ML091870829

Calculation No. FAI/09-44R, Rev. 0, Post-Test Analysis of the Fai Millstone 3 RWST Scale Gas Entrainment Test, Enclosure 2

ML091870829
Person / Time
Site:	Millstone
Issue date:	03/13/2009
From:	Ramsden K Fauske & Associates
To:	Office of Nuclear Reactor Regulation
References
09-186A FAI/09-44R, Rev 0
Download: ML091870829 (52)

Text

Serial No. 09-186A Docket No. 50-423 ENCLOSURE 2 (Non-Proprietary)

ONE COPY OF THE NON-PROPRIETARY REPORT FAI/09-44R "POST-TEST ANALYSIS OF THE FAI MILLSTONE 3 RWST 1/4 SCALE GAS ENTRAINMENT TEST" MILLSTONE POWER STATION UNIT 3 DOMINION NUCLEAR CONNECTICUT, INC.

ENCLOSURE 4 TO THIS LETTER CONTAINS PROPRIETARY INFORMATION - WITHHOLD UNDER 10 CFR 2.390. UPON SEPARATION OF ENCLOSURE 4, THIS LETTER IS DECONTROLLED.

FAI/09-44R Page 1 of 51 Rev. 0 Date: 03/13/09 FAUSKE & ASSOCIATES, INC.

CALCULATION NOTE COVER SHEET SECTION TO BE COMPLETED BY AUTHOR(S):

Calc-Note Number: FAI/09-44R Revision Number: 0

Title:

Post-Test Analysis of the FAI Millstone 3 RWST 1/4 Scale Gas Entrainment Test Project: Assess the transport of non-condensable gas in the Project Number Millstone 3 ECCS suction lines. Or Shop Order:

Purpose:

The purpose of this report is to assess the ability of RELAP5 to predict the entrainment and transport of gas in the FAI Millstone 3 RWST 1/4 scale test facility.

Results Summary: The comparisons of RELAP5 model results to test data shows that RELAP5 hydrodynamic models predict entrainment conservatively and reasonably.

References of Resulting Reports, Letters, or Memoranda (Optional)

Author(s): Completion Name (Print or Type) Signature: Date:

A Kevin Ramsden ,- M 3/13/09 Z=ýZzz SECTION TO BE COMPLETED BY VERIFIER(S):

Verifier(s): Completion Name (Print or Type) Signature: Date:

Damian Stefanczyk ý2 ex- 3/13/09 Independent Review or Method of Verification: Design Review Alternate Calculations X Testing Other (specify):

SECTION TO BE COMPLETED BY MANAGER:

Responsible Manager:

Name (Print or Type) Signature: Approval Date:

Robert E. Henry _ _ _'o _ _ 3/13/09

FAI/09-44R Page 2 of 51 Rev. 0 Date: 03/13/09 CALC NOTE NUMBER FAI/09-44R PAGE 2 CALCULATION NOTE METHODOLOGY CHECKLIST CHECKLIST TO BE COMPLETED BY AUTHOR(S) (CIRCLE APPROPRIATE RESPONSE)

1. Is the subject and/or the purpose of the design analysis clearly stated? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . .

. NO

2. Are the required inputs and their sources provided? NO N/A

3. Are the assumptions clearly identified and justified? .................... ........ NO N/A

4. Are the methods and units clearly identified? ...................................... NO N/A

5. Have the limits of applicability been identified? ................................... C NO N/A (Is the analysis for a 3 or 4 loop plant or for a single application.)

6. Are the results of literature searches, if conducted, or other background data provided? ........................................ YES NO

7. Are all the pages sequentially numbered and identified by the calculation note num ber? ...................... NOO............................................

8.. Is the project or shop order clearly identified? NO N/A

9. Has the required computer calculation information been provided? ....... NO N/A

10. Were the computer codes used under configuration control? ....... ........ NO

11. Was the computer code(s) used applicable for modeling the physical N/A and/or computational problems identified? .......... N............O.................

N (i.e., Is the correct computer code being used for the intended purpose.)

12. Are the results and conclusions clearly stated? ................................... C NO

13. Are Open Items properly identified ................................... . ..... ......... YES NO

14. Were approved Design Control practices followed without exception? ...... YES NO (Approved Design Control practices refers to guidance documents within Nuclear Services that state how the work is to be performed, such as how to perform a LOCA analysis.)

N/A

15. Have all related contract requirements been met? ...................... N.........

NO NOTE: If NO to any of the above, Page Number containing justification:

FAI/09-44R Page 3 of 51 Rev. 0 Date: 03/13/09 AFAUSKE WORLD LEADER IN NUCLEAR AND CHEMICAL PROCESS SAFETY Report No.: FAI/09-44R Post-Test Analysis of the FAI M/listone 3 R WST 1/4 Scale Gas EntrainmentTest Submitted to:

Dominion Prepared by:

K Ramsden Reviewed by:

D. Stefanczyk

FAI/09-44R Page 4 of 51 Rev. 0 Date: 03/13/09 ABSTRACT A gas volume detected in the Millstone Unit 3 RWST suction line led to installation of additional piping vents and required that an understanding of the transport of that gas in postulated ECCS initiation events be developed. The utility performed detailed RELAP5 modeling of the ECCS suction piping as part of their assessment. In addition, a 1/4 scale test of the RWST suction piping was built at FAI and exercised to experimentally investigate the fluid dynamic response and serve as a test for the analytical predictions. As part of this testing effort, a RELAP5 model was developed and used to analyze the scale model results. This report documents the results of that effort.

This report evaluates the RELAP5 model of the 1/4 scale test loop and compares the gas entrainment predictions to the test data. It also shows that the test loop gas separators introduce a velocity oscillation on pump start in the test facility. The results of the comparison demonstrate that RELAP5 gas entrainment models conservatively over-predict the gas transport observed.

F-AI/09-44R Page 5 of 51 Rev. 0 Date: 03/13/09 TABLE OF CONTENTS ABST R AC T ............................................................................................................

1.0 INT R O DUC T IO N ....................................................................................... 8

2.0 DESCRIPTION

OF TEST FACILITY/RESULTS ...................................... 9 3.0 RELAP5 Calculations for the Experiment .................................................. 12 3.1 Supplemental Test Comparison ......................... 15 4.0 C O NC LU S IO NS ........................................................................................ 24

5.0 REFERENCES

......................................... 25 APPENDIX A: Pretest Predictions for Case 1 ........................................ 26 APPENDIX B: Model Development Calculations .................................. 31 APPENDIX C: RELAP5 Model Input Deck .......................... 38

FAI/09-44R Page 6 of 51 Rev. 0 Date: 03/13/09 LIST OFFIGURES Figure 3-1 Test facility RELAP5 model diagram... ........................................................ 13 Figure 3-2 Case 1 charging flow velocity ....................................................................... 16 Figure 3-3 Case 1 RH line gas transport .......................................................................... 17 Figure 3-4 Case 9 RH line gas transport .......................................................................... 18 Figure 3-5 Case 10 RH line gas transport ...... .................................................................. 19 Figure 3-6 Test Case 12 charging line velocity .............................................................. 20 Figure 3-7 Test Case 12 gas void fraction near charging line elbow ................ 21 Figure 3-8 Test Case 12 fluid velocity near charging line elbow .................................... 22 Figure A-1 Void fractions in 6 inch header 5% void case ..................................... 27 Figure A-2 Void fractions exiting the time dependent junctions - 5% initial void case.

.... ..... . ,o,°o,,ooo.............

o .. ,........... ,,,,,,...... ............................................... 28 Figure A-3 Void fractions in 6 inch header 10% void case ..................................... 29 Figure A-4 Void fractions exiting the time dependent junctions - 10% initial void case.

....... , ................. .......................................... . . .. ............ ......... 3 0

FAI/09-44R . Page 7 of 51 Rev. 0 Date: 03/13/09 LIST OF TABLES Table 2-1 Test Matrix (Experimental Planning) .................................................... 10 Table 3-1 Air Transport Result Comparison ............................ 23 Table 3-2 Supplemental Tests ..................................... 23 Table 3-3 Supplemental Test Com parison .......................................................... 23

FAI/09-44R Page 8 of 51 Rev. 0 Date: 03/13/09

1.0 INTRODUCTION

Recently, a gas void was identified in the Millstone 3 ECCS suction piping adjacent to the RWST. The gas void was conservatively estimated at 8% in a section of 24 inch piping upstream of the RHR, SIH, and Charging pumps. Analytical predictions of the void transport were prepared by utility personnel using the RELAP5 Mod 3.3 computer code as part of their assessment of the safety significance of this event. A 1/4 scale test was also performed at FAI to provide additional understanding of the gas transport phenomena in this condition. The RELAP5 code was also used to perform post-test analysis of the 1/4 scale testing. This report documents the results obtained and insights gained in this analysis work.

The objective for this report is to demonstrate the capability of RELAP5 Mod 3.3 to predict the major phenomena observed in the scale tests. One result of this work was the development of insights with respect to the key phenomena observed in the test as well as the application of the code to properly capture them.

It should be noted that both the utility personnel as well as FAI applied the RELAP5 Mod 3.3gl (patch 03) configuration in performing this work. This is significant due to the changes implemented in the horizontal stratification vapor pull through models with this release.

FAI/09-44R Page 9 of 51 Rev. 0 Date: 03/13/09

2.0 DESCRIPTION

OF TEST FACILITY/RESULTS The Millstone 3 ECCS suction test facility is described in detail in (FAI, 2009). In summary, the facility consisted of a main 6" loop with prototypic connections for the RH, SIH, and Charging systems. The RH suction connection is a 4 inch diameter 45 degree downward tee connection. The SIH suction is a 2 inch pipe that connects to the bottom of the 6 inch main header, just upstream of the RH connection. The charging line is a horizontal connection downstream of the RH suction tee.

The RH piping continues after the 45 degree downward tee to a horizontal header that leads to a gas water separator. The test RH pump takes suction on a line from the side of the separator. Test RH flow measurement is taken between the pump and the separator outlet. It should be noted with respect to the much larger flow rate for the scaled RH system compared to the SIH and charging systems, the gas volume in the RH separator was comparable to the gas volumes used in the SIH and charging separators during the testing. .

The SIH piping drops vertically 9 inches and then is routed horizontally to a gas water separator. In the test SIH pump takes suction on a line from the side of the separator. The test SIH flow measurement is taken between the pump and the separator outlet. A fairly large gas volume was applied in the SIH and Charging separators during testing.

The Charging line drops vertically approximately 39 inches from the elbow connecting it to the header downstream of the RH suction takeoff. There is a short horizontal run of approximately 12 inches before the line enters the separator. The test Charging pump takes suction on a line from the side of the separator. The test flow measurement is taken between the pump and the separator outlet. The charging separator and SIH separator were constructed nearly identically and were operated comparably.

The test matrix and observed results is shown on the next page.

FAI/09-44R Page 10 of 51 Rev. 0 Date: 03/13/09 Table 2-1 Test Matrix (iExperimental Planning)

Flow Flow, Flow F Rate ti Initial Gas Gas Volumnes N' For tihe Collected (in")

Test Rate For Rate For Fo t Void Number tihe RHS the SIl Charging Fraction Purpose of the Test gm) gp) (gpmi

• System __________________ SIH Charging I 315 23 22 8 Rmiwesenl the max ESF case. 0 8.6 2 310 23 22 8 Rt!pcat olTest #1. 0 0 3 315 24 21 8 RzpL,,at ofTest #1. 0 4.3 Irwwease the Froude numlners for thie SIH 4 310 38 34 8 ancharging flows by 50%) as 0 17.2

_*_o:mmnended by the Hydraulics Institute.

5 310 38 34 8 Rlijivat Test #4. 0 17.2 6 310) 38 34 8 R-eaI Test #4. 0 17.2 7 0 40 34 8 Inwdsligate small break LOCA response. 0 159.8 8 0 .40 34 8 RLpoat Test #7. 0 137.6 9 170 24 21 8 lamlesligale R HS. a single train response for the 0 30.1 W0 97 24 21 8 Wimehmark case for RIELAP5. 0 98.9 II 0 27 22 5 lamcstigate smaller avet\'age void for small 0 0 kwaik LOCA conditions.

12 0( 27 22 8 tumcstigate small break LOCA without a 0 30.1 5'1CIAincrease in) thie FroIude nube11r.

13 0 27 22 8 RWpca Test # 2.? 0 34.4 14 172 26 22 8 Rpyal 'Test#9. 0 51.6 15 172 25 22 8 Wqp.al Test #9. 0 47.3 16 175 25 212 5 n~e stigate single RHS train hehavior with 6.5 Z1stmal lr averlae void.

17 175 25 22 5 Rqoat Test# 116. 0 4.3

FAI/09-44R Page 11 of 51 Rev. 0 Date: 03/13/09 Cases selected for post test analysis comparison were:

1) Case 1

2) Case 9

3) Case 10

4) Case 11

5) Case 12 The pump flow transient data was reviewed to develop approximations of the flow transient imposed by the pump start. Two observations were significant in the review of the data and photographic results for the various cases:

a. The test loop tended to accumulate more gas in the charging line separator than would have originally been expected, based on the RELAP5 calculations for the Millstone 3 plant.

b. The gas transport in the charging line was related to the initial pump start transient. Once steady conditions were achieved, a stable gas bubble would be formed at the charging line elbow with water flowing underneath it. Very little gas stripping occurred in this condition, which is consistent with the relatively low flows present.

FAI/09-44R Page 12 of 51 Rev. 0 Date: 03/13/09 3.0 RELAPS Calculations for the Experiment Pretest predictions were made for Case 1 with a simplified model. (Appendix A) The pretest model reasonably predicted the results for Case 1, namely that the gas would primarily be transported down the RH line with little or no gas transported to the SI and charging pumps.

Further application of this model to other cases quickly demonstrated that the simplified model was not adequately capturing the startup transient deposition of gas to the charging line in the experiment. The steady state behavior of the model was consistent with the test data. The model was revised to include the following:

1) Gas Separators were added for all three suction lines. Initial air volumes were modeled for the charging and SI. The RH accumulator was assumed full of water, due to the smaller ratio of the gas volume the the RH flow rate in the scaled tests. (Initially, the separators applied the stacked volume level option, but this was dropped when non-physical temperature oscillation was observed in volumes experiencing void boundaries crossing the junction)

2) Actual pump start times were developed from the test data to provide more accurate simulation of the start transient. For the purposes of this comparison, the pumps were assumed to start simultaneously. In the tests, the electrical knife switches were activated with a single bar to make the pump initiation as simultaneous as practical.

3) Void volumes of 5 and 8% were applied in the main header.

4) Actual piping runs to the gas separators were added, with pump suction taken from the side of the separator.

5) A control system and TDJ were added to maintain the supply tadnk at" constant elevation, consistent with the tests.

Model development calculations are provided in Appendix B. A diagram of the model is provided in Figure 3-1.

The model is exercised for a 100 second null transient prior to the pump start to enable stable pressures and void fractions to be attained. The.pumps are started and run for times comparable to the test. The model is exercised for 10 seconds beyond the pump stop to enable a rest condition to be achieved in the gas separators comparable to what was measured shortly after each test.

FAI/09-44R Page 13 of 51 Rev. 0 Date: 03/13/09 H r-------------

I-i D~

Fiue3zi U-M facill'Ity RELAPS m6deal dlagfam.

FAI/09-44R Page 14 of 51 Rev. 0 Date: 03/13/09 Cases 1, 9, 10, 11, and 12 were simulated with the RELAP5 model. As noted previously, the pre-test model was upgraded to add more detail, particularly with respect to physically modeling the gas separators and their non-condensable gas volumes. The effect of doing this is shown in Figure 3-2, which provides the flow velocity part way down the vertical drop from the main header and compares it to the measured flow on the line between the separator and the pump suction (note that velfj452 is the measured velocity, applied at a time dependent junction). The presence of a gas volume in the system clearly induces a dynamic response in the piping. The velocity oscillation resulting from the start transient can result in differences in the computed entrainment, and are the most likely reason we observed gas entrainment in the charging line that exceeded our expectations. In addition, it provides an explanation of the sinusoidal early oscillatory behavior we observed in the videos taken of this line, as well as the visual observation that virtually any gas entering the charging line did so very early in the test. The initial negative flow shown in Figure 3-2 is due to the dominance of the RH flow transient.

The predicted air transport as measured in the separators is compared to the actual measurements in Table 3-1. The RH gas separator proved to experience too much carryunder to provide a reliable measurement of gas transport to that loop so RH gas transport was calculated separately using a command file with APTPLOT. The RELAP model predicted this behavior as well. The pattern that emerges is that the code correctly predicts no SI entrainment, and tends to send all the gas down the RH 45 degree inclined pipe. In the absence of RH flow claiming the air (Tests 11 and 12), the entrainment in the charging line is predicted quite well and slightly on the conservative side. The RH gas transport for Cases 1, 9, and 10 are shown in Figure 3-3, Figure 3-4 and Figure 3-5. In case 1, virtually all the gas in the system is transported and fully expelled from the RH suction piping. In case 9, most of the gas in the system is transported into the RH suction piping, but a large percentage is held up in the suction line and returns to the system after pump trip. In case 10, a significant fraction of the gas in the system is pulled into the RH line and held up, but very little actually is entrained and transported through the system.

Figure 3-6 provides the charging line flow velocity for Case 12, the case with no RH pump running. As can be seen, a velocity oscillation occurs due to the effects of the gas separator. There is no initial negative flow, which confirms the hypothesis that the RH start transient causes an initial reverse flow in the charging piping. As noted above, RELAP5 does a very good job at predicting the entrainment and transport of gas for this case. Figure 3-7 shows the gas void fractions at the down-turning elbow in the charging suction line. The plot clearly shows that gas is being held up in the elbow. Figure 3-8 shows the liquid velocities for the same volumes. This shows that the liquid is running at higher speeds under the gas void at the elbow, supplying the flow required by the pump. This feature, sometimes referred to as a kinematic shock, was clearly seen in the test.

FAI/09-44R Page 15 of 51 Rev. 0 Date: 03/13/09 3.1 Supplemental Test Comparison Based on the results obtained in the test comparison and the observed impact of the gas-separator dynamic contributions, a decision was made to repeat some of the tests with the intent of more closely observing the transient level response of the gas separators. The base conditions present in Test 12 were selected (RH=0, Chg=21gpm, SIH=25gpm with an 8% initial header void). Three additional tests were run at these flows and the gas separator level behavior was confirmed to exhibit sinusoidal oscillation during pump start. The measured gas transport to the charging separator was very consistent with that observed in Test 12. Four additional tests were performed to allow additional data points at increased charging pump flow rates. Test 21 was run at a charging flow of 55 gpm (SIH at 25 gpm). At this flow rate, some gas bubbles were observed to entrain in the charging gas separator and be transported towards the pump. The next test point selected was with a charging flow rate of 35 gpm (SIH=25gpm).

This flow proved to remain within the capability of the gas separator to fully retain the gas entering the separator. This test condition was repeated twice more to establish repeatability.

(Tests 22-24). The tests run and observed gas transport are provided in Table 3-2.

The model was configured to reflect the flows of Test 21 and Test 24 and cases were run for comparison. The gas transport to the charging line separator comparison is provided in Table 3-3. For Test 21, the RELAP model predicted a small amount of carryunder from the gas separator, which was consistent with the behavior observed. The results from this case need to be treated with some circumspection, since there is no way to accurately compare the carryunder gas flows. There was no carryunder observed in the Test 24 case. The gas transport predicted compared very favorably with the test measured values.

FAI/09-44R Page 16 of 51 Rev. 0 Date: 03/13/09 Millstone 3 Test Case 1 Velocity in Chg line/Pump flow 6

C 0*

-2 100 110 120 130 Time (s)

Figure 3-2 Case 1 charging flow velocity.

FAI/09-44R Page 17 of 51 Rev. 0 Date: 03/13/09 Millstone 3 Test Case I Integrated Gas flow in RH suction Li Li.

400

-E

-J Li 0 t 0

I , I 50 i

100 I I 150 200 Time (s)

Figure 3-3 Case 1 RH line gas transport.

FAI/09-44R Page 18 of 51 Rev. 0 Date: 03/13/09 Millstone 3 Test Case 9 Integrated Gas Flow in RH Swuilion 400 0 50 100 150 200 Time Is!

Figure 3-4 Case 9 RH line gas transport.

Note: The downturn and decay to approximately 100 cu in reflects the return of gas to the 4 and 6 inch headers following pump trip.

FAI/09-44R Page 19 of 51 Rev. 0 Date: 03/13/09 Millstone 3 Test Case 10 IntegL-ated Gas Flow in RH Suction 600 500 400

- 300 200 100 0

0 50 100 150 200 Time (s)

Figure 3-5 Case 10 RH line gas transport.

Note: The downturn and decay to approximately 30 cu in reflects the return of gas to the 4 and 6 inch headers following pump trip.

FAI/09-44R Page 20 of 51 Rev. 0 Date: 03/13/09 Millstone 3 Test Case 12 Velocity in Chg line/pump flow 3

"C 2

-o 72

()

100 110 120 130 140 150 Time (s)

Figure 3-6 Test Case 12 charging line velocity.

FAI/09-44R Page 21 of 51 Rev. 0 Date: 03/13/09 Millstone 3 Test Case 12 Gas Void fraction at charging line elbow 0.7 0.6 0.5 00.4 0.3 0.2 0.1 90 100 110 120 130 140 130 160 Time (s)

Figure 3-7 Test Case 12 gas void fraction near charging line elbow.

FAI/09-44R Page 22 of 51 Rev. 0 Date: 03/13/09 Millstone 3 Test Case 12 Fluid Velocities at Charging line elbow 4

0 100 120 140 160 Time (s)

Figure 3-8 Test Case 12 fluid velocity near charging line elbow.

FAI/09-44R Page 23 of 51 Rev. 0 Date: 03/13/09 Table 3-1 Air Transport Result Comparison Case Chg (Meas./Pred) SIH Flows Modeled cu-in cu-in RH/SI/CHG 1 0-9/0 0/0 315/23/22 9 30-50/0 0/0 170/25/22 10 98.9/18.3 0/0 100/25/22 11 0/12.4 0/0 0/25/22 12 30-35/42 0/0 0/25/22 Table 3-2 Supplemental Tests Case Chg (Measured gas SIH Flows Modeled accumulation cu-in) cu-in RH/SI/CHG 18 29 0/0 025/21 19 34.6 0/0 0/25/21 20 29 0/0 0/25/21 21 393 0/0 0/25/55 22 185 0/0 0/25/35 23 231 0/0 0/25/35 24 185 0/0 0/25/35 Table 3-3 Supplemental Test Comparison Case Chg (Measured gas SIH Flows Modeled accumulation cu- cu-in RH/SI/CHG in)/Predicted 21 393/371 0/0 0/25/55 24 185/167 0/0 0/25/35

FAI/09-44R Page 24 of 51 Rev. 0 Date: 03/13/09

4.0 CONCLUSION

S A detailed model of the 1/4 scale Millstone 3 RWST suction piping test was prepared and exercised using the measured pump flow rates and known initial void conditions. The following observations are salient:

1) The model does a good job of predicting the entrainment in the Charging line in the absence of RH flows, and for a range of charging flows, demonstrating that RELAP correctly handles entrainment at downward elbows in pipe models.

2) The horizontal stratification vapor pull through model on the downward pointing SI takeoff matches the test observations of no gas entrainment in any of the test cases.

3) The horizontal stratification vapor pull through model on the 45 degree RH takeoff works well, and may be somewhat over-conservative in its prediction of gas pull through. The results match the test data in that RH was observed to entrain virtually all the gas available.

4) The additional dynamic behavior produced by the presence of the gas separators highlights the importance of capturing any such effects when performing analysis of gas transport in piping systems.

5) The results obtained in these comparisons support a conclusion that RELAP5 Mod 3.3gl (patch 03) demonstrates the ability to predict gas transport for this configuration with reasonable fidelity.

FAI/09-44R Page 25 of 51 Rev. 0 Date: 03/13/09

5.0 REFERENCES

FAI/09 (2009). "Test Results for the Millstone 3 Gas/Water Transport Tests". FAI/09-22:

Fauske & Associates, LLC.

FAI/09-44R Page 26 ,f 51 Rev. 0 Date: 03/13/09 APPENDIX A: PretestPredictionsfor Case 1 Pretest Predictions of the Millstone 3 FAI 1/ Scale Test Using RELAP5 Introduction A RELAP5 model was prepared to simulate the gas void transient in the FAI 1/4 scale experiment. The model was based on as built measurements of the test apparatus. Two cases with different initial void content in the 6 inch piping were run for the same pump flow combinations. The results obtained provide insight into what may be expected to occur in the test assembly.

Model Description A diagram of the RELAP5 model is attached. Generally, the piping was subdivided into node lengths such that the L/D ratio was approximately 1, in keeping with code developer guidelines. The horizontal stratified flow pull through model was applied for a 45 degree outtake condition for the RH line, and for a 90 degree outtake condition for the SI line. The CV line did not specifically apply the horizontal stratified pull through model. Time steps of 2 milliseconds were applied for both cases. The pump outflows were modeled as time dependent junctions with specified velocity condition vs time.

Case Description Two cases were performed: 5% initial void present in the 6 inch header, and 10% initial void present in the 6 inch header. The pump flows were computed based on 25 gpm each through the SI and Charging headers, and 310 gpm flow through the RHR line.

Results The model was initialized with the desired void fraction present in the 6 inch header. A linear ramp of pump flow was initiated at 1 second with the pumps reaching steady condition in one second. The runs were continued until the void was transported through the system.

Figure A-1 shows the void fraction at two locations in the 6 inch header for the 5% initial void case. Figure A-2 shows the void fractions exiting the TDJ's representing the pump outtakes.

As can be readily seen, the RHR line carries virtually all of the gas. A very minute amount of gas is transported in the charging header, on the order of 0.05%.

Figure A-3 provides the void fraction at two locations in the 6 inch header for the 10%

initial void case. Figure A-4 shows the void fractions exiting the TDJ's representing the pump outtakes. The model predicts virtually all the void entering the RHR header, with a small amount (less than 1% void fraction) entering the charging line.

FAI/09-44R Page 27 of 51 Rev. 0 Date: 03/13/09 Millstone 3 Pretest Prediction 5%initial void Void fractions in middle and end of 6 inch header 0.08 0.06 0.04-0.02 0

5 10 15 20 Time (s)

Figure A-1 Void fractions in 6 inch header 5% void case.

FAI/09-44R Page 28 of 51 Rev. 0 Date: 03/13/09 Millstone 3 Pretest Prediction 5% initial void Void fractions at exit TIDjs

-- oidg;-450000000

-voidgi-451000000 0.15

-voidgj-452000000 I

U 0.1H C

U 0.05F 0

0 10 15 20 Time (s)

Figure A-2 Void fractions exiting the time dependent junctions - 5% initial void case.

FAI/09-44R Page 29 of 51 Rev. 0 Date: 03/13/09 Millstone 3 Pretest Prediction 10% void Void fraction at middle and end of 6 inch header 0.2 0.15 0.1 0.05 0

0 10 15 20 25 30 Time (s)

Figure A-3 Void fractions in 6 inch header 10% void case.

FAI/09-44R Page 30 of 51 Rev. 0 Date: 03/13/09 Millstone 3 Pretest Prediction 10% initial Void Void fractions at e.sal TDJs 0.25) voidgi-450000000 voidgj-451000000 voidg j-4.52000000 Z0.2 o.15 0.

-5 C

.o 0.1 0

0.031 0 5 10 15 20 23 30 Time (s)

Figure A-4 Void fractions exiting the time dependent junctions- 10% initial void case.

FAI/09-44R Page 31 of 51 Rev. 0 Date: 03/13/09 APPENDIX B: Model Development Calculations FAI PROPRIETARY aa a

FAI/09-44R Page 32 of 51 Rev. 0 Date: 03/13/09 aa a

i FAI/09-44R Page 33 of 51 Rev. 0 Date: 03/13/09 aa a

FAI/09-44R Page 34 of 51 Rev. 0 Date: 03/13/09 aa a

FAI/09-44R Page 35 of 51 Rev. 0 Date: 03/13/09 aa a

FAI/09-44R Page 36 of 51 Rev. 0 Date: 03/13/09 aa a

FAI/09-44R Page 37 of 51 Rev. 0 Date: 03/13/09 aa a

FAI/09-44R Page 38 of 51 Rev. 0 Date: 03/13/09 APPENDIX C: RELAP5 Model Input Deck FAI PROPRIETARY bb b

FAI/09-44R Page 39 of 51 Rev. 0 Date: 03/13/09 I.b bb

FAI/09-44R Page 40 of 51 Rev. 0 Date: 03/13/09 bb b

FAI/09-44R Page 41 Qf 51 Rev. 0 Date: 03/13/09 bb b

FAI/09-44R Page 42 of 51 Rev. 0 Date: 03/13/09 bb b

FAI/09-44R Page 43 of 51 Rev. 0 Date: 03/13/09 bb b

FAI/09-44R Page 44 of 51 Rev. 0 Date: 03/13/09 bb b

FAI/09-44R Page 45 of 51 Rev. 0 Date: 03/13/09 bb b

FAI/09-44R Page 46 of 51 Rev. 0 Date: 03/13/09 bb b

FAI/09-44R Page 47 of 51 Rev. 0 Date: 03/13/09 bb b

FAI/09-44R Page 48 of 51 Rev. 0 Date: 03/13/09 bb b

FAI/09-44R Page 49 of 51 Rev. 0 Date: 03/13/09 bb b

FAI/09-44R Page 50 of 51 Rev. 0 Date: 03/13/09 bb b

FAI/09-44R Page 51 qf 51 Rev. 0 Date: 03/13/09 bb b