ML20205K564
| ML20205K564 | |
| Person / Time | |
|---|---|
| Site: | Davis Besse  |
| Issue date: | 04/07/1999 |
| From: | Campbell G CENTERIOR ENERGY |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| Shared Package | |
| ML20205K567 | List: |
| References | |
| 2582, GL-96-06, GL-96-6, TAC-M96803, NUDOCS 9904130273 | |
| Download: ML20205K564 (38) | |
Text
_
EN QY Davis-Besse Nuctaar Power Sta00n
$501 North State Route 2 m Oak Harbor, ohio 43449-9760 Quy G. Campbell 419-3214S88 Vice Presktent- Nuclear Fax:419-3214337 Docket Number 50-346 License Number NPF-3 Serial Number 2582 April 7, 1999 United States Nuclear Regulatory Commission Document Control Desk Washington, D. C. 20555-0001
Subject:
Response to NRC Request for Additional Information pertaining to Generic Letter 96-06: Assurance of Equipment Operability and Containment Integrity During Design-Basis Accident Conditions (TAC No. M96803)
Ladies and Gentlemen:
On September 30,1996, the Nuclear Regulatory Commission (NRC) issued Generic Letter (GL) 96-06. That letter requested licensees, such as those for the Davis-Besse Nuclear Power Station (DBNPS) Unit Number 1, to address the following generic issues:
(1) Cooling water systems serving the containment air coolers (CACs) may be exposed to the hydrodynamic effects of water hammer during either a loss-of-coolant accident (LOCA) or a main steam line break (MSLB). These cooling water systems were not designed to withstand the hydrodynamic effects of water hammer and corrective actions may be neede i to satisfy system design and operability requirements.
Licensees are to determine if their plant's CACs cooling water systems arc susceptible to water hammer during postulated accident conditions.
(2) Cooling water systems serving the containment air coolers may experience two-phase flow conditions during postulated LOCA and MSLB scenarios. The heat removal l assumptions for design-basis accident scenarios were based on single-phase flow conditions. Corrective actions may be needed to satisfy system design and operability requirements. Licensees are to determine if their plant's CACs are susceptible to two-phase flow conditions during postulated accident conditions. I i
i}.' (
9904130273 990407
~
l yDR ADOCK 05000346 PDR
WY Docket Number 50-346 i License Number NPF-3 l Serial Number 2582 Page 2 I
(3) Thermally-induced overpressurization ofisolated water-filled piping sections ir.
containment could: 1) jeopardize the ability of accident-mitigating systems to perform their safety functions, and 2) could also lead to a breach of containment integrity via bypass leakage. Corrective actions may be needed to satisfy system operability requirements. Licensees are to determine if the piping systems that penetrate the containment of their plant are susceptible to thermal expansion of fluid so that overpressurization could occur.
On January 28,1997, the DBNPS provided by letter (Serial Number 2439) an interim response to GL 96-06. On February 28,1997, the DBNPS provided by letter (Serial Number 2442) a summary report describing the actions taken to that date and the results of those actions. On July 28,1997, the DBNPS provided by letter (Serial Number 2473) an update on the status ofits progress in resolving the issues of GL 96-06. On September 30,1997, the DBNPS provided by letter (Serial Number 2488) a report on the resolution of post-LOCA thermal overpressurization of containment penetrations, and the water I hammer and two phase flow effects on the CAC trains.
f l
1 On March 18,1998, the NRC issued a Request for Additional Information (RAI) I concerning questions posed by the NRC Office of Nuclear Reactor Regulation (NRR) staff regarding the 13 pipe lines penetrating the DBNPS containment that were identified by the DBNPS as possibly susceptible to thermally induced overpressurization. On August 28,1998, the DBNPS provided its response regarding the requested information (Serial Number 2554).
)
On April 15,1998, the NRC issued another RAI (DBNPS Log Number 5253) containing additional questions posed by the NRC NRR staff. These extensive questions relate to the Containment Air Coolers, possible water hammer, and two-phase flow in piping systems servicing the Containment Air Coolers. In its letter (Serial Number 2554) dated August 28,1998, the DBNPS informed the NRC of its plans and schedule to respond to these questions. During January through March 1999, the DBNPS staff periodically notified the NRC staff by telephone ofits progress in preparing the response.
The DBNPS originally enlisted the consulting services of Fauske and Associates, Inc. to determine potential water hammer loads on the containment air cooler system piping.
Since the initial issuance of GL 96-06, the NRC has refined the objectives of the effort and better defined the methods that are considered acceptable. The RAI dated April 15, 1998, reflected this more focused definition and required that the DBNPS obtain the services of the original consultant to describe how the previously completed work met the revised guidelines of the NRC. This work is completed and included in this letter as !
Attachment 1. It provides the response to the questions contained in the April 15,1998, NRC RAI. Attachment 2, Fauske & Associates document FAI/98-126, Waterhammer Phenomena in Containment Air Cooler Service Water Systems, is provided as a supporting document.
l
]
[;
Docket Number 50-346 License Number NPF-3 Serial Number 2582 Page 3 Should y ,u have any questions or require additional information, please contact i Mr. James L. Freels, Manager - Regulatory Affairs, at (419) 321-8466.
l \
l Very truly yours,
)
MW(
FWK/laj I 4 cc: S. J. Campbell, (Acting) NRC Region III, DB-1 Senior Resident Inspector l J. E. Dyer, Regional Administrator, NRC Region 111 W. O. Long, NRC/NRR Senior Project Manager Utility Radiological Safety Board l
1
i Docket Number 50-346 License Number NPF-3 Serial Number 2582 Page 4 l
RESPONSE
TO REQUEST FOR ADDITIONAL INFORMATION TO NRC GENERIC LETTER 96-06 FOR THE DAVIS-BESSE NUCLEAR POWER STATION UNIT 1 1, Guy G. Campbell, state that (1) I am Vice President - Nuclear of the FirstEnergy Nuclear Operating Company, (2) I am duly authorized to execute and file this certification on behalf of the Toledo Edison Company and The Cleveland Electric Illuminating Company, and (3) the statements set forth herein are true and correct to the best of my knowledge, information and belief.
B e; GuyD. Campbell, Vice President - Nuclear l
Affirmed and subscribed before me this 7thday of April,1999.
/Wei sw H Notary Public,$Iate of Ohio - Nora Lynn Flood j My commission expires September 4, 2002.
i
Docket Number 50-346 License Numbcr NPF-3 Serial Number 2582 Page1 DAVIS BESSE NUCLEAR POWER STATION RESPONSE TO NRC'S GENERIC LETTER 96-66 REQUEST FOR ADDITIONAL INFORMATION Introduction This attachment provides the Davis-Besse Nuclear Power Station's (DBNPS) responses to the NRC's request for additional information dated April 15,1998, pertaining to NRC Generic Letter 96-06. In order to provide a detailed response to each NRC question, the NRC's Questions 1 and 2 were subdivided into several parts and each part was responded to separately.
Question 1 was subdivided into two parts and Question 2 was subdivided into 5 parts. In order to facilitate the review, the corresponding NRC question was also indicated along with each DBNPS response.
Response 1 (NRC Question 1, Part 1):
If a methodology other than that discussed in NUREG/CR-5220, " Diagnosis of Condensation-Induced Waterhammer," was used in evaluating the effects of waterhammer, describe this alternate methodology in detail.
METHODOLOGY OVERVIEW The methodology used to analyze the Davis-Besse plant in response to Generic Letter 96-06 consists of two parts. First, scaled experiments and independent theoretical investigations were performed to investigate the dominant containment fan cooler two-phase flow and waterhammer phenomenon [ reference RAI 1.1]. Second, detailed computer modeling of the affected portions of the service water system was performed to determine the transient thermal hydraulic response of the service water system, the response and long term performance of the containment air coolers, and the forcing functions resulting from the condensation-induced water hammer events.
Additional details of the experimental and theoretical investigations are provided in FAI/98-126 in response to RAI Item 2, while the detailed computer modeling approach is discussed below.
COMPUTER CODE MODEL DESCRIPTION Detailed computer modeling in the form of two-phase flow transient analysis was performed with the TREMOLO Revision 1.0 computer code developed by Fauske & Associates, Inc (FAI).
TREMOLO has been used previously to perform transient thermal hydraulic analyses for several nuclear plants in the U.S. and Canada, including analyses in support of Generic Letter 89-10 and 96-06 issue resolution. TREMOLO is a transient thermal hydraulic code developed to analyze single- and two-phase flow conditions in plant piping systems. TREMOLO - Thermal hydraulic Response of a Motor-Operated valve Line - was so named since it was originally developed to evaluate the pressure oscillations associated with valve closures and openings in piping segments that could be exposed to two-phase Dow conditions. For the current effort, the TREMOLO code was run with Davis-Besse specific plant models for the three separate containment air cooler (CAC) piping circuits.
Docket Number 50-346 License Number NPF-3 Serial Number 2582 Attachment 1 Page 2 TREMOLO is a node and junction code that has a well-defined code architecture that clearly separates the " instantaneous picture", the rate dependent processes, the integration function, and the updating of the variables as the integration progresses. Such architecture is fundamental to the success of a transient thermal hydraulic code. Additionally, the TREMOLO computer code uses a classic 4th order Runge-Kutta integration scheme coupled with a variable time step.
The TREMOLO Revision I model uses a "one and a half" fluid model which implies separate mass and energy conservation equations for each of the two fluid phases and a single momentum equation *.o describe the fluid mixture. TREMOLO considers two fluid phases (liquid and vapor) which may exist in a non-equilibrium state. Thus, for each fluid node, TREMOLO evaluates five transport " rate" equations (liquid phase mass, liquid phase energy, gas phase mass, gas phase energy, and mixture momentum). To provide closure to this system of equations, Huid transport between the phases is defined and an equation of state is used. The equation of state calculates the sonic velocity in the single and two-phase fluids, then combines that information with the nodal mass and energies to estimate node pressures. An essential part of the equation of state is the retention of a residual void following the larger scale void collapse. This assumption is discussed in additional detail in Response 4.
pAVIS-BESSE PLANT MODEL The physical plant model used in TREMOLO is contained in a TREMOLO parameter file. The parameter file defines the pipe system geometry, flow element locations and type, heat exchanger characteristics, empirical model parameters, boundary and initial conditions, and code control parameters. The CAC model descriptions and parameter files are documented in verified design calculations.
The plant model also incorporates sequence-specific assumptions, initial conditions, and boundary conditions. For the current effort a conservative set of sequence-specific parameter values was selected. For additional information regarding the sequence-specific assumptions and input, please refer to Responses 4 and 5.
The TREMOLO model of a pipe system consists of a series of pipe sections assembled in series between the upstream and downstream boundaries. For instance, Figure RAI l-1 gives a pictorial l representation of the TREMOLO nodalization scheme of the first pipe section in the CAC-1 I model which interfaces with the upstream boundary. As shown, a pipe section is characterized by a total length, constant diameter, constant wall thickness, linear elevation change from inlet to outlet, flow element locations relative to the pipe section inlet, and a fixed number of equally spaced axial fluid nodes and radial pipe wall nodes. The remaining pipe sections have identical numerical treatments, only with varying inner diameters, lengths, flow elements locations, and l pipe wall thicknesses. For additional information on the plant physical model, please refer to the I
simplified diagram of the CAC circuit provided in Response 10.
In addition to the fluid transport equations, discussed previously, convective heat transfer between the fluid and the metal piping is considered. To account for the conduction of heat through the piping, multiple radial pipe wall nodes were incorporated into the model.
Docket Number 50-346 License Number NPF-3 Serial Number 2582 Attachment 1 Page 3 REACTION FORCES A key part of the Davis-Besse analysis is the calculation of reaction forces on the pipe walls that result from the dynamic fluid behavior over the course of the design-basis accident. TREMOLO Revision 1 provides a calculation of the reaction force at the inlet of each defined flow element as well as the reaction force resulting from fluid flow across the upstream and downstream boundaries of the modeled pipe circuit.
The reaction force for each of the internal flow elements is calculated assuming the geometry of a 90-degree bend. Considering the various flow element types (90-degree elbow,45-degree bend, pipe tee branch, pipe tee mn, reducer), the largest force occurs at the 90-degree bend since the fluid velocity vector undergoes the maximum change with this geometry.
Reaction forces calculated by TREMOLO were then used as input to the piping stress analysis code. The stresses resulting from the water hammer event were combined with pressure and weight stresses and compared to ASME Section III allowable stresses for a Faulted Condition.
The restraint and anchor loads consisted of a combination of weight, thermal and water hammer loads.
References RAI 1.1 FAI/97-2, "FAI Experience on Waterhammer Phenomena in Containment Air Cooler Service Water Systems," (January,1997).
g XZP = 64 ' a r i 7Fi ; i 3 XIDP : :* : :
a = 0.833 8 v .
jg# .
- jgg*
j n
XWALLP T
" = 0.0304 8 I
t =0 R 2 = 18 a" Approximate flow elementlocations FIGURE RAI 1-1 TREMOLO Model of CAC-1 Pipe Section 1
~ Docket Number 50-346 License Number NPF-3
,' ' Serial Number 2582 Attachment 1 Page 4 Response 2 (NRC Question 1, part 2):
Also, explain why this methodology is applicable and gives conservative results for the Davis-Besse plant (typically accomplished through rigorous plant-specific modeling,
. testing and analysis).
I
! As described in Response 1, the methodology for evaluating the effects of waterhammer and two-phase flow in the cooling water serving the Davis-Besse containment air coolers consists of two parts. First, experimental investigations were performed to obtain a better understanding of the controlling physical processes related to two-phase flow in service water systems. Second, detailed thermal hydraulic transient analyses of the service water piping were performed, utilizing a conservative set of assumptions, to quantify the effects of waterhammer and two-phase flow on
- the affected piping and components.
. The bases for the methodology, including assurance that conservative results are obtained for the Davis-Besse plant, are provided in the attached report, FAU98-126. The report stresses:
- 1. The importance and impacts of the thermal boundary layer (see FAU98-126 Section 2.2);
- 2. The local entrainment issue which, per FAI experiments, demonstrates that a one-dimensional (not stratified) propagation occurs (see FAU98-126 Sections 2.2 and 2.3); and
- 3. The presence and influence of residual gas bubbles in the fluid following void collapse (see FAU98-126 Section 2.1).
Additionally, FAU98-126 discusses the types of waterhammer events expected to occur under the postulated accident conditions for the Davis-Besse-specific service water piping geometry (see FAU98-126 Section 2.2). Finally, FAU98-126 includes benchmarks of TREMOLO code predictions against experimental data for relevant waterhammer events to demonstrate that the TREMOLO methodology is applicable and to quantify the conservatism in the TREMOLO results (see FAU98-126 Section 3.0).
Another aspect of the analysis that ensures a conservative result is the conservative definition of the accident scenarios selected for the analyses. The basis for selection of the accident scenarios is addressed in Response 5, below.
-In summary, the methodology used to anelyr the Davis-Besse plant in response to Generic letter 96-06 is shown to be applicable and conservative through detailed plant specific modeling with a conservative set of input parameters, computer code benchmarking, scaled laboratory testing, and independent analyses of the dominant physical processes.
t
Docket Number 50-346 License Number NPF-3 Serial Number 2582 l
Page 5 Response 3 (NRC Question 2a):
Identify any computer codes that were used in the waterhammer and two-phase flow analyses and describe the methods used to benchmark the codesfor the specific loading conditions involved (see Standard Review Plan Section 3.9.1).
Detailed two-phase flow and waterhammer calculations were performed with TREMOLO 1.0, a proprietary code developed by Fauske & Associates, Inc. (FAI). TREMOLO is designed to perform transient analyses of single- and two-phase flow conditions in nuclear power plant piping systems.
Please refer to Response 1 for a discussion of the application of the TREMOLO code to the Davis-Besse plant configuration and a discussion of additional code modeling details.
Benchmarking of the TREMOLO code has been performed by comparing the TREMOLO calculated results against data obtained from scaled experiments d,.siped to investigate column separation, two-phase flow, and condensation-induced waterhammer events in power plant piping systems. Benchmarking consists of separate effects tests to validate individual phenomena models as well as integrated code tests to validate the overall code applicability to applied two-phase flow problems. Please refer to Section 3 of FAI/98-126 for a more thorough discussion of the TREMOLO benchmarking efforts, i
l I
)
Docket Number 50-346 License Number NPF-3 Serial Number 2582 Attachment 1 Page 6 Response 4 (NRC Question 2b):
Describe andjustify all assumptions and input parameters (including those used in any computer codes) such as amplifications due tofluid structure interaction, cushioning, speed of sound, force reductions, and mesh sizes, and up:cin why the values selected give conservative results. Also, provide justification for omitting any effects that may be relevant to the analysis (for example, fluid structure interaction, flow induced vibration, erosion).
OVERVIEW The Davis-Besse two-phase flow and waterhammer analysis is dependent on detailed computer modeling using the TREMOLO Revision 1.0 computer code, as described in Response 1. There are two general categories of assumptions and input parameters used in the Davis-Besse TREMOLO calculations. First, there are assumptions regarding which phenomena are to be included in the computer code phenomena models and the valid range of any empirical modeling parameters. Second there are inputs related to the plant-specific piping systems and accident sequences. Each of these two categories ofinputs and assumptions will be addressed separately, beginning with the phenomena models.
MODELING OF THE DOMINANT PHYSICAL PROCESSES The TREMOLO computer code contains models of the phenomena relevant to the study of transient two-phase flow and condensation-induced waterhammer events that could occur in process piping systems. These models were selected and developed based on an understanding of the dominant physical processes expected during the accident conditions postulated to occur in service water cooling systems of nuclear power plants. Namely, based on the scaled experimental investigations performed by Fauske & Associates, Inc, and a review of the available literature, the dominant phenomena include one-dimensional, non-equilibrium, two-phase fluid flow; presence of residual gas bubbles in the fluid following large scale void collapse due to non-condensible gas exiting from solution at low pressures; and the thermal gradient in the fluid between the cold refill water and the hot steam void region. Experimental investigations of condensation-induced waterhammer events in condenser cooling water systems using prototypic test apparatus have not indicated that fluid structure interactions, flow induced vibrations, or erosion play a significant role in the accident analysis relevant to Generic Letter 96-06. Thus, models for these phenomena are not included in the TREMOLO code. Please refer to Section 2 of FAI/98-126 for additional information on the relevant physical processes.
Integral code benchmarks against relevant experimental data suggest that the TREMOLO code is applicable to the types of accident scenarios and waterhammer events addressed under Generic Letter 96-06 given the current set of phenomena models utilized by TREMOLO. Please refer to Section 3.3 of FAl/98-126 for additional information on the TREMOLO integral code benchmark 3.
TREMOLO MODELING ASSUMPTIONS A description of the specific modeling assumptions used by TREMOLO for the Davis-Besse analyses are as follows.
r Docket Number 50-346 License Number NPF-3 l Serial Number 2582
! Attachment 1 Page 7 l
Friction Losses, Pipe Roughness, and Nominal Flow through the CAC Circuit Friction pressure drop is accounted for by coalbining pipe wall friction for straight pipe sections and friction losses through miscellaneous flow elements. Pipe wall friction is characterized by l
the term fUD where f is the Moody friction factor, L is the actual pipe length, and D is the pipe hydraulic diameter. Friction losses through flow elements are characterized by loss coefficients, l
K. The pipe diameters (nominal) in the Davis-Besse CAC-1 circuit range from 3 inches up to 10 inches (not including the CAC tubes, themselves). Assuming fully developed turbulent flow in clean commercial steel, this results in Moody friction factors of 0.017 to 0.013 [ reference RAI
- 4. I].
For the accident scenarios considered here, flow rates might deviate substantially from those that are characteristic of fully developed turbulent flow. Thus, friction factors are evaluated within TREMOLO for every fluid node at every time step based on smooth- pipe correlations [ reference RAI 4.1] as a function of the current flow rate in the fluid node. Furthermore, pipe roughness is accounted for in the TREMOLO model by applying a linear multiplier to the calculated smooth pipe friction factor.
A steady state, cold water flow rate of approximately 1600 gpm is used in the citrrent analysis.
This is achieved by modeling pipe wall friction, losses through flow elements and a friction factor multiplier of 1.6 for CAC-1 and CAC-3, and 1.175 for CAC-2 to account for pipe roughness. Reference RAI 4.2 reports a minimum measured flow rate through the CAC circuit of about 1350 gpm, a design flow rate of 1600 gpm, and a clean pipe, maximum flow rate of about 1850 gpm. The current analysis approximates the design flow rate.
Heat Transfer in the CAC TREMOLO Revision 1 provides a mechanistic heat exchanger calculation which is used to model the CAC heat transfer rate under low flow and two-phase flow conditions The calculated heat transfer rate is a function of heat exchanger geometry, cooling water flow rate, gas temperature, and tube fouling factor. To maximize heat transfer, a lower bound on CAC fouling factor of .00045 is used in the TREMOLO model [ reference RAI 4.2]. This fouling factor combined with the other CAC heat exchanger design parameters input to TREMOLO, yields heat removal rates as a function of containment gas temperature, as shown in Figure RAI 4-1.
The TREMOLO results are compared to the Davis-Besse data provided in reference RAI 4.2. As shown, TREMOLO calculates more heat transfer for the same conditions than the reference data, which is conservative for the current analysis.
r Docket Number 50-346 1
License Number NPF-3 Serial Number 2582 l l
Attachment 1 l Page 8 FIGURE RAI 41 Davis-Besse CAC-1 Calculated Heat Transfer Rates !
CAC-1: 85 F,1600 gpm, .00045 Fouling 275 i , , , , ,
,e O
@. 250 -
/ -
,s
/
t s 8- 225 -
/ -
Au s'
/
5 s' 3 MO -
f# -
.h #
5 /
s' O 175 -
j# -
/
4 '
130 20 30 40 SG 60 70 80 90 Heat Transfer to Cooling Water (MBtu/hr)
- TREMOLO Air Cooler Model
+ Davis Bess- Pulldown Data i
Containment Gas Temperature The containment gas temperature is assumed to vary with respect to time according to the design basis containment response documented in Toledo Edison calculation C-NSA-60.05-002
[ reference RAI 4.2]. Ten data points are used to model the containment gas temperature profile for the first 60 seconds of the design basis analysis. These 10 data points are plotted in Figure ,
RAI 4-2.
Supply and Return Header Pressures TREMOLO uses pressure boundary conditions at the upstream and downstream boundaries of
' the pipe circuit. The current analysis accounts for service water pump trip and restart by varying the upstream and downstream boundary pressures as a function of time. Service water pump test data was collected to determine the duration and pressure history characteristic of the pump 4 coastdown following trip and ramp up following restart [ reference RAI 4.2]. Data was collected for both the supply and return headers. TREMOLO boundary pressures were then varied according to these test data. However, linear pressure changes with respect to time were used to simulate the more complex header pressure profiles. The duration of both the coastdown and ramp up, as well as the maximum and minimum pressures used for the upstream boundary in TREMOLO, are based on the CAC header test data.
L
l Docket Number 50-346 License Number NPF-3 )
Serial Number 2582 Attachment 1 Page 9 l
Parallel Flow WCu and Pipe Circuit Boundaries The CAC pipe circu.ts are actually complex pipe networks consisting of numerous parallel flow paths and several thousand feet of pipe. The portion of the network selected for the current l model simplifies the complex pipe network into a single pipe circuit consisting of individual pipe {
sections of varying diameter, length, and elevations connected in series. Furthermore, the CAC pipe circuit models start at upstream and downstream " header" points rather than the actual service water pump discharge and return cenal discharge points (see Figure RAI 10-1). Parallel ,
flow paths are accounted for by selecting an equivalent pipe diameter that conserves the total mass flow rate and actual fluid velocity through each branch. Equivalent pipe diameters are used !
to model parallel flow paths in the 6-inch,3-inch, and CAC cooling tube regions. !
I Fluid Compressibility Due to Presence of Steam and Non Condensible Gas Experimental evidence presented in Section 2 of FAl/98-126 indicates that non-condensible gas ;
and steam voids appear rather rapidly following sudden depressurization transients in water- I filled fluid systems, and that these gas bubbles go back into solution over several tens of seconds.
The lingering presence of the gas bubbles accounts for observed waterhammer pressure loads that are significantly less than the calculated loads based on solid liquid conditions. In essence, the gas bubbles provide compressibility to the otherwise incompressible liquid system. This is accounted for in TREMOLO Revision I by maintaining a minimum void fraction once steam and non-condensible gases are releaed into the pipe system.
t
f l Docket Number 50-346 License Number NPF-3 Serial Number 2582 Attachment 1 Page 10 FIGURE RAI 4-2 Davis-Hesse DBA Containment Temperature used in TREMOLO LOOP +LOCA Analysis Davis-Besse DBA Containment Profile 300 i i , , , ,
- =
250 -
5 j 200 -
I
$ 150 -
I e
3
.l100 -
3 O
50 -
1 0
-10 0 10 20 30 40 50 60 Time (seconds)
There is some uncertainty in the degree to which retained gas increases the compressibility of the fluid. Therefore, benchmarks of TREMOLO with experimental data [FAI/98-126, Section 3.3]
and independent hand calculations [FAI/98-126, Section 2.1], are used to select the appropriate value for the controlling TREMOLO model parameter, VFMIN (minimum void fraction following gas evolution). Please refer to Section 3.3.2 and 3.3.3 of FAI/98-126 for a discussion of the integral code tests, and Section 2.1 for an independent analysis of the non-condensible gas content, used to establish the bounding values of VFMIN.
Model Parameters The TREMOLO code requires several user-specilled, empirically derived model parameters.
Default values for these parameters are specified in the *MODEL section of the parameter file, while sequence specific model parameter settings are specified in the sequence input decks.
Table RAI 4-1 summarizes all parameters from the *MODEL section of the Davis-Besse CAC parameter files, including Se specific values used in the current analysis.
l l
l i
l 1
1 Docket Number 50-346 i i License Number NPF-3
! Serial Number 2582 i Attachment i Page11 l TABLE RAI 4-1 Model Parameters Used l'or Davis Besse Analysis
! TREMOLO Parameter Davis Besse Name Value Parameter Description Technical Basis NWALL 2. NUMBER OF NODES FCR PIPE WALL Maximize wall heat transfer to i l IIEAT TRANSFER CALCULATIONS reduce void in0uence i NSECT 3. NUMBER OF NODES PER PIPE Modeler discretion to balance SECTION FOR FLUID FLOW precision and computa' ion I CALCULATIONS requirements. See Ncte 1 i FFRIC 0.013 NOMINAL FRICTION FACTOR FOR 10" clean commercial steel STRAIGHT PIPE SECTIONS pipe value [ reference RAI 4.1]
HTCO 0. HEAT TRANSFER COEFFICIENT AT Outer surface assumed PIPE WALL OUTSIDE SURFACE adiabatic FHTCI 1.0 WALL INSIDE SURFACE HEAT Allow fluid-pipe wall heat TRANSFER COEFFICIENT transfer MULTIPLIER; = 1.0 FOR NORMAL CALCULATION; = 0.0 FOR ADIAB ATIC BC FCDUP 1.0 DISCHARGE COEFFICIENT FOR Maximize inlet flow rate FLOW AT BOUNDARY OF i UPSTREAM PIPE SECTION l FHXFC 0 FAN COOLER HEAT EXCHANGER Mechanistic model must be MODFLSELECTION used because oflow flow and two-phase Gow conditions.
=0; MECHANISTIC HEAT
)
EXCHANGER MODEL
=l: LOOK-UP TABLE FQMULT 1.0 FAN COOLER HEAT EXCHANGER Only 1 active CAC modeled HEAT TRANSFER MULTIPLIER.
CALCULATED HEAT TRANSFER WILL BE MULTIPLIED BY FQMULT ITDBUB 0.001 CHARACTERISTIC TIME FOR Based on photographic studies BUBBLE GROWTH reported in [ reference RAI 4.3]
NBM3 1.E9 INITIAL BUBBLE DENSITY Based on best estimate (BUBBLES /M^3) [ references RAI 4.4-7]
l RBO 1.E-6 INITIAL BUBBLE RADIUS (M) Based on best estimate i
[ references RAI 4.4-7] l PPN2MN 3.3E4 INITIAL MINIMUM NITROGEN Conservatively selected less PARTIAL PRESSURE (PA) than 1 atmosphere. See Note i
Docket Number 50-346 License Number NPF-3 Serial Number 2582 Attachment 1 Page 12 TABLE RAI 4-1 Model Parameters Used for Davis-Besse Analysis (continued)
TREMOLO Parameter Davis hs=e j Name Value Parameter Description Technica! Basis CCOND 0.6 CONDUCTION COEFFICIENT Based on comparison to fan cooler test data;[see also, Section 3 2.2 of FAl/98-126]
VFMIN 1.E-3 MINIMUM VOID FRACrlON Based on two-phase FOLLOWING GAS RELEASE waterhammer benchmark [see Section 3.2 of FAl/98-1261 FROUGli 1.6 (CAC-1,3) FRICrlON FACTOR MULTIPLIER TO Selected to match design flow 1.175 (CAC-2) ACCOUNT FOR PIPE WALL rate of 1600 gpm ROUGHNESS Note 1: Justification for use of " engineering judgment" or "modeler discretion" is provided in Rcsponse 7.
DAVIS-BESSE PLANT MODEL The physical plant model used by TREMOLO is contained in a TREMOLO parameter file. The parameter file defines the pipe system geometry, flow element locations and type, heat exchanger characteristics, empirical model parameters, boundary and initial conditions, and code control parameters. Separate Davis-Besse plant models were developed for each of the three CAC circuits based on plant drawings and additional plant design inputs [ reference RAI 4.2].
The plant models are stored electronically in parameter filer for use by TREMOLO and the parameter file development is documented in verified design calculations. The development and I verification of the design calculations ensures that valid, plant-specific models are used for the current analysis.
References RAI 4.1 CRANE Technical Paper No. 410, " Flow of Fluids Through Valves, Fittings, and )
Pipe," 1988 Editien.
l l
l RAI 4.2 G. Thomas Elicson and Jim P. Burelbach, memo to file, " Summary of Design Inputs i
Received from Toledo Edison to Suppott Generic Letter 96-06 CAC Analysis, dated February 20,1997.
RAI 4.3 Bennett and Myers, Momentum. Heat. and Mass Transfer. Third Ed. p. 411, McGraw-Hill Book Company, New York (1982).
RAI 4.4 Abdollahian, D., Healzer, J., Janssen, E., and Amos, C.,1982, " Critical Flow Data Review and Analysis," EPRI NP-2192.
I RAI 4.5 Ardon, K. H.,1978, "A Two-Fluid Model for Critical Vapor-Liquid Flow," Int. J.
Multiphase Flow 4, p. 323
Docket Number 50-346 License Number NPF-3 Serial Number 2582 Attachment 1 Page 13 l-RAI 4.6 Richter, II. J.,1981, " Separated .Two-Phase Flow Model: Application to Critical Two-Phase Flow," EPRI NP-1800 (April).
RAI 4.7 Rivard, W. C. and Travis, J. R.,1980, "A Nonequilibrium Vapor Production Model for Critical Flow," Nuclear Science and Engineering B, pp. 40-48.
l l
l l l I l I l I
l I
i
{
Docket Number 50-346 License Number NPF-3 Serial Number 2582 Attachment 1 Page 14 i
Response 5 (NRC Question 2c):
Provide a detailed description of the " worst case" scenarios for waterhammer and two-phase flow, taking into consideration the complete range of event possibilities, system configur.itions, and parameters. For example, all waterhammer types and water slug scenarios should be considered, as well as temperatures, pressures, flow l rates, load combinations, and potential component failures. Additional examples include:
e The effects of voidfraction m: flow balance and heat transfer; \
e The consequences ofsteamformation, transport, and accumulation; e Cavitation, resonance, andfatigue effects; and e Erosion considerations NUEEG/CR-6031, " Cavitation Guide for Control Valves," may be helpful in addressing some aspects of the two-phaseflow analyses.
OVERVIEW The worst case scenarios for waterhammer and two-phase flow for Davis-Besse are based on independent TREMOLO code calculations for of each of the three CAC loops and the conservative definition of accidem scenarios. The waterhammer loads calculated by TREMOLO are consistent with those observed in the scaled experiments performed by Fauske & Associates, Inc. [ reference RAI 5.1]. In addition to the condensation-induced waterhammer events considered by TREMOLO, a detailed review of the service water piping systems supplying the three Davis-Besse air coolers was performed to see whether or not additional types of waterhammer events, such as those discussed in NUREG/CR-5220, could occur. A summary of the TREMOLO calculations for the wor
- case scenario and a detailed discussion of the NUREG/CR-5220 review are presented below.
ACCIDENT SCENARIO DEFINITION FOR TREMOLO ANALYSES For significant voiding to occur in the cooling water system servicing the containrr.ent air coolers, the service water pump would have to trip and coast down to zero flow and then remain in that state on the order of 10 seconds or more. For the current analysis, a pump trip of an extended duration is postulated to occur as the result of an unexpected loss of offsite power (LOOP). If a loss of offsite power is anticipated, such as during inclem " weather, then the i emergency diesel generators will be running prior to the loss of power anu the duration of the pump trip will not be sufficient to cause voiding in the system. Such was the case on June 24, 1998 when tornado strikes near the Davis Besse plant led to a loss of offsite power. Because the diesel generators were running, the duration of the service water pump trip was only a few seconds.
I For the current analysis, the possibility of heat transfer from the containment to the air cooler cooling tubes augmenting the voiding caused by the extended duration of the pump trip is also considered. The maximum voiding would occur either during a design basis loss of coolant i l
accident (LOCA) or a main steamline break (MSLB). The main steamline break scenario is
Docket Number 50-346 I
License Number NPF-3 Serial Number 2582 Attachment i Page 15 short lived relative to the design basis LOCA event, and the heatup of the air cooler cooling water during a main steamline break event is bounded on the low side by the LOOP and on the high side by the LOCA+ LOOP scenarios. Nonetheless, a sensitivity calculation assuming the i
MSLB containment conditions is presented in Response 8. The original TREMOLO analysis considered the LOOP and LOCA+ LOOP scenarios, only.
l Each of the three air cooler piping circuits was analyzed separately for their response to postulated conditions of a LOOP and a LOCA+ LOOP. The specific set of initial and boundary conditions used for each of the accident scenarios is summarized below. The conditions remained the same for analysis of each of the CAC circuits unless otherwise noted The basis for
! the assumed initial and boundary conditions is provided in Response 4.
Loss of Olisite Power A loss of offsite power event is postulated to occur which results in a trip of the service waer I pump supplying the CAC piping. It is assumed that 30 seconds are required prior to service water pump restart. This time delay accounts for the time required to start the emergency diesel generators plus the sequencing delay for service water pump actuation.
The cooling water conditions are temperature of 85 degrees F and initial, cold water flow rate of 1600 gpm through the CAC circuit. Since no other accident initiators are present, the containment is assumed to be at a nominal temperature and therefore no heat is transferred between the service water piping and the containment atmosphere.
The initial upstream and downstream boundary pressures are based on plant data from Davis-Besse [ reference RAI 5.2] which indicate values of 74.3 psia and 22.1 psia, respectively.
Following pump trip and coastdown, the minimum header pressures are both set at 12.1 psia.
The header pressure data is approximated by using a linear change in pressure between the values listed above. Pump coastdown is assumea to occur over 5 seconds, while pump restart is assumed to occur over 1 second.
The end time for the calculation was set at 60 seconds, which was sufficient to see a return to all liquid equilibrium conditions in the service water piping.
LOCA Coincident with Loss of Offsite Power A LOCA coincic' nt with a loss of offsite power event is postulated to occur which results in a l trip of the servici water pump supplying the CAC piping. It is assumed that 30 seconds are required prior to ervice water pump restart. This time delay accounts for the time required to start the emergency diesel generators plus the sequencing delay for service water pump actuation.
The initial cooling water conditions are a temperature of 85 degrees F and initial, cold water flow
. rate of 1600 gpm through the CAC circuit. The LOCA condition results in an increase in the l containment gas temperature to a maximum of 264 degrees F at 15 seconds. As the containment l
i I
Docket Number 50-346 License Number NPF-3 Serial Number 2582 Attachment i Page 16 l
temperature rises, heat transfer across the CAC cooling tubes increases. Then, as the serv ce i water flow decreases due to the pump trip, boiling is expected to occur in the CAC tube region.
The initial upstream and downstream boundary pressures are based on plant data from Davis-Besse [ reference RAI 5.2] which indicate values of 74.3 psia and 22.1 psia, respectively. i Following pump trip and coastdown, the minimum header pressures are both set at 12.1 psia. l The data is approximated by using a linear change in pressure between the values listed above. ;
Pump coastdown is assumed to occur over 5 seconds, while pump restart is assumed to occur l over i second.
l The end time for the calculation was set at 60 seconds, which was sufficient to see a retum to all liquid equilibrium conditions in the service water piping. l RESULTS OF T11E TREMOLO THERM AL liYDRAULIC ANALYSES Thermal hydraulic results and forcing functions are provided as a function of time for each fluid node and each defined flow element. Please refer to Response 1 for a description of the plant physical model. Results indicate that waterhammer events occur during the initial draindown of system high points as well as during the refill stage of the accident. Furthermore, significant waterhammer events are calculated prior to the final column rejoining event. This results from the local, rapid condensation of steam on cold piping and lower temperature water during both the draindown of system high points and during the initial refill stage, and is consistent with experimental observations [ reference RAI 5-4]. A detailed discussion of the types of waterhammer events expected during the draindown and refill phases is provided in Sections 2.2 and 2.3 of FAI/98-126 provided in Response 2.
For all scenarios considered, a voided region developed in the service water piping nearest the air cooler region. Typically, the a'r cooler cooling tubes and the 6" feeder pipes would become completely voided. Also, voids of up to 50% were calculated to the bottom of the supply riser pipe at the 599' elevation and to the bottom of the return downcomer pipe at the 586'-6" elevation. For reference, these are points 4 and 13 in Figure RAI 10-1. In no case were voids j observed outside of the containment walls. I Thermal hydraulic calculations indicate the formation and collapse of voids for all sequences analyzed. In all sequences, a return to equilibrium conditions cccurred prior to the sequence end 3 time of 60 seconds. At that time only small residual gas bubbles remained in the system. i Furthermore, the return to equilibrium for the LOCA sequences indicated a return of the air l l coolers to their design service conditions. ;
j i As a result of the postulated accident conditions, waterhammer events were calculated to occur, ;
causing loads at the pipe system elbows and tees. These loads arc - norted in TREMOLO in the j form of reaction forces at the flow elements. The largest force. typically observed in the 1 elevated regions of the 6" supply and return piping and in the s piping adjacent to the 6" !
branches following pump restart, l
p Docket Number 50-346 License Number NPF-3 l Serial Number 2582 Attacnment 1 Page 17 l
l Reaction forces calculated by TREMOLO were then used as input to the piping stress analysis I code. The stresses resulting from the water hammer event were combined with pressure and weight stresses and compared to ASME Section III allewable stresses for Faulted Conditions.
The restraint and anchor loads consisted of a combination of weight, thermal and water hammer loads. The resulting calculation indicates that the system stresses remain within the ASME mtion III allowables.
i y ' EG/CR-5220 REVIEW xEG/CR-5220 discusses five types of condensation-induced waterhammer events. These are,
- 1. Subcooled water slugs
- 2. Watercannon
- 3. Trapped void collapse
- 4. Saturated water slugs
- 5. Thermalinversion A detailed review of the service water system piping geometry in conjunction with an estimate of the fluid state in the piping system prior to pump restart was performed to determine the likelihood of any of these five types of waterhammer events occurring in the Davis-Besse service water piping.
Subcooled Water Slugs Waterhammer events caused by subcooled water slugs require a large area of steam and subcooled water contact, such as would occur in a pipe with an initially stratified two-phase flow pattern. Rapid condensation of the steam on the subcooled water could result in the formation of a water slug that could cover the entire pipe cross-section, thus trapping a steam void between the water slug and a solid water column. As the trapped steam vo.. collapses, the water slug can be accelerated into the void region and come in contact with the water column causing a waterhammer event.
I As shown in Figure RAI 10-1, long horizontal runs of pipe exist in elevated portions of the ;
Davis-Besse service water system near the inlet and outlet of the air coolers. Furthermore, as discussed above, these piping sections are predicted to contain significant steam voids under the postulated accident conditions prior to the service water pump restart. However, as described in Section 2.2 and 2.3 of FAI/98-126 provided in Response 2, the specific accident conditions will !
not be conducive to the formation of subcooled water slugs. First, during the draindown phase of !
the accident and prior to pump restart, the flow rates of water spilling down the downcomer pipes are expected to be high enough to allow the horizontal pipes at the lower elevations to run full of water. Thus, stratified now patterns would not develop in the horizontal piping segments.
This is consistent with the TREMOLO calculations that indicate no progression of high
r
{
{
Docket Number 50-346 I License Number NPF-3 Serial Number 2582 Attachment I i Page 18 i
l l
temperature steam voids into the horizontal piping runs depicted as pipe sections 3-4 and 13-14 I in Figure RAI 10-1.
Following pump restart, the anti.:7ated refill velocity would be substantial and the subcooled cold water column originating at the pump would move down the pipe in essentially a plug manner, preventing the formation of stratified flow patterns at the liquid-steam void interface.
)
Thus, waterhammer events induced by the formation of subcooled water slugs are not likely to
{
occur and are eliminated from further consideration for the Davis-Besse analysis in response to 1 Generic Letter 96-06.
Watercannon Watercannon events occur when a steam void is trapped in a dead-ended pipe section, such as between a subcooled water column and a closed valve. Subsequent condensation of the steam void on the cold water surface or doe to heat losses can draw liquid into the voided region.
When the closed valve suddenly stops the water, a waterhammer event occurs. Watercannon events can also occur if steam flows through a constriction, such as would be present in a throttled flow control valve.
A review of the piping servicing the Davis-Besse containment air coolers indicates that there are no dead-ended piping volumes where steam voids could be trapped. Additionally, in the Davis- .
Besse analysis no steam voids progress past the containment boundaries (refer to Figure RAI 10- l 1). Thus, steam voids cannot ingress into portions of the service water system not directly servicing the containment air coolers. Note, for example, that while tees leading to other portions of the service water system are located on the supply and return piping at points 0,2,4, )
and 16 in Figure RAI 10-1, the voided region calculated by TREMOLO does not extend into l these flow junctions even in the most limiting case. Therefore, there is no possibility of trapping steam voids in dead-ended piping segments in other portions of the system.
Finally, the only location that might be susceptible to watercannon would be the air-operated flow control valves located on the return piping, outside of containment (see valves TV-1356, 1357, and 1358). During loss of power events, the air supply to the valve is lost and it fails in the full open position. During LOCA events, these valves go to their 'ull open position. Even if the flow control valve would remain partially closed, no steam voids are predicted to progress outside of containment, therefore no steam flow through the throttled valve would occur. There are no other throttled valves located on the piping servicing the containment air coolers. Thus, given the Davis-Besse piping geometry and accident conditions, there is no possibility for the occurrence of watercannon events and they are eliminated from further consideration l Trapped Void Collapse A void initially trapped between a water column and a closed valve or another stagnant water column can be collapsed due to repressurization of the system such as would occur during a pump restart. The pressurization accelerates one of the water columns which then moves through the voided region. As the voided region vanishes, the water column is decelerated leading to a waterhammer event.
f L
Docket Number 50-346 License Number NPF-3
- Serial Number 2582 Attachment 1 Page 19 This particular type of condensation-induced waterhammer event is most like the waterhammer events modeled by TREMOLO. Typically, during an accident scenario, a voided region of either high temperature steam generated in the air cooler region or low temperature steam formed from draindown of liquid from the piping highpoints, is located betwe.en an upstream and a downstream water column. Pressurization of the system then occurs following restart of the I service water pump or due to the onset of boiling in the air cooler region. In some instances, TREMOLO indicates the presence of multiple voided regions (e.g., one in the supply pipe high point, one in the return pipe high point, and one in the piping nearest the air cooler region), thus multiple waterhammer events could occur as each voided region cellapses. The magnitude of the resulting pressure rise is influenced by the thermal boundary between the steam void and the subcooled water column, the relative velocities of the two water columns surrounding the steam void, and the presence of residual voids following the larger scale void collapse. The TREMOLO code provides detailed transient calculations of dominant phenomena that would influence the occurrence of trapped void collapse waterhammer events. Thus, no separate analysis of trapped void collapse is necessary.
Saturated Water Slug If a condensate slug which has collected at the low point of a piping system, such as in a piping loop seal or in the vicinity of a closed valve at the bottom of a vertical piping run, is sudden:y accelerated due to valve openi..g or system pressurization, then a waterhammer event could be initiated.
A review of the Davis-Besse plant configuration indicates that there are no instances of slugs of condensate collecting at the bottom of an otherwise steam filled pipe being accelerated through the steam void due to a sudden pressurization of the system. Thus, saturated water slug events are not expected in the Davis-Besse service water piping and no further consideration of saturated water slug waterhammer events is necessary.
Thermal Inversion Thermal inversion occurs when warm water enters and begins to flow upwards in a vertical riser
. pipe while colder water is located in a reservoir at the top of the riser. Due to the static pressure drop in the warm fluid as it moves up the riser, flashing of the fluid may occur. This l
- hydraulically unstable situation of liquid supported above a steam column may lead to cold water draining back down the riser pipe. If a cold water column drains down the riser pipe and impacts the warm water column, below, then a waterhammer event will occur.
l Although the Davis-Besse configuration can lead to heated fluid flowing upward through vertical piping sections initially filled with subcooled fluid, flow reversal of the colder water in the elevated portions of the riser pipes is unlikely. Specifically, the vertical piping near the piping system high points (see segments 5-6 and Il-12 in Figure RAI 10-1) could see such a situation.
However, since the cold fluid is initially draining from the highpoints away from the air cooler region, the momentum of these flowing water columns is unlikely to be reversed once hot fluid begins to flow from the air cooler region. If the cold water columns stagnate to the point where L
r I
Docket Number 50-346 License Number NPF-3 Serial Number 2582 Attachment i Page 20 I flow reversal is a possibility, then the flow reversal and column rejoining waterhammer event will be modeled by TREMOLO. Benchmarks of TREMOLO against condenser pressure surge tests documented in Section 3.3.3 of FAI/98-126 provided in Response 2 indicate that TREMOLO will model flow reversal and any subsequent waterhammer events caused by column rejoining. Thus, no separate analyses of this class of waterhammer event, beyond what is inherently provided by TREMOLO, is needed.
SUMMARY
Details of the worst case scenarios have been presented. Results fo detailed, independent analyses of each of the three CAC trains under these postulated wea e , -e conditions indicate the CACs have remained operable.
An additional review of waterhammer events described in NUREG/CR-5220 indicate that the TREMOLO two-phase flow and waterhammer analyses were complete and no additional waterhammer events need be considered in the Davis-Besse response to Generic Letter 96-06.
References RAI 5.1 FAl/97-2, "FAI Experience on Waterhammer Phenomena in Containment Air Cooler Service Water Systems," (January,1997).
RAI 5.2 G. Thomas Elicson and Jim P. Burelbach, memo to file, " Summary of Design Inputs Received from Toledo Edison to Support Generic Letter 96-06 CAC Analysis, dated February 20,1997.
l l
l l
l l
___ l
Docket Number 50-346 License Number NPF-3
( . Serial Number 2582 L Attachment 1 l Page 21 1
Response 6 (NRC Question 2d):
Confinn that the analyses included a complete failure modes and effects analysis (FMFA) for all components (including electrical and pneumatic failures) that could impact performance of the cooling water system and confinn that the FMEA is documented and availablefor review, or explain why a complete andfully documented FMEA was notperformed.
A failure modes and effects analysis (FMEA) was not performed, nor is one necessary for developing the worst case scenario for waterhammer and two phase flow in the cooling water systems serving the containment air coolers during design-basis accident conditions. Rather, to assure a worst case scenario within the plant design basis, the detailed thermal hydraulic analyses, including sensitivity analyses presented in Response 8, consider design basis accid:nt conditions plus conservative assumptions to maximize void formation. Additional scenarios that might be derived from an FMEA would be bounded by the set of analyses obtained from the curret approach, therefore an FMEA is not warranted.
For instance, with design cooling water flow rates and peak DBA containment temperature and pressure conditions, there will be no two-phase flow conditions in the cooling water piping serving the containment air coolers out to the 30" return header. Thus, the principle assumption used in the analysis to obtain two-phase flow conditions is that the service water pump trips and begins to coast down to zero flow at the onset of the postulated accident. The pump is then assumed to remain in the tripped state until it is loaded onto the diesel generator and restarts.
This minimizes the cooling water flow through the containment air coolers and maximizes the time for water drain down.
To obtain an upper bound estimate of the two-phase flow conditions, design gas flow across the containment air cooler cooling coils is maintained throughout the postulated accident scenario !
(i.e. no fan coastdown is modelled) and a peak DBA containment pressure and temperature profile is imposed. This approach provides an upper bound within the plant design basis of the heat load and void formation in the cooling water piping serving the containment air coolers.
To obtain an estimate of the waterhammer effects resulting from a lower bound estimaic of the void formation in the containment air cooler piping, analyses were performed with the service water pump trip and zero heat load from the containment (i.e., conditions consistent with a postulated loss-of-offsite power and no LOCA). This case is representative of simple column !
rejoining on pump restart. i Finally, separate analyses were perfonned for each of the three containment air cooler piping l circuits to account for the influence of piping geometry on the calculated results.
In summary, analyses were performed that take into account the detailed geometry of each of the air cooler circuits, allow for significant void formation, and represent the possible range of containment heat loads to the containment air cooler cooling coils. Additional scenarios that l
l -
Docket Number 50-346 License Nuraber NPF-3 Serial Number 2582 Attachment i
! ' Page 22 might be derived from an FMEA would be bounded by the current set of analyses, therefore an
' FMEA is not warranted.
1 1
l l
l 1
l t
1 Docket Number 50-346 Ucense Number NPF-3 i
Serial Number 2582 Attachment 1 Page 23 Response 7 (NRC Question 2c); Explain andjustify all uses of "engineeringjudgement. "
Uses of engineering judgment are enumerated and described in additional detail below.
1
- 1. Selection of model boundaries at the supply and return main headers. This was done to focus the analysis on the specific ponions of piping servicing the containment air cooler.
Analysis results indicate that significant forces were limited to the 6" and 8" piping closest to the air cooler region and that steam voids were limited to the 6" and 8" piping, as well. Thus, selection of the main headers as the model boundaries places the boundaries sufficiently far away from the region of two-phase flow and waterhammer events such that the boundary placement will not adversely affect the analysis.
- 2. Treatment ofparidfeljlow paths as a single equivalentflow path. The equivalent now paths i were created to preserve the fluid velocity through any individual flow branch while delivering the total design flow to the CAC region. Thus, calculations of any column l separation, boiling, or the timing and magnitude of any condensation-induced waterhammer events will not be adversely affected by this modeling approach.
- 3. Nodalization scheme. Selection of the nodalization scheme is governed by several factors that require judgment on behalf of the computer modeler. First, the current problem considers pipes of varying lengths and diameters. Thus, varying node sizes are necessary to adequately model the approximately 1000 feet of pipe located between the main supply and return headers. The nodal length-to-diameter ratios (UD) used in the models for the three CACs range from 5 to 40. Typically, smaller node UD ratios are used in the model region closest to the air coolers. For the current model the node UD in the 6" and 8" pipe is approximately 10. Use of nodes with reduced UD ratios in the air cooler regions is appropriate since the two-phase flow and waterhammer events are expected to occur in these regions and the smaller UD will provide greater numerical resolution. Second, the nodalization scheme is selected based on experience from performing code benchmarks against available experimental data. During the benchmark process, numerous nodalization schemes are attempted until a range of node sizes is determined that will balance numerical l
stability, code run time, and reasonable results as compared to the experimental data. The nodal UD ratios used in the Davis-Besse model are consistent with those used in the benchmarking activities documented in Section 3 of FAl/98-126. Finally, the time step selection must be adjusted based on the selected node size to ensure that the time step is well below the Courant limit. Consideration was given to the factors discussed above to select an appropriate nodalization scheme for use in the Davis-Besse analysis.
- 4. Initial nitrogen partialpressure. The TREMOLO code models non-condensible gas coming out of solution. This model is based on an assumed initial non-condensible gas panial pressure. The gas then comes out of solution as the total fluid pressure approaches (but while it is still typically a few percent higher than) the initial gas pressure. Based on experiments performed by Fauske & Associates, .Tc. on room temperature water initially saturated with air at I atmosphere, gas was observe ' o come out of solution when the fluid pressure was decreased to 6 psia or less. The initial nitrogen partial pressure assumed for the Davis-Besse analysis is slightly less than 5 psi. Thus, this modeling assumption is consistent with
I' Docket Number 50-346 License Number NPF-3 Serial Number 2582 l Attachment 1 -
Page 24 experimental observations. Please refer to Section 2.1 of FAI/98-126 for additional details regarding non-condensible gas content in subcooled water.
l i
1 Docket Number 50-346 License Number NPF-3 Serial Number 2582 Attachment 1 Page 25 Response 8 (NRC Question 3):
Determine the uncertainty in the waterhammer and two-phase flow analyses, explain how the uncertainty was determined, and how it was accountedfor in the analyses to assure conservative resultsfor the Davis-Besse plant.
Two general classes of uncertainties have been accounted for in the detailed waterhammer and two-phase flow analyses developed for Davis-Besse: uncertainties in the definition of the postulated accident scenario; and engineering uncertainties (i.e., uncertainties and limitations in modeling of the physical processes and in the development of the computer model for the service water piping system).
i Engineering uncertainties have been determined through experimental observations and independent investigations of the controlling physical processes, and have been accounted for in the detailed calculations by appropriate selection of "modeling parameters" (see response to RAI 4 for a discusdon of key modeling parameters and assumptions). Please refer to FAI/98-126 for i a detailed discussion of the controlling physical processes (see FAI/98-126, Section 2.0). For r summary of relevant experimental observations, including computer code benchmarks, see FAI/98-126, Section 4.0.
l l
Uncertainties in the accident scenario definition have been accounted for by performing a series of detailed waterhammer and two-phase flow calculations that consider a conservative set of l
design basis accident conditions to obtain two-phase flow in the cooling water system. Please '
refer to Response 6 for a general description of the conservative approach used in the scenario definition, and please refer to Response 5 for additional details regarding the conservative l assumptims used in the detailed analyses. Also, results of a sensitivity study to consider additional variations in the sequence definitions are presented below.
Table RAI 8-1 provides a summary of key parameters used in the sequence definition for the detailed GL 96-06 thermal hydraulic analyses. The column entitled " Analysis value" indicates
! the values used for the base case analyses for which results have already been previously l
reported. Results of > Mitional analyses are presented here to examine the sensitivity of the reported results to ' .wiation in the key parameters used in the sequence definition across a credible range of val es.
The sensitivity calculations presented here were performed with the Davis-Besse CAC-3 plant model. This model is typical of all three CAC's and similar results of this sensitivity analysis would be expected for CAC's 1 and 2. Table RAI 8-2 describes the additional sensitivity runs, while results from the sensitivity runs are e ampared to the base case analysis in Figures RAI 8-1 ami D AI 8-2.
Figure RAI 8-1 compares the total void volume in the service water pipe as a function of time.
Time zero corresponds to the onset of the LOCA and the time of pump trip. The service water pump is then assume.d to have a 5 second coastdown. In all cases, voiding first appears between 4 and 5 seconds after pump trip. The void volume continues to grow until shortly after the time
i Docket Number 50-346 License Number NPF-3 ;
Serial Number 2582 l Attachment 1 Page 26 i
of pump restart. The peak void volume for case 3A (sensitivity to pump restart at 25 seconds) occurs at about 26 seconds. For the remaining cases, the peak void is observed at 31 seconds.
The case producing the maximum void is case 3D, with the maximum lake water temperature (90 degrees F), while the sensitivity run using 48 degrees F lake water produced the minimum -
total void volume.
TABLE RAI 8-1 Summary of Sensitivity Parameters for Davis-Besse GL %-06 Analysis Analysis Conservative Lower Upper Best i Description ralue / Pessimistic Optimistic bound bound estimate l Pump coastdown time 3 5 5 5 5 5 (sec)'
Pump restart: time to 1 --- --
1 4 4 )
full speed (sec)
Cooling water initial 85 -- ---
48 90 80 temperature (F) 2 Containment peak gas 260 260 255 - -- 260 temperature during LOCA (F) 3 Containment peak gas --
289 --- - - - - - . ---
temperature during MSLB (F)
Time prior to pump 30 - ---
25 40 30 !
restart (sec) d Pumped flow rate per 1600 --- - - -
1150 1600 1350 CAC (rpm)
CAC fouling factor .00045 -- - - -
0.000 0.002 0.00N5 Residual void fraction .001 .001 .005 --- ---
N/A following void collapse NOTES:
- 1. Pump restart to full speed is based on value given for similar pump in the USAR.
- 2. Davis-Besse is in the process ofincreasing maximum allowable temperature to 90F. The best estimate value 80F is based on average value for July and August.
- 3. Containment vapor temperature of 260F is based on computer output.
- 4. The lower bound value for pump restart is based on minimum delays associated with instrumentation and diesel start and sequencer.
Results indicate that the extent of the voiding is more sensitive to the lake water temperature than it is to the containment conditions (note that the void volume for the limiting lake water l temperature cases bound the MSLB -case 3E - void volume). In all cases, a return to all liquid l
conditions occurs prior to the run end time of 60 seconds.
Although the maximum void volume for case 3D is approximately 45% larger than the case 3C maximum void volume (see Figure RAI 8-1), the maximum distance the voided region has traveled on either side of the fan coil region is the same for both cases. The void travel is controlled by the piping geometry, which includes a significant portion of piping near the air
f Docket Number 50-346 License Number NPF-3 Serial Number 2582 Attachment 1 Page 27 cooler in a loop seal-type geometry. Figure RAI 10-1 provides an elevation schematic of the air cooler service water piping. Comparing the axial void profile from Figure RAI 8-2 to the schematic indicates that the hot steam produced in the air cooler remains trapped between the piping high points at 614'-6". Water drain down on the outside of the high points results in some additional " cold void" due to column separation.
Finally, Figure RAI 8-2 compares the peak magnitudes of the reaction forces at the piping 90-degree elbows and tees for the sensitivity study. Although the timing and duration of these reaction forces must be considered to develop estimates of the loading on pipe hangers and supports, the summary information in the figure provides a general idea of the sensitivity of the results to changes in key input parameters. Since the changes in sequence definition can lead to changes in the void size, void distribution, and renti times, the location and timing of each individual condensation-induced waterhammer event may vary from one sensitivity case to another. Also, as indicated in Figure 8-2, there is no single sequence dennition that produces the bounding force at all locations. In general, however, the base case sequence definition provides an estimate of the peak forces that is representative of all CAC-3 sensitivity analyses. Finally, results for case 3E, which are obtained by applying the main steam line break containment temperature profile, are overstated because the air cooler heat transfer model is conservative for a MSLB compared to . LOCA. This is due to the fact that, although the containment atmosphere is superheated with a o w mass fraction of steam, the heat transfer model does not consider the steam superheating m reduced steam mass fraction in the gas passing over the fan coils for a MSLB.
The results for case 3b (for 40 second delay on pump restart) shows a peak void volume that is less than the base case ( for 30 second delay on pump restart). The curve for case 3b is also displaced in time as compared to the other cases. This change occurred as a result of a slight computer code improvement which was used only for case 3b. Using the revised code for the base case (data not presented), it was verified that the total void volume was slightly less for the 30 second delay than for the 40 second delay. Thus, although a longer delay would be expected to physically translate into slightly larger loads than a shorter delay, the result remains bounded by the original, more conservative calculation with the earlier code version.
l l
r Docket Number 50-346 License Number NPF-3 Serial Number 2582 Attachment 1 Page 28 I
TABLE 8-2 Description of Sensitivity Analyses Performed for Davis Besse CAC-3 Case ID Description 3A LOCA+ LOOP with minimum pump restart time (25 seconds) 3B' LOCA+ LOOP with maximum pump restart time (40 seconds) 2 3C LOCA+ LOOP with cold lake water temperature (48 F) 3D LOCA+ LOOP with maximum lake water temperature (90 F) 3 3E Main steam line break containment temperature profile + LOOP NOTES:
- 1. A different version of the TREMOLO code was used to perform the 40 second pump restart analysis. All other sensitivity analyses were performed with the same code version as was used in the base case analysis.
- 2. Although the minimum allowable lake water temperature is 40 F, the TREMOLO equation's, lower limit is 48 F, thus the cold water sensitivity case was performed at 48 F rather than 40 F.
- 3. Although superheated gas conditions exist in containment during the MSLB, the containment temperature profile for the MSLB was applied assuming saturated steam conditions at the given gas temperature. The MSLB containment temperature profile is based on assumed single failure conditions where auxiliary feedwater to the faulted steam generator was not isolated.
l l
l i
l
)
l l
vS E ZC3 $ [t0" 2 .
n0
>A 5g ~
C
- $ yC 0
._ T _. - _ _! _ - - - 2 ___~ - _ I- 6 _
~___
5 s
g
.'A 0 -
E 5 _
M \,
U _
' \ ',' s' _
L O
)p
)
t i .' ',. s' '
V
)t
',.' '., s c DI t
ipv_ s s '
O V
ip .l
. v_3 v_ O a c
',', \ .'..'. \ \
) - s L
A abc 33o t
ip
.' s l 0
aal ) ' 4 T ccbt e. .. \
O lood+p i
3 ., x '
T l .
a ,
s )
- )ddb b v_ v_
c o
N
' S Y it((cd 33 l
b D _
D p w
U .CCea d
(
g N _
dEEcc T
S l
oSSeol l K
\ O _
\
C e _
v_5 0bb A *
\
Y dd \
T 3 24(( E -
\
\
I 0 E a R I
V cTTRR B
. 3 S I
oAAEE - \
q >
(
TI :b E l
TT N ,-
SAdTTAA :N BI -
~ '
E S OETT EE MA SM '
M LSSSKK E
> I T
3+AEEAA CRRLL +T S - , >
C R ,
AREPPFF EN CESMM I i-i 0
WAUU80 WA EOBPP49 OM -
2 SP P S
EF F '
BO -
S
- O-IS -
0E SS 5
- l VS -
AO - -
DL -
- O-L l i 0 1
' /'
i '
- _. _____2- .-
_ - [ _
0 o" o ae n o o" oie o o
_ m < & u. -
l3 m E .n. o y Q O >
J_ < H O r M$ ga"I gOeE g a 40 0 o"Cm1B aEa Og d@,O5 oygC$n% l, M5"
S3J!EI NnmqaJ gg8E ynnaywaul I d"33 EO o
-i ii i.i
,,iil I,,ii
!iii ;;;
i,,,Iiii,I,iiili,!,
m?
i t
- s
! f
. .. p . . , . . .q . . . . . . . p . . . 8. -
. (
l P i
j o o3 <
.7... ..,.. . . . . .
3o- j i ! ! . . .!8 P
- i - -
i ,r e g
- sI
. g.. p l4g
~ ' !
]
35 6 5
. . _ .i
- J 5 " 6 i <
5 ..3"Z, c ~
3R! M-
~^* . . v ... .. .. . .. .. . . . . . . . . - . . . . . . . . . . _
g
- x
. ;a ,
- 3 a - a9 si. ._ =
- i
- o=* o_ - :c
%",a : :E_E .
- - z m w I -t - -i- > d-
- h ,- [-g g-}i'NO 4.. .._
E n" w ". n: n: . ,
M S< 00 < . -
c s>"m * ; !
i I 88!
id z m
= ot o E;0***M*
- s s s: ;$
i
'. = :
- i
- 0. 4 -
c mdn E o . . ,: m m
,." c .'
i 3
O m m 1 .r m
3 E+;g ;; +" i
> mmu.mm ,, m2 a "EE gp -
>cc=o .! ,
. -. . . .j . . . . . . . . . .
..g .. 4 gg mm , .
.c o o . i. -
j"eoo<o "x ;
og g _. . . + . . ..
..+.. ..+.. . . . . . . . _,,g
- g
, 4
- i t ;
o
- i oi o -
I Liiiji46li.ijiiiiiii iiiiliiiil,iiiiiiil..._
01 9 8 7 6 5 4 3 2 1 O
'01 x )FBL( EDUTINGAM ECROF MUMIX AM l
l dID1fM3 MVI 8-g ygux!mmu jon.aas ni a!qou.s nup taas y.om I(ups-Itassa 3V3-C l
,juanoj o saus!ila!;Inun! Is!s
Docket Number 50-346 License Number NPF-3 Serial Number 2582 Attachment 1 Page 31 Response 9 (NRC Question 4):
Confirm that the waterhammer and two-phaseflow loading conditions do not exceed any design specifications or recommended service conditionsfor the piping system and components, including those stated by equipment vendors. Also, confirm that the system will continue to perform its design-basis functions as assumed in the safety analysis reportfor thefacility.
As previously stated in Davis-Besse letters (Serial Numbers 2439,2442,2473,2488, and 2554) ,
the piping systems and components under consideration remain within the design specifications and recommended service conditions for the postulated events. The CAC cooling coils have been evaluated for temperatures and pressures credible during the postulated events.
- Structurally, the housing, frame and mounting of the CACs have been evaluated for water hammer and two phase flow induced loading. These evaluations show that the CACs will continue to perform their design-basis functions and therefore remain operable.
I i
l l
L
Docket Number 50-346 License Number NPF-3 Serial Number 2582 Attachment i Page 32 Response 10 (NRC Question 5):
Provide a simplified diagram of the system, showing major components, active components, relative elevations, lengths ofpiping runs, and the location of any orifices andjlow restrictions.
A simplified diagram of the service water piping systems analyzed for the Containment Air Cooler (CAC) trains 1,2, and 3 is shown in Figures 10-1. The diagram is typical for all three CAC trains, however valve numbers specific to CAC-1 are indicated. Numbers enclosed by circles identify the end points of pipe runs in the figure. The lengths of the identified pipe runs are summarized in Tables 10-1,10-2, and 10-3, respectively for the three CAC trains. The discharge valves are normally open, air operated valves, which fail open. There are no intentional flow restrictions or orifices.
TABLE 101 DAVIS-BESSE CAC-1 LENGTH OF PIPE RUNS APPEARING IN FIGURE 10-1 Length Length Length Length Pipe Run (ft) Pipe Run (ft) Pipe Run (ft) Pipe Run (ft) 0-1 64 5-6 74 10-11 74 15-16 52 ;
l-2 77 6-7 74 11-12 76 16-17 12 l 2-3 69 7-8 33 12-13 56 17-18 51 3-4 37 8-9 24 13-14 11 18-19 23 4-5 51 9-10 33 14-15 44 19-20 64 l TABLE 10-2 DAVIS-HESSE CAC-2 LENGTH OF PIPE RUNS APPEARING IN FIGURE 10-1 Length Length Length Length Pipe Run (ft) Pipe Run (ft) Pipe Run (ft) Pipe Run (ft) 0-1 37 5-6 72 10-11 39 15-16 49 l-2 85.7 6-7 37 11-12 74 16-17 40 2-3 71.3 7-8 30 12-13 54 17-18 32 3-1 45.3 8-9 24 13-14 10.7 18-19 45 4-5 46.4 9-10 30 14-15 48 19-20 40
Docket Number 50-346 License Number NPF-3 Serial Number 2582 Attachment 1 Page 33 TABLE 10-3 DAVIS-BESSE CAC-3 LENGTH OF PIPE RUNS APPEARING IN FIGURE 10-1 Length Length Length Length Pipe Run (ft) Pipe Run (ft) Pipe Run (ft) Pipe Run (ft) 0-1 64 5-6 74 10-11 60 15-16 56 1-2 77 6-7 60 11-12 76 16-17 12 2-3 69 7-8 33 12-13 56 17-18 51 3-4 41 8-9 24 13-14 11 18-19 23 4-5 51 9-10 33 14-15 44 19-20 64 i
ogwa"Z gk 9ye r 8aarG e
s Z y (
- m c, ma u*~ Zc3
~
5
>::un7gaE '
=
2, o.
Taa tAo l l' rL e
- - E idE )
8" 7
sD e y
- - ~~ -- 5 0 l n
1 HON 0"M o ,
- - ~ - ,
3 1 9- -
- - ~~
7 a C
- - - - D
. A C
6 - r
- - - 3 -
4 o
- 3' - 5 7 s f
- - 9 - 5 - r 1 n
- 5 c S w o
- - , - - - C C h
- - "6- - E s
- - e 4
6
- 9 5
- 6' 8 - mo r a
Q- s g
- '-4 F r r 6" -
1
- - = e 1 - b
- 6 -
. x 2 2 -.
- m 1'
1 6
03 8
r 6 -
u n .
- d -- e
- - 9 '6 e. - - v 5 - 3" l a
- - C v
-- - g ' A C
s
- - n W" , 2 r -
i a -
- ' 7 6 3
v- t r
- - '- A- -
C FI.
6" ' 7 C-
- - s 9 3 6
'f -
A
- 2'-@-
1 5
- 2 2
- '5 8
o- - C
- 6 - - F - 5m- - e e
-=
- '- "8 t o-to -
B-ht f
r 6
- 6%
4 1
6 9' -
p 6
3 1
3-l o
a 5_ V2 1 c
- - 7 HF- - - - i p
4 C y
- f - A -
C
(
t
- - 0 - n
- - - 9 5
i o
t
- -- - 8"- - - a W ve
- '4 ) -
C
- - - 9 5 (3 -
7' - C' lE g-o- -
- ' 5 8
T . C
- A C
- - - - ' - 2 4 jx 5 2L .
e
- - - '- ' - ET s s
DR 0
- - - - 1 7 1 '- O TA e B-
'3 7 @MS i s
5 8 v
- - - - - ~6 5 a D
0
- - - - - - 7 1
- . 5
- - -- - 3 - - 0
- 7 -
1 i-g - 2' I 6
- - - - - A g'
gA- -
- 5- _
R e
r u
g
'0
'0 0' '0 '
0 8
0 0 6
i F
2 1 0 9 7 6 6- 6 5 5 5 5
Docket Number 50-346 License Number NPF-3 Serial Number 2582 FAI/98-126 Waterhammer Phenomena in Containment Air Cooler Service Water Systems J