ML20249B969
| ML20249B969 | |
| Person / Time | |
|---|---|
| Site: | Brunswick  |
| Issue date: | 06/11/1998 |
| From: | Arterburn J, Mast P, James Smith CAROLINA POWER & LIGHT CO. |
| To: | |
| Shared Package | |
| ML20249B959 | List:
|
| References | |
| IEB-96-003, IEB-96-3, VC44.F02, VC44.F02-R01, VC44.F2, VC44.F2-R1, NUDOCS 9806250117 | |
| Download: ML20249B969 (37) | |
Text
- - _ _ - - - _ - _ - _ _ _ - - _ _ - _ - _ _ _ - - _ _ - _ _ _ _ _ - _
l l
Document No.: VC44.F02 Revision:
1 J
Brunswick Steam Electric Plant, Unit Nos.1 and 2 l
ECCS Suction Strainers Replacement Project NRC Bulletin 96-03 Final Report Prepared by: \\d d /6 983 Date:
Jon R. Arterburn, P.E.
[S!lI !48 Reviewed by:
jfwn Date:
J. Aaros' Smith, P.E.
6[Il!18 Reviewed by:
Date:
Peter K. Mast, Ph.D.
Reviewed by:
M
(/'A Date:
6 // 78 Curtis Varchol, P.E.
[
l i
4 {l 9006250117 990619 PDR ADOCK O G
j
BSEP, Unit Nos. I and 2, ECCSSuction Strainer Replacement Project l
NRCBulletin 96-03 FinalReport i
I i Brunswick Steam Electric Plant, Unit Nos. I and 2 ECCS Suction Strainers Replacement Project NRC Bulletin 96-03 Final Report l
l l
l I
m____._
l l
i BSEP, Unit Nos. I rnd 2, ECCSSuction Striiner Ref'facement Project NRC Bulletin 96-03 FinalReport Table of Contents 1.0 Introduction...............................................................-.........................
................1
' 2.0 Potential ECCS Suction Strainer Blockage Issue Resolution Approach...............................)
2.1 S u c tio n St ra i ne r Re pla ce m e nt......................................................................................... 2 2.2 Suppression Pool and ECCS Strainer Cleanliness............................................................. 3 23FMEProgram......................................................................................................................J 2.4 ECCS Pump Technical Specification Surveillance Data Review.............
....................4 2.5 Primary Containment Coating Monitoring................................................................ 4 l
3.0 Plaat Configuration and Design Parameters...
.. 4 3.1 C o n t ai n m e n t D es ig n.................................................................................................
- 4 3.2ECCSDesign.........................................................................................................5 3 3 D rywell Ins ula tio n......................................................
....................................10 3.4 Other Potential Drywell Debris Sources..........................
..............................10 3.5 Poten tial Wetwell Debris Sou rces............................................................................ I 1 4.0 Replacement Strainer Design Requirements.............
.....................................12 4.1 St rainer Fu netional Req uiremenis..........................................,...................................... 12 4.2 S t ra in e r Cod e Req ui re m en1s.............................................................................. 12 4 3 Strainer St ructu ral Req uire ments.......................................................................... 13 5.0 Replacement Strainer /ECCS Long Term Recirculation Evaluation Methodology...........14 I
J 5.1 Deb ris G ene ratio n...............................................................................
. 14 5.2 D rywell Debris Transport..............................................................
15 53 Wetwell Deb ris Tran spo rt..........................................................................................
... I 5 5.4 Suction Strainer Blockage and Head Loss............................................
....................I6 I
5.5 ECCS Pumps NPSH Margin Assessment..........................................
................ 1 7 5.6 Suction Strainer Structural Evaluntion.......................................
.............................I8 5.7 Other Plant Structural, Systems and Components Evaluation................................... 20
)
i
)
6.0 Li ce ns i n g Co n sid e ra t io us............................................................................................. 2 0 APPENDICES A - Modified NUREG/CR-6224 Head Loss Correlation
)
B - Sure-Flow" Strainer Performance Test Reports C - BSEP Unit Nos. I and 2 Replacement Suction Strainers Design Drawings D - BSEP Unit Nos.1 and 2 Torus Design Drawing E - Strainer Hydrodynamic Mass Test Reports l
l it L
BSEP, Unit Nm. I and 2 ECCSSuction Strainer Replaccinent Project NRC Bulletin 96-03 Final Report 1.0 Introduction This document constitutes Carolina Power and Light (CP&L) Company's Gnal report in response to the NRC Bulletin 96-03," Potential Plugging of Emergency Core Cooling Suction Strainers by Debris in Boiling Water Reactors," dated May 6,1996, for the Brunswick Steam Electric Plant (BSEP), Unit Nos. I and 2. An initial response to Bulletin 96-03 was provided in CP&L Letter, No.BSEP 96-0364, dated November 1,1996. A supplemental response was provided with CP&L Letter, No. BSEP 97-0209, dated June 5,1997. The Gnal report for BSEP, Unit No. 2 was provided with CP&L Letter, No. BSEP 97-0411, dated November 6,1997. This report is a revision to the previously submitted BSEP, Unit No. 2 Final Report and provides additional information identiGed or developed since its original submittal. This revision of the report also provides the Final Report information applicable to BSEP, Unit No. l.
This final report provides the details of: 1) the approach CP&L has implemented to resolve the potential Emergency Core Cooling System (ECCS) suction strainer blockage issue identified in Bulletin 96-03; 2) the criteria CP&L has imposed in the design and evaluation of the replacement strainers; and 3) the methodology CP&L has employed in the evaluation of the replacement strainers and demonstration of the long term ECCS recirculation capability. This Gnal report is submitted to the NRC as required by Bulletin 96-03.
As described in Section 2, CP&L's primary action for resolution of the potential ECCS suction strainer blockage issue was the replacement of the previously installed suction strainers with large, passive strainers. CP&L, together with Duke Engineering and Services, Inc. (DE&S),
Innovative Technology Solutions Corporation (ITS) and Performance Contracting, Inc. (PCI),
has completed the design, engineering, fabrication, installation and testing o-f the replacement strainers for BSEP Unit Nos. I and 2.
Installation and testing of the new ECCS Suction Strainers was completed on October 7,1997 in Unit 2, and May 21,1998 in Unit 1.
2.0 Potential ECCS Suction Strainer Blockage Issue Resolution Approach CP&L's approach for resolution of the potential ECCS suction strainer blockage issue includes the following:
Replace the existing ECCS suction strainers on the Residual ifeat Removal (RilR) and Core Spray (CS) systems, with larger, passive strainers.
Perform periodic suppression pool and ECCS suction strainer inspections and cleanings, based on the predicted sludge generation rate.
Maintain an effective foreign materials exclusion (FME) program to ensure the cleanliness of the primary contamment, j
l Perform reviews of data obtained during the quarterly Technical SpeciGcation surveillance of the ECCS pumps.
l 1
a
l l
l RSEP, l! nit Nm. I and 2, ECCS Suction Strainer Replacensent Project NRC Bulletin 96-W Final Report I
Primary containment coatings will be periodically monitored to ensure the quantity of unqualiGed coatings remains within the approved design limits.
l These actions are presented in more detail below.
2.1 Suetion Strainer Replacement i
l CP& L has replaced the existing suction strainers on the RiiR and CS Systems with large, passive
)
i strainers. CP&L selected PCI's Sure-Flow suction strainers, a stacked disk strainer design, for j
the replacement strainers (see Appendices B and C). The main design features of the Sure-l Flow" suction strainers are as follows:
The disks are fabricated from stainless steel, perforated plate. The holes in the perforated l
plate are selected such that a maximum particle size that can pass through the strainer is less l
than the minimum ori6ce size in the RilR and CS systems. The perforated plate can be i
inspected and cleaned underwater. Stainless steel is utilized to prevent degradation of the strainers due to corrosion.
l The disks are attached to an internal core tube that provides How control capabilities. Hojes are cut in the core tube and are designed such that the uniform now is achieved along the entire strainer length and the now across the perforated plate is laminar, thereby reducing head loss.
The core tube also acts as a structural backbone, capable of resisting the large submerged structure hydrodynamic loads that are postulated in Boiling Water Reactor (BWR) suppression pools. Use of the core tube and suppons allows relatively long strainer lengths to be installed. The supports attach directly to the core tube.
l The strainers are designed to prevent vortexing. This has been confirmed by prototype I
testing.
m strainers have been tested to demonstrate their hydraulic The prototype Sure-Flow performance. These tests were conducted at the Electric Power Research Institute (EPRI) NDE Center in Charlotte, NC. The tests included low and high Eber quantities, with and without particulate. Testing of mixed fiber and reucctive metal insulation beds was also performed.
These tests demonstrate the ability of the Sure-Flow strainers to perform under various debris loadings, and also demonstrate that the analytical methods developed to evaluate strainer performance provide acceptable results. The reports documenting the tests performed and the results obtained are included in Appendix B.
The replacement strainers have been sized to provide suf6cient surface area to ensure that the head loss through the fouled strainers, for the design basis events, is acceptable. The strainer sizes for each system loop, in terms of strainer diameter, length and surface area, are provided in the table shown below. Also shown is the approximate increase in surface area, as compared to the existing strainers.
I L
l CSEP, Unit Nos. I and 2, ECCSSuction Stritner Replacement Project NRCBulletin 96-03 FinalReport i
Approx.
Approx.
Approx. Increase System Replacement Strainer Replacement Strainer in Surface Area
- l Diameter x Length
- Surface Area
- RHR 45" dia. x 16'-9" long 529 ft2 1,550 %
CS 45" dia. x 6'-0" long 245 ft2 1,530 %
- the total values per system loop suppression pool penetration 2.2 Suppression Pool and ECCS Strainer Cleanliness i
In response to the NRC Bulletin 95-02," Unexpected Clogging of Residual Heat Removal (RHR)
Pump Strainer While Operating in Suppression Pool Cooling Mode," CP&L committed to develop a long term program for inspecting and cleaning the BSEP suppression pools and ECCS suction strainers.
This program has been developed and is documented in Preventive Maintenance (PM) Routes APUH 001 and APUL 001, for Unit Nos. I and 2, respectively. These routes require periodic inspections of the suppression pool for sludge accumulation. If the sludge accumulation limit specified in the routes is reached, then suppression pool cleaning is scheduled and performed in accordance with guidelines given in the routes. The routes also require periodic physical condition inspections of the ECCS suction strainers, at the same interval as the suppression pool inspections. The strainers are cleaned, if required by the route criteria.
l The sludge accumulation limit specified in the PM routes was selected to ensure that the maximum sludge accumulation in the suppression pool does not exceed the design value used in l
the strainer performance evaluation (refer to Section 3.5). The sludge accumulation limit was l
established based on sludge generation rates determined from data obtained from previous BSEP suppression pool inspections /cleanings and from review of other industry data. These sludge generation rates will be reviewed as additional data is obtained from upcoming suppression pool inspections and the inspection intervals and sludge accumulation limit may be adjusted accordingly.
2.3 FME Program Also as discussed in CP&L's response to Bulletin 95-02, CP&L maintains FME and cleanliness administrative procedures to ensure the cleanliness of the Primary Containment. These l
procedures are OAl-125, " System Cleanliness /FME," and 0Al-127, " Primary Containment Inspection and Closcout." These procedures control materials in the drywell, suppression pool, and systems that interface with the suppression pool, and ensure that materials that could potentially impact ECCS operations are properly controlled and prevented from entering the suppression pool.
i Currently performed each refueling outage 3
I L __-_- _ _
BSEP, Unit Nos. I and 2, ECCSSuction Strainer Replacement Project NRC Bulletin WO3 FinalReport 2.4 ECCS Pump Technical Specification Surveillance Data Review As discussed in the response to Bulletin 95-02, CP&L will continue to collect and review pump suction data obtained during the quarterly Technical Specification surveillance. This collection and review of the Technical Specification surveillance data will identify potential degradation of the ECCS systems during the period between the inspections.
2.5 Primary Containment Coating Monitoring Baseline Primary Containment coatings inspections have been performed during the B112R1 and B213R1 outages while the torus was drained for ECCS Suction Strainer Replacement. Coatings are periodically monitored to assure that the quantity of unqualified coatings remain within the design limits, and to evaluate service level I coating degradation.
i 3.0 Plant Configuration and Design Parameters 3.1 Containment Design 3.1.1 Containment System Design Both units have Mark I containment designs. The Mark I containment consists of the following:
Drywell Pressure suppression chamber (torus)
Vent system between the drywell and torus e isolation valves Vacuum breakers In the event of a process system piping failure within the drywell, reactor water and steam are released into the drywell atmosphere. The resulting increased drywell pressure then forces a mixture of drywell atmosphere, steam, and water through the vent system, which opens beneath the surface of the suppression pool water in the torus. The water quenches the steam, thereby controlling the pressure rise in the drywell. The drywell steam-air-nitrogen atmosphere, which is transferred to the torus, pressurizes the torus. In the event that the drywell pressure drops below the torus pressure, the torus is vented to the drywell through vacuum breakers to equalize the pressure between the two vessels.
Cooling systems are provided to remove heat from the reactor core, the drywell, the water in the torus, and from the atmosphere in the vapor space above the water. This provides continuous cooling of the primary containment under accident conditions to maintain the pressure suppression capabilities and containment integrity.
4
i l
BSEP, Unit Nos. I cnd 2, ECCSSuction Stritner Replacement Project l
. NRC Bulletin 96-03 FinalReport l
3.1.2 Pressure Suppression Chamber (Torus) Mark i Design 1
1 The pressure suppression chamber (torus) is a hollow, reinforced concrete shell encircling the lower portion of the drywell containment structure. The concrete shell encloses a continuous 16-sided steel torus liner of circular cross sections. The major centerline diameter of the torus is 109 feet and the cross-sectional diameter of the circular liner is 29 feet. A paper joint is provided between the bottom of the torus and the mat foundation, to allow radial expansion of I
the torus. Vertical keys are provided along the outside perimeter of the drywell pedestal to allow independent, unrestrained radial expansion of the torus, when subjected to symmetric loadin :
Under asymmetric loads, the keys force the drywell and torus to respond as a single unit.
The pressure suppression vent system connects the drywell and torus together. Eight circular vent pipes, equally spaced around the periphery of the drywell, connect the drywell to a vent header contained within the air space of the torus. Projecting downward from the header are 96 downcomer pipes, which terminee approximately 3 feet below the water level of the torus. Jet l
de0ectors are provided at the inlet of each vent pipe to prevent possible damage to the vent line and header due to jet forces and/or missiles which might accompany a pipe break inside the drywell. Steam discharged into the torus causes the pool to swell and surge upward. Vent header deflectors are provided to deflect this surge away from the vent header toward the torus liner.
Steam discharged through the safety relief valves (SRV's) enters the torus through T-quenchers.
i The T-quenchers contain numerous orifices through which the steam exits into the suppression pool.
l l
A drawing showing section views through the torus is provided in Appendix D.
3.2 ECCS Design The ECCS Systems include the R11R System (operating in the low pressure coolant injection mode), CS, liigh Pressure Core injection (1IPCI), and the Automatic Depressurization System (ADS). The combination of these systems is designed to satisfy the 10CFR50.46 ECCS Acceptance Criteria. The RHR and CS Systems are the ECCS Systems utilized for long-term cooling. These systems take suction from the suppression pool, and, as such, are potentially impacted by the suction strainer plugging issue. Therefore, the RilR and CS suctions strainers have been replaced. As the llPCI System is not utilized for long-term cooling, its suction strainer was not replaced.
A brief description of the RHR and CS Systems is provided below.
3.2.1 RilR System The RilR system consists of two essentially complete and independent loops, identified as Loop A (Division 1) and Loop B (Division 11), with each loop containing two pumps and the necessary piping, valves and controls to support the various modes of operation. The suction piping of each pump on a loop ties into a common header (24" diameter pipe) which singly l
l 5
l l
BSEP, Unit Nos. I and 2, ECCSSuction Strxiner Replacement Project NRCBulletin 96-03 FinalReport penetrates the torus. The suction strainer is attached to a flanged connection immediately adjacent to the torus penetration.
The RHR system is designed for the following six modes of operation:
Low Pressure Coolant injection (LPCI)
Shutdown Cooling Containment (Drywell and Suppression Pool) Spray Suppression Pool Cooling RHR Service Water injection Fuel Pool Cooling Assist j
The LPCI mode of operation of the RilR System is the only mode providing an emergency core cooling function. This mode of operation provides a low pressure source of core cooling for the entire range of postulated break sizes. If the break is small, yet high pressure systems are unable to recover level, ADS will depressurize the Reactor Vessel, thus allowing LPCI and core spray injection. For larger breaks, the Reactor Vessel depressurizes via the break which allows almost immediate (approximately 20 sec) LPCI injection. The LPCI mode injects via the Reactor Recirculation System discharge piping to provide a volume of water to flood the Reactor core.
Automatic initiation of the system occurs from low reactor vessel water level or high drywell pressure (approximately 1.8 psig) " coincident" with low reactor vessel pressure (approximately 410 psig).
After the core is flooded to at least two-thirds core height, one RHR or Core Spray Pump is normally required to make up for Jet Pump throat to diffuser slipjoint leakage. The other pumps are stopped so that the emergency power (if there is a loss of offsite power) that was being used by these pumps may be shifted to other plant loads meluding RHR Service Water Booster Pumps. One RilR Pump and Heat Exchanger are normally placed in the Containment Cooling Mode aner a line break in the drywell as per the Emergency Operating Procedures (EOPs).
Upon automatic initiation of the A(B)2 Loop of LPCI, the associated pumps should start approximately 10 seconds after the initiation signal is received. If a loss of power has occurred l
resulting in the Emergency Diesel Generators powering the pumps, the pumps will start 10 seconds afler the associated diesel ties onto the emergency bus, j
The A(B) and C(D)3 RilR Pumps take suction from the Suppression Pool through the normally open motor operated Suppression Pool Suction Valve, Ell-F020A(B), and the associated motor-operated RHR Pump Suppression Pool Suction Valve, Ell-F004A(B) or Ell F004C(D),
and discharge through the Pump Discharge Check Valve, El1-F031 A(B) or Ell-F031C(D). The discharge check valves are designed to prevent backflow through the pump and to maintain a water leg in the discharge piping.
I With the CS and RHR Pumps running at minimum flow or dead headed, indicated pump discharge pressure on CS should increaae to approximately 305 psig and RhR should increase to 2
A(B) Loop signifies Loop A and/or Loop B.
A(B) and C(D) pump signifies Pump A and/or Pump C on Loop A, and Pump B and/or Pump D on Loop B. A similar identification applies to valves.
6 l
)
BSI:P, I! nit Nos I anti 2, ECCS Suction Strainer Replacenrent Project NRC Bulletin 96-03 FinalReport approximately 202 psig. As reactor pressure decreases to approximately 410 psig, the LPCI Inboard injection Valve, El1-F015A(B), should automatically open. As reactor pressure continues to decrease, the discharge of the RilR Pumps should overcome reactor pressure below approximately 200 psig, allowing the Dowpath to continue from the RilR Pumps' discharge check valve directly into the Reactor Vessel through the normally open LPCI Outboard injection Valve, Ell-F017A(B), the LPCI inboard injection Valve, Ell-F015A(B), the LPCI Injection Line Check Valve, El l-F050A(B), the locked open LPCI Manual Injection Valve, El1-F060A(B), and into the Reactor Recirculation System discharge lines. Once reactor pressure is reduced to approximately 20 psig, RilR How should reach approximately 17,000 ppm per operating loop with two pumps.
The LPCI Outboard injection Valve, F017A(B), is a throttle valve which may be adjusted to control now into the vessel, whereas the inboard injection Valve, El 1.F015 A(B), is designed for either full open or full close service. El1-F017A(B)is normally open, but with a LPCI initiation signal present, this valve cannot be closed or throttled for 5 minutes to ensure a discharge path exists from the pumps to the vessel. Ell-F015A(B), cannot be closed as long as the LPCI initiation signal is present. In addition, the RilR heat exchanger is automatically bypassed via the RiiR lleat Exchanger Bypass Valve, Ell-F048A(B), for the first three minutes to ensure that now gets to the reactor through the most direct route. During the interval of time when the RilR pumps are operating to restore the reactor vessel lesel, suppression pool heat removal is not necessary.
Each LPCI loop is provided with a minimum Dow line to the Suppression Pool to protect the pumps from damage due to overheating as a result of low or no-Dow operation. This feature allows the pumps to operate with a closed discharge valve, without overheating, by recirculating Suppression Pool water through the minimum Dow bypass line.
Each of the two separate loops of RilR in the Suppression Pool Cooling (SPC) mode, provide the primary source of containment cooling. Following a LOCA, SPC is manually initiated to limit Suppression Pool temperature and Primary Containment pressure within design limits.
The RilR pumps receive power from the 4160V emergency auxiliary buses. For each loop, the RilR pump motor and associated automatic motor valves receive AC power from different buses.
3.2.2 Core Spray System J
The Core Spray (CS) System provides a low pressure source of core cooling for the entire range l of postulated break sizes. If the break is small, yet high pressure systems are unable to recover level, ADS will depressurize the reactor vessel, thus allowing CS injection. For larger breaks, the reactor vessel depressurizes via the break which allows immediate CS injection. The CS System injects via a dedicated set of spray spargers (one per loop) in a spray pattern directly over the reactor core.
The CS System consists of two redundant and independent loops. Each loop contains one 100%
l capacity pump and the necessary piping, valves, and controls. The suction piping (14" diameter pipe) on each loop independently penetrates the torus. The suction strainer attaches to a Hanged I
l 1
7 o
_____a
l l
HSEP, Unit Nm. I anti 2. ECCS Suction Strainer Replaccinent Project NRC Rulletin 96-0.1 Final Report I
connection on the torus penetration.
l The CS System controls support vessel injection when reactor pressure is reduced below the l
injection permissive. Automatic initiation of the system occurs from low reactor vessel water level or high drywell pressure coincident with low reactor vessel pressure l
The two loops are designated as "A" and "II", with the "A" Loop assigned to Division I and the f
"II" Loop assigned to Division 11. Water from the suppression chamber is discharged into the I
reactor pressure vessel via the CS spargers. An alternative water supply to the CS loops is available from the Condensate Storage Tank, via a manual valve.
I The CS System relies upon station AC power sources to power the pumps. Pump power is l
supplied from the associated station emergency 4160 VAC llus. Each pump is powered from an independent bus with each emergency bus capable of being powered from one of three sources of AC power. Normally during plant operation, the buses will be powered from the Main Generator output via the Unit Auxiliary Transformer (UAT). During plant shutdown conditions, and in the event this power source is lost, the emergency buses are automatically or manually transferred to the unit's Startup Auxiliary Transforn.er (SAT). Should both of the above fail, the station Emergency Diesel Generators automatically power the buses.
Failure of a single loop of the CS System is backed up by the other loop. Failure of one or both l
loops is also backed up by the LPCI Mode of RiiR.
3.2.3 Limiting Conditions for Strainer Performance l
The specific, limiting (i.e., worst case) conditions that are to be used in the evaluation of the strainer performance are provided below. These conditions are associated with a LOCA and loss of one emergency power system. For the governing case for RllR, this results in two RilR pumps on one loop operating at run-out conditions, and one CS pump operating at minimum Technical Specification flow. For the governing case for CS, one CS pump is operating at run-out conditions, with one RilR operating at minimum Technical Specification flow. Run-out flow is assumed to continue until up to ten minutes into the accident, when coolant level in the reactor is reestablished. Operator action is taken at that time to reduce RilR and/or CS flows, as any single RilR or CS pump is able to maintain water level in the reactor. This action is required to change the RilR mode of operation to containment cooling and/or containment spray and also assures adequate steam condensation capability by limiting suppression pool temperatures.
Maximum RllR flow until ten minutes into the accident is either: 1) 21,000 gpm through one loop, coincident with total CS flow of 4,625 ppm for a Recirculation Line Ilreak; or 2) 18,750 gpm through one loop, coincident with total CS flow of 4,625 gpm for normal RilR injection.
At ten minutes, the total RilR flow is reduced to 11,550 gpm, which is the nominal, single pump RilR flow.
Maximum CS flow of 6,700 gpm, coincident with total RilR flow of 7,700 ppm, until ten l
minutes into the accident. At ten minutes, the CS flow is reduced to 4,725 gpm, which is the nominal CS flow.
8
RSEP, finit Nos. I and 2, ECCS Suction Strainee Replacernent Project NRC Bulktin 96-03 FinalReport These conditions are based on a Design Basis Accident (DBA) LOCA. The DBA conditions, in conjunction with the debris loading generated as a result of a DBA, represent the governing suction strainer design conditions.
3.2.4 Net Positive Suction licad Margins The Net Positive Suction llead (NPSH) margins available for the replacement strainers for the RiiR Loops A and B pumps (2 pumps and one strainer per loop) and CS Loops A and B pumps l (one pump and one strainer per loop) were determined. The pump NPSH is affected by the wetwell pressure, suppression pool temperature and line losses. The BSEP NPSH evaluation does not take credit for wetwell pressure increases, so the pressure is held at 14.7 psia, for the design basis events.
UFSAR Section 6.2.1.1.3.2.1 defines the most limiting credible scenario for long term containment cooling For a RIIR failure, this is a LPCI recirculation coolant discharge line break where the pump is initially running in an unthrottled state into the broken pipe loop at 21,000 gpm. For a CS failure this is a recirculation coolant suction line break where the pump is initially running in an unthrottled state into the reactor vessel core spray spargers at 6,700 gpm.
l At ten minutes into the accident, the available NPSil margins are as follows:
=
accident condition: @ 21,000 gpm,161.8 F,14.7 psia:
9.4 ft. head or normal condition: @ l8,750 gpm,162.5 F,14.7 psia:
12.0 ft. head c5 1
accident condition: @ 6,700 gpm,162.5 F,14.7 psia:
10.9 ft. head After ten minutes into the accident, the RilR and CS pumps are assumed to be manually throttled by an operator and the available NPSil margins are as follows:
RilR @ l1,550 gpm,189.4 F,14.7 psia:
9.0 ft. head CS @ 4,725 gpm,189.4 F,14.7 psia:
9.6 ft. head l
These NPSil margins after ten minutes into the accident are based on the maximum predicted suppression pool temperatures.
As the suppression pool temperatures decreas<r beyond that l
point in time, both the available NPSil margin and the debris head loss will increase. Ilowever, l
analysis has shown that the limiting conditions are predicted to occur at the time of maximum j
suppression pool temperature.
(
l l
i 1
l i
BSEP, Unit Nos. I end 2, ECCSSuction Strainer Replacement Pro}ect NRCBulletin 96-03 FinalReport 1
3.3 Drywell Insulation The insulation systems used in the drywell ir.clude both fibrous and reflective metal insulation.
l The specific types and approximate quantities are summarized below:
Approx. Quantity Type Unit 1 l
Unit 2 3
Fibrous Insulation (quantities in ft )
NUKON -Jacketed 1,406 1,232 NUKON - Unjacketed 1,145 1,036 Temp Mat 46 58 Reflective Metal Insulation (quantities in ftz)
Transco RMI(l mit Al) 33,021 44,609 Diamond Power RMI (2 mit SS) 44,320 44,320 J
Other (quantities in ft )
Calcium Silicate 4
4 Micro Therm 2
2 Vention (fiberglass fabric) 0.2 0.2 Asbestos Fabric Wrapping 0.0 0.03 l
3.4 Other Potential Drywell Debris Sources The other potential drywell debris sources can be divided into three categories. These categories l
are defined in Section 3.2.2 of NEDO-32686," Utility Resolution Guidance for ECCS Suction Strainer Blockage,"(URG), Revision 0, dated November 1996.
3.4.1 Fixed Debris Fixed debris is material that is part of the permanent plant that becomes a debris source only after exposure to the effects of a LOCA. This includes material that is blown or stripped off as a result of impingement forces. Material in this ca'egory are listed in Section 3.2.2.2.2 of the URG.
The other potential drywell debris applicable to BSEP are as follows.
Quantity Debris Type (Ibm)
Reference LOCA generated dirt / dust 150 URG, Section 3.2.2.2.3.2 Qualified paint / coatings 71 URG, Section 3.2.2.2.2.1.1 (100 % epoxy)
Rust flakes from unpainted steel 50 URG, Section 3.2.2.2.3.6 Other Drywell fixed debris 15 Contingency (assumed to be paint)
Total 286 10
BSEP, U:lt Nos. I cnd 2, ECCSSuction Strainer Replacement Project i
NRC Bulletin 96-03 FinalReport l
)
i 3.4.2 Latent Debris l
Latent debris is debris that would not be present until later in the LOCA event progression after l
prolonged exposure to a LOCA environment. For example, unqualified coatings (not directly
)
impacted by the LOCA jet) have the potential to detach from the surface where applied, but only after prolonged exposure to the LOCA environment and after containment pressure is reduced later in the event. This type of debris is discussed in detail in Section 3.2.2.3 of the URG.
For strainer design purposes in each unit, 6.67 cubic feet of unqualified paint / coatings is permitted. Primary Containment Coatings baseline inspections have been performed on both units during the Bil2R1 and B213R1 outages. The current unqualified coatings values are less than 90% of the design limits. Periodic monitoring is performed to assure that the quantity of unqualified coatings remain within the design limits.
3.4.3 Transient Debris i
Transient debris is non-permanent plant material brought into the drywell, typically during an outage (e.g., tools, rags, sheets, plastic bags, temporary filters, dirt / dust, etc.). Transient debris is principally controlled through FME and housekeeping programs. This type of debris is described in more detail in Section 3.2.2.2.1 of the URG.
CP&L maintains an effective FME and housekeeping program. Dirt / dust debris is included in the fixed drywell debris and wetwell sludge quantities. A contingency of an additional 4 f3 of j
fibrous debris is included in the strainer design to account for fibrous transient debris.
3.5 Potential Wetwell Debris Sources Potential wetwell debris sources are rust particles (sludge) from the carbon steel, including the torus liner and interior surfaces of torus attached piping systems, and unqualified paint.
Based on limited BSEP specific data and with comparisons to other industry data, the sludge l
generation rate is currently being conservatively set at 100 lbm (dry) per year. A total sludge i
quant:ty of 600 lbm (dry) is to be used in the strainer evaluation, based upon an assumption of three,24 month cycles of sludge generation before cleaning. It should be noted that the last sludge measurement at BSEP Unit Nos. I and 2 showed a sludge generation rate of 33 lbm (dry) per year.
The quantity of unqualified paint in the wetwell is included in the quantity of unqualified paint specified in Section 3.4.2.
u_____________________--
CSEP, U;:lt Nos. I cnd 2, ECCSSuction Strdner Rept: cement Project NRCCulletin 96-03 FinalReport 4.0 Replacement Strainer Design Requirements 2
4.1 Strainer Functional Requirements
'Ihe functional requirements for the suction strainers are as follow:
The suction strainers are to screen out debris particles greater than 0.095".
l This particle size limit ensures that any particle passing through the strainer will be smaller l
than the orifices associated with the cyclone separators, and all other small flow restrictions in the pump seal flush piping. This requirement is consistent with the General Electric Company Service Information Letter No. 323, which provided recommendations for sizing the strainer hole si7e.
l The suction strainers, including the strainer elements itself, the collected debris and any interconnecting piping components, produce a total head loss that does not exceed:
l For the first ten minutes into the LOCA:
I 8.4 ft. water at 21,000 gpm at 161.8 F, and i1.0 ft. water at 18,750 gpm at 162.5 F l
for PJIR; l
9.9 ft. water at 6,700 gpm at 162.5*F for CS After ten minutes:
8.0 ft. water at 11,550 gpm at 189.4*F for RHR 8.6 ft water at 4,725 gpm at 189.4"F for CS 1
These total head loss values provide an available NPSil that exceeds the required NPSli for I
the ECCS modes of operation for RIIR and CS. The debris generation, transport and head loss determinations is in accordance with Bulletin 96-03 and Regulatory Guide 1.82, l
Revision 2, unless otherwisejustified.
4.2 Strainer Code Requirements l
The strainer code requirements are in accordance with the requirements listed below. These l
code requirements are considered appropriate for the intended applications and function of the i
suction strainers. The code edition and addenda utilized are consistent with those specified in l
the UFSAR.
Quality Assurance l
ASME Certificate not required Materials 1
Conform to ASTM material specifications Certified Material Test Reports (CMTRs) are to be provided for all materials (except where ASME Section Ill Code would permit Certificates of Compliance) 12 L___________-_________________.
BSEP, Unit Nos. I and 2, ECCSSuction Strainer Replacement Project NRC Bulletin 96-03 FinalReport Design e
- Qualified to ASME Section 111, Nuclear Power Plant Components, Subsection NC Welding e
Weld procedures and personnel qualified to ASME Section IX.
- NDE Critical structural welds examined by PT or MT per ASME Section 111 All other welds visually examined per ASME Section 111 Stamping e
NPT Stamp is not required The RIIR and CS piping systems are classified as ASME Class 2 piping systems. The suction strainers are components which attach to these piping systems. As the suction strainers are non-pressure boundary components, the suction strainers are not required to be constructed to the l
requirements of the ASME Code (refer to Paragraphs NCA-1130(b) and NC-2121,1986 l
Edition). In fact, no code and standards have been developed or adopted that specifically define the requirements for the construction of the strainers for this application. Therefore, the above requirements have been developed and are specified for the construction of the strainers. These requirements rely on good engineering practice and have adopted appropriate sections of the ASME Code to assure the fabricated strainers are robust and of high quality.
Suction strainers constructed to the above requirements are of a quality and reliability l
appropriate and commensurate with their intended application and function.
4.3 Strainer and Strainer Support Structural Requirements The strainers and strainer supports are qualified for the loads, load combinations and acceptance criteria established under the Mark I Containment Reevaluation Program. The governing documents for these requirements are as follows:
NEDO-21888," Mark 1 Containment Program Load Definition Report,"(LDR) Revision 2, November 1981, including Addenda, Sheet 1, April 1982.
l NEDO-24583-1, " Mark I Program, Structural Acceptance Criteria, Plant Unique Analysis Application Guide,"(PUAAG), October 1979.
NUREG-0661, " Safety Evaluation Report, Mark I Containment Long-Term Program, Resolution of Generic Technical Activity A-7," July 1980, including Supplement 1, August 1
1982.
13
BSEP, Unit Nm. I and 2. ECCS Suction Strainer Replacement Project l
NRC Bulletin 96-03 Final Report 5.0 Replacement Strainer /ECCS Long Term Recirculation Evaluation Methodology 1
This section presents the methodology used to evaluate the replacement strainers and ECCS long-term recirculation. This methodology meets the replacement strainer design requirements discussed in Section 4. This methodology is based on information contained in NUREG/CR-6224, in the URG and in other published literature. In some instances, the information in these i
publications is modified to provide appropriate application to this evaluation for BSEP.
5.1 Drywell Debris Generation l
The types and quantities of debris in the drywell generated as a result of a LOCA is a function of:
Postulated pipe break locations Break blast / jet shapes, sizes and intensities (i.e., Zone ofInfluence)
Insulation debris types and quantities within the Zone ofinfluence l
Other debris types and quantities Each of these items can be evaluated explicitly for each postulated pipe break location to determine the quantity of debris generated by that break.
Alternately, simplified, but conservative evaluations can be made that provide bounding debris quantities to be utilized for the strainer and ECCS long-term recirculation evaluation.
CP&L performed a more simplified evaluation of the bounding debris quantities generated as a result of a LOCA. This simplified evaluation approach is described below.
Determine the type, thickness and application (i.e., the piping systems and pipe sizes) of the insulation systems utilized in the drywell. (Note that this information, as well as the general condition of the insulation, has been confirmed through the performance of field walkdown conducted during the last two refueling outages.)
Determine the Zone ofInfluence for each insulation type by pipe break size using Method 3 of the URG (refer to Section 3.2.1.2.3.3).
l The URG uses a spherical Zone of influence, which is consistent with guidance given in i
Regulatory Guide 1.82, Revision 2. As defined in the URG, the volume of the sphere is I
determined for each insulation type based on the pipe diameter where the break is postulated, the amount of break separation, and the destruction pressure for that insulation type. All breaks are conservatively assumed to be unrestrained (i.e., having a radial offset greater than 3D/2). Destruction pressures are provided for the required insulation types.
Once the volume of the sphere is known, the equivalent sphere radius is calculated.
14 l
BSEP, U lt Nos. I gnd 2, ECCS Suction Strainer Replacement Project NRCBulletin 96-03 FinalReport Determine the routing of the high energy piping and their location relative to the majority of the piping insulation in the drywell using the plant composite piping drawings.
Identify potentially controlling pipe breaks using the above information.
This step involves inspecting the composite piping drawings, and with the knowledge of the Zones of Influence for various pipe break sizes, selecting break locations which would generate large volumes of debris. The location of the insulation debris relative to platforms and grating is also accounted for due to different transport factors being applied (see Section 5.2). As BSEP is mostly homogeneous in its fiber-type insulation, this identification of potentially controlling pipe breaks from visual inspection of the composite drawings is possible.
Determine the insulation debris types and quantities associated with each potentially controlling pipe break, and associated RIIR/CS demand (i.e., flow rates).
Select the break location (s) that results in the bounding debris volumes to be utilized in the strainer performance evaluation. More than one break location may be required to be evaluated to address all insulation types and corresponding RIIR/CS demand.
The reactor vessel is insulated with reflective metal insulation (RMI). Any pipe breaks inside the bioshield wall would be bounded by the drywell pipe break quantities of RMI, since there are guard pipes installed on the reactor vessel nozzles which direct 85% of the pressure out through the bioshield walls. This was confirmed by evaluating breaks inside the bioshields walls.
l Other non-insulation drywell debris is generated as a result of the LOCA. This debris and the potential quantities generated for each are provided in Section 3.4.
5.2 Drywell Debris Transport The drywell insulation debris transported as a result of a LOCA were determined using the combined generation and transport factors given Section 3.2.3.2.5, Tables 5 and 6, of the URG.
The basis for these combined generation and transport factors are also provided in the URG, and provide the justification for utilizing transport factors less than 100%.
All particulate debris generated in the dryw:ll was assumed to be transported to the suppression pool with 100% efficiency, and in the suppression pool at time zero into the accident.
5.3 Wetwell Debris Transport The suppression pool is highly turbulent during the blowdown phase of a LOCA.11owever, after the first several minutes of the transient, the degree of turbulence is significantly reduced. The key phenomenon that can occur at that time is settling of debris to the bottom of the suppression pool. The fibrous debris is typically too light for settling to be important, and thus fiber settling was not considered.
I.
15
BSEP, Unit Nos. I and 2, ECCS Suction Strainer Replacement Project NRC Bulletin WO3 FinalReport Particulate debris, especially larger particulate species, such as paint chips and rust fiakes, have been shown to settle very rapidly. The key parameters for calculating settling, other than the paniculate sizes and densities, are the suppression pool volume, the volumetric flow rate through the suppression pool, and the effective particulate settling velocities (based on degree of pool 1
turbulence). For conservatism, this settling was not included in the strainer head loss analysis.
However,it was considered in the strainer margin assessment performed as described in Section 5.5.
5.4 Suction Strainer Blockage and IIcad Loss i
l The strainer head loss, given the deposition of a certain amount (and type) of fibrous debris and a certain amount (and type) of paniculate debris on the strainer, was calculated using a modified NUREG/CR-6224 head loss model. The head loss model in NUREG/CR-6224 was modified by our consultants, ITS, to provide a correlation for strainers having different geometries and for strainers with a heavy fiber loading. This modified head loss model explicitly treats the effects of:
stacked disk strainer geometry for both light fiber loads (when the entire stacked disk surface e
area is accumulating debris) and for heavy fiber loads (when the fibrous debris is simply building up on the outside of a cylindrical shape);
debris deposition on the outside of a cylinder rather than a flat surface (resulting in a reduced l
bed thickness);
different fibrous debris constituents; e
different types and quantities of particulate debris constituents; e
overall bed porosity (e.g., compression); and e
the effect of different fluid temperature and flow rate.
a It should be noted that the last four items listed above are also explicitly treated in the original NUREG/CR-6224 head loss model.
The model was developed to represent the physics and fluid flow behavior expected for thick l
fiber beds (with paniculates) on large passive strainers. A comparison of model predictions to the results of the PCI strainer tests performed at EPRI has been performed. This comparison showed agreement to within ~10% for a wide range of fiber quantities, and sludge to fiber mass i
ratios.
l l
A complete description of the modified head loss model is presented in Appendix A.
This modified head loss model and a comparison of the model predictions with results from tests performed at EPRI were presented to the NRC Staff by ITS on February 18,1997.
16 l
l
BSEP, U:lt Nos. I and 2, ECCSSuction Strainer Replacement Project NRC Bulletin 96-03 FinalReport A copy of the PCI Test Reports is included in Appendix B.
As discussed in Appendix A, the strainer head loss was calculated considering the time-dependent debris build-up on the strainer. This buildup of debris on the strainers, which occurs over a finite time interval, is primarily determined by the volumetric flow rate through the strainer as compared to the suppression pool volume.
The strainer head loss, due to the deposition of RMI across the strainer surface, was calculated using RMI head loss model presented in Appendix B to the URG. The combined effect of RMI and fiber was approximated by summing the calculated head loss due to each material alone. For Brunswick, the debris source term is dominated by fibrous debris, and the ratio of RMI to fiber is very low. Testing conducted on the prototype Sure-flow" strainer at the EPRI facility demonstrated that, for such low RMI to fiber ratios, the sum of the head loss due to each material alone is a conservative approximation to the total head loss.
5.5 ECCS Pumps NPSil Margin Assessment The existing NPSli margin calculations were revised to incorporate the head loss through the i
debris, calculated as described above, the head loss through the strainer itself, and the head loss through any interconnecting piping components required to connect the strainer to the existing ECCS pipe connections. The revised NPSli margins were then determined and shown to be acceptable.
An assessment was also made as to the conservatism contained in the NSPli margins calculated above and the additional margins that would be shown to exist, without these conservatism included. This included consideration of other time-dependent aspects of strainer performance, including:
the efficiency with which the fibrous debris can trap particulate debris while the debris build-e up is occurring and prior to the development of very thick beds; and the settling of particulate in the suppression pool, which occurs during the finite time of i
debris build-up on the strainer, ne BLOCKAGE computer program was utilized to quantify this debris bed efficiency and particulate sedimentation.
This assessment provides additional assurance that RHR and CS systems can perform their intended functions.
The potential for air / steam ingestion into the strainers during an SRV T-Quencher discharge and the potential for vortexing was assessed and was determined to be acceptable. Minimum water levels for various RiiR and CS flow conditions specified in the Emergency Operating Procedures (EOPs) were revised for the replacement strainer design and performance requirements, without reliance on drywell pressure.
17
1 BSEP, Unit Nos. I c:d 2, ECCS Suction Str:Iner Replacement Project l
NRCBulletin 96-03 FinalReport 5.6 Suction Strainer Structural Evaluation j
i The replacement suction strainers and supports were qualified for the Mark I submerged structure loads defined in the LDR, and other applicable loads acting on the strainers (i.e.,
l deadweight, thermal expansion and seismic). This qualification utilized methods and parameters that are the same as described in the BSEP Plant Unique Analysis Report (PUAR) and in the UFSAR.
l In the implementation of these methods, specific analytical techniques were employed to reduce unnecessary conservatism that was included in the qualification of other BSEP submerged structures during the Mark I Program. These analytical techniques are discussed below.
l In the determination of the acceleration drag volumes for the hydrodynamic loads acting on the replacement strainer assemblies, an effective hydrodynamic mass coefficient, Cm, was l
determined for the strainers taking into account their geometry and perforated nature. As the Mark I LDR Application Guides do not provide a coefficient for submerged structures with geometries similar to the replacement strainers, one had to be developed to eliminate excessive conservatism. The Mark I LDR Application Guides do provide the methodology that was used for determination of the Cm coefficient for other submerged structures. This same methodology was employed in the development of the Cm coefficient for the replacement strainers.
Testing of prototypical strainers was performed to measure Cm coefficients for the Sure-Flow" stacked disk strainer design. The Cm coefficients obtained from the testing of the prototype strainer were adjusted to account for differences between the prototype design and the BSEP-specific design. This methodology used to determine Cm assures no reduction in the safety margin for the BSEP containment and ECCS design basis.
I Non-proprietary copies of the test reports on the prototype strainers, " Test Report for Hydrodynamic Inertial Mass Testing of ECCS Suction Strainers," No. TR-ECCS-GEN-01, l
Revision 2, and " Supplement 1 to Hydrodynamic Inertial Mass Testing of ECCS Suction Strainers - Free Vibration Analysis," No. TR-ECCS-GEN-05, Revision 0, are included in Appendix E. Report TR-ECCS-GEN-01 justified that, for a strainer with perforated plate l
with 1/8" and 40% open area, the Cm coefficient could be conservatively taken as 62% of the coefficient for the same strainer without holes (i.e., Cm=1.24). Report TR-ECCS-GEN-05 further examined the test results and concluded that a Cm coefficient of 50% of the coefficient for the same strainer without holes (i.e., Cm=1.00) could also be justified and shown to be conservative.
j The BSEP strainers utilize perforated plate with 3/32" holes and 33% open area. The effect of the reduced hole size and flow area will tend to increase the differential pressure across the perforated plate, thereby potentially increasing the Cm coefficient. This effect is equivalent to the phenomenon commonly referred to as a vena contracta. A ratio of the areas of the vena contracta was used to conservatively account for the change in the reduction in hole sire and flow area. This evaluation determined a Cm coefficient for the 3/32" l
18
RSEP, Unit Nin. I cnct 2, ECCS Suction Strainer Replacement Project NRC Bulletin 96-0.1 FinalReport plate to conservatively be 80% of the coefficient that would be applied to the same strainer without holes (i.e., C,,,=1.60 for an in6nitely long strainer).
j Other design differences, i.e., the disk and gap sizes, and the eccentricity of the disks, were conservatively not included in the adjustment, as these items, considered together, were judged to reduce the C,,, factor.
Subsequent to the development of the C,,, coefficient used in the qualineation of the BSEP strainers, additional testing was performed on several Sure-Flow strainers. The tested strainers had significant geometric differences to enable the evaluation of the effect of these differences on the response of the strainers to accelerated Gow Gelds. The geometric differences included overall length, disk diameter, perforated plate hole size, disk and gap widths, internal disk stifTening and core tube diameters.
These tests allowed the development of empirical formulations for C,,, as a function of dimensionless parameters which characterize the strainer geometry.
The tests demonstrated that the lateral C,,, coef6cient (using the volume displaced by the strainer enselope as a basis) is relatively constant over a wide range of parameters, varying between 0.48 and 0.63. For a single BSEP Core Spray strainer module, a C, factor of 0.63 has been determined based on these tests. including test uncertainty and 10% margin. For a complete BSEP Core Spray strainer assembly (considering the total length of two strainer modules attached to a ramshead) using the conservative correlation developed from the test, the predicted C,,, value is 0.85. Therefore, the C,,, value of 1.35 used in the qualification is conservative. For the longer RilR strainers, the C,,, value would be slightly higher (on the order of 1.0), but still well below the 1.54 value used. A copy of the report documenting these additional tests and the results obtained is also included in Appendix E.
In the development of specific load combinations, the BSEP PUAR typically used conservative, bounding combinations to simplify the analysis processes. Typically, jet loads and drag loads (due to either LOCA and SRV discharge) were conservatively added together.
These components ofload do not occur simultaneously. Thereforc, this practice was overly conservative. The analysis of the replacement strainer assemblies more appropriately considers the maximum of the jet load or the drag load when detennining the maximum loads due to a specific phenomena.
Additionally, the analysis for the replacement strainer assemblies combined independent j
loads using a square-root-sum-of-the-squares (SRSS) combination method. Although this i
combination method was not used in the analysis of the previously installed strainers, it was psed on cenain components in the original Mark i design at BSEP, when it was required to remove excessive conservatism, it was pan of the original Mark I design basis and its use was documented in the PUAR.
For the Chugging load case, the original analysis conservatively combined the effects of each harmonic absolutely, regardless of sign or phase. The analysis for the replacement strainer assemblies takes credit for the random nature of the phasing of each of the harmonics using a combination method developed during the BWROG efforts to resolve the Mark i Program. This combination method has been shown to meet NUREG-0661 generic 19
RSEP, Unit Nos. I ant! 2, ECCSSuction Strainer Replacement Project NRC Bulletin WO3 Final Report 1
l acceptance criteria, thus maintaining a consistent margin of safety with the original Mark I program.
5.7 Other Plant Structural, Systems and Components Evaluation The installation of the replacement strainers affected the qualification of other existing structures, systems and components. The affected items were:
the RHR and CS torus penetrations; e
the strainer support attachment points on the torus liner; e
other torus internal structures and systems in close proximity to the replacement strainers; e
and
+ torus water volume.
The requalification of torus penetration, liner, internal structures and piping components utilized the same methods as described in the BSEP PUAR and UFSAR. In some caces, the specific analytical techniques described above (i.e., SRSS of dynamic loads) for the strainer qualification were applied in the requalification of some of these components.
The torus water volume is decreased less than 100 ft', which is approximately 0.11% of the total water volume in the torus, due to the installation of the replacement strainers. This change has been evaluated and shown to have a negligible effect in the torus temperatures during accident conditions.
I 6.0 Licensing Considerations CP&L has implemented the strainer replacement, and evaluated the acceptability of the strainers and the ECCS long-term recirculation capability, under the 10CFR50.59 process.
A Technical Specification change, to address the change in suppression pool water volume due the installation of the replacement strainers, submitted under TSC #97TSB06 has been approved and implemented.
No other Technical Specification change is contemplated at this time. CP&L does not plan to include torus and suction strainer surveillance requirements in the Technical Specifications. As discussed in Section 2, these surveillance requirements are already contained in the appropriate BSEP preventive maintenance procedures (routes). Placing these surveillance requirements in the maintenance procedures provides the appropriate level of control to insure that adequate I
inspection and cleaning is performed.
20
BSEP, Unit Nos.1 c::d 2, ECCS Suction Stritner Rept: cement Project NRC B:illetin 96-03 Fin:IReport APPENDICES l
1 l
l
l' BSEP, Unit Nos. I cad 2, ECCS Sxction Strainer Rept: cement Project NRCB:l!etin 96-03 FinalReport l
l l
Appendix A Modified NUREG/CR-6224 Head Loss Correlation I
BSEP, Unit Nos.1 and 2, ECCS Suction Strainer Replacement Project AppenditA NRCBulletin 96-0.1 FinalReport Table of Contents A1.0 Background and Objectives......
......]
Al.1 Background.
..... 1 A1.2 Objectives of the Analysis....
.2 Al.3 Assumptions....
...... 3 A2.0 Methodology Description....
.. 3 A2.1 Basic IIead Loss Correlation.....
... 3 A2.2 Calculation of Effective Strainer Surface Area and Approach Velocity.........
..... 5 A2.3 Calculation of Debris Deposition on the Strainer..
.6 A2.4 Debris Thickness Calculation - Effective Surface Area Consideration..
.6 A2.5 Debris Bed Thickness Calculation - Surface Geometry Considerations....
...............7 A2.6 Consideration ofImpact of Strainer Supports..
.8 A2.7 Calculation of Average Debris Surface to Volume Ratio..
..8 A3.0 References.
..10 ii
BSEP, Unit Nos.1 and 2, ECCS Suction Strainer Replacement Project AppenditA NRCBulletin 96-03 FinalReport A1.0 Background and Objectives There is a concern that during a design basis loss of coolant accident (LOCA) in a Boiling Water Reactor (BWR), the strainers at the suction inlet to the Emergency Core Cooling System (ECCS) could become sufficiently clogged with debris generated during the LOCA as to cause cavitation and failure of the ECCS pumps. This debris consists of drywell piping insulation (fibrous or metallic) loosened as a result of the LOCA forces and transported to the suppression pool, sludge that has built up in the suppression pool during the reactor's steady-state operation, and other particulate debris sources such as dirt and dust, loosened paint chips, and loosened rust. In order to assess the performance and adequacy of such strainers, one must be able to predict the head loss across those strainers during the accident as a function of the time dependent debris buildup and the time-dependent ECCS flow and coolant temperature. At the present time no single tool exists that can predict such head loss under the full range of potential strainer debris loading conditions.
A1.1
Background
The most comprehensive assessment of the phenomenological issues that impact strainer head loss and the potential computational models that could be used to predict such head loss was conducted by the Nuclear Regulatory Commission (NRC) and documented in the NUREG/CR-6224 report entitled " Parametric Study of the Potential for BWR ECCS Strainer Blockage Due to LOCA Generated Debris"[Zigler,1995]. The correlations presented therein were shown to be applicable to a wide variety of situations involving head loss due to flow through a fibrous debris layer. The BLOCKAGE computer code [Rao,1996]was subsequently developed to implement the modeling approaches recommended in NUREG/CR-6224.
The BLOCKAGE code provides a very comprehensive assessment of the time-dependence of debris transport from the BWR drywell to the suppression pool, e
the buildup of debris on the strainers as a function of pump flow rate and pool water volume, e
the potential reduction in debris buildup as a result of sedimentation to the floor of the e
suppression pool, the potential reduction in the buildup of particulate debris as a result of less than perfect e
filtration of such particulate by the fibrous debris, and the head loss resulting from the flow through the deposited debris.
e However, the BLOCKAGE code was developed under the assumption that the surface area of the strainer could be treated as a constant, user-supplied input to the analysis, with the debris buildup being calculated as though the strainer could be represented as a flat surface with the same surface area. This simplifying assumption is very valid in the case where one has a large surface area relative to the debris volume, such that only a thin debris layer would be calculated.
However, in the case where one has a large volume of debris, with a complex strainer geometry involving stacked disks and curved surfaces, the BLOCKAGE approach to debris deposition is no longer valid. There are two principal reasons for this:
1
BSEP, U:lt N:s.1 c:d 2, ECCS S:ction Strainer Repl: cement Project Appe: dita NRCBulletin %03 Fin:IReport l
l
- 1. A stacked disk strainer has a very large surface area relative to the overall strainer volume.
' With large volumes of fibrous debris, the interstitial gaps between the disks can become filled with debris. When that occurs, the effective surface area of the strainer for additional debris deposition is reduced to the circumscribed area of the strainer.
l
- 2. For thick layers of debris on the outside of a cylindrical shape, the debris thickness relative l
to the debris volume is a function of the surface curvature, and is less than the thickness that would result from deposition on a flat surface of the same area.
l The NUREG/CR-6224 methodology accounts for the head loss due to particulate material l
trapped within a fibrous debris bed through a decrease in the bed porosity as well as through the use of a modified (" average") debris bed surface to volume ratio. A suggested approach to determining the average surface to volume ratio is not provided. However, the BLOCKAGE code allows for the use of either a simple volume averaged value, or simply the fiber value. For a particulate material such as sludge, whose surface to volume ratio is not too different from that of Nukon fiber, these two approximations do not have a large impact on the calculated result.
This is not the case, however, for a small surface to volume ratio particulate species such as paint chips. In this case, the approximation of keeping the surface to volume ratio fixed at the value for fiber has the effect of treating the paint chips as though it were simply more fiber. This tends to significantly overpredict the impact of paint chips on head loss. Even worse, however, is the use of a simple volume average. It can be shown that because of the form of the NUREG/CR-6224 head loss correlation, the use of simple volume averaging actually leads to the non-physical result that the addition of a small surface to volume ratio particulate such as paint chips actually reduces the calculated head loss. This is clearly an unacceptable, non-conservative result. It had not been identified as a problem previously, because the NUREG/CR-6224 correlations had not previously been applied to paint chip particulate.
i A1.2 Objectives of the Analysis In light of the limitations in BLOCKAGE identified above, and the fact that no near-term revision of the BLOCKAGE code has been publicized, a modified NUREG/CR-6224 head loss correlation was developed that could be used to assess stacked-disk strainer performance under heavy fiber load with significant quantities of arbitrary particulate (including small surface to volume ratio material such as paint chips). This methodology incorporates the following
)
features:
head loss estimates based on the identical basic head loss correlations used in BLOCKAGE, time-dependent debris build-up on the strainers based on strainer flow rate and pool water e
volume as in BLOCKAGE (with all debris assumed to be suspended in the suppression pool at time zero),
use of the full strainer surface area for debris deposition until the gaps between the stacked e
disks are filled with debris, l
use of the strainer circumscribed area for further debris deposition after the gaps are filled, a calculation of debris thickness on the outside of the circumscribed area that accounts for e
the surface curvature, and l
2
BSEP, Uzit Nos.1 and 2, ECCSSuction Str:iner Repizcement Project AppendixA NRC Bulletin %0.1 FinalReport implementation of an averaging algorithm for the debris surface to volume ratio that is e
consistent with the basic head loss correlations.
A1.3 Assumptions l
l The methodology relies on the same basic head loss correlation documented in NUREG/CR-l 6224 and implemented in the BLOCKAGE code. Validation of that correlation as documented I
in NUREG/CR-6224 and in the BLOCKAGE Code validation report [Shaffer,1996) is assumed
)
to be applicable herein.
1 1
A detailed discussion of stacked disk strainers is not provided herein. The reader who is not familiar with the basics of that strainer geometry is referred to the PCI report discussing the performance characteristics of their stacked disk strainer design [ Hart,1996).
A2.0 Methodology Description As discussed in the section on objectives, this methodology is based on the identical head loss correlation described in NUREG/CR-6224. Any enhancements are implemented exclusively in the calculation of certain terms in that correlation. The methodology is set up to perform the l
strainer performance assessment for one strainer at a time. Thus, if multiple strainer designs are being evaluated (RHR and Core Spray, for example) a separate analysis would need to be performed for each one.
1 A2.1 Basic Head Loss Correlation The NUREG/CR-6224 head loss correlation is described in detail in Appendix B to that report and is a semi-theoretical head loss model. The correlation is based on the theoretical and experimental research for the pressure drops across a variety of fibrous porous media carried out since the 1940s. This head loss model, proposed for laminar, transient and turbulent flow regimes through mixed debris beds (i.e., debris beds composed of fibrous and particulate matter) is given by:
m.5 (1+57 a 3 )
AH =A [3.5 Sy2a 1
2 U + 0.66 Sy a /(1-a ) p U ) Alm m
m m
i
- where, AH is the head loss, Sy is the average surface to volume ratio of the debris, is the dynamic viscosity of water, l
U is the fluid approach velocity, p is the density of water, um is the mixed debris bed solidity (one minus the porosity),
Alm is the mixed debris bed thickness, and A is a unit conversion factor (A = 1 for SI units).
3
BSEP, Uxit Nos.1 cxd 2, ECCS Suctio2 Strainer Replacement Project AppenditA NRCC:lletin W03 FinalReport The mixed debris bed solidity is given by:
r 3
1+ &
M.
a=
11 a Y Pp
\\
)
m
- where, I
a, is the as fabricated fiber bed solidity, do is the theoretical fibrous debris bed thickness, f
t1 = mp/mf s the particulate to fiber mass catio in the debris bed, i
pf is the fiber density, and p, is the average particulate material density.
For N classes of particulate materials, mp nd p are defined by:
a p
p N,
m, = E m, l
o.s and N,
EPo Yo
'~',
Pp
=
N E,.i l
Vo where mi, pi and V are the mass, density and volume of a particulate material i.
i l
Compression of the fibrous bed due to the pressure gradient across the bed is also accounted for.
The relation that accounts for this effect, which must be satisfied in parallel to the previous equation for the head loss, is given by (valid for (AH/Mo) > 0.5 ft-water / inch-insulation):
Mo = 13 Mm (AH / Mo)0.38 l
i For very large pressure gradients, the compression is limited such that a maximum solidity is not j
exceeded. In the NUREG/CR-6224, this maximum solidity is defined to be 3
am = 65 lb/ft /pp i
L
r I
l BSEP, UnitNos.1 e d2, ECCSS:ction Strainer Replaceme t Project AppendixA NRCE:lletin WO3 FiniiReport l
3 3
which is equivalent to having a debris layer with a density of 65 lb/ft. Note that 65 lb/ft is the macroscopic density of a granular media such as sand or gravel and clay [Baumeister,1958). In this methodology, am is considered to be a user-specified input parameter.
It should be noted that there are indications that this formulation for debris bed compression may overpredict compression significantly in the case of very thick debris layers. As such, it is possible to ignore compression in this analysis if so desired.
l The NUREG/CR-6224 model assumes that the debris is uniformly distributed on the strainer surface. For flat-disk strainers and thin layer beds (ALo s 0.125 inches), that correlation is known to over-predict the results [Zigler, et al,1995). Thus, a minimum debris thickness equal to the strainer perforated plate hole diameter is assumed to be necessary to result in any measurable head loss.
In the above formulation, three parameters have as of yet not been specified; the approach velocity, U, the theoretical (uncompressed) fibrous debris bed thickness, ALo, and the average surface to volume ratio, Sv. The approach velocity is a function of the user specified volumetric i
flow rate through the strainer of interest and the effective strainer surface area. The debris bed thickness is a function of the volume of debris on the strainer, the effective strainer surface area, and the strainer surface curvature.
A2.2 Calculation of Effective Strainer Surface Area and Approach Velocity The fluid approach velocity, U, is given simply in terms of the volumetric flow rate and the effective surface area as Q
U=A
- where, Q is the volumetric flow rate through the strainer, and A is the effective strainer surface area.
The effective area, A, is a function of the debris bed thickness, Alm. While the debris bed thickness is calculated to be less than half the width of the gaps between the disks (a user-defined input parameter), the effective area is simply set equal to the full (perforated) surface area of the strainer, As. This area is calculated based on simple geometrical considerations of the strainer dimensions. Once the debris thickness exceeds half the gap width, the effective surface area is reduced to the circumscribed area of the strainer, Ac (i.e., ignoring the surface area within the gaps between the disks). Note that in both cases, this area calculation accounts for the perforated surfaces on the ends of the strainer.
5 l-
1 BSEP, U:lt Nos.1 and 2, ECCS S:ction Strainer Rept: cement Project Appe dit A NRCEr!!etin 96-03 Fin:1 Report A2.3 Calculation of Debris Deposition on the Strainer The total quantities of fibrous debris and particulate assumed to be suspended in the pool at time zero must be specified. In order to calculate the rate at which this debris accumulates on the strainer being analyzed, one has to specify the volumetric flow rate through the strainer, Q, the total volumetric flow rate through all strainers, Qtot, and the total pool water inventory, V ool.
p If the flows are constant, the fractional debris deposition on the strainer at any given time, t (user input), into the accident can be calculated assuming that all debris in the pool stays uniformly distributed in the pool, e
debris is removed from the pool (deposited on all strainers) in proportion to the rate at which e
water is removed from the pool, and debris is deposited on the strainer being analyzed in proportion to the flow rate through that e
strainer versus the total flow rate.
This results in the following equation to describe the fractional debris buildup on the strainer of interest:
r 3f 3
FRAC =
Ga
- - exp( V'"
- t)
\\
s<
m This fraction is applied uniformly to the total quantity of each debris constituent (fiber and all particulate species) initially in the pool.
In the case where a time-dependent (but step-wise constant) flow is specified in the input, each time interval of constant flow (out to the user-specified problem time) must be treated separately.
The fractional deposition during an interval is applied to the total debris quantity in the pool at the start of the interval. These deposition quantities can then be summed to yield the total deposition for the problem time ofinterest.
i A2.4 Debris Thickness Calculation - Effective Surface Area Consideration As a rough approximation, it would be expected that the theoretical fibrous debris bed thickness l
on the strainer would be given by the volume of fibrous debris divided by the strainer surface i
area. This is in fact the approach taken in the BLOCKAGE code. In the case of a stacked-disk strainer, however, it is clear that the debris thickness would be underpredicted by such an
{
approximation once the gaps between the disks are filled. Thus, the calculation of theoretical 1
fibrous debris bed thickness is accomplished in two steps.
Sten 1 - It is first necessary to calculate a theoretical debris thickness using the entire strainer surface area, As. Thus, for a total fiber volume on the strainer of Vr, the thickness would be calculated as l
ALo = Vr/ As.
6 L__ _ _ _
j
i BSEP, U:lt Nos.1 cnd 2, ECCSSuction Str:iner Replaceme:t Project Appe:dixA NRCBulletin 96-03 FinalReport 1
If following the correction for bed compression, the debris bed thickness, Mm, is less than half the width of the gaps between the disks, the calculation of theoretical debris bed thickness is complete.
Steo 2 - If the thickness calculated in step i exceeds the maximum allowable, a correction must be made. To do this it is necessary to first calculate the volume of fiber in the gaps between the disks. This is simply given by V' = V ap
- Mo/Mm g
- where, V
is the total volume associated with the interstitial gaps between the disks. This quantity is cake lated explicitly based on the strainer geometry.
The theoretical thickness of the fibrous debris on the outside of the strainer circumscribed surface is then given by Mo = (V - V') / Ac f
- where, Ac is the circumscribed surface area of the strainer.
A2.5 Debris Bed Thickness Calculation - Surface Geometry Considerations The discussion on fibrous debris bed thickness above is only qualitatively correct. Simply dividing the fiber vclume by the strainer surface area would accurately predict theoretical thickness if the surface geometry were planar. This is a good approximation when one is dealing with the entire strainer surface area (step 1 above), since much of the disk area is associated with the flat surfaces of the disks. However, once the gaps are filled, most of the deposition effectively occurs on the outside of a cylinder, and the debris bed thickness is less than that predicted above.
To account for this effect, the debris bed thickness can be calculated assuming that the thickness of the fiber is the same on the ends of the strainer (flat surfaces) as on the outside of the l
circumscribed cylindrical surface. With this assumption, it is a simple matter to calculate the total volume of a debris layer of thickness Mo, equate this to the known volume of the fiber (reduced by the fiber in the interstitial gaps), and solve the resulting quadratic equation for the debris bed thickness. This result is given by:
AL, = }(xDH+ Ap)2 + 4(y,, y.) xH -(xDH + Ay) 2xH 7
RSEP, U:lt Nos.1 c d 2, ECCS S::ction Strainer Replacement Project AppenditA NRCBulletin W(U FinalReport
- where, D is the outer diameter of the strainer disks (assumed to be uniform axially),
H is the active length of the strainer, and fl i A at s the flat surface area of the ends of the strainer (calculated automatically).
This equation is used in lieu of the simpler expression previously presented in Section 2.4 under step 2.
A2.6 Consideration ofImpact of Strainer Supports For relatively small strainer.4, the only connection to the strainer is the inlet tube to the strainer.
However, for larger strainers, it may be necessary to provide additional supports for structural reasons. Depending on the details of these supports, they can have a small impact on the effective strainer total surface area, AAs, circumscribed area, AAc, and interstitial gap volume, AV ap. To account for this effect, it is necessary to quantify these reductions. These reductions g
are then taken into account in the calculation of theoretical debris bed thickness by setting A, -+ A, - M, in the equation describing the debris thickness over the entire strainer surface area (Section 2.4, Step 1), and setting xDH H -+ H
- xDH-M,
~bY y,4 y,, Yw n
V, in the equation describing the debris thickness once the interstitial gaps have been filled (Section 2.5).
A2.7 Calculation of Average Debris Surface to Volume Ratio The intuitive choice for the average surface to volume ratio is to use a volume weighted average, which is equivalent to defining the average by the total surface area divided by the total volume.
Unfortunately, such a choice is inconsistent with the formulation of the basic head loss correlation and can lead to non-physical results. Presented below is a simple derivation of an alternate averaging scheme that provides consistency with the basic form of the correlation.
Consider a fixed volume V that contains a mixture of two types of fiber, I and 2. We take Si = the surface to volume ratio of fiber type i, and 8
L -- -___________ _-_-__
- BSEP, U:lt Nos.1 and 2, ECCSSuction Strainer Rept: cement Project Appe::dirA NRC Bulletin 96-03 FinalReport qi = the quantity (microscopic volume) of fiber type i.
If we consider the dominant terms in the relationship for the pressure drop, AP, we have AP ~ 2
- a1.5
- where,
is the average surface to volume ratio, and a is the total solidity which is (q) + q2)/V.
We now consider the following argument for determining the average . Rather than having the two fiber types well mixed in region V, consider that the fibers are separated into adjacent regions V and V (with the constraint that the sum of the two volumes be V), with part of the I
2 flow going through each region (in parallel rather than in series). We now postulate that the flow velocity through each of these two regions is the same, and that the pressure drop across the two regions is also the same. This allows us to solve for the relative volumes V and V. One form I
2 of this relationship can be expressed as 3 4/3
[ = 1 + 92(82 V,
q, r So if we once again consider the two fiber constituents to be well mixed, and again postulate the pressure drop to be unchanged and the flow velocity to be unchanged, we can equate the pressure drops calculated from the two different viewpoints to yield I
e ><,
(S)' = Sl
< 9 > < Vu Substituting for the volume ratio, we are left with
- 3/4 A l4S'
(S) =
S 2
9 9
1 1
It is obvious how this result could be extended to more than two such fiber species. We further make the assumption that the same formulation can be applied when some of the species are
{
particulate rather than fiber.
l j
i 9
1 t
________-____a
BSEP, Uzit Nis.1 cxd 2, ECCS Srction Strainer Repl: cement Pro}ect Appe:dixA NRCC:lletin 96-03 Fin:1 Report A3.0 References Baumeister, T., Mechanical Engineers ' Handbook, Sixth Edition, McGraw-Hill Book Company, 1958.
Ilart, G., "The Development and Testing of Performance Contracting, Inc. 's Sure-Flow StackedDisk Suction Strainer", Performance Contacting, Inc., February 1,1996.
Rao, D. V., et. al., " BLOCKAGE 2.5 User's Manuaf', NUREG/CR-6370, U.S. Nuclear Regulatory Commission, December,1996.
Shaffer, Clint, et. al., " BLOCKAGE 2.5 Reference Manual", NUREG/CR-6371, U.S. Nuclear l
Regulatory Commission, December,1996.
1 Zigler, G., J. Brideau, D.V. Rao, C. Shaffer, F. Souto, and W. Thomas, " Parametric Study of the Potentialfor BWR ECCS Strainer Blockage Due to LOCA Generated Debris", NUREGICR-6224, U.S. Nuclear Regulatory Commission, October 1995.
i 10
v BSEP, Unit Nos. I c:d 2, ECCS S:ction Str:lner Repl1 cement Pro}ect NRC Bulletin 96-03 FinalReport l
l l
l l
Appendix B Sure-Flow" Strainer Performance Test Reports 1
l Reports:
1 1.
Performance Contracting, Inc. Report," Summary Report on Performance of Performance I
Contracting,Inc.'s Sure-Flow Suction Strainer with Various Mixes of Simulated Post-l LOCA Debris," Revision 1, dated September 19,1997 2.
Continuum Dynamics, Inc. Report, " Performance Contracting, Inc. ECCS Sure-Flow" Strainer Data Report," No. WO4536-01, dated December 1996 l
i l
I 1
l 1
I
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _,