ML17229A900
| ML17229A900 | |
| Person / Time | |
|---|---|
| Site: | Saint Lucie  |
| Issue date: | 10/30/1998 |
| From: | Stall J FLORIDA POWER & LIGHT CO. |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| GL-96-06, GL-96-6, L-98-276, NUDOCS 9811050017 | |
| Download: ML17229A900 (77) | |
Text
I CATEGORY 1 f,~i " 1 REGULA'I RY INFORMATION DISTRIBUTION SYSTEM (RIDS)
ACCESSION NBR:9811050017 DOC.DATE: 98/10/30 NOTARIZED: NO DOCKET FACIL:50-335 St. Lucie Plant, Unit 1, Florida Power & Light Co. 05000335 50-)89 St.'ucie Plant, Unit 2, Florida Power & Light Co. 05000389 AUTR. NAME 'UTHOR AFFILIATION STALL,J.A. Florida Power &. Light Co.
RECIP.NAME RECIPIENT AFFILIATION Records Management Branch (Document Control Desk)
SUBJECT:
Provides supplement to licensee response to GL 96-06, C specifically event scenario, event description, method analysis,imput parameters, assumptions, engineering judgement
&. uncertainties within analysis.
DISTRIBUTION CODS: A072D TITLE: GL COPIES RECEIVED: LTR 96-06, "Assurance of Equip Oprblty &
l ENCL J SIZE:
Contain.Integ. during Design NOTES: Q RECIPIENT COPIES RECIPIENT COPIES 0 ID CODE/NAME LTTR ENCL ID CODE/NAME LTTR ENCL NRR/WETZEL, B. 1 1 PD2-3 PD .1 1 GLEAVES,W 1 1 INTERNAL LE ENTER 01 1 1 NRR/DE/EMEB 1 1 NRR 1 1 NRR/DSSA/SPLB 1 1 EXTERNAL: NOAC 1 1 NRC PDR 1 1 D
C U
M E
N NOTE TO ALZi "RZDS>> RECIPIENTS PLEASE HELP'US TO REDUCE WASTE. TO HAVE YOUR NAME OR ORGANIZATION REMOVED FROM DISTRIBUTION LISTS OR REDUCE THE NUMBER OF COPIES RECEIVED BY YOU OR YOUR ORGANIZATION, CONTACT THE DOCUMENT CONTROL DESK (DCD) ON EXTENSION 415-2083 TOTAL NUMBER OF COPIES REQUIRED: LTTR 9 ENCL 9
t r
Florida Power & Light Company, 6351 S. Ocean Drive, Jensen Beach, FL 34957 October 30, 1998 L-98-276 10 CFR 50.4 U. S. Nuclear Regulatory Commission Attn: Document Control Desk Washington, D. C. 20555 RE: St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 Request for Additional Information Generic Letter 96-06 Res onse This letter provides a supplement to Florida Power & Light Company (FPL) response to Generic Letter (GL) 96-06, Assurance ofEquip)nent Operability and Containment Integrity During Design-Basis Accident Conditions, for St. Lucie Units 1 and 2. By letter dated July 28, 1998, the NRC issued a request for additional information (RAI) after reviewing the original FPL responses L 280 dated October 28, 1996, L-97-18 dated January 28, 1997, and L-97-114 dated April 22, 1997. The NRC letter established a target for FPL to respond to the RAI by September 30, 1998.
In a September 22, 1998, conference call, Mr. W. C. Gleaves (NRC) agreed to extend the due date to October 30, 1998. FPL provides the attached responses to the RAI concerning the event scenario, event description, method of analysis, input parameters, assumptions, engineering judgement, and uncertainties within the analysis. FPL has modified the analysis to address the following issues.
During the development of response to the NRC RAI, FPL determined that the initial evaluation was not conservative regarding the projected timeframe for component cooling water (CCW) system pressurization following the design basis accident with loss of offsite power. SpecificaHy, the evaluation used a nominal value for emergency diesel generator (EDG) start time and CCW pump load block loading. This approach failed to consider time delays for the receipt of a safety injection actuation signal (SIAS) or an undervoltage (UV) signal or tolerances in relay settings.
The increase in the time duration for CCW system pressurization would result in additional heating prior to the increase in system pressure and would exacerbate the potential for voiding and waterhammer.
A subsequent evaluation has been completed addressing the additional CCW pump coastdown and the time for EDG start/CCW system pressurization. To conservatively analyze the effect'of pump coastdown, the revised analysis utilized enhanced heat transfer rates resulting from the higher velocities; however, the additional heat capacity of cold water entering the containment fan coolers (CFC) was not credited. To conservatively analyze the time for CCW pump restart, the maximum tolerance for each timeblock (e.g., relay tolerance) was utilized. The times used within the analysis (18.4 seconds for Unit 1 and 18.5 seconds for Unit 2) are conservative with respect to typical test times.
This analysis based on the design basis pump start times and on mechanistic modeling of pump coastdown found that CFC voiding could be expected foHowing a design basis accident. Voiding creates the potential for waterhammer loading on pump restart which has been addressed by additional analysis. The resulting system pressures, pipe stresses, and support loads have been addressed under functionality rules established for the St. Lucie Plant. System operability in the recovery period immediately following pump restart was reviewed and previous conclusions that two-phase flow would not occur and CCW pump NPSH would be adequate were found to be valid.
98ii050017 98i030 .'
PDR ADGCK 05000335 P PDR an FPL Group company
't. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Page 2
's peart of the long term corrective action, FPL willjoin the group of utilities working with Nuclear Energy Institute (NEI) and the Electric Power Research Institute (EPRQ to develop a design based version of the NUREG/CR-5220, Diagnosis of Condensation-Induced Waterhammer, Case Studies. This approach was presented to the NRC at an industry workshop in Washington, DC in May 1998, and is intended to assist the utility industry in developing waterhammer analysis techniques which are better suited for design purposes. Accordingly, FPL will develop a plan for the resolution of this issue in the long-term based on the outcome and schedule developed for the work scope of this industry group.
Please contact us ifthere are any questions regarding this supplement.
Very truly yours,
. A. Stall Vice President St. Lucie Plant JASIGRM Attachment cc: Regional Administrator, Region II, USNRC Senior Resident Inspector, USNRC, St. Lucie Plant
I
~l
'L ts V
I s
't.Docket Lucie Units Nos.
1 and 2 50-335 and 50-389 L-98-276 Attachment 1 Page 1 St. Lucie Units 1 and 2 Generic Letter 96-06 NRC Request for Additional Information Response Introduction FPL provided the initial response to NRC Generic Letter (GL) 96-06 containment fan cooler voiding and waterhammer concerns in FPL letter L-97-18 dated January 28, 1997. For this response, FPL analyzed the affects on the St. Lucie Units 1 and 2 component cooling water system (CCW) for a loss-of-coolant accident (LOCA) and main steam line break (MSLB) accident with loss of offsite power (LOOP). The analysis was intended to determine ifvoiding would occur in the CCW system within the containment fan coolers (CFC) during a design basis accident (DBA) transient. The response also addressed whether sustained two-phase flow would occur during the recovery phase from the initial transient and whether CCW pump net positive suction head (NPSH) would be assured during the recovery phase.
The initial response concluded that no steam generation would occur within the bounding HVS-1C cooler for either Unit prior to the time (16 seconds) the CCW pumps restarted. Of the cases reviewed, the Unit 1 LOCA scenario (Case 2) had the smallest subcooling margin - the margin to boil was 0.1 seconds. The potential for two-phase flow during the Case 2 recovery period was evaluated for CFC HVS-1C at two critical locations in the discharge piping; at the highest location in the discharge piping and at a position just downstream of the throttle valve and flow element orifice in the CFC return piping. Two-phase flow did not occur at either location. A CCW pump NPSH margin review determined that CCW pumps would not be adversely affected by flushing hot CFC water, mixed with flow from the shutdown cooling water (SDC) heat exchanger, into the CCW pump suction during the recovery period.
NRC issued an RAI concerning the St. Lucie Plant response on July 28, 1998. During the course of developing the additional information requested, a full review of the FPL evaluation was conducted to provide complete documentation of input parameters, methodology, and results and to respond to the NRC request for discussion of assumptions, engineering judgement, and uncertainties within the analysis.
The review of the analyses found the methodology to be conservative and appropriate for the given design and input assumptions. The bases for a majority of the assumptions were also found to be conservative. However, this review identified certain inputs to the original analysis that were non-conservative and would require changes.
~ The design basis value for the CCW pump start was found to be longer than that assumed in the initial analysis. This allows a longer period for heatup of the water in the CFC tubes and increases the potential for voiding.
The change in the pump restart time was a result of further investigations which indicated that the bounding time for restart of the pumps are 18.4 seconds and 18.5 seconds for Units 1 and 2 respectively. In comparison to the 16 second time used as a basis to develop the conclusions presented in the initial response, the additional delay time for the pump restart exceeded the margin in the time to boil determined in the original analysis. The extended time for pump restart is primarily a result of a time delay for receipt of an emergency diesel generator (EDG) start signal (SIAS or UV) and relay tolerances. The source of this non-conservatism was an inappropriate use of nominal design information without the consideration of signal delays and relay tolerances.
1
~
~
'I h
"St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 I 98-276 Attachment 1 Page 2
~ Mechanistic analysis of the flow coastdown transient indicates that flows to the CFCs persist for considerably longer periods than the two seconds assumed in the initial analysis. The increase in flow rate raises heat transfer coefficients within the tubing and increases the potential for voiding.
In the initial analysis, it was believed that since the voiding phenomenon was due to loss of system flow to the CFC, quicker coastdown would result in less time to reach boiling. The coastdown time assumed was somewhat arbitrary, but was based on observations of the time required to see a pump stop spinning when tripped. In addition, the two second assumption was consistent with the assumed pump coastdown times for other stations performing similar analyses. For short pump coastdown times, this was not considered to be a sensitive parameter.
However, by more rigorously examining intermediate coastdown times, certain flow versus time'rofiles resulted in a more severe heatup transient and resulted in voiding within the CFCs at earlier times.
This can be understood as follows:
~ For short coastdowns, CFC tube water stagnates early in the transient, inside heat transfer coefficients are relatively low and extend the time to boiling.
~ For extremely long coastdown times, CFC tube velocity is maintained throughout the transient (i.e., until pump restart). CFC tube water is continuously flushed from the tube as cold water is introduced. While inside heat transfer coefficients are high, CFC tube water residence time is reduced.
~ For intermediate coastdown periods, high inside heat transfer coefficients are maintained throughout the transient, but tube velocities decay at such a rate that the hot water in the tube is not completely flushed. The increased heat transfer rates for this case more than compensate for the introduction of cold water, and the time to reach boiling is reduced relative to the stagnant water case.
As discussed by telephone with the NRC Project Manager, W.C. Gleaves on September 22, 1998, and the NRC Senior Resident, T. Ross, on September 25, 1998, FPL has identified that the initial response analysis failed to accurately assess these issues.
A subsequent evaluation has been completed which considers the additional CCW pump coastdown and timeframe for EDG start/CCW system pressurization. To conservatively analyze the effect of pump coastdown, the revised analysis utilized the enhanced heat transfer rates resulting from the higher velocities, however, the additional heat capacity of cold water entering the CFCs was not credited. To conservatively analyze the time for CCW pump restart, the maximum tolerance for each timeblock (e.g., relay tolerance) was utilized. The times used within the analysis (18.4 seconds for Unit 1 and 18.5 seconds for Unit 2) exceed typical test times by 1.5 seconds to 2.0 seconds.
Analysis based on the new design basis pump start times and on mechanistic modeling of pump coastdown found that CFC voiding could be expected following a design basis accident (DBA).
Voiding creates the potential for waterhammer loading on pump restart which has been addressed by additional analysis. The resulting system pressures, pipe stresses, and support loads can be
S l
~.
ll I ~
II aS
~c'
~
,Yf
<<fa I'
,a
St. Lucie Units 1 and 2
,, Docket Nos. 50-335 and 50-389 I 98-276 Attachment 1 Page 3
'n the'recover period accommodated under functionality rules established for the immediately following pump restart St. Lucie Plant. System operability was reviewed and previous conclusions that two-phase flow would not occur and CCW pump NPSH would be adequate were found to be valid.
FPL willjoin the group of utilities working with Nuclear Energy Institute ~
Power Research Institute (EPRQ to develop a design based version of NUREG/CR-5220, and the Electric Diagnosis of Condensation-Induced Waterhammer, Case Studies. This approach, as presented to the NRC in an industry workshop in Washington, DC in May 1998, is expected to assist the utility industry in developing waterhammer analysis techniques which are better suited for design purposes. Accordingly, FPL will develop a plan for the resolution of this issue in the long-term based on the outcome and schedule developed for the work scope of this industry group.
This Attachment provides the FPL response to the RAI concerning the event scenario, event description, method of analysis, input parameters, assumptions, engineering judgement, and uncertainties within the analysis. FPL has modified the operability analysis to address the previously described issues.
Res onse to S ecific RAI uestions:
NRC Question 1:
Provide a detailed description of the "vorst case" scenario for vaterluunmer and tvo-phase flow, taking into consideration the complete range of event possibilMes, system configurations, and parameters. For example, the worst-ease temperatures, pressures, flov rates, load combinations, and component failures shouM be considered. Describe hov mucls margin to boiling villexist for these scenarios.
FPL Response 1:
Back round S stem Confi ration The CCW systems for Units 1 and 2 are similar in layout and design. They are closed loop systems which provide cooling water flow to essential plant auxiliaries via two parallel safety related trains. The two essential trains are connected via return and supply headers with a non-essential train (N-header) arranged between them, During normal operation, all three trains are supplied with cooling water via two CCW pumps. A third pump functions as an installed spare.
Automatic pump start of the spare pump is not a design feature.
On a SIAS, the two essential trains are isolated from the N-headers by spring operated butterfly valves. Closure of these valves also serves to separate the essential trains into two redundant safety related A and B trains. The A and B trains serve essential services such as the containment coolers, shutdown heat exchangers, ECCS pump seals, and the control room air-conditioners PJnit 2 only]. The common atmospheric pressure surge tank for each unit is subdivided by a baffle and has an independent surge line serving each essential train.
A spent fuel pool heat exchanger is located on the N-header for Unit 1. The fuel pool heat exchangers for Unit 2 are normally served by a single essential header; these heat exchangers are isolated on SIAS. For both units, containment fan coolers HVS-1A and 1B are supplied from CCW Train A and HVS-1C and 1D are supplied from CCW Train B.
I 4i-k
\ Y A
4I
'H 1 J
sC ill
St. Lucie Units 1 and 2
,, Docket Nos. 50-335 and 50-389 L-98-276 Attachment 1 Page 4
'he CCW system is assumed to be operating in the normal configuration prior to the event with flow through both essential headers and the N-header. Within the design basis for each Unit, water is always flowing to all four CFCs when the unit is on line. On a SIAS, the two essential trains ("A" and "B") are isolated from each other and from the non-essential header ("N-header")
by redundant, spring-to-close, fail-close butterfly valves. On a SIAS, a single spring-to-open, fail-open butterfly valve is opened within each essential train to provide flow to the shutdown heat exchangers. The transient hydraulic analysis considers these automatic valve position changes.
There are no other special system configurations to be considered.
"Worst Case" Scenario The worst case scenarios evaluated for waterhammer and two-phase flow considered the most severe containment response coupled with a loss-of-offsite power (assumed to occur concurrent with the start of the DBA event). These UFSAR containment analyses for LOCA and MSLB provide the most severe containment response for containment peak pressure. These containment analyses provided the maximum post DBA temperature profiles for use in the waterhammer and two-phase flow evaluations. The assumption of a LOOP coincident with the DBA provided the longest duration for a loss of CCW and is consistent with the approach used in other design basis analyses.
The analysis assumes normal operation of system components in order to demonstrate that the event can be accommodated by the system (i.e., the system remains operable and a common mode failure is not credible.) Single failures within one train can be accommodated with respect to loss of cooling function (i.e., bounded by loss of the train) provided containment integrity is maintained. Accordingly, single failures were treated in a matrix as a range of event possibilities with their effect considered. See the response to RAI Question 3.
For the 16 possible cases (2 Units, 4 CFCs, LOCA/MSLB), evaluation results indicate that prior to CCW pump restart, boiling and void generation will occur in 5 of the 16 cases affecting both CCW trains in both units. For the remaining 11 cases, the margin to voiding for CFCs with no computed voids is stated in terms of the number of seconds to reach voiding relative to the assumed pump start and system pressurization time. Eight of the eleven cases which did not boil prior to pump restart had a margin to boiling of greater than 6 seconds. One of the scenarios did not have any margin in the time to boil (i.e., boiling begins immediately after the time at which pump restart is assumed to occur). The two remaining cases had margins to the onset of boiling greater than 1 second. Twenty-three figures providing plots of system input and output parameters with respect to time for the duration of the event transient are provided in Attachment 2.
l t
.P
~g p ~
Ip 4
1 V
At I
I l
I
~
Pa ly, Jt
St. Lucie Units 1 and 2
,, Docket Nos. 50-335 and 50-389 L-98-276 Attachment 1 Page 5 Time-to-Boil and Void Size Determination LOCA Cases wfth CFC Time-to-Boil Void Sze based on CGA System Repressurization Timeframe LOOP Unit I Train A 14.5 sec 21.95 sec 6.1 ft'oid expected in HVS-1A No voiding expected in HVS-1B Unit 1 Train B 12.25 sec >25 sec 12.9 fts void expected in HVS-1C No voiding expected in HVS-ID Unit 2 Train A 17.70 sec >25 sec 0.6 ft'oid expected in HVS-1A No voiding expected in HVS-1B Unit 2 Train B 14.7 sec >25 sec 5.5 ft'oid expected in HVS-1C No voiding expected in HVS-1D MSLB Cases arith CFC Time-to-Boil Void Size based on CGA System Repressurization Timeframe LOOP Unit 1 Train A >25 sec No voiding expected in HVS-1A No voiding expected in HVS-1B Unit 1 Train B >25 sec >25 sec No voiding expected in HVS-1C No voiding expected in HVS-1D Unit 2 Train A 19.65 sec >25 sec No voiding expected in HVS-1A No voiding expected in HVS-1B Unit 2 Train B 17.9 sec 18.5 sec>> 0.2 ft'oid expected in HVS-1C No voiding expected in HVS-1D>>
>> Zero Margin for this case Design input and modeling relative to pressures, temperature, flow rates, surge tank operating level, and event timing are addressed under RAI Question 2. The impact on the calculated margins due to uncertainties in the calculation are discussed in the response to RAI Question 4.
A discussion of the modeling methodology used for the current operability analysis is provided below.
Methodolo for Time-to-Boil and Voidin Anal sis The analysis was performed to model the CCW system response for DBA/LOOP transient in the time period immediately following CCW pump trip. Two aspects of system operation were examined in this analysis:
- 1. Determination of the time-to-boil for each CFC for DBA/LOOP transients
The Sargent & Lundy CFC program employed in the calculation conservatively treats the CFC fluid volume as a single node at a uniform temperature and pressure for any given time step. The highest temperature fluid in the CFC will exist near the exit of the tube (i.e., at the return side.)
Following CCW pump coastdown, the static pressure distribution will be nearly uniform along the tube length. Therefore, the fluid temperature near the return will approach the saturation temperature more quickly than locations upstream. This may result in localized boiling at the CFC return, which occurs before bulk boiling of the entire CFC. This tendency is more pronounced for longer duration CCW pump coastdown periods where a greater CFC fluid temperature gradient can be established along the CFC tube length.
The vapor produced during any localized boiling willlikely rise into and condense in the relatively colder water in the return-side collection manifold of the CFC unit. Since lower elevation CFC tubes or supply/return manifolds are not modeled, this effect cannot be observed or credited.
However, detailed analyses performed by Sargent & Lundy for other clients have shown that the duration of localized vapor voiding in the CFC tubes is short-lived, with a relatively quick transition to saturated bulk boiling along the CFC tube length. This is followed by bulk boiling in lower elevation CFC coils, progressing from the top down with time. Significant vapor generation, leading to steam migration to the CCW system piping, does not occur until saturated bulk boiling begins in the CFC unit.
l hi+
V il 'pi a,,
tt Wv r&
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment 1 Page 6
'he c'ode used to model the CFC neglects the heat capacity of the incoming water. As a result, any local boiling that would occur at the exit of the tubes would be conservatively applied to the entire mass of water in the CFC. Furthermore, the flushing of hot water from the CFC tubes was also not credited. As a result, the temperature response calculated should conservatively bound the expected response of the CFCs following a DBA event concurrent with a LOOP.
Methodolo for H draulic Transient Anal sis Analysis was performed to model the CCW system hydraulic response for the time period immediately following CCW pump trip and following pump restart after DBA/LOOP. Two aspects of system operation were examined in this analysis:
- 1. Determination of CCW pump coastdown profile for input to Time-to-Boil and Voiding Analysis
- 2. Determination of waterhammer forcing functions following CCW pump restart (ifvoids are present) for input into Piping Stress Analysis The Sargent & Lundy HY'IRAN computer program is used to perform hydraulic analyses and waterhammer fluid analysis. The HYTRAN program implements the fixed-grid method of characteristic solution technique to calculate pressures and flow velocities in a network of pipes.
These quantities are then used to calculate the time-varying forces on the pipe segments that comprise the network.
An analysis for the Unit 1 CCW B train piping system was performed for a pump trip followed by restart after a coastdown interval of 18.5 seconds. The pressure and flow time histories generated for the HVS-1C and HVS-1D cooler outlets by this analysis were provided as an input to the Time-To-Boil and Voiding Analysis, as previously described. This latter analysis determines the void size that occurs due to boiling in the CFCs, as well as, the pressure time history during the boiling process. The pressure time history is provided as input into the HYTRANanalysis and the Voids Collapse Analysis is developed by HYTRANfor the time period following pump restart. The resulting waterhammer force time histories are used as input for the piping stress analysis.
The Voiding Analysis shows that once boiling begins, the pressure in the CFC discharge nozzle ramps upward to about 30 psia (saturation pressure). This pressure remains constant throughout the entire event until the steam void collapses. As the refill and downstream columns of water compress the void, the dissolved gas in the fluid that has come out of solution during boiling will be compressed. This cushion of compressed gas will allow the pressure in the void to increase slowly relative to water rejoining without an intermediate pocket. Consequently, the waterhammer forces on individual piping segments and the pipe support loads are significantly reduced. This effect was demonstrated by the GL 96-06 work Sargent & Lundy performed for Zion in Calculation No. 22S-B-022M-631, Rev 0, For Zion, a MK.AP analysis was made which took credit for the non-condensable air in solution prior to boiling. The effect of this air becoming released from the water after boiling occurred, slowed the pressure rise on void collapse to more than 60 milliseconds. The HYTRANanalysis for the St. Lucie GL 96-06 RAI response is based on the RELAP work completed for Zion.
l~-
E J
~ll i"7 7
'k
St, Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389
~ '-98-276 Attachment 1 Page 7 While this cushioning effect is known to occur in practice, the HYTRAN analysis only takes
'artia'1 credit for the phenomena. The HYATTanalysis makes a conservative assumption that the pressure remains at the saturation value of 30 psia and does not use the higher pressures (about 230 psi) that occur just prior to void closure that would reduce waterhammer loads. In order to compensate for this neglect of the cushioning effect, a reduced value of sound speed of 1000 fps is used in the pipe segments where boiling has occurred. Also, the theoretically instantaneous void closure that would be used for waterhammer with a cold fluid is assumed to take 60 milliseconds to complete. For purposes of calculating conservative pipe support loads, these are the same assumptions that were used in Sargent & Lundy's analysis work for Zion.
Methodolo for Pi e Stress A Pi e Su ort Anal sis The HYTRANprogram is used to simulate a hydraulic transient in train B of the Unit 1 CCW system. The developed waterhammer force time histories are then applied to a structural model of the piping associated with the containment fan coolers using the Sargent & Lundy PIPSYS program. The PIPSYS program uses a linear 3-dimensional finite element analysis program to analyze piping systems. For this calculation, a direct integration option of PIPSYS is used to determine the effects of fluid forcing functions acting on each pipe segment.
Waterhammer pipe stresses and support loads were calculated for Unit 1 CCW Train B as the void sizes were largest for this train. The piping geometry, diameter, thickness, and support/restraint locations, type, and direction are modeled within PIPSYS based on the FPL Stress Analysis of Record stress isometrics.
The scope of the piping modeled within the Unit 1B train for the waterhammer analysis included both supply piping for the HVS-1C and HVS-1D CFC inlet nozzles and the return piping for the HVS-1C and HVS-1D cooler outlet nozzles through the containment penetration to the 20-inch train B supply and return headers.
Termination of the model(s) at this point was determined to be reasonable since: (1) a sufficient number of restraints in the three orthogonal directions are installed on the 20-inch header piping, in the vicinity of the tie-in, which provide the necessary "anchoring effect" to terminate the analysis at this node; (2) the pressure pulse subsides in the 20-inch header; (3) although slightly larger loads may be developed in the 20-inch header, the substantially larger sizes and capacities of pipe supports on the 20-inch header will more than accomodate the increase in loads.
Longitudinal pressure stress in the pipe is calculated from the maximum pressure pulse obtained from the hydraulic analysis. The maximum deadweight stresses are then taken from the respective Stress Analysis of Record, combined with the longitudinal pressure stress, and the maximum hydrodynamic waterhammer stresses. UFSAR Section 3.1.2 states that systems and components vital for the mitigation of accident conditions shall be designed to withstand the effects of a LOCA coincident with the effects of design basis event (DBE). Hydrodynamic loading occurring concurrently with DBE loading is not considered credible since the waterhammer event is a consequence of the LOCA, both the waterhammer and DBE events are of relatively short duration, and the evaluation is for functionality considerations only.
Conservatively, pipe stresses are combined considering the highest value from each of the respective analyses, regardless of the node at which they occur, within the bounds of affected modeled) piping, and compared to the functionality set allowables established for 'hydrodynamic the St. Lucie Plant.
J,
~ C I
04
St. Lucie Units 1 and 2
, Docket Nos. 50-335 and 50-389 I 98-276 Attachment 1 Page 8
'oAffectedoriginal the pipe supports/restraints design are evaluated for applicable design loads by either comparison for which they are qualified or further rigorous analysis, as necessary.
These analysis results are compared to the functionality set allowables established for the St. Lucie Plant.
The functionality criteria utilized at St Lucie Units 1 and 2 is largely based on the Service Level D design criteria. It relies on ASME Section IIIAppendix F criteria for the evaluation of supports and the plastic (strain-contmlled) behavior of piping systems which tend to reduce loading through energy absorption and frequency shifting. This is consistent with the functionality evaluation criteria endorsed by the NRC in Generic Letter 91-18.
Methodolo for Recove Anal sis Two-Phase Flow and NPS Analysis was performed to model the CCW system response in the time period immediately following pump restart (and void collapse, if voids are present) for the DBA/LOOP transient.
Three aspects of system operation are of interest for this condition and were examined in the calculation:
- 1. Heat removed from the containment atmosphere by the CFCs in the early portion of the LOCA/MSLB transient and how it relates to that assumed in the design basis containment analysis,
- 2. Potential for (sustained) two-phase flow and associated two-phase frictional pressure drop in the discharge piping between the CFCs and the main CCW return header and how this may affect CFC flow and system operation, and
In general, there is substantial margin in the system design to recover to design performance following pump restart. Due to this, a simplified, bounding analysis was used. A composite CCW system using limiting input from either unit was analyzed as an enveloping case. Margin was included in the recovery analysis, both in the use of conservative assumptions and design input: (1) CFC performance was based on design fouling (to estimate recovery timeframe); (2) system flows were combined in a conservative way (maximum, minimum); (3) composite models were used (taking conservative, enveloping input for either unit), surge tank level was taken as low level alarm setpoint; (4) CFC discharge water temperatures were set to upper bound containment LOCA limits; (5) CCW pump was degraded curve by 10% on head; (6) pressure drops were conservatively calculated using maximum or minimum flowrates, as appropriate; (7)
NPSHwas required to exceed NPSH~ by 25%.
NRC Question 2:
Describe and justify all assumptions and input parameters (including those used in any computer codes) that vere used in the waterfuvnmer aIul two phase flow analyses. Confirm that these assumptions and input parameters are consistent with the existing design and licensing basis of the plant. Allexceptions should be explained andj ustified.
FPL Response 2:
Design input, and modeling relative to pressures, temperature, flow rates, surge tank operating level, and event timing used within the analysis are consistent with the existing design and licensing bases. Curves from the UFSAR for LOCA and MSLB containment transient response (pressure and temperature) are utilized as design inputs for these analyses. System flowrates are
t 04
,a
~ 'I f
8 ~
v 'A' R
h(
l 14 b gQ
<<..l P
VQ
) 'E
St. Lucie Units 1 and 2
, Docket Nos. 50-335 and 50-389 L-98-276 Attachment 1 Page 9 based on values from Technical Specifications, UFSAR, plant setpoints, equipment performance
'ata, 'equipment geometry, and surveillance procedures as appropriate. Where a range of values are allowed, the minimum, maximum or representative value is utilized as appropriate.
Input parameters for analyses are described in the Table below and in the response to RAI Question 4. Assumptions are described in response to RAI Question 7 concerning "engineering judgement."
sis In ut Parameters::::.:::::: Piseusstpn "'i'I~i'"~'i""~ "" """"""~"""'~")""'" ':i .ii.i ~"'t'i:: ':
Contauunent DBA Pressure and Temperatures Curves Per UFSAR curves or worst case scenarios ui ment an S stem Geome 2 evations Per i ment an ant win s Routin o CCW es Per ui mentan ant rawin s S stem ow Rates Per ran e ow in OP I 2%310020 A F Ec G Valve Stro e Times Per ran e ow in OP 1 2~101 A Sur e Ta Level Per low arm s oint eve nunus se oint tolerance CFC Fou n Factors Assumed to be zero Imti Containment Te erature 120'F- Tec S ec Maximum Imti Contamment Re ative Hunu 4 FSAR CCW Pu Hea NPSH Vendor Manu Drawin s CCW Pum Coastdown Pro e Calculated bas on Unit 1 Train B con ration CCW S stem R ressunzation Time 18.4 sec 18. sec calculat to be 18.224 Sc18.286 Imti CFC Inlet Tem or Vouhn Waterhammer 100'F as controlled b tern erature contro er SAR Imti CFC Inlet Tem or Two-Phase F ow NPSH 260 'F as bounded b Umt 1 LOCA FSAR lmti CFC Outlet Tem for Voi in Waterhammer 102'F er vendor manu s Norm Acci ent CFC ow Rates Per UFSA Vendor Manu Tran ort Pro cities ASMB Steam Tab es Materi erties Textboo Data Crane Tec ruc Pa er 410 V ve Worlan Pressures ANSI B16.34 Pi in Co ration Stress Anal sis o Reco Su ort Confi rations Detail Drawm s NRC Question 3:
Confirm that the water hammer and two-phase flow analyses included a complete failure modes and effects analysis (FMEA) for all components (including electrical and pneumatic failures) that could impact performance of the cooling water system and confirm that the FMEA is documented and available for review, or explain why a complete and fully documented FMEA was not performed. For example, failure of a CCW pump to start, or a delayed start of a CCW pump, could have a significant impact on the analyses that have been completed.
FPL Response 3:
Single failures were reviewed to determine the effect that potential component failures will have on waterhammer and two-phase flow analyses. This FMEA of single failures was developed to better or equal the documentation level of single failure reviews contained within the UFSARs.
The FMEA reviews all major components (including electrical and pneumatic failures) that could impact performance of the cooling water system. The FMEA is documented and available for review. The single failures described in the FMEA are generally in addition to the those single failures already imbedded in the post DBA temperature profiles (which are based on the UFSAR containment analyses).
L n
~A yA I'4'
St. Lucie Units 1 and 2
, Docket Nos. 50-335 and 50-389 L-98-276 Attachment 1 Page 10 NRC Question 4:
Deterlnine the uncertainty in the analyses that have been performed, explain how the uncertainty was determined, and how it was accounted for in the analyses to assure conservative results. Also describe the conservatians that have been reEaxed in the CFC model and exp/ain why conservative results are still assured.
FPL Response 4:
Analysis uncertainty is evaluated by considering the conservatism of the individual inputs, assumptions, and modeling methodology as well as the sensitivity of the analysis results to modification of these inputs. In this manner, it can be concluded that the outcome of the analysis is reasonable and conservative.
~ Individual inputs were selected to represent the plant accident analyses, initial conditions, plant geometry, and physics of the event. Inputs are based on design values and are generally conservative and bounding for each parameter. Individual inputs are discussed in the response to RAI Question 2.
~ Assumptions were made with regard to inputs and modeling to streamline the analysis.
Assumptions allowed ignoring small effects and using analysis of one cooler/train to represent other configurations, Assumptions are discussed in the response to RAI Question 7.
~ The modeling methodology was developed to represent the physics of the event. Modeling decisions reflect a need to accurately predict the event while employing simplifications and approximations to allow complex situations to be developed in a straightforward manner.
Iterative solutions are avoided where possible by bounding simplifications. Modeling is discussed in the response to RAI Question 1.
The following table presents a review of major input parameters, assumptions, and methodology, reviews the conservatism of individual parameters, and provides a weighting of its importance to the outcome of the analysis. Review of the matrix of parameters, their individual conservatism, and weighting of importance to the outcome of the analysis indicates the degree of conservatism in the analysis. This review supports a conclusion that the analysis provides conservative results.
Parameters Wt.:+ Discussion Contauunent DBA Curves H Conservative because curves deve oped to create maximum containment transient CCW System Repressurization Time based on H Conservative value o 18.4 seconds or Uiut 1 bounds a summation of time durations and tolerances -
decade of test values by 1.5 2.8 seconds. Value of 18.5 seconds for Unit 2 bounds a decade of test values by 1.0 to 4.2 seconds Containment atmosphere entermg CFC mo e ed to Conservative y maxmuzes extern eat trans er contain the same concentration of water vapor found in coefficient the containment ressurization anal sis.
Containment atmosphere maintained at esign ow rate Conservatively maxinuzes extern eat trans er even thou the fans would be coastin down coefficient Uiut 1 B tmn was mode ed and the coastdown used to Reasona e due to si arity etween piping runs resent the coastdown time for Unit 1 Train A.
CCWpumpcoast own owpro es oreac o e mt Reasona e as on system si arity: no credit en 1 CFCs are assumed to be applicable to the for expected CFC flow in lower range of acceptable co ondin Unit 2 CFCs. values rior to Fuel Pool HX valve closure CFC mo e assumes ou et temperature o ig est tu e H Conservative y overstates expec voi size with no credit for flushin b cold inlet flow Outside and inside CFC fouling resistances are assumed H Conservatively overstates expected heat trans er to be zero for heat transfer calculation Outsi e and inst e CFC ou resistances are assum Conservative un erstates ex ec recove bme m
St. Lucie Units 1 and 2
. Docket Nos. 50-335 and 50-389 L-98-276 Attachment 1 Page 11 Discussion'-:."::::::-::-:::;.,;:::::: i:::::::::;';:-;::,';!;,;:i',:::;::.:.'.:.:::.:::;::.::::;::.---::::::::
~ to be'desi n or recove determmation transient For MSLB, sm e-phase orced convection one Conservative y overstates expec eat trans er as it determines fin tern erature distribution and efficienc . maximizes fin efficienc Flow in the common return header is o y e resu t o Conservatively overstates steam generation and void size the CFC bein anal zed OCA Cases as it minimizes credit taken for bac ressure Flow in the common return header is not an yzed Conservatively overstates steam generation an voi size SLB Cases as no credit is taken for bac ressure All steam genera m the CFCs is assumed to exit e H Conservatively overstates steam generation and voi size CFCandformavoidinthereturn i in . as no credit is taken for steam blanketin of CFC tubes Top CFC tu e represents the entire CFC Conservatively overstates steam generation as no credit is taken for suppression of steam generation in lower tubes CCW Surge Ta Level assumed to be at ow eve Conservative y overstates expec steam generation as alarm elevation minus uncertain it minimizes sur e tank ove ressure Imtial CFC tube and return pipe w temperatures Conservative y overstates expected steam generation as al to the initial containment te erature of 120'F it does not resent a ilibrium condition CFCmo e const ers T~ o owmgsteampressurea r Conservatively overstates expec steam generation as steam generation it does not consider water sensible heat gain (small effect Voi con ensation withm Return Lmes inc u es 0 Conservative y overstates net voi size as it conservatism conservativel understates steam consum tion b 50%
Two-P ase ow ev uation uses m nun parameter H Consetvativelyun erstatestwo-p ase owmargmwi m values &om Unit 1 & 2 to devel boundin anal sis both lants Two-Phase ow ev uation assum CC pump curve Conservative y un erstates two-p ase ow margm or wasde dedb 10% onhead. the case of the hi est return line elevation PSH ev uation uses m nun parameter v ues m Conservative y understates NPSH margm wi o Unit 1 Sc 2 to develo boundin anal sis lants Hydrauhc Transient ev uation uses a 30 psia v ue or H Conservatively un erstates cus omng e ect o air cushionin . colla sin void Hydrauhc Transient ev uation uses a 1000 sec v ue H A uced soun sp an eng en c osure time is for the speed of sound in three segments to model two- used to reasonably model the effect of air that has come phase conditions and uses a 60 millisecond time period out of solution during the steam generation process.
for final corn ression of air bubbles Waterhammer stresses generated to not consi er .7 Conservative, overstates waterhammer pipe stress or reduction factor in 1977 version of ASMB BEcpv code cases where the stress intensification factor SI is used.
Pipe stresses are combined considers i est v ue H Conservatively overstates maximum pipe stress
&om each load case re ardless of location on model Su ort oadsarecombinedusm a soutev ues Conservativel overstates su ort esi n oads
- Wtre ersto ere ativeim acton ean sis. H-Hi M-Medium L-Low NRC Question 5:
Discuss speci/re system operating parameters and other operating restrictions that must be maintained to assure that the ivaterhammer and tlvo-phase floe analyses remain valid (e.g.,
surge tank level, pressure, temperature), and explain ivhy it vrould not be appropnate to establish Technical Speci+cation requirements to acknoivledge the importance of these parameters and operating restrictrons. Also describe andjustify use of any non-safety related instrumentation and controls for maintaining these parameters.
FPL Response 5:
The system operating parameters used both in the original response and in the current operability analysis were based on current design and operating requirements. No new or additional system restrictions have been identified or credited. Therefore, no new or additional restrictions on operating parameters are required to ensure that the input assumptions used in the submittal and current operability assessment) remain valid. analyses'original Although no new restrictions are required to support the analyses performed for the GL 96-06 response, concerns relative to the assumptions used in the original response have been identified.
Specifically, the analysis performed for the original response was intended to demonstrate that the
A 4 k,J'
St. Lucie Units 1 and 2
, Docket Nos. 50-335 and 50-389 I 98-276 Attachment 1 Page 12 existing plant design precluded waterhammer and two-phase flow concerns. As such, the system
'perating paiameter inputs were intended to represent bounding design basis values. However, as discussed in the introduction, it was identified that the system operating parameters used in the original response related to restart of a CCW pump after a postulated DBA were non-conservative.
Subsequent analysis of the potential for waterhammer using bounding values for CCW flow restart concluded that the potential for waterhammer was not precluded in all cases. Further evaluation demonstrated that although steam voiding could not be precluded in all cases, system functionality would not be compromised. The operational values used in the functionality evaluation were based on existing Technical Specification limits, or represented operation at a conservative value within normal control bands. Where applicable, tolerances and uncertainties were also applied to these values. Based on the methodology used, there are no additional controls required to ensure that that assumed operational parameters remain valid.
None of the operating parameters used in the analysis, which are not aheady contained in the plant Technical Specifications, have been identified which meet the criteria of 10 CFR 50.36 for addition to the plant Technical Specifications. The result of future analyses, as discussed in the response to RAI Question 9, will consider the implications of this question.
NRC Question 6:
Implementing measures to assure that waterhammer willnot occur, such as maintaining a minimum required surge tank level to prevent boiling, is an acceptable approach for addressing the waterhmnmer concento. However, all applicable scenarios must be considered to assure that the vulnerability to waterhammer and two-phase flow has been eliminated. Confir that the measures that exist or that have been established to prevent the occurrence of waterluanmer and two-phase flow are adequate for all applicable accident scenarios.
FPL Response 6:
The current operability analysis demonstrates that under the evaluated scenarios, waterhammer in the CCW system cannot be prevented. The current operability analysis concludes that the evaluated resulting stresses and loads will not result in loss of function, but may result in allowable limits. As such further actions will be required to resolve this issue exceeding design justify as outlined in response to RAI Question 9. These future actions will address any additional measures which may need to be established.
NRC Question 7:
Explain and all uses of "engineering judgement" that were credited in the analyses.
FPL Response 7:
Assumptions and uses of engineering judgement within the analyses are discussed in the following Table. The Table is orgaruzed to address the various assumptions and engineering judgements by analysis.
I
/
W ll I
~
g 1
J~
~ I 1
l,,r r
p'P
~r
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L98-276 Attachment 1 Page 13
'-.:,': .. ',.:tEn cent Jud emen Assnm tioas', ':.:;::;:: :-::::.-. .":::..:,::::,.:::.;,.'.:.:::;: ~::,::::: ';. '.Review/Rationale,
-:=:;.:---:=:;=:-:::-::;";-::.:.-:::,::,: .
.-::-::.=::-:::::-:-:::;::-.:::-::Time-To-Boil
...:.::::.:::-::::.:::::::::-..:.::::::::,=::::: and Voidin Anal sis::::::-:.:::::::-:::::::::::-::::-:::::::::..'::.::..'-:
Containment atmosphere catering a CFC contains the same The post-LOCA contaiament atmosphere is assumed to be homogeneous, with equal air/steam mass ratios at all concentratioa of water vapor fouad in the containment elevations. This is conservative since condensation on other heat sinks in coatainment near the CFC inlet locations, pressurization analyses in UFSAR Section 6.2. This and condensation on the inside of the CFC ducting, would tead to result in a lower steam/air mass ratio, aad therefore, assumption increases heat transfer to the CFCs and, lower coadensation heat traasfer coefficients at the CFC coils. In effect, the calculation assumes direct contact of the therefore is conservative. st-LOCA containment steam/air mixture onto the CFC coil surface.
Containment atmosphere is supplied at full rate of flow Coastdown time for the CFC fans is assumed to be much loager than the transients beiag evaluated. Assumption (design fiow rate) even though the fans would be coasting results in a steady supply of water vapor to the CFCs, increases heat transfer, and is, therefore, conservative.
down followiag a LOOP while the EDGs are beiag started/loaded.
Air-side velocity has no significant effect on outside heat Condeasation dominates the post-LOCA CFC heat removal performance as demonstrated by vendor performance data traasfer process after the LOCA DBA. This assumption is Condensation rate of water from a mixture of steam and nonwondensable gases is dominated by the concentration of reasonable for the high water vapor concentration in post- non~ndensables and does not have a strong depeadence on air side velocity (per Uchida).
LOCA contaiameat atmos here.
Single-phase forced convection heat transfer in the CFC Dittus-Boelter correlation was used for all portions of the coastdown when forced convection heat transfer is applicable.
Tubes is modeled using the Dittus-Boeltei correlation. Approach is conservative as it ignores the transition from turbulent to laminar fiow regime at reduced Reynolds numbers, aad therefore results in an over-prediction of the forced convective heat transfer coefficient during the latter portions of the CCW pump coastdowa. The duration of the CCW pump coastdown for which laminar conditions are applicable is relatively short, so this assumption does not add a great deal of coaservatism to the calculation. Although the Dittus-Boelter correlation is used throughout the industry to calculate the forced convection heat transfer coefficient for circular tubes, an additional margin was conservatively applied as a result of references which indicate that the Dittus-Boelter can under- redict heat transfer coefficients b 10 rcent.
Free convection heat transfer in CFC tubes is assumed for For uniformly heated CFC tubes with no forced flow, heat transfer inside the tubes is governed by conduction heat stagnant flow after CCW pump coastdown. transfer from the CFC tube wall to the stagnant fluid. The analysis models this portion of the transient with a free convection heat transfer coeKcient applicable to an enclosure heated on one side and cooled on the other. This modeling approach will over-predict the heat transfer coefficient, since this correlatioa is based on the presence of buoyancy-induced convection currents within the enclosure established as a result of the temperature difference betwem the plates. Since the pump coastdown time has been extended in the current analysis, there generally is no period of sta t, free convection in the anal sis and this model is not invoked.
The condensation heat transfer coeflicient on outer CFC Heat transfer coeflicieat associated with condensation on the outside of the CFC tubes is taken as four times the Uchida tube walls is assumed to be four times the Uchida data value. The cited reference is applicable to minimum containment pressure analyses for ECCS performance. In this presented in SRP Branch Technical Position CSB 6-1. case, it is conservative to maximize the heat removal from the containment atmosphere, and the factor of four is appropriate. For analysis pertaining to Generic Letter 9646, the application of the factor of four is judged to be reasoaable. A steady-state analysis of CFC performance was performed to compare the expected outside heat transfer coefficient to the four times Uchida model. Results showed the four times Uchida model to be very close to the vendor data, slightly over. predicting CFC performance for portions of the LOCA response profile. This analysis iadicates that the model for outside heat transfer is conservative while not severel nalizin results of the aaal sis.
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment 1 Page 14 En eerin Jud ement/Assum tzons Review/Rationale Radiative heat transfer is not modeled in the calculation. Radiative heat transfer from the containment atmosphere to the CFCs is not modeled in the calculation. because the peak temperatures in containment are relatively low ( 300 'F), this heat transfer mechanism will be small in comparison to the convection/condensation modes of heat transfer to the CFC coils. 'Ibis exclusion is bounded by the inherent conservatism associated with using four times the Uchida value for condensing heat transfer in the CFC coils and on the return i in inside containment.
CCW Pump Coastdown time was modeled for Unit 1 Train Pump coastdown times for HVS-1C and HVS-1D coolers (B Train) were obtained from a HYTRAN analysis.
B (vs. previous use of 2 second coastdown time.) HFGVN analysis based on system piping configuration, sizes, component hydraulic parameters, elevations, volumes, inertia etc. Model res nse reflected normal and accident stead state flows.
Unit 1 B train was modeled and the Pump coastdown used This appxoach is acceptable given the similarities in the pipe routing and sizes of the different trains. Minor elevation to represent the coastdown time for Unit 1 Train A (HVS- difference between the two loops is not important for pump coast down determination.
1A 8c HVS-1B .
CCW pump coastdown flow profiles for each of the Unit 1 CCW system piping and components are similar between units. Differences between units (Fuel Pool Hxs, CR HVAs, CFCs are assumed to be applicable to the corresponding etc.) would tend to reduce the coastdown profile. Therefore this assumption is conservative.
Unit 2 CFCs.
Outside and inside CFC fouling resistances are assumed to Increases heat transfer rate to the CFC water, reduces the time to onset of boiling, and increases void size be zero for heat transfer calculations. determination. Therefore, this assumption is conservative.
For MSLB cases, single-phase forced convection acts alone Lower air-side heat transfer coefficients result in higher fin efficiencies. Assumption niaximizes fin efficiency and to determine fin temperature distribution and fin efficiency. therefore increases calculated heat transfer and void size determination. Assumption is employed to xemove iterative a roachotherwise ired. CFCfinefficienc is over- xedicted. Therefore this assum ionis conservative.
Flow inthe common returnheaderis only the result of the Per a sensitivity analysis, this assumption conservatively max'nnizes the calculated void size as it minimizes CFC bein anal zed. bac ressuxe enerated within the return i in .
Allsteam generated in the CFCs is assumed to exit the CFC Maximizes the un-voided heat transfer area inside the CFCs, ignores potential for steam blanketing to reduce heat and form a void in the return i in . transfer, and thus maximizes the heat transfer rate to the CCW water.
Properties of air/steam mixture consists of the mass average Accepted method for determining mixture properties. Does not adversely impact the margin in the results of the values of air and steam separately at their respective partial calculation.
res sures.
CCW surge tank pressure is assumed to remain at Reduces force required by growing steam void to displace CCW water. Increases the calculated void size. Assumpfion atmos heric conditions throu out the transient. effect is sli t as two inch vent line willminimize ressurization effect.
CCW Surge Tank Level assumed to be at or above low level CCW surge tank is controlled within a control band well above the low level alarm elevation. It is assumed that alarm elevation includin uncertain Control Room willtake action to recover sur e tank level on xecei t of alarm.
Return line piping within both units is assumed to be Assumption maximizes heat transfer to return line piping and decreases heat capacity available to condense steam uninsulated. (Unit 1 return piping is uninsulated, Unit 2 generated &om CFC. For the timeframe of interest, assumption conservatively increases the calculated void size.
return i in is insulated.
Initial temperature of CFC tube/fin and return pipe wall is As 12'xceeds the equilibrium temperature, this assumption increases heat transfer rate to the CCW water during to the initial containment tern rature of 120K. earl sta es of transient. Therefore, assum ion is conservative.
Initial temperature of water in the CFC return piping is Assumption is appropriate given the system piping configuration, controlled inlet temperature of 100 F and normal assumed to be equal to the exiting temperature of the CFCs 100'F + 2 F Rise.
(
rise of 2 'F.
Recove -Evaluation:.(Two-'Phase Flow:.and 8)
The length of the recovery period for the fan cooler unit to The quasi-steady state transient refers to the normal containment/CCW temperature transient for non-LOOP DBAs.
reach the quasi-steady state condition is based the amount of The heat transferred with CCW at full flow is oxders of magnitude above the heat stored in the CFC metal at the heat stored in the CFC metal above normal quasi-steady elevated LOOP conditions. The recovery period before quasi-steady state conditions are re-established (metal state accident tern ratures without loss of flow. tern rature is redu is shown to be shozt 3 seconds .
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment1 Page15 En eerm Jud ement/Assum ttons ReviewlRation ale Outside and inside CFC fouling resistances are assumed to Decreases comparative heat flux and overstates time for transient recovery. 'Iherefore, this assumption is conservative.
be desi for recove time corn arison ses.
The review of the recovery phase for the worst case The worst case transient results in the highest CFC Cooler temperature and therefore the other cases are bounded.
transient bounds the recovery from the other transient cases The DBA case without LOOP is bounded as no loss of flow would occur.
as well as the i-stead state DBA case without LOOP.
Limiting inputs were used from both Units to develop a High amount of margin for two-phase flow and NPSH calculations allowed a bounding analysis to be performed.
sin leboundin anal sis. Conservative a roach for either case.
CCW pipe routing, post-SIAS flows and line sizes are Limiting routing is used in the analysis. Line sizes are similar and plant layout, routing similar.
nearl identical for the two units.
The potential for two-phase flow is evaluated at two critical Locations: 1) highest discharge piping location for HVS-1C cooler and 2) just downstream of the CFC return line locations in the discharge piping from the CFC. throttling valve and flow element orifice. Other postulated locations have increased pressures relative to these two locations; their lower elevation provides greater elevation head which exceeds the hydraulic pressure drop within the i in to reach the location. Conservative.
The analysis of the HVS-1C CFC bounds the other coolers The HVS-1C CFC has the highest elevation and therefore the lowest margin to saturation - the onset point of two-phase flow. This assum tion is conservative.
Calculated NPSH must exceed NPSH b 25% Conservative, as this results in 25% a lied mar in be ond vendor test data
.:-.::..--:-:::i::-. Waterhammer. Evaluation:(HYTRAN) -=..:::=.:::=-".--.-". ':--=----.'- --=:"--::-:
CCW Pump curves are essentially the same for both units Guarantee curves are identical, specific pump test curves essentially the same. Applied margin bounds minor pump differences.
CCW Pumps are assumed to re-start and reach full speed in Start time typical of large centrifugal pumps. Supported by FPL Eval. JPN-PSLSEEP-94-051,Rev 1 a relativel short time =2 sec.
Model was terminated at the tie-in to the 20" supply and 20" The pressure pulse in the 20" headers is expected to decrease due to the transition to much larger piping, return headers. the distance and friction from the point of void collapse. Although support loads may be slightly larger, substantially larger sizes and capacities of pipe supports on the 20" header will more than accomodate the increase in loads.
1B CCW train void size, being the largest, is considered to Generally, small size voids willgenerate smaller pipe stresses and support loads. As the 1A, 2A and 2B trains utilize bound the 1A, 2A and 2B train conditions similar system design, pipe size, materials, construction techniques, support methodology, degree of design conservatism etc. a boundin evaluation ma be used.
Air cushioning is assumed based on steam saturation Once boiling begins the pressure in the CFC discharge nozzle ramps upward to about 30 psia (saturation pressure).
pressure (30 psia) This pressure remains constant throughout the entire event until the steam void collapses. Pressure then rises to a value of 231 psia. While this air cushioning effect is known to occur in practice, the HYIRANanalysis only takes partial credit for the phenomena. The HYTRAN analysis makes a conservative assumption that the pressure remains at the saturation value of 30 psia and does not use the higher pressures that occur just prior to void closure that would otherwise further reduce waterhammer loads.
The theoretically instantaneous void closure that would be A reduced value of sound speed of 1000 fps is used in the pipe segments where boiling has occurred to addiess the used for cold water piping hammer is assumed to take 60 quantity of air within the system. Also, the theoretically instantaneous void closure that would be used for cold water milliseconds to complete and a reduced value of sound speed piping waterhammer is assumed to take 60 milliseconds to complete. The same assumptions are used in Sargent &
of 1000 fps is used in the pipe segments where boiling has Lundy's Design Basis Calculation for Zion 9646 analysis (Zion Calculation No. 22S-B-022M-631, Rev. 0 based on occurred. RELAP anal sis .
Evaluation:.::::::.:.:-.:: ...:::::::::-:::.:: ..-'-'=:.'-:.::':.:.-':.'-':-:::::-:.- ':::= '=.
Waterhammer stresses generated to not consider .75 Conservative, as no credit was taken for high stress intensification factors in the fmal stress tabulation.
reduction factor in 1977 version of ASME B&PV code.
t F
tl
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment 1 Page 16 Zn ee -"Jud ement/Assum ons:::.':.':::-:::: '-".: '"..:::.:::::::: ~::;.:-':..':::::: -::.:::.: ".-';::::::;:-::':--':::::::::::::-..'::Review/Rationale': '.'::::.'.:'.:::.'::.-:. =
Piping is evaluated for combined dead weight, pressure, and FSAR Sect. 3.1.2 states SSC vital for mitigation of accident conditions shall be designed to withstand effects of LOCA waterhammer stresses. Seismic stresses were not coincident with DBE. Hydrodynamic loading concurrent with seismic (DBE) is not considered credible since the considered concurrently. waterhammer event is a consequence of the LOCA, both waterhammer and DBE events are of relatively short duration and the evaluation is not in the desi basis realm functionali onl Combined dead weight, pressure and waterhammer stresses Conservative, results will clearly bound worst occurring stress in the affected piping.
are taken as maximum regardless of where each may occur ontheaffected i in .
u rt Evaluation Support, nozzle and penetration design loads include dead FSAR Sect. 3.1.2 states SSC vital for mitigation of accident conditions shall be designed to withstand effects of LOCA weight and waterhammer loads. Seismic loads were not coincident with DBE. Hydrodynamic loading concurrent with seismic (DBE) is not considered credible since the considered. waterhammer event is a consequence of the LOCA, both waterhammer and DBE events are of relatively short duration and the evaluation is not in the desi basis realm functionali onl Supports are evaluated to the reqmrements of Appendix F of Criteria is largely based on the Service Level D design criteria. This is consistent with the functionality evaluation the ASME McPV code criteria endorsed b the NRC in Generic Letter 91-18.
Su rt loads cases are added as absolute values. Conservative a roach ensures maximum desi load is utilized in all directions
4 A
M ft p, I4 g p!'Q k
II tg
. F1 I'iv l
I'P
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment 1 Page 17
'RC Question 8:
Provide a simplified diagram of the system, shoving major components, active components, relative elevatrons, lengths of piping runs, and the location of any orifices and flov restrictions.
FPL Response 8:
The figure (Attachment 3) depicts the layout of major components within the Unit 1 CCW system drawn to scale. The figure includes the containment fan coolers, SDC heat exchangers, CCW heat exchangers, CCW surge tank, CCW pumps, and associated piping. Relative elevations, lengths of piping runs and location of orifices and flow restrictions are noted. The Unit 2 configuration is similar in layout.
NRC Question 9:
Describe in detail any plant modifications or procedure changes that have been made or are planned to be made to resolve tire vaterharnmer and tvo-phase flov issues.
FPL Response 9:
No plant modifications have been made to date to address waterhammer and two-phase flow concerns expressed in Generic Letter 96-06.
As discussed by phone with the NRC Project Manager, W.C. Gleaves on 9/22/98 and in person with the NRC Site Resident, T. Ross, on 9/25/98, FPL identified that the previous response analysis failed to accurately assess the time for CCW pressurization. A subsequent evaluation that considers the additional timeframe for EDG restart have established that waterhammer is anticipated in the CCW lines serving the CFCs. The resulting system pressures, pipe stresses, and support loads can be accommodated under functionality rules established for the St. Lucie Plant.
FPL has decided to join the group of utilities working with NEI and EPRI to develop a design based version of NUREG/CR-5220, Diagnosis of Condensation-Induced Waterhammer, Case Studies. This approach, which was presented to the NRC in an industry workshop in Washington, DC in May 1998, is expected to assist the utility industry to develop waterhammer analysis techniques which are better suited for design purposes.
Accordingly, FPL will develop a plan for the resolution of this issue in the long-term based on the outcome and schedule developed for the work scope of this industry group.
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 I 98-276 Attachment 2 Page 1 Figure 1. Temperature vs. Time For Case U1LA (Unit 1 LOCA, CFC 1HVS-1A) 280 260 240 220 F- 200
~ 180 160 140 Cont. Drtr Bulb 120 100 --
Cont. Saturation
- CFCTube/Fin CFC Water 80 10 15 20 25 Time (seconds)
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment 2 Page 2 Figure 2. Temperature vs. Time For Case U1MA (Unit 1 MSLB, CFC 1HVS-1A) 300 280 260 240 220 r'ont.
200
- o. 180 E
I-160 ~o 140 Dry Bulb 120 Cont. Saturation
--- - CFCTube/Fin 100 CFC Water 80 0 10 15 20 25 Time (seconds)
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment2 Page3 Figure 3. Temperature vs. Time For Case U1LB (Unit1 LOCA, CFC1HVS-1B) 280 260 240 220 oU- 200 Cl e 180
~o 160 140 Cont. Dry Sutn 120 100 Cont. Saturation
- - CFC Tube/Fin CFC Water 80 0 10 15 20 25 Time (seconds)
I R
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment2 Page4 Figure 4. Temperature vs. Time For Case U1LC (Unit 1 LOCA, CFC 1HVS-1C) 280 260 240 220 ou- 200 m 180 160 140 Cont. Dry Bulb 120 100 --
Cont. Saturation CFC Tube/Fin CFC Water 80 10 15 20 25 Time (seconds)
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment2 Page5 Figure 5. Temperature vs. Time For Case U1MC (Unit 1 MSLB, CFC 1HVS-1C) 300 280 260 240 220 200 ~ I o 180 E
I- ~ I 160 140
~a Cont. Dry Bulb 120 Cont. Saturation
- - - CFCTube/Fin 100 CFC Water 80 0 10 15 20 25 Time (seconds)
S avgaane - Luaanarly' a
l I
l
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment 2 Page 6 Figure 6. Temperature vs. Time For Case U1LD (Unit 1 LOCA, CFC 1HVS4D) 280 260 ~ ~
240 220 Du- 200 O
m 180 CL E
160 140 120 Cont. Dp Buib 100 Cont. Saturation
- CFC Tube/Fin CFC Water 80 0 10 15 20 25 Time (seconds)
J I
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment2 Page7 Figure 7. Temperature vs. Time For Case U1MD (Unit 1 MSLB, CFC 1HVS4D) 300 280 260 240 220 200 C$
I 180 E ~ e O
I ~ o 160 140 ~0 Coot. Dry Bulb 120 Cont. Saturation
- - - - CFC Tube/Fin 100 CFC Water 80 0 10 15 20 25 Time (seconds)
0 I
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment 2 Page 8 Figure 8. Temperature vs. Time For Case U2LA (Unit 2 LOCA, CFC 2HVS4A) 280 260 240 220 ou- 200 6
180 CL E
160 140 cont. Dry Burn 120
~ ~
Cont. Saturation
- -"CFCTube/Fin 100
.CFCWater 80 0 10 12 14 16 18 20 Time (seconds)
Luna~I ~ ~
r
'I
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment 2 Page 9 Figure 9. Temperature vs. Time For Case U2MA (Unit 2 MSLB, CFC 2HVS4A) 400 360 320 280 0
m 240 200 160 Cont. Dry Bulb 120 Cont. Saturation
-CFC Tube/Fin CFC Water 80 0 10 12 14 16 18 20 Time (seconds)
. ~
Lasaaaubr<<
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment2 Page 10 Figure 10. Temperature vs. Time For Case U2LB (Unit 2 LOCA, CFC 2HVS-1 B) 280 260 240 220 ou- 200 O
m 180 160 140 Cont. Drtr Bulb 120 100 Cont. Saturation CFCTube/Fin CFC Water 80 0 10 15 20 25 Time (seconds)
~aata Luna aaartba a~ ~
P II P
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment2 Page 11 Figure 11. Temperature vs. Time For Case U2LC (Unit 2 LOCA, CFC 2HVS-1 C) 280 260 240 220 ou- 200 a
m 180 160 140 Cont. Drtr Butb 120 100 Cont. Saturation
- CFCTube/Fin CFC Water 80 0 10 15 20 25 Time (seconds)
Gargaang" Lraanoly' ~
II
>>I I'
l
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment 2 Page 12 Figure 12. Temperature vs. Time For Case U2MC (Unit 2 MSLB, CFC 2HVS4C) 400 360 320 280
~ 240 200 160 ~ r Cont. Dry Bulb Cont. Saturation
~ ~
120 - - - - CFC Tube/Fin CFC Water 80 10 12 14 16 18 20 Time (seconds)
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment 2 Page 13 Figure 13. Temperature vs. Time For Case U2LD (Unit 2 LOCA, CFC 2HVS4D) 280 260 240 220
~a 200 O
m 180 160 140 Cont. Dry Bulb 120 100 Cont. Saturation CFCTube/Fin CFC Water 80 0 10 15 20 25 Time (seconds)
}
N I
1 8
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment 2 Page 14 Figure 14. Temperature vs. Time For Case U2MD (Unit 2 MSLB, CFC 2HVS-1D) 400 360 320 280 O
m 240 200 160 Cont. Dry Bulb Cont. Saturation 120 - - -. -CFC Tube/Fin CFC Water 80 0 10 12 14 16 18 20 Time (seconds)
Luna anarata r r ~
I W
4
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment 2 Page 15 Figure 15. Return Pipe Wall Temperature vs. Time For Case U1LC (Unit1 LOCA, CFC 1HVS-1C) 280 260 240 220 o~ 200 m 180 160 140 Cont. Dry Bulb 120 100 Cont. Saturation
- - - - -Voided Section Water Solid Section 80 0 10 15 20 25 Time (seconds)
~aaaat Luna anartta a aa
k r
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment2 Page16 Figure 16. CFC and Return Pipe Internal Heat Transfer Coefficient vs. Time For Case U1LC (Unit 1 LOCA, CFC 1HVS-1C) 2250 2000 1750 Water Solid Return Pipe CFC Voided Return Pipe 0
ee 1500 I-IQ c 1250 0
oo 1000 e
m 750 500 250 10 15 20 25 Time (seconds)
IL St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment2 Page 17 Figure 17. Steam Volumes vs. Time For Case U1LA (Unit 1 LOCA, CFC 1HVS4A) 120 Condensed Generated 100 Net 80 ee~
0 E
o 60 E
C5 0
M 40 20 14 15 16 17 18 19 20 21 22 23 24 25 Time (seconds)
~eeesa 'ose~see
I 0
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment2 Page18 Figure 18. Steam Volumes vs. Time For Case U1LC (Unit 1 LOCA, CFC 1HVS-1C) 120 Condensed Generated 100 Net 80 6
E
)o E
60 tg V) 40 20 12 13 14 15 16 17 18 19 20 21 23 24 25 Time (seconds)
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment 2 Page 19 Figure 19. Steam Volumes vs. Time For Case U2LA (Unit 2 LOCA, CFC 2HVS-1A) 3.5 Condensed Generated 3.0 Net 2.5 e 20 E
0 m 1.5 V) 1.0 0.5 0.0 17 17.5 18 18.5 19 19.5 20 Time (seconds)
C3m~eaeee: . Lass~" e
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment 2 Page 20 Figure 20. Steam Volumes vs. Time For Case U2LC (Unit 2 LOCA, CFC 2HVS4C) 70 Condensed Generated 60 Net 50 ee o 40 E
0 m 30 V) 20 10 15 16 17 18 19 20 21 22 23 24 25 Time (seconds)
~ ~
I I II i
V I
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment 2 Page 21 Figure 21. Steam Volumes vs. Time For Case U2MC (Unit 2 MSLB, CFC 2HVS-1C) 3.0 Condensed Generated 2.5 Net 2.0 ee~
tD E
o 1.5 E
CJ 0
M 1.0 0.5 0.0 17 17.5 18 18.5 19 19.5 20 Time (seconds)
LA>eaeeae ee~
~ C
~ U
~
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment 2 Page 22 Figure 22. Steam Volumes vs. Time For Case U2MD (Unit 2 MSLBs CFC 2HVS 1D) 3.5 Condensed Generated 3.0 Net 2.5 e 2.0 E
)0 Po 1.5 tO 1.0 0.5 0.0 17 17.5 18 18.5 19 19.5 20 Time (seconds)
St. Lucie Units 1 and 2 Docket Nos. 50-335 and 50-389 L-98-276 Attachment 2 Page 23 (Final)
Figure 23. CFC Pressure vs. Time For Cases U1LA, U1LC, U2LA, U2LC, and U2MC 35
""""Case U2LA Case U2LC
- - - -Case U1LC 30 Case U1LA Case U2MC r
Pr es 25 rrrrrrrrnrrrr nrr rrnrr rn rrrrrrr r. r r. rrrrrrrrrrnrrnrrnrn.rrr..... ~.. nr rrr rrnn rn rrnrrns nrr rn rc rc .r r...; rr. rrr rrr rn Sll re (ps ia) 20 15 10 0 10 12 14 16 18 Time (seconds)
c
/
y" IC
PQ LES 0
EP2 DEIL Q E
1
~o CORPOROIT COOLNO WATER PUUPS IA EL. 25.75 TT.
g Lrt~
Qa I IC SI. LUCK Ural I IS OEPICItO TO SCILE.
PIPE ROV le fOR IPPT 2 IS SRLIAR, IS ELEVAIINIS fCR PRINCIPLE COUPOIKRIS ARE SINWR fOR OQIII UI4TS.
0 CTC WS-IA URR I EL 51.5 TT.
Q~ igg fed'QQLKBS URR 2 0 50'I fT Eir OO
~a PIPPIN WS-IB Igtf I EL $ 13 ff.
UMT 2 EL 50.4 ff.
I I@-IC CfC Ll'ill I EL 229 ff.
WS 1O CfC LTIIT 2 LL Tl 4 fl IRIII I Et. SI.S ff.
IRNT 2 EL. St.t fl,
~
0 hpP iy
\y gP'T.
P~gi 1St'LLVAROHS k SEIPOtkfS tttLI IIIL2 rr-r Iran HOI ttIIC JIAhl rr C INK cotrKL BiRr iran wW t '4 lr.r 7$ rs' rr-x .
LUCIE COMPONENT COOLING WATER Ir-$ @fr PIPING LAYOUT rr-lf rt-r I
QXQKP L RS ff,