ML20086F229
| ML20086F229 | |
| Person / Time | |
|---|---|
| Site: | 05200004 |
| Issue date: | 07/11/1995 |
| From: | Quinn J GENERAL ELECTRIC CO. |
| To: | Quay T NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM), Office of Nuclear Reactor Regulation |
| References | |
| MFN-100-95, NUDOCS 9507130062 | |
| Download: ML20086F229 (17) | |
Text
_
GEKuclearEnergy YM h165 San Jose, CA 05125-1014 B
rri 75 C no Aven 408 925-1005 (phone) 408 925-3991 (facsimile)
July 11,1995 MFN 100-95 Docket STN 52-004 Document Control Desk U. S. Nuclear Regulatory Commission Washington DC 20555 I
Attention: Theodore E. Quay, Director Standardization Project Directorate
Subject:
SBWR - Responses to Open Items from the March 8 and 9,1995, Meeting.
Reference:
1.
Letter from S. Q. Ninh (NRC) to GE Nuclear Energy (GE), MEETING
SUMMARY
OF MARCH 8 and 9,1995, dated March 28,1995.
The attachment to this letter responses to Open Items noted in the Reference letter. For convenience in managing closure of the Open Items, GE has put the Open Items and the Responses into Open Item / Response format similar to NRC Rt. quests for Additional Information (RAIs).
Sincerely l
1 l
Jamef[Quinn, Projects Manager fR and SBWR Programs
Attachment:
Responses to Items From March 28,1995, NRC/GE Letter.
i-cc:
P. A. Boehnert (NRC/ACRS) (2 paper copies w/att. plus E-Mail w/att.)
I. Catton (ACRS)
(1 paper copy w/att. plus E-Mail w/att.)
S. Q. Ninh (NRC)
(2 paper copies w/att. plus E-Mail w/att.)
J. H. Wilson (NRC)
(1 paper copy w/att. plus E-Mail w/att.)
1 30GW po 9507130062 950711 PDR ADOCK 05200004 A
PDR lg
GENuclearEnergy j
~
MFN 100-95 bec: (E-Mail w/att. except as noted)
J. A. Beard P. F. Billig R. H. Buchholz i
T. Cook (doe)
(2 paper copies w/att, plus E-Mail w/att.)
l J. D. Duncan R. T. Fernand+z (EPRI)
J. R. Fitch J. E. Leatherman J. E. Quinn T. J. Mulford (EPRI)
(2 paper copies w/att. plus E-Mail w/att.)
P. E. Novak F. A. Ross (doe)
K. T. Schaefer B. Shiralkar i
R. Srinivasan (EPRI)
J. E. Torbeck l
GE Master File M/C 747 (1 paper copy w/att. plus E-Mail w/att.)
j SBWR Project File (1 paper copy w/att. plus E-Mail w/att.)
l l
i 1
L
. Attachment to MFN No.100-95 Responses to Items From March 28,1995, NRC/GE Letter GIRAFFE / Helium Test Conditions and Basis l
01328.01 Address the potential for condensation in the drywell to PCC line and the effect on the l
accuracy of non-condensable gas measurements in the GIRAFFE / Helium Tests.
l Discuss the effect of non-condensables on the flow measurement instrumentation on l
this line. Discuss the type of instrument to be used and it's sensitivity to the density of the mixture being measured. Discuss how the addition of non<ondensables will affect l
the density of the mixture being measured in the GIRAFFE / Helium Tests and how it l
will not result in an inaccurate measurement of the flow.
RESPONSE TO OI328.01 Originally, GE did consider collecting a nonsondensable gas sample from the PCC steam supply line, later GE decided to instead collect the gas sample from the upper drywell. Therefore, the first part of this open item is no longer applicable.
A venturi flow meter is installed near the inlet to the PCC, which is used to measure the inlet flow to the PCC.
In addition, near the PCC inlet there is a pressure transducer used to measure the total fluid pressure, and near the drywell outlet there is a thermocouple used to measure the fluid temperature.
By using the measured temperature and pressure and assuming satisfated conditions the partial pressures of steam and non-condensable gases can be determined.
The steam density can be referenced from a steam table at the pressure corresponding to the saturated temperature. For tests H1, H2, T1 and T2, where there is only one type of non-condensable gas used, the steam quality and the non-condensable gas quality can be calculated based on the previously determined partial pressures. For tests H3 and H4, where both nitrogen and helium are used, the direct gas sampling results can be used to estimate the relative amounts of nitrogen and helium, once that is known the non-condensable gas quality and the steam quality can be determined. The density of the mixture of steam and non-condensable gases can then be calculated using the steam quality and saturated steam density.
i The error analysis for the venturi flow meter will be included in the Data Transmittal Report.
-~q un n
~,-r-em s-v-n-,e-,
w
,,w~--,,2e-
--tw w
e m>-
r-
,vv-~s-w nn~wwee-r---
-e w
w
-~~--
---Y
- --wr
=-~-www~-
+rv=--
-s M
wn-s+
+
I i
01328.02 Document to the staff the basis for the total mass of helium for tests H3 and H4 in the GIRAFFE / Helium Tests. The total SBWR zirconium mass and mass oxidized used in l
the limiting design basis accident should be considered along with a discussion of its relationship to the amount of helium selected for the tests.
- RESPONSE TO 01328.02 The SBWR fuel cladding zirconium mass is equal to 23,120 kilograms. Two moles of hydrogen gas are generated per mole of zirconium reacted, therefore, 0.02192 l
kilogram moles of hydrogen gas are generated per kilogram of zirconium reacted. Due
{
to a 100% fuel-clad metal water reaction 506.8 kilogram moles of hydrogen gas are i
generated, l
i A total helium mass of 0.97 kilograms is used for tests H3 and H4, this is equivalent to l
20 % by volume of the scaled amount of hydrogen gas that would be generated by a I
100% fuel-clad metal water reaction.
1 The. purpose of the H3 and H4 tests is to demonstrate the effects of a high concentration of a lighter than steam gas on the performance of tne PCC. At the same time, the pressure capability of the facility must be considered. The design pressure for the Drywell and Suppression chamber is 0.6 MPa. In order to assure that the design pressure is not exceeded during tests H3 or H4,. a helium mass less than the scaled amount determined for a 100% fuel-clad metal water reaction is used.
Therefore, GE decided for tests H3 and H4 to use a helium mass of 0.97 kilograms, which is equivalent to 20% by volume of the scaled amount of hydrogen gas that would be generated by a 100% fuel-clad metal water reaction. This results in an initial GIRAFFE Drywell concentration of approximately 23% helium by volume. Since this quantity is equal to approximately 60 times the PCC volume it is a sufficiently high quantity of helium to capture the prototypical behavior of lighter-than-steam gases on the performance of the PCC.
GIRAFFE Facility Characterintion Tests OI328.03 Document to the staff the conservative heat loss strategy in GIRAFFE, and the importance of net heat into the system.
RESPONSE TO 01328.03 i
Facility heat loss tests were performed to determine the required microheater power and additional RPV bundle heater power in order to maintain a constant vessel pressure i
of 0.4 MPa and to thereby minimize the effects of heat loss to the environment.
During the heat loss tests, the Drywell temperature was monitored in order to prevent the generation of superheated steam which would result if excessive amounts of microheater power were input.
~-
l 1
The generation of superheated steam is not desirable because it is expected that it could
[
result in additional circulation in the vessels which would not be representative of the SBWR design.
OD28.04 l
Document the location and control of the GIRAFFE microheaters to the staff.
RESPONSE TO OD28.04 The locations of and the controls for the microheaters are provided in Section 4.6 and Figure 7.5 of the GIRAFFE Heat Removal Performance Tests Test Plan and i
Procedures. (Toshiba document No. TOGE110-T07, Rev. 2) l 1 01328.05 I
Provide the NRC with heat loss and pressure loss data for the GIRAFFE Facility.
RESPONSE TO OD28.05 GE will provide the NRC the Apparent Test Results for the heat loss and pressure loss tests by July 31,1995.
i 1
GIRAFFE-Non-Condencable Gas Measurements (Method and Locations) 2 OD28.%
Finalize the number (3 or 4) and location of GIRAFFE Non-Condensable Gas Measurement sample points and submit a description to the staff.
I RESPONSE TO 01328.06 i
There are three non-condensable gas measurement sample points. There locations are shown in Figure 3-1 of the GIRAFFE Helium Test Specification. (GE document No.
25A5677, Rev.1)
-., =...,,.. -
~
01328.07 i.
The accuracy of the GIRAFFE thermocouple measurements as used in calculation of non-condensable fraction should be addressed.
1 i
l RESPONSE TO 01328.07 i
s The accuracy of all GIRAFFE thermocouple measurements is plus or minus 1.0 K, and j
is specified in Section 7.4(4) of the GIRAFFE Heat Removal Performance Tests Test Plan and Procedures. (Toshiba document No. TOGE110-T07, Rev. 2) 4 1
1 PANDA-Non-Condencable Gas Measurements (Method and Locations)
J f
j 01328.08 i
l GE has committed to additional PANDA instrumentation for Non-Condensable Gas Measurements; the location and type need to be finalized and described to the staff.
Provide discussion of the final PANDA non-condensable gas measurement approach.
RESPONSE TO 01328.08 The non-condensable gas concentration will be measured continuously at eight locations during the integral systems tests in PANDA. These locations are shown in Figure A.3-13d of the SBWR Test and Analysis Program Description, NEDO-32391, Rev B. The sensors will be positioned near the vertical centerlines of the two drywells and two wetwell gas spaces, at three elevations in each drywell and one elevation in each wetwell gas space. The sensors measure the oxygen partial pressure. The oxygen partial pressure is determined by measuring the voltage generated across a zirconia element. The voltage is a function of the sensor temperature and the differential oxygen pressure across the zirconia element.
The air partial pressure can be determined from the oxygen partial pressure, and the non-condensable gas concentration can be determined from the air partial pressure and the total pressure.
Data Reoorrino 01328.09 GE and NRC lead engineers need to be identified and agree on format for electronic transmittal of test data. Define the GIRAFFE, PANDA, and PANTHERS data reporting method, scope, and format.
RESPONSE TO OI328.09 The GE responsible engineers are as follows: PF Billig for PANTHERS PCC and IC tests, M Herzog for GIRAFFE Helium series tests, JD Duncan for GIRAFFE Systems Interaction tests, and JE Torbeck for PANDA Steady State and Transient tests.
l
. The electronic transmittals of test data for each of the test programs will be provided on either 3.5 inch floppy disks or on 4 mm digital audio tape. The data files will be provided in an easily understood format, and in all cases, a description of the data file format will be included in the Data Transmittal Report.
Provided below, as an example is a brief description of the electronic transmittal for PANTHERS PCC Test data:
Appendix E of. the PANTHERS-PCC Data Report (SIET Document 00393RP95) describes the format of the data tape for the PANTHERS-PCC tests. Copies of the data tapes were sent to the NRC in MFN 057-95 (April 14, 4
1995) and MFN 086-95 (June 30,1995), and the appendices to the report were sent in MFN 075-95 (May 11,1995). Below is a brief description of the tape format from that appendix.
All PANTHERS thermal-hydraulic data are stored on 4 mm 120 Mbyte tapes using a Colorado Memory Systems "Trakker 250 backup device. The data files contain all the directly acquired thermal-hydraulic data (e.g., temperatures, pressures, etc.) and derived thermal-hydraulic quantities (e.g.,
power condensed, mass flowrates, water levels, etc.). In addition to these files, other data files are included which give instrument zeroes, constants used for calculation of derived quantities, historical data covering the entire test day, and files of fast acquired thermal-hydraulic data. All of these files are explained in Appendix E of the PANTHERS-PCC Data Report.
All files are in ASCII format with a comma (,) or semicolon (;) used as a separator.
For each column of data, the first four lines indicate, the measurement name, the measurement unit, and the channel number. For the directly acquired data, the measurement names are the plant codes of the instmments. For the derived quantities, the measurement names are defined in l
Appendix E.
Scaling of GIRAFFE / Helium 01328.10 Address the relative heat loss between the GIRAFFE / Helium PCC heat exchanger headers and tubes.
RESPONSE TO OI328.10 The relative heat loss through the GIRAFFE / Helium PCC heat exchanger headers is very small. The estimated header heat loss is >0.5% of the total PCC heat removal at rated conditions. This is accomplished by insulating the steam header with a 4 mm air gap on the radial surface and a 15mm air gap on the top. For comparison the SBWR heat losses under similar conditions are about 0.5% of the total PCC heat removal. In both cases the header heat losses are small enough that there is no significant effect on the overall system behavior.
1
..s-w
---,v.-r
---,y
,y,
,,,-~e.w-%
.-,.v4-
,-4.,.,,,m,
-.wm.ytu
. ar
-...,~--w=--,-nv
--m-mm-em-n-,we.d,-
w
i 01328.11 Further define the GIRAFFE / Helium scaling analysis for 2 component non-condcasables; manometer oscillations between pools need to be addressed.
RESPONSE TO 01328.11 There is the possibility that the presence of non-condensable gases will result in pressure oscillations due to the PCCS alternately filling with non-condensables and venting them to the suppression pool. The concern here is that these pressure oscillations may result in pool level oscillations and further that these level oscillations will in turn provide feedback for further pressure oscillations. GE believes that the NRC has more of a concern about this for lighter-than-air gases because they may accumulate at the top of the PCC and purge in large burps rather than a slow bleed off through the vent.
The magnitude of pressure oscillations in the PCCS will be limited to a maximum value of about 7000 Pa by the PCCS vent submergence of 0.75 m. This pressure difference will be between the Drywell and Wetwell. Using this as a bounding driving pressure between both the DW-to-RPV and WW-to-RPV, GE has performed an analyses which shows that any water level oscillations will be highly denped and that no significant water level oscillations will occur. This would be tme for larger driving pressures also.
The details of this analysis will be provided in the next revision of the scaling report (NEDC-32288, Rev.1) to be issued in September 1995.
OI328.12 The amount of helium to be used in GIRAFFE does not appear to have been adequately scaled. Provide the basis for the total mass of helium to be used.
RESPONSE TO OI328.12 Covered by the response to OI328.02.
OI328.13 Provide the basis for the heat distribution associated with GIRAFFE / Helium drywell 1I(pch) of 48.
RESPONSE TO OI328.13 The value of interest is P n from the March 8 & 9 NRC meeting. The value of P n =
pc pc 0.48 for the drywell indicates that 48% of the decay heat at this point in time is removed by the PCC. The question is what is happening to the other 52% of the heat.
This value is based on a heat balance of the SBWR.
The point in time at which this heat balance was taken is approximately 4000 s into a main steam line break. This is at the transition from the GDCS phase into the long-term
_~_
cooling phase. The remaining 52% of the decay heat is accounted for by several sources, including: warming subcooled GDCS flow that enters the RPV; heat passed to j
i the wetwell by the flow of uncondensed steam in the PCC vent; and heating containment structures. A review of the TRACG SBWR predictions for this point in time show that the percentage of heat removed by the PCC is consistent with P n pc (Decay heat = 26 MW, PCC heat removal = 12 MW).
This point in time was selected because it is representative of the start time for the GIRAFFE / Helium tests. The heat balance averaged over the period from 4000 to 6000 sec is shown in Table RI328.13. Average values are used because results taken graphically from TRACG predictions are difficult to quantify at a specific point in time. The average values are indicative of the ratios at any point in time during that period.
Table RI328.13. Averaged Heat Balance for RPV and DW from 4000 s to 6000s into a main steam line break Decay heat 21 M W 4
Condensation in PCC 13 M W Heat up sub-cooled GDCS flow to saturation 5.5 MW temperature i
Uncondensed steam flow through PCC vent 2.5 MW Heat up DW structures 1MW Since the numbers are taken graphically from the TRACG predictions they are approximate and the total heat usage adds up to 22 MW instead of the 21 MW of decay heat. The level of accuracy is adequate for scaling analysis. Additionally, there are other heat sources and sinks of less than 1 MW that are not shown in the table since they are not significant. These include, heat released by the RPV vessel, vacuum l
breaker leakage and pressurizing the drywell.
TRACG Model for SBWR Containment Analysis 01328.14 The GE TRACG Model needs to address overall margins in the containment design (compare PCC capacity vs. decay heat input and demonstrate containment safety margins).
RESPONSE TO 01328.14 The SBWR containment pressure and temperature time histories are calculated using TRACG. The methodology and application of TRACG is detailed in GE Licensing Topical Reports NED-32176, 32177, and 32178, submitted to the NRC for review in February,1993. The SBWR containment structural and dynamic loads are calculated using methods and load combinations used in previous product line plants and reviewed by NRC on ABWR, GESSAR, and Operating Plant dockets.
The overall SBWR containment design margins are addressed Appendix 3E of the SBWR SSAR. Analyses of containment design basis events are presented in section 6.2 of the SBWR SSAR and
show' that considering both decay and sensible hn.t sources and tbc expected distribution of energy and available heat sinks, triat there is still excess capability in the PCCS for accommMating limiting Design Basis Accident (DBA) event and single failure combination.
01328.15 Prepare comparison of heat transfer degradation by non-condensables as predicted by the TRACG correlation and the Uchida correlation.
RESPONSE TO 01328.15 The question arose when discussing condensation heat transfer in the drywell. A comparison of heat transfer coefficients is more appropriate than a comparison of heat transfer degradation. The graph below shows a comparison of the correlation in
(
TRACG verm.:s the "Uchida" correlation for a typical containment pressure of approximately 3.3 bar. Uchida is an average heat transfer correlation. However, the TRACG correlation based on the Vierow-Schrock or "Tsukuba" correlation is a local heat transfer correlation.
So, in order to make a comparison, the heat transfer coefficient from the TRACG correlation was averaged over a vertical slab 5 meters in height.
i Comparison of TRACG vs. Uchida Correlation 10000 h
to00em \\
[
ti..a j
'E 6@ !
...TRACG y4000 o..
L
, 3000
~
\\
0 0 01 01 1
Noncondensible Mass Fraction n
{'
The figure compares the heat transfer coefficients of the TRACG and Uchida correlations over a wide range of noncondensible gas mass fractions. The TRACG correlation predicts less heat transfer for noncondensible mass fractions less than 10%
fi (this corresponds to a majority of the containment conditions) and more heat trar.sfer for mass fractions greater than 10%.
The TRACG prediction of the performance of the containment should not depend significantly on the particular cctrelation used for drywell condensation. The drywell structures must eventually come to thermal equilibrium with the drywell atmosphere.
To do thi.s heat must be transferred to the wall by condensation. The only difference
the correlation makes is in the timing of the drywell condensation, not in the total amount of drywell condensation. In addition, the amount of heat transfer (and thus, condensation) to thick structures in the drywell very quickly becomes controlled by conduction of heat away from the surfaces and not by the condensation heat transfer resistance.
OI328.16 Define and justify the basis of the temperature difference used in the TRACG correlation for heat transfer, both for tubes and on exterior surfaces.
1 RESPONSE TO 01328.16 This question arose during the discussion of the condensation heat transfer correlation used in TRACG. The temperature difference used in the TRACG correlation for condensation heat transfer is the saturation temperature at the steam partial pressure for the node nearest the surface minus the wall temperature. This is consistent with the definition of the Vierow-Schrock correlation. The description of the data reduction technique for calculating the saturat on temperature and heat transfer coefficients in the i
Vierow experiment can be found on pages 2-2 and 2-3 of NEDC-32301 " Single Tube Condensation Test Program." In that document it describes how the steam partial pressure (tmd therefore the saturation temperature) is calculated by determining how mu/1 steam has already condensed Hor to the axial position of interest in the tube.
TFKG also calculates heat tran m t w to conduction / convection from the steam to the interface if the steam temperatw k a'ferent from the saturation temperature. This is small and in parallel to the condmAnn heat transfer.
Drvwell Mixing OI328.17 Provide discussions of GE's approach to address drywell mixing.
RESPONSE TO OI328.17 The complexity of the drywell mixing combined with the fact that TRACG is not a CFD code suggests that a conservative treatment of the drywell mixing should be used in qualifying TRACG for use on the SBWR. This can be done through the use of a bounding model in the otherwise best estimate predictions from TRACG or by the addition of a bias to the final best estimate predictions. Additional discussion on this topic is included in RT328.21.
The determination of the bias or bounding model to be applied for mixing will be determined through a combination of sensitivity studies using TRACG or other systems codes and review of the results form the SBWR certification tests with various demands on the drywell and PCC. In the code sensitivities the best estimate of drywell mixing will be overridden to force a range of behaviors involving gas concentration r.d locations to determine the maximum effect on the containment pressure and c.
c j.
. temperature. This difference can then be applied as a conservative bias to the TRACG results.
}
The Test and Analysis Program, as desc-iW in the TAPD, includes a wide range of i
tests which will give information on the effects of various non-condensable gas j-conditions. These will be used to confirm that no important effects have been missed in j
the analytical sensitivity studies.
i i
i-Plans for TRACG Validation of Rhort Term Containment Recnonce
}
01328.18 i
Consider additional TRACG qualification against PSTF data including pool swell back pressure and HCU floor pressure drop effects for TRACG Validation of Short Term j
Containment Response.
j RESPONSE TO 01328.18:
l As part of the TRACG qualification effort for short-term containment response,
~
comparisons are being made between TRACG calculations and both PSTF MARK II and MARK III test results. The tests to be analyzed and the comparisons to be performed are described in Paragraphs A.3.1.8.4, A.3.1.8.5, and A.3.1.8.6 of the TAPD. Comparisons of calculated and measured vent flow rate will be made for both MARK II (closed wetwell) and MARK III (open wetwell) tests. A satisfactory comparison between TRACG and both these data sets will confirm the ability of TRACG to model horizontal vent flow into a closed wetwell. Additional comparisons between TRACG and PSTF data will include drywell and wetwell pressures and suppression pool temperature distribution.
S 1
i i
i
,,..._.._.,.~.,,,,.4
_.,,.m,...,.
,,,...,e
. ~ r, r w,.m
.,.. -,-.._.,-...,, _..-.........,._ ~,. -,,..
,-..4...,~
,., ~,.., -.,, -.,,
1 01328.19 How will containment structural and dynamic loads be incorporated into the SBWR design.
RESPONSE TO 01328.19 The SBWR containment pressure and temperature time histories are calculated using the TRACG code. The methodology and application of TRACG is detailed in GE Licensing Topical Reports NED-32176, 321277, and 32178, submitted to the NRC for review in February,1993. The SBWR containment structural and dynamic loads are calculated using methods and load combinations used in previous product line plants l
and reviewed by the NRC on ABWR, GESSAR, and Operating Plant dockets. The l
SBWR containment hydrodynamic loads due to a postulated Loss Of Coolant Accident (LOCA) and Safety Relief Valve (SRV) / Depressurization Valve (DV) actuation are defined in Appendix 6A of the SSAR. The structural analysis and design of the SBWR Reactor Building, including containment, is documented in Appendix 3E of the SBWR l
SSAR.
01328.20 Document how the droplet flow model in TRACG applies to fogging in the containment for TRACG Short Term Containment Response.
RESPONSE TO 01328.20 When TRACG calculates a vapor temperature below saturation, condensation in the form of droplets takes place with the droplet size governed by a critical Weber number.
I In effect, TRACG assumes that the droplets are created at their final size. TRACG does not simulate the rucleation of small droplets which subsequently grow by agglomeration. In consequence, TRACG allows the droplets to rain out faster than would be expected when fogging occurs.
Apolication Methodology OI328.21 l
Provide additional Application Methodology discussions m the approach to CSAU for containment.
RESPONSE TO 01328.21 There is a significant amount of regulatory guidance in the use of best estimate codes for ECCS/LOCA (meeting the regulations in 50.46 of 10CFR50) through Regulatory Guideline 1.157 and the supporting references. For the application of best estimate codes to containment design, on the other hand, there is very little guidance. Therefore ongoing dialogue between GE and the Containment Systems Branch of NRR will be needed to agree on an acceptable application methodology. GE plans to do this through
periodic phone calls and mcctings. This OI response will provide a starting point for those discussions.
The Federal Regulation of interest for containment code licensing is General Design Criteria (GDC) 50 in Appendix A of 10CFR Part 50. The key point in this criteria is that the containment must accommodate the pressures and temperatures resulting from any LOCA without exceeding the design leakage rate and with sufficient margin.
Further this sufficient margin should include consideration of:
effects of potential energy sources not included in the determination of peak e
conditions limited experience and experimental data available for defining accident phenomena e
and containment responses the conservatism of calculational model and input parameters e
The applicable sections of the Standard Review Plan (SRP) for standard plants (NUREG/CR-0800) provides some guidance for the review of containment performance. Section 6.2 gives guidance on the review of the containment portion of engineered safety features. In this document sufficient margin is identified as 15% for standard plants. GE believes that the guiding principles behind Reg Guide 1.157, applied to the containment, will be sufficient to meet the requirements of GDC 50.
Therefore it is our intention to use a CSAU-like approach. The details of which will have to be resolved with the NRR-CSB.
The CSAU methodology provides a rigorous method for determining the uncertainty of a best estimate code to a given level of confidence. This includes the uncertainties resulting from code uncertainty, experimental data used including scale up effects, boundary and initial conditions, fuel behavior, and simplifying assumptions. The figures of merit for the containment will be pressure and temperature. Biases can be added to these uncertainties for things such as insufficiently detailed models, as described in 01328.17 for drywell mixing.
The 95% probability value, which includes the statistically determined combination of all of the individual uncertainties plus any biases, will be compared against the design limits for the containment. In addition, the nominal best estimate of performance will be compared against the design limit with an additional 15% margin. This is shown conceptually in Figure 1.
~
P or T e
Right at limit
/
Design Limit
, ___ ' ' ' Best enimate plus Design Limit
---T~~~-~~~
- ~ ~ ~ ~ -
Dias plas Ursertainty I'*
8" Minus 15%
l
/
Nominal Best Estimate Prediction l
l l
l t
\\
Figure 328.21-1. Schematic of application of design margin for TRACG containment application.
In the example shown in the figure the containment is right at the design limit based on best estimate plus uncertainty type application and has some additional margin to the 15% below design limit level based on the best estimate results.
1 Because there is less certainty in the modeling of some of the phenomena associated with containment performance some parameters will need to be treated in a i
conservative manner through the use of conservative models or biases. The following table provides a preliminary list of the model uncertainties to be addressed and whether they will be addressed by sensitivity analysis or bounding analysis.
l l
i
Table 328.21-1 Preliminary list of phenomena and method of incorporating in
- TRACG model uncertainty.
PHENOMENA Uncertainty / Bounding BRI Break flow Sensitivity SQl SRV flow Sensitivity DW2 Condensation on DW walls Sensitivity Wall / Structure conduction Sensitivity DW3 3 D effects - phase distribution Bias or Bounding Model 3 D effects - noncondensibles distribution Bias or Bounding Model Buoyancy / natural circulation Sensitivity WW3 Condensation / evaporation of PCC vent discharge Sensitivity WW5 Condensation on WW walls Sensitivity Conduction through WW walls Sensitivity WW6 Pool mixing and stratification Bounding Model WW7 3 D effects in gas space-temperature distribution Bias or Bounding Model 3 D cffects in gas space-noncondensible distribution Bias or Bounding Model WW8 Containment spray condensation Sensitivity WW9 Containment hydrodynamic loads Done with existing Licensing basis using M3CPT GD2 GDCS flow Sensitivity PCI PCC flow / pressure drop Sensitivity PC2 Condensation on primary side Sensitivity Degradation by n/c Sensitivity PC3 Secondary side heat transfer-pool temperature dist.
Sensitivity Secondary side heat transfer-pool void dist.
Sensitivity Secondary side heat transfer-natural circulation Sensitivity Secondary side entrainment Sensitivity PC4 Paralle! PCC tube effects Sensitivity PC5 Parallel PCC unit effects Sensitivity PC8 PCCS startup with n/c Sensitivity DWB1 Leakage between drywell and wetwell Bounding Model/ Sensitivity EQ2 Sloshing through equalization line Sensitivity OCl Heat transfer to safety envelope Sensitivity Additional Application Methodology discussions in the approach to CSAU for containment will be provided in a revision to GE LTR NED-32178.
l l
l l
01328.22 i
Provide additional Application Methodology discussions that assure the containment pressure and temperature are sufficient figures of merit for the containment CSAU.
RESPONSE TO OI328.22 Further Application Methodology discussions have been provided in the response 01328.21. This provides some insight into how pressure and temperature will be used for code qualification. GE does not see that any criteria other than pressure and temperature should be used as figures of merit. Our review of 10CFR50 and the Standard Review Plan for standard plants supports this positions.
Additional Application Methodology discussions that assure the containment pressure and temperature are sufficient figures of merit for the containment CSAU will be provided in a revision to GE LTR NED-32178.
1 01328.23 GE needs to assure that the TRACG model report distinguishes applicability of correlations used in TRACG to containment modeling.
RESPONSE TO 01328.23 GE will assure that the applicability to containment of correlations used in TRACG is clear in the TRACG LTRs.
01328.24 List which GIRAFFE and PANDA tests will have blind analyses performed with TRACG prior to and/or after conduction of the tests.
RESPONSE TO OI328.24 The GIRAFFE / Helium TRACG Analysis Plan is discussed in the GE LTR NED-32391, SBWR Test and Analysis Program Description (TAPD), Revision-B Appendix A, paragraph A.3.1.6.5 and Table TA.2-3. The GIRAFFE / SIT TRACG Analysis Plan l
is discussed in paragraph A.3.1.7.5 and Table TA.2-3.
The PANDA TRACG l
Analysis Plan is discussed in paragraph A.3.1.3.5 and Table TA.3-13.
PANTHERS /PCC TRACG Analyses are listed in Table TA.3-3 and PANTHERS /IC Analyses are listed in Table TA.3-6. These sections of the report stipulate pre / post and blind analyses.
l l
l!.,,..,._., _.
_ _ _ _.. _. _ _,. _.., _.