, piping code operability has been validated based on the interim l

operability criteria of ASME Code Section III. These criteria permit operation for an interim period only. Our probabilistic safety assessment for this condition calculated that the

VPNPD-96-065 . Attachment A Summary Evaluation Page 8 of 8 incremental increase in risk of operating 14 months under this postulated condition is not significant.' Therefore, corrective actions have been planned for the scheduled refueling outages in 1997.

Attachments:

B. Sargent & Lundy Report," Operability Assessment for Transient Conditions in the Point Beach Nuclear Plant Service Water System During a Coincident Loss of Coolant Accident and Loss of Offsite Power", dated September 9,1996 C. Fauske & Associates Report FAI/96-75," Evaluation of Possible Water-Hanuner Loads in the Service Water System for DBA Conditions", September,1996 D. PSA Evaluation of Water Hammer in SW Supply to Containment Accident Fan Coolers, i Rev.3 i l

l l

!