NSD-NRC-96-4780, Forwards Three Copies of Licensee Responses to NRC RAI on AP600

Forwards Three Copies of Licensee Responses to NRC RAI on AP600

ML20116C716
Person / Time
Site:	05200003
Issue date:	07/26/1996
From:	Mcintyre B WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP.
To:	Quay T NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
NSD-NRC-96-4780, NUDOCS 9608010081
Download: ML20116C716 (11)

Text

, - . _ - _ . - - -

.. ' s" cu 1

. Westinghouse Energy Systems Box 355 Electric Corporation Pittsburgh Pennsytvania 15230-0355 NSD-NRC-96-4780 DCP/NRC0559 Docket No.: STN-52-003 I

July 26,1996 Document Control Desk i U.S. Nuclear Regulatory Commission  ;

l Washington, D.C. 20555 ATTENTION: T.R. QUAY l

SUBJECT:

WESTINGHOUSE RESPONSES TO NRC REQUESTS FOR ADDITIONAL INFORMATION ON THE AP600 ,

l l

Dear Mr. Quay:

l l

Enclosed are three copies of the Westinghouse responses to NRC requests for additional information on the AP600 topics. Responses to 410.28.2,410.285, and 440.286 respond to questions on the S3AR. The NOTRUMP CAD is discussed in RAIs 440.332,440333 and 440.340 while EGOTHIC l

is discussed in responses to RAls 480.334 and 480.342.

l The NRC technical staff should review these responses as a part of their review of the AP600 design.

These responses close the eight RAls.

l Please contact Brian A. McIntyre on (412) 374-4334 if you have any questions concerning this l ]

l transmittal. i A f$/

Brian A. McIntyre, Manager Advanced Plant Safety and Licensirg i

/nja Enclosures cc: T. Kenyon, NRC (w/o enclosures) \I

D. Jackson, NRC (IEl)

E. Throm, NRC (w/o enclosures) >

J. Kudrick, NRC (w/o enclosures) )) l P. Boehnert, ACRS (4EI)

B. Grimes, NRC (w/o enclosures) ~

N. Liparulo, Westinghouse (w/o enclosures) 2859A 9608010001 960726 PDR ADOCK 05200003 A PDR

m NRC REQUEST FOR ADDITIONAL INFORMATION i

l l Question 410.282 I l

The demineralized water transfer and storage system was designed with three 50-percent capacity motor-driven centrifugal horizontal pumps to provide the plant demineralized water system pressure and capacity. In Revision ,

( 3 of the SSAR, Westinghouse reduced the number of pumps to two of the same capacity and same type pumps to l perform the same fur:ction as the three pumps. Discuss the reason for reducing the number of pumps. 1

? l l

Response

l SSAR subsection 9.2.4.2.2, Revision 6, states: "Each [of the two] pump [s] provides full flow recirculation through l the catalytic oxygen reduction unit as well as providing the required system demand." This is the proper design requirement for the demineralized water system pumps.

SSAR Revision: NONE l

l l

1 l

l I

1 l

l i 1 i

1 1

l l

W Westinghouse

l NRC REQUEST FOR ADDITIONAL INFORMATION R

Question 410.285 In Revision 3 of the SSAR, Westinghouse changed the design of the waste water system by collecting the lew volume wastes in the drain tanks rather than in the turbine building sumps and changed the waste water collection systems for the auxiliary building, diesel generator building, and diesel fuel oil storage area. Explain the reason for and advantages of these changes. Will the drain tanks replace the sumps in the design change?

Response

SSAR subsection 9.2.9.2.1, Revision 6, provides a more complete description of this nonsafety-related waste water system and its drain collection philosophy. There is better control of the waste water, its contents and processing if it is collected in tanks. Sumps still exist in the design to act as initial collection sites for drainage from some areas of the plant.

SSAR Revision: NONE l

l T Westinghouse

^

4

NRC REQUEST FOR ADDITIONAL INFORMATION 1

Question 410.286 The hot water heating system was initially designed with two 100-percent capacity pumps to distribute hot water to the various systems. In Revision 3 of the SSAR, Westinghouse changed the design to two 50-percent capacity pumps. Provide reasons for reducing the pump capacity.

Response

There are no design requirements that dictate the use of two 100-percent capacity pumps for this nonsafety-related system. During most at-power operations, a single 50-percent capacity pump is sufficient to protect equipment and piping. Two 50-percent capacity pumps are sufficient to prevent equipment and water filled piping from freezing i during cold weather plant outages.

SSAR Revision: NONE I

)

i l

l I

1 l

l 1

^

W Westinghouse

i i

J j

i I

NRC REQUEST FOR ADDITIONAL INFORMATION i

l Question 440.332 l

Re: WCAP-14206 (NOTRUMP CAD)

On Page 4-3, reference is made to the "long term cooling code" which is used once the IRWST begins draining.

No references nor description and use of this code is given. Please provide a document describing the methods and models comprising the long term cooling code. PIRT components and applicable phenomena should also be addressed. This document should also describe the comparisons of the long term cooling code predictions to separate effects and integral tests. The document should further describe how the code is initialized from the NOTRUMP code. Please also describe the containment model and the modeling of the subcompartments/ sump and their effect on recirculation during the long term.

Response

The WCOBRA/ TRAC computer code is employed in the long-term cooling analyses of the AP600. It is applied in a " window mode" approach as described in Reference 440.332-1. That reference also describes the model to be used and its validation by comparison with data from the Oregon State University integral test facility. Reference 440.332-1 will be superseded by the OSU Final Validation Report when it is issued during the fourth quarter of 1996.

The Final Validation Report will include the final PIRT for long-term cooling.

ECCS performance analyses of small break LOCA events are performed using the NOTRUMP computer code from the initiation of the break through initial IRWST draining. Conditions at the start of a long-term cooling window for a small break LOCA event are identified from the NOTRUMP-calculated values for the break. The interface between the WCOBRA/ TRAC ECCS performance analysis and the containment transient predictions of WGOTHIC is detailed for the AP600 long-term cooling application in Reference 440.332-2.

References:

440.332-l Chow, S.K. et.al. LTCT-GSR-003, "WCOBRA/ TRAC OSU Long-Term Cooling Preliminary Validation Report," August,1995.

l 440.332-2 RAI 440.554, Revision i Response submitted June,1996.

l l

\

l SSAR Revision: NONE

)

440.332 i 3 Westfrigt100Se l

NRC REQUEST FOR ADDITIONAL INFORMATION Question 440.333 Re: WCAP-14206 (NOTRUMP CAD)

On Page 4-3 last paragraph of Section 4.0 states that a simple boundary condition is used for the containment building back pressure. Pleasejustify the use of a nxed pressure boundary condition since the response of the safety systems, particularly initiation ofIRWST injection, depend on containment pressure. Please show AP-600 response to several back pressures and demonstrate the value chosen is appropriate. Also, please explain the interaction of the sump during the long term given this simplifying boundary condition. The dynamics of the interactions of the sump and IRWST appear to be omitted with this modeling approach, which could affect long term behavior in the reactor coolant system. That is, communication paths between the many containment subcompartments could also setup an oscillatory flow that could affect RCS behavior during the long term. Fluid needed for long term circulation could also be trapped in the subcompartments reducing the long term source of cooling water source. A description of how the containment subcompartments communicate and their influence on the long term sump behavior is also desired. Please explain and provide more detailed information justifying the use of a fixed pressure.

Response

The effect of containment pressure on the AP600 passive safety systems response has been addressed in Reference l 440.333-1. As reported therein, the two-inch cold leg break LOCA was analyzed at an assumed containment pressure of 14.7 psia, and also using the AP600 containment pressure transient calculated for that break size with the WGOTHIC computer code (Reference 440.333-2); the WGOTHIC containment pressure values were calculated to vary between 28 and 26 psia when the pressurizer pressure had decreased below 100 psia. Input for the two i

NOTRUMP cases was identical except for the different containment pressure input values. Reference 440.333-1

! indicates that the use of atmospheric pressure is conservative in NOTRUMP ECCS performance analyses of AP600 because it maximizes the RCS depressurization necessary to achieve injection from the IRWST. Therefore,14.7 psia

is established as a conservative containment pressure boundary condition to apply in the AP600 SSAR small break LOCA analysis.

i For long-term cooling, the sump is isolated from the direct vessel injection line piping by squib valves until the sump l fills beyond the design basis fillup level of 107'2" A stable, uniform water level is established within the sump l before the sump isolation valves open upon receipt of a Low-3 IRWST level signal. Once sump injection begins, j change in the sump water level is very small because the cross-sectional area at the fillup elevation (slightly above i the lower compartment wall level)is huge. After the sump has become the primary source of safety injection water, l no significant potential exists for the preferential filling of any containment subcompartment to establish the

! difference in level needed to create an oscillatory water flow within the containment sump. The methodology of using WGOTHIC to calculate containment phenomena during long-term cooling is provided in Reference 440.333-3.

W-Westinghouse

l

. i I

NRC REQUEST FOR ADDITIONAL INFORMATION ll l

References:

l 4

440.333-1 Westinghouse letter NTD-NRC-94-4239, September,1994.

l l

440.333-2 WCAP-14382, "WGOTHIC Code Description and Validation." 1 440.333-3 RAI 440.554, Revision 1 Response, dated June,1996.

SSAR Revision: NONE i l l

l l

l l

l l

440.333-2 W Westingh00Se

[

l t

i

. I NRC REQUEST FOR ADDITIONAL INFORMATION l

l E l l i l

Question 440.340 1

Re: WCAP-14206 (NOTRUMP CAD) l l

In view of the lengthy transient times for AP600, there is the potential for boiling to persist for extended periods of time at very low pressure. As such, boric acid can build-up in the core and vessel lower plenum regions and l potentially precipit-te. Please provide an analysis showing the boron concentration in the core for the break showing the highest accuraubted concentration. Also, please explain the influence of the sump on this event during the long l term. Also, can subcooled conditions in the lower plenum and lower core regions develop after long periods of boiling that could cause precipitation while the hotter upper core conditions maintain solubility? Please explain.

Response

i l Please consult the response of Reference 440.340-1 for a discussion of the continuous flushing of the core which I occurs throughout the long-term cooling phase of AP600 LOCA events. Many of the Oregon State University facility tests were carried out for 6000-15,000 seconds of sump injection. As reported in Reference 440.340-2, throughout this extended duration every test exhibited continuous injection from the sump into the reactor vessel and liquid flow exiting continuously through the ADS fourth-stage flow paths. This behavior was observed to be stable no matter the extent to which boiling occurred in the core, and boiling did occur in the core in every extended l duration case in Reference 440.340-2.

l Flow continuously flushing the AP600 core during IRWST and sump injection prevents boron from concentrating l to any significant extent in the core. So, during long term cooling no occurrence of a subcooled liquid condition in the lower core region or in the reactor vessel lower plenum could cause boron to precipitate due to reduced solubility at a lower temperature. Furthermore, the AP600 sump cannot become depleted of boron due to solute concentration within the core; such a core recriticality scenario during long term cooling does not apply to AP600.

l

Reference:

l l 440.340-1 Response to RAI 440.355, May,1996.

i l

440.340-2 Andreychek, T. S. et al., "AP600 Low-Pressure Integral Systems Test at Oregon State University Test Analysis Report," WCAP-14292, Revision I, September 1995.

SSAR Revision: NONE 440.340 W-Westingt100Se 4

NRC REQUEST FOR ADDITIONAL INFORMATION Question 480.334 Re: WGOTUIC MODELS AND PHENOMENA) FINITE DIFFERENCE (FD) CAL.CULATIONS:

Has WEC assessed the impact of the first order method on the ability of the code to predict stratification? i l

Response

There has been no specific study on the ability of WGOTHIC to predict stratification relative to the ability of other codes employing higher order methods. One recognized shortcoming of the first order method used in WGOTHIC is that it can overpredict diffusion and smear out gradients. This numerical diffusion is e.r:ctly proportional to fluid velocity across the gradient of interest. In the case of stratification, the velocity across the gradient and the l

corresponding convection terms must be small ou.tive to the terms responsible for the stratification in order for the j gradient to persist. Therefore we don't expect st#.iication to be any more of a problem for WGOTHIC than it would I be for a code that uses a higher order method.

WGOTHIC's, or any codes, ability to predict stratification hinges on its ability to predict distributions in buoyancy dominated problems. The GOTHIC Qualification Report (Reference 480.334-1) documents several tests and models where buoyancy is a dominant mechanism. These include Test 6,12 and 20 in the Battelle Model Containment, Test

]

5 and 6 in the HEDL ice condenser containment model and HDR test V44 and T31.5. These test show that GOTHIC 1 predicts, with reasonable accuracy in complex modeling arrangements, the effects of buoyancy. The Battelle Model containment tests 12 and 20 and the HDR tests were modeled using the lumped volume approach. The rest of the I tests were modeled using subdivided volumes. Both the lumped and subdivided modelling approaches use the same first order upwind numerical scheme. The tests in the Battelle Model Containment demonstrate the importance of the stored heat in the solid conductors in determining when and where stratification occurs. GOTHIC is able to demonstrate this same effect. The HDR tests showed significant variation in the intermediate time range vertical temperature profiles following a stream release. GOTHIC is in reasonable agreement with the test data and demonstrates the importance of tracking the flow of condensate to lower regions of the containment in determining temperature distribution. HDR Test T31.5 included a hydrogen release. Except for the break room, where the data is questionable, GOTHIC is reasonably accurate in modeling the hydrogen distribution. In this test, GOTHIC overpredicts the amount of hydrogen that migrates down to rooms just below the break room, but at lower levels of the containment GOTHIC predicts very little hydrogen in agreement with the test data. For the classical prediction of buoyancy driven flow in a closed cavity (one side heated, one side cooled), it has been shown that first order methods give reasonable results. Second order methods perform better, but the advantage is not overwhelming.

Higher order methods perform no better than the first order method for ths problem, Reference 480.334-2.

Although there are clear advantages to using second order methods to reduce numerical diffusion, their application to complex multiphase, multicomponent problems in complex geometries such as the blowdown experiments in HDR that include flashing of two phase jets, condensation on walls and equipment, revaporization, drop entrainment and transport, drop deposition, pool formation and runoff, gas diffusion, evaporation from pools, films and drops with the attendant numerical difficulties of phase depletion, flow regime prediction and water packing has not been proven.

SSAR Revision: NONE T westingtiouse i

I

. . l l

NRC REQUEST FOR ADDITIONAL INFORMATION I h l

References:

480.334-1 NTD-NRC-95 4563, " GOTHIC Version 4.0 Documentation," September 21, 1995, Enclosure 1, l

" GOTHIC Containment Analysis Package Qualification Report."

l 480.334-2 See G. De Vahl Davis, " Natural Convection of Air in a Square Cavity: A Benchmark Numerical Solution", Int. J. Numerical Methods in Fluids, Vol. 3,1983.] l l

i 480.334-2 W-Westinghouse

a .

e NRC REQUEST FOR ADDmONAL INFORMATION l, l Question 480.342 Re: (WCAP-13246) FOG FORMATION IN THE EXTERNAL ANNULUS How does WGOTHIC model fog formation? Is the mass of the water present as fog taken into account in calculating the gas densities? Is its effect upon buoyancy taken into account?

Response

WGOTHIC does not include a fog formation model that would spontaneously generate drops in a supersaturated atmosphere. The only mechanisms for drop formation are spray nozzles, sources from injected flows and entrainment from films. Drops, regardless of how they are formed, are treated as a separate field with their own mass, energy and momentum balances. They influence the vapor flow through the interphase drag. When the drops are very small, such as in a fog, the motion of the drops is controlled by the interphase drag term and the drops will move with nearly the same velocity as the vapor. The drag force on the drops is about the same (but opposite in sign) as the body force. This same drag force, but opposite in direction, is applied to the vapor momentum balance. Thus, the weight of the drops is effectively transferred to the vapor phase through the interphase drag. The net result is that when drops are present in the atmosphere, their effect on buoyancy forces is properly taken into account.

SSAR Revision: NONE W westinghouse 4sa342a l

l l

_ _ _ . _ _ _ _ . . _