ML20080Q191

Provides Responses to NRC Requests for Addl Info on Sbwr Design Re Topical Repts NEDE-32176P,NEDE-32177P & NEDE-32178P.Encl Withheld (ref10CFR2.790)

ML20080Q191
Person / Time
Site:	05200004
Issue date:	02/27/1995
From:	Quinn J GENERAL ELECTRIC CO.
To:	Borchardt R NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM), Office of Nuclear Reactor Regulation
Shared Package
ML19325F475	List: ML20080Q191
References
MFN-034-95, MFN-34-95, NUDOCS 9503080278
Download: ML20080Q191 (37)

Text

_ - _ _ _ _ _ _ .

o GENuclearEnergy MR B ra$s Y5C erAvenueh165 San Jose, CA 951251014 408 925-1005 (phone) 408 925-3991 (facsimile)

February 27,1995 MFN No. 034-95 Docket STN 52-004 Document Control Desk U. S. Nuclear Regulatory Commission Washington DC 20555 Attention: Richard W. Borchardt, Director  !

Standardization Project Directorate ]

Subject:

Responses to NRC Requests for Additional Information (RAIs) on the Simplified Boiling Water Reactor (SBWR) Design, RAIs 901.52-901.72 Related to Licensing Topical Reports i (NEDE-32176P, NEDE-32177P, and NEDE-32178P)  !

Reference:

1) Letter, Melinda Malloy (NRC) to Patrick W. Marriott (GE), Request for Additional Infonnation (RAI) Regarding the SBWR Design, 1 November 25,1994 (Q901.52 - Q901.74).  !

2) Letter,J. P. Klapproth (GE) to USNRC, Submittal of Licensing Topical Reports (NEDE-32176P and 32177P) for Review and Acceptance for Referencing, February 23,1993,JFK93-013, MFN-029-93.

3) Letter, P.W. Marriott (GE) to USNRC, Submittal of Licensing To alcal Report (NEDE-32178P) for Review and Acceptance for i Referencing, February 25,1993, Docket No.52-004, MFN-029-93.

4) Letter,J. P. Klapproth (GE) to USNRC, Submittal of Revised Licens,mg Topical Report (32177P) for Review and Acceptance for Referencmg, July 12,1993,JFK93-034, MFN-115-93.

This letter provides the responses to the RAls requested by Reference 1. These RAI responses are attached, and provide additional information on the subject Licensing Topical Reports (References 2,3 and 4).

I hr

~

9dO3080278 950227 4 hd/-

PDR ADOCK 0520 j  %

d pA / g#

/ dl

c GENuclearEnergy MFN No. 034-95 Page 2 Please note that Figures 60.1,64.1,64.2,64.3,66.1 and the RAI 901.63 figures contained in the attachment are of the type which GE maintains in confidence and withholds from public disclosure. They have been handled and classified as proprietary to GE as indicated in the attached aflidavit. We hereby request that this mformation be withheld from public disclosure in accordance with the provisions of 10CFR2.790.

Sincerely, 1@

f k). James LMRE.and Quinn, SBWR Projects Programs Manager h

Enclosure cc: P. A. Boehnert (NRC/ACRS)

I. Catton (ACRS)

S. Q. Ninh (NRC) l J. H. Wilson (NRC) l

4 l! s General Electric Company AFFIDAVIT I, DavidJ. Robare, being duly sworn, depose and state as follows:

1 .

l l (1) I am Manager, ALMR Project Manager, General Electric Company ("GE")- l and have been delegated the function of reviewing the information described in paragraph (2) which is sought to be withheld, and have been authorized to - ,

apply for its withholding. 1 (2) The specific information (Figures 60.1, 64.1, 64.2, 64.3, 66.1 and the figures l in RAI 901.63) sought to be withheld is contained in the attachment entitled

I Submittal of Additional Information on Licensing Topical Reports (NEDE-l 32176P, NEDE 32177P and NEDE-32178P), MFN-034-95, dated February 27, l 1995. .l l

, (3) In making this application for withholding.of proprietary information.of which it is an owner, GE relies upon the exemption from disclosure set forth in the Freedom ofInformation Act ("FOIA"),5 USC Sec. 552(b)(4), and the 4

Trade Secrets Act,18 USC Sec.1905, and NRC regulations 10 CFR 9.17(a)(4),  !

2.790(a)(4), and 2.790(d)(1) for " trade secrets and commercial or financial information obtained from a person and privileged 'or confidential" (Exemption 4). The material for which exemption from disclosure is here i sought is all " confidential commercial information", and some portions also qualify under the narrower definition of " trade secret", within the meanings assigned to those terms for puraoses of FOIA Exemption 4 in, respectively, Critical Mass Eneruv Proiect v. Muclear Regulatory Commission. 975F2d871 (DC Cir. 1992), and Public Citizen Health Research Group v. FDA. ~

704F2dl280 (DC Cir.1983).

(4) Some examples of categories ofinformation which fit into the-definition of proprietary mformation are:

a. Information that discloses a process, method, or apparatus, including 3 supporting data and analyses, where prevention of its use by General

Electric's competitors without license from General Electric constitutes a competitive economic advantage over other companies;

b. Information which, - if used by a competitor, would . reduce his

, expenditure of resources or improve his competitive position in the design, manufacture, shipment,-installation, assurance of quality, or licensing of a similar product;

c. Information which reveals cost or price information, production capacities, budget levels, or commercial strategies of General Electric, its customers, or its suppliers;

. .- - . ~ .- .

J .

r d. -Information which reveals aspects of past, present, or future' General

_ Electric customer-funded development plans and programs, of potential commercial value to General Electric;

e. Information which discloses patentable subject matter for which it may be desirable to obtain patent protection.

V The information sought to be withheld is considered to be proprietary for the ,

reasons set forth in both paragraphs 4.b and 4.d, above.

(5).The information sought to be withheld is being submitted. to NRC in-i confidence. The information is of a sort customarily held in confidence by GE, and is in fact so held. The information sought to be withheld has, to the -

best of my knowledge and belief, consistently been held in' confidence by GE, .

no public disclosure has been made, and it is not available in public sources.

All disclosures-to third parties-including any required' transmittals toLNRC, have been made,:or must be made,. pursuant to regulatory provisions or -

proprietan a

' confidence. greements Its ' initial which provideasforproprietag designation maintenance of the information information, and the -in-subsequent steps taken to arevent its unauthorized disclosure, are as set forth-in paragraphs (6) and (7) Pollowing.

.] ]

i (6) Initial approval of proprietag treatment' of a' document is made by. the-manager of the ' ongmating component, the ~ person most likely- to :be . 1 acc uamted with the value and sensitivity of the information in relation to industy knowledge. ~ Access to such documents within GE is limited on a "need to know" basis.

J,

. (7) The procedure for approval of external release of such a document typically

}

requires review by the staff manager, project manager, principal scientist or other equivalent authority, by the manager of the cognizant marketing i function (or his delegate), and by the Legal Operation, for technical content,

competitive effect, and deternunation of the accuracy of the proprietary designation. Disclosures outside GE are~ limited to~ regulatog bodies,

customers, and potential customers, and their agents, suppliers, and licensees, and others with a legitimate need for the information, and then only in accordance with appropriate regulatory provisions or proprietary agreements.-

i (8) The information identified in paragraph (2), above,is classified as proprietary i

because it would provide other parties, . includin with information related to General Electric fuel designs,g . competitors, analysis results a potential commercial which were developed at a considerable expense to General Electric.

1 (9) Public disclosure of the information sought to be withheld is likely to cause substantial harm to GE's competitive position and foreclose or reduce the availability of profit-making opportunities. The information is part of GE's

comprehensive BWR technology base, and its commercial value extends -

beyond the original development cost. The value of the technology base goes beyond the extensive physical database and analytical methodology and

,. 2-

- )

4 includes development of .the expertise to determine and apply the ~'

i appropriate evaluation process.

4

)

The research, development, engineering, and analytical costs compnse a  :

substantial investment of time and money by GE. -l 4 The precise value.of the expertise to devise an evaluation process and aaply 4

the correct analytical methodology is difIlcult to quantify, but it clearly is substantial.

GE's competitive advantage will be lost if its com etitors are able to use the l

results of the GE experience to normalize or veri their own process or if they l are able to claim an equivalent understanding by emonstraung that they can l' arnve at the same or similar conclusions.

The value of this infonnation to GE would be lost if the information were

disclosed to the public. Making such information available to competitors' without their having been rec utred to undertake a similar expenditure of-

! resources would unfairly provic.e competitors with a windfall, and deprive GE

! of the opportunity to exercise its competitive advantage to seek an adequate return on its large investment in developing these very valuable designs.

I i

4 i

u j l o

STATE OF CALIFORNIA )ss:

COUNTY OF SANTA CLARA )

l DavidJ. Robare, being duly sworn, deposes and says:

That he has read the foregoing affidavit and the matters stated therein are  ;

true and correct to the best of his knowledge, Executed at SanJose, California, this V day of flanWrn ,19}5 I bd DavidJ. Robare l General Electric Company Subscribed and sworn before me this OYday of K kw y .1936 OdY &[tt. alto Notary Public, State of Califorifia -

I l

pat;iA F. HUS$EY s COMM.d!CM6120 7 g Notcry h.t4c - Californ:o ~$

B SANTA CLARA COUNTY y

{ My Comm. Ex:*es DEC 1.1998 g 4

y . - .- .-. _ . . .- -

i w

RAI No. 901.52 Question:

l NEDE-32177P, "TRACG Qualification Report," uses the word " qualification" without any specific definition. The commonly used terms " verification" and " validation" are defined ,

i and it appears that the term " qualification." as used in this report, is essentially identical

, to " validation." If this is the case, the report should so state. If not, then explain what constitutes " qualification" and how it differs from " validation."

i GE Response:

The term " Computer Code Qualification" has been in use at GE for many years and incorporates the process of validation of the code against data or other engineering j calculations. This is part of the process of " qualifying" the code for design application.

j In the context of the Staff review, " qualification" is equivalent to " validation". This clarification will be provisd in the revised report.  ;

1 r

4 1 s I

f

A ,

s 4

} RAI No. 901.53  ;

l Question:

In Figure 5-3 of NEDE-32178P, " Application of TRACG Model to SBWR Licensing Safety l j Analysis," the containment model is apparently significantly simplified, when compared

to the model of Figure 41. Some of these simplifications may be well justified in j evaluations of reactor vessel transients. However, the lack of passive containment cooling l system (PCCS) models were not anticipated, since an evaluation of the long-term cooling i period to 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> can only be achieved if the PCCS, which is the only remaining heat j l

sink, is included. (Actually, PCCS performance was also ranked high in the GE phenomena identification and ranking tables for large breaks and for small breaks.)

t l GE Response: i i

l The model shown in Figure 5-3 of NEDE-32178P is used for the short term evaluation of j the post-LOCA transient with emphasis on the RPV water level. The primary feedback is i j from containment pressure. During the early part of the event the bulk of the blowdown  !

flow is through the horizontal vents. As such, it is expected that the results will not be affected by the PCCS. (The PCCS is not ranked highly in the LOCA/ECCS PIRT for this i reason.) The long-term cooling period is evaluated with the more detailed containment j model shown :n Figure 41 of NEDE-32178P. However, the PCCS (and ICS) have now 4

been included in the short-term LOCA model. The modified model will be documented e in Revision 1 of NEDE-32178P.  ;

j  !

4 .

1 1

1 4

j i

i 1

i I '

9 RAI No. 901.54 Question:

Early and Rapid Drvwell Atmosphere Mixing. Early in blowdown transients, rapid mixing of the original drywell atmosphere with incoming reactor coolant takes place. Section 2N3 of the Mark III containment report (NEDO-20533) outlines a classical solution of the mass and energy consenation equations for gas mixtures, including an iterative procedure, to determine at each time step, whether the gas mixture is one of inert gas and superheated vapor or one ofinter gas, saturated vapor, and liquid.

NEDE-32176P, "TRACG Model Description," does not describe any such model for gas l mixtures. .A mass consenation equation for noncondensibles is given, however no inclusion of the corresponding terms in the energy equations is indicated, nor is~ a. .

solution procedure outlined on how to assess the proper state of the gas mixture in the ,

drywell. (See Equations 3.1-10, 6.2-20, or 6.3-18 and Section 6.3.3.1 of NEDE 32176P.)

Such modeling may not be important when very small fractions of noncondensibles are present in specific components, but the dqwell undergoes a change from all inert gas to l virtually all vapor and then back to a mixture, all within a relatively short time.

i Page 6-11 mentions that Dalton's law applies but no procedure for state evaluation is l outlined. Also on page 6-11, it is stated " air" (N, most likely, OK) and stem are " assumed l to be perfectly mixed," which is not necessarily conservative (see RAI 901.55).

l l GE Response:

(See following pages.)

l

l' I

l l

n i

l. RAI 901.54

! TRACG solves the conservation equations for the liquid and gas phases given by Equations 1 3.1-1.-3.1-7. r l

Vanor mass:

1 .

f (aQv) = - V * (aQvVv) + Tg+Mmix,v (3.1-1)

Mixture mass: i f'(1-a)pg+agy

= - V - [(1 - a)pg gY~ + agvVv]

(3.1-2) l + M,ix,,

Noncondensible gas mass:

(G0a) = - V * (a0aYv) + Mmix,a (3.1-3) l Woor momentum: i I

0 Yv + k g(Vy - Vg) = - Vy V7, - OCyd y

" Y(Yv - Y t)

(3.1-4) l _ _LVp _ fw pv Pv fev +8+g mix,v upy l

Liauid momentum:

! l I 8

--V oc a Mc at g + k (1 - a)pg at[p _ yv) , _ pt . ypt _ (1 - a)et

- - (3.1-5)

Vd

• V(Ye - Yv) - VP-h+g_'" + g + 5,gx,e Woor energy-0 v + P U=-V- v

( agv(e + 2 )j ( agvVv(e + 2 )j at at (3.1-6) i l - V - (PaVv) + qwy + qiy +Phgs+Emix,v l

2

L.

p I

l o

. Liouid energv: ,

f (1 - a)pg(eg + - P h = -. V - (1 - a)pg Ve(eg + -(3,l,7)

- V - (P(1 - a)V g) + q,e + qig - T gt h + Emix,g The gas phase is assumed to be a mixture of steam and non-condensible gas.' q l

Ov = Os + On )

The non-condensible gas is assumed to be a perfect gas with a gas constant specified by the user, .

the non-condensible mixture can thus be air, Nitrogen or any other non-condensible gas. ; In' each

- node, the steam and the non-condensible gas are assumed to be perfectly mixed, i.e. they have the same velocity and the same temperature. Different inodes, however, can t have different concentrations of steam and non-condensible. The calculation of the distribution of steam and non-condensible is thus controlled by the nodalization of the different containment volumes. The ,

mixture energy is given by:

Ov e v = . 0s es + Qa ca TRACG solves the equations for conservation of mass for the steam and non-'condensible mixture (Eq. 3.1-1), the total mixture (Eq. 3.1-2) and the non-condensible. gas (Eq'. 3.1-3). . The mass conservation for the steam is obtained by subtracting Equation 3.1-3 from Equation 3.1-1, and the mass conservation for the liquid is obtained by subtracting Equation 3.1-1 from Equation 3.1-2.

TRACG solves the momentum equations for the steam and non-condensible mixture

! (Eq. 3.1-4) and for the liquid (Eq. 3.1-5). The steam and the non-condensible in a node are--

assumed to have the same velocity.

TRACG solves the equations for conservation of energy for the steam and non-condensible mixture (Eq. 3.1-6) and for the liquid (Eq. 3.1-7). The steam and non-condensible in a node ~am assumed to have the same temperature.

l Neglecting the momentum equations, TRACG solves the equations for conservation of the l' steam and non-condensible gas mass, the total mass, the non-condensible gas mass, the steam and non-condensible gas energy, and the liquid energy. This gives five equations for the calculation of the five primary dependent variables in TRACG:

P , a , Ty , Te and Pa

l. 3 L .~ . ._ _.- _ - . _ . _

i l

The partial pressure.of the steam is given by Dalton's Law:

P = Ps + Pa l j TRACG thus has five equations with five unknowns, and the solution is obtained through a simple Newton-Ralphson iteration, which is described in detail in Section 6 of the TR ACG Model .-

Description.

i i

. Note, that TRACG does not rely on the assumption of thennal equilibrium for the mixture.-  ;

1 i

4 i

l 1

l l

i l

l' f

l l 4 i

l RAI No. 901.55 Question:

Deeree of Early Drvwell Atmosphere Mixine. Section 3 of the Mark III containment report (NEDE-20533) points out that perfectly mixed assumption can result in lower peak pressures than would be observed with partial mixing. There does not appear to be any procedure for a best-estimate evaluation of the degree of mixing or a bounding assumption to account for partial mixing in NEDE-32176P, "TRACG Model Description."

l (See also RAI 901.54 above and page 6-11 of NEDE-32176P.) '

GE Response:

TRACG does not force perfect mixing within the drywell. Perfect mixmg of the-constituent gases is only assumed within each cell. As the break flow discharges into the drywell, the noncondensibles in the cells near the PCCS inlet and the pipes feeding the-horizontal vents will be first purged into the wetwell. Because of the large volume of steam discharging from the break, the remaining noncondensibles in the upper drywell-will be mixed with the steam and also purged rapidly into the wetwell. However, the calculations show that noncondensibles remain in the lower drywell and are continuously discharged into the upper drywell, and through the PCCS to the wetwell, over a long time into the transient. In the SBWR calculation, the short term pressure increase is not limiting. Thus, the slow discharge of noncondensibles form the drywell serves to limit the PCCS heat removal and increase the long term pressure. Early purging of all noncondensibles, on the other hand, would result in pure steam condensation in the PCCS, and reduction in drywell pressure. This would tend to open the vacuum breakers and bring noncondensible back into the drywell, which have to be recycled through the PCCS to the wetwell. It is not obvious which of these two ' scenarios (i.e. continuous bleeding of noncondensibles or periodic purging) would lead to a higher calculated long term pressure. Sensitivity studies will be made to bound this effect.

l 1

_ , - . . - . - - . . . .- - ._ . ~ , . _ -

~.  ;

1 RAI No.' 901.56 1

Question: l Horizontal Vent Clearine. Section 4 of the Mark III containment report (NEDE-205'33) presents an elaborate model of the early downward acceleration of the liquid in an annulus, correspon' ding to the eight individual vertical vent pipes in the SBWR design. l The model includes tracking of the water level in the vertical vents, to initiate drywell to-  !

suppression pool gas flow, as the horizontal vents clear.'

-]

It appears that " Tee" sections are used in the TRACG model, but NEDE-32176P, "TRACG Model Description," does not describe any level tracking capability in components like pipes or Tees.

i I

'Section 6.2.2, last paragraph, of NEDE-32178P, " Application of TRACG Model to SBWR j Licensing Safety Analysis," states: " Improved input modeling of the loss-of-coolant j l accident vents made it possible for TRACG to predict vent clearing." It is assumed that -

this refers to the horizontal vents. However, neither of the_ reports provides any .

~

f information regarding these models, nor is it clear how one can obtain a level tracking l capability via " input modeling." )

l l Note the validation of the Mark III model predictions against PSTF data. Have similar ~ l l

comparisons been made with proposed TRACG models? ? -l l The last paragraph of Section 6.2.2 of NEDE-32178P states: "It should be noted...'that l M3CPT is still used for prediction of the design vent clearing and dynamic loads..." It is assumed that the dynamic loads mentioned here are the transient forces addressed in Appendix 6A of the SBWR Standard Safety Analysis Report (SSAR). However, what is meant by " design vent clearing?" Does this mean that TRACG gets some input from MSCPT simulations as to when the vents clear?

GE Response:

]

l J l The reference to " improved input modeling" in NEDE-32178P referred to the detail of l the vent input model. The earlier (SSAR) model included only the tope horizontal vent.

l The improvement was to include all three rmvs of horizontal vents.

TRACG predictions will be compared with PSTF data as part of the supplemental qualification effort now in progress.

The reference to " design vent clearing" was to peak drywell pressure which occurs in the first few seconds of the blowdown as the vents clear. The TRACG input model does not utilize any data from M3CPT calculations.

. RAI No. 901.57 Question:

Early Peak Pressure (SBWR SSAR. Ficure 6.2-13). Based on initial evaluations and results represented by others (Arai & Nagasaka,1991), we would expect an early peak pressure

~,

of about 2.5-3.5 bar within the first 3-10 seconds and maybe without significant pressure l decrease thereafter. The M3CPT results shown in Figure 6.2-13 of the SSAR differ 1 significantly from this expectation, shown a peak drywell pressure of 3.5 bar only at 36 )

seconds. Comment on the early pressure history of this scenario. j i i GE Response-4 There are characteristically two pressurization events in the dnwell following the LOCA.

The first is associated with the initial clearing of the main vents. This pressure peak typically occurs in the 3-10 second range. Evidence of the vent clearing peak can be seen within this time period by close examination of Figure 6.2-13 of the SSAR. The attached  ;

Figure 57.1 shows an expanded view of the typical short-term response. The magnitude 3

of the peak is approximately 3 bar. The second pressurization event is caused by the combination of suppression pool heatup and the purging of nitrogen from the dqwell to j

the wetwell gas space. Figure 57.1 shows this second peak occurring at about 36 seconds i 1

with a magnitude of 3.5 bar.

i 1

4 I

,1 i

4

't 4

4 I

4 d

J

a_

6 .

T P

C _

3 M

E S

N O

P S

E R

_ E F R R

- U G.

S S m G R

E T_

~

R 1 S

E

_ P 7 R

P N 5 W e r

O u D i g

W F O  %

L B

R _

W -.

B S _

m _

0 0 3 0 0 0 0 _

0 5 G 5 5 0 5 .

4 3 3 2 1 1

_ S _

=- .

M _

i

. 1 i

t-RAI No. 901.58 4

l Question:

l

Vent Flow. The model outlined in Section 5 of the Mark III containment report NEDE-20533) uses quasi-steady compressible adiabatic flow with several simplifying assumptions.

The proposed TRACG model applies three Tees for these vents. From NEDE-32176P, a "TRACG Model Description,

it is not clear whether the horizontal portion of the Tees j can handle an equivalent flow model. Provide sufficient details on the Tee and pipe .

f models so the availability of an adequate vent flow model can be ascertained. The lowest l Tee should actually be an elbow. Is an elbow a special case of a Tee?

GE Response

I The conservation equations and constitutive relations described in Sections 3.1 and 3.2 of

i. NEDE-32176P are, in general, utilized for all TRACG components. To assist in the  !

interpretation of the conservation equations for PIPE and TEE components, reduction of l

the conservation equations to one-dimensional form has been shown in Equations 3.1-8 j through 3.1-14. The combination of these equations with the appropriate constitutive 1 f

relations from Section 3.2 yields an adequate description of blowdown flow through the ,

l vents. I i

The vertical sections of the SBWR vents extend a small distance below the bottom l l tangent of the lowest horizontal vent to provide support for the vertical to horizontal l pipe weld. However, the comment is correct in noting that the geometry of the bottom -  ;

, section is closer to an elbow than a tee. A TEE component was used in the TRACG l model because is provides a clearer representation of the flow area change between the

vertical and horizontal sections.

1 l

4 I

1

)

1

j I

RAI No. 901.59 1

) ' Question:

t Vent Back Pressure. The' Mark III containment report (NEDE-20533) provides a model

for gas bubbles expanding into the suppression pool, with reducing bubble pressure and i

pool level swell. The resulting back pressure will affect the horizontal vent flow i significantly. Similar modeling capabilities are not described in NEDE-32176P, "TIU,CG

Model Description."

Note also the validation of the Mark III model predictions against PSTF data. Have.

5 similar comparisons be made with proposed TRACG models? .

i GE Response:

I .

p The Mark III containment report (NEDE-20533) provides a model for calculating vent .

i back pressure. This pressure is different from than the hydrostatic pressure at the vent i elevation because the fluid above the vent must be accelerated to allow for the expansion i of the gas bubble. In the referenced model, the gas bubble is assumed to occupy the -

j entire pool cross section, and remain distinct from the pool liquid which is accelerated as j j a slug. This provides for a conservative treatment of the back pressure. l i

j TRACG does not have an explicit model for this phenomenon. . In: the -_TRACG l calculation, the gas discharged through the vent will be mixed with the liquid in the cell j i in which it is injected. The resultant expansion of fluid in the cell will accelerate the

[

1 fluid in the cells above, and the cell pressure will reflect this effect. Thus, ' the phenomenon is modeled with the TRACG conservation equations. The accuracy of the -

j simulation depends on the number of cells to characterize the pool. .This is being

validated through comparisons with selected Mark Ill data as part of TRACG validation

for short-term containment response, j It is important to keep in mind the differences between the Mark III containment and the SBWR containment with respect to the relative importance of different phenomena. I

! In the Mark III containment, the wetwell airspace is very large and there is essentially no 1

pressurization of the wetwell. Thus, the maximum drywell pressure is set by the timing of vent clearing. For the SBWR containment, the vent clearing and establishment of vent i flow reduces the drywell pressure for a short period of time, and the pressure again j increases as the wetwell discharged with the noncondensibles initially in the drywell. The peak pressure in the blowdown period is tims set by the ratio of the volumes of the drywell to the wetwell., and the steam partial pressure corresponding to the' pool surface j temperature. It is independent of the details of the vent clearing process.

I-

)

l i

L a

e q w ~ -

,-r-w., 0v -

e-+ - "sege

l 1

RAI 901.59 Response (Continued) l l

Even for the Mark III containment, the vent back pressure does not affect the peak

drywell pressure, because that occurs prior to vent clearing. The back pressure affects

the pressure response immediately after vent clearing (NEDO-20553, Appendix A). This phenomenon will have only a minor effect on the SBWR pressure response.

i 1

i 1

I

+

1 1

l l

I E

1

.i l

I b

J 4

d d

1 j RAI No. 901.60 -

Question:

Suppression Pool Condensation. The Mark III containment report (NEDE-20533) uses a very simple model for condensation in the suppression pool (all incoming water vapor condenses, the poolis at one average temperature).

4

{ Page 6.2-5 of the SBWR SSAR, last paragraph, states: " Performance of the pressure i suppression concept in condensing steam...has been demonstrated...as described in j Appendix 6A." No such description appears in SSAR Appendix 6A., except for one sentence on page 6A-2, third paragraph, that' states: "Overall, prior test data have 1 indicated that the steam is condensed in the horizontal vent exit region." Describe the

! intended TRACG models and if the same assumption (all condenses) is to be made, e

provide the appropriate references.

] The last sentence of Section 6.2.2 of NEDE-32178P, " Application of TRACG_Model to

SBWR Licensing Safety Analysis," refers to a "special procedure" to conservatively acco mt for thermal stratification in the pool. Provide a description of this proceduie.

GE Response:

l No "ad-hoc" assumption is made regarding steam condensation in the suppression pool.

l The process is governed by the interfacial heat transfer equations described in Paragraph 3.2.9.2 of NEDE-32176P. It is recognized that the single sentence in Appendix 6A is less than one would expect from the reference on page 6.2-5 of the SSAR. However, it is true that none of the blowdown testing done by GE had indicated anything other than

complete condensation of the steam within the pool.

l The procedure used for the treatment of suppression pool stratification was previously described in the response to RAI 901.49. This procedure was also described in a presentation made at the ACRS meeting in Los Angeles on December 16,1994. Attached

Figure 60,1, presented at that meeting, compares predictions of pool surface temperature i using the SBWR stratification procedure with test data from Mark II and ABWR i blowdown tests. These results support the use of the stratification procedure for use in j SBWR design calculations.

a a

3

RAI No. 901.61 Question:

Wetwell Pressure and Temnerature. The question of mixed gas atinospheres was raised with the drywell atmosphere. It is assumed that once that has been clarified, there should be no problem in applying the same model, maybe with slightly differing assumptions, to the wetwell region. Confirm or, if not so, provide alternative details.

GE Response:

The discussion of drywell modeling of noncondensibie gases (901.54) applies also to the

wetwell vapor space. The modeling, at the basic level, is ide;,dcal. The drywell and wetwell are both represented by cells which are part of the .3-dimensional VESSEL j component.

4 s

i i

4 4

J

l j .

1  ;

i- RAI No. 901.62 i

!- .i

Questiom j Transient Wetwell Pressurization. As ' outlined in Section 9 of the Mark III containment - )

i report (NEDE-20533), initial pool level swell can cause a transient pressure rise. Since _i the SBWR wetwell is a -(nearly) closed cavity, this pressure rise may' be even more l

pronounced. Describe the intended TRACG models and provide a comparison of their  !

l response to the PSTF data, l GE Response:

j TRACG does not have an explicit model for a multidimensional representation of the a initial air bubble and its break through the pool surface. The pool will be in the bubbly ..

} flow regime, and interfacial shear correlations which are applicable for this flow regime (see Section 3.2.2.1 of NEDE-32176P) are used to calculate the relative velocity of the

vapor bubble with respect to the liquid. The movement of the two-phase level is calculated in TRACG by the level tracking model described in Section 3.2.7 in NEDE-32176P. The pressurization of the vapor space is calculated by the TRACG conservation equations, which are generalized for twophase flow, and treat compressibility effects.

Currently, GE is in the process of comparing results of TRACG calculations with a

number of PSTF tests. The most suitable tests for looking at wetwell pressurization are-l the Mark II tests with a confined airspace. Results of these calculations will be provided when available in the form of a PreliminaryValidation Report.

i j Please note that TRACG is uni being used to calculate pool swell (air clearing) loads.

The previously approved methodology is applied for this calculation.

4 I

i 4

d

}

l 4

i 2

3

l . -

i RAI 901.63: An important parameter strongly affecting the containment building i pressures is the amount of energy absorbed by walls and solid structures, j including condensation on these surfaces. The heat transfer process will be a governed by the convective heat transfer coefficient with condensation in i the presence of noncondensables. To establish the heat transfer rates that i can be expected under such conditions, GE has sponsored recent - )

, experiments at the Massachusetts Institute of Technology (MIT) (Dehbi et j al,1991). Neither NEDE-32176P, "TRACG Model Description," nor

NEDE-32177P, TRACG Qualification Report," give any information about

! aux heat transfer coefficient correlations which would apply for containment transients. Provide reference and/or details of the

correlatioins currently in use for this purpose in TRACG.

! Reply: The heat transfer correlation for condensation in the presence of

noncondensables is described in Ref. (1), Section 3.2.10.5. TRACG does 1

l not have an explicit containment component. The containment modelis .  !

composed of vessel, heat slab and other components which are al1  !

- described in Ref(1). The correlation described in Section 3.2.10.5 of Ref. . )

,j (1) is used for all steam condensation in the presence of nitrogen. Ref. (2),'  ;

Ref. (3) and RAI 901.68 contain more information about the Vierow-l Schrock correlation,' on which the TRACG correlation is based. As j described in RAI 901.68, fi accounts for the degradation of heat transfer due to the presence of noncondensables and f2 accounts for the . i

! enhancement of heat transfer due to interfacial shear. (In Ref. (1) the '!

, definitions of fi and f2 are interchanged). In the TRACG calculations for  !

the containment, the shear enhancement, fi, will be very close to unity due j j to the low velocities. The only relevant component of the correlation then l is the f2 factor. Figure I shows a plot of the f2 factors from the TRACG -  !

] correlation and the K-S-P (Kuhn) correlation. As described in RAI 901.68, i the Kuhn data set is the latest data set from U.C. Berkeley. The figure ;

j. shows that the TRACG correlation predicts more heat transfer degradation -

. from the presence of noncondensables.

I Figure 2 shows a comparison of the Dehbi correlation [Ref(4)] and the -

I TRACG correlation for conditions typical of the SBWR drywell 20 minutes after a LOCA. The figure shows that the TRACG correlation predicts less j heat transfer than the Dehbi correlation for typical drywell conditions. This -

l is conservative from the standpoint of containment performance because

! less steam will be condensed in the drywell in the early stages and because .

heat transfer from the drywell to the wetwell (which may be a concern in i the long term) will be dominated by the thermal resistanc'e of the'

conduction through concrete, not by the condensation process.

3 Much of the steam condensation that occurs in the drywell occurs during the blowdown period. In this period the fluid flow conditions are such that t

4 4

i

t i

RAI No. 901.64 Question: ,

, The bottom drain line break scenario chosen in NEDE-32177P, "TRACG Qualification Report," is GIST Test A07. Figure 5.S12 of NEDE-32177P shows the gravity-driven - ,

i cooling system (GDCS) flow rate. The test says onset of the GDCS flow for TRACG and Test A07 occurred within 12 seconds (see page 5-76). The figure shows a difference of- -

about 70 seconds. GEFR 00850, Figure 4.S54 for the same event appears to give quite different results and, with some scatter of the data, the onset of GDCS flow is not clearly identifiable. Explain these discrepancies.

! GE Response:

, The GIST facility contains several GDCS lines. In GEFR 00850 Figure 4.S5, a comparison is made for one of the GDCS lines, while in the TRACG Qualification report Figure 5.S 12 the comparison is made for the total GDCS. The attached figures show the GDCS flow -

i in line A, which is used ir. GEFR 00850, all the lines, and the total flow which is used in the TRACG Qualification report. It should also b no:-d, tint the calculation in GEFR 00850 was made with an earlier version of TRACG. ,

The statement about the timing difTerence for the onset of GDCS flow contains a -

typographical error, it should read approximately 72 seconds and not 12 seconds.

1 The GIST calculations have been repeated since the calculations reported in the TRACG

Qualification report. In these calculations, deficiencies in the TRACG input decks were corrected. The agreement with the data improved slightly and all major conclusions concerning the comparison remained unchanged.

4 i

4 a

i l

i RAI No. 901.65 Question:

GEFR 00850, Section 4, refers to a revised two-phase level model for TRACG (page 4-2).

NEDE-32176P, "TRACG Model Description," page 3-15, introduces a revised two-phase ]

level model and refers discussion of the modification to Section 3.3.7.3. There is no l Section 3.2.7.3. The summary of code modifications, given in Appendix A of NEDE-32176P also does not mention a revised two-phase level model Provide the description.  !

GE Response:

Section 3.3.7.3 was inadvertently left out of the TRACG Model description, the missing text has been supplied through the response to RAI 901.11bb.

The statement "rgvised two-phase level model" refers to the fact, that earlier GE-versions of TRAC did not properly account for the void distribution below and above the two-phase level in the evaluation of the constitutive correlations.

1 i

I l

l

i RAI No. 901.66 Question: .

I i The TRACG drywell nodalization used in GEFR 00850 to simulate GIST (Figure 4.1-2) is l l much finer than the nodalization of the.SBWR drywell (Figure ~4.1-5). Explain this 4

difference.

1 j GE Response:

The TRACG nodalization used for GIST had 16 axiallevels. This was done to capture any 1

thermal stratification in the fluid that might have occurred in the test due; to the ,

relatively small cross-sectional areas and also to accommodate break flow from the GDCS

} line break and the bottom drain line break. In contrast, the current SBWR TRACG deck ,

has 9 axial levels. Before the noding of the SBWR is finalized, comparisons will be made  ;

1 with all the SBWR test facilities (GIST, GIRAFFE and PANDA) with comparable noding ]

] as the plant. Any differences necessary because of facility unique characteristics will be j noted andjustified.  !

) It should be noted that the containment nodalization for the SBWR has evolved since ,

i NEDE-32178P was submitted. The current nodalization is attached as Figure 66.1. The i major differences as compared to the previous nodalization are the addition of a second'

] PCCS heat exchanger component and the representation of the IC/PCC pool.' A second

vacuum breaker component has also been added. None of these changes have resulted in an increase in the predicted peak pressure following the design-limiting LOCA. I 1 l

]

i i

i d

1 4

~-

.%. . , -..m., ,

i

' l l ,

l RAI No. 901.67 Question:

The TRACG model apparently models the active section of each PCCS module as a single pipe, with interior and exterior heat transfer, as implied by NEDE-32178P, " Application of TRACG Model to SBWR Licensing Safety Analysis," Figure 41. Confirm this.

I GE Response: j l

The 496 tubes in each PCC unit are collapsed into a single PIPE component in the l

TRACG model. The PIPE component includes inside and outside heat transfer and wall heat conduction.

l I

4 l

l l

i i

1

'. RAI 901.68

Since PCCS condensation will be strongly affected by the presence of -

noncondensable gases and in view of the GE-sponsored extensive -

p experiments at the University of California at Berkeley and at MIT, it is

, expected that any correlations to be used in TRACG will consider the '

. results of these studies. Some of the correlations developed under these

programs appear to overpredict the heat transfer, in particular, in the ini

t -

i region of the PCCS tube bundle. Provide a description andjustification of j-- the correlations to be used.

- y

[ Reply: As described in Ref. (1) five experiments using four different test rigs were ]

conducted. The first experiment was.the Vierow experiment at UCB JA )

) modified version of the Vierow-Schrock correlation is used in TRACG for 'I

[ the PCCS condensation and is described in more detail in Ref. (1).' Like ;

! correlations used previously in containment codes, the.Vierow-Schrock' .)

{ correlation is in the form of a correction factor, f, to a reference heat - l j transfer coefficient called the "Nusselt" heat transfer coefficient.-  !

I I

.. I

h,z a = f Ibr  ;

2-

. The correction factor is divided into two parts such that f = fif2.L The factor - ,

j fi si the correction factor which accounts for the enhancement of heat. l transfer due to interfacial shear, interfacial waves and deviations from the -

Nusselt model. ;The factor f2 is the correction factor which ~ accounts for

the effects of heat transfer degradation'due to the noncondensable gas mass' j fraction.

$ Subsequent experimental data sets (Kuhn at UCB and Siddique at MIT)

{ ~ have shown that the Vierow-Schrock correlation is accurate over the range i of data in the Vierow experiment. Ths Kuhn experiment is vie ved by GE j to be the best database by virtue ofits high reproducibility, low data scatter l and large range of experimental conditions. Figure 1 (Figure 7-1 of Ref.

! (2)) is a plot of heat transfer coefficients predicted by both the Vierew- _

1-d Schrock and the K-S-P (Kuhn-Schrock-Peterson) correlations vs. the experimentally determined heat transfer coefficient of the Vierow 3

experiment. While the scatter in the Vierow-Schrock correlation is larger, I j the two correlations predict very closely the same values.

j Initially, the Siddique results (3) seemed to show much higher heat transfer coefficients at the inlet of the tube. As explained in Ref. (1) and Ref. (2),

j; the data at the axial location of 10 cm from the top have large uncertainties  !

. associated with them. When this data are excluded, the large differences in i heat transfer are eliminated. The K-S-P correlation predicts well Siddique's j experimental data for distances greater than 10 cm from the top.

I ,

i -

1 l

i i a

5 J

w r e. + e - , ,.,.-..e --v. , ,..e . - ,--.7m

.a As described in Ref. (1), page 2-11, the Ogg experiment (4) (the second of three at UCB) experienced some mechanical problems and lacked reproducibility. The results were never published. For these reasons, the Ogg results are not be considered in this RAI (Ogg's results are similar in :

magnitude to the other results).

The Hasanein experiment at MIT (5) focused on steam-helium 'and steam-helium-air tests and also on investigating the coolant side. An additional .

steam-air correlation was not developed. The steam-helium results compare well with the Siddique air-helium correlation. i GE believes that the Vierow data are confirmed by the subsequent >

experiments at UCB and MIT. However, it is also clear that the Vierow -

Schrock correlation overpredicts heat transfer when the mixture Reynolds.- -

number exceeds the experimental Reynolds numbers tested in the Vierow-experiment. The mixture Reynolds number is a parameter in the if factor in .

the Vierow-Schrock correlation: .;

4 fi = 1 + 2.88x10 Re.2 "

TRACG limits the fi factor to f i5; 3. Even with this limitation, the -

modified Vierow-Schrock correlation.overpredicts heat transfer (compared -

to the K-S-P correlation) at high Reynolds numbers.L fii s modified further for high film Reynolds numbers (not seen in the PCCS tubes). The impact of the overprediction of the Vierow-Schrock correlation is described in -

Section 4.1 ofRef. (1) and will be further evaluated by TRACG comparisons with the PANTHERS PCC test data. In fact, the impact is likely to be less than described in Section 4.1 because the correlation in TRACG overpredicts the in-tube heat transfer coefficient only when the mixture Reynolds number is high. This occurs at the inlet of the tube where the local steam mass fraction is also highest and therefore the heat transfer coefficient is a maximum. Under these conditions the TRACG' correlation will be a very small part of the overall thermal resistance from the steam inside the tube to the pool outside. Section 4.1 states that the correlation accounts for typically 25% of the overall resistance over the whole tube length. For the conditions just described (high Re, high h) the _

tube wall thermal resistance and the outside pool resistance will dominate and the inside resistance will account for less than 25% of the total. The resulting error in using the current correlation will, therefore, be small.

REFERENCES (1) Usry, W. R., Single Tube Condensation Test Program, NEDC-32301, March i 1994.

2

(2) Kuhn, S. Z., Schrock, V. E., and Peterson, P. F, Final Report on U.C.

Berkeley Single Tube Condensation Studies, UCB-NE-4201, August 1994.

(3) Siddique, M., The Effects of Noncondensable Gases on Steam Condensation under Forced Convection Conditions, Ph.D. Dissertation, MIT,1992.

(4) Ogg, D. G., Vertical Downflow Condensation Heat Transfer in Gas-Steam .

Mixtures, M.S. Thesis, U.C. Berkeley Dept. ofNuclear Engineering,1991.

8 (5) Hasanein, H., Steam Condensation in the Presnece of Noncondensable Gases under Forced Com>ection Conditians, Ph.D. Dissertation, MIT,1994.

3-ll ll '

-a. 4Tr

!!!I - l/

l ll l +; .N '

O hm l

+

l y

o 10000-

+ h.rse w .

F 4

[Mk' . , ,

bYM "E 7 i 6 j s-l -p +p>g. '-[ , go 4-:-

I. I I

. . __ .-h.-_; 1

.9 3- --t  ?

i, *..

p q. it. -I

} 2 ..

k **

7 o g O y

>i 5m A $kY 'I r 1000_ '

m g d-t---j

-+ y y; gg- -y p . .

..-.y. . - . - ,

- o- '

>!  ! i ,

8- t;y.+/ Poi g i i 1-t- i i I i 1

/c 4 I i :i 2- -,

7

} ,

o l l,ll 1

i .;

r  :

100- - -

I i i l i s l i i l i i l i ii 100 1000 10000 h,,,w,,,, W/m^2 'K Figure 1: Heat transfer coefficients predicted by K-S-P and Vierow-Schrock Correlations vs. Vierow's Experimental Data. Taken from Ref. (2).

3

g 1 l

l . RAI No. 901.69 l l

r Question:

i No information is given on the pool-side heat transfer correlations to be used for the PCCS, where the possibility exists that the inner tubes may not get suflicient flow. Itis J l

expected that any such insufficiency in performance would be revealed during the PANTHERS test program. However, for the assessment of TRACG's capability to model containment transients, provide a description andjustification of the correlations to be used for the pool-side heat transfer.

GE Response:

\

l The heat transfer correlations used ' to . calculate secondary-side heat transfer. are described in Section 3.2.10 of NEDE-32176P. Generally speaking, the heat transfer regime is nucleate boiling and the Chen correlation is used as described in Paragraph

3.2.10.2. Thejustification for the TRACG modeling of secondary-side heat transfer will l be based on the results of comparisons between TRACG predictions and PANTHERS test i data.

l l

{

l l I i 1 l

l l

l l

l l

l l l l' I l  ;

4 i

e I

RAI No. 901.70 Question: ]

In NEDE-32176P, "TRACG Model Description," Figure 5.1-1 is described in the text to be a pipe with venturi tube and abrupt area change. No such thing is visible in Figure 5.1-1.

Is the text wrong, or the figure?

GE Response:

Figure 5.1-1 will be replaced with the attached version.

s l

i 1

d j

i

2 i

i

+

, l 4 ,

l -

4  ;

i l

i i

] JUN2 1

j _ _ _ _ _ _ _ _ ___-_____.,

7l --

ll l NCELLSq,

.(l+ [g~ g i ll *  ; I I 1

i j

!i .c f I I l i;

Ij o 9 $ i l j

o I k I!2 I s j i lI I

I I l t

l HD l= NODES i

7 < _ _ _ _ __ _ _ _ _) .a' I l

! l i

i l; o .

\

i I

i TH  % /

l 4. . .- _ _ x ... ~

3 ll -_ h_ _ _ ll l

} l DX o l lf--}/

I l i RADI( e l

l i

r h-

\\ //

\ o //

\\ //

\\ //

p I Il 1.

ll o 1:

ll

lI I il ll 13 l h ---- ----

4

// \

i

// 2 o g 2 /.,/. _____ ______M i il l ll 4 1 1-1 Il i:

0 li l l

j l Figure 5.1-1 PIPE Noding Diagram 4

.Jll.-____--____-ll.

4 i JUNI e

1 E

i a

m - ren-e, w ,

.e

• l RAI No. 901.71 Question:

The text of NEDE-32176P, "TRACG Model Description," on page 5-52 explaining Figure 5.8-4, refers to Cells 1 and 5 as being connected to the slab. As shown, it should be Cells ,

2 and 6.

GE Response:

The text in the TRACG Model Description (Page 5-52, fifth paragraph, third sentence) should read: "In this figure, the outside surface of the double-sided heat slab associated with fluid Cell 2 in actually in contact with the fluid Cell 6.

l l l l

l l

l

.s O

RAI No. 901.72 Question:

Figure 6-1 of NEDE-32178P, " Application of TRACG Model to SBWR Licensing Safety Analysis," shows a noncondensible return line from the isolation condenser (s) to the suppression pool. It is understood that this venting must be done manually and intermittently during station blackout. If this is so, inclusion in the model for design-basis accidents appears superfluous, since no credit should be taken for it (see bullet 11 of SSAR Section 6.2.1.1). Clarify.

GE Response:

No credit has been taken for the ICS in licensing design basis calculations. The ICS is included in the model to provide flexibility for a full range of design calculations but it is valved out for licensing calculations.

1 l

l

, .s

! !t e

i l

RAI No. 901.73 ,

l Question: ,

In NEDE-32178P, " Application of TRACG Model to SBWR Licensing Safety Analysis," the phrase "inside containment" should be added to the last sentence of Section 6.1 after '

"...of a main steamline" (a break outside containment would certainly not constitute a l design-basis accident).

l l GE Response:

t

This modification is accepted and will be used in future references to the limiting break.

l l l

! l l i l

i 1

i

yg*

.(.

RAI No. 901.74 Question:

In Figure 5-3 of NEDE-32178P, " Application of TRACG Model to SBWR Licensing Safety Analysis," the containment portion shows the center line symbol at the outside of the GDCS and suppression pools. The containment should be an annulus, with the center line to the right of the right border.

GE Response:

The centeiline designation in Figure 5-3 refers to the TRACG r-z model, not to the physical containment. The TRACG r-z model for the short-term LOCA analysis essentially models the reactor vessel and containment as two separate vessels. (This is:

achieved in the input by placing the containment vessel above the reactor vessel and then adjusting the elevation changes in the connecting lines to achieve the correct thermodynamic relationship between the two regions.) This approach is a valid alternative to the use of concentric annuli for a situation in which conduction heat transfer between the annuli need not be considered. It has the advantage of allowing the axial nodalization of the containment to be set independently from that for the reactor pressure vessel.

l

.