ML20196G033
| ML20196G033 | |
| Person / Time | |
|---|---|
| Site: | Byron, Braidwood, Quad Cities, Zion, LaSalle  |
| Issue date: | 05/02/1997 |
| From: | Hosmer J COMMONWEALTH EDISON CO. |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| GL-96-06, GL-96-6, NUDOCS 9705130280 | |
| Download: ML20196G033 (40) | |
Text
.~..
Commonwealth Edison Company
(
lex) Opus Place h
Downers Gros e, II. 60515-501 a
(
.May 2,1997 j
U.S. Nuclear Regulatory Commission Washington, DC 20555-0001 Attention:
Document Control Desk
Subject:
Braidwood Station Units 1 and 2 Byron Station Units 1 and 2 LaSalle County Station Units 1 and 2 Quad Cities Station Units 1 and 2 Zion Station Units 1 and 2 Commonwealth Edison Company (Comed) Response to Nuclear Regulatory Commission (NRC) Generic Letter (GL) 96-06, " ASSURANCE OF EQUIPMENT OPERABILITY AND CONTAINMENT INTEGRITY DURING DESIGN-BASIS ACCIDENT CONDITIONS," dated September 30,1996.
NRC Dockets 50-454 and 50-455 i
NRC Dockets 50-456 and 50-457 NRC Dockets 50-373 and 50-374 NRC Dockets 50-254 and 50-265 NRC Dockets 50-295 and 50-304
References:
1.
NRC Generic Letter 96-06, " ASSURANCE OF EQUIPMENT l[f OPERABILITY AND CONTAINMENT INTEGRITY DURING l
DESIGN-BASIS ACCIDENT CONDITIONS ", dated September 18, 1996
/hl 2.
John Hosmer Letter to USNRC, Response to Generic Letter 96-06,
)
" ASSURANCE OF EQUIPMENT OPERABILITY AND
/
CONTAINMENT INTEGRITY DURING DESIGN-BASIS ACCIDENT CONDITIONS," dated October 28,1996 3.
John Hosmer Letter to USNRC, Response to Generic Letter 96-06,
" ASSURANCE OF EQUIPMENT OPERABILITY AND CONTAINMENT INTEGRITY DloUNG DESIGN-BASIS ACCIDENT CONDITIONS," dated January 28,1997 4.
J. S. Perry Letter to USNRC, Dresden Nuclear Power Station Units 2 and 1'"*q p/
3 Supplemental Response to NRC Generic Letter (GL) 96-06,
" ASSURANCE OF EQUIPMENT OPERABILITY AND CONTAINMENT INTEGRITY DURING DESIGN-BASIS ACCIDENT CONDITIONS," dated March 28,1997 l
9705130280 970502 i
.fff'ff';ffff' f' f.ff'ff.ff",.f.l.
l PDR ADOCK 05000254-P PDR r r l
A Umcom Company I-
1 1
g e
USNRC May 2,1997
)
In the Reference 1 Letter the NRC staff notified all addressees about safety-significant issues that could affect containment integrity and equipment operability during accident conditions. In NRC l
Generic Letter 96-06 the NRC staff required all addressees to submit within 30 days of the generic letter a written response indicating if the requested actions would be completed and submitted within the requested time period. Reference 2 transmitted Comed's 30 day response to j
the Generic Letter.
Reference 3 transmitted Comed's 120-day response summarizing our actions taken in response to the requested actions. In that letter Comed committed to provide additional details on implementing hardware changes at each of our sites. Reference 4 described the work to be done at Dresden Station. The plans for Byron, Braidwood, Quad Cities, and Zion Stations are provided in Attachments 1 through 4 of this letter. LaSalle Station is committed to provide a detailed response by May 30,1997, which will describe their analysis and any design changes required.
l To the best of my knowledge and belief, the information contained herein is true and accurate.
Please direct any questions c; neerning this response to Marcia Lesniak, Nuclear Licensing Administrator at 630-663-6481 Sincerely, i
& 8' Y, John B. Hosmer Vice President cc:
A. Beach, Regional Administrator-RIII R. Capra, Director of Directorate III-2, NRR G. Dick, Byron /Braidwood Project Manager, NRR D. Skay, LaSalle Project Manager, NRR R. Pulsifer, Quad Cities Project Manager, NRR C. Shiraki, Zion Project Manager, NRR C. Phillips, Senior Resident Inspector (Braidwood)
S. Burgess, Senior Resident Inspector (Byron)
M. Huber, Senior Resident Inspector (LaSalle)
C. Miller, Senior Resident Inspector (Quad Cities)
A. Vegel Senior Resident Inspector (Zion)
B. Wetzel, Project Manager, NRR Office of Nuclear Facility Safety - IDNS kngewic\\gl\\gt9hueW16SUP. Don GL 964)6 120 DAY RESPONSE 4
t i
Braidwood Station NRC Docket 50-456 and 50-457 Response to NRC Generic Letter 96-06,
" Assurance of Equipment Operability and Containment Integrity During Design-Basis Accident Conditions," dated September 30,1996 NRC REQUEST:
"Within 120 days of the date of this generic letter, addressees are requested to submit a written summary report stating actions taken in response to the requestedactions noted above, conclusions that were reached relative to susceptibilityfor waterhammer and two-phaseflow in the containment air cooler cooling water system and overpressuri:ation of piping that penetrated containment, the basisfor continued operability of affected systems and components as applicable, and corrective actions that were implemented or are planned to be implemented. Ifsystems werefound to be susceptible to the conditions that are discussedin this generic letter, identify the systems affectedanddescribe the specific circumstances involved. "
NRC REQUESTED ACTIONS:
" Addressees are requested to determine:
(1) if containment air cooler cooling water systems are susceptible to either waterhammer or two-phaseflow conditions duringpostulated accident conditions; (2) ifpiping systems thatpenetrate the containment are susceptible to thermal expansion offluidso that overpressuri:ation ofpiping could occur. "
RESPONSE TO ITEM 1:
Potential for Waterhammer Transient due to LOCAILOOP scenario l
In the 120-day response of January 28,1997, Byron /Braidwood provided the preliminary results I
of the review on potential waterhammer within the reactor containment fan cooler service water system. The calculations have been completed and have confirmed the acceptability of the
)
existing configuration without the need for changes to the system physical configuration or system operation. The approach used to reach this conclusion is discussed in detail below.
l l
kngenerungrg!9e06t960b5UP.m GL 964M 120 DAY RESPONSE i
t
e System Considerations l
The Essential Service Water (SX) system is an open loop system with the water source and the i
discharge water level located at elevations above the top elevation of the reactor containment fan I
cooling coils. The Reactor Containment Fan Coolers (RCFCs) consist of two banks of SX and chilled wner coils. The SX water coils provide the post accident heat removal while the chilled w::ter coils are not used post accident. There are 5 SX and 5 chilled water coils per bank. In normal operation, both sets of coils have flow, and the RCFC fan is operating in a high speed mode. Design service water flow through the RCFC is 2650 gpm, divided equally among the 10 l
coils.
Based upon the concern identified in Generic Letter 96-06, and using the above system considerations, a calculation was developed to determine a best estimate quantification of the thermal hydraulic behavior of the RCFC coils during the initial phase of a Loss of Coolant Accident / Main Steam Line Break (LOCA/MSLB) with Loss of Offsite Power (LOOP). As previously noted, the RCFCs are initially running at high speed and, upon loss of offsite power, begin coasting down. The service water flow also coasts down and is restored approximately 43 seconds into the event when the SX pump starts following bus energization by the diesel generator. Prior to restoration of the SX flow, heet transfer from the containment to the RCFC coils has the potential to create steam voids in the coils. The principal concern being addressel is whether upon SX pump restart the rapid collapse of these voids will result in significant waterhammer loads to the RCFC coils or piping.
Methodology A detailed model of one side of an RCFC SX coil set was developed for use with the RELAP5M3 computer code. The model provided a best estimate predictive capability of the thermal hydraulic behavior that would be anticipated during this type of event. The model purposely employs boundary condition definitions that will lead to conservative overprediction of the initiation of voiding, the extent ofvoiding, and the rate ofvoid collapse due to SX pump restart. The model was used for the limiting DBA LOCA containment conditions. The MSLB events were also examined to ensure that the most limiting conditions relative to coil void generation had been enveloped.
The calculation quantifies the behavior of the SX side of the RCFCs during the initial phases of the design basis LOCA/MSLB with ' LOOP. In this situation, the SX flow will decay rapidly on loss of offsite power and allow boiling to occur in the RCFC tubes, prior to restart of the SX flow during the diesel generator loading sequence. The calculation demonstrates that for a representative piping configuration some voiding will occur, particularly in the upper coils, but that conditions leading to severe waterhammer will not occur. In addition, the calculation develops the force time history information for the RCFC coils and outlet piping to allow stmetural evaluation of the loading conditions calculated.
L l
l bgenerrW%00%065UP. DOC)
GL 96-06 120 DAY RESPONSE
i Use of a Representative Piping Configuration The Braidwood RCFC installation was selected for the development of the detailed model. The general arrangement of all the RCFCs in the Byron /Braidwood containments is very similar. In both plants, the RCFC sits at a lower elevation than the lake / cooling tower, and the RCFC coil is efTectively at the base of a large U formed by the inlet header piping and the exhaust header piping. The choice of plant modeled was based on the fact that the Byron SX piping configuration connects to the cooling tower basin which is at a slightly higher elevation than the Braidwood SX connection to the cooling lake (approximately 4 feet). Therefore, there is less static pressure available in the Braidwood RCFC coils following SX pump coastdown, and more voiding is anticipated. This arrangement allows a bounding analysis to be performed that is applicable to both sites.
Model Description Hydraulic Model The model provides a very detailed modeling of one bank of SX cooling coils and the connecting piping. To assure appropriate boundary conditions, the inlet and outlet time dependent volumes are specified as pressure conditions only, and are connected to the piping with normal single junction components. This allows flow to develop in either direction based on developed pressure drop. The initial pressure on the upstream time dependent volume (TDV)is selected to provide the design flow through the system of 1325 (2650/2) gpm. The downstream TDV is set at a constant pressure that is equivalent to the elevation distwee from the volume to the lake level.
While discharge piping losses would cause a slightly highet pressure at this point in normal operation, it is conservative and appropriate to neglect them in this analysis. The pump coastdown is simulated by reducing the pressure in the upstream volume to atmospheric conditions in a 5 second time interval. This provides a rapid flow coastdown that is bounding compared to pump vendor input which indicates a pump coastdown time on the order of 10-20 seconds. The flow control valves are not modeled in this calculation, since they would provide an additional restriction to flow, requiring additional inlet TDV pressure, potentially delaying and/or limiting the extent of voiding predicted. In addition, assuming the flow control valves are wide open allows calculation of maximum potential loads on the exhaust riser section, as the two phase mixture is accelerated during the pump start.
Pump restart is simulated by rapidly increasing (linearly over one second) the inlet TDV to twice the steady state pressure calculated above and then allowing it to decay over approximately 10 seconds to the nominal value. This approach was selected to provide a rapid initiation of flow and resultant void collapse. The maximum discharge pressure of the SX pumps is approximately 40 -
100 psig, with approximately 30 psi of elevation head loss alone anticipated prior to reaching the cooler inlet header. Additional head losses due to strainers, pipe fittings and pipe friction would reduce the pressures even further. Therefore, the use of 43 psig as a spike pressure on an unthrottled system is reasonable with respect to actual system capabilities. A review of the pump motor torque curves and pump / motor inertias shows that the pump will reach full speed in 1 k \\generr\\gr,glWWW16SUP DOC GL 964M 120 DAY RESPONSE
I secend when starting at 100% rated voltage conditions. The pressure rise in the inlet TDV is conseivatively modeled to occur over one second. A more gradual increase in pressure would be expected as the pump proceeds to accelerate fluid throughout the open system, operating at the high flow end of the pump curve and moving back to the operating point as flow losses build up with fluid velocity.
Coil Heat Transfer Model The coils are divided into 2 foot long piping segments over the 40 foot run of the coil (4 pass,10 feet per pass, with 60 tube side circuits). The tubing is modeled as a two-sided conducting cylinder of Cu-Ni material. The inside of the tubing employs the normal RELAP5 heat transfer l
map. The outside of the tube represents the containment boundary condition by providing a table of temperature vs time in conjunction with a specified convective boundary condition. A heat 2
transfer coefficient of 500 Btu / hr-ft _.F was employed as a constant value for this problem. This conservatively bounds the values that would be predicted by the Uchida condensing correlation, and effectively sets the outside of the tube to the containment temperature. The Uchida 2
correlation returns a range of heat transfer coefficients from 2-280 Btu / hr-ft,.F based on the air / steam ratio. Calculation of the air / water mass ratios at 50 seconds for the containment under LOCA and MSLB conditions shows that the appropriate coefficient should be approximately 92 2
2 Btu / hr-ft _.F for LOCA and 41 Btu / hr-ft *F for MSLB applications. Since the coils are finned, a multiplier on the outside heat transfer coefficient is employed to account for the increased efficiency. A model of the tubing and fins has been developed to calculate appropriate heat transfer multipliers for the coils as a function of assumed outside heat transfer coefficient. The use of the multipliers causes the effective outside heat transfer coefficient to be approximately 230 2
2 Btu / hr-ft _.F for the LOCA case and 130 Btu / hr-ft,.F for the MSLB case. As noted above, in 2
the final loads development case used for LOCA, 500 Btu / hr-ft.F was employed to ensure additional conservatism in the heat transfer modeling.
A review of the results of recent Westinghouse calculations to support increased initial containment temperatures of 130*F was performed. The limiting case for DBA LOCA was established to be the Double Ended Pump Suction (DEPS) Break with maximum safety injection.
This break establishes relatively high temperatures very quickly into the event, and sustains temperatures in the range of 260*F throughout the period ofinterest for this analysis. A review of 2
the MSLB breaks performed for the same study demonstrated that the 0.942 ft split mpture is bounding with respect to other break sizes analyzed. A comparison of the DEPS temperature profile and the MSLB profile shows that the DEPS profile reaches higher temperatures earlier, but the MSLB profile passes the DEPS over time, reaching a higher peak temperature. Both breaks were analyzed to ensure that the maximum void predictions would be obtained.
j i
UpwrWWWWmSUP.DOo GL 964)6 120 DAY RESPONSE
i
]
1 Development of Piping Forces i
. In order to assess the effects of the hydraulic transients calculated for the RCFC coils during the restart of the SX flow in conjunction with a LOOP /LOCA scenario, it is necessary to derive the force time history that would result. Methodologies have been previously demonstrated to provide the force time histories associated with hydraulic transients calculated using the RELAP5 computer code. These methodologies typically employ a post processing code such as R5 FORCE or REFORC to operate on the RELAP5 data and develop the forces on piping segments. A method has been developed using the XMGR software package to perform a similar calculation.
These loads were then utilized in structural evaluations of the systems.
i I
- Two Phase Flow i
i The calculations demonstrate that under the limiting assumptions applied, voiding will occur in the J-RCFC coils prior to restart of the pump. The presence of the voids leads to forces on the structures as the voids are compressed and fluid accelerates following pump restart. Key points that have been determined include:
i 4
The forces on the system are relatively small and do not cause the equipment to exceed e
allowable stresses Conditions leading to rapid void collapse and resultant waterhammer do not exist for this configuration i
Additionally, due to the configuration of the system, particularly with respect to the location of the throttle valves on the exit side of the RCFC, combined with the SX pump capability and the geometry of the RCFC piping, the potential for flow stall due to increased two phase pressure drop is considered highly unlikely.
i Conclusion The results of the above analysis have concluded that the resultant loads are relatively small, demonstrating that while there is some additional structural loading due to the two phase transient, the effect is relatively small. The piping stresses, support and structural loads, and i
equipment loads that result from this transient are within design basis limits. This c<mclusion, j
j combined with the lack of two phase flow problems associated with the above scenaiio, 5
completes the response on this issue.
I l
l
)
kigener 6stigl%fE%0651IP.Doct GL 964)6 120 DAY RESPONSE
'l i
1
RESPONSE TO ITEM 2:
Potential for Overpressurization of Isolatable Piping Sections The review performed to address thermally induced pressurization resulting from the Loss of Coolant Accident / Main Steam Line Break (LOCA/MSLB) was described in the Braidwood attachment to the response dated January 28,1997. The January 28th response described the screening evaluation performed to identify the containment penetrations susceptible to thermally induced overpressurization.
The penetrations that were identified in the January 28th response as susceptible to overpressurization are listed below:
P-5, P-6, P-8, and P Containment Chilled Water system P Reactor Coolant Drain Tank Pump Discharge P Containment Demineralized Water Supply P Fuel Pool Cooling Return to Refueling Cavity P Containment Fire Protection Supply P Reactor Coolant System Loop Fill Header P Primary Water Supply to Reactor Coolant Pump Seal #3 and Pressurizer Relief Tank P Containment Floor Drain Sump Pump Discharge P Safety Injection Accumulator Fill Line P Fuel Pool Cooling Suction from Refueling Cavity Based on later versions of the screening calculations, detailed evaluation of containment penetration pressure conditions, and component availability for the orderly recovery from a MSLB event, the following penetrations have been added to the list of penetrations for which l
thermally induced pressure conditions will be addressed:
P Process Sampling (Containment isolation valves -
normally closed during plant operation).
P Chemical and Volume Control System Normal Charging Flow Path (Containment isolation valves -
close on a containment isolation signal).
l P Component Cooling Water Return from the Reactor Coolant Pump Thermal Barrier Heat Exchanger (Containment isolation valves - close on a containment phase B isolation signal).
e,emi is i9erme(msur. con GL 96 )6 120 DAY RESPONSE ss
= -
On each side of the containment wall, these penetrations generally have a containment isolation valve which is either closed during normal plant operation (P-30, P-32, P-37, P-55, P-57 and P-
- 70) or automatically closes on a containment isolation signal (P-5, P-6, P-8, P-10, P-11, P-24, P-34, P-44, P-47 at:d P-71). These penetrations could be potentially heated during a LOCA or a MSLB inside containment, either of which would provide the containment isolation signal.
The overpressure mitigation methodologies presently planned for penetratioits susceptible to overpressurization fall into four categories as listed below:
A. Design changes are required to mitigate thermally induced overpressure conditions - 11 penetrations (P-5, P-6, P-8, P-10, P-11, P-24, P-34, P-37, P-44, P-47, and P-55).
B. Overpressure conditions that are intended to be mitigated by procedural changes for valve lineups - one penetration (P-30).
C. Overpressure conditions that are intended to be mitigated by penetration draining - two penetrations (P-32, P-57).
D. A more detailed analysis is warranted to determine if any over pressure concern actually exists - two penetrations (P-70, P-71).
The design changes required for Category A items listed above are being implemented through i
the Braidwood Station design change procedures and processes. The design change process f
insures that all aspects of materials, installation, system operation and system performance / interaction are properly considered and factored into the required design changes.
Preliminary design activities are presently in process for these changes. Identified overpressure conditions will be mitigated by the installation of spring loaded relief valves, check valve arrangements or modifications to containment isolation valves. Installation of these design changes is presently anticipated at Braidwood during outages AIR 07 (Fall 1998) and A2R07 (Spring 1999). These outages represent the earliest scheduled refueling outages where it is practical to properly evaluate and install design changes of this magnitude.
Category B and C items are currently being reviewed and, if confirmed appropriate, mitigating actions will be implemented no later than the unit return to operation from Braidwood outages AIR 06 (Spring 1997) and A2R06 (Fall 1997). If the reviews being conducted determine that the proposed actions are not appropriate, design changes will be implemented to mitigate thermally induced overpressure conditions utilizing the methods and schedules described above for Category A items.
UgmtW4MMMUP.Doo GL 96-06 120 DAY RESPONSE
1 l
Category D items are being analyzed in detail (beyond the original screening process that l
conservatively enveloped many conditions and circumstances) to determine if there is an actual l
overpressure condition and the magnitude of such a condition, ifit exists. These further analyses are expected to be completed such that, if any design changes are required to mitigate thermally induced overpressure conditions, they will be able to utilize the methods and schedules described above for Category A items.
A basis for operability has been determined for the affected penetrations and included consideration of one or more of the following: expansion of the trapped fluid in voided areas of the isolated piping section; leakage from one of the containment isolation valves (seat, packing, or l
body to bonnet flange); existing insulation of piping inside containment or heat dissipation from l
piping outside containment to prevent the temperature induced pressure increase; lifting of the air l
operated valves due to the prcssure increase and plastic straining of the affected pipe to accommodate the pressure increase This basis for operability provides the justification for continued unit operation as well as unit restart from refueling and other outages until the appropriate design changes are installed.
Additionally, Comed is actively working with EPRI, BWROG, and NEI in determining the best options with respect to resolution of the concerns identified in GL 96-06. The outcorne of the ongoing efforts by these industry groups will be used in adjusting, as necessary, the long-term plans for resolution of this issue at Braidwood Station.
l t
i k tgeners\\gr4%*9WUP. DOG GL 964)6 120 DAY RESPONSE
Byron Station NRC Docket 50-454 and 50-455 Response to NRC Generic Letter 96-06,
" Assurance of Equipment Operability and Containment Integrity During Design-Basis Accident Conditions", dated September 30,1996 I
f NRC REQUEST:
Within 120 days of the date of this generic letter, addressees are recpiested to submit a written summary report stating actions taken in response to the requested actions noted l
above, conclusions that were reachedrelative to susceptibilityfor waterhammer and two-phaseflow in the containment air cooler cooling water system and overpressuri:atrion of piping that penetrates containment, the basisfor continued operability of affected systems and components as applicable, and corrective actions that were implemented or l
are planned to be implemented Ifsystems werefound to be susceptible to the conditions l
that are discussed in this generic letter, identify the systems affected and describe the specific circumstances involved.
NRC REQUESTED ACTIONS:
i
" Addressees are requested to determine:
l if containment air cooler cooling water systems are susceptible to either water hammer l
or two-phaseflow conditions duringpostulated accident conditions:
l l
ifpiping systems thatpenetrate the containment are susceptible to thermal expansion of j
fluidso that overpressuri:ation cfpiping could occur. "
l t
l RESPONSE TO ITEM 1:
Potential for Waterhammer Transient due to LOCA/ LOOP scenario In the 120-day response of January 28,1997, Byron /Braidwood provided the preliminary results of the review on potential waterhammer within the reactor containment fan cooler service water system. The calculations have been completed and have conhrmed the acceptability of the existing configuration without the need for changes to the system physical configuration or system operation. The approach used to reach this conclusion is discussed in detail below.
i.
I kngenetryg!%EmmSUPJX)Cs GL 96 06 120 DAY RESPONSE l
l
l i
i
[
System Considerations l
The Essential Service Water (SX) system is an open loop system with the water source and the discharge water level located at elevations above the top elevation of the reactor containment fan cooling coils. The Reactor Containment Fan Coolers (RCFCs) consist of two banks of SX and chilled water coils. The SX water coils provide the post accident heat removal while the chilled 4
water coils are not used post accident. There are five SX and five chilled water coils per bank. In
- =
normal operation, both sets of coils have flow, and the RCFC fan is operating in a high speed mode. Design service water flow through the RCFC is 2650 gpm, divided equally among the ten i
coils.
Based upon the concern identified in Generic Letter 96-06, and using the above system considerations, a calculation was developed to determine a best estimate quantification of the 4
thermal hydraulic behavior of the RCFC coils during the initial phase of a Loss of Coolant Accident / Main Steam Line Break (LOCA/MSLB) with Loss of Offsite Power (LOOP). As previously noted, the RCFCs are initially running at high speed and, upon loss of offsite power, begin coasting down. The. service water flow also coasts down and is restored approximately 43 seconds into the event when the SX pump starts following bus energization by the diesel generator. Prior to restoration of the SX flow, heat transfer from the containment to the RCFC coils has the potential to create steam voids in the coils. The principal concern being addressed is whether upon SX pump restart the rapid collapse of these voids will result in significant waterhammer loads to the RCFC coils or piping.
4 Methodology A detailed model of one side of an RCFC SX coil set was developed for use with the RELAP5M3
)
computer code. The model provided a best estimate predictive capability of the thermal hydraulic behavior that would be anticipated during this type of event. The model purposely employs boundary. condition definitions that will lead to conservative overprediction of the initiation of voiding, the extent of voiding, and the rate of void collapse due to SX pump restart. The model was used for the limiting DBA LOCA containment conditions. The MSLB events were also examined to ensure that the most limiting conditions relative to coil void generation had been enveloped.
The calculation quantifies the behavior of the SX side of the RCFCs during the initial phases of the design basis LOCA/MSLB with LOOP. In this situation, the SX flow will decay rapidly on i
loss of offsite power and allow boiling to occur in the RCFC tubes, prior to restart of the SX flow during the diesel generator loading sequence. The calculation demonstrates that for a representative piping configuration some voiding w.~ti occur, particularly in the upper coils, but i
that conditions leading to severe waterhammer will not occur. In addition, the calculation
^
develops the force time history information for the RCFC coils and outlet piping to allow structural evaluation of the loading conditions calculated.
i k:Menertyrgl%06N065UP DOC \\
GL %06 120 DAY RESPONSE
i Use of a Representative Piping Configuration l
The Braidwood RCFC installation was selected for the development of the detailed model. The l
general arrangement of all the RCFCs in the Byron /Braidwood containments is very similar. In l
both plants, the RCFC sits at a lower elevation than the lake / cooling tower, and the RCFC coil is effectively at the base of a large U formed by the inlet header piping and the exhaust header l
piping. The choice of plant modeled was based on the fact that the Byron SX piping configuration connects to the cooling tower basin which is at a slightly higher elevation than the Braidwood SX connection to the cooling lake (approximately four feet). Therefore, there is less static pressure availal.% in the Braidwood RCFC coils following SX pump coastdown, and more voiding is anticipated. This arrangement allows a bounding analysis to be performed that is applicable to both sites.
Model Description l
Hydraulic Model The model provides a very detailed modeling of one bank of SX cooling coils and the connecting piping. To assure appropriate boundary conditions, the inlet and outlet time dependent volumes l
are specified as pressure conditions only, and are connected to the piping with normal single l
junction components. This allows flow to develop in either direction based on developed pressure drop. The initial pressure on the upstream time dependent volume (TDV) is selected to provide l
l the design flow through the system of 1325 (2650/2) gpm. The downstream TDV is set at a j
constant pressure that is equivalent to the elevation distance from the volume to the lake level.
While discharge piping losses would cause a slightly higher pressure at this point in normal operation, it is conservative and appropriate to neglect them in this analysis. The pump l
coastdown is simulated by reducing the pressure in the upstream volume to atmospheric conditions in a five second time interval. This provides a rapid flow coastdown that is bounding compared to pump vendor input indicating a pump coastdown time on the order of 10-20 seconds. The flow control valves are not modeled in this calculation, since they would provide an additional restriction to flow, requiring additional inlet TDV pressure, potentially delaying and/or i
limiting the extent ofvoiding predicted. In addition, assuming the flow control valves are wide open allows calculation of maximum potential loads on the exhaust riser section, as the two phase mixture is accelerated during the pump start.
Pump restart is simulated by rapidly increasing (linearly over one second) the inlet TDV to twice l
the steady state pressure calculated above and then allowing it to decay over approximately ten seconds to the nominal value. This approach was selected to provide a rapid initiation of flow and resultant void collapse. The maximum discharge pressure of the SX pumps is approximately 90 -
100 psig, with approximately 30 psi of elevation head loss alone anticipated prior to reaching the l
cooler inlet header. Additional head losses due to strainers, pipe fittings, and pipe friction would reduce the pressures even further. Therefore, the use of 43 psig as a spike pressure on an unthrottled system is reasonable with respect to actual system capabilities. A review of the pump motor torque curves and pump / motor inertias shows that the pump will reach full speed in one second when starting at 100% rated voltage conditions. The pressure rise in the inlet TDV is kigeserrWgIME%%SUP DOC \\
oL 9606 120 DAY RESPONSE
conservatively modeled to occur over one second. A more gradualincrease in pressure would be expected as the pump proceeds to accelerate fluid throughout the open system, operating at the high flow end of the pump curve and moving back to the operating point as flow losses build up with fluid velocity.
Coil IIcat Transfer Model The coils are divided into two foot long piping segments over the 40 foot run of the coil (four pass, ten feet per pass, with 60 tubes side circuits). The tubing is modeled as a two-sided conducting cylinder of Cu-Ni material. The inside of the tubing employs the normal RELAP5 heat transfer map. The outside of the tube represents the containment boundary condition by providing a table of temperature versus time in conjunction with a specified convective boundary 2
condition. A heat transfer coeflicient of 500 Btu / hr-R *F was employed as a constant value for this problem. This conservatively bounds the values that would be predicted by the Uchida condensing correlation, and effectively sets the outside of the tube to the containment temperature. The Uchida correlation returns a range of heat transfer coeflicients from 2 - 280 Btu / hr-n2,.F based on the air / steam ratio. Calculation of the air / water mass ratios at 50 seconds for the containment under LOCA and MSLB conditions shows that the appropriate coeflicient 2
should be approximately 92 Btu / hr-fl _.F for LOCA and 41 Btu / hr-n2_.F for MSLB applications. Since the coils are finned, a multiplier on the outside heat transfer coeflicient is employed to account for the increased efliciency. A model of the tubing and fins has been developed to calculate appropriate heat transfer multipliers for the coils as a function of assumed outside heat transfer coeflicient. The use of the multipliers causes the effective outside heat 2
2 transfer coefficient to be approximately 230 Btu / hr-ft _.F for the LOCA case and 130 Btu / hr-fl,
- F for the MSLB case. As noted above, in the final load development case used for LOCA,500 Btu / hr-n2 *F was employed to ensure additional conservatism in the heat transfer modeling.
l A review of the results of recent Westinghouse calculations to support increased initial containment temperatures of 130*F was performed. The limiting case for DBA LOCA was j
established to be the Double Ended Pump Suction (DEPS) Break with maximum safety injection.
j This break establishes relatively high temperatures very quickly into the event, and sustains temperatures at approximately 260*F throughout the period ofinterest for this analysis. A review 2
of the MSLB breaks performed for the same study demonstrated that the 0.942 ft split rupture is bounding with respect to other break sizes analyzed. A comparison of the DEPS temperature profile and the MSLB profile shows that the DEPS profile reaches higher temperatures earlier, but the MSLB profile passes the DEPS over time, reaching a higher peak temperature. Both breaks were analyzed to ensure that the maximum void predictions would be obtained.
Useneru grgw*mESUP DOO.
GL 96JM 120 DAY RESPONSE d
Development of Piping Forces In order to assess the effects of the hydraulic transients calculated for the RCFC coils during the restart of the SX flow in conjunction with a LOOP /LOCA scenario, it is necessary to derive the force time history that would result. Methodologies have been previously demonstrated to provide the force time histories associated with hydraulic transients calculated using the RELAP5 computer code. These methodologies typically employ a post processing code such as R5 FORCE or REFORC to operate on the RELAPS data and develop the forces on piping segments. A method has been developed using the XMGR software package to perform a similar calculation.
l These loads were then utilized in structural evaluations of the systems.
Two Phase Flow l
The calculations demonstrate that under the limiting assumptions applied, voiding will occur in the 1
l RCFC coils prior to restart of the pump. The presence of the voids leads to forces on the l
structures as the voids are compressed and fluid accelerates following pump restart. Key points l
that have been determined include:
l The forces on the system are relatively small and do not cause the equipment to exceed e
allowable stresses Conditions leading to rapid void collapse and resultant waterhammer do not exist for this l
configuration l
l Additionally, due to the location of the throttle valves on the exit side of the RCFC combined with the SX pump capability and the geometry of the RCFC piping, the potential for flow stall due to increased two phase pressure drop is considered highly unlikely.
i l
l i
kngerarcs t'p960996065UP.!XO GL 964)6 120 DAY RESPONSE s
1
c Conclusion The results of the above analysis have concluded that the resultant loads are relatively small, demonstrating that while there is some additional structural loading due to the two phase transient, the effect is relatively small. The piping stresses, support and structural loads, and equipment loads that result from this transient are within design basis limits. This conclusion, combined with the lack of two phase flow problems associated with the above scenario, completes the response on this issue.
RESPONSE TO ITEM 2:
Potential for Overpressurization of Isolatable Piping Sections The review performed to address thermally induced pressurization resulting from the Loss of Coolant Accident / Main Steam Line Break (LOCA/MSLB) was described in the Byron attachment to the response dated January 28,1997. The response described the screening evaluation performed to identify the containment penetrations susceptible to thermally induced overpressurization.
The penetrations that were identified in the January 28th response as susceptible to overpressurization are listed below-P-5, P-6, P-8, and P Containment Chilled Water system P Reactor Coolant Drain Tank Pump Discharge P Containment Demineralized Water Supply P Fuel Pool Cooling Return to Refueling Cavity P Containment Fire Protection Supply P Reactor Coolant System Loop Fill Header P Primary Water Supply to Reactor Coolant Pump Seal #3 and Pressurizer Relief Tank P Containment Floor Drain Sump Pump Discharge P Safety Injection Accumulator Fill Line P Fuel Pool Cooling Suction from Refueling Cavity Based on later versions of the screening calculations, detailed evaluation of containment penetration pressure conditions, and component availability for the orderly recovery from a MSLB evee, the following penetrations have been added to the list of penetrations for which thermally induced pressure conditions will be addressed:
P Process Sampling (Containment isolation valves -
normally closed during plant operation).
P Chemical and Volume Control System Normal Charging Flow Path (Containment isolation valves -
close on a containment isolation signal).
k$ generic \\grgl%066UP. DOC \\
GL 96-06 120 DAY RESPONSE
e P Component Cooling Water Return from the Reactor Coolant Pump Thermal Barrier Heat Exchanger (Containment isolation valves - close on a containment phase B isolation signal).
l 1
On each side of the containment wall, these penetrations generally have a containment isolation valve which is either closed during normal plant operation (P-30, P-32, P-37, P-55, P-57 and P-
- 70) or automatically closes on a containment isolation signal (P-5, P-6, P-8, P-10, P-11, P-24, P-34, P-44, P-47 and P-71). These penetrations could be potentially heated during a LOCA or a MSLB inside containment, either of which would provide the containment isolation signal.
The overpressure mitigation methodologies presently planned for penetrations susceptible to overpressurization fall into four categories as listed below:
A. Design changes are required to mitigate thermally induced overpressure conditions - 11 penetrations (P-5, P-6, P-8, P-10, P-11, P-24, P-34, P-37, P-44, P-47, and P-55).
B. Overpressure conditions that are intended to be mitigated by procedural changes for valve lineups - one penetration (P-30).
C. Overpressure conditions that are intended to be mitigated by penetration draining - two penetrations (P-32, P-57).
D. A more detailed analysis is warranted to determine if any over pressure concern actually exists - two penetrations (P-70, P-71).
The design changes required for Category A items listed above are being implemented through the Byron Station design change procedures and processes. The design change process insures that all aspects of materials, installation, system operation and system performance / interaction are properly considered and factored into the required design changes. Preliminary design activities are presently in process for these changes. Identified overpressure conditions will be mitigated by the installation of spring loaded relief valves, check valve arrangements or modifications to containment isolation valves. Installation of these design changes is presently anticipated at Byron during outages BIR09 (Spring 1999) and B2R08 (Fall 1999). These outages represent the earliest scheduled refueling outages where it is practical to properly evaluate and install design changes of this magnitude.
l 1
k3gewcigrgIwwwihSUP.fXr\\
GL 964)6 120 DAY RESPONSE
1 l
Category B and C items are currently being reviewed and, if confirmed appropriate, mitigating l
actions will be implemented no later than the unit return to operation from Byron outages BIR08 l
(Fall 1997) and B2R07 (Spring 1998). If the reviews being conducted determine that the proposed actions are not appropriate, design changes will be implemented to mitigate thermally induced overpressure conditions utilizing the methods and schedules described above for Category Aitems.
Category D items are being analyzed in detail (beyond the original screening process that conservatively enveloped many conditions and circumstances) to determine if there is an actual overpressure condition and the magnitude of such a condition, ifit exists. These further analyses i
are expected to be completed such that, if any design changes are required to mitigate thermally induced overpressure conditions, they will be able to utilize the methods and schedules described above for Category A items.
j A basis for operability has been determined for the affected penetrations and included consideration of one or more of the following: expansion of the trapped fluid in voided areas of the isolated piping section; leakage from one of the containment isolation valves (seat, packing, or body to bonnet flange); existing insulation of piping inside containment or heat dissipation from piping outside containment to prevent the temperature induced pressure increase; lifting of the air operated valves due to the pressure increase and plastic straining of the affected pipe to accommodate the pressure increase.
This basis for operability provides the justification for continued unit operation as well as unit restart from refueling and other outages until the appropriate design changes are installed.
Additionally, Comed is actively working with EPRI, BWROG, and NEI in determining the best options with respect to resolution of the concerns identified in GL 96-06. The outcome of the ongoing efforts by these industry groups will be used in adjusting, as necessary, the long-term plans for resolution of this issue at Byron Station.
l j
l I
l l
I k igenerOgrgtMW%0hsUP DOC \\
GL 9M)6 120 DAY RESPONSE i
v
Quad Cites Station NRC Docket 50-254 and 50-265 Response to NRC Generic Letter 96-06,
" Assurance of Equipment Operability and Containment Integrity l
During Design-Basis Accident Conditions," dated September 30,1996 NRC REQUEST:
Within 120 days of the date of this generic letter, addressees are requested to submit a I
written summary report stating actions taken in response to the requested actions noted above, conclusions that were reached relative to susceptibilityfor waterhammer and two-phaseflow in the contamment air cooler cooling water system and overpressuri:atrion of j
l piping thatpenetrates containment, the basisfor continued operability ofaffected i
systems and components as applicable, and corrective actions that were implemented or are planned to be implemented. Ifsystems werefound to be susceptible to the conditions l
that are discussedin this generic letter, identify the systems affectedanddescribe the i
specific circumstances involved.
I i
l NRC REQUESTED ACTIONS:
1
" Addressees are requested to determine:
l ifcontainment air cooler cooling water systems are susceptible to either water hammer or two-phaseflow conditions during postulated accident conditions:
ifpiping systems thatpenetrate the containment are susceptible to thermal expansion of fluidso that overpressuri:ation ofpiping could occur. "
RESPONSE TO ITEM 1:
The Quad Cities response to the water hammer and two phase ilow issues dated January 28,1997 explained why the Reactor Building Closed Cooling Water (RBCCW) piping inside the Drywell will not develop any significant voids due to the post accident Drywell environment. The response also indicated that the water hammer and two phase flow issues identified in NRC Generic Letter 96-06 are not significant for Quad Cities Units 1 and 2. In subsequent telephone conversations on March 6,1997 the NRC restated their concerns related to reestablishing RBCCW flow to the Drywell following an accident.
k.\\senme\\gr Ne06\\96065UP1xxT GL 9606 120 DAY RESPONSE a
. -... -... -.. - = - -.
i A
{
l The RBCCW containment isolation valves do not receive any autornatic isolation signals. Under specific accident conditions, the RBCCW Pumps will trip. By procedure, the containment isolation valves would be isolated by operator action if conditions indicative of a RBCCW break inside the containment exist. It is unlikely that an operator would attempt to reestablish RBCCW flow to the Drywell if the system had been previously isolated due to a suspected line break, 4
I however existing procedures do not preclude such action. Although RBCCW is not safety related, the ability to use the system following a LOCA is desirable. The benefits ofinitiating the j
RBCCW system post LOCA outweigh the low potential for water hammer in the system. If indications of significant RBCCW system damage (i.e. Iow pressure, low expansion tank level)
I exist following reinitiation of flow, the operators would isolate the system as noted above. To l
ensure Operations personnel are aware of the potential for water hammer when reinitiating flow following a LOCA, an appropriate discussion will be added to applicable operating procedures.
1
)
j I
l NEN$ jMM\\QQf{
GL % 120 DAY RESPONSE I
-- u---
2
- +
m, v
R
l l
Response to item 2 ii A summary of the reviews performed to address the post LOCA thermal pressurization issue 1
j identified in NRC Generic Letter 96-06 was included in the Quad Cities response dated January 28,1997. The systems reviewed, scope of the review, susceptible piping sections, basis of operability, long term resolutions under consideration and the implementation schedule are 1
discussed below:
j i
j A. Thermally induced over pressunzation is applicable to liquid filled piping systems which penetrate containment. The following systems were reviewed.
Core Spray (CS) Piping System l
Control Rod Drive (CRD) Hydraulic Piping System i
l Demineralized Water Piping System High Pressure Coolant Injection (HPCI) Piping System
)
Reactor Core Isolation Cooling (RCIC) Piping System
}
Reactor Building Equipment Drain System i
Reactor Feed Piping System i
Recirculation Piping System Reactor Water Clean Up (RWCU) Piping System Residual Heat Removal (RHR) Piping System Standby Liquid Control (SBLC) Piping System I
l Reactor Building Cooling Water (RBCCW) Water Piping System i
j B. A review was performed to identify the piping sections for which such pressurization could j
^
jeopardize the ability of accident mitigating systems to perform their safety functions or could l
lead to breach of containment or bypass leakage. The portions of the systems that are not j
susceptible to thermally induced pressurization due to system configuration, valve lineup, i
check valves, relief valves, and piping open to the reactor were excluded from further review.
i k
4 I
t
]
l 4
i f
i ij 1
I k%enericWgl9f4EWeSUP.Doo GL 964)6 120 DAY RESPONSE 2
e C. Eight penetrations were determined to be susceptible to thermally induced pressurization conditions discussed in Generic Letter 96-06. The penetrations and associated descriptions are identified below:
j Penetration Description Number X-24 Reactor Building Closed Cooling Water (RBCCW) Outlet From Drywell X-23*
Reactor Building Closed Cooling Water (RBCCW) Inlet To Drywell X-41 Reactor Recirculation System Sample Line Piping
)
X-12 Residual Heat Removal (RHR) Shutdown Cooling (SDC) Suction Piping X-14 Reactor Water Clean-up(RWCU) Suction Piping X-20 Clean Demineralized Water Piping to the Drywell X-18
Drywell Equipment Drain Sump (DWEDS) Discharge Piping
- Not identified in original response. Added based on further review and evaluation; The potentially isolable piping sections have been determined to be operable, but are considered degraded. A summary of the basis for operability and actions taken or planned to mitigate or i
eliminate the potential for overpressurization are also discussed below.
k:\\geneneigtst9606%)65UP.txo GL 96 M 120 DAY RESPONSE
a Penetration X-24, Reactor Building Closed Cooling Water (RBCCW) Outlet From Drywell Discussion:
This penetration involves the RBCCW Outlet piping from the Drywell. The primary reason that RBCCW to the Drywell would be isolated during a LOCA/MSLB would be a RBCCW break inside containment. QCOA 3700-06, RBCCW Line Break Inside Containment directs closure of the isolation valves in that situation. The RBCCW pumps trip on a LOCA signal but the isolation valves require operator action to close. Due to the piping orientation, it is possible that the volume between the MO 3703 and MO 3706 could be water solid following a RBCCW break i
inside the Drywell, therefore the is a potential that thermal pressurization could occur. There is no post accident requirement to reopen these valves following a DBA.
i Nuclear Design Information Transmittal (NDIT) MSD-97-002, Rev.1, " Evaluation of Pipmg under Water Solid Conditions and LOCA Temperatures (GL96-06)", provides the results of an evaluation of the pressurization resulting from heating a water solid pipe. The calculation summarized, Calculation DRE97-0001, Rev.1, shows that the permanent strain in the piping is on the order of 6% (maximum). The bases for that calculation are three very conservative assumptions:
no leakage through the isolation valves exists e
properties of pure water, with no uncondensibles or trapped air, are used e
the worst possible temperature difference (cold to hot)is used AND
- hot temperature is e
permitted to exist for a sufficient length of time to heat the water to that temperature Even with these conservative assumptions, the maximum strain is well within the ultimate strains for the piping materials used at Quad Cities. Thus, the pressure boundary is acceptable.
Calculation DRE97-0001 shows that the permanent strain is on the order of 6% even assuming the end connections (valves) do not leak at all. This is well within ultimate strain for typical piping materials. Thus, the piping pressure boundary is acceptable.
i 1
)
l k:\\generica 4 rammUP. Don GL 96-06 120 DAY RESPONSE rN
Valves are typically as thick or thicker than the mating pipe. Thus, the external pressure boundary of the valve would see strains on the same order as the pipe, again assuming no valve leakage.
Valves, however, willleak at the pressures which cause yield in the piping. The MO 3703 & MO 3706 valves at Quad Cities are of the bolted bonnet variety. The valves have a 150 pound pressure class rating. As the pressure increases beyond the design rating, the bolts will experience additional strain. These bolts will begin to elongate and allow for the internal pipe pressure to be reduced by leaking water from inside the pipe through the packing and/or bonnet. If the inboard valve bolted bonnet is the weaker link, then primary containment would be maintained by the piping and valve outside of containment. The reverse happens if the outboard valve is the weaker link. No matter where the postulated leakage occurs, primary containment is maintained. If one of the valves did not close or seat properly for other reasons, pressurization will not occur and the other valve will provide for primary containment. The valve stem packing will hold against only a certain amount ofinternal pressure. As the internal pressure increases, the packing may begin to leak, thereby reducing the internal pressure. Either the inboard or outboard valve stem packing will be the weaker link and the other valve will maintain primary containment.
Additionally, a calculation has been performed which indicates that based on recent LLRT measurements, thermal overpressurization of this penetration is not expected due to existing valve leakage.
On the above basis, this penetration will continue to maintain pressure under accident conditions and will remain operable.
Corrective Actions:
A design change is in progress which will provide a relief valve for this volume. Installation of this change for Unit 2 will be completed during the current refueling outage, Q2R14.
A similar modification is planned for installation during the next Unit 1 Refueling Outage, QlR15 which is currently scheduled to begin in Spring,1998.
I l
l l
I k4generrigrgf%wwmSUP.D(n GL 9646 120 DAY RESPONSE
- ~
Penetration X-23, Reactor Building Closed Cooling Water (RBCCW) Inlet To Drywell This penetration involves the RBCCW Supply piping to the Drywell. The valve configuration associated with this penetrations consists of a motor operated gate valve outside containment and a check valve inside containment. Pressure buildup in this volume would be limited by expansion of the copper heat exchanger tubes in the Drywell coolers or valve leakage. The Drywell Cooling system is not required to operate to mitigate a design basis accident.
Additionally, RBCCW to the Drywell would not be isolated during a accident unless conditions indicative of a RBCCW break inside containment existed. In such a situation, operating procedures direct closure of the RBCCW isolation valves. Under specific accident conditions, the RBCCW pumps will trip automatically, but the isolation valves require operator action to close.
If the valves were isolated and a leak inside containment existed between inlet and outlet isolation l
valves, any pressure buildup associated with penetration X-23 would be relieved through the inboard check valve through the leak.
1 Corrective Actions:
l A design change is in progress which will provide a relief valve for this volume. Installation of this change for Unit 2 will be completed during the current refueling outage, Q2R14.
j i
A similar modification is planned for installation during the next Unit 1 Refueling Outage, QlR15 l
which is currently scheduled to begin in Spring,1998.
l Penetration X-41, Reactor Recirculation Sample Line Piping l
This penetration is bounded by the Recirc system sample valves AO 0220-44 and AO 0220-45.
One valve is inside containment and the second valve is outside containment. During normal plant operation, the isolation valves are open and the line has continuous flow. As a result the piping will be normally hot. These valves would be expected to close due to a Group 1 Isolation signal during a LOCA, however thermal pressurization of the isolated volume would not be expected due to its higher initial temperature.
If the line is isolated due to any reason prior to a LOCA, the containment isolation valves A0 220-44 and 45 and piping between them will cool down and the isolated volume may be susceptible to thermally induced pressurization. However, the isolation valves are air operated.
J As pressure increases between the valves, the air operator spring tension that keeps the valves l
closed will be overcome. As pressure increases, seat leakage will relieve the pressure buildup.
I On the above basis, this penetration is operable.
Corrective Actions:
A design change is in progress which will provide a relief valve for this volume. Installation of this change for Unit 2 is scheduled for the current refueling outage, Q2R14. The Station is expediting procurement of a qualified relief valve to support this installation period, however the current expected valve delivery date is beyond the projected outage completion date. Because of 4
this uncertainty regarding the availability of a qualified valve, the design change has incorporated k:\\generegnglww%MUP.txn GL 96-06 120 DAY RESPONSE
L provisions to install the necessary piping (i.e. blanked flange) such that the relief valve can be l
installed at the next outage of sufficient duration following receipt of the valve.
A similar modification is planned for installation during the next Unit 1 Refueling Outage, QlR15 -
i j
which is currently scheduled to begin in Spring,1998.
j l
Penetration X-12, RHR Shutdown Cooling Suction Piping i
This penetration includes the RHR Shutdown Cooling Suction piping and is bounded by the 1001-l 47 and 1001-50 valves. One valve is inside containment and the second valve is outside containment. This boundary is isolated prior to start-up at a relatively cold condition. Review of QCOP 1000-5, Shutdown Cooling Operation, indicated that the volume would be expected to be l
water solid when the system is shutdown in preparation for ' Unit Start-up, therefore the potential l
of thermal pressurization is a concern.
1 i
The discussion associated with RBCCW is also applicable to this penetration. There is no j
requirement to reopen these valves following a LOCA. A calculation has been performed which indicates that based on recent LLRT measurements, thermal pressurization of the Unit i volume is not expected due to existing valve leakage. The Unit 2 volume had minimal LLRT leakage and a similar calculation indicates that the volume may experience thermal overpressurization, however j
based on the discussion presented in NDIT Number MSD-97-002, Evaluation of Piping Under Water Solid Conditions and LOCA Temperatures (GL 96-06), the penetration is' operable.
Corrective Actions:
A design change is in progress which will provide a relief valve for this volume. Installation of this change for Unit 2 will be completed during the current refueling outage, Q2R14.
A similar modification is planned for installation during the next Unit 1 Refueling Outage, QlR15 which is currently scheduled to begin in Spring,1998. Administrative procedures have been implemented to partially drain the Unit 1 Shutdown Cooling Suction piping to mitigate the potential for overpressurization pending installation of a design change.
l l
l 1
)
i l
Lgmu4Bst9emo606sur Don GL %06 120 DAY RESPONSE T
-.. +. -
l w
Penetration X-14, Reactor Water Clean-up Suction Piping The only section of the reactor water clean-up piping that may be susceptible to thermally induced pressurization due to post-LOCA heat-up is the containment penetration piping associated with valves MO 1201-2 and MO 1201-5. One valve is inside containment and the second valve is outside containment. During normal plant operation the isolation valves are open and the line has continuous flow. As a result the piping will be normally hot. If the isolation valves close due to a 1
LOCA, the temperature of the isolated line would not be expected to increase due to its higher initial temperature and therefore will not be subject to thermally induced pressurization.
If the line is isolated due to any reason prior to the LOCA, the containment isolation valves and associated piping will be cooler and the isolated volume may be susceptible to thermally induced pressurization. The reactor water clean-up piping inside Drywell is insulated. The rate at which the water temperature inside the isolated pipe volume increases due to the post LOCA Drywell environment is significantly reduced by the insulation on the pipe. As a result, the rate at which the water inside the isolated volume expands will also be significantly reduced by the insulation and the additional volume of water created by expansion would leak past the penetration isolation i
valves.
A calculation has been performed which indicates that based on the most recent LLRT measurements, thermal overpressurization is not expected due to existing valve leakage. The operability discussions associated with RBCCW Drywell Outlet piping are also applicable to this penetrationL There is no post accident requirement to reopen these valves following a DBA. On the above basis, the penetration is operable.
Corrective Actions:
A design change is in progress which will provide a relief valve for this volume. Installation of this change for Unit 2 will be completed during the current refueling outage, Q2R14.
A similar modification is planned for installation during the next Unit 1 Refueling Outage, QlR15 L
which is curremly scheduled to begin in Spring,1998.
l 4
u enue nvumwoisur.txc GL 96-06 120 DAY RESPONSE s
e
l l
Penetration X-20, Clean Demineralized Piping This volume includes the piping associated with the Clean Demineralized Water Service drop inside the Drywell. Calculations performed in response to Generic Letter 96-06 indicated the potential for thennal pressurization of this piping. This piping is provided for convenience during outage work inside the Drywell and is not used during operation. The operability discussions presented for RBCCW penetration X-24 are also applicable to this piping. The valves associated with this penetration are considered to be the weak link in this isolated volume. If valve seat leakage does not relieve thermal pressurization, as pressure increases, leakage will develop leakage through the weaker valve via the body / bonnet interface or the valve packing. The otlier valve will remain intact and ensure containment integrity. There is no post accident requirement to open these valves following a DBA.
l l
Corrective Actions:
l This piping will be drained to prevent thermal pressurization. Operating procedures have been revised to require that this volume to be drained prior to start-up following every refueling l
outage. The volume will also be drained prior to start-up from non-refueling outages if the clean demineralized water piping associated with penetration X-20 has been valved into service. The Unit I piping associated with this volume has been drained. The Unit 2 volume will be drained l
prior to start-up from the current refueling outage.
l l
1 kn ewie4r mw9esuraxo GL 96 )6 120 DAY RESPONSE s
a 4
Penetration X-18, Drywell Floor Drain Sump Discharge Piping
)
Penetration X-19, Drywell Equipment Drain Sump Discharge Piping Penetrations X-18 and X 19 are subject to thermal pressurization on both Units. The containment isolation valves associated with these penetrations are both located outside of containment.
Following a LOCA, the piping and associated valves from the sump pump discharge check valves to the isolation valves could experience stresses beyond design allowables. The sump pump check valves are not containment isolation valves and are not leak rate tested. A review of maintenance i
history determined that the check valves are routinely replaced every other refueling outage. Due to the piping configuration and the quality of the sump water, debris tends to collect around the check valve seats making a leak-tight boundary unlikely. In addition to the sump discharge check valves, each volume has additional leakage paths available through other valves making up the isolated volume. Based on the multiple leakage paths available and the small leakage rate needed to mitigate the thermal pressurization, there is reasonable assurance that significant pressurization of the 6p: Ag will not occur.
In the event the leakage paths are not sufficient to prevent pressurization, the piping and l
associated components would be expected to maintain integrity at pressures significantly above their design ratings, however they may experience deformation and leak. Catastrophic failure of the piping coir penents would not be expected. If the piping on components in the Drywell leak, primary containroent is unaffected. If the leakage occurs outside the Drywell, the size of the leakage path needed to relieve the pressure would be expected to be small enough such that containment leakage requirements would not be exceeded. Therefore based on these discussions, it is reasonable to conclude that penetrations X-18 and X-19 are operable and primary containment will perform as required.
Corrective Actions:
During the initial review performed in response to GL 96-06, these penetrations were not identified as susceptible to thermal pressurization. During a subsequent internal review, errors in the original evaluation were identified. The corrective actions to address these penetrations have not been finalized. Quad Cities will provide overpressurization protection for these penetrations.
The primary options being considered are include use of a relief valve or drilling a small continuous hole in the sump pump discharge check valve to provide a continuos relief path for the discharge piping. It is expected that these actions will be completed on Unit 2 by the end of the current refueling outage.
Corrective actions for the Unit 1 penetrations will be completed during the next Unit 1 Refuelig Outage. Q1D 15 which is currently scheduled to begin in Spring,1998, l
k:c=es sn wmwimsurnon GL W6 120 DAY RESPONSE l
i
i i
l j
Zion Station NRC Dockets 50-295 and 50-304 Response to NRC Generic Letter 96-06,
" Assurance of Equipment Operability and Containment Integrity During Design-Basis Accident Conditions," dated September 30,1996 1
NRC REQUEST:
i 5
Within 120 days of the date of this generic letter, addressees are requested to submit a written summary report stating actions taken in response to the requested actions noted 4
above, conclusions that were reached relative to susceptibilityfor waterhamirer and two-phaseflow in the containment air cooler cooling water system and overpressu;i:ation of piping thatpenetrates containment, the basisfor continuedoperability ofaffectedsystems and components as applicable, and corrective actions that were implemented or are planned to be implemented. Ifsystems werefound to be susceptible to the conditions that are discussedin this generic letter, identify the systems affectedanddescribe the specific circumstances involved.
NRC REQUESTED ACTIONS:
" Addressees are requested to determine:
Ifcontainment air cooler cooling water systems are susceptible to either water hammer or two-phaseflow conditions duringpostulated accident conditions:
ifpiping systems thatpenetrate the containment are susceptible to thermal expansion of s uid so that overpressuri:ation ofpiping could occur. "
RESPONSE TO ITEM 1:
Potential Water llammer Transient Analyses are complete which evaluate the potential for water hammer in the Zion Station service water (SW) system supply and return piping for the Reactor Containment Fan Coolers (RCFCs) (Ref.1). Specifically (+ing evaluated are the effects of a loss of offsite power (LOOP) occurring simuhaneously with a large break loss of coolant accident (LOCA) or a Main Steam line break (MSLB) inside containment.
kngenerrigf#W6W45UP.txo GL 9MM 120 DAY RESPONSE 8-
~
The analyses determined that a potential for steam void formation in the RCFCs exists and that the large break LOCA condi1 ions govern. When the coils of the RCFCs are exposed to the atmosphere inside containment following a large break LOCA, the water in the coils of the RCFCs is subject to boiling once the SW pumps stop following a LOOP and the pressure is reduced. The steam void that results from the boiling has the potential to j
cause water hammer in the supply and return piping to the fan coolers under two conditions. First, prior to restart of the SW pumps, as the boiling occurs, water hammer can result from condensation induced water hammer. Then, once power is supplied from the diesel generators and the SW pumps are restarted, water hammer can also result from impact of water columns as the piping refills and the steam voids collapse.
Review Approach
_Syst_em Configuration and Analysis The system configaration considered bounding is, a LOOP /LOCA occurance while one of the units is in cold shutdown and the other unit is operatmg. In this case, the non-LOCA unit's SW pumps may not be available due to unit status. Instead of having two of the SW pumps start in approximately 15 seconds from the Blackout sequence. timer on the non-LOCA unit, three SW pumps would start frem the Safety Injection (SI) sequence timer on the unit with the LOCA after approximately 30 to 36 seconds. The three SW pump restart would provide higher flow ari create greater water hammer loads than a two pump restart. Also, the time delay before the pumps start would allow for maximum steam void creation. This configuration was conservatively used to determine the water hammer loads in the piping servicing the RCFCs.
The start. time of the RCFC fans was ignored for analysis purposes. The analysis assumed full condensation heat tr.nsfer from an infinite source. Thus, the rate of fan coastdown and startup do not arTect the results of this analysis. The control valves which close off loads to the turbine building, booster pumps, traveling screens, and fire protection system were assumed to close as quickly as possible. This assumption conservatively maximized flow to the RCFCs when the SW pumps restart. The high end of the range of pump restart times was used. This allows for maximum steam void creation in the RCFC coils.
1 It is possible that one of the valves to the turbine building loads may be closed off from the shutdown unit. While there is crosstie piping between the turbine building piping which allows flow from one unit's SW pumps to supply both unit's turbine building loads, it is assumed that only half the design flow will be sent to the turbine building loads. This will j
cor.servatively increase the flow sent to the RCFCs.
The hydraulic transient analysis was performed using Sargent & Lundy Program HYTRAN (Ref. 2).
)
t yminenewsur txe GL %06 120 DAY RESPONSE
1 i
. Heat Transfer Consequences of the LOCA Event During the LOOP /LOCA wenario, cooling water flow to the RCFCs is lost as the service l
water pumps trip. Increased heat transfer to the RCFCs due to condensation of steam in the post LOCA containment atmosphere, coupled with loss of coolant flow, results in boiling of the water in the RCFCs. A RELAP (Ref. 3) model was developed to model heat transfer to the RCFCs during the LOOP /LOCA in order to predict the size and location of voids prior to pump restart (Ref. 4).
i The LOCA transient is deemed to be more limiting than a Main Steam Line Break (MSLB) with respect to heat addition to the RCFC coils because substantially more steam j
is released to contair. ment during the LOCA than for the MSLB during the time ofinterest in this analysis. This results in a higher steam / air mass ratio for a LOCA than for a MSLB, with corresponding higher air side condensation heat transfer coefficients.
Water hammer from potential condensation induced water hammer and the resulting pipe loads are predicted to be less than those resulting from rejoining of the water columns following restart of the SW pumps (Ref.1). Therefore column rejoining water hammer loads were used in operability evaluations of the RCFCs and the associated supply and return piping and restraints.
Review Results The analyses have shown that column rejoining water hammer may occur, which results in potentially significant piping loads. The effects of these loads are evaluated by analytically applying them to the supply and return piping of the RCFC located at the highest elevation and having the longest pipe lengths inside containment (IRV003). This RCFC and associated piping is expected to experience the most severe water hammer loads.
Calculations demonstrate (Ref.1) that the piping and RCFCs are operable following a LOOP /LOCA. There are a few instances where the pipe stresses exceed design allowables; however, all of the piping stresses are less thar. the ASME Class 2/3 faulted allowable of 2.4Sn which is more restrictive than Appendix F of the ASME Code.
Similarly, there are a few instances where the pipe support stresses exceed design allowables; however, all of the pipe support stresses are within the limits specified in Appendix F of the ASME Code.
These results indicate that operability is demonstrated (pressure retaining boundary, structural stability and required flow area) based on the fact that the calculated pipe stresses are within operability limits. Zion Operability Assessment ER9604802 documents the basis for continued operability of the service water piping.
k% =re w m m esuP D C GL %C6 120 DAY RESPONSE
References i
1.
Sargent & Lundy Calculation 22S-B-022M-597, Rev. O.
2.
HYTRAN, Sargent & Lundy Program No. 03.7.263-1.4, Released February, 1997.
3.
RELAP5/ MOD 3.1, Sargent & Lundy Program No. 03.7.459-1.1.
4.
Sargent & Lundy Calculation 22S-B-022M-598, Rev. O.
i h
i E
1 e
J i'
L i
k:vansw'emamsor. Don GL %4M 120 DAY RESPONSE
Potential for Two-Phase Flow A calculation has been completed using the RELAP model developed to analyze the initial part of the LOOP /LOCA transient, modified as necessary to determine the steady-state 1
conditions following pump restart (Ref.1). Steady-state heat transfer to the RCFC coils and containment piping was modeled.
Review Approach Once the service water piping refills upon pump restart and. the RCFC fans restart, the j
RCFC cooling water exit temperature will be maintained at an elevated level due to the high condensation heat transfer rates. This condition could result in a two phase flow j
condition downstream of the RCFCs at the throttling devices. Frictional pressure drop in the return piping would increase, either because ofincreased two phase hydraulic resistance, and/or flashing at high resistive components (orifices and throttle valves) due to the increased vapor pressure of the cooling water. With an input pressure and temperature in the supply piping, the RELAP model calculates the temperature and flow rate of water through the RCFC units and return piping to the 20 inch collection manifold outside of containment.
Review Results The results of the evaluation demonstrate that two-phase flow conditions will not occur.
Based on the RCFC outlet pressures from the various SW pump start scenarios contained in the flow model (Ref. 2), the minimum pressure downstream of the thrattle valves and -
orifice, where the RCFC discharge water mixes with other streams, is approximately 4 psig. The maximum temperature of water leaving the RCFC units and exiting the heated containment piping is less than the saturation temperature of water at this pressure. This ensures that two phase flow will not occur.
References 1.
Sargent & Lundy Calculation 22S-B-022M-598, Rev. O.
2.
Comed Calculation 22S B-0220 525, Rev.1.
e mervr twswresor.Doci GL *MM 120 0 AY RESPONSE s
e I
PLANNED CORRECTIVE ACTIONS FOR ITEM 1 For the potential water hammer in the SW piping as a result of voids created by a LOOP /LOCA, two options are under investigation. The first option is to modify the support configuration as necessary to show both the piping and supports meet the design limits. Based on the results of the calculations used to demonstrate operability, it is anticipated that the system can be qualified by analysis without significantly changing the support arrangement.
The second option is to install modifications to ensure a slow refill of the system and significantly reduce the water hammer loads in the system. The reduced water hammer loads will be analyzed as described above, but it is anticipated that hardware changes would be minimal depending on the design of the slow fill system.
Zion Station will complete the analysis and install the design changes necessary during the current Unit 1 outage (ZlR15) and during the next Unit 2 outage (Z2R15).
No further action is required to address the potential of two-phase flow occurring once flow has been established to the RCFCs. The calculations demonstrate that two-phase flow will not occur and the single-phase flow assumption used in the Service Water Flow Model is acceptable.
i l
l l
4 k.w me r,siwwwnswt.ocs GL %06 120 DAY RESPONSE e
l RESPONSE TO ITEM 2:
1 A summary of the review performed to address the post accident thermally induced pressurization j
issue identified in NRC GL 96-06 was included in the Zion response dated January 28,1997.. The
')
systems reviewed, scope of the review, susceptible piping sections, basis of operability, long term
. resolutions and schedule are discussed below; l
i Review Approach j
Thermally-induced over pressurization is applicable to liquid-filled piping systems which penetrate containment. The following systems were reviewed:
Fire Protection Waste Disposal Liquid Blowdown l.
Waste Disposal ' Auxiliary Building Drain j
Reactor Coolant Chemical and Volume Control Residual Heat Removal r
Safety Injection Demineralized Flushing Water Primary Process Sampling Heating System t
Waste Drain A review was performed to identify the liquid-filled piping sections for which such pressurization j
could jeopardize the ability of accident mitigating systems to perform their safety functions or could lead to breach of containment or bypass leakage. The poitions of the systems that are not susceptible to thermally-induced pressurization due to syster' configuration (e.g. valve line-up, check valves, relief valves) were excluded from further review.
t k
i
[
d
- j.
i t
kagenseicgn tweenossur.mc GL 96 06 120 DAY RESPONSE e
5 s
Review Results Nine liquid-filled systems were found to be susceptible to thermally-induced pressurization conditions discussed in Generic Letter 96-06. The penetrations and systems are as follows:
Penetration No.
System P-1 Fire Protection P-34 Demineralized Water P-43 Waste Drain P-88/89 Station Heating l
P-102 Primary Water
)
P-76 Accumulator Test Line P-76/86 Sample System P-76 RCS Loop Fill P-71/91/94/115/118 S/G Blowdown For these piping sections, operability of the piping was demonstrated by showing that the permanent strain developed would be on the order of 2% to 3%, but no higher than 6% in the worst case, which is well below the ultimate strain for the material (Ref.1).
At these locations, the thermally-induced pressure increase is greater than the design pressure of the containment isolation valves. These valves are considered operable and capable of performing their safety function of providing containment isolation and maintaining structural integrity. This determination is based on the position that the valves can also withstand these higher pressures.
In addition, high pressure may cause leakage through the bonnet or out to the lower pressure piping through the disc / seat. The gradual pressure build-up that would occur would be relieved l
by this leakage. Calculations (Ref. 2) have been completed to determine the pressure capability of l
valve models typical of some containment isolation valves (CIVs) installed on the systems which l
may be subject to overpressure. These calculation show 1) when air operated valves (AOVs) are pressurized under the seat, the plug will lift ofTthe seat at pressures much less than the overpressure calculated assuming perfect valve isolation and,2) valves can withstand significantly higher pressures than the ANSI rating.
l l
None of the systems subject to overpressure perform an active post-accident cafety-related I
function and the ability of the boundary valves to open is not a safety issue. Zicn Operability Assessment ER9606789 documents the basis for continued operability of these alTected penetrations and systems / components.
k VeurungrglWEWESUP.tvo GL 9M)6 120 DAY RESPONSE
s 1
References:
4
- 1. Comed Calculation DRE97-0001, R.ev.1.
l
- 2. EMS Calculations 22S-B-026M-030,031,032, Rev. O.
N PLANNED CORRECTIVE ACTIONS FOR ITEM 2 Each segment of a piping system that could potentially be affected by post accident thermally-induced pressurization including the specific circumstances involved and the planned corrective actions are given below. For installations involving relief valves, the effects of valve discharges will be evaluated in the design.
1.
The isolable portion of the Fire Protection system at P-1 is between normally closed inboard valves and outboard CIV FCV-FP08 which will close on a Safety Injection signal.
j l
Planned Corrective Action:
)
Install a relief valve inside containment with the discharge of the relief valve directed to the floor.
2.
The isolable portion of the Demineralized Water system at P-34 is between normally closed inboard valves and outboard CIVs DWOO30 and DW0038.
Planned Corrective Action:
Install a relief valve inside containment with the discharge of the relief valve directed to the floor.
3.
The isolable portion of the Waste Drain system at P-43 is between the inboard check valves and outboard CIVs LCV-DT1003 and AOV-DT9170 which close on a Safety Injection signal.
Planned Corrective Action:
Install a relief valve inside containment with the discharge of the relief valve routed to the Reactor Coolant Drain Tank (RCDT) via line DT030-3".
l kv=regr iwwwmsur Exc GL 964 120 DAY RESPONSE a
-..... _ _ _. _ _ _ _. ~
)
i 4.
The isolable portion of the Station Heating system at P-88/89 is between the four normally closed CIVs FCV-RV111,112,113,114; Planned Corrective Action:
3 Install a relief valve inside containment with the discharge of the relief valve directed to the floor.
5.
The isolable portion of the Primary Water system at P-102 is between several normally closed inboard valves and inboard CIV RC8046 (check valve) and potentially between RC8046 and outboard CIVs AOV-RC8028 and AOV-RC8029 which close on a Safety l
Injection signal.
i Planned Corrective Action:
Install a relief valve inside containment, downstream of check valve RC8046 with the discharge of the relief valve directed to the floor.
t 6.
The isolable portion of the accumulator test line at P-76 is between several inboard j
normally closed AOVs and normally closed outboard CIVs AOV-SI8888, SI8961, SIO245.
L i
Planned Corrective Action:
Install a relief valve inside containment with the discharge of the relief valve routed to the PRT via line RC158-4".
c f
7.
The isolable portion of the Primary Sample system at P-76 is between normally closed inboard AOVs at the accumulators and normally closed outboard CIVs AOV-SS9357A and B. There are two isolable portions at P-86. One portion is between several normally j
closed inboard AOVs at the Reactor Coolant System (RCS) hot and cold legs and normally closed outboard CIVs AOV-SS9356A and B. The other portion is between a normally closed inboard AOV at the pressurizer and normally closed outboard CIVs AOV-SS9355 A and B.
i L
Planned Corrective Action:
Install one relief valve for each sample line inside containment with the discharge of the RCS and pressurizer sample line relief valves routed to the PRT via line RC158-4" and the discharge of the accumulator sample line relief valve routed to the trough inside containment.
i 1
l 8.
The isolable portion of the RCS Loop Fill system at P-76 is between several normally closed inboard AOVs at the RCS cold legs and inboard CIV VC8224 (check valve) and j
potentially between VC8224 and normally closed outboard CIVs VC8480A and B.
4 bgmwry,glws96065UP Dro GL 9M)6 120 DAY RESPONSE l
t
7
]
s,,s 9
Planed Corrective Action:
Install a relief valve inside containment, downstream of check valve VC8224 with the discharge of the relief valve routed to the PRT via line RC158-4" 9.
The isolable ponion of the S/G Blowdown system at P-71/91/94/115/118 is between outboard CIVs MOV-BD09 through BD16 and CIV FCV-BD17. Although this system is normally in service and has an elevated normal operating temperature,' it is taken out of service and allowed to cool for maintenance. The out-of-service boundaries would normally include these CIVs, thus isolating the piping inside containment.
Planned Corrective Action:
Install a rupture disc with the discharge routed to an adequately sized chamber. This assembly will be located inside the valve house. This design is being used because the relief device is being installed between CIVs and a failure of a relief valve (stuck open) would cause a breach of containment integrity.- The chamber will include a test tap which can be used to periodically ensure the rupture disc integrity.
I The overpressure protection devices under consideration are classified as devices that are l
intended to mitigate the consequences of a service condition. If the penetration portion of the L
system is normally isolated during plant operation and is subject to overpressurization from j
l normal temperature fluctuations, then the overpsessure event will be considered a normal
~
operation event. If the system is normally in service or is often placed in service during operation (e.g. Sample System, S/G Blowdown, Was'e Drain, Fire Protection), then it is not credible that the overpressure event will occur from r.armal temperature fluctuations. For these systems, the thermally induced overpressure is caused by a LOCA/MSLB and the applicable service condition is the faulted condition.' Consequently, the set pressure chosen to limit the thermally induced l
pressure will correspond to the faulted condition allowable pressure.
L Zion Station will complete the analysis and install the design changes necessary during the current Unit 1 outage (ZlR15) and during the next Unit 2 outage (Z2R15).
i l
i
- I i
i 1
].4 l
4 k4e=reargnee s o ur.r m G1,96-06 120 DAY RESPONSE 1
1
-