relap5/mod3.3 code manual volume iii: … · relap5/mod3.3 code manual volume iii: developmental...

RELAP5/MOD3.3 CODE MANUALVOLUME III:

DEVELOPMENTAL ASSESSMENT

PROBLEMS

Nuclear Systems Analysis Division

December 2001Information Systems Laboratories, Inc.

Rockville, MarylandIdaho Falls, Idaho

Prepared for the Division of Systems Research

Office of Nuclear Regulatory ResearchU. S. Nuclear Regulatory Commission

Washington, DC 20555

ABSTRACT

The RELAP5 code has been developed for best-estimate transient simulation of light water reactorcoolant systems during postulated accidents. The code models the coupled behavior of the reactor coolantsystem and the core for loss-of-coolant accidents and operational transients such as anticipated transientwithout scram, loss of offsite power, loss of feedwater, and loss of flow. A generic modeling approach isused that permits simulating a variety of thermal hydraulic systems. Control system and secondary systemcomponents are included to permit modeling of plant controls, turbines, condensers, and secondaryfeedwater systems.

In this volume the results of a wide range of test applications of RELAP5/MOD3.3 are presented forthe purpose of demonstrating the applicability of the code. The applications include results from tenphenomenological tests, nineteen separate effects tests, and five integral experiments. Thephenomenological problems are primarily used to determine whether the code produces qualitativelycorrect results; however two of the problems have analytical solutions that serve as quantitative tests aswell. The separate effects problems all provide quantitative as well as qualitative tests for correctsimulation. Each problem is selected to emphasize a particular physical effect that provides a test forcorrect functioning of a particular model or group of models. The integral problems provide tests forqualitative and, if data is available, quantitative correctness of all the code models working in concert.

The primary objective of the work reported in this edition of Volume III is to provide a comparisonof results obtained with the current version of RELAP5, MOD3.3 to results obtained using the previousversion MOD3.2. However, for those problems having analytical solutions or data, the MOD3.3 resultsare also compared to those data as well.

iii NUREG/CR-5535/Rev 1-Vol III

NUREG/CR-5535/Rev 1-Vol III iv

EXECUTIVE SUMMARY

EXECUTIVE SUMMARY

The light water reactor (LWR) transient analysis code, RELAP5, was originally developed at theIdaho National Engineering Laboratory (INEL) for the U.S. Nuclear Regulatory Commission (NRC).Code uses include analyses required to support rulemaking, licensing audit calculations, evaluation ofaccident mitigation strategies, evaluation of operator guidelines, and experiment planning analysis.RELAP5 has also been used as the basis for a nuclear plant analyzer. Specific applications have includedsimulations of transients in LWR systems such as loss of coolant, anticipated transients without scram(ATWS), and operational transients such as loss of feedwater, loss of offsite power, station blackout, andturbine trip. RELAP5 is a highly generic code that, in addition to calculating the behavior of a reactorcoolant system during a transient, can be used for simulation of a wide variety of hydraulic and thermaltransients in both nuclear and nonnuclear systems involving mixtures of steam, water, noncondensable,and solute.

The MOD3 version of RELAP5 has been developed jointly by the NRC and a consortium consistingof several foreign and domestic organizations that were members of the International Code Assessmentand Applications Program (ICAP) and its successor organization, the Code Applications and MaintenanceProgram (CAMP). Credit also needs to be given to various Department of Energy sponsors, including theINEL laboratory-directed discretionary funding program. The mission of the RELAP5/MOD3development program was to develop a code version suitable for the analysis of all transients andpostulated accidents in LWR systems, including both large- and small-break loss-of-coolant accidents(LOCAs) as well as the full range of operational transients.

The RELAP5/MOD3 code is based on a nonhomogeneous and nonequilibrium model for the two-phase system that is solved by a fast, semi-implicit numerical scheme to permit economical calculation ofsystem transients. The objective of the RELAP5 development effort from the outset was to produce a codethat included important first-order effects necessary for accurate prediction of system transients but thatwas sufficiently simple and cost effective so that parametric or sensitivity studies were possible.

The code includes many generic component models from which general systems can be simulated.The component models include pumps, valves, pipes, heat releasing or absorbing structures, reactor pointkinetics, electric heaters, jet pumps, turbines, separators, accumulators, and control system components. Inaddition, special process models are included for effects such as form loss, flow at an abrupt area change,branching, choked flow, boron tracking, and noncondensable gas transport.

The system mathematical models are coupled into an efficient code structure. The code includesextensive input checking capability to help the user discover input errors and inconsistencies. Alsoincluded are free-format input, restart, renodalization, and variable output edit features. These userconveniences were developed in recognition that generally the major cost associated with the use of asystem transient code is in the engineering labor and time involved in accumulating system data anddeveloping system models, while the computer cost associated with generation of the final result is usuallysmall.

The development of the models and code versions that constitute RELAP5 has spannedapproximately 23 years from the early stages of RELAP5 numerical scheme development to the present.RELAP5 represents the aggregate accumulation of experience in modeling reactor core behavior during

v NUREG/CR-5535/Rev 1-Vol III

EXECUTIVE SUMMARY

accidents, two-phase flow processes, and LWR systems. The code development has benefited fromextensive application and comparison to experimental data in the LOFT, PBF, Semiscale, NRU, and otherexperimental programs.

Several new models, improvements to existing models, and user conveniences were added toRELAP5/MOD2 to produce RELAP5/MOD3.0. Additional models, model improvements, and correctionshave been incorporated with the MOD3.2 and MOD3.3 releases. The primary focus of this document willbe the performance of the RELAP5/MOD3.3 version of the code in comparison to previous versions. Mostcomparisons will be to the MOD3.2 results and to the data or analytical solutions where they exist.

The new models that have been added since the conclusion of the MOD2 development include:

• Several counter-current flow limiting correlations that can be activated by the user at eachjunction in the system model.

• The ECCMIX component for modeling of the mixing of subcooled emergency corecooling system (ECCS) liquid and the resulting interfacial condensation.

• A zirconium-water reaction model to model the exothermic energy production on the

surface of zirconium cladding material at high temperature.

• A surface-to-surface radiation heat transfer model with multiple thermal radiationenclosures defined through user input.

• A level tracking model.

• A thermal stratification model.

Improvements to existing models include:

• New correlations for interfacial friction for all types of geometry in the bubbly-slug flow

regime in vertical flow passages.

• Use of junction-based interphase drag.

• An improved model for vapor pullthrough and liquid entrainment in horizontal pipes toobtain correct computation of the fluid state convected through a break.

• A new critical heat flux correlation for rod bundles based on tabular data.

• An improved horizontal stratification inception criterion for predicting the flow regime

transition between horizontally stratified and dispersed flow.

• A modified reflood heat transfer model.

NUREG/CR-5535/Rev 1-Vol III vi

EXECUTIVE SUMMARY

• Improved logic for vertical stratification inception to avoid excessive activation of the

water packing model.

• An improved boron transport model.

• A mechanistic separator/dryer model.

• An improved crossflow model.

• An improved form loss model.

• The addition of a simple plastic strain model with a clad burst criterion to the fuelmechanical model.

• The addition of a radiation heat transfer term to the gap conductance model.

• Modifications to the noncondensable gas model to eliminate erratic code behavior and

failure.

• Improvements to the downcomer penetration, ECCS bypass, and upper plenumdeentrainment capabilities.

• An improved equation of state that includes the meta-stable regions and usesthermodynamically consistent interpolation.

Additional user conveniences include:

• Modifications that place both the vertical stratification and water packing models underuser control so they can be deactivated.

• Removal of bit packing and vectorization to improve portability and readability.

• Computer portability through the conversion of the FORTRAN coding to adhere to the

FORTRAN 77 standard.

• Code execution and validation on a variety of systems. The code should be easily installed(i.e., the installation script is supplied with the transmittal) on the CRAY X-MP(UNICOS), DECstation 5000 (ULTRIX), DEC Alpha Workstation (OSF/1), IBM

Workstation 6000 (UNIX), SUN Workstation (UNIX), SGI Workstation (UNIX), and HPWorkstation (UNIX). The code has been installed (although the installation script is notsupplied with the transmittal) on the IBM 3090 (MVS) and IBM-PC (DOS). The code can

be installed easily on all 64-bit machines (integer and floating point operands) and any 32-bit machine that provides for 64-bit floating point.

vii NUREG/CR-5535/Rev 1-Vol III

EXECUTIVE SUMMARY

The RELAP5/MOD3 code manual consists of seven separate volumes. The modeling theory andassociated numerical schemes are described in Volume I, to acquaint the user with the modeling base andthus aid in effective use of the code. Volume II contains more detailed instructions for code application andspecific instructions for input data preparation. Both Volumes I and II are expanded and revised versions

of the RELAP5/MOD2 code manuala and Volumes I and III of the SCDAP/RELAP5/MOD2 code

manual.b

This volume, Volume III, presents the results of developmental assessment cases run with RELAP5/MOD3.3 to demonstrate and validate the models used in the code. The assessment matrix containsphenomenological problems, separate-effects tests, and integral systems tests.

Volume IV contains a detailed discussion of the models and correlations used in RELAP5/MOD3. Itpresents the user with the underlying assumptions and simplifications used to generate and implement thebase equations into the code so that an intelligent assessment of the applicability and accuracy of theresulting calculations can be made. Thus, the user can determine whether RELAP5/MOD3 is capable ofmodeling his or her particular application, whether the calculated results will be directly comparable tomeasurement or whether they must be interpreted in an average sense, and whether the results can be usedto make quantitative decisions.

Volume V provides guidelines for users that have evolved over the past several years fromapplications of the RELAP5 code at the Idaho National Engineering Laboratory, at other nationallaboratories, and by users throughout the world.

Volume VI discusses the numerical scheme in RELAP5/MOD3, and Volume VII is a collection ofindependent assessment calculations.

a. V. H. Ransom et al., RELAP5/MOD2 Code Manual, Volumes I and II, NUREG/CR-4312, EGG-2396, Idaho

National Engineering Laboratory, August 1995 and December 1985, revised March 1987.

b. C. M. Allison and E. C. Johnson, Eds., SCDAP/RELAP5/MOD2 Code Manual, Volume I: RELAP5 Code

Structure, System Models, and Solution Methods, and Volume III: User’s Guide and Input Requirements,

NUREG/CR-5273, EGG-2555, Idaho National Engineering Laboratory, June 1989.

NUREG/CR-5535/Rev 1-Vol III viii

ACKNOWLEDGMENTS

ACKNOWLEDGMENTS

Development of a complex computer code such as RELAP5 is the result of team effort and requires

the diverse talents of a large number of people. Special acknowledgment is given to those who pioneeredand continue to contribute to the RELAP5 code, in particular, V. H. Ransom, J. A. Trapp, and

R. J. Wagner. A number of other people have made and continue to make significant contributions to thecontinuing development of the RELAP5 code. Recognition and gratitude is given to the members of theINEL RELAP5 team:

V. T. Berta C. E. Lenglade R. A. Riemke

K. E. Carlson M. A. Lintner R. R. Schultz

C. D. Fletcher C. C. McKenzie A. S-L. Shieh

E. E. Jenkins G. L. Mesina R. W. Shumway

E. C. Johnsen C. S. Miller C. E. Slater

G. W. Johnsen G. A. Mortensen S. M. Sloan

J. M. Kelly P. E. Murray M. Warnick

H-H. Kuo R. B. Nielson W. L. Weaver

N. S. Larson S. Paik G. E. Wilson

The list of contributors is incomplete, as many others have made significant contributions in the past.

Rather than attempt to list them all and risk unknowingly omitting some who have contributed, weacknowledge them as a group and express our appreciation for their contributions to the success of theRELAP5 effort.

The list of contributors while the RELAP5 Program was at SCIENTECH, Inc., included:

Bill Arcieri Doug Barber Robert Beaton

Robert Copp Byron Hansen Scott Lucas

Glen Mortensen Dan Prelewicz Rex Shumway

Randy Tompot Weidong Wang

The current list of contributors to the RELAP5 Program at Information Systems Laboratories, Inc.,include:

Bill Arcieri Doug Barber Robert Beaton

Mark Bolander Don Fletcher Glen Mortensen

Dan Prelewicz Rex Shumway Victor Ransom

The RELAP5 Program is indebted to the technical monitors from the U. S. Nuclear Regulatory

Commission and the Department of Energy-Idaho Operations Office for giving direction and managementto the overall program. Those from the NRC include W. Lyon, Y. Chen, R. Lee, R. Landry, H. Scott,M. Rubin, D. E. Solberg, D. Ebert, S. Smith, T. Lee, V. Mousseau, and Weidong Wang. Monitors from

ix NUREG/CR-5535/Rev 1-Vol III

ACKNOWLEDGMENTS

DOE-ID when the RELAP5 program was at the INEL include N. Bonicelli, C. Noble, W. Rettig, and D.

Majumdar.

The technical editing of the RELAP5 manuals when the RELAP5 program was at the INEL wasdone by D. Pack and E. May.

Finally, acknowledgment is made of all the code users who have been very helpful in stimulating

timely correction of code deficiencies and suggesting improvements.

NUREG/CR-5535/Rev 1-Vol III x

CONTENTS

Page

1 INTRODUCTION.........................................................................................................................1

1.1 ADDED CAPABILITY OF RELAP5/MOD3.3 ..............................................................1

1.2 INTENDED APPLICATIONS OF RELAP5/MOD3.3....................................................3

1.3 KNOWN LIMITATIONS OF RELAP5/MOD3.3............................................................3

1.4 Developmental Assessment Objective .............................................................................4

2 DEVELOPMENTAL ASSESSMENT PROBLEMS....................................................................7

2.1 PHENOMENOLOGICAL PROBLEMS ....................................................................... 11

2.1.1 Nine-Volume Water Over Steam ....................................................................... 112.1.2 Nitrogen-Water Manometer Problem................................................................162.1.3 Branch Reentrant Tee Problem..........................................................................182.1.4 Cross-Flow Tee Problem ...................................................................................202.1.5 Cross Tank Problem ..........................................................................................212.1.6 Three-Stage Turbine ..........................................................................................262.1.7 Workshop Problem 2 .........................................................................................292.1.8 Workshop Problem 3 .........................................................................................302.1.9 Horizontally Stratified Countercurrent Flow ....................................................362.1.10 Pryor’s Pipe Problem .........................................................................................422.1.11 References .........................................................................................................48

2.2 SEPARATE - EFFECTS PROBLEMS...........................................................................49

2.2.1 Edwards Pipe Problem ......................................................................................492.2.2 Dukler Air-Water Flooding Tests ......................................................................512.2.3 Marviken Test 24...............................................................................................582.2.4 Marviken Test 22...............................................................................................612.2.5 LOFT Test L3-1 Accumulator Blowdown ........................................................652.2.6 Bennett’s Heated Tube Experiments..................................................................712.2.7 Royal Institute of Technology Tube Test 261 ...................................................732.2.8 ORNL Bundle Tests ..........................................................................................762.2.9 Christensen Subcooled Boiling Test 15.............................................................792.2.10 Shoukri Subcooled Flow Boiling and Condensation Test.................................812.2.11 MIT Pressurizer Test ST4..................................................................................832.2.12 FLECHT-SEASET Forced Reflood Tests .........................................................862.2.13 FLECHT-SEASET Boil Off Test 35658. ........................................................1042.2.14 Summary of Separate Effects Assessment ......................................................1052.2.15 References .......................................................................................................108

2.3 INTEGRAL TEST PROBLEMS ................................................................................. 111

2.3.1 LOFT Small-Break Test L3-7 ......................................................................... 1112.3.2 LOFT Large-Break Test L2-5 ......................................................................... 1182.3.3 Semiscale Natural Circulation Tests S-NC-2 and S-NC-3..............................1362.3.4 Zion-1 PWR Small Break................................................................................1452.3.5 References .......................................................................................................149

xi NUREG/CR-5535/Rev 1-Vol III

3 CONCLUSION .........................................................................................................................153

3.1 Phenomenological Problems ........................................................................................153

3.2 Separate Effects Problems............................................................................................154

3.3 INTEGRAL TEST PROBLEMS .................................................................................155

NUREG/CR-5535/Rev 1-Vol III xii

FIGURES

Page

2.1-1. RELAP5 Nodalization Diagram of the Nine-Volume Water Over Steam Problem...............................................................................................................12

2.1-2. The History of Void Distribution for the Nine-Volume Problem (Volume 1) ....122.1-3. The History of Void Distribution for the Nine-Volume Problem (Volume 3) ....132.1-4. The History of Void Distribution for the Nine-Volume Problem (Volume 5) ....132.1-5. The History of Void Distribution for the Nine-Volume Problem (Volume 7) ....142.1-6. The History of Void Distribution for the Nine-Volume Problem (Volume 9) ....142.1-7. Void Distribution in the Vertical Pipe at Various Times as Calculated by

MOD3.2 ..............................................................................................................152.1-8. Void Distribution in the Vertical Pipe at Various Times as Calculated by

MOD3.3 ..............................................................................................................152.1-9. RELAP Nitrogen-Water Manometer Nodalization.............................................162.1-10. Liquid Velocity at the Bottom of the Manometer (junction 777110000) ...........172.1-11. MOD3.2 and MOD3.3 Calculated Water Level Comparison for the

Manometer Problem ..........................................................................................182.1-12. Nodalization of the Branch Tee Problem Using Two Non-Sink Junctions ........192.1-13. Nodalization Diagram for the Cross Flow Tee Problem.....................................212.1-14. Calculated Mass Error Comparison for the Crossflow Tee Problem..................222.1-15. RELAP5 Cross Tank Nodalization Diagram ......................................................222.1-16. Mass Flow Rate Comparison: Pipe 200, Volume 14 (Liquid Region) ...............232.1-17. Mass Flow Rate Comparison: Pipe 400, Volume 14 (Liquid Region) ...............242.1-18. Mass Flow Rate: Cross Flow Junction 315 (Connecting Pipe 200 to Pipe 400

in the Liquid Region) ..........................................................................................242.1-19. Mass Flow Rate Comparison: Pipe 200, Volume 18 (Gas Region)....................252.1-20. Mass Flow Rate Comparison: Pipe 400, Volume 18 (Gas Region)....................252.1-21. Mass Flow Rate: Cross Flow Junction 319 (Connecting Pipe 200 to Pipe 400

in the Gas Region) ..............................................................................................262.1-22. Nodalization Used for the Three-stage Group Turbine Problem........................272.1-23. Mass Error Comparison for the Three-stage Turbine Problem...........................282.1-24. Relap5 Nodalization Diagram for Workshop Problems 2 and 3.........................292.1-25. Pressurizer Pressure History of Workshop Problem 2........................................302.1-26. Core Outlet Pressure History of Workshop Problem 2.......................................312.1-27. Steam Generator Steam Dome Pressure History of Workshop Problem 2 .........312.1-28. Steam Generator Secondary Side Liquid Level for Workshop Problem 2 .........322.1-29. Mass Flow in the Primary Loop Hot Leg for Workshop Problem 2...................322.1-30. Mass Flow in the Primary Side of Steam Generator for Workshop

Problem 2............................................................................................................332.1-31. Reactor Core Pressure Response for Workshop Problem 3................................34

xiii NUREG/CR-5535/Rev 1-Vol III

2.1-32. Pressurizer Pressure Response for Workshop Problem 3 ...................................342.1-33. Mass Flow in the Reactor Core for Workshop Problem 3 ..................................352.1-34. Mass Flow in the Pressurizer for Workshop Problem 3 .....................................352.1-35. Mass Flow in the Steam Generator Secondary Side (Riser) for Workshop

Problem 3............................................................................................................362.1-36. Mass Flow in the Steam Discharge Line for Workshop Problem 3....................372.1-37. Steam Generator Steam Dome Pressure Response for Workshop Problem 3 ....372.1-38. Relap5 Nodalization Diagram for a Horizontally Stratified Countercurrent

Flow Problem......................................................................................................382.1-39. Relap5-Calculated Junction Liquid Velocity at the Left End .............................392.1-40. Relap5-Calculated Junction Liquid Velocity at the Mid-section ........................392.1-41. Relap5-Calculated Junction Liquid Velocity at the Right End ...........................402.1-42. Relap5-Calculated Junction Vapor Velocity at the Left End ..............................402.1-43. Relap5-Calculated Junction Vapor Velocity at the Mid-Section ........................412.1-44. Relap5-Calculated Junction Vapor Velocity at the Right End ............................412.1-45. Relap5 Nodalization Diagram for Pryor’s Pipe Problem ...................................422.1-46. Void Fraction Response for Volume 2 of the Pryor’s Pipe Problem...................432.1-47. Void Fraction Response for Volume 4 of the Pryor’s Pipe Problem...................432.1-48. Void Fraction Response for Volume 6 of the Pryor’s Pipe Problem...................442.1-49. Void Fraction Response for Volume 8 of the Pryor’s Pipe Problem...................442.1-50. Void Fraction Response for Volume 10 of the Pryor’s Pipe Problem.................452.1-51. Volume Pressure Response for Volume 2 of the Pryor’s Pipe Problem..............452.1-52. Volume Pressure Response for Volume 4 of the Pryor’s Pipe Problem..............462.1-53. Volume Pressure Response for Volume 6 of the Pryor’s Pipe Problem..............462.1-54. Volume Pressure Response for Volume 8 of the Pryor’s Pipe Problem..............472.1-55. Volume Pressure Response for Volume 10 of the Pryor’s Pipe Problem............472.2-1. Relap5 Hydrodynamic Nodalization for Edwards’ Pipe Experiment.................492.2-2. Pressure Comparison at Left Section of Edwards Pipe Blowdown

Experiment..........................................................................................................502.2-3. Vapor Void Fraction Comparison at Left Section of Edwards Pipe Blowdown

Experiment..........................................................................................................502.2-4. Pressure Comparison at Left Section of Edwards Pipe Blowdown

Experiment, Heavy Water Medium ....................................................................512.2-5. Vapor Void Fraction Comparison at Left Section of Edwards Pipe Blowdown

Experiment, Heavy Water Medium ....................................................................522.2-6. Pressure Comparison at Left Section of Edwards Pipe Blowdown Experiment,

Nearly Implicit Advancement Scheme ...............................................................522.2-7. Vapor Void Fraction Comparison at Left Section of Edwards Pipe Blowdown

Experiment, Nearly Implicit Advancement Scheme ..........................................532.2-8. Edwards Pipe Blowdown Pipe Transient - Restart at 0.1 Second, Mod3.3........532.2-9. Schematic of the Dukler Air/water Test Facility ................................................54

NUREG/CR-5535/Rev 1-Vol III xiv

2.2-10. Relap5 Nodalization for Dukler’s Air/water Test Problem ................................552.2-11. Data for the Liquid Downflow Rate Versus Air Flow Injection Rate for Dukler’s

Air/Water Problem ..............................................................................................562.2-12. Measured and Mod3.2 Calculated Liquid Downflow Comparison for Dukler’s



Air/Water Problem With Gas Constant and Slope Fitted to Data.......................582.2-15. Schematic, Nodalization and Initial Temperature Profile for Marviken

Test 24................................................................................................................602.2-16. Measured and Calculated Pressure in the Top of the Vessel for Marviken

Test 24.................................................................................................................602.2-17. Measured and Calculated Mass Flow Rate at the Nozzle for Marviken

Test 24.................................................................................................................622.2-18. Measured and Calculated Density in the Middle of the Discharge Pipe for

Marviken Test 24 ................................................................................................622.2-19. Measured and Calculated Pressure (Including Phase Slip at Choked

Conditions) in the Top of the Vessel for Marviken Test 24 ................................632.2-20. Measured and Calculated Mass Flow Rate (Including Phase Slip at Choked

Conditions) at the Nozzle for Marviken Test 24.................................................632.2-21. Measured and Calculated Density (Including Phase Slip at Choked

Conditions) in the Middle of the Discharge Pipe for Marviken Test 24.............642.2-22. Schematic, Nodalization and Initial Temperature Profile for Marviken

Test 22................................................................................................................652.2-23. Measured and Calculated Pressure in the Top of the Vessel for Marviken

Test 22.................................................................................................................662.2-24. Measured and Calculated Mass Flow Rate at the Nozzle for Marviken

Test 22.................................................................................................................662.2-25. LOFT L3-1 Accumulator A and Surgeline Schematic .......................................672.2-26. RELAP5 LOFT L3-1 Accumulator Model Schematic .......................................682.2-27. Measured and Calculated Accumulator Gas Dome Pressure Versus Volume,

Loft Test L3-1 .....................................................................................................682.2-28. Measured and Calculated Accumulator Gas Dome Pressure, Loft Test L3-1 ....692.2-29. Measured and Calculated Accumulator Liquid Level, LOFT Test L3-1............692.2-30. Measured and Calculated Accumulator Gas Temperature, LOFT Test L3-1 .....702.2-31. RELAP5 Nodalization Diagram for Bennett’s Heated Tube Experiment ..........732.2-32. Measured and Calculated Axial Wall Temperature Profiles for Bennett’s

Heated Tube Low Mass Flux Experiment - Test 5358 .......................................742.2-33. Measured and Calculated Axial Wall Temperature Profiles for Bennett’s

Heated Tube Intermediate Mass Flux Experiment - Test 5294 ..........................742.2-34. Measured and Calculated Axial Wall Temperature Profiles for Bennett’s

Heated Tube High Mass Flux Experiment - Test 5394.......................................75

xv NUREG/CR-5535/Rev 1-Vol III

2.2-35. Measured and Calculated Critical Heat Flux Position for the Royal Institute of Technology Tube Test 261 ..................................................................................76

2.2-36. Measured and Calculated Surface Temperature for Ornl Bundle CHF Test 3.07.9B ........................................................................................................77

2.2-37. Measured and Calculated Surface Temperature for Ornl Bundle CHF Test 3.07.9N........................................................................................................78

2.2-38. Measured and Calculated Surface Temperature for Ornl Bundle CHF Test 3.07.9W .......................................................................................................78

2.2-39. Measured and Calculated Axial Void Fractions for the Ornl Void Profile Test ......................................................................................................................79

2.2-40. Measured and Calculated Rod Temperature for the Ornl Void Profile Test .......802.2-41. Measured and Calculated Steam Temperature for the Ornl Void Profile Test ....802.2-42. Measured and Calculated Axial Void Fractions for the Christensen Subcooled

Boiling Test 15....................................................................................................812.2-43. Measured and Calculated Axial Void Fractions for the Shoukri 3c

Experiment..........................................................................................................822.2-44. Schematic of the Experimental Apparatus for the Mit Pressurizer Test .............832.2-45. Measured and Calculated Rate of Pressure Rise for the Mit Pressurizer Test....842.2-46. Calculated Vapor Generation Rate for Volume 3080000, MIT Test ST4 ...........852.2-47. Tank fluid and Inside Wall Temperature at 35s into the MIT Pressurizer Test...862.2-48. Tank Fluid and Inside Wall Temperature at 35s into the Transient During the

MIT Pressurizer Test with the Thermal Front Tracking Model Active ..............872.2-49. Measured and Calculated Rate of Pressure Rise for the MIT Pressurizer Test,

Thermal Front Tracking Model Active...............................................................872.2-50. RELAP5 Nodalization for the FLECHT-SEASET Forced Reflood Tests..........882.2-51. Measured and Calculated Rod Surface Temperature Histories for

FLECHT-SEASET Forced Reflood Run 31504 at the 0.61 m (2 ft) Elevation .............................................................................................................89

2.2-52. Measured and Calculated Rod Surface Temperature Histories for FLECHT-SEASET Forced Reflood Run 31504 at the 1.22 m (4 ft) Elevation .............................................................................................................89



2.2-55. Measured and Calculated Rod Surface Temperature Histories for FLECHT-SEASET Forced Reflood Run 31504 at the 2.85 m (9.25 ft) Elevation .............................................................................................................91

2.2-56. Measured and Calculated Rod Surface Temperature Histories for FLECHT-SEASET Forced Reflood Run31504 at the 3.08 m (10 ft)

NUREG/CR-5535/Rev 1-Vol III xvi

Elevation .............................................................................................................912.2-57. Measured and Calculated Rod Surface Temperature Histories for

FLECHT-SEASET Forced Reflood Run31504 at the 3.38 m (11 ft) Elevation .............................................................................................................92

2.2-58. Measured and Calculated Steam Temperatures for FLECHT-SEASET Forced Reflood Run 31504 at 1.23 m (4 ft) Elevation ...................................................93



2.2-61. Measured and Calculated Steam Temperatures for FLECHT-SEASET Forced Reflood Run 31504 at 2.85 m (9.25 ft) Elevation ..............................................94

2.2-62. Measured and Calculated Steam Temperatures for FLECHT-SEASET Forced Reflood Run 31504 at 3.08 m (10 ft) Elevation .................................................95

2.2-63. Measured and Calculated Steam Temperatures for FLECHT-SEASET Forced Reflood Run 31504 at 3.54 m (11 ft) Elevation..................................................95

2.2-64. Measured and Calculated Total Bundle Mass Inventory for FLECHT-SEASET Forced Reflood Run 31504.................................................................................96

2.2-65. Measured and Calculated Void Fractions at 0.92 to 1.23 m (3 to 4 ft) for FLECHT-SEASET Forced Reflood Run 31504 .................................................96





2.2-70. Measured and Calculated Axial Void Profile at 100s for FLECHT-SEASET Forced Reflood Run 315504...............................................................................99

2.2-71. Measured and Calculated Axial Void Profile at 200s for FLECHT-SEASET Forced Reflood Run 31504.................................................................................99

2.2-72. Measured and Calculated Axial Void Profile at 300s for FLECHT-SEASET Forced Reflood Run 31504...............................................................................100

2.2-73. Measured and Calculated Rod Surface Temperatures for FLECHT-SEASET Forced Reflood Run 31701 0.62 m (2 ft) Elevation .........................................101

2.2-74. Measured and Calculated Rod Surface Temperatures for FLECHT-SEASET Forced Reflood Run 31701 1.22 m (4 ft) Elevation .........................................101

2.2-75. Measured and Calculated Rod Surface Temperatures for FLECHT-SEASET Forced Reflood Run 31701 at the 1.83 m (6 ft) Elevation ...............................102

2.2-76. Measured and Calculated Rod Surface Temperatures for FLECHT-SEASET Forced Reflood Run 31701 at the 2.46 m (8 ft) Elevation ...............................102

xvii NUREG/CR-5535/Rev 1-Vol III

2.2-77. Measured and Calculated Rod Surface Temperatures for FLECHT-SEASET Forced Reflood Run 31701 at the 3.08 m (10 ft) Elevation .............................103

2.2-78. Measured and Calculated Total Bundle Mass Inventory for FLECHT-SEASET Forced Reflood Run 31701 ...............................................103

2.2-79. Measured and Calculated Void Fraction History at the 0 to 1ft Level for Test 35658.........................................................................................................104




2.2-83. Measured and Calculated Void Fraction History at the 3 to 4ft level for Test 35658.........................................................................................................107


2.2-85. Measured and Calculated void fraction history at the 5 to 6ft level for Test 35658.........................................................................................................108

2.3-1. Schematic of LOFT Test Facility......................................................................1122.3-2. RELAP5 Nodalization for LOFT Test L3-7: Vessel and Broken Loop............1142.3-3. RELAP5 Nodalization for LOFT Test L3-7: Intact Loop ................................1152.3-4. CPU Time Versus Simulated Time for LOFT Test L3-7 ..................................1162.3-5. Measured and Calculated Primary System Pressure for LOFT Test L3-7........1172.3-6. Measured and Calculated Secondary System Pressure for LOFT Test L3-7....1172.3-7. Measured and Calculated Liquid Velocity in the Intact Loop Hot Leg for

LOFT Test L3-7 ................................................................................................1182.3-8. Measured and Calculated Vapor Velocity in the Intact Loop Hot Leg for

LOFT Test L3-7 ................................................................................................1192.3-9. Measured and Calculated Liquid Temperature at the Core Inlet for

LOFT Test L3-7 ................................................................................................1192.3-10. Measured and Calculated Liquid Temperature at the Core Outlet for LOFT

Test L3-7 ...........................................................................................................1202.3-11. Measured and Calculated Mass Flow Rate at the Break for LOFT

Test L3-7 ...........................................................................................................1202.3-12. Measured and Calculated Density in the Intact Loop Hot Leg for LOFT

Test L3-7 ...........................................................................................................1212.3-13. RELAP5 Nodalization for LOFT Test L2-5: Vessel and Broken Loop............1232.3-14. RELAP5 Nodalization for LOFT Test L2-5: Intact Loop ................................1242.3-15. Measured and Calculated Primary System Pressure for LOFT Test L2-5........1252.3-16. Measured and Calculated Secondary System Pressure for LOFT Test L2-5....1252.3-17. Measured and Calculated Mass Flow Rate in the Broken Loop Cold Leg for

LOFT Test L2-5 ................................................................................................126

NUREG/CR-5535/Rev 1-Vol III xviii

2.3-18. Measured and Calculated Mass Flow Rate in the Broken Loop Hot Leg for LOFT Test L2-5 ................................................................................................126

2.3-19. Measured and Calculated Mass Flow Rate in the Intact Loop Cold Leg for LOFT Test L2-5 ................................................................................................127

2.3-20. Measured and Calculated Mass Flow Rate in the Intact Loop Hot Leg for LOFT Test L2-5 ................................................................................................127

2.3-21. Measured Intact Loop Hot Leg Differential Pressure; LOFT Test L2-5 ..........1282.3-22. Absolute Value of Measured and Calculated Mass Flow Rate in the Intact Loop

Hot Leg for LOFT Test L2-5 ............................................................................1282.3-23. Measured and Calculated Density in the Intact Loop Hot Leg for LOFT

Test L2-5 ...........................................................................................................1292.3-24. Measured and Calculated Accumulator Level for LOFT Test L2-5.................1302.3-25. Calculated Mass Flow Rate from the Accumulator for LOFT Test L2-5.........1302.3-26. Measured and Calculated Pump Speed for Primary Coolant Pump 2

(Hydrodynamic Volume 165) for LOFT Test L2-5 ..........................................1312.3-27. Measured and Calculated Upper Plenum Temperature Below the Nozzle

for LOFT Test L2-5 ..........................................................................................1322.3-28. Measured and Calculated Lower Plenum Temperature for LOFT Test L2-5 ...1322.3-29. Measured and Calculated Fuel Centerline Temperature 0.69m (27in.) Above

the Bottom of the Core for LOFT Test L2-5.....................................................1332.3-30. Measured and Calculated Fuel Cladding Temperature 0.13m (5in.) Above



the Bottom of the Core for LOFT Test L2-5.....................................................1342.3-33. Measured and Calculated Fuel Cladding Temperature 0.99m (39in.) Above the

Bottom of the Core for LOFT Test L2-5 ..........................................................1352.3-34. Measured and Calculated Fuel Cladding Temperature 1.37m (54in.) Above


the Bottom of the Core for LOFT Test L2-5.....................................................1362.3-36. Semiscale Mod-2A Single-loop Configuration ................................................1382.3-37. Schematic of RELAP5/MOD3 Natural Circulation Test Model ......................1392.3-38. Measured and Calculated Primary System Mass Flow Rate at the 60kW

Core Power for Test S-NC-2.............................................................................1402.3-39. Measured and Calculated Primary System Hot Leg Fluid Temperature at

the 60kW Core Power for Test S-NC-2 ............................................................1412.3-40. Measured and Calculated Primary Side Steam Generator Outlet Fluid

Temperature at the 60kW Core Power for Test S-NC-2 ...................................1412.3-41. Measured and Calculated Primary System Pressure at the 60kW Core

Power for Test S-NC-2......................................................................................142

xix NUREG/CR-5535/Rev 1-Vol III

2.3-42. Measured and Calculated Primary System Mass Flow Rate Versus Steam Generator Secondary Side Heat Transfer Area for Test S-NC-3 ......................143

2.3-43. Measured and Calculated Primary Side Hot Leg Fluid Temperature Versus Steam Generator Secondary Side Heat Transfer Area for Test S-NC-3 ......................143

2.3-44. Measured and Calculated Primary Side Steam Generator Outlet Temperature Versus Steam Generator Secondary Side Heat Transfer Area for Test S-NC-3 ......................................................................................................144

2.3-45. Measured and Calculated Primary System Pressure Versus Steam Generator Secondary Side Heat Transfer Area for Test S-NC-3 .......................................144

2.3-46. RELAP5 Nodalization for the Zion-1 PWR: Vessel Model .............................1462.3-47. RELAP5 Nodalization for the Zion-1 PWR: Loop Model ...............................1472.3-48. MOD3.2 and MOD3.3 Primary System Pressure (Core Outlet) Comparison

for a Small Break Transient in a Typical PWR.................................................1482.3-49. MOD3.2 and MOD3.3 Break Mass Flow Rate Comparison for a Small Break

Transient in a Typical PWR ..............................................................................1492.3-50. MOD3.2 and MOD3.3 Intact Loop Accumulator Liquid Volume Comparison

for a Small Break Transient in a Typical PWR.................................................1502.3-51. MOD3.2 and MOD3.3 Broken Loop Accumulator Liquid Volume Comparison

for a Small Break Transient in a Typical PWR.................................................1502.3-52. MOD3.2 and MOD3.3 Core Outlet Void Fraction Comparison for a Small

Break Transient in a Typical PWR ...................................................................1512.3-53. MOD3.2 and MOD3.3 Intact Loop Steam Generator System Pressure

(Steam Dome) Comparison for a Small Break Transient in a Typical PWR....1512.3-54. MOD3.2 and MOD3.3 Mass Error Comparison for a Small Break Transient

in a Typical PWR..............................................................................................152

NUREG/CR-5535/Rev 1-Vol III xx

TABLES

Page

2.0-1. Developmental Assessment Matrix ......................................................................72.1-1. Results for the Branch Tee Problem ...................................................................192.1-2. Results for the Crossflow Tee Problem ..............................................................202.1-3. Turbine Parameters for the Assessment Problem ...............................................272.2-1. RELAP5/MOD3 Nonequilibrium Model Developmental Assessment Matrix ..71

xxi NUREG/CR-5535/Rev 1-Vol III

NUREG/CR-5535/Rev 1-Vol III xxii

INTRODUCTION

1 INTRODUCTION

The RELAP5 computer code is a light water reactor transient analysis code developed for the U.S.Nuclear Regulatory Commission (NRC) for use in rulemaking, licensing audit calculations, evaluation ofoperator guidelines, and as a basis for a nuclear plant analyzer. Specific applications of this capabilityhave included simulations of transients in LWR systems, such as loss of coolant, anticipated transientswithout scram (ATWS), and operational transients such as loss of feedwater, loss of offsite power, stationblackout, and turbine trip. RELAP5 is a highly generic code that, in addition to calculating the behavior ofa reactor coolant system during a transient, can be used for simulation of a wide variety of hydraulic andthermal transients in both nuclear and nonnuclear systems involving steam-water noncondensable solutefluid mixtures.

The MOD3.0 version of RELAP5 was developed jointly by the NRC and a consortium consisting ofseveral of the countries and domestic organizations that are members of the International Code Assessmentand Applications Program (ICAP). The mission of the RELAP5/MOD3.0 development program was todevelop a code version suitable for the analysis of all transients and postulated accidents in PWR systems,including both large- and small-break loss-of-coolant accidents (LOCAs), as well as the full range ofoperational transients. Although the emphasis of the RELAP5/MOD3.0 development was on large-breakLOCAs, improvements were made to existing code models based on the results of assessments againstsmall-break LOCAs and operational transient test data.

Since the completion of the RELAP5/MOD3.0 development, error correction, refinement, andimprovement of model has continued. These improvements have been incrementally included in theMOD3.x releases of the code. RELAP5/MOD3.3 is the latest in this series and this version is the focus ofthis developmental assessment.

RELAP5/MOD3.3 contains new and improved modeling capability as well as additional userconveniences compared to RELAP5/MOD2. The purpose of this volume is to document thedevelopmental assessment problems performed using RELAP5/MOD3.3. The remainder of this sectionbriefly describes the newly developed features and improvements in RELAP5/MOD3.3, the intendedapplication of the code, and its known limitations. Section 2 presents the developmental assessmentproblems. The problems are divided into three categories: phenomenological problems (Section 2.1);separate effects-problems (Section 2.2); and integral problems (Section 2.3). Finally, Section 3 presentsconclusions drawn from the assessment.

1.1 ADDED CAPABILITY OF RELAP5/MOD3.3

The added capability in RELAP5/MOD3.3 since the release of RELAP5/MOD2 includes newmodeling, improvements to existing models, and new user conveniences. A detailed description of theMOD3.3 capabilities can be found in Volumes I and II of the RELAP5/MOD3.3 code manual.

The new models that have been added since the initiation of the RELAP5/MOD3 developmentinclude:

• A counter-current flow limiting model that uses correlations, which are based on actual geometry and can be activated by the user at each junction in the system model.

1 NUREG/CR-5535/Rev 1-Vol III

INTRODUCTION

• The ECCMIX component for modeling of the mixing of subcooled emergency core cooling sys-tem (ECCS) liquid and the resulting interfacial condensation.

• A zirconium-water reaction model to model the exothermic energy production on the surface of zirconium cladding material at high temperature.

• A surface to surface radiation heat transfer model with multiple radiation enclosures defined through user input.

• A level tracking model.

• A thermal stratification model.

Improvements to existing models include:• New correlations for interfacial friction for all types of geometry in the bubbly-slug flow regime in

vertical flow passages.

• Use of junction based interphase drag.

• An improved model for vapor pull-through and liquid entrainment in horizontal pipes to obtain correct computation of the fluid state convected through a break.

• A new critical heat flux correlation for rod bundles based on tabular data.

• An improved horizontal stratification inception criterion for predicting the flow regime transition between horizontally stratified and dispersed flow.

• A modified reflood heat transfer model.

• Improved logic for vertical stratification inception to avoid excessive activation of the water pack-ing model.

• An improved boron transport model.

• A mechanistic separator/dryer model.

• An improved crossflow model.

• An improved form loss model.

• The extension of water packing logic to horizontal volumes.

• A new default critical flow model (Henry-Fauske).

• The addition of a simple plastic strain model with clad burst criterion to the fuel mechanical model.

• The addition of a radiation heat transfer term to the gap conductance model.

• Modifications to the noncondensable gas model to eliminate erratic code behavior and failure.

NUREG/CR-5535/Rev 1-Vol III 2

INTRODUCTION

• Improvements to the downcomer penetration, ECCS bypass, and upper plenum deentrainment capabilities.

• An improved equation of state that includes the meta-stable regions and uses thermodynamically consistent interpolation.

Additional user conveniences include:• Modifications that place both the vertical stratification and water packing models under user con-

trol so they can be deactivated.

• Removal of bit packing and vectorization to improve portability and readability.

• Computer portability through the conversion of the FORTRAN coding to adhere to the FORTRAN 77 standard.

• Code execution and validation on a variety of systems. The code should be easily installed (i.e., the installation script is supplied with the transmittal) on the CRAY X-MP (UNICOS), DEC station 5000 (ULTRIX), DEC Alpha Workstation (OSF/1), IBM Workstation 6000 (UNIX), SUN Work-station (UNIX), SGI Workstation (UNIX), and HP Workstation (UNIX). The code has been installed (although the installation script is not supplied with the transmittal) on the IBM 3090 (MVS) and IBM-PC (DOS). The code can be installed easily on all 64-bit machines (integer and floating point operands) and any 32-bit machine that provides for 64-bit floating point.

1.2 INTENDED APPLICATIONS OF RELAP5/MOD3.3

RELAP5/MOD3.3 is designed for use in the analysis of pressurized water reactor transients resultingfrom large- and small-break loss-of-coolant accidents and operational transients such as anticipatedtransients without scram, loss of feedwater, and turbine trip. Both primary and secondary systems,including balance of plant components, can be modeled. The code has generic modeling capability so thatseparate effects experiments can also be modeled for use in the assessment of code capability and forextrapolation of separate effects results to integral system behavior.

The modeling philosophy to be followed in using RELAP5/MOD3.3 is the same as for MOD2 exceptwhere new modeling capability requires special treatment. In general, all modeling capability present inthe MOD2 version of the code has been retained in the MOD3 versions, and it has been an objective tokeep input decks which have been developed for MOD2 compatible with MOD3.3. With a fewexceptions, this has been achieved.

1.3 KNOWN LIMITATIONS OF RELAP5/MOD3.3

This discussion of known limitations of RELAP5/MOD3.3 is based on experiences with previousversions, the developmental assessment work reported herein, and known or suspected limitations of themodels used in the code.

The Henry-Fauske critical flow model does not consider slip between the phases as it is currentlycoded. This can lead to an over-prediction of the mass flow rate under critical flow conditions. Inaddition, the Henry-Fauske model is not coded in the nearly implicit numerical scheme.


INTRODUCTION

The reflood model developed for RELAP5/MOD3 has shown good agreement with nonuniformlyheated rod bundle data with respect to time to maximum temperature. The liquid entrainment appears tobe about right, since the liquid inventory in the core is predicted well. However, the axial drag distributionis not correct because an inverted void profile develops at low reflood rates. This results in quench frontvelocity stalling at about core midplane.

The code tends to over-calculate the interfacial heat transfer in mist flow. In the case where thedroplet size is small, the surface area is large and the heat transfer between the two phases is over-stated.

The energy dissapation due to form loss is not a default in RELAP5. It was included in earlierversions of the code, but was removed because of several problems it caused. It is now a users option(option 41).

The pump model does not have an integral model to represent cavitation. The pump model doesallow the user the option of accounting for the cavitation or two-phase degradation effects on pumpperformance through the input of homologous, two-phase curves for head and torque, curves that are in theform of difference curves. This approach is not generally acceptable, particularly when correlations for agiven pump are available for determining the cavitating pump head as a function of available net positivesuction head. Users have resorted to using the RELAP5 control system, or hardwired updates for aparticular pump in order to base the pump head on the net positive suction head.

The code has only approximate two- and three-dimensional capability, which must be invoked byusing cross-flow junctions to cross connect a matrix of volumes. This approach appears to be adequate forfriction-dominated cases. This method is approximate, in that all convected momentum terms areneglected at the cross flow junctions and the primary flow direction only includes the axially convectedmomentum terms.

For horizontal flow, the plug flow regime is not generally available in the code. It is only present inthe ECC mixer component.

The reactor kinetics model uses point kinetics. There are situations where one- and three-dimensional capability is required but is not available in the code. To handle these situations, the PARCScode has been coupled to RELAP5/MOD3.3 to provide 1- and 3-D neutronic effects.

1.4 Developmental Assessment Objective

The objective of the developmental assessment is to determine the qualitative and quantitativeaccuracy of the code for problems that are consistent with the intended application of the code. This isaccomplished using three types of problems: phenomenological problems, modeling of separate effectsexperiments, and modeling of integral experiments. The phenomenological problems are used todemonstrate that the code is in qualitative agreement with the physics of the problem and in cases whereanalytical solutions exist, the qualitative accuracy of the code can be judged as well.

The separate effects tests are designed to provide data on a primary physical effect. These problemsare selected to test a key model or models of the code. Qualitative agreement with the data is the first


INTRODUCTION

criteria that must be satisfied, i.e. the correct trends are predicted. Given this, then the code results can bequantitatively compared to the data.

The integral problems provide evidence that the collection of models in the code function in concert.Code predictions of intergral system parameters, such as pressure, clad temperature, and mass inventory,are used to assess overall accuracy of the code.

The scope of this developmental assessment is limited to the suite of test problems that werepreviously used to assess the RELAP5 MOD2 and MOD3.0 versions of the code, and that are notproprietary. This suite of problems provides tests of the major code models. While beyond the scope ofthis effort, a PIRT analysis to select a suite of problems covering all significant models of the code wouldimprove the level of confidence relative to the general accuracy of the code.


INTRODUCTION


DEVELOPMENTAL ASSESSMENT PROBLEMS

2 DEVELOPMENTAL ASSESSMENT PROBLEMS

A total of 34 developmental assessment calculations were performed using RELAP5/MOD3.3.

Table 2.0-1 shows the developmental assessment matrix, which includes a brief description of theobjective of each problem. The matrix contains 10 phenomenological problems, 19 separate-effects

problems, and 5 integral test problems. The phenomenological problems are presented in Section 2.1, theseparate-effects problems are presented in Section 2.2, and the integral problems are presented in Section2.3.

A code to code or code to data assessment is made for each of the problem types in the development

assessment matrix. The codes used in the assessment are RELAP5/MOD3.2 and RELAP5/MOD3.3 (thesecode versions are referred to throughout the text as MOD3.2 and MOD3.3 respectively).

Table 2.0-1 Developmental Assessment Matrix

ProblemType

AssessmentObjective(s)

Input DeckName(s)*

Phenomenological Problems

Nine-Volume Water Over Steam

Gravitation head effect,Two fluid kinematics

ninevol.i

Nitrogen-Water Manometer Problem

Noncondensable state,Oscillatory flow

manom3.i

Branch Reentrant Tee Problem Tee model using branch component brtee.i

Crossflow Tee Problem Tee model using crossflow feature crfltee.i

Cross Tank problem Crossflow feature, Recirculating flow

crosstank.i

Three-Stage Turbine Turbine component turbine.i

Workshop Problem 2 Hypothetical two-loop PWR, System Modeling,

Control system,Steady-state option

wrkshp_prob2.i

Workshop Problem 3 Hypothetical two-loop PWR, System Modeling,

Control system,Transient option

wrkshp_prob3.i

Horizontal Stratified Countercurrent Flow

Countercurrent flow model stwave.i

Pryor’s Pipe Problem Water packing pryors.i

Separate-Effects Problems



Edwards Pipe Problem Vapor generation model edhtrk.iedhtrkd.iedhtrkn.i

Dukler Air-Water Flooding Tests

Noncondensable,Interphase drag mode,

Countercurrent flow model

dukler100.idukler250.idukler500.i

dukler1000.i

Marviken Test 24 Subcooled choking model marv24.i

Marviken Test 22 Subcooled choking model marv22.i

LOFT Test L3-1 Accumulator Blowdown

Accumulator model l31acc.i

Bennett’s Heated Tube Experiments

Experiment 5358Experiment 5294Experiment 5394

Nonequilibrium heat transfer,Vapor generation model,

CHF correlation

ben5294.iben5358.iben5394.i

Royal Institute of Technology Tube Test 261


CHF correlation

rit261.ir261-pg.i

ORNL Bundle CHF TestsTest 3.07.9BTest 3.07.9NTest 3.07.9W


CHF correlation

3079B.i3079N.i3079W.i

ORNL Void Profile TestTest 3.09.10i

Nonequilibrium heat transfer,Vapor generation model

or3910i.i

Christensen Subcooled Boiling Test 15

Subcooled boiling model chris15.i

Shoukri Subcooled Boiling Experiment at Low Pressure

Subcooled boiling model sh3c.i

MIT Pressurizer Test Wall condensation model,Stratified interfacial heat transfer

mitst4base.i

FLECHT-SEASET Forced Reflood Tests

Test 31504Test 31701

Reflood model fs31504.ifs31701.i

Table 2.0-1 Developmental Assessment Matrix (Continued)

ProblemType


Input DeckName(s)*



* Most of these input decks are provided with the RELAP5 MOD3.3 Beta code distribution.

FLECHT-SEASET Boiloff Test 35658

Saturated BoilingInterphase Drag

fs35658.i

Integral - Effects Problems

LOFT Small-Break Test L3-7 Small-Break LOCA loftl37ss.iloftl37tr.i

LOFT Large-Break Test L2-5 Large-Break LOCA l25ss.il25ss2.i

l25.i

Semiscale Natural Circulation Test S-NC-2Test S-NC-3

Natural Circulation snc02-1.i through snc02-16.isnc03-10.i through snc03-19.i

Zion-1 PWR Small Break Small-Break LOCA typ1200.i

Table 2.0-1 Developmental Assessment Matrix (Continued)

ProblemType


Input DeckName(s)*


PHENOMENOLOGICAL PROBLEMS

2.1 PHENOMENOLOGICAL PROBLEMS

These are a collection of simple problems that have been used to demonstrate qualitatively correctfunctioning of the code. In some cases they are problems that have been used to diagnose numericalproblems that have been encountered during code development. The rationale for the selection of thisparticular set lies in the historical evolution of the code.

2.1.1 Nine-Volume Water Over Steam

The 9-volume water-over-steam problem is simply a vertical pipe with the upper one-third of thepipe initially filled with water and the bottom two-thirds of the pipe filled with steam. The problem wasdeveloped to check the gravitational head effect and the development of countercurrent flow from an initialsharp liquid-vapor interface. Since no analytical solution exists for this problem, it is primarily aqualitative test, i.e. the liquid should fall to the bottom of the pipe. A quantitative limit on the time atwhich the liquid should first reach the bottom of the pipe is provided by considering the liquid to be in freefall. The models for convective flow in a confined pipe, interphase and wall friction should result in agreater time for code calculations.

Figure 2.1-1 shows the RELAP5/MOD3 nodalization diagram. The vertical pipe (the total pipe

length is 4.16448 m and the pipe area is 1.0 m2) was modeled using nine volumes and eight junctions. Theupper three volumes were initially filled with saturated water at a pressure of 413 kPa, and the bottom sixvolumes were filled with saturated steam at a pressure of 413 kPa. During the transient, the liquid fallsfrom the top three volumes and displaces the vapor in the bottom three volumes. Counter current flowdevelops which restricts the rate of fall of the liquid, both due to frictional effects and due to the pressuregradient associated with the upward vapor flow.

The local void fraction history for volumes 1, 3, 5, 7, and 9 as calculated by both MOD3.2 andMOD3.3, is shown in Figure 2.1-2, through Figure 2.1-6. The results indicate that the interphase frictionmodeling in MOD3.3 produces less interphase drag allowing the two phases to slip by each other easierand the liquid to fall to the bottom of the pipe more quickly. This is illustrated in Figure 2.1-7 and Figure2.1-8 which show the axial void profile at various times of the transient for the MOD3.2 and MOD3.3calculations respectively. The liquid in the top three volumes tended to hang-up in the MOD3.2calculation rather than drain smoothly as it did in the MOD3.3 calculation. An analytical solution to givecredence to the calculation is to assume the liquid acts like a plug and the vapor is pushed out of the piperather than percolate up through the liquid. With no friction, the time it takes the plug to reach the bottom

of the pipe is given as where s is the distance for the plug to travel and a is the gravitationalconstant. At s equal to 2.78 meters, the time is 0.75 seconds. Factoring in interphase drag as the twophases slip past each other, the calculated results appear to be reasonable. Figure 2.1-6 shows liquidappearing in the bottom volume at about 0.4 seconds and is full by about 2.75 seconds. The slightly earlyarrival of the liquid is the result of numerical diffusion associated with the finite number of nodes. Thevoid fraction in the intermediate volumes (see Figure 2.1-4) drops no lower than 0.80 indicating that theliquid forms a thin stream as it flows to the bottom volumes. All of these features are qualitatively inagreement with the physics of this problem.

2 s g⁄( )



Pipe 1 Steam

Water

Figure 2.1-1 RELAP5 Nodalization Diagram of the Nine-Volume Water Over Steam Problem

1

2

3

4

5

6

7

8

9

0 2 4 6 8 10Time (sec)

0

0.25

0.5

0.75

1

Vapo

r Voi

d Fr

actio

n

Nine Volume Volume 1

voidg-1010000 3.2voidg-1010000 3.3

Figure 2.1-2 The History of Void Distribution for the Nine-Volume Problem (Volume 1)



0 2 4 6 8 10Time (sec)

0

0.25

0.5

0.75

1Va

por V

oid

Frac

tion

Nine Volume Volume 3

voidg-1030000 3.2voidg-1030000 3.3


0 2 4 6 8 10Time (sec)

0

0.25

0.5

0.75

1

Vapo

r Voi

d Fr

actio

n

Nine VolumeVolume 5

voidg-1050000 3.2voidg-1050000 3.3




0 2 4 6 8 10Time (sec)

0

0.25

0.5

0.75

1Va

por V

oid

Frac

tion

Nine VolumeVolume 7

voidg-1070000 3.2voidg-1070000 3.3


0 2 4 6 8 10Time (sec)

0

0.25

0.5

0.75

1

Vapo

r Voi

d Fr

actio

n

Nine VolumeVolume 9

voidg-1090000 3.2voidg-1090000 3.3




0 1 2 3 4 Distance from bottom (meters)

0

0.2

0.4

0.6

0.8

1 V

apor

Voi

d Fr

actio

n Nine Volume

MOD3.2

Time = 0.0 sTime = 2.0 sTime = 4.0 sTime = 6.0 sTime = 8.0 s

Figure 2.1-7 Void Distribution in the Vertical Pipe at Various Times as Calculated by MOD3.2

0 1 2 3 4 Distance from bottom (meters)

0

0.2

0.4

0.6

0.8

1

Vap

or V

oid

Frac

tion

Nine VolumeMOD3.3

Time = 0.0 sTiime = 2.0 sTime = 4.0 sTime = 6.0 sTime = 8.0 s

Figure 2.1-8 Void Distribution in the Vertical Pipe at Various Times as Calculated by MOD3.3



In conclusion, the nine volume problem demonstrated that the calculated gravitational effect andkinematic models for phase continuity provide a qualitatively correct result. The results indicate interphasefriction modeling in MOD3.3 results in less interphase drag, compared to MOD3.2, allowing the liquid tofall to the bottom of the pipe more quickly and in a smoother manner.

2.1.2 Nitrogen-Water Manometer Problem

A 20-volume nitrogen-water manometer, as shown in Figure 2.1-9, was set up to check thenoncondensable state calculation, the code momentum formulation for periodic flow, and the leveltracking model. The manometer problem has an analytical solution so the calculated period of theoscillation can be checked.

The manometer was modeled using a pipe component. Each volume had an area of 1.0e-02 m2

(0.108 ft2) and a length of 1.0 m (3.2808 ft). The first 10 volumes were oriented vertically downward andthe last 10 volumes were oriented vertically upward. A time dependent volume and single junction wasconnected to both the pipe inlet and outlet. The bottom five volumes on the left side (777060000 to777100000) and the bottom five volumes on the right side (777110000 to 777150000) were filled initially

with water at 100.11 kPa (14.5 psi) and 323 K (122oF). The remaining volumes were initialized with drynitrogen at the same pressure and temperature (including the time dependent volumes). The wall frictionflag was set to 1, to turn off wall friction effects, and the mixture level tracking model was turned on. Aninitial null problem was run to initialize the pressure gradient in the pipe component to a hydrostatic state.To initiate the oscillation, an initial velocity of -1.0 m/s was placed at each junction. The expected

123456789

10 11121314151617181920

Pipe 777 Nitrogen

Water

SJ 665 SJ 815

Figure 2.1-9 RELAP Nitrogen-Water Manometer Nodalization

TDV888

TDV555



behavior for this problem is that, without wall friction, the liquid will oscillate back and forth between thetwo vertical columns with a non-decaying maximum height. The frequency of oscillation is analyticallypredictable and can be used as a measure of the correctness of the code calculated body force to inertiaratio.

The behavior of the manometer can be illustrated by examining the calculated liquid velocity andcollapsed liquid level. It is noted that in order to make the MOD3.2 calculation run, the mixture leveltracking model had to be deactivated. A comparison of the liquid velocity at the bottom of the manometeris shown in Figure 2.1-10. The initial liquid velocity was input at -1.0 m/s that resulted in an amplitude of+/-1.0 m/s. Analytically, the amplitude of the liquid velocity should remain at the +/-1.0 m/s rate. Asobserved, the MOD3.2 calculated liquid velocity decayed until a steady amplitude of about +/-0.4 m/s wasreached. On the other hand, the MOD3.3 calculated amplitude remained at +/-1.0 m/s. Figure 2.1-11shows the liquid level comparison. Again, the MOD3.2 calculation showed a decrease in the amplitude ofthe level. The calculated period of oscillation for both MOD3.2 and MOD3.3 is shown to be about 4.5

seconds. The theoretical period of the oscillation is given by: where L is the length of theliquid in the manometer and g is the gravitational constant. With L = 10 meters, the theoretical period iscalculated to be 4.486 seconds. Thus the period of oscillation evidenced by the two code versions agreeswell with the theoretical value.

This problem shows that qualitatively, MOD3.2 and MOD3.3 agree well with the theoretical periodof oscillation and that MOD3.3 maintained the amplitude of the oscillation. Thus the ratio of gravitationalforce to inertia calculated by the code is correct. The problem also demonstrates that improvements to the

2π 2g L⁄( )⁄ 1 2⁄

0 25 50 75 100Time (sec)

-2

-1

0

1

2

Liqu

id V

elocit

y (m

/s)

Manometer - Liquid Velocity ComparisonBottom of manometer

velfj-777110000 3.2velfj-777110000 3.3

Figure 2.1-10 Liquid Velocity at the Bottom of the Manometer (junction 777110000)



MOD3.3 mixture level tracking model allowed it to function properly. Further, the MOD3.3 versionshows almost zero numerical damping or decay in the maximum velocity and level.

2.1.3 Branch Reentrant Tee Problem

The branch reentrant tee problem is a conceptual problem that was set up during the development of

RELAP5/MOD12.1-1 to test the ability of the branch component to accurately model a tee for single- andtwo-phase inflows. At that time, it illustrated a code problem that was encountered in modeling a tee.Several configurations were originally tried for modeling a tee. While some configurations producedacceptable results, the configuration used here produced unphysical oscillatory and asymmetrical flow inspite of inherent symmetry. For these reasons, this problem has been used to assess the stability andsymmetry of each new code version.

The branch reentrant tee model is constructed of two equal length, equal area side by side horizontalpipes (Components 200 and 300) connected to the outlet side of a horizontally oriented branch component(100) that represents the tee. A time dependent volume (supply source) is connected to the upstream sideof each of the pipes by a single junction. A horizontally oriented single volume (400) is also connected tothe outlet side of the branch component. A time dependent volume (401) is then connected to thedownstream side of the single volume (sink volume). The nodalization diagram for this case is shown inFigure 2.1-12.

The fluid conditions set in the supply sources and all components up to the sink volume (401) aresingle-phase, saturated water, at 2.63962 MPa and 500 K. The fluid conditions set in the sink volume are

0 25 50 75 100Time (sec)

4

4.5

5

5.5

6

6.5

7W

ater L

evel

Manometer - Level ComparisonWater Level

cntrlvar-1 3.2cntrlvar-1 3.3

Figure 2.1-11 MOD3.2 and MOD3.3 Calculated Water Level Comparison for the Manometer Problem



single-phase saturated water at 2.5 MPa and 497.09 K. Fluid flows from the pipes through the tee and intothe sink volume. Theoretically the flows through the pipes should be symmetrical and the out-flow to the

sink volume should be the sum of the two pipe flows. The problem was run to 20 seconds with two-phaseconditions developing in the pipes and branches due to depressurization caused by the lower pressure inthe sink volume.

The results of the MOD3.2 and MOD3.3 calculations are shown in Table 2.1-1 The values shown(taken at 20 seconds) are the mass flow rates in the two pipes upstream of the branch component (PipeComponents 200 and 300) and the junction to the sink volume (Single Junction 402). Symmetrical flow inthe two pipes upstream of the modeled Tee component (branch) was calculated by both MOD3.2 andMOD3.3. Thus it is demonstrated that the unphysical characteristics of oscillatory and unsymmetricalflow that have been encountered with earlier versions of the code do not occur in either the MOD3.2 orMOD3.3 code versions. The difference in the total mass flow between MOD3.2 and MOD3.3 isinsignificant.

Table 2.1-1 Results for the Branch Tee Problem

Variable MOD3.2 MOD3.3

Mass flow rate in Pipe Component 200 898.99 (kg/s) 897.88 (kg/s)

Mass flow rate in Pipe Component 300 898.99 (kg/s) 897.88 (kg/s)

Summation of mass flow rate 1797.98 (kg/s) 1795.76 (kg/s)

Figure 2.1-12 Nodalization of the Branch Tee Problem Using Two Non-Sink Junctions

100

1 2 3

1 2 3

Pipe 300

Pipe 200

SV 400 Branch

SJ202

SJ302

SJ402

TDV301

TDV401

TDV201

SinkVolume



The calculated discharge mass flow rate for MOD3.2 is somewhat lower than the combined pipemass flow rates. The combined MOD3.3 calculated mass flow rates agree with the discharge mass flowrate. This is probably a result of mass error corrections that have been implemented between MOD3.2 andMOD3.3.

2.1.4 Cross-Flow Tee Problem

The cross-flow tee problem is a conceptual problem that was developed to test how the new cross-

flow junction feature of RELAP5/MOD22.1-1 could be used to model a variant of the tee problempresented in the previous section. It was run using both MOD3.2 and MOD3.3 to ascertain whether non-oscillatory and symmetrical flows result in either version.

The model nodalization for this problem is similar to the model presented in the previous sectionexcept that the outflow is modeled using a vertical single volume (single-volume component 400). Thenodalization diagram is shown in Figure 2.1-13. The tee volume (single-volume component 250) wassquare and oriented from left to right. The perpendicular junctions 251 and 254 were made crossflowjunctions. Forward and reverse loss coefficients of 1.0 were used for the crossflow junctions. The fluid

conditions were the same as the branch-tee model presented previously. The problem was run to 20seconds with two-phase conditions developing in the pipes and branches.

The results of the MOD3.2 and MOD3.3 calculations are shown in Table 2.1-2. The values shown

(taken at 20 seconds) are the mass flow rates in the two pipes upstream of the tee volume (Component 250)and the junction to the sink volume (Component 401). As in the previous case, there was symmetricalflow calculated in the two pipes upstream of the tee volume and the calculated mass flow rate downstreamof the tee volume was the combination of the two pipe flows. The calculated discharge mass flow rate forMOD3.3 agrees well with the combined mass flow rates of the two upstream pipes. The calculated

Mass flow rate in Single Junction 402 1795.2 (kg/s) 1795.8 (kg/s)

Mass error 0.9911 (kg) -0.575 (kg)

Table 2.1-2 Results for the Crossflow Tee Problem


Mass flow rate in pipe 200 656.37 (kg/s) 656.0 (kg/s)

Mass flow rate in pipe 300 656.37 (kg/s) 656.0 (kg/s)

Summation of mass flow 1312.74 (kg/s) 1312.0 (kg/s)

Mass flow rate in single junction 402 1316.9 (kg/s) 1312.8 (kg/s)

Mass error -1.0454 (kg) -0.15201 (kg)

Table 2.1-1 Results for the Branch Tee Problem (Continued)




discharge mass flow rate for MOD3.2 is somewhat higher than the combined pipe mass flow rates. This isprobably a result of the mass error calculated between MOD3.2 and MOD3.3 as shown in Figure 2.1-14.The mass flow rates obtained in the two calculations are fairly close.

It is demonstrated with this problem that crossflow junctions can be used to model the behavior of atee without encountering unphysical behavior and minimal differences exist in results obtained usingMOD3.2 and MOD3.3, except that MOD3.3 is more accurate in terms of mass conservation.

2.1.5 Cross Tank Problem

The cross tank problem is a conceptual problem that was set up to test for flow anomalies that mayappear as recirculating two-phase or single-phase flows using the cross-flow junction feature in the code tomodel multidimensional effects. Multidimensional effects may occur in a number of places in a powerplant under transient conditions and it is desirable to be able to approximate such effects by coupling aspatially distributed network of volumes using the cross-flow junction.

The cross tank model consists of two vertically oriented pipe components equally divided into 19volumes and is shown in Figure 2.1-15. Cross-flow junctions connected the two pipe components at eachvolume; thus there are 19 cross-flow junctions. The bottom 15 volumes in each pipe component were

initialized with water at 0.1014 MPa (14.7 psia) and 305 K (90oF). Volume 16 in each pipe componentcontained a mixture of air and water in equilibrium condition with a static quality of 0.5 and a pressure andtemperature the same as the liquid. The remaining three top volumes of each pipe component wereinitialized with air at the same pressure and temperature as the liquid, with a static quality of 1.0 in

Figure 2.1-13 Nodalization Diagram for the Cross Flow Tee Problem

SV250

SV400

1 2

Pipe 200

2 1

Pipe 300

SJ 254(crossflow)

SJ 251(crossflow)

SJ 402

SJ302

SJ253

SJ252

SJ202

SV100

TDV301

TDV401

TDV201



0 10 20 30Time (sec)

-1.5

-1

-0.5

0M

ass E

rror (

kg)

Cross Tee

emass-0 3.2emass-0 3.3

Figure 2.1-14 Calculated Mass Error Comparison for the Crossflow Tee Problem

Figure 2.1-15 RELAP5 Cross Tank Nodalization Diagram

Pipe 400Pipe 200

123

456789

10111213141516171819

123

456789

10111213141516171819

Vapor

Liquid

SJ 301SJ 302SJ 303SJ 304SJ 305SJ 306SJ 307SJ 308SJ 309SJ 310SJ 311SJ 312SJ 313SJ 314SJ 315SJ 316SJ 317SJ 318SJ 319



equilibrium conditions. The problem was run using null flow conditions (vg= vf = 0.0). The expected

result is that the system will remain near the null condition throughout the calculation.

The calculated mass flow rates just below and above the liquid level in each of the two pipecomponents and the corresponding connecting cross-flow junctions are shown in Figure 2.1-16 throughFigure 2.1-21. As observed, spontaneous recirculation flows were established in the MOD3.2 calculationwhereas the MOD3.3 calculated flows remained near null. The MOD3.2 flow anomalies were discovered

to be caused by three problems with the code formulation2.1-2. The first problem related to the way inwhich the donored velocity was computed for inclusion in the momentum flux terms. The second problemrelated to the Co coefficient in the drift flux model and its value for special cases that included low flow

and single- phase liquid. The third problem related to the large coefficients in the difference momentumequation for single-phase flow. Modifications to the code to address these problems have corrected amajority of the recirculation problems as observed by the MOD3.3 calculated mass flow rates shown inFigure 2.1-16 through Figure 2.1-21.

The results demonstrate that the problem or problems causing the non-physical recirculation flowscalculated by MOD3.2 have been corrected in MOD3.3. All of the cross-flow junction mass flow ratesremain near zero for the MOD3.3 case.

0 10 20 30Time (sec)

-200

-100

0

100

200

300

Mas

s Flo

w Ra

te (k

g/s)

Cross TankCheck flow recirculation

mflowj-200140000 3.2mflowj-200140000 3.3

Figure 2.1-16 Mass Flow Rate Comparison: Pipe 200, Volume 14 (Liquid Region)



0 10 20 30Time (sec)

-300

-200

-100

0

100

200M

ass F

low

Rate

(kg/

s)Cross Tank

Check flow recirculation

mflowj-400140000 3.2mflowj-400140000 3.3

Figure 2.1-17 Mass Flow Rate Comparison: Pipe 400, Volume 14 (Liquid Region)

0 10 20 30Time (sec)

-100

0

100

200

300

Mas

s Flo

w Ra

te (k

g/s)


mflowj-315000000 3.2mflowj-315000000 3.3

Figure 2.1-18 Mass Flow Rate: Cross Flow Junction 315 (Connecting Pipe 200 to Pipe 400 in the Liquid Region)



0 10 20 30Time (sec)

-1

-0.5

0

0.5

1M

ass F

low

Rate

(kg/

s)Cross Tank


mflowj-200180000 3.2mflowj-200180000 3.3

Figure 2.1-19 Mass Flow Rate Comparison: Pipe 200, Volume 18 (Gas Region)

0 10 20 30Time (sec)

-1

-0.5

0

0.5

1

Mas

s Flo

w Ra

te (k

g/s)


mflowj-400180000 3.2mflowj-400180000 3.3

Figure 2.1-20 Mass Flow Rate Comparison: Pipe 400, Volume 18 (Gas Region)



2.1.6 Three-Stage Turbine

A steam turbine is a device that converts thermal energy contained in high-pressure and high-temperature steam to mechanical work. A special volume centered model for the turbine is included inRELAP5. A turbine can be modeled using a single stage-group, i.e., a single volume and junction, orseveral stage-groups linked together, depending upon the resolution required. The number of actual stagesthat a stage-group corresponds to is not explicitly specified, but is simply implicit in the specification ofthe design operating conditions. Thus a stage group could consist of a single stage or several stageslumped together. For the three-stage turbine model used for assessment, each stage-group corresponds to asingle turbine stage.

The three-stage-group turbine problem is a conceptual problem and is shown in Figure 2.1-22. Inthis case three turbine stage-groups are modeled, denoted on as Stage 1, Stage 2, and Stage 3 respectivelyin Figure 2.1-22. The complete turbine is then modeled using the three stage groups (three singlevolumes) connected in series. This problem is used to check-out the turbine component for each newversion of the code by comparing the results to results from previous code versions. It does not correspondto a real turbine design and no performance data exist for comparison.

Single-phase, superheated vapor was supplied to the turbine from the upstream source and exhaustedinto a downstream sink. The pressure and temperature in the upstream source and the downstream sinkvolumes are 6 MPa, 748 K, and 0.5 MPa, 430 K, respectively, and are constant with time.

0 10 20 30Time (sec)

-1

-0.5

0

0.5

1M

ass F

low

Rate

(kg/

s)Cross Tank


mflowj-319000000 3.2mflowj-319000000 3.3

Figure 2.1-21 Mass Flow Rate: Cross Flow Junction 319 (Connecting Pipe 200 to Pipe 400 in the Gas Region)



The corresponding RELAP5 nodalization diagram is also indicated on Figure 2.1-22. Two time-dependent volumes were used to model the steam source and discharge plena of the turbine, timedependent volumes 200 and 900, respectively. Three turbine components were used to simulate the threestage-groups. Artificial turbines having zero efficiency (η=0 for stage 0 and stage 4 corresponding tovolumes 400 and 800) were placed before and after the three normal turbines (stages 1, 2, and 3,corresponding to volumes 500, 600, and 700). The turbine inlet line was modeled using a single volume,volume 300. No loss coefficient was used for the exit junction.

The MOD3.2 and MOD3.3 calculated stage pressures, torques, and efficiencies for each stage areshown in Table 2.1-3. There is a pressure drop through the turbines and a slight pressure rise from stage 3

Table 2.1-3 Turbine Parameters for the Assessment Problem

Volume NumberPressure (MPa)

Torque (n-m)

Efficiency

SV - 3000010000 MOD3.2 2.816 ---- ----

SV - 3000010000 MOD3.3 2.784 ---- ----

Artificial turbine - 400010000 (Stage 0) MOD3.2 2.372 0 0

Artificial turbine - 400010000 (Stage 0) MOD3.3 2.351 0 0

Turbine - 500010000 (Stage 1) MOD3.2 1.312 429986.0 0.800


Figure 2.1-22 Nodalization Used for the Three-stage Group Turbine Problem

SV300 400 500 600 700 800

SJ850

TDJ250

Stage 0Stage 1

Stage 2Stage 3 Stage 4

TDV200

TDV900



to the exit volume. The differences between MOD3.2 and MOD3.3 tabulated results are small andinsignificant. However, the mass error calculated by the two code versions (see Figure 2.1-23) shows thatthe mass error corrections implemented in the code since MOD3.2, have significantly reduced the masserror for this problem.





Artificial turbine - 800010000 MOD3.2 0.449 0 0

Artificial turbine - 800010000 MOD3.3 0.449 0 0

Table 2.1-3 Turbine Parameters for the Assessment Problem (Continued)

Volume NumberPressure (MPa)

Torque (n-m)

Efficiency

0 0.5 1 1.5Time (sec)

0

0.5

1

1.5

Mas

s Erro

r (kg

)

Turbine


Figure 2.1-23 Mass Error Comparison for the Three-stage Turbine Problem



2.1.7 Workshop Problem 2

This problem was named Workshop Problem 2 as it was used as the second demonstration problem

during the RELAP5/MOD12.1-3 workshop held in April 1982. It was set up to simulate one loop of a two-loop pressurized water reactor (PWR) and to assess the system modeling capability of the RELAP5computer code, including the control features, for steady-state operation. The system consists of a primarysystem loop and a steam generator secondary recirculation loop. The primary loop contains a reactorvessel, including an electrically heated reactor core, a pressurizer, a hot leg, a once-through steamgenerator, a coolant pump and a cold leg. The steam generator secondary loop consists of the heatexchanger tube region, separator, and downcomer. Feedwater is supplied from a time dependent junctionand volume. Steam is discharged from the secondary loop through a control valve connected to a timedependent volume.

The RELAP5 nodalization diagram for the system is shown in Figure 2.1-24 The model consists offour time-dependent volumes, twenty-six regular volumes, twenty-one junctions, one pump, two valves,and six heat structures. The built-in, homologous curves for the Westinghouse pump were used to model

the pump behavior. The primary system pressure is initially set at 1.5x107 Pa with an average systemtemperature of 550 K and a primary system flow at 131 kg/s. The core power is set at 50 MW. The

secondary side pressure is set at 2.0x106 Pa with feedwater flow and temperature set at 26.1 kg/s and 478K, respectively.

Figure 2.1-24 Relap5 Nodalization Diagram for Workshop Problems 2 and 3

Core

Pipe 110

1

2

3

SV 100

Core Inlet

Pump 170SV 160

Cold Leg PipeSJ151

1

2

3

Pipe 150Primary SideSG Tubes

Pipe 140

Hot Leg Pipe12

Branch 120

Core Outlet

HS 111ElectricHeater

Pipe 130Pressurizer

SJ 131

TDV 132 TDV 134

Valve 133

Pipe 132

WorkshopProblem 3changes

TDV 240TDV 280

Valve 231Valve 232Steam Dome

SV 230Branch 290

SJ 141Branch 220

Separator 260

Sec. SideSG TubesPipe 210

1

2

3

Pipe 250DCBottom

1

2

21

1234

SJ 251

Annulus 270DC Top

TDV 200Feed Supply

HS 115SG Tube Metal

SRV MSV



Here again, this problem has been used to compare results obtained with each new version of thecode with results from past versions. The calculations were performed with the steady-state option.Figure 2.1-25 through Figure 2.1-27 show the MOD3.2 and MOD3.3 calculated volume pressure in thepressurizer, core outlet, and steam dome of the steam generator secondary side, respectively. Thecalculated pressure response for both codes was similar in both the primary side and the secondary side.

The steam generator secondary liquid level comparison is shown in Figure 2.1-28. The calculatedliquid levels for the two codes agree well, although there was a small difference at the end of thecalculation. This difference is considered insignificant, and it appears that had the problem run further intime, the calculated levels would converge.

Figure 2.1-29 and Figure 2.1-30 show the mass flow rate in the hot leg to the pressurizer and in theprimary side of the steam generator. The MOD3.2 and MOD3.3 results are similar, which implies that theheat transfer from the primary side to the secondary side is similar in the two codes.

2.1.8 Workshop Problem 3

Workshop Problem 3 simulates a modified station blackout transient for the system described inWorkshop Problem 2. As with Workshop Problem 2, this is a hypothetical problem which was originallyused to demonstrate the transient simulation capability of RELAP5 when applied to a system problem. Itis used here to compare with results from the previous version of the code. The model of WorkshopProblem 2 was used for the transient analysis except that the pressurizer time-dependent volume(Component 132) was replaced by a vertical, two-volume pipe (Component 132), a pressurizer relief valve

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150Time (sec)

15

15.002

15.004

15.006

15.008

15.01

Pres

sure

(MPa

)

Pressurizer pressure historyWorkshop Problem 2

p-130040000 3.2p-130040000 3.3

Figure 2.1-25 Pressurizer Pressure History of Workshop Problem 2



0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150Time (sec)

15

15.1

15.2

15.3

15.4

Pres

sure

(MPa

)Core Outlet pressure history

Workshop Problem 2

p-120010000 3.2p-120010000 3.3

Figure 2.1-26 Core Outlet Pressure History of Workshop Problem 2

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150Time (sec)

1.8

2.3

2.8

3.3

3.8

Pres

sure

(MPa

)

Steam Generator steam dome pressure historyWorkshop Problem 2

p-230010000 3.2p-230010000 3.3

Figure 2.1-27 Steam Generator Steam Dome Pressure History of Workshop Problem 2



0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150Time (sec)

4.4

4.8

5.2

5.6

6

6.4

6.8

7.2Li

quid

Lev

el (m

)Steam Generator secondary side liquid level

Workshop Problem 2

cntrlvar-21 3.2cntrlvar-21 3.3

Figure 2.1-28 Steam Generator Secondary Side Liquid Level for Workshop

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150Time (sec)

-4

0

4

8

12

16

20

Mas

s Flo

w Ra

te (k

g/s)

Mass flow into the pressurizerWorkshop Problem 2

mflowj-120030000 3.2mflowj-120030000 3.3

Figure 2.1-29 Mass Flow in the Primary Loop Hot Leg for Workshop Problem 2



(Component 133 - trip valve), and a time-dependent volume (Component 134) as shown by the bracket andarrow in Figure 2.1-24. In the model, the steady-state conditions calculated from Workshop Problem 2were used as initial conditions; and the transient sequence used the following boundary conditions: 1)relief valve opens when the pressurizer pressure exceeds 18.0 MPa, and 2) reactor shutdown occurs after 5seconds continuous actuation of the pressurizer relief valve.

Figure 2.1-31 and Figure 2.1-32 show the volume pressure in the core and the pressurizerrespectively. The primary pressures calculated by MOD3.2 and MOD 3.3 are similar out to about 20seconds. After 20 seconds the pressure calculated by MOD3.3 was lower than the pressure calculated byMOD3.2. By about 70 seconds the pressure calculated by MOD3.3 had leveled off whereas the pressurecalculated by MOD3.2 began to increase. A couple of reasons for the difference in the pressure responsebetween the two code versions are given. First, the primary loop mass flow rate calculated by MOD3.3was higher as shown in Figure 2.1-33, effectively removing more energy through the steam generator andthus lowering the primary pressure. Second, the mass flow rate out of the pressurizer through the reliefvalve calculated by MOD3.3 was higher as shown in Figure 2.1-34, thus more energy was removed andthe pressure decreased. At about 35 seconds the MOD3.3 calculated pressure dropped below thepressurizer relief valve closing setpoint and the flow out the valve stopped. This did not occur until about45 seconds in the MOD3.2 calculation. After about 70 seconds the MOD3.3 calculated primary loop flowwas steady at about 10 kg/s (natural circulation) with a liquid full system. The MOD3.2 calculated loopflow, however, stopped momentarily then was virtually zero throughout the remainder of the transientunder two-phase conditions.

0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150Time (sec)

100

150

200

250

300

Mas

s Flo

w Ra

te (k

g/s)

Mass Flow in the primary side of steam generatorWorkshop Problem 2

mflowj-151000000 3.2mflowj-151000000 3.3

Figure 2.1-30 Mass Flow in the Primary Side of Steam Generator for Workshop Problem 2



0 50 100 150 200 250 300Time (sec)

8

10

12

14

16

18

20Pr

essu

re (M

Pa)

Volume Pressure of the reactor coreWorkshop Problem 3

p-110030000 3.2p-110030000 3.3

Figure 2.1-31 Reactor Core Pressure Response for Workshop Problem 3

0 50 100 150 200 250 300Time (sec)

8

10

12

14

16

18

20

Pres

sure

(MPa

)

Volume Pressure of the pressurizerWorkshop Problem 3

p-132020000 3.2p-132020000 3.3

Figure 2.1-32 Pressurizer Pressure Response for Workshop Problem 3



0 50 100 150 200 250 300Time (sec)

0

50

100

150

200

250

300M

ass F

low

Rate

(kg/

s)Mass flow in the reactor core

Workshop Problem 3

mflowj-110020000 3.2mflowj-110020000 3.3

Figure 2.1-33 Mass Flow in the Reactor Core for Workshop Problem 3

0 50 100 150 200 250 300Time (sec)

0

5

10

15

20

25

Mas

s Flo

w Ra

te (k

g/s)

Mass flow into the pressurizerWorkshop Problem 3

mflowj-133000000 3.2mflowj-133000000 3.3

Figure 2.1-34 Mass Flow in the Pressurizer for Workshop Problem 3



Figure 2.1-35 and Figure 2.1-36 show the secondary loop mass flow in the steam generator riserand steam exit line. At the initiation of the transient, the steam line control valve was closed, along with thefeedwater flow, and the secondary side pressure relief valve opened. The pressure temporarily decreaseduntil the energy being transferred to the secondary side from the primary side exceeded the energy beingremoved through the relief valve and the pressure began to increase as shown in Figure 2.1-37. As theenergy being removed from the primary side declined, the energy removal rate through the secondary siderelief valve began to dominate and the pressure response turned over. The turn over point occurred earlierfor the MOD3.2 calculation. The pressure decreased until the relief valve closing pressure was reachedand the valve closed. The pressure then increased to the relief valve open setpoint, the valve opened andthe pressure declined. The valve cycled in this manor through the remainder of the transient.

There have been several changes implemented in the code between MOD3.2 and MOD3.3 that couldexplain the differences observed. These changes include the following: a) an improved wall friction modeland b) defaulting the following card 1 user options: 1) Option 55 - heat transfer correlation improvements,2) Option 57 - phasic partitioning, 3) Option 53 - Henry-Fauske critical flow model, and 4) Option 47 -linear implicit logic for the wall drag, form loss, and interphase drag terms in the sum and differencemomentum equations. In general, the MOD3.3 calculation showed smoother trends and less erraticbehavior than the MOD3.2 calculation.

2.1.9 Horizontally Stratified Countercurrent Flow

The horizontally stratified countercurrent flow problem is a conceptual problem involving ahorizontal pipe closed at both ends with a linearly graduated liquid level. Because of the gravitational head

0 50 100 150 200 250 300Time (sec)

-200

-100

0

100

200

300

400

500

Mas

s Flo

w Ra

te (k

g/s)

Mass flow in the SG secondary sideWorkshop Problem 3

mflowj-210020000 3.2mflowj-210020000 3.3

Figure 2.1-35 Mass Flow in the Steam Generator Secondary Side (Riser) for Workshop Problem 3



0 50 100 150 200 250 300Time (sec)

-10

-5

0

5

10M

ass F

low

Rate

(kg/

s)Mass flow in the steam discharge line

Workshop Problem 3

mflowj-231000000 3.2mflowj-231000000 3.3

Figure 2.1-36 Mass Flow in the Steam Discharge Line for Workshop Problem 3

0 50 100 150 200 250 300Time (sec)

1

1.5

2

2.5

3

3.5

Pres

sure

(MPa

)

Volume Pressure of the SG steam domeWorkshop Problem 3

p-230010000 3.2p-230010000 3.3

Figure 2.1-37 Steam Generator Steam Dome Pressure Response for Workshop Problem 3



difference, the liquid tends to flow from the higher-level side to the lower-level side due to the hydrostaticpressure difference caused by stratification. The vapor is forced to flow in the opposite direction from theliquid and a countercurrent flow is developed. The problem was designed to check the RELAP5countercurrent flow model and to verify that the speed of propagation for a void wave is qualitativelycorrect.

Figure 2.1-38 shows the RELAP5 nodalization diagram. The pipe was modeled using 20 volumes

and 19 junctions (total pipe length 10 m, pipe flow area = 0.19635 m2). The pipe is initially filled with alinearly distributed, two-phase, saturated, liquid/vapor mixture at a pressure of 10 MPa; and the qualityvaries from 0.083 to 0.067 which corresponds to an average void fraction of approximately 0.5.

The calculated history of the liquid and vapor junction velocities at three locations (left, middle, andright) are shown in Figure 2.1-39 through Figure 2.1-44. The liquid flows from right to left and the vaporflows from left to right and the void profile propagates as a wave. As observed, there is no differencebetween MOD3.2 and MOD3.3. Thus the code modifications implemented between these two codeversions does not affect the results of this problem. The speed of propagation from the calculated results is0.74 m/s while the theoretical value for a stratified wave in frictionless flow is 2.8 m/s. The lower speed ofpropagation determined from the calculations is qualitatively in the right direction as a result of theinterphase drag and virtual mass models (which are inherent in the codes). These results showqualitatively that the horizontal stratified flow model in MOD3.3 is functioning properly.

Figure 2.1-38 Relap5 Nodalization Diagram for a Horizontally Stratified Countercurrent Flow Problem

Saturated LiquidSaturated Vapor

Pipe 31 201918171615141312111098765432



0 10 20 30Time (sec)

-0.05

0

0.05Li

quid

Velo

city

(m/s)

HorizCCFLiquid Velocity

velfj-3010000 3.2velfj-3010000 3.3

Figure 2.1-39 Relap5-Calculated Junction Liquid Velocity at the Left End

0 10 20 30Time (sec)

-0.05

0

0.05

Liqu

id V

elocit

y (m

/s)


velfj-3090000 3.2velfj-3090000 3.3

Figure 2.1-40 Relap5-Calculated Junction Liquid Velocity at the Mid-section



0 10 20 30Time (sec)

-0.05

0

0.05Li

quid

Velo

city

(m/s)


velfj-3190000 3.2velfj-3190000 3.3

Figure 2.1-41 Relap5-Calculated Junction Liquid Velocity at the Right End

0 10 20 30Time (sec)

-0.06

-0.04

-0.02

0

0.02

0.04

0.06

Gas

Velo

city

(m/s)

HorizCCFGas Velocity

velgj-3010000 3.2velgj-3010000 3.3

Figure 2.1-42 Relap5-Calculated Junction Vapor Velocity at the Left End



0 10 20 30Time (sec)

-0.06

-0.04

-0.02

0

0.02

0.04

0.06G

as V

eloc

ity (m

/s)HorizCCFGas Velocity

velgj-3090000 3.2velgj-3090000 3.3

Figure 2.1-43 Relap5-Calculated Junction Vapor Velocity at the Mid-Section

0 10 20 30Time (sec)

-0.06

-0.04

-0.02

0

0.02

0.04

0.06

Gas

Velo

city

(m/s)

HorizCCFGas Velocity

velgj-3190000 3.2velgj-3190000 3.3

Figure 2.1-44 Relap5-Calculated Junction Vapor Velocity at the Right End



2.1.10 Pryor’s Pipe Problem

The Pryor’s Pipe Problem was developed to check the water packing problem that can occur in thefinite difference scheme. The problem consists of a horizontal pipe section, a water injection system, and aconstant pressure exit system. The nodalization diagram is shown in Figure 2.1-45. Initially, the pipe wasfilled with slightly superheated vapor, at a pressure of 0.4 MPa and a temperature of 418.2 K. Subcooledliquid at a pressure of 0.4 MPa and a temperature of 353 K was injected into the inlet side of the pipe andthe opposite end is open to a time dependent volume at a pressure 0.4 MPa. The model consisted of 20volumes and 19 junctions for the pipe section, a time-dependent volume and junction for the waterinjection source, and a time-dependent volume and single junction for the discharge system.

Local void fraction and pressure in the even numbered volumes out to the middle of the pipe areshown for the MOD3.2 and MOD3.3 calculations in Figure 2.1-46 through Figure 2.1-55. The

calculations were run with the water packer turned on. A RELAP5/MOD22.1-3 calculation was made withthis model with the water packing option turned off. The results showed that when each volume filled withwater, pressure spikes resulted with a maximum amplitude of 0.4 MPa. These pressure spikes wereattributed to water packing. The MOD3.2 and MOD3.3 calculations (with water packer on) do not showthese pressure spikes as the volumes fill with water. It is also noted that the code calculated response forthe two versions (MOD3.2 and MOD3.3) is essentially the same. Thus, the MOD3.3 modifications did notaffect the response of the water packing model.

Figure 2.1-45 Relap5 Nodalization Diagram for Pryor’s Pipe Problem

1 202 3 4 5 6 7 9 10 11 12 13 14 15 16 17 18 198

Pipe 110

TDV120

TDV100 TDJ

105SJ

115Water InjectionSource



0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1Time (sec)

0

0.5

1Va

por V

oid F

racti

onPryors - Voidg Comparison

voidg-110020000 3.2voidg-110020000 3.3

Figure 2.1-46 Void Fraction Response for Volume 2 of the Pryor’s Pipe Problem

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1Time (sec)

0

0.5

1

Vapo

r Voi

d Fra

ction

Pryors - Voidg Comparison

voidg-110040000 3.2voidg-110040000 3.3




0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1Time (sec)

0

0.5

1Va

por V

oid F

racti


voidg-110060000 3.2voidg-110060000 3.3


0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1Time (sec)

0

0.5

1

Vapo

r Voi

d Fra

ction

Pryors - Voidg Comparison

voidg-110080000 3.2voidg-110080000 3.3




0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1Time (sec)

0

0.5

1Va

por V

oid F

racti


voidg-110100000 3.2voidg-110100000 3.3


0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1Time (sec)

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Pres

sure

(MPa

)

Pryors - Pressure Comparison

p-110020000 3.2p-110020000 3.3

Figure 2.1-51 Volume Pressure Response for Volume 2 of the Pryor’s Pipe Problem



0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1Time (sec)

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1Pr

essu

re (M

Pa)


p-110040000 3.2p-110040000 3.3


0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1Time (sec)

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Pres

sure

(MPa

)


p-110060000 3.2p-110060000 3.3




0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1Time (sec)

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1Pr

essu

re (M

Pa)


p-110080000 3.2p-110080000 3.3


0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1Time (sec)

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Pres

sure

(MPa

)


p-110100000 3.2p-110100000 3.3




2.1.11 References

2.1-1. V. H. Ransom et al., RELAP5/MOD2 Code Manual, Volume 3: Developmental AssessmentProblems, EGG-TFM-7952, December 1987.

2.1-2. G. A. Mortensen et al., Flow Anomaly Final Report, Idaho National Engineering Laboratory,Lockheed Martin Idaho Technologies Co., Prepared for the Office of Nuclear RegulatoryResearch, U. S. Nuclear Regulatory Commission, February 1997.

2.1-3. V. H. Ransom et al., RELAP5/MOD1 Code Manual, Volumes 1 and 2, NUREG/CR-1826, EGG-2070, March 1982.


SEPARATE - EFFECTS PROBLEMS

2.2 SEPARATE - EFFECTS PROBLEMS

This section presents the results of the application of RELAP5/MOD3.3 to 13 separate-effectexperiments in order to give qualitative and quantitative assessment of the adequacy and accuracy of theMOD3.3 code version. Simulations of the experiments using RELAP5/MOD3.3 are compared to the datasets for the experiments and to the results obtained using prior versions of the code, where appropriate. Inall cases, results from MOD3.2 and MOD3.3 are compared. The assessment objectives of each simulationare listed in Table 2.0-1.

2.2.1 Edwards Pipe Problem

The Edwards2.2-1 pipe experiment was used to verify several models in RELAP5. In this assessmentthe problem is used to verify the blowdown behavior (including flashing), heavy water properties, and thenearly implicit time advancement scheme. A few artificial boundary conditions were added and includeheat loss at the pipe wall, a couple of control components, a few trips, and a reactor kinetics model. Thesewere added so that more components are exercised, since this problem is run whenever a new version ismade. It is expected these additions have little if any effects on the key parameters examined. The

hydrodynamic nodalization scheme that was used in the RELAP5/MOD12.2-2 assessment is the same forthis assessment and is shown in Figure 2.2-1.

A comparison of the measured and calculated pressure and void fraction is shown in Figure 2.2-2and Figure 2.2-3. The MOD3.2 and MOD3.3 results are in good agreement with the data. It is noted thatMOD3.2 over-calculated the pressure where as MOD3.3 under-calculated the response. The void fractionhowever, was under-calculated by MOD3.2 and over-calculated by MOD3.3. The MOD3.3 results tend tofollow somewhat closer to the measured values. The pressure and void fraction behavior within the pipeare tied to the flow out of the pipe. The MOD3.3 code version uses the Henry-Fauske critical flow modelas a default model. It is suspected that this model over-calculated the liquid discharge rate, thus affectingthe calculated pressure and void fraction behavior.

Figure 2.2-1 Relap5 Hydrodynamic Nodalization for Edwards’ Pipe Experiment

1 2019181716141312111098765432 15

Pipe 3

SJ 4TDV 5

Area = 0.00456 m2 Area = 0.003967 m2

Pipe Initial Conditions:P = 7.0 MPaT = 502 KVg = Vf = 0.0

SJ and TDV Initial Conditions:P = 0.1 MPaX = 1.0Vg = Vf = 0.0



0 0.1 0.2 0.3 0.4 0.5Time (sec)

0

1.25

2.5

3.75

5

6.25

7.5Pr

essu

re (M

Pa)

Edward’ Pipe Blowdown - Pressure Comparison

p-3080000 3.2p-3080000 3.3P-GS-5-data

Figure 2.2-2 Pressure Comparison at Left Section of Edwards Pipe Blowdown Experiment

0 0.1 0.2 0.3 0.4 0.5Time (sec)

0

0.2

0.4

0.6

0.8

1

Vapo

r Void

Fra

ction

Edward’ Pipe Blowdown - Voidg Comparison

voidg-3080000 3.2voidg-3080000 3.3voidg-GS-5-data

Figure 2.2-3 Vapor Void Fraction Comparison at Left Section of Edwards Pipe Blowdown Experiment



Heavy water was substituted for light water in the second calculation set. Figure 2.2-4 and Figure2.2-5 show the pressure and void fraction behavior, using heavy water as the medium, compared with data,although, the data are for light water. The calculated results exhibits the same behavior as it did with lightwater.

The nearly implicit advancement scheme was used in place of the semi-implicit advancementscheme in the next calculation set. The results are shown in Figure 2.2-6 and Figure 2.2-7. The MOD3.2and MOD3.3 calculations exhibited behavior similar to the semi-implicit scheme. The Henry-Fauskecritical flow model is not coupled with the nearly implicit advancement scheme. This combination is notallowed in MOD3.3 and results in a code termination. Therefore, the original critical flow model(Ransom/Trapp) was used.

A restart case was performed with the MOD3.3 code version where the problem was restarted at 0.1second and run out to 0.5 seconds. The purpose of this calculation was to show that the results of atransient calculation are replicated if a restart is done at an earlier time during the transient (with no modelchanges at restart). The results should be the same as the original transient case. The comparison betweenthe new transient case and the restart case is shown in Figure 2.2-8. The two calculations lie on top ofeach other, as expected. Thus it is shown that calculations are unchanged following restarts.

2.2.2 Dukler Air-Water Flooding Tests

Dukler and Smith2.2-4 conducted a simple flooding experiment at the University of Houston to studythe interaction between a falling liquid film with an upflowing gas core. A sketch of the test loop is shown

0 0.1 0.2 0.3 0.4 0.5Time (sec)

0

2

4

6

8

Pres

sure

(MPa

)

Edward’ Pipe Heavy - Pressure Comparison

p-3080000 3.2p-3080000 3.3P-GS-5-data

Figure 2.2-4 Pressure Comparison at Left Section of Edwards Pipe Blowdown Experiment, Heavy Water Medium



0 0.1 0.2 0.3 0.4 0.5Time (sec)

0

0.2

0.4

0.6

0.8

1Va

por V

oid F

racti

onEdward’ Pipe Heavy - Voidg Comparison

voidg-3080000 3.2voidg-3080000 3.3Voidg-GS-5-data

Figure 2.2-5 Vapor Void Fraction Comparison at Left Section of Edwards Pipe Blowdown Experiment, Heavy Water Medium

0 0.1 0.2 0.3 0.4 0.5Time (sec)

0

2

4

6

8

Pres

sure

(MPa

)

Edward’ Pipe Nearly - Pressure Comparison

p-3080000 3.2p-3080000 3.3P-GS-5-data

Figure 2.2-6 Pressure Comparison at Left Section of Edwards Pipe Blowdown Experiment, Nearly Implicit Advancement Scheme



0 0.1 0.2 0.3 0.4 0.5Time (sec)

0

0.2

0.4

0.6

0.8

1Va

por V

oid F

racti

onEdward’ Pipe Nearly - Voidg Comparison

voidg-3080000 3.2voidg-3080000 3.3Voidg-GS-5-data

Figure 2.2-7 Vapor Void Fraction Comparison at Left Section of Edwards Pipe Blowdown Experiment, Nearly Implicit Advancement Scheme

0 0.1 0.2 0.3 0.4 0.5Time (s)

0

1.25

2.5

3.75

5

6.25

7.5

Pres

sure

(Pa)

Edward’ Pipe Blowdown - Restart

p-3080000 basecasep-3080000 restart at 0.1 second

Figure 2.2-8 Edwards Pipe Blowdown Pipe Transient - Restart at 0.1 Second, Mod3.3



in Figure 2.2-9. The flow system consisted of a 1.52 m (5 ft) length of 0.051 m (2 in.) I.D. plexiglass pipeused as a calming section for the incoming air, a 0.305 m (1 ft) I.D. section of plexiglass pipe for bothintroducing the air to the test section and removing the falling liquid film, a 3.96 m (13 ft) test sectionconsisting of 0.051 m (2 in.) I.D. plexiglass pipe, and an exit section for removing the air, entrainment, andthe liquid film flowing up. Measurements were taken of the pertinent flow rates, pressure gradients, andthe liquid film thickness over a wide range of gas and liquid flow rates in the flooding region. The liquidfilm upflow, downflow, and entrainment rates were determined by weighing the liquid flow for a fixedperiod of time (see discharge lines to weigh tanks labeled B in Figure 2.2-9). Most of the instantaneousmeasured parameters oscillated once quasi-steady state conditions were reached, and it was necessary totime-average these parameters. Dukler and Smith indicate that the countercurrent flow limiting (CCFL)process is basically an unstable process that is driving the oscillations.

PT

PT

A

A

A

A

B

B

B

AIR OUT

EXIT SECTION

UPFLOW FILM UPFLOWENTRAINMENT

FLOWMETERS

PUMP

WATER TANK

AIR INLETORIFICE METERS

AIR INLETSECTION

LIQUID FEED

A indicates measuring stations, B indicates discharge to weigh tank

Figure 2.2-9 Schematic of the Dukler Air/water Test Facility



Of particular interest is the CCFL phenomenon. The experiment was modeled using the nodalizationshown in Figure 2.2-10. The homogeneous (single-velocity momentum equation) option was specified atthe air injection point (Junction 10103) to prevent liquid from flowing down the air injection pipe. Thefalling liquid film drained through Junction 10102. The forward and reverse form loss factors were set to20 for the junction and a pressure of 0.104 MPa was specified for the drain tank (Single Volume 200).

Experimental data for liquid downflow versus air flow injection rates for given liquid injection rates(WL = 0.0126, 0.0315, 0.0630, and 0.1260 kg/s) are shown in Figure 2.2-11. Dukler discussed more than

one CCFL correlation, but the one that appeared to be best for his test is a Wallis2.2-5 form of the

correlation: , where j is defined as the non-dimensional superficial velocity, subscripts g

and f refer to gas and liquid respectively, m is the slope, and C is the gas intercept constant. Thiscorrelation was found to be reasonable for air/water systems where standing waves appeared on the surface

of the liquid film. Dukler found this to occur in his experiment. Wallis2.2-5 indicated that m = 1, and Cvaried between 0.88 and 1.0. The RELAP5 input model activated the CCFL model at the junction betweenComponents 104 and 105 (see Figure 2.2-10). The CCFL input data card for this junction was used with

the following values: junction hydraulic diameter = 0.0508 m, flooding correlation form =0.0 (WallisCCFL form), gas intercept C = 0.88, and slope m = 1.0.

Figure 2.2-10 Relap5 Nodalization for Dukler’s Air/water Test Problem

TDV107

TDV103

SV 200

WATER

AIRTDJ102

AIR INLET

Pipe100

123Pipe

19012 BOTTOM PIPEWATER DRAIN PIPE

Pipe104

TEST SECTION PIPE

123456789

10

Branch 101

TDJ 106

Water Inlet

Branch105

JUNCTION FOR FLOODING TEST

Pipe108

TOP PIPE

12345678

SJ 109

TDV 110

Brj 105-1

Brj 105-2

Brj 101-1

Brj 101-2 Brj 101-3

jg1 2⁄ mjf

1 2⁄+ C=

β



MOD3.2 and MOD3.3 calculations were run with values of liquid and air injection flows consistentwith the data. Figure 2.2-12 and Figure 2.2-13 compare the calculated liquid downflow rates with data atthe given liquid injection flow rate for the MOD3.2 and MOD3.3 simulations, respectively. Goodagreement with the data is observed in both the MOD3.2 and MOD3.3 calculations. As shown, thecalculated results were in excellent agreement with the flooding correlation of Wallis. Thus the code isworking properly based on the intercept and slope values input to the model. At the higher liquid injectionrates, the MOD3.3 calculated liquid downflow was less than the data indicating more of the injected liquidwas entrained and exited through the top. A possible reason for the under-calculated liquid downflow isthat the values for the gas intercept and slope do not fit the data.

A linear regression was used to fit the data of Dukler to determine optimal constants for the Wallisflooding correlation. The fit was based on all of the data from the tests. Dukler did not fit the data to anyparticular flooding correlation, but referenced correlations developed by Wallis. The resulting constantswere 0.8915 for the gas intercept and 0.9364 for the slope. Figure 2.2-14 shows the results match the datamuch better when the new constants are used.

In summary, the CCFL model implemented in RELAP5 is working properly. Improved results wereobtained when a flooding correlation based on fitting the Dukler data was used. Code modifications toMOD3.3 have not significantly altered the calculated CCFL behavior.

0.015 0.02 0.025 0.03 0.035 0.04Air flow rate (kg/s)

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14Li

quid

film

dow

nflo

w (k

g/s)

Dukler Air/Water Tests

WL = 0.0126 kg/s WL = 0.0315 kg/s WL = 0.0630 kg/s WL = 0.1260 kg/s

Figure 2.2-11 Data for the Liquid Downflow Rate Versus Air Flow Injection Rate for Dukler’sAir/Water Problem



0.015 0.02 0.025 0.03 0.035 0.04Air flow rate (kg/sec)

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14Li

quid

film

dow

nflo

w (k

g/se

c)Dukler Air/Water Tests

MOD3.2 (0.0126 kg/s)MOD3.2 (0.0315 kg/s)MOD3.2 (0.0630 kg/s)MOD3.2 (0.1260 kg/s)WL = 0.0126 kg/secWL = 0.0315 kg/secWL = 0.0630 kg/secWL = 0.1260 kg/sec

Figure 2.2-12 Measured and Mod3.2 Calculated Liquid Downflow Comparison for Dukler’s Air/Water Problem

Filled = TestUnfilled = MOD3.2jg

*1/2 + jf*1/2 = 0.88


0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

Liqu

id fi

lm d

ownf

low

(kg/

sec)

Dukler Air/Water Tests

MOD3.3 (0.0126 kg/s)MOD3.3 (0.0315 kg/s)MOD3.3 (0.0630 kg/s)MOD3.3 (0.1260 kg/s)WL = 0.0126 kg/secWL = 0.0315 kg/secWL = 0.0630 kg/secWL = 0.1260 kg/sec

Figure 2.2-13 Measured and Mod3.3 Calculated Liquid Downflow Comparison for Dukler’s Air/Water Problem

Filled = TestUnfilled = MOD3.3jg

*1/2 + jf*1/2 = 0.88



2.2.3 Marviken Test 24

Marviken III Test 24, a full-scale critical flow test, was selected to check out and evaluate theRELAP5 choked flow model. Because of the short nozzle design (L/D = 0.33) and the long duration(about 20 seconds) of subcooling at the break, the test is particularly well-suited for validating theRELAP5 subcooled choking model. The RELAP5/MOD3.3 model is based on the Henry-Fauskecorrelation for non equilibrium and non homogeneous flows, while previous versions of the code haveused the Trapp-Ransom choked flow model, in which the Alamgir-Lienhard-Jones subcooled nucleation

correlation was used as the basis for subcooled choking.2.2-3

Marviken III Test 24 is the twenty-fourth test in a series of full-scale critical flow tests performed as

a multi-national project at the Marviken Power Station in Sweden.2.2-6 The test equipment consisted offour major components: a pressure vessel, a discharge pipe, a test nozzle, and a rupture disc assembly.

The pressure vessel was originally a part of the Marviken nuclear power plant. Of the original vesselinternals, only the peripheral part of the core superstructure, the cylindrical wall, and the bottom of themoderator tank remained. Gratings were installed at three levels in the lower part of the vessel to preventthe formation of vortices that might enter the discharge pipe. The vessel had an inside diameter of 5.22 mand was 24.55 m high as measured from the vessel bottom to the top of the top-cupola. The net available

internal volume was 420 m3.


0

0.02

0.04

0.06

0.08

0.1

0.12

0.14Li

quid

film

dow

nflo

w (k

g/se

c)Dukler Air/Water Tests

MOD3.3 (0.0126 kg/s)MOD3.3 (0.0315 kg/s)MOD3.3 (0.0630 kg/s)MOD3.3 (0.1260 kg/s)WL = 0.0126 kg/secWL = 0.0315 kg/secWL = 0.0630 kg/secWL = 0.1260 kg/secMOD3.3 (0.0126 kg/s) ModifiedMOD3.3 (0.0315 kg/s) ModifiedMOD3.3 (0.063 kg/s) ModifiedMOD3.3 (0.1230 kg/s) Modified

Figure 2.2-14 Measured and Mod3.3 Calculated Liquid Downflow Comparison for Dukler’s Air/Water Problem With Gas Constant and Slope Fitted to Data

Filled = TestUnfilled = MOD3.3small unfilled symbol

jg*1/2 + jf

*1/2 = 0.88

large unfilled symbol

jg*1/2 + 0.9364jf

*1/2 = 0.8915



The discharge pipe consisted of seven elements, including an axisymmetric inlet section, aconnection piece, two pipe stools, two instrumentation rings, and an isolation ball valve. The internaldiameters of the connection piece, pipe stools, and instrumentation rings were all 752 mm. The internaldiameter through the ball valve was 782 mm with a fairly abrupt diameter change at each end. The axialdistance from the discharge pipe entrance to the end of the discharge pipe (nozzle entrance) was 6.3 m.The test nozzle was connected to the lower end of the discharge pipe and for Test 24 the nozzle consistedof a rounded entrance section, followed by a test section 500 mm in diameter, with a length-to-diameterratio (L/D) of 0.33.

A rupture disc assembly was attached to the downstream end of the test nozzle. The assemblycontained two identical rupture discs, and the test was initiated by overpressurizing the volume betweenthe discs. This overpressure caused the outer disc to fail, which subsequently resulted in the failure of theinner disc. Failure of the discs was designed to occur along the entire periphery so that they werecompletely removed from the nozzle exit.

A schematic of the pressure vessel and discharge pipe is shown in Figure 2.2-15 (see Reference 2.2-6 for detailed drawings). The initial water level in the vessel was at an elevation of 16.7 m above thedischarge pipe inlet. A warm-up process was applied to produce a temperature profile, also shown inFigure 2.2-15. From the top of the vessel down, the fluid conditions were as follows: a steam dome(above 19.88 m) saturated at 4.96 MPa, a saturated liquid region extended for about 2 m, and a transitionregion where the temperature dropped rapidly down to 504 K at the discharge pipe inlet. The fluid at thebottom of the vessel was about 32 K subcooled relative to the steam dome temperature. The test wasinitiated by releasing the rupture disc. The ball valve started to close after 55 seconds and was fully closedat 65 seconds.

The RELAP5 nodalization is also shown in Figure 2.2-15. The vessel was represented by 39volumes and subdivided from the top as follows: one volume for the top-cupola, one volume for the steamdome, one volume for the two-phase interface region, 36 volumes of equal length (0.5 m) for the mainportion of the vessel, and one volume for the bottom of the vessel, which takes into account the standpipeentrance. All junctions in the vessel were modeled using the smooth area change option. The dischargepipe was modeled by six volumes. The third and fifth junctions of the discharge pipe were modeled usingthe abrupt area change option, while the rest were modeled with the smooth area change option. Thenozzle was modeled as a single junction with a smooth area change with no special nodalization being usedin the nozzle region. Modeling the nozzle in this manor was possible because RELAP5 includes ananalytical choking criterion, which is applied at the throat of the nozzle. The time-step control cards used0.05 seconds as the user-specified maximum time step from 0 to 5 seconds, and 0.25 seconds for theremainder of the run. The small time step in the early part was used to force the code to follow the rapidacceleration phenomena in the first part of the test.

Figure 2.2-16 shows a comparison of the measured to calculated pressure in the top of the vessel.MOD3.2 calculated the trend of the measured pressure response reasonably well. MOD3.3 under-calculated the response for the first 18 seconds. After 18 seconds, the MOD3.3 calculation agreed quitewell with the data until about 35 seconds. After 35 seconds, the calculated value deviated somewhat fromthe data.



Figure 2.2-15 Schematic, Nodalization and Initial Temperature Profile for Marviken Test 24

19.88 m24.55 m

493 533Temperature (K)

InitialTemperature Profile

NodalizationRELAP5

Marviken Vessel

DischargePipe

Test Nozzle(L/D = 0.33)

0 10 20 30 40 50Time (sec)

1000

2000

3000

4000

5000

Pres

sure

(KPa

)

Pressure in top of vesselMarviken Test 24

p-3010000 3.2p-3010000 3.3001M103

Figure 2.2-16 Measured and Calculated Pressure in the Top of the Vessel for Marviken Test 24



The delayed pressure response at the initiation of the transient is caused by the time delay associatedwith nucleation and subsequent flashing. Clearly the vapor generation modeling in MOD3.3 results in alarge nucleation delay.

The pressure response of the Marviken vessel is governed by the flashing of the hot layer of water atthe top of the vessel. The calculated pressure response is strongly affected by the numerical diffusion ofthis hot layer into the colder water below as the gradient is convected down the vessel due to massdischarge. This invariably leads to underpredicting the pressure early in the transient (the hot water coolsdue to numerical mixing) and over predicts the pressure (MOD3.2) late in the transient due to more hotwater existing at a higher temperature (due to less energy loss from earlier flashing). Generally, this effectcan be reduced by finer axial nodalization in the vessel. The numerical diffusion of the hot fluid layer willalso affect the time at which the nozzle flow transitions to two-phase choked flow.

The comparison of calculated to measured nozzle mass flow rate shown in Figure 2.2-17 agreeswell out to the initiation of two phase flow. After the initiation of two phase flow MOD3.3 over-calculatedthe mass flow rate, whereas MOD3.2 under-calculated the mass flow rate. Consequently, more mass andenergy was expelled from the vessel in the MOD3.3 calculation and thus a lower pressure as shown inFigure 2.2-16. The differences in the calculated mass flow rates appear to be related in part to the criticalflow models used in the two code versions. MOD3.2 uses the original RELAP5 critical flow model as adefault model. MOD3.3 uses the Henry-Fauske critical flow model as a default model. The default valuesfor the subcooled and two-phase discharge coefficients were used: 1.0 and 1.0 for the original RELAP5critical flow model and 1.0 for the discharge coefficient and 0.14 for the nonequilibrium parameter in theHenry-Fauske critical flow model. Slip between the two phases is not considered in the Henry-Fauskemodel. Thus more mass was expelled from the vessel in the MOD3.3 calculation. At about 47 seconds,the vessel liquid mass in the MOD3.3 calculation was nearly gone and the nozzle mass flow rate rapidlydeclined.

The density measurement in the middle of the discharge pipe is compared to the calculated value inFigure 2.2-18. The calculations showed good agreement with the data out to about 23 seconds. At about23 seconds initiation of two phase flow at the nozzle was calculated to begin. The voiding in the dischargepipe after this time appeared to be greater for the MOD3.3 calculation as evident in the large decrease inthe mixture density. After about 47 seconds the vessel was nearly voided in the MOD3.3 calculation andthe density rapidly declined.

As stated above, the liquid and vapor velocities are coupled together when choking occurs in theHenry-Fauske model as implemented in MOD3.3. However, if slip is allowed between the two phaseswhen choking occurs, the MOD3.3 results are closer to the data as shown in Figure 2.2-19 through Figure2.2-21. Investigation to include slip in the Henry-Fauske model in MOD3.3 appears warranted.

2.2.4 Marviken Test 22

Marviken III Test 222.2-7 was conducted by expelling water and steam-water mixtures from a full-size reactor vessel through a large-diameter discharge pipe that supplied the flow to a reactor pipesimulator (test nozzle). The test nozzle consisted of a rounded entrance section followed by a 500 mm,constant diameter test section having a length-to-diameter ratio of 1.5.



0 10 20 30 40 50Time (sec)

0

5000

10000

15000M

ass

Flow

Rat

e (k

g/s)

Mass flow rate at nozzle outletMarviken Test 24

mflowj-6000000 3.2mflowj-6000000 3.3FAVE

Figure 2.2-17 Measured and Calculated Mass Flow Rate at the Nozzle for Marviken Test 24

0 10 20 30 40 50Time (sec)

0

500

1000

Mixt

ure

Dens

ity (k

g/m

^3)

Mixture Density in middle of discharge pipeMarviken Test 24

rho-5030000 3.2rho-5030000 3.3003M601

Figure 2.2-18 Measured and Calculated Density in the Middle of the Discharge Pipe for Marviken Test 24



0 10 20 30 40 50Time (s)

1000

2000

3000

4000

5000Pr

essu

re (K

Pa)


p-3010000 3.2p-3010000 3.3001M103p-3010000 3.3 slip

Figure 2.2-19 Measured and Calculated Pressure (Including Phase Slip at Choked Conditions) in the Top of the Vessel for Marviken Test 24

0 10 20 30 40 50Time (s)

0

5000

10000

15000

Flow

Rat

e (k

g/s)


mflowj-6000000 3.2mflowj-6000000 3.3FAVEmflowj-6000000 3.3 slip

Figure 2.2-20 Measured and Calculated Mass Flow Rate (Including Phase Slip at Choked Conditions) at the Nozzle for Marviken Test 24



The initial steam dome pressure was 4.93 MPa, and the initial subcooling at the nozzle entrance was95 K relative to the steam dome saturation temperature. Saturation conditions were recorded in thedischarge pipe from 26 to 33 seconds. The system needed 1.2 seconds to establish a stable rate ofdepressurization. Saturation conditions were present everywhere in the discharge pipe from 33 secondsuntil the test was terminated at 48 seconds with a steam dome pressure of 2.4 MPa.

A schematic of the pressure vessel and discharge pipe is shown in Figure 2.2-22 (see Reference 2.2-7 for detailed drawings). The warmup process produced the temperature profile shown in Figure 2.2-22.A sketch of the RELAP5 nodalization is also shown in Figure 2.2-22. It is the same as Test 24 except thatthe nozzle is modeled in Test 22 because it is longer. In addition, the choking flag was turned off in allvolumes upstream in the discharge pipe. This is a common necessity when choking occurs, and it wasneeded for this model. As in the simulation of Test 24, the Henry-Fauske model for choked flow is usedfor the MOD3.3 simulation, while the Trapp-Ransom model (original choking model) is used for theMOD3.2 simulation. Discharge coefficients of 1.0 were used for both subcooled and two-phase chokedflow and the empirical nonequilibrium constant in the Henry-Fauske model was 0.14.

A comparison of the MOD3.2 and MOD3.3 calculated pressure in the top of the vessel to measureddata is shown in Figure 2.2-23. The comparison is similar to the results for Test 24. The calculated valueswere low in the beginning of the transient. After 23 seconds the calculations agreed well with the data andbeyond that time the MOD3.2 calculated pressure was higher than the data. Here again the explanation forthis characteristic is numerical diffusion of the sharp thermal gradient in the vessel as the gradient isconvected down the vessel. Figure 2.2-24 shows a comparison of the calculated mass flow rate at the

0 10 20 30 40 50Time (s)

0

500

1000De

nsity

(kg/

m3 )

Mixture Density in middle of discharge pipeMarviken Test 24

rho-5030000 3.2rho-5030000 3.3003M601rho-5030000 3.3 slip

Figure 2.2-21 Measured and Calculated Density (Including Phase Slip at Choked Conditions) in the Middle of the Discharge Pipe for Marviken Test 24



nozzle with measured data. Initially, the calculations agree with the measured data. However, after about20 seconds, the MOD3.2 calculated mass flow rate was under-calculated. The MOD3.3 calculated massflow rate continued to agree well with the data.

In summary, for Marviken Test 22, RELAP5/MOD3 results compared well with the data. To obtainthese results, it was necessary to turn choking off in the upstream discharge pipe.

2.2.5 LOFT Test L3-1 Accumulator Blowdown

This problem simulates the blowdown of the LOFT Accumulator A and surge line during Test L3-1.

Test L3-1 was a nuclear small-break experiment conducted at the LOFT Facility.2.2-8 During thisexperiment the LOFT PCS underwent a blowdown simulating a small break. As the pressure decreased at

the intact loop cold leg emergency core cooling (ECC) injection point, it became less than the pressure inthe accumulator. Flow from the accumulator was initiated injecting cold water into the primary systemcold leg. The purpose of this problem was to assess the performance of the RELAP5 accumulator modelby simulating the blowdown of the LOFT accumulator that occurred during LOFT Test L3-1 andcomparing the predicted behavior to the measured parameters.

Figure 2.2-25 is a schematic showing the arrangement of the LOFT Accumulator A and surge linerelative to the cold leg ECC injection point. The accumulator is a 1.25 m (49 in.) diameter cylindrical tankwith elliptical ends. The effective volume of liquid and gas available for injection is adjustable by varyingthe height of a standpipe inside the tank. For Test L3-1, the standpipe height was approximately 0.76 m

(30 in.), giving an effective liquid-gas volume of approximately 2.88 m3 (103 ft3). The standpipe exit is

Figure 2.2-22 Schematic, Nodalization and Initial Temperature Profile for Marviken Test 22

19.64 m24.55 m

493 533Temperature (K)

InitialTemperature Profile

NodalizationRELAP5

Marviken Vessel

DischargePipe

Test Nozzle(L/D = 1.5)



0 10 20 30 40 50Time (sec)

0

2000

4000

6000Pr

essu

re (K

Pa)


p-3010000 3.2p-3010000 3.3001M103

Figure 2.2-23 Measured and Calculated Pressure in the Top of the Vessel for Marviken Test 22

0 10 20 30 40 50Time (sec)

0

5000

10000

15000

Mas

s Fl

ow R

ate

(kg/

s)


mflowj-6000000 3.2mflowj-6000000 3.3FAVE

Figure 2.2-24 Measured and Calculated Mass Flow Rate at the Nozzle for Marviken Test 22



attached to the surge line. For Test L3-1, the length of the combined standpipe and surge line was 2.74 m

(107.7 ft), with an average flow area of 0.01 m2 (0.147 ft2). From the standpipe entrance to the primarysystem ECC injection point, there was a 2.14 m (7 ft) rise in elevation.

Figure 2.2-26 is a schematic of the RELAP5/MOD3.3 model, which consists of four components.Component 1 is the accumulator component, which represents the LOFT A Accumulator. Component 2 isa pipe component and represents the stand pipe and surge line. Component 3 is a single junction thatrepresents the connection of the surge line to the cold leg ECC injection piping. Component 4 is a time-dependent volume, which imposes the pressure history at the LOFT cold leg ECC injection point for LOFTTest L3-1. It should be pointed out that the accumulator tank volume is consistent with the accumulatorvolume above the standpipe height for Test L3-1. The surge line area and length are consistent with theTest L3-1 piping arrangement and include the entire surge line from its entrance in the standpipe to itsinjection point in the primary system cold leg. The forward and reverse loss factors represent all of thesurge line orifice, bend, and contraction/expansion losses distributed over the length of the surge line.

Calculated results are shown in Figure 2.2-27 through Figure 2.2-30, which are plots of theaccumulator gas dome pressure versus volume; the gas dome pressure versus time; the accumulator liquidlevel, and the accumulator gas temperature.

In Figure 2.2-27, curves for isothermal (cntrlvar 8) and isentropic (cntrlvar 9) expansion of the gasdome (calculated by MOD3.3 only) and the measured data from the LOFT L3-1 test are also shown forcomparison. As shown, the calculated results agree reasonably well with the data and lie between theisothermal and isentropic expansion curves, as expected. Initially the expansion is nearly adiabatic and as

Figure 2.2-25 LOFT L3-1 Accumulator A and Surgeline Schematic

Accumulator A

Standpipe

Surgeline

Primary system piping



Figure 2.2-26 RELAP5 LOFT L3-1 Accumulator Model Schematic

1

2

3

4

5

Pipe 2

TDV4

SURGELINE

Accum 1

ACCUMULATOR A

PRIMARY PIPE

1 1.5 2 2.5 3Gas Volume (m^3)

1

2

3

4

5

Pres

sure

(MPa

)

LOFT Test L3-1Accumulator Pressure vs Volume

acvdm-1 -vs- p-1010000 3.2acvdm-1 -vs- p-1010000 3.3isothermal (acvdm-1 vs cntrlvar-8 3.3)isentropic (acvdm-1 vs cntrlvar-9 3.3)l31pv

Figure 2.2-27 Measured and Calculated Accumulator Gas Dome Pressure Versus Volume, Loft Test L3-1



0 200 400 600 800 1000 1200 1400 1600Time (sec)

1.5

2

2.5

3

3.5

4

4.5Pr

essu

re (M

Pa)

LOFT Test L3-1Accumulator Pressure

p-1010000 3.2p-1010000 3.3PT-P120-043

Figure 2.2-28 Measured and Calculated Accumulator Gas Dome Pressure, Loft Test L3-1

0 200 400 600 800 1000 1200 1400 1600Time (sec)

0

0.5

1

1.5

Leve

l (m

)

LOFT Test L3-1Accumulator Liquid Level

acvliq-1 3.2acvliq-1 3.3LIT-P120-044

Figure 2.2-29 Measured and Calculated Accumulator Liquid Level, LOFT Test L3-



the gas temperature decreases due to expansion (see Figure 2.2-30), the heat transfer from the walls andliquid interface cause the expansion to approach isothermal conditions.

The accumulator pressure data from the LOFT L3-1 test are shown in Figure 2.2-28 for comparisonto the calculated values. As shown, the calculated accumulator pressure response agreed well with the datafor both versions of the code.

The measured and calculated accumulator liquid level is shown in Figure 2.2-29. Again, thecalculated response agreed reasonably well with the data. Although some differences exist between themeasured data and the calculated values, there was no difference between the two code versions. Thus themodifications made to the code between the two versions did not affect the calculated accumulatorbehavior.

The accumulator gas temperature comparison is shown in Figure 2.2-30. It appears that themeasured data is in error, perhaps due to a broken thermocouple. As the accumulator blows down, the gasdome temperature decreased due to the expansion of the gas. At about 220 seconds, the gas temperatureturned over and began to increase, possibly from wall heat transfer and energy transfer from the liquid. Itis noted the change in the gas temperature is small and appears to behave as expected. The perturbation atabout 900 seconds appears to be related to when the accumulator empties and the accumulator componentis converted to an equivalent single volume.

0 200 400 600 800 1000 1200 1400 1600Time (sec)

290

295

300

305

310Te

mpe

ratu

re (K

)LOFT Test L3-1

Accumulator Gas Temperature

tempg-1010000 3.2tempg-1010000 3.3TE-P120-041

Figure 2.2-30 Measured and Calculated Accumulator Gas Temperature, LOFT Test L3-1



In conclusion, the results of the calculation indicate that the accumulator model in MOD3 reasonablyapproximates the behavior of an accumulator out to about 900 seconds. After 900 seconds, theaccumulator emptied and the accumulator component was converted to an equivalent single volume.

2.2.6 Bennett’s Heated Tube Experiments

The Bennett heated tube experiments2.2-10 were conducted using a vertical, 1.26 cm diameter tubethat was electrically heated. Initial and boundary conditions for the tests are presented in Table 2.2-1.Water at a pressure of 6.9 MPa flowed upward in the tube. The main objectives of the experiment were tomeasure the dryout location [or critical heat flux (CHF) location] where liquid ceased to adhere to theinside wall and the surface temperature profiles in the region beyond the dryout point.

Table 2.2-1 RELAP5/MOD3 Nonequilibrium Model Developmental Assessment Matrix

Data Source and Test Conditions Primary Feature Assessed

Bennett experiment 5358Pressure = 6.9 MPa

Heat flux = 0.512 MW/m2

Mass flux = 380. kg/s-m2

Subcooling = 34.41 K

Nonequilibrium heat transferCHF correlation




Subcooling = 18.8 K







Royal Institute of Technology Test 261 Pressure = 7.0 MPa




Noneqilibrium heat transferCHF correlation

ORNL Test 3.07.9BPressure = 12.7 MPa


Mass flux.- 713. kg/s-m2





Three of the Bennett tests were simulated; low (Test 5358), intermediate (Test 5294), and high (Test5394) mass flux tests. The RELAP5 nodalization diagram for the Bennett experiments is shown in Figure2.2-31. The model consisted of a pipe with 32 vertical volumes, 31 junctions, and two time-dependentvolumes to simulate the inlet and outlet boundary conditions. The heat generated by electric power in thetest section was modeled using 32 heat slabs, and the initial and boundary conditions given by the test (seeTable 2.2-1) were input to the model.

A comparison of the calculated wall temperature and the Bennett post-CHF data for the low,intermediate, and high mass flux tests is shown in Figure 2.2-32 through Figure 2.2-34. In general, thecalculated wall temperature and CHF location are reasonable compared to the data for all three test cases.The MOD3.3 calculation showed the beginning of dryout at a lower elevation than the MOD3.2calculation. For the low mass flux case, MOD3.3 matched this location quite well. The other two casesshow the calculated dryout occurring at a higher elevation (Test 5294) or at a lower elevation (Test 5394).At elevations above the CHF point, the surface temperature was under-calculated indicating too much

ORNL Test 3.07.9NPressure = 8.52 MPa





ORNL Test 3.07.9WPressure = 12.55 MPa





ORNL Test 3.09.lOiPressure = 4.50 MPa

Heat flux - 0.38 MW/m2

Mass flux - 29.76 kg/s-m2


Axial void profile

Christensen Test 15Pressure = 5.512 MPa


Mass flux = 907.3 kg/s-m2

Inlet subcooling = 12.5 K

Subcooled nucleate boiling

Shoukri Experiment 3cPressure = 0.153 MPa

Heat Flux = 0.7237 MW/m2

Mass Flux = 450.0 kg/s-m2

Inlet Subcooling = 26.25 K

Subcooled nucleate boilinglow pressure

Table 2.2-1 RELAP5/MOD3 Nonequilibrium Model Developmental Assessment Matrix (Continued)

Data Source and Test Conditions Primary Feature Assessed



cooling was available. This may possibly be due to too much liquid entrainment calculated by the code,thus providing too much cooling capability.

2.2.7 Royal Institute of Technology Tube Test 261

The Royal Institute of Technology in Sweden obtained experimental data2.2-11 similar to the Bennettdata. The experiment used a vertical 7 m long, 0.015m diameter heated tube. The RELAP5 model wassimilar to the RELAP5 model for the Bennett experiment as shown in Figure 2.2-31 except that 47 fluidcells were used. The boundary conditions for the model are given in Table 2.2-1.

Figure 2.2-35 shows the surface temperature versus elevation for the measured data and thecalculations. The initial conditions for this experiment are quite close to the Bennett intermediate massflux experiment, thus similar calculated behavior was expected. The measured CHF position was 4.65 m.The CHF position for MOD3.2 and MOD3.3 was 5.2 m and 5.15 m, respectively. As in the Bennettexperiment, the calculated peak temperature fell below the measured peak temperature. Some possibleinfluences on the differences between the measured and calculated CHF position include the interphase

Figure 2.2-31 RELAP5 Nodalization Diagram for Bennett’s Heated Tube Experiment

TDV 110TDJ 100

Pipe 1

SJ 200 TDV 210

Heat Structure1003

1.98 m

3.56 m



0 1 2 3 4 5 6Elevation (m)

400

600

800

1000

1200Su

rface

Tem

pera

ture

(K)

Bennett - 5358Low Mass Flux Experiment

MOD3.2MOD3.3Data

Figure 2.2-32 Measured and Calculated Axial Wall Temperature Profiles for Bennett’s Heated Tube Low Mass Flux Experiment - Test 5358

0 1 2 3 4 5 6Elevation (m)

500

600

700

800

900

Surfa

ce T

empe

ratu

re (K

)

Bennett - 5294Medium Mass Flux Experiment

MOD3.2MOD3.3Data

Figure 2.2-33 Measured and Calculated Axial Wall Temperature Profiles for Bennett’s Heated Tube Intermediate Mass Flux Experiment - Test 5294



drag model, the interphase mass transfer model, and wall friction. Improvements to these models inMOD3.3 have led to a modest improvement in the comparison with the test data, as noted in Figure 2.2-35.

A new optional set of CHF correlations developed by the Nuclear Research Institute Rez in the

Czech Republic2.2-12,2.2-13 are implemented in MOD3.3 (PG-CHF Correlation). The correlations arebased on data in the Czech Republic data bank from 173 different sets of tube data, 23 sets of annular data,and 153 sets of rod bundle data. The correlations use the critical heat flux divided by the local heat flux, orthe departure from nucleate boiling ratio. There are four different formulations of the correlations (basic,flux, geometry, and power) with three different internal coefficient sets which are user selected. The"basic" form uses the local equilibrium quality and the local heat flux. The "flux" form uses the local heatflux and the heated length including the axial power peaking factor. The "geometry" form uses the localequilibrium quality and the heated length including the axial power peaking factor. The "power" formcomes from a heat balance method and can be used to calculate the critical power ratio (for a detaileddiscussion of this model see Volume 4, Section 4.3 of this document). The PG-CHF correlation isactivated by entering a 2 on heat structure cards 800 and 900 and using the 13-word format for heatstructure cards 801-899 and 901-999.

The PG-CHF correlation was activated in the MOD3.3 model of this problem using the power formof the correlation. Figure 2.2-35 shows the results of this calculation compared to the data and theMOD3.2 and MOD3.3 calculations presented above. As shown, the MOD3.3-PG calculation under-calculated the CHF position (about 4.26 m), but calculated the peak temperature and trend closer to the

0 1 2 3 4 5 6Elevation (m)

500

600

700

800

900Su

rface

Tem

pera

ture

(K)

Bennett - 5394High Mass Flux Experiment

MOD3.2MOD3.3Data

Figure 2.2-34 Measured and Calculated Axial Wall Temperature Profiles for Bennett’s Heated Tube High Mass Flux Experiment - Test 5394



data. Additional sensitivity studies are needed to understand this model and the different forms of thecorrelation.

In conclusion, RELAP5 does a reasonable job in calculating the CHF position when compared withthe data. Improvements to the MOD3.3 code version have improved the accuracy of the code incalculating CHF. Additional sensitivity studies using the PG-CHF correlation are needed to understandthis model’s feature.

2.2.8 ORNL Bundle Tests

Four ORNL tests were also used to assess the RELAP5 code and code improvements for CHF.These tests were 3.07.9B, 3.07.9N, 3.07.9W, and 3.09.10i. The first three tests examined rod dryoutbehavior (CHF). The fourth test examined void and steam temperature profiles, as well as rod walltemperature behavior.

2.2.8.1 ORNL Bundle CHF Tests. ORNL Tests 3.07.9B, 3.07.9N, and 3.07.9W2.2-14 wereperformed with an 8 x 8, full-length, electrically heated rod bundle. The rod diameter was the same sizeused in the 17 x 17 bundles for PWRs (0.0095 m). The tests were performed by adjusting the power untila steady-state dryout point had been established. Table 2.2-1 gives the test conditions. The tests areconsidered low mass flux runs.

0 2 4 6 8Elevation (m)

500

600

700

800

900Su

rface

Tem

pera

ture

(K)

RITTest 261

MOD3.2MOD3.3MOD3.3 PGData

Figure 2.2-35 Measured and Calculated Critical Heat Flux Position for the Royal Institute of Technology Tube Test 261



The RELAP5 model represented the test section as a vertical pipe divided into 24 volumes. The first23 volumes were 0.15 m long and the last volume was 0.208 m. Initial and boundary conditions for themodel are given in Table 2.2-1.

Figure 2.2-36 through Figure 2.2-38 show the measured and calculated rod surface temperatureversus elevation for each of the tests examined. The comparisons are similar to the Bennett low mass fluxcomparisons presented above. The calculated CHF position was lower in each of the tests by about 0.4 m.Generally, MOD3.3 calculated the upper axial temperature response closer to the data than did MOD3.2.Modeling improvements to the interphase drag, the interphase mass transfer, and wall friction arepotentially responsible for the improved MOD3.3 results.

2.2.8.2 ORNL Void Profile Test. ORNL tests series 3.09.102.2-15 was similar to the above tests; butthe axial void profiles and steam temperatures are reported, as well as the rod temperatures above the CHFpoint. The axial position of the CHF point was not given. The rod wall temperature at a particularelevation is the average over all the rod thermocouples at that elevation.

The measured and calculated void fraction profile is compared in Figure 2.2-39. The measured axialvoid fraction was calculated from differential pressure measurements with a reported accuracy of +3%.The calculated void fraction below 2 m was slightly higher than the measured data. However, above 2 mthe calculation agreed well with the data.

Figure 2.2-40 shows the measured and calculated rod surface temperature above the CHF point. Thedata showed a dip in the average rod wall temperature. The dip was attributed to enhanced heat transfer


600

700

800

900

1000

1100

1200

Surfa

ce T

empe

ratu

re (K

)

ORNL Bundle CHF Tests3079B

MOD3.2MOD3.3Data

Figure 2.2-36 Measured and Calculated Surface Temperature for Ornl Bundle CHF Test 3.07.9B




400

500

600

700

800

900

1000

1100

1200Su

rface

Tem

pera

ture

(K)

ORNL Bundle CHF Tests3079N

MOD3.2MOD3.3Data

Figure 2.2-37 Measured and Calculated Surface Temperature for Ornl Bundle CHF Test 3.07.9N


400

500

600

700

800

900

1000

1100

1200

Surfa

ce T

empe

ratu

re (K

)

ORNL Bundle CHF Tests3079W

MOD3.2MOD3.3Data

Figure 2.2-38 Measured and Calculated Surface Temperature for Ornl Bundle CHF Test 3.07.9W



effects downstream of a grid spacer. Since the code has no mechanism to enhance the heat transfercoefficients downstream of grids, the calculated rod temperature showed no dip. The calculated rod walltemperature generally was higher than the data, due to a higher calculated steam temperature at elevationsabove the CHF point as shown in Figure 2.2-41. The higher calculated steam temperature indicates theinterfacial heat transfer is too low. At elevations greater than 3 m, the calculated void fraction was 1.0whereas the measured void fraction indicated some liquid droplets as shown in Figure 2.2-39. At voidfractions of 1.0 the ability to cool the steam through interfacial heat transfer does not exist and the steamtemperature was calculated to be higher, thus affecting the rod wall temperature. Although the calculatedsteam temperature was higher than the data, the trend of the reported data was well represented by thecalculations.

In conclusion, the assessment of RELAP5 using ORNL data has shown the code does reasonablywell in calculating the CHF phenomenon. Improvements to the code between MOD3.2 and MOD3.3 haveenhanced the ability of the code to calculated CHF given a variety of flow and energy conditions.

2.2.9 Christensen Subcooled Boiling Test 15

The interphase mass transfer and wall heat flux partitioning model was assessed using the data from

a separate-effects, subcooled nucleate boiling experiment conducted by Christensen.2.2-16 The test sectionwas a rectangular tube with a 1.11 x 4.44 cm cross section and a 127 cm height. The tube was initially

liquid full with an inlet liquid mass flux of 907.3 kg/s-m2, a 12.5 K inlet subcooling, and a pressure of


0

0.2

0.4

0.6

0.8

1Vo

id F

ract

ion

ORNL Void Profile Test3.09.10i

MOD3.2MOD3.3Data

Figure 2.2-39 Measured and Calculated Axial Void Fractions for the Ornl Void Profile Test




400

600

800

1000

1200Ro

d Te

mpe

ratu

re (K

)ORNL Void Profile Test

3.09.10i

MOD3.2MOD3.3Tw-ave-data

Figure 2.2-40 Measured and Calculated Rod Temperature for the Ornl Void Profile Test


500

600

700

800

Stea

m T

empe

ratu

re (K

)

ORNL Void Profile Test3.09.10i

MOD3.2MOD3.3Tg-est-data

Figure 2.2-41 Measured and Calculated Steam Temperature for the Ornl Void Profile Test



5.512 MPa. The tube was heated by passing an AC current through the tube walls. The void fraction alongthe test tube was measured by a gamma densitometer.

The RELAP5 modeled test section consisted of a 20 volume vertically oriented pipe with 19junctions, two time-dependent volumes, two time-dependent junctions, and 20 heat slabs simulating heatgenerated at the walls. The test conditions given in Table 2.2-1 were used in the model.

A comparison of the RELAP5/MOD3 axial void fractions with measured data from Christensen Test15 is shown in Figure 2.2-42. The importance of the subcooled boiling model is illustrated on this testbecause the water was subcooled over the complete length of the test section. Generally, the comparisonshows good agreement between the calculations and the data. The code to code comparison showedinsignificant differences indicating improvements to the code between MOD3.2 and MOD3.3 had littleeffect on the simulation of this test. The calculated void profile is slightly higher than the data indicatingthat interphase mass transfer may be slightly over-stated by both codes.

2.2.10 Shoukri Subcooled Flow Boiling and Condensation Test.

The new model for subcooled boiling was assessed using data from the Shoukri subcooled flow

boiling and condensation experiments2.2-17. The new subcooled boiling model was implemented in thecode to consider low pressure situations. The Shoukri experiments were performed under low pressureconditions (0.153 MPa). The Shoukri experiment test apparatus consisted of a vertical concentric annulartest section through which subcooled water was pumped vertically upward. The inner tube (12.7 mmdiameter) consisted of three sections. The middle section (30.6 cm long) was electrically heated to

0 0.25 0.5 0.75 1 1.25Elevation (m)

0

0.1

0.2

0.3

0.4

0.5

0.6

Void

Fra

ctio

n

Christensen Subcooled Boiling TestTest 15

MOD3.2MOD3.3Data

Figure 2.2-42 Measured and Calculated Axial Void Fractions for the Christensen Subcooled Boiling Test 15



produce void formation in the test section. This heated section was preceded by a 34 cm long non-heatedsection and followed by a 50 cm long non-heated section. The outer tube was a 25 mm inner diameterplexiglass tube which permitted visual observation. The experiment simulated was identified as 3c(Reference 2.2-17, page 417) and the test conditions are listed in Table 2.2-1.

The RELAP5 model consisted of a 20 volume vertically oriented pipe with a heat structurerepresenting the inner annulus heated section. A time dependent volume and junction were placed at thepipe inlet to set the fluid conditions and flow rate into the pipe. A single junction and time dependentvolume were connected to the top to set the pressure boundary. The test conditions for this experiment(listed in Table 2.2-1) were used as initial and boundary conditions in the model.

Figure 2.2-43 shows the measured axial void fraction compared to the calculations. The subcooled

boiling model in MOD3.2 was mainly based on high pressure subcooled boiling data. Extrapolating highpressure data to low pressure conditions resulted in an under-calculation of the void fraction with theMOD3.2 code version. However, the implementation of the low pressure subcooled boiling model inMOD3.3 showed reasonable agreement with the data. Downstream of the heated section, MOD3.3calculated a faster bubble collapse than shown by the data. It was reported (Reference 2.2-17) that whilecondensation was taking place down stream of the heated test section bubble coalescence took placeresulting in larger bubbles. The larger bubbles resulted in a smaller interfacial surface area for heattransfer. A possible reason for the rapid calculated bubble collapse is a result of too much interfacial heattransfer due to a larger interfacial surface area. The code calculates the interfacial surface area based onthe bubble diameter which is a function of the relative velocity between the two phases. The closer the twophases are coupled together (interphase drag) the larger the bubble diameter and consequently the smaller

0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4Elevation (m)

0

0.05

0.1

0.15

0.2

0.25

Void

Fra

ctio

n

Shoukri - 3c

MOD3.2MOD3.3Data - 3c

Figure 2.2-43 Measured and Calculated Axial Void Fractions for the Shoukri 3c Experiment



the interfacial surface area becomes. It is suggested the interphase drag is too small in this calculation,leading to a higher interfacial surface area, a subsequent higher interfacial heat transfer, and a more rapidbubble collapse. An investigation into the way the code treats interphase drag is warranted.

2.2.11 MIT Pressurizer Test ST4

The Massachusetts Institute of Technology (MIT) Pressurizer Test ST42.2-18,2.2-19 involved a small-scale, low-pressure pressurizer that was initially partially filled with saturated water. The test was initiatedby opening two quick-opening valves, which resulted in the insurge of subcooled water into the bottom ofthe pressurizer. The accurate calculation of data from this test depends on accurate modeling of steamcondensation on the wall as well as interfacial heat transfer between the stratified liquid and the vaporabove the liquid.

The experimental apparatus is shown schematically in Figure 2.2-44. It consisted of two cylindricalsteel tanks: the primary tank, 1.14 m tall and 0.203 m ID, and the storage tank. The primary tank had sixwindows and was equipped with six immersion heaters with a total power of 9 kW. The storage tank waspressurized with nitrogen to force the liquid into the primary tank.

Test ST4 began with the liquid level in the primary tank at 0.432 m from the bottom. The averagelevel rise velocity was 0.0115 m/s over a 41 second time period. The primary tank was modeled with a tencell pipe, each with a heat slab set at the saturation temperature of the initial pressure. The water and steamin the tank were also set at saturation conditions. A time-dependent junction fed water at the specified ratefrom a time-dependent volume to the pipe. The time-dependent volume conditions were those of the

Figure 2.2-44 Schematic of the Experimental Apparatus for the Mit Pressurizer Test

N2

StorageTank

Water

Drain

Quick-OpeningValve

D/PCell

D/PCell

Orifice ControlValve

Steam

Water/Drain

PrimaryTank

Computer



subcooled water in the storage tank. The initial subcooling as the water entered the tank was 129 K. It wasnoted that as the liquid level in the pressurizer rose during the subcooled liquid injection, little mixingoccurred between the initial saturated fluid in the pressurizer and the incoming highly subcooled liquid.The pressurizer was insulated to diminish energy losses. Calibration tests were used to estimate the lossesat 1.1 kW.

The vessel was modelled using 10 fluid cells. A more accurate prediction could be obtained withmore cells, however, models of reactor pressurizers usually have less than 10 cells. The water level wasinitially in cell 4 (the void fraction was 0.22) and reached its maximum value in cell 8 (the void fractionwas 0.69). The experimenters did not report on the type and thickness of the insulation covering thevessel. The code model used 8.9 cm of fiber glass insulation. Steady state calculations were performed toadjust the insulation conductivity so that the steady state heat loss agreed with the reported value.

Figure 2.2-45 shows that the calculated rate of pressure rise is close to the measured value. As thecold water was injected into the pressurizer, the pressure increased due to compression of the steamvolume. As the pressure increased the saturation temperature also increased. Energy transferred from thevapor to the wall and condensation at the liquid/vapor interface mitigated the pressure rise. The calculatedpressure rise was slightly under-predicted. An atypical decrease in the calculated pressure in the MOD3.2calculation at about 36 seconds was a result of rapid condensation as subcooled liquid droplets entered asaturated steam environment near the top of the pressurizer. It is suspected the problem is associated withthe interfacial heat transfer model. However, this behavior was not exhibited in the MOD3.3 calculation.

0 10 20 30 40 50Time (sec)

0.48

0.5

0.52

0.54

0.56

0.58

0.6

Pres

sure

(MPa

)

MIT Pressurizer TestMITST4

p-3100000 3.2p-3100000 3.3Data

Figure 2.2-45 Measured and Calculated Rate of Pressure Rise for the Mit Pressurizer Test



At about 41 seconds, the flow into the pressurizer was stopped. The pressure response turned overand began to decline. The MOD3.2 calculated pressure did not decline as fast as the MOD3.3 calculatedpressure. Differences in the interfacial heat transfer package of the two code versions appear to be a factorin the behavior between the two calculations. The condensation rate calculated by MOD3.2 during thistime was much less than that calculated by MOD3.3 as shown in Figure 2.2-46 (condensation is a negativevapor generation). The smaller condensation rate resulted in a smaller depressurization rate as shown inFigure 2.2-45. The MOD3.3 calculated pressure behavior after the flow into the pressurizer was stoppedis more typical of the data. Several parameters can affect the calculated response. Thermal stratification inthe fluid appears to occur in the experiment as the liquid in the pressurizer moves up. When the flow intothe pressurizer is cut off, the pressure responds to the heat removal process at the liquid/vapor interface(which decays rapidly) and the vessel wall. Numerical mixing of the cold and saturated liquid in thecalculation resulted in a lower liquid temperature near the interface, thus affecting the calculatedcondensation/vaporization rate. A finer nodalization would minimize the numerical mixing of the liquid.Use of more optimal noding (both hydrodynamic and wall conduction) would be expected to better predictthe response of this test.

Figure 2.2-47 is a snapshot of the tank fluid and inside wall temperature at 35 seconds into thetransient. The water level is at the 0.79 m elevation. The data showed that the initial 0.432 m of saturatedwater rose with little mixing of the injected subcooled water. The calculation however, showed mixing ofthe cold and saturated liquid as a result of numerical mixing within the volume boundary. Although thefluid and wall temperature at the intermediate levels misses the actual temperature from the experiment,the trend of the data is represented.

0 10 20 30 40 50Time (sec)

-0.8

-0.6

-0.4

-0.2

0

0.2

Vapo

r Gen

erat

ion

Rate

(kg/

m3s

)


vapgen-3080000 3.2vapgen-3080000 3.3

Figure 2.2-46 Calculated Vapor Generation Rate for Volume 3080000, MIT Test ST4



A thermal front tracking model is implemented in the MOD3.3 code version (the t-flag in the volumecontrol flags) that was exercised with this experiment. This feature was included to improve the accuracyof solutions when there is warm fluid appearing above cold fluid in a vertical stack of cells (see Volume Iof this manual). Figure 2.2-48 shows the fluid and wall temperature response at 35 seconds of theMOD3.3 calculation with the thermal front tracking model activated compared to the base case MOD3.3calculation and data. As observed, the model greatly improved the calculated temperature response. Thepressure behavior, with the thermal front tracking model on, is compared with the original calculation anddata in Figure 2.2-49. The response with the thermal front tracking model on is improved during and afterthe liquid insurge. After the inflow was stopped, the pressure responded more like the MOD3.2calculation. Because of the nature of this model, little condensation occurred at the liquid/vapor interface.Additional investigation into the use of the thermal front tracking model for simulating pressurizerbehavior appears to be warranted.

2.2.12 FLECHT-SEASET Forced Reflood Tests

Forced reflood test data2.2-20, Runs 31504 and 31701 from the 161-rod FLECHT-SEASET facility,were used to assess the reflood model at low and high reflood rates. The electrically heated rodconfiguration was typical of a full-length Westinghouse 17 x 17 rod bundle. The rods had a uniform radialpower profile and a cosine axial power profile. The primary component in the test facility was a testsection that consisted of a cylindrical, low-mass housing 0.19 m (7.625 in.) inside diameter by 3.89 m(1530.0 in.) long and attached upper and lower plenums. For the tests selected, the flooding water entered

0 0.25 0.5 0.75 1 1.25Elevation (m)

300

350

400

450Te

mpe

ratu

re (K

)MIT Pressurizer Test

MITST4

Liquid Temperature 3.2Liquid Temperature 3.3Fluid Temperature - DataWall Temperature 3.2Wall Temperature 3.3Wall Temperature - Data

Figure 2.2-47 Tank fluid and Inside Wall Temperature at 35s into the MIT Pressurizer Test



0 0.25 0.5 0.75 1 1.25Elevation (m)

250

300

350

400

450Te

mpe

ratu

re (K

)MIT Pressurizer Test

MITST4

Liquid Temperature 3.3Liquid Temperature 3.3 T FrontFluid Temperature - DataWall Temperature 3.3Wall Temperature 3.3 T FrontWall Temperature - Data

Figure 2.2-48 Tank Fluid and Inside Wall Temperature at 35s into the Transient During the MIT Pressurizer Test with the Thermal Front Tracking Model Active

0 10 20 30 40 50Time (sec)

0.48

0.5

0.52

0.54

0.56

0.58

0.6

Pres

sure

(MPa

)


p-3100000 3.2p-3100000 3.3p-3100000 3.3 Thermal FrontData

Figure 2.2-49 Measured and Calculated Rate of Pressure Rise for the MIT Pressurizer Test, Thermal Front Tracking Model Active



the lower plenum at a nominal temperature of 325 K (125oF), with the rods at an initial nominal

temperature of 1140 K (1592oF), and an initial average power of 2.3 kW/m (0.7 kW/ft). The heatedcoolant exited to the upper plenum at a pressure of 0.28 MPa (40 psia).

The test section was modeled using 20 cells as shown in Figure 2.2-50. Measured fluid conditionswere used to define the conditions in the upper and lower time-dependent volumes, which represented theupper and lower plenums, respectively. The measured flow injection velocity was used to define the flowconditions at the time-dependent junction that connected the lower plenum and the pipe, which representedthe low mass housing. The measured power, which decreased during the test, was used as input to the heatstructures representing the rods. The reflood model option was not operational in MOD3.2 andconsequently not used. However, the reflood model was operational in MOD3.3 and was used.

2.2.12.1 FLECHT-SEASET Run 31504 Low Reflood Rate.

The low reflood test, Run 31504, used an injection velocity of 0.0246 m/s (0.97 in./s). Comparisonsof measured and calculated rod surface temperature histories at various elevations are presented in Figure2.2-51 through Figure 2.2-57. The legend for the rod temperature data is the rod number followed by the

Figure 2.2-50 RELAP5 Nodalization for the FLECHT-SEASET Forced Reflood Tests

123

56789

1011121314151617181920

4

TDV 7

TDV 5

TDJ 301

SJ 302

Pipe 6HS 61

A = 0.015478 m2

H = 0.1829 m



0 100 200 300 400 500 600 700Time (sec)

200

400

600

800

1000

1200

1400Cl

addi

ng T

empe

ratu

re (K

)FSET-31504Low Reflood Rate

httemp-6100407 3.2httemp-6100407 3.38N-024-data-2ft

Figure 2.2-51 Measured and Calculated Rod Surface Temperature Histories for FLECHT-SEASET Forced Reflood Run 31504 at the 0.61 m (2 ft) Elevation

0 100 200 300 400 500 600 700Time (sec)

200

400

600

800

1000

1200

1400

Clad

ding

Tem

pera

ture

(K)

FSET-31504Low Reflood Rate

httemp-6100707 3.2httemp-6100707 3.38H-048-data-4ft




0 100 200 300 400 500 600 700Time (sec)

200

400

600

800

1000

1200

1400

1600Cl

addi

ng T

empe

ratu

re (K


httemp-6101107 3.2httemp-6101107 3.37J-072-data-6ft


0 100 200 300 400 500 600 700Time (sec)

200

400

600

800

1000

1200

1400

Clad

ding

Tem

pera

ture

(K)


httemp-6101407 3.2httemp-6101407 3.38K-096-inch-data




0 100 200 300 400 500 600 700Time (sec)

200

400

600

800

1000

1200

1400Cl

addi

ng T

empe

ratu

re (K


httemp-6101607 3.2httemp-6101607 3.311E-111-inch-data

Figure 2.2-55 Measured and Calculated Rod Surface Temperature Histories for FLECHT-SEASET Forced Reflood Run 31504 at the 2.85 m (9.25 ft) Elevation

0 100 200 300 400 500 600 700Time (sec)

200

400

600

800

1000

1200

1400

Clad

ding

Tem

pera

ture

(K)



Figure 2.2-56 Measured and Calculated Rod Surface Temperature Histories for FLECHT-SEASET Forced Reflood Run31504 at the 3.08 m (10 ft) Elevation



elevation in inches; i.e., 7J-072 was from a thermocouple in a rod near the center of the bundle at the axialmidplane 1.83 m (72 in.) from the inlet. The MOD3.3 calculated rod surface temperature agreed quite wellwith the data especially below the mid-plane of the rod. Above the rod mid-plane the trend of themeasured rod temperature was calculated correctly (i.e. heatup, turn over, and quench), but the calculatedpeak temperature was below the data and turned over earlier. The MOD3.2 calculated rod temperatureexceeded the peak rod temperature with a delayed quench time. The under-calculation of the rodtemperature and the early turn over of the rod temperature above the mid-plane for MOD3.3 can beattributed to a cooler calculated vapor temperature as shown in Figure 2.2-58 through Figure 2.2-63. Thecooler calculated vapor temperature is an indicator that the interfacial heat transfer was too high. Thecalculated rod temperature behavior with the two code versions illustrates improvements in the refloodmodel as well as other model improvements such as interphase drag and entrainment.

The total bundle mass inventory is compared in Figure 2.2-64. The MOD3.3 calculated value oftotal bundle mass agreed well with the data. However, the MOD3.2 calculated total bundle mass wasbelow the measured data indicating more liquid mass exited the system. An examination of the voidfraction history throughout the test bundle, shown in Figure 2.2-65 through Figure 2.2-69, and the voidfraction profile, shown in Figure 2.2-70 through Figure 2.2-72, at different times during the experimentverified this behavior. MOD3.3 showed good agreement with the measured data, while the MOD3.2calculated void fraction history through the test bundle showed not enough coolant remaining in the systemcompared to the data.

0 100 200 300 400 500 600 700Time (sec)

200

400

600

800

1000

1200

1400Cl

addi

ng T

empe

ratu

re (K


httemp-6101907 3.2httemp-6101907 3.311E-132-data-11ft

Figure 2.2-57 Measured and Calculated Rod Surface Temperature Histories for FLECHT-SEASET Forced Reflood Run31504 at the 3.38 m (11 ft) Elevation



0 100 200 300 400 500 600 700Time (sec)

200

400

600

800

1000

1200

1400St

eam

Tem

pera

ture

(K)


tempg-6070000 3.2tempg-6070000 3.3SP7I-4-data

Figure 2.2-58 Measured and Calculated Steam Temperatures for FLECHT-SEASET Forced Reflood Run 31504 at 1.23 m (4 ft) Elevation

0 100 200 300 400 500 600 700Time (sec)

200

400

600

800

1000

1200

1400

Stea

m T

empe

ratu

re (K

)


tempg-6110000 3.2tempg-6110000 3.3SP4F-6-data




0 100 200 300 400 500 600 700Time (sec)

200

400

600

800

1000

1200

1400St

eam

Tem

pera

ture

(K)


tempg-6140000 3.2tempg-6140000 3.3SP10L-8ft-data


0 100 200 300 400 500 600 700Time (sec)

200

400

600

800

1000

1200

1400

Stea

m T

empe

ratu

re (K

)


tempg-6160000 3.2tempg-6160000 3.3SP13F-9.25ft-data

Figure 2.2-61 Measured and Calculated Steam Temperatures for FLECHT-SEASET Forced Reflood Run 31504 at 2.85 m (9.25 ft) Elevation



0 100 200 300 400 500 600 700Time (sec)

200

400

600

800

1000

1200

1400St

eam

Tem

pera

ture

(K)


tempg-6170000 3.2tempg-6170000 3.3SP131-10ft-data


0 100 200 300 400 500 600 700Time (sec)

200

400

600

800

1000

1200

1400

Stea

m T

empe

ratu

re (K

)


tempg-6200000 3.2tempg-6200000 3.3SP5K-11.5ft-data




0 100 200 300 400 500 600 700Time (sec)

0

5

10

15

20

25

30

35Bu

ndle

Fluid

Mas

s (kg


tmass-0 3.2tmass-0 3.3Data

Figure 2.2-64 Measured and Calculated Total Bundle Mass Inventory for FLECHT-SEASET Forced Reflood Run 31504

0 100 200 300 400 500 600 700Time (sec)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Void

Frac

tion


voidg-6060000 3.2voidg-6060000 3.3voidg-6070000 3.2voidg-6070000 3.3Data 3 - 4 foot elevation

Figure 2.2-65 Measured and Calculated Void Fractions at 0.92 to 1.23 m (3 to 4 ft) for FLECHT-SEASET Forced Reflood Run 31504



0 100 200 300 400 500 600 700Time (sec)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1Vo

id F

ract

ionFSET-31504Low Reflood Rate

voidg-6080000 3.2voidg-6080000 3.3Data 4 - 5 foot elevation


0 100 200 300 400 500 600 700Time (sec)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Void

Frac

tion






0 100 200 300 400 500 600 700Time (sec)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1Vo

id F

ract

ionFSET-31504Low Reflood Rate



0 100 200 300 400 500 600 700Time (sec)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Void

Frac

tion


voidg-6130000 3.2voidg-6130000 3.3Data 7 - 8 foot elevation





0

0.2

0.4

0.6

0.8

1Vo

id F

ract

ion

FSET - 31504Low Reflood Rate

MOD3.2 - 100 sMOD3.3 - 100 sData - 100 s

Figure 2.2-70 Measured and Calculated Axial Void Profile at 100s for FLECHT-SEASET Forced Reflood Run 315504


0

0.2

0.4

0.6

0.8

1

Void

Fra

ctio

n


MOD3.2 - 200 sMOD3.3 - 200 sData - 200 s




In conclusion, it has been demonstrated that improvements to the models in the code havestrengthened the code’s ability to calculate more accurately the thermal/hydraulic phenomenon associatedwith low rate reflood.

2.2.12.2 FLECHT-SEASET Run 31701 High Reflood Rate.

The high reflood test, Run 31701, used an injection velocity of 0.16 m/s (6.1 in./s). Comparisons ofmeasured to calculated rod surface temperatures for Run 31701 are presented in Figure 2.2-73 throughFigure 2.2-77. The calculated results below the mid-plane agreed well with the data. The calculated rodtemperature from the mid-plane and above show that MOD3.3 agreed with the data trend although the rodtemperature was higher than the data. The MOD3.2 calculated rod temperature at and above the mid-planeshowed a slight heatup of the rods except at the top where an earlier quench occurred. As in the discussionpresented above, MOD3.2 under-calculated the bundle fluid mass, as shown in Figure 2.2-78. With lessmass in the system, the available cooling capacity for the rods was diminished and the rods continued aslight heatup. The MOD3.3 calculated bundle mass was also lower than the data out to about 65 seconds.This under calculation of the bundle mass resulted in a rod temperature higher than the data. After 65seconds, the MOD3.3 calculated bundle mass was more in line with the measured data and the calculatedrod temperature was reasonable compared to the data.

As in the previous section, it has been demonstrated that improvements to the code between MOD3.2and MOD3.3 have strengthened the code’s ability to calculate more accurately the thermal/hydraulicphenomenon associated with high rate reflood.


0

0.2

0.4

0.6

0.8

1Vo

id F

ract

ion


MOD3.2 - 300 sMOD3.3 - 300 sData - 300 s




0 20 40 60 80 100 120 140 160Time (sec)

200

400

600

800

1000

1200Cl

addi

ng T

empe

ratu

re (K

)FSET-31701High Reflood Rate


Figure 2.2-73 Measured and Calculated Rod Surface Temperatures for FLECHT-SEASET Forced Reflood Run 31701 0.62 m (2 ft) Elevation

0 20 40 60 80 100 120 140 160Time (sec)

200

400

600

800

1000

1200

Clad

ding

Tem

pera

ture

(K)

FSET-31701High Reflood Rate


Figure 2.2-74 Measured and Calculated Rod Surface Temperatures for FLECHT-SEASET Forced Reflood Run 31701 1.22 m (4 ft) Elevation



0 20 40 60 80 100 120 140 160Time (sec)

200

400

600

800

1000

1200Cl

addi

ng T

empe

ratu

re (K


httemp-6101107 3.2httemp-6101107 3.37J-072-data-6ft

Figure 2.2-75 Measured and Calculated Rod Surface Temperatures for FLECHT-SEASET Forced Reflood Run 31701 at the 1.83 m (6 ft) Elevation

0 20 40 60 80 100 120 140 160Time (sec)

200

400

600

800

1000

1200

Clad

ding

Tem

pera

ture

(K)


httemp-6101407 3.2httemp-6101407 3.35H-096-inch




0 20 40 60 80 100 120 140 160Time (sec)

200

400

600

800

1000

1200Cl

addi

ng T

empe

ratu

re (K


httemp-6101707 3.2httemp-6101707 3.35J-120-inch


0 20 40 60 80 100 120 140 160Time (sec)

0

10

20

30

40

50

60

Bund

le Fl

uid M

ass (

kg)


tmass-0 3.2tmass-0 3.3Data

Figure 2.2-78 Measured and Calculated Total Bundle Mass Inventory for FLECHT-SEASET Forced Reflood Run 31701



2.2.13 FLECHT-SEASET Boil Off Test 35658.

The FLECHT-SEASET boil off test procedure was somewhat similar to the reflood test procedure.The bundle was filled with saturated water at the test pressure. The rod power was turned on to its presetvalue and the bundle was allowed to boil until the bundle was essentially dry. No bundle inlet flow wasprovided. The test conditions consisted of a uniform radial power profile with a rod peak power of 1.38

kw/m (0.422 kw/ft) and an initial rod temperature of 377 K (220oF), a system pressure of 0.14MPa (20

psia), and a coolant temperature of 381 K (227oF). The bundle was liquid full. These initial and boundaryconditions were applied in the RELAP5 model.

The RELAP5 model for the boil off test was similar to the model for the forced reflood, except thenumber of axial nodes in the rod bundle was reduced from 20 to 12. An examination of the calculatedresults compared to data showed a significant difference in the early part of the transient. As shown inFigure 2.2-79, void appeared almost immediately after the initiation of the transient in the calculationswhereas there was a delay in the void generation in the data. It appears the difference between thecalculations and the data may be a time delay in the heatup of the rods. A study was done where the powerto the test section was ramped over 60 seconds. These results were more reasonable. According to the test

description2.2-21, the power was turned on at its preset value at the initiation of the test. It is not fullyunderstood how the power was applied, whether it was ramped up to its preset value or turned on all atonce. The calculations where the power was ramped up will be presented below.

0 100 200 300Time (sec)

0

0.1

0.2

0.3

0.4

0.5

Gas V

oid F

racti

on

FLECHT-SEASET Boiloff test-356580-1ft.

voidg-6010000 3.2voidg-6010000 3.3Data

Figure 2.2-79 Measured and Calculated Void Fraction History at the 0 to 1ft Level for Test 35658



Figure 2.2-80 through Figure 2.2-85 show the calculated void fraction history at variouslevels in the test section compared to data. Typically the calculations represented the trend of thedata reasonably well. Early in time and at the lower levels it appears the calculated entrainmentrate is too high and thus the void fraction is over-calculated. The entrained liquid is carried upand out of the test section as evidenced by the lower calculated void fraction at elevations abovethe bottom cell during that time. This behavior also persists at later times as observed in thefigures. Simulation of the boiloff test seems to indicate that the interphase drag calculated by thecode is too large. The rate of coolant lost out the bundle top in the calculation is greater thanshown by the data. The MOD3.3 results show some improvement in this regard when comparedto the MOD3.2 results.

In summary, it was shown that RELAP5 reasonably calculated the reflood and boiloff phenomenon.Model enhancements to the code between MOD3.2 and MOD3.3 show improved behavior when comparedto data. The interphase drag model appears to be important in predicting the reflood and boiloffphenomena and additional work on this model is warranted.

2.2.14 Summary of Separate Effects Assessment

The MOD3.2 and MOD3.3 simulations of the separate effects tests examined showed reasonablerepresentation of the data. The MOD3.3 simulation of the FLECHT-SEASET low flow reflood testsshowed the most improvement over MOD3.2. Generally, model improvements enhanced the calculatedresults of MOD3.3 over MOD3.2 when compared to data. Several modeling enhancements needed to

0 100 200 300Time (sec)

0

0.1

0.2

0.3

0.4

Void

Frac

tion


voidg-6010000 3.2voidg-6010000 3.3Data




0 100 200 300Time (sec)

0

0.2

0.4

0.6

0.8

1Vo

id F

ract

ionFLECHT-SEASET Boiloff test-35658

1-2ft.

voidg-6020000 3.2voidg-6020000 3.3Data


0 100 200 300Time (sec)

0

0.2

0.4

0.6

0.8

1

Void

Frac

tion


voidg-6030000 3.2voidg-6030000 3.3Data




0 100 200 300Time (sec)

0

0.2

0.4

0.6

0.8

1Vo

id F

ract


3-4ft.

voidg-6040000 3.2voidg-6040000 3.3Data

Figure 2.2-83 Measured and Calculated Void Fraction History at the 3 to 4ft level for Test 35658

0 100 200 300Time (sec)

0

0.5

1

1.5

Void

Frac

tion


voidg-6050000 3.2voidg-6050000 3.3Data




improve the MOD3.3 response were identified. These include the interphase drag model, interphase masstransfer, the Henry-Fauske critical flow model relative to phase slip and coupling to the nearly implicitadvancement scheme, and numerical mixing.

2.2.15 References

2.2-1. A. R. Edwards and F. P. O’Brien, "Studies of Phenomena Connected with the Depressurization ofWater Reactors," Journal of the British Nuclear Energy Society, 9, 1970, pp. 125-135.

2.2-2. R. A. Reimke, H. Chow, V. H. Ransom, RELAP5/MOD1 Code Manual, Volume 3: CheckoutProblems Summary, EGG-NSMD-6182, February 1983.

2.2-3. N. Abauf, O. C. Jones, Jr., and B. J. C. Wu, Critical Flashing Flow in Nozzles with SubcooledInlet Conditions, BNL-NUREG-27512, 1980.

2.2-4. A. E. Dukler and L. Smith, Two Phase Interactions in Counter-Current Flow: Studies of theFlooding Mechanism, NUREG/CR-0617, 1979.

2.2-5. G. F. Hewitt and G. B. Wallis, Flooding and Associated Phenomena in Falling Film Flow in aTube, UKAEA Report AERE-R 4022, 1963.

2.2-6. L. Erickson et al., The Marviken Full-Scale Critical Flow Tests Interim Report: Results fromTest 24, MXC224, May 1979.

0 100 200 300Time (sec)

0

0.5

1

1.5Vo

id F

ract


5-6ft.

voidg-6060000 3.2voidg-6060000 3.3Data

Figure 2.2-85 Measured and Calculated void fraction history at the 5 to 6ft level for Test 35658



2.2-7. L. Erickson et al., The Marviken Full-Scale Critical Flow Test Interim Report: Results from Test22, MXC-222, March 1979.

2.2-8. D. L. Reeder, LOFT System and Test Description (5.5 ft. Nuclear Core/LOCEs), NUREG/CR-0247, TREE-1208, July 1978.

2.2-9. P. D. Bayless et al., Experiment Data Report for LOFT Nuclear Small Break Experiment L3-1,NUREG/CR-1145, EGG-2007, January 1980.

2.2-10. A. W. Bennett et al., Heat Transfer to Steam-Water Mixtures Flowing in Uniformly Heated Tubesin Which the Critical Heat Flux has been Exceeded, AERE-R5373, October 1976.

2.2-11. A. Sjoberg, and D. Caraher, Assessment of RELAP5/MOD2 Against 25 Dryout ExperimentsConducted at the Royal Institute of Technology, NUREG/IA-0009, October 1986.

2.2-12. R. Pernica and J. Cizek, "General Correlation for Prediction of Critical Heat Flux Ratio,"Proceedings of the 7th International Meeting on Nuclear Reactor Thermal-Hydraulics,NURETH-7, Saratoga Springs, NY, September 10-15, 1995, NUREG/CP-0142, Vol. 4.

2.2-13. R. Pernica and J. Cizek, PG General Correlation of CHFR and Statistical Evaluation Results,NRI Report, UJV-10156-T, February 1994.

2.2-14. G. L. Yoder et al., Dispersed Flow Film Boiling in Rod Bundle Geometry-Steady State HeatTransfer Data and Correlation Comparisons, NUREG/CR-2435, ORNL-5822, March 1982.

2.2-15. T. M. Anklam, R. J. Miller, and M. D. White, Experimental Investigations of Uncovered-BundleHeat Transfer and Two-Phase Mixture-Level Swell Under High-Pressure Low Heat-FluxConditions, NUREG/CR-2456, ORNL-5848, March 1982.

2.2-16. H. Christensen, Power-to-Void Transfer Function, ANL-6385, 1961.

2.2-17. M. Shoukri et al., Experiments on Subcooled Flow Boiling and Condensation in Vertical AnnularChannels, pp 413-422, Phase-Interface Phenomena in Multiphase Flow, Hemisphere PublishingCorporation, 1991.

2.2-18. H. R. Saedi and P. Griffith, "The Pressure Response of a PWR Pressurizer During an InsurgeTransient," Transactions of ANS, 1983 Annual Meeting, Detroit, MI, June 12-16, 1983.

2.2-19. H. R. Saedi, Insurge Pressure Response and Heat Transfer for a PWR Pressurizer, MIT METhesis November, 1982.

2.2-20. M. J. Loftus et al., PWR Flecht-Seaset Unblocked Bundle, Forced and Gravity Reflood TaskData Report, NUREG/CR-1532, EPRI NP-1459, WCAP-9699, June 1980.

2.2-21. S. Wong and L. E. Hochreiter, Analysis of the FLECHT-SEASET Unblocked Bundle SteamCooling and Boiloff Tests, NUREG/CR-1533, EPRI NP-1460, WCAP-9729, January 1981.


INTEGRAL TEST PROBLEMS

2.3 INTEGRAL TEST PROBLEMS

RELAP5/MOD3.3 simulations of five integral problems are presented in this section. They areLOFT small-break Test L3-7, LOFT large break Test L2-5, Semiscale natural circulation Tests S-NC-2and S-NC-3, and a ZION-1 PWR postulated small break transient. The first four tests are presented withcode calculations compared to data while the last problem is a hypothetical plant accident and acomparison is presented between RELAP5 calculations using MOD3.2 and MOD3.3.

2.3.1 LOFT Small-Break Test L3-7

The Loss-of-Fluid Test (LOFT) Facility2.3-1 was a 50-MWt, volumetrically scaled PWR system.The LOFT facility was designed to obtain data on the performance of the engineered safety features of acommercial PWR system for postulated accidents including LOCA’s.

The LOFT nuclear core was approximately 1.68 m in length and 0.61 m in diameter and wascomposed of nine fuel assemblies, containing 1300 fuel rods of representative PWR design. Threeunbroken PWR coolant loops were simulated by using a volume/power ratio scaled by the singlecirculating (intact) loop in the LOFT primary system, and the postulated broken PWR loop was simulatedby the scaled LOFT blowdown (broken) loop (see Figure 2.3-1).

The LOFT broken loop was orificed to simulate various break sizes and contains steam generator andpump simulators to model the hydraulic resistance of these components. Either hot-leg (reactor vesseloutlet piping) or cold-leg (reactor vessel inlet piping) breaks could be simulated by relocating the steamgenerator and pump simulators. Quick-opening valves (with opening times adjustable from approximately20 to 50 ms) simulated the initiation of primary coolant piping ruptures. Primary blowdown effluent wascollected in a blowdown suppression tank, which was used to simulate the significant portions of thebackpressure transient for various PWR containments.

An emergency core cooling (ECC) system was provided to simulate the loss-of-coolant engineeredsafety features in PWRs. The ECC was supplied by a high-pressure injection system (HPIS) centrifugalpump and a nitrogen-pressurized accumulator. The low pressure injection system (LPIS) and accumulatordischarge lines were orificed as required to simulate the delivery characteristics of various PWRemergency coolant injection systems. The accumulator was equipped with an adjustable height standpipe,which allowed the liquid and gas volumes to be varied. Five ECC injection points were built into theprimary coolant system (PCS). These injection points were located in the intact loop hot leg, intact loopcold leg, upper plenum, lower plenum, and vessel downcomer.

Fluid pressure, temperature, velocity, and density were monitored by extensive instrumentation atkey locations in the primary coolant, emergency core coolant, blowdown, and secondary coolant systems.Thermocouples monitored fuel rod cladding temperatures and support tube temperatures at 196 corelocations. Four fixed nuclear detectors and a four-location traversing in-core nuclear detector systemdetermined core power profiles and transient response.

LOFT small break Test L3-72.3-2 was performed to analyze the effects of a single-ended, offset shearbreak of a small [2.54-cm (1-in.) diameter] pipe connected to the cold leg of a large, 4-loop PWR. The testwas conducted at 49 MW, yielding a maximum linear heat generation rate of 52.8 kW/m.



Figure 2.3-1 Schematic of LOFT Test Facility



The primary objectives of Test L3-72.3-2 were to establish a break flow approximately equal to HPISflow when the primary pressure was in the range of 6.9 MPa, to establish conditions for steam generatorreflux cooling, to isolate the break and stabilize the plant at cold shutdown condition, and to analyze thedata obtained to investigate associated phenomena.

Prior to opening the break, the nuclear core was operating at a steady-state maximum heat generationrate of 52.8 + 3.7 kW/m. Other significant initial conditions for Test L3-7 were: system pressure, 14.90 +0.25 MPa; core outlet temperature, 576.1 + 0.5 K; and intact loop flow rate, 481.3 + 6.3 kg/s. At 36seconds after the break occurred, the reactor scrammed on a low system pressure signal. Within 10seconds after scram verification, the pumps were manually tripped and coasted down. Pump coastdownwas followed by the inception of loop natural circulation. Between 1800 (30 min) and 5974 seconds (1 h,40 min), the HPIS was turned off to hasten the loss of fluid inventory and to establish the conditionsconsidered favorable for reflux flow in the primary loop. Starting at 3600 seconds (1 h),operator-controlled steam bleeding (opening the main steam bypass valve early and the main steam valvelater in the transient) and steam generator feeding (using both the auxiliary and main feedwater systems)were used to decrease primary system pressure. Steam generator secondary feed and bleed maintained aneffective heat sink throughout the experiment.

Later in the experiment, at 7302 seconds (2 h, 2 min), the blowdown isolation valve was closed,which isolated the break. System mass depletion stopped, and all decay heat energy not lost to theenvironment was removed by the steam generator. Primary system pressure gradually increased, causingthe fluid in the system to become subcooled. Subsequently, the purification system was used to bring thereactor to a cold shutdown condition and the experiment was terminated.

The RELAP5 simulation of LOFT Test L3-7 is presented here to assess the ability of MOD3.3 tocalculate the important parameters in a small-break transient for a full system test and to compare withMOD3.2 results.

This RELAP5 input model was developed from the LOFT base model2.3-3 The L3-7 model containsa filler gap (small flow path parallel to the downcomer), as this was found to be important in small breaks.The RELAP5 model of the LOFT facility for Test L3-7 included 129 fluid control volumes and 135 flowjunctions. The system nodalization is illustrated schematically in Figure 2.3-2 and Figure 2.3-3. In themodel, a total of 137 heat slabs were used to represent heat transfer in the intact loop steam generator,vessel, core, primary system piping, and pressurizer. The values of the two-phase and subcooled dischargecoefficients for the break used in the input deck were both 1.0. The nonequilibrium factor for theHenry-Fauske critical flow model used in MOD3.3 was 0.14 (default value).

The input deck for this developmental assessment was modified to include crossflow junctions infour places: the pressurizer, the pressurizer spray, the inlet annulus, and the upper head. The connectionsbetween these four places and the primary system piping, as well as the leak path between the inlet annulusand the upper head, were modeled as crossflow junctions.

A steady-state control system developed by the LOFT group was used in conjunction with this modelto obtain a steady state. Additional components, such as pump speed controllers, a pressurizer spray valvecontroller, pressurizer heaters, a pressurizer level controller using charging and letdown components, and asteam valve controller, were provided in the RELAP5 model. The steady-state controller allows the user to



Figure 2.3-2 RELAP5 Nodalization for LOFT Test L3-7: Vessel and Broken Loop

SV

260

Br

255

Br

252

Br

250 Br

245SV

251

Br

240

123

456 3 2 1

Pip

e23

0

Annulus 210

Pipe 235

Br

225

Br

215

SV

220

Annulus 2231 2 3 4

Br 2

05

Br

202

Br

200 S

J 22

4

Brj

200

-1

1 2 3 54 6

Brj

202

-1

Brj

202-

2

Brj

Brj

215

-1

Brj

215

-2Brj

215

-3

Brj

225

-1B

rj 2

25-2

Brj

240

-1B

rj 2

40-2

Brj

245

-1Brj

245

-2B

rj 2

50-1

Brj

252

-1Brj

255

-1

Brj

255

-2

205-

1

Br

335

Br

340

Br

345

Pip

e 35

0V

alve

SV

360

Val

ve

Pip

e 37

0

SJ

375

Pip

e 38

0B

r 30

0B

r 30

5B

r 31

0Pipe

315

12

123456

7 8 9 1011

12

12

3

32

1

355

365

TD

V

805

Brj

300-

1 Brj 335-1

Brj

Brj

Brj

Brj

Brj

Brj

Brj

Brj

310-

231

0-1

305-

130

0-2

335-

234

0-134

5-1

345-

2



Figure 2.3-3 RELAP5 Nodalization for LOFT Test L3-7: Intact Loop

Valve995

Valve985

TDV990

TDV980

Valve600

Valve610

Letdown and Charging

Accum620

Pipe615

SV605

TDV635

TDJ640

TDV625

TDJ630

LPISHPIS

Vessel

Vessel

202223

252Br 100Br 105

Br 107Brj 100-1

Brj100-2

Brj105-1Brj

107-1

Brj110-1

Br 110Pipe 11212

BrBrj114-1

Brj114-2

Pipe

1

2

3

45

6

7

8

114

115

Br116

Brj116-1

Brj116-2

Pipe 118

1

2

3

Br 120

Brj120-1

Brj120-2

Brj120-3

Br125

SV130

Pump135

SV 140

Br125-1

Br 145

Brj145-1

Brj145-2

Br 150

Brj150-1

Brj150-2

Br155

Brj155-1

SV160

Pump165 Br 170

Br 172

Brj172-1

Pipe 1751 2

Br 180

Brj180-1

Br 185Brj185-1

Brj185-2

Br 400

Brj 400-1

Brj 400-2

Pipe 4051

2

SJ 410

Pipe 415

123456

SJ 417

Pipe 42012

Br430

Brj430-1

Valve435

500-1

500-3

503-2

503-1 500-2Br

503 Separatr500Br

505

Br 520Brj 525-1

Br 525Brj 525-1

SV 530

505-1

Br

508-1

SJ513

Pipe515

1

2

3

4

5

508

Annulus510

Valve540

Br541

TDV542

Valve545

TDV546

TDV 565TDJ566

TDV 568

TDJ569

Main Feed

Aux Feed

MainSteam

Bypass



specify seven set points: primary system mass flow rate, pressurizer pressure, pressurizer level, primarysystem cold leg temperature, steam valve relative steam position, steam generator level, and feed massflow rate. A successful steady state was run out to 200 seconds, and the set points for the L3-7 test werereached. A user guideline that came to light during the RELAP5/MOD2 assessment is that the chokingoption should be turned off in the secondary system steam valve.

The transient calculation was carried out to 1000 seconds. Figure 2.3-4 shows a comparison of cputime versus simulated time for the MOD3.2 and MOD3.3 calculations. A SUN Workstation was used toperform these calculations. Generally, the MOD3.3 calculation ran slightly faster than the MOD3.2calculation.

Figure 2.3-5 shows a comparison between the measured and code calculated primary systempressures. In general the comparison showed good agreement with the data. Between 200 and 500seconds the decrease in pressure was delayed in the calculations. The two code calculations agreed wellwith each other. A comparison of the secondary system pressure is shown in Figure 2.3-6. Although thesecondary system pressure was slightly over calculated, the calculated response agreed well with the data.The difference between the calculated secondary pressure and the data is similar to the difference seen withprevious versions of the code and probably due to generic modeling problems with the secondary side ofthe steam generator.

Figure 2.3-7 compares the measured and calculated liquid velocity and Figure 2.3-8 compares themeasured and calculated vapor velocity in the hot leg intact loop. The comparisons show good agreement.Figure 2.3-9 and Figure 2.3-10 shows the comparison for the measured and calculated liquid temperature

0 100 200 300 400 500 600 700 800 900 1000Simulated Time (sec)

0

20

40

60

80

100

120

140

160

180

200

CPU

Tim

e (s

ec)

LOFT Test L3-7CPU Time

cputime-0 3.2cputime-0 3.3

Figure 2.3-4 CPU Time Versus Simulated Time for LOFT Test



0 100 200 300 400 500 600 700 800 900 1000Time (sec)

0

2

4

6

8

10

12

14

16

18

20Pr

essu

re (M

Pa)

LOFT Test L3-7Primary system pressure

p-100010000 3.2p-100010000 3.3PE-PC-002

Figure 2.3-5 Measured and Calculated Primary System Pressure for LOFT Test L3-7

0 100 200 300 400 500 600 700 800 900 1000Time (sec)

5

6

7

8

Pres

sure

(MPa

)

LOFT Test L3-7Secondary system pressure

p-515050000 3.2p-515050000 3.3PE-SGS-001

Figure 2.3-6 Measured and Calculated Secondary System Pressure for LOFT Test L3-7



at the core inlet and outlet respectively (upper plenum to lower plenum). The calculated response agreed

well with the data, although the temperature was slightly under-calculated near the end of the transient.Figure 2.3-11 compares the measured and calculated mass flow rate at the break. The agreement is good.The oscillatory behavior of the measured data is not exhibited by the calculated response and dose notappear to be propagated to other measured parameters. Finally, a comparison of the hot leg intact loopdensity is shown in Figure 2.3-12. The code calculated value for the density tended to be higher than thedata.

In summary, the code calculated response showed good agreement with some of the key parametersof the L3-7 test out to 1000 seconds. No significant differences exist between the MOD3.3 and theMOD3.2 calculated results for LOFT L3-7.

2.3.2 LOFT Large-Break Test L2-5

The LOFT Test L2-52.3-4 was performed to establish the effects of a 200% double-ended cold legbreak with an immediate primary coolant pump trip. The test was conducted at 36 MW, yielding amaximum linear generation rate of 40.1 kW/m. The LOFT facility is described in the preceding discussionof the LOFT L3-7 test.

The initial conditions for significant parameters for Test L2-5 were: systems pressure, 14.94 + 0.06MPa; core outlet temperature 589.7 + 1.6 K; and intact loop flow rate, 192.4 + 7.8 kg/s.

0 100 200 300 400 500 600 700 800 900 1000Time (sec)

-1

0

1

2

Junc

tion

liqui

d ve

locity

(m/s

)LOFT Test L3-7

Liquid velocity in the intact loop hot leg

velfj-112010000 3.2velfj-112010000 3.3PNE-PC-002

Figure 2.3-7 Measured and Calculated Liquid Velocity in the Intact Loop Hot Leg for LOFT Test L3-7



0 100 200 300 400 500 600 700 800 900 1000Time (sec)

-1

0

1

2Ju

nctio

n va

por v

eloc

ity (m

/s)

LOFT Test L3-7Vapor velocity in the intact loop hot leg

velgj-112010000 3.2velgj-112010000 3.3PNE-PC-002

Figure 2.3-8 Measured and Calculated Vapor Velocity in the Intact Loop Hot Leg for LOFT Test L3-7

0 100 200 300 400 500 600 700 800 900 1000Time (sec)

500

520

540

560

580

600

Volum

e liq

uid

tem

pera

ture

(K)

LOFT Test L3-7Liquid temperature at core inlet

tempf-225010000 3.2tempf-225010000 3.3TE-4LP-003

Figure 2.3-9 Measured and Calculated Liquid Temperature at the Core Inlet for LOFT Test L3-7



0 100 200 300 400 500 600 700 800 900 1000Time (sec)

500

520

540

560

580

600Vo

lume

liqui

d te

mpe

ratu

re (K

)LOFT Test L3-7

Liquid temperature at core outlet

tempf-240010000 3.2tempf-240010000 3.3TE-4UP-001

Figure 2.3-10 Measured and Calculated Liquid Temperature at the Core Outlet for LOFT Test L3-7

0 100 200 300 400 500 600 700 800 900 1000Time (sec)

0

1

2

3

Mas

s flow

rate

(kg/

s)

LOFT Test L3-7Break mass flow rate

mflowj-365000000 3.2mflowj-365000000 3.3FR-BL-111

Figure 2.3-11 Measured and Calculated Mass Flow Rate at the Break for LOFT Test L3-7



The experiment was initiated by opening the quick-opening blowdown valves in the broken loop hotand cold legs thus simulating a 200% double-ended cold leg break (see Figure 2.3-1). The reactorscrammed on low pressure at 0.24 + 0.01 seconds. Following the reactor scram, the operators tripped theprimary coolant pumps at 0.94 + 0.01 seconds. The pumps were not connected to the flywheels during thecoastdown.

The fuel rods began to heat up almost immediately following the break initiation (approximately 1second). A rewet of the upper portion of the center fuel assembly began at approximately 12 seconds andended at approximately 23 seconds. Accumulator injection of ECC to the intact loop cold leg began at16.8 + 0.1 seconds. Delayed ECC injection from the HPIS and LPIS began at 23.90 + 0.02 and 37.32 +0.02 seconds respectively. The fuel rod peak cladding temperature of 1078 + 13 K was attained at 28.47 +0.02 seconds. The cladding was quenched at 65 + 2 seconds, following the core reflood. The LPISinjection was stopped at 107.1 + 0.4 seconds, after the experiment was considered complete.

The RELAP5 model of the LOFT facility for Test L2-5 included 129 fluid control volumes and 139flow junctions. The system nodalization is illustrated schematically in Figure 2.3-13 and Figure 2.3-14.In the model, a total of 32 heat slabs were used to represent heat transfer in the intact loop steam generatorand core. The values of the break two-phase and subcooled discharge coefficients were 0.93 and 0.84,respectively for the MOD3.2 calculation which used the original RELAP5 critical flow model. For theMOD3.3 calculation, which used the Henry-Fauske choked flow model, a single value of 0.84 is used forthe discharge coefficient in both two-phase and subcooled choked flow. In addition, the empirical

0 100 200 300 400 500 600 700 800 900 1000Time (sec)

400

600

800

1000To

tal d

ensit

y (k

g/m

^3)

LOFT Test L3-7Density in the intact loop hot leg

rho-112020000 3.2rho-112020000 3.3Avg DE-PC-002-A-B-C

Figure 2.3-12 Measured and Calculated Density in the Intact Loop Hot Leg for LOFT Test L3-7



nonequilibrium factor of the Henry-Fauske model is set to 0.14. This deck contains a split downcomer,which was found to be important for large breaks.

The input deck was modified previously to include crossflow junctions in the same places as theL3-7 deck. The only difference is that the L2-5 deck doesn’t contain a pressurizer spray for the transientcalculation (just for steady state; i.e., volume 430 and junction 435), so there is no need for a crossflowjunction there. The deck was also modified to include a 12-volume core rather than a 6-volume core. Itwas felt that this finer nodalization was necessary to obtain a good comparison with the data.

As with L3-7, the steady-state control system was used to obtain a good steady state. The sameseven set points were used, except that the values were changed to match the L2-5 test specifications. Asuccessful steady state was run out to 867 seconds.

The transient calculation was carried out to 50 seconds. Comparison plots are shown for MOD3.2,MOD3.3, and the data. Figure 2.3-15 and Figure 2.3-16 show the calculated primary and secondarypressure compared to measured data. The calculated primary pressure for both code versions agreed fairlywell with the data. The calculated pressure was slightly lower than the data for the first 20 seconds andafter 20 seconds the calculated pressure was slightly higher than the data. The two calculated pressureresponses are nearly identical and neither show the knee seen in the data and also predicted with previousversions of the code. This difference is probably due to the choked flow model since primary systempressure is directly affected by break flow and the MOD3.3 predicted flow in the broken loop cold leg isquite different than the data or the MOD3.2 calculation. The calculated MOD3.3 secondary pressureagreed quite well with the data. The trend of the calculated MOD3.2 secondary pressure response agreedwith the data, however, the initial secondary pressure was lower and this trend continued throughout thetransient. The difference between the MOD3.2 and MOD3.3 secondary pressure is most likely a result ofthe new subcooled boiling model that was added to MOD3.3.

Figure 2.3-17 through Figure 2.3-20 show the MOD3.3 and MOD3.2 calculated mass flow rates inthe broken loop cold leg, broken loop hot leg, intact loop cold leg, and intact loop hot leg, each comparedto the data. The flow spikes in the broken loop cold leg mass flow rate between 20 and 40 seconds were aresult of slugs of liquid passing through to the break. They occurred during the time of core refloodingwhile the accumulators were injecting. The calculated spikes were more pronounced than shown in thedata. The calculated mass flow rates compared well with the measured data with two exceptions. First, theMOD3.3 calculated mass flow in the broken loop cold leg at the beginning of the transient showed asignificantly different trend than either the MOD3.2 calculation or the data. This difference is most likelydue to changes in the choked flow model between MOD3.2 and MOD3.3.

Second, the calculated intact loop hot leg mass flow rate between 5 and 25 seconds showed a largereversal with both codes (see Figure 2.3-20) whereas the data appears to show a large positive flow aroundthe loop. The instrument used to measure the flow only measured the magnitude and not its direction.However, the flow direction can be inferred by the measured differential pressure shown in Figure 2.3-21.Positive flow is indicated by a positive differential pressure and reverse flow is indicated by a negativedifferential pressure. The absolute values of the flow rates shown in Figure 2.3-22 give a more directindication of the agreement between the measured and calculated results. The calculated intact loop hotleg flow was positive (toward the steam generator) until about 4 seconds. The flow reversed, going towardthe reactor vessel, due to the pump trip and subsequent flow coastdown. Flashing in the steam generator



Figure 2.3-13 RELAP5 Nodalization for LOFT Test L2-5: Vessel and Broken Loop

SV

260

Br

255

Br

252

Br

250 Br

245SV

251

Br

240

123

Pipe 235

Br

225

Br

215

SV

220

SJ 7

09

Brj

215

-2

Brj

215

-3

Brj

225

-1B

rj 2

25-2

Brj

240

-1B

rj 2

40-2

Brj

245

-1Brj

245

-2B

rj 2

50-1

Brj

Brj

255

-1

Brj

255

-2

Br

335

Br

340

Br

342

Pipe

370

SJ 3

75

Pipe

380

Br

300

Br

305

Br

310P

ipe

315 123

4 5 6 78

12

3

32

1

TD

V 8

05B

rjB

rj

Brj

Brj

Brj

Brj

Brj

Brj

310-

231

0-1

305-

130

0-2

335-

234

0-134

2-1

344-

1

Val

ve31

7

TD

V80

0

Br

344

Val

ve34

7P i p e 2 3 0

SJ

719

Brj

252-

225

2-1

710

740

712

714

716

718

Ann

Ann

Ann

Ann

Ann

Ann

SJ

711

SJ

741

SJ

713

SJ

715

SJ

717

Ann

700

Ann

730

Ann

702

Ann

704

Ann

706

Ann

708

SJ

701

SJ

703

SJ

705

SJ

707

SJ

731

Brj

300-

1

Brj

335-

1 Not

e: T

he A

nnul

us c

ompo

nent

s re

pres

ent t

he d

ownc

omer

a

nd a

re in

terc

onne

cted

with

cro

ssfl

ow ju

ncti

ons

at

ea

ch a

xial

leve

l (no

t sho

wn)

.

VE

SSE

L

BR

OK

EN

LO

OP



Figure 2.3-14 RELAP5 Nodalization for LOFT Test L2-5: Intact Loop

Valve995

Valve985

TDV990

TDV980

Valve600

Valve610

Letdown and ChargingAccum620

Pipe615

SV605

TDV635

TDJ640

TDV625

TDJ630

LPISHPIS

Vessel

Vessel

252Br 100Br 105

Br 107Brj 100-1

Brj100-2

Brj105-1Brj

107-1

Brj110-1

Br 110Pipe 11212

BrBrj114-1

Brj114-2

Pipe

1

2

3

45

6

7

8

114

115

Br116

Brj116-1

Brj116-2

Pipe 118

1

2

3

Br 120

Brj120-1

Brj120-2

Brj120-3

Br125

SV130

Pump135

SV 140

Br125-1

Br 145

Brj145-1

Brj145-2

Br 150Brj170-1

Brj150-1

Br155

Brj155-1

SV160

Pump165 Br 170

Pipe 175

1 2 Br 180

Brj180-1

Br 185

Brj185-3

Brj185-2

Br 400

Brj 400-1

Brj 400-2

Pipe 4051

2

SJ 410

Pipe 415

123456

SJ 417

Pipe 42012

Br430

Brj430-1

Valve435

500-1

500-3

503-2

503-1 500-2Br

503 Separatr500Br

505

Br 520Brj 525-1

Br 525Brj 525-1

SV 530

505-1

Br

508-1

SJ513

Pipe515

1

2

3

4

5

508

Annulus510

Valve540

Br541

TDV

542

TDV 565TDJ566

Main Feed

MainSteam

730SV602

Brj185-1



0 10 20 30 40 50Time (sec)

0

5

10

15Pr

essu

re (M

Pa)

LOFT Test L2-5Primary system pressure

p-250010000 3.2p-250010000 3.3PE-1UP-001A1

Figure 2.3-15 Measured and Calculated Primary System Pressure for LOFT Test L2-5

0 10 20 30 40 50Time (sec)

5

5.5

6

Pres

sure

(MPa

)

LOFT Test L2-5Secondary system pressure

p-530010000 3.2p-530010000 3.3PE-SGS-001

Figure 2.3-16 Measured and Calculated Secondary System Pressure for LOFT Test L2-5



0 10 20 30 40 50Time (sec)

-250

0

250

500

750M

ass

Flow

Rat

e (k

g/s)

LOFT Test L2-5Mass flow in the broken loop cold leg


Figure 2.3-17 Measured and Calculated Mass Flow Rate in the Broken Loop Cold Leg for LOFT Test L2-5

0 10 20 30 40 50Time (sec)

-100

0

100

200

Mas

s Fl

ow R

ate

(kg/

s)

LOFT Test L2-5Mass flow in the broken loop hot leg


Figure 2.3-18 Measured and Calculated Mass Flow Rate in the Broken Loop Hot Leg for LOFT Test L2-5



0 10 20 30 40 50Time (sec)

-200

-100

0

100

200

300M

ass

Flow

Rat

e (k

g/s)

LOFT Test L2-5Mass flow in the intact loop cold leg

mflowj-180010000 3.2mflowj-180010000 3.3FR-PC-101

Figure 2.3-19 Measured and Calculated Mass Flow Rate in the Intact Loop Cold Leg for LOFT Test L2-5

0 10 20 30 40 50Time (sec)

-100

0

100

200

300

Mas

s Fl

ow R

ate

(kg/

s)

LOFT Test L2-5Mass flow in the intact loop hot leg

mflowj-100010000 3.2mflowj-100010000 3.3FR-PC-201

Figure 2.3-20 Measured and Calculated Mass Flow Rate in the Intact Loop Hot Leg for LOFT Test L2-5



0 10 20 30 40 50Time (s)

-50

-40

-30

-20

-10

0

10

20

30

40

50D

iffer

entia

l Pre

ssur

e (kP

a)

LOFT Test L2-5Measured intact loop hot leg DP

PDE-PC-003

Figure 2.3-21 Measured Intact Loop Hot Leg Differential Pressure; LOFT Test L2-5

0 10 20 30 40 50Time (sec)

-100

0

100

200

300

Mas

s Fl

ow R

ate

(kg/

s)

LOFT Test L2-5Mass flow in the intact loop hot leg

abs (mflowj-100010000) 3.2abs (mflowj-100010000) 3.3abs (FR-PC-201)

Figure 2.3-22 Absolute Value of Measured and Calculated Mass Flow Rate in the Intact Loop Hot Leg for LOFT Test L2-5



u-tubes forced flow toward the reactor vessel resulting in the peak negative flow at about 10 seconds in thecalculations. Draining of the pressurizer also contributed to the flow from the intact loop to the vessel. Asimilar flow reversal almost ceartinly occurred in the experiment based on the comparisons shown inFigure 2.3-22 and the reversal in the measured differential pressure shown in Figure 2.3-21. The trends ofthe data were calculated reasonably well by MOD3.2 and MOD3.3. The oscillatory behavior exhibited inthe data after 35 seconds (corresponding to the time of core reflood) was not exhibited in the calculations.

The calculated and measured density in the intact loop hot leg is shown in Figure 2.3-23. Thecalculated density agreed well with the measured data.

The accumulator level and the mass flow rate from the accumulator are shown in Figure 2.3-24and Figure 2.3-25. The calculated trend for both MOD3.2 and MOD3.3 agree fairly well with themeasured data. However, the under calculated primary pressure (see Figure 2.3-15) resulted in an earlieractuation of the accumulator and beyond 20 s the higher pressure resulted in a slower rate of emptying.

The pump speed for primary coolant pump 2 (hydrodynamic volume 165) is shown in Figure2.3-26. The agreement is quite good until about 15 seconds. After 15 seconds the calculated MOD3.2pump speed began to decrease again. The calculated MOD3.3 pump speed did not begin to decrease againuntil about 20 seconds. The measured pump speed began to increase after 25 seconds. During thetransient, the pumps were tripped about 1.2 seconds after transient initiation and allowed to coastdown.The modeled pumps were tripped at 1.75 seconds and allowed to coastdown based on the torque-inertiaequation. The measured data shows a positive coolant flow increase around the loop (see Figure 2.3-19and Figure 2.3-20) that caused the pump angular velocity to increase. This loop flow increase was not

0 10 20 30 40 50Time (sec)

-500

0

500

1000

Tota

l Den

sity

(kg/

m^3

)

LOFT Test L2-5Density in the intact loop hot leg

rho-100010000 3.2rho-100010000 3.3DE-PC-205

Figure 2.3-23 Measured and Calculated Density in the Intact Loop Hot Leg for LOFT Test L2-5



0 10 20 30 40 50Time (sec)

0.5

1

1.5

2

2.5Le

vel (

m)

LOFT Test L2-5Accumulator liquid level

cntrlvar-59 3.2cntrlvar-59 3.3LE-ECC-01A

Figure 2.3-24 Measured and Calculated Accumulator Level for LOFT Test L2-5

0 10 20 30 40 50Time (sec)

0

50

100

Mas

s Flow

Rat

e (k

g/s)

LOFT Test L2-5Mass flow from the accumulator

mflowj-610000000 3.2mflowj-610000000 3.3

Figure 2.3-25 Calculated Mass Flow Rate from the Accumulator for LOFT Test L2-5



observed in the calculations. A sensitivity calculation was performed which set the pump friction torque tozero (zero value for the friction torque coefficients on pump input card 302). The results showed that thepump velocity increased, driven by the loop flow. This suggests that the user input friction torquecoefficients in the model may be in error. It was noted that changing the torque friction did notsignificantly change the loop behavior.

The upper and lower plenum liquid temperature comparisons are shown in Figure 2.3-27 and Figure2.3-28 respectively. The calculated liquid temperature was at the saturation temperature corresponding tothe primary system pressure. The deviation between the calculated temperatures and the measuredtemperatures reflects the difference in the calculated and measured primary system pressure.

A comparison of the fuel centerline temperature at 0.69 m (27 in.) above the bottom of the core isshown in Figure 2.3-29. The calculated trend of the centerline temperature agreed fairly well with the dataout to about 30 seconds when the calculated cooling rate increased. This did not occur until about 44seconds for the measured data. Although not shown, the centerline temperature data began to drop at afaster rate around 60 seconds. The earlier calculated cooloff was probably a result of the accumulatordischarging sooner.

Figure 2.3-30 through Figure 2.3-35 compare the calculated cladding temperatures to the measuredtemperatures at various elevations above the bottom of the core. At the lower elevations the calculatedclad temperatures for MOD3.2 and MOD3.3 are in good agreement, but both are lower than the data. Atthe 0.13 m elevation, shown in Figure 2.3-30, a quench occurs about 10 s earlier than shown by the data.At the core mid plane both MOD3.2 and MOD3.3 predict heatups and maximum temperatures in

0 10 20 30 40 50Time (sec)

-100

0

100

200

Angu

lar V

eloc

ity (r

ad/s

ec)

LOFT Test L2-5Primary coolant pump 2 speed

pmpvel-165 3.2pmpvel-165 3.3RPE-PC-002

Figure 2.3-26 Measured and Calculated Pump Speed for Primary Coolant Pump 2 (Hydrodynamic Volume 165) for LOFT Test L2-5



0 10 20 30 40 50Time (sec)

300

400

500

600Vo

lum

e Liq

uid

Tem

pera

ture

(K)

LOFT Test L2-5Upper plenum liquid temperature

tempf-250010000 3.2tempf-250010000 3.3TE-4UP-001

Figure 2.3-27 Measured and Calculated Upper Plenum Temperature Below the Nozzle for LOFT Test L2-5

0 10 20 30 40 50Time (sec)

300

400

500

600

Volu

me

Liquid

Tem

pera

ture

(K)

LOFT Test L2-5Lower plenum liquid temperature

tempf-215010000 3.2tempf-215010000 3.3TE-4LP-001

Figure 2.3-28 Measured and Calculated Lower Plenum Temperature for LOFT Test L2-5



0 10 20 30 40 50Time (sec)

0

500

1000

1500

2000M

esh

Poin

t Tem

pera

ture

(K)

LOFT Test L2-5Fuel centerline temperature - .69 m

httemp-232000501 3.2httemp-232000501 3.3TC-5D07-27

Figure 2.3-29 Measured and Calculated Fuel Centerline Temperature 0.69m (27in.) Above the Bottom of the Core for LOFT Test L2-5

0 10 20 30 40 50Time (sec)

200

600

1000

Mes

h Po

int T

empe

ratu

re (K

)

LOFT Test L2-5Fuel clad temperature - 0.13 m

httemp-232000110 3.2httemp-232000110 3.3TE-5I06-005

Figure 2.3-30 Measured and Calculated Fuel Cladding Temperature 0.13m (5in.) Above the Bottom of the Core for LOFT Test L2-5



0 10 20 30 40 50Time (sec)

250

750

1250M

esh

Poin

t Tem

pera

ture

(K)




0 10 20 30 40 50Time (sec)

400

600

800

1000

1200

Mes

h Po

int T

empe

ratu

re (K

)


httemp-232000510 3.2httemp-232000510 3.3TE-5D07-027




0 10 20 30 40 50Time (sec)

400

600

800

1000M

esh

Poin

t Tem

pera

ture

(K)




0 10 20 30 40 50Time (sec)

200

400

600

800

1000

Mes

h Po

int T

empe

ratu

re (K

)






agreement with the data, but the maximum temperatures occur earlier than shown by the data in Figure2.3-32. At all the higher elevations two heatups followed by quench occur in the data, while only oneseparate modest heatup and quench are shown in the calculations (see Figure 2.3-33 and Figure 2.3-34).In general there is no significant difference between MOD3.3 and MOD3.2. MOD3.3 results show noheatups at the two highest elevations. This is in contrast to the data in which two consecutive heatups andquench are shown at both locations. The reason for the under-predicted rod temperature in the upperportion of the core may be two-fold. First, the interphase drag may be too high, thus providing an effectivecooling medium for the rods at the top of the core. Second, the reflood model may be over-predicting theheat transfer between the rods and the coolant. Further improvements to these models appears to bewarranted.

In summary, for the complexity of these types of transients, the code reasonably calculated thetransient behavior. Specific differences between the measured and calculated values noted were related tothe way the code handles choked flow and interphase drag. It is clear that further improvements areneeded in the reflood model.

2.3.3 Semiscale Natural Circulation Tests S-NC-2 and S-NC-3

Natural circulation experiments were performed in the Semiscale Mod-2A test facility, a small-scalemodel of the primary system of a four-loop PWR nuclear power generating plant (scaling factor 1/1705).The Mod-2A system incorporates the major components of a PWR including steam generators, vessel,downcomer, pumps, pressurizer, and loop piping. Detailed descriptions of the Semiscale Mod-2A testfacility and operation procedure are given in Reference 2.3-6 and Reference 2.3-7.

0 10 20 30 40 50Time (sec)

200

400

600

800M

esh

Poin

t Tem

pera

ture

(K)


httemp-232001110 3.2httemp-232001110 3.3TE-5H07-058




The natural circulation experiments reported here used a single-loop configuration, including theintact loop with steam generator and vessel/downcomer, as shown in Figure 2.3-36. In the single-loopconfiguration, the intact loop pump was replaced with a spool piece containing an orifice that simulated thehydraulic resistance of a locked pump rotor. In addition, the vessel was modified from the normal Mod-2Aconfiguration for all the experiments by removing the vessel upper head to ensure a uniform heatup of theentire system and to avoid condensation in upper-head structures (the upper vessel was capped).

Ten test series made up the Semiscale natural circulation experiments. The various phenomenasimulated in these test series included: transition from single-phase to two-phase to reflux naturalcirculation modes by varying primary side system mass; effect of steam generator secondary side massinventory on two-phase natural circulation; effect of secondary conditions on reflux; effect ofnoncondensable gas (nitrogen) on both reflux and two-phase, single, and two-loop effects; transient smallbreak LOCAs; and ECCS effects. Tests S-NC-2 and S-NC-3 were used to assess the MOD3.2 andMOD3.3 versions of the code. The S-NC-2 test examined single-phase, two-phase, and reflux steady statemodes by varying the primary side system mass at different core powers (30, 60, 100 kW) with a constantsteam generator secondary side condition. The S-NC-3 test examined primary side two-phase naturalcirculation behavior under varying steam generator secondary side mass inventory (defined as Case 3 inthis test) at a core power of 62 kW.

The RELAP5 model used for the natural circulation study was a modified version of the standardSemiscale Mod-2A RELAP5 model representing the system configuration for the S-NC Test Series.Figure 2.3-37 shows a schematic of the model, which includes the following modifications to the standardmodel:

1. The broken loop piping, pump, and steam generator and the intact loop pump were removed.2. The upper head volumes were removed.3. The pressurizer was removed to reflect actual operating conditions for all measured conditions except

a liquid-full system. Calculations for a liquid-full condition employed a time-dependent pressure boundary rather than a pressurizer.

4. The crossflow junction was utilized, connecting the primary loop piping to the vessel upper plenum, the vessel downcomer upper annulus, and the pressurizer surge line.

5. The secondary system steam separator was renodalized to represent its actual location at the top of the steam generator riser.

6. An annular bypass volume is included, which allows communication between the steam generator upper downcomer and the steam dome.

7. The steam dome was renodalized as two volumes connected by a crossflow junction. This configura-tion is a more precise geometrical representation of the physical steam dome in the Semiscale system.

8. The inside surfaces of the primary system pressure boundaries were modeled as adiabatic boundaries in lieu of modeling piping guard heaters, metal mass, and mechanistic heat loss.

The model consisted of 113 hydrodynamic volumes and 69 heat structures. All volume parameters

were calculated with nonequilibrium code models. The specified core axial power profile2.3-7 wasmodeled in 12 consecutive heat structures over six hydrodynamic volumes.

The results of the S-NC-2 test comparison are presented first. The core power level used in thisassessment study was 60 kW. At this power level, 17 different steady-state conditions were achieved.



Figure 2.3-36 Semiscale Mod-2A Single-loop Configuration

Steam Generator

Pressurizer

Vessel Top Cap

Hot Leg

Vessel

Cold Leg

Pump ReplacementSpool

Downcomer

PumpSuction



Figure 2.3-37 Schematic of RELAP5/MOD3 Natural Circulation Test Model

Br 640

TDV650

Valve 634Valve 635

SV 613

Br 612SV 611

Separatr 601Br605

Annulus600

Annulus602

Annulus603

Br604

Pipe220

1234

7

8

9

1011

887

76

6 65 5 5

4 43 32 21 1

10

1112131415161718

9TDJ610

TDV630

TDV695

TDV696

TDJ691

TDJ690

Fill/Drain

Feedwater

1231231

2

3

4

5 6

7

8

91 2 1 2

123456789

10

123456

Br 163

SV 165

Br 162

Br 161

Pipe150

SV 140

Br 130

SV 120

Pipe110

Br101 Br

185

Br184

Br182

Pipe 201Pipe 203

Br993

Valve991

TDV989

Br 202

SJ205

SJ235

Br 215SV 210

Br 225

SV 230

Pipe240

Pipe 261 Br 262 Pipe 263

SJ250

TDV95

TDV96

TDJ290

TDJ291

Fill/Drain

Intact Loop



Each steady-state was at a different primary side mass inventory that ranged from 100% down to 61.2%full. The primary system pressure was allowed to vary as the system was drained (the pressurizer wasvalved out) while the steam generator secondary pressure was held constant and acted as an effective heatsink throughout the test. It was noted from the experiment that for mass inventories between 94% and100% the primary system behavior was similar to single-phase natural circulation. It was also noted thatwhen the primary side mass inventory was decreased below 66 to 70% the system transitioned fromtwo-phase natural circulation to the reflux mode of cooling. The reflux mode of natural circulation is atwo-phase state roughly characterized by a vapor-continuous condition entering the inlet of the steamgenerator primary with counter-current liquid flow returning to the vessel.

Figure 2.3-38 through Figure 2.3-41 show the calculated results compared to data for mass flowrate, hot leg fluid temperature, primary side steam generator outlet fluid temperature, and primary systempressure. The calculated results compare well with the data in the 97% to 100% mass inventory range. In

the two-phase region (between 70% and 94% of inventory) the mass flow rate increased with a decrease inthe mass inventory out to 92% for the data and about 88% for the calculations. The increased mass flowrate was a result of boiling in the core. As a consequence of boiling in the core a larger density differencebetween the core and the steam generator (heat sink) existed and the mass flow rate increased. Furtherreduction in the mass inventory caused void formation in the steam generator (down side of the U-tubes)and subsequently the mass flow rate decreased due to a smaller density difference between the core and thesteam generator. The calculated mass flow rate showed a slower increase with decreasing mass inventory.There may be several reasons for the slower increase in the calculated mass flow rate, but the major factorappeared to be related to the interphase drag. It is suspected that the interphase drag allowed more liquid tobe carried up in the hot leg, thus affecting the density head difference and resulting in a slower mass flow

60 70 80 90 100Primary Coolant System Inventory (%)

0

0.2

0.4

0.6

0.8

1

Mas

s Fl

ow R

ate

(kg/

s)

Semiscale Natural Circulation TestS-NC-2

MOD3.2MOD3.3Data

Figure 2.3-38 Measured and Calculated Primary System Mass Flow Rate at the 60kW Core Power for Test S-NC-2




550

560

570

580Fl

uid

Tem

pera

ture

(K)


MOD3.2MOD3.3Data

Figure 2.3-39 Measured and Calculated Primary System Hot Leg Fluid Temperature at the 60kW Core Power for Test S-NC-2


546

547

548

549

550

551

Flui

d Te

mpe

ratu

re (K

)


MOD3.2MOD3.3Data

Figure 2.3-40 Measured and Calculated Primary Side Steam Generator Outlet Fluid Temperature at the 60kW Core Power for Test S-NC-2



rate. The slower mass flow rate in turn affected the fluid temperature and pressure response as shown inFigure 2.3-39 through Figure 2.3-41. MOD3.3 showed improvement in the system response at lowermass inventories during the two-phase natural circulation mode. The calculated response appearedreasonable during the reflux mode (between 60% and 70% mass inventory). The calculated mass flow wasoscillatory in this mode showing both positive and negative values. The mean value of the oscillation isshown. It is noted the MOD3.3 calculation failed at the primary side mass inventory of 61%.

The calculated results of the S-NC-3 test are presented next. There were three cases involved withthis test. The first case involved variations in core power and primary system mass inventory with aconstant secondary side mass inventory. The second case involved variation in primary system mass at aconstant secondary side mass inventory and a core power at 62.6kW. The third case was at a constant corepower of 62kW and a constant primary system mass inventory 91.8% of full value (two phase flow) with avariation in secondary side mass inventory. There were a series of 9 steady state conditions achieved forCase 3. By varying the steam generator secondary mass, the effective heat transfer area from primary tosecondary was changed. The code assessment calculations were for the Case 3 steady state series.

Figure 2.3-42 through Figure 2.3-45 compare the calculated results with data for mass flow rate, hotleg fluid temperature, primary side steam generator outlet fluid temperature, and primary system pressureversus the steam generator effective heat transfer area. The measured mass flow rate and the fluidtemperature at the primary side steam generator outlet in the region from 15% to 45% steam generatoreffective surface area was oscillatory. The minimum and maximum of the oscillations observed are shownin Figure 2.3-42 and Figure 2.3-44. The calculated mass flow rate compared well with the data ateffective heat transfer areas above 55% as observed in Figure 2.3-42. Typically, MOD3.3 showed


6

7

8

9

10

11Pr

essu

re (M

Pa)


MOD3.2MOD3.3Data

Figure 2.3-41 Measured and Calculated Primary System Pressure at the 60kW Core Power for Test S-NC-2



0 10 20 30 40 50 60 70 80 90 100SG U-Tube Heat Transfer Area (%)

0

0.2

0.4

0.6

0.8M

ass

Flow

Rat

e (k

g/s)

Semiscale Natural Circulation TestS-NC-3, case 3

MOD3.2MOD3.3DataData max oscillationsData min oscillations

Figure 2.3-42 Measured and Calculated Primary System Mass Flow Rate Versus Steam Generator Secondary Side Heat Transfer Area for Test S-NC-3


550

555

560

565

570

575

580

Flui

d Te

mpe

ratu

re (K

)


MOD3.2MOD3.3Data

Figure 2.3-43 Measured and Calculated Primary Side Hot Leg Fluid Temperature Versus Steam Generator Secondary Side Heat Transfer Area for Test S-NC-3




540

550

560

570

580Fl

uid

Tem

pera

ture

(K)


MOD3.2MOD3.3DataData max oscillationData min oscillation

Figure 2.3-44 Measured and Calculated Primary Side Steam Generator Outlet Temperature Versus Steam Generator Secondary Side Heat Transfer Area for Test S-NC-3


6

7

8

9

10

Pres

sure

(MPa

)


MOD3.2MOD3.3Data

Figure 2.3-45 Measured and Calculated Primary System Pressure Versus Steam Generator Secondary Side Heat Transfer Area for Test S-NC-3



improvement over MOD3.2. Subsequently, the calculated hot leg and primary side steam generator outletfluid temperature and the primary system pressure compared well with the data as shown in Figure 2.3-43through Figure 2.3-45. As the secondary side mass inventory was further reduced, the effective heattransfer surface area was degraded (less than 55%) thus minimizing the heat removal capability of thesteam generator. Consequently, the driving potential (density difference between the core and the steamgenerator) was diminished and the mass flow rate slowed. The calculated mass flow at the reduced steamgenerator effective heat transfer surface area was higher than the data. This implies a higher calculateddensity driving head than indicated by the data. It is suspected that the calculated liquid entrainmentcarried more liquid over the top of the steam generator U-tubes and thus maintained the higher densitydifference to drive the flow. The higher mass flow rate resulted in lower calculated fluid temperatures anda lower primary system pressure as shown in Figure 2.3-43 through Figure 2.3-45.

In conclusion, RELAP5/MOD3 simulated the Semiscale natural circulation tests reasonably well forthe higher PCS mass inventories. Also, at the higher steam generator mass inventories, the codecalculations are in good agreement with the measured data. Typically, MOD3.3 showed improvementover MOD3.2. It appears the interphase drag model allowed too much liquid to be entrained thus affectingthe results of the calculation. Further investigation into the interphase drag model is warranted.

2.3.4 Zion-1 PWR Small Break

The Zion-1 PWR plant is a Westinghouse 4-loop PWR, and the RELAP5 model has been used to

analyze loss of offsite power2.3-8 and instrument tube rupture2.3-9 scenarios. The model has been modifiedto remove proprietary information and to model a 2% [0.1 m (4 in)] cold leg break. Since the RELAP5/MOD3 development, this model has been used as a quality assurance test problem that is run when newversions of RELAP5 are created. The RELAP5 nodalization diagram is shown in Figure 2.3-46 andFigure 2.3-47. The model contains 139 volumes, 142 junctions, and 83 heat structures. Two primarycoolant loops were modeled. One loop, called the broken loop, represented a single primary coolant loop.The break was modeled in the pump discharge piping of the broken loop cold leg. The other loop, calledthe intact loop, represents three primary coolant loops lumped as one. The pressurizer was attached to theintact loop. The intact and broken loops were modeled symmetrically except for differences due to thelocation of the break and pressurizer. Component numbers for the intact loop were between 100 and 194;component numbers for the broken loop were between 200 and 294. The Zion-1 nodalization is similar toSemiscale and LOFT models. Heat structures were used to represent heat transfer from fuel rods, U-tubes,pressure vessel wall, vessel downcomer wall, core shroud, and internals in the upper head and lower andupper plena.

The transient was initiated by an instantaneous opening of the break. This was accomplished byusing a trip valve, which was open for times greater than 0.01 seconds. A scram signal was generatedwhen the pressurizer pressure decreased to 12.82 MPa (1860 psia). Scram occurred 3 to 4 seconds after thescram signal was generated. The reactor coolant pumps began coasting down simultaneously with thescram signal. Valves in the steam generator feedwater and steam lines began closing simultaneously withthe scram signal. The steam line and feedwater valves were linearly closed within 1 second and 10seconds, respectively, of the scram signal. Safety injection and charging began 5 seconds after thepressurizer pressure decreased to 12.62 MPa (1830 psia). Secondary side auxiliary feedwater flow wasinitiated 14 seconds after the scram signal was generated. Automatic control of the feedwater flow basedon steam generator downcomer liquid level was simulated.



Figure 2.3-46 RELAP5 Nodalization for the Zion-1 PWR: Vessel Model

SnglVol 323

Branch 322

Branch 330

SnglVol 325

Pipe335

SnglJun 336

Pipe320

Annulus315

SnglVol 340Triple LoopCold Leg

Pipe118

Pipe100

Triple LoopHot Leg

Pipe200

Pipe214

Single LoopHot Leg

Single LoopCold Leg

Branch 305

Branch 345Branch 300

Branch 355

Pipe310

Pipe350

Pipe356



Figure 2.3-47 RELAP5 Nodalization for the Zion-1 PWR: Loop Model

158

TDV

Pipe150

1

2

3

4

5

6

Valve 157PORV

Pressurizer

Pipe152

SJ 151

Pipe 10012

Br 102Pipe 10412

SJ 107SV 106

SJ 105

Pipe 108

1

2

3

45

6

8

7

SJ 109SV 110

SJ111

Pipe 112

1

2

3

4

5

Pump

113

Pipe 114 Br 116 Pipe 1181 2 300

345

VesselUpperPlenum

VesselDowncomer

SJ380

SJ385

Accum190

TDV

193

TDV

191

TDJ 194TDJ 192

Charging SI

1 1

2 2

3 3

4 4

Pipe170

5

6

Steam Generator

Separatr 171

SV 178

Br 180

TDV188TDJ

187

TDV186 Valve

185

Main Steam

TDV182

TDV

184

TDJ181TDJ183

MainFeed

AuxFeed

Pipe176

Br174

SV172

Triple Loop

Note: The single loop is similar only without the pressurizer. The single loop numbering scheme uses 200’s. The break is located in the single loop cold leg.

Hot Leg

Cold Leg

PumpSuction



Figure 2.3-48 through Figure 2.3-54 show comparisons for MOD3.2 and MOD3.3 results forselected parameters. Figure 2.3-48 shows the pressure in the upper plenum (Volume 345) and is

representative of the primary system pressure. Energy removed at the break and by the steam generatorsexceeded the core decay heat and stored energy in the vessel heat structures and the system pressuredecreased. At about 400 seconds, the MOD3.2 calculated break mass flow rate spiked to about 3900 kg/sas a slug of coolant exited the system (see Figure 2.3-49), resulting in a more rapid decrease in the systempressure as can be seen in Figure 2.3-48. This rapid decrease in system pressure resulted in an increase ofthe safety injection flow rate, further cooling down the system. The MOD3.3 calculation did not exhibitthis behavior. The system pressure was driven by the break mass flow rate and choking at the break planeoccurred in both calculations. One of the significant differences between the two code calculations was thedefault break flow model. MOD3.2 used the original RELAP5 critical flow model, whereas MOD3.3 usedthe Henry-Fauske critical flow model. Differences between these two models were noted earlier in Section2.2.3 where the addition of phase slip in the Henry-Fauske model showed improvements when comparedto the data. Other differences are improvements in the interfacial drag terms in the sum and differencemomentum equations, formerly user option 47, which have been made default in MOD3.3. The MOD3.2calculated system pressure continued to decrease at a faster rate than the MOD3.3 calculated systempressure. At about 550 seconds accumulator injection was initiated in the MOD3.2 calculation as shown inFigure 2.3-50 and Figure 2.3-51. The MOD3.2 calculated break flow again increased, more systemenergy was removed, and a more rapid depressurization rate resulted. The accumulators emptied atapproximately 700 seconds in the MOD3.2 calculation. At that time, the break mass flow rate declinedand the mass flow rate from the safety injection and charging systems exceeded the break flow rate and thesystem pressure began to rise. The accumulators in the MOD3.3 calculation did not begin injecting until

0 400 800 1200Time (sec)

0

5

10

15

20

Pres

sure

(MPa

)Typical PWR - Typ1200

Primary System Pressure

p-345010000 3.2p-345010000 3.3

Figure 2.3-48 MOD3.2 and MOD3.3 Primary System Pressure (Core Outlet) Comparison for a Small Break Transient in a Typical PWR



about 800 seconds. By that time, the mass flow into the system from the safety injection and chargingsystem was nearly equal to the break mass flow rate and the system depressurization rate declined.

The void fraction at the core outlet is shown in Figure 2.3-52. This figure indicates that more liquidremained in the upper core volume in the MOD3.3 calculation than in the MOD3.2 calculation.

The steam generator secondary side pressure response is shown in Figure 2.3-53. The response ofthe secondary followed that of the primary system due to reduced primary to secondary heat transfer. Theonly mass entering the secondary side was from the auxiliary feed system. The secondary pressure did notrise enough to activate the secondary side pressure relief valve in either calculation.

The mass error for the two calculations is shown in Figure 2.3-54. Although mass error correctionshave been made in the code, other modifications to improve the calculational stability and accuracy of thecode have affected the mass error for this particular problem. Compared to the total mass of the systemthe mass error represents only 0.02% and is insignificant.

2.3.5 References

2.3-1. D. L. Reeder, LOFT System and Test Description (5.5 ft. Nuclear Core/LOCEs), NUREG/CR-0247, TREE-1208, July 1978.

2.3-2. D. L. Gillas and J. M. Carpenter, Experimental Data Report for LOFT Nuclear Small BreakExperiment L3-7, NUREG/CR-1570, EGG-2049, August 1980.

0 400 800 1200Time (sec)

0

1000

2000

3000

4000

Flow

Rat

e (k

g/s)

Typical PWR - Typ1200Break Mass Flow Rate

mflowj-505000000 3.2mflowj-505000000 3.3

Figure 2.3-49 MOD3.2 and MOD3.3 Break Mass Flow Rate Comparison for a Small Break Transient in a Typical PWR



0 400 800 1200Time (sec)

0

20

40

60

80Ac

cum

ulat

or L

iquid

Volum

e (m

^3)

Typical PWR - Typ1200Accumulator Blowdown

acvliq-190 3.2acvliq-190 3.3

Figure 2.3-50 MOD3.2 and MOD3.3 Intact Loop Accumulator Liquid Volume Comparison for a Small Break Transient in a Typical PWR

0 400 800 1200Time (sec)

0

10

20

30

Accu

mul

ator

Liqu

id Vo

lume

(m^3

)

Typical PWR - Typ1200Accumulator Blowdown

acvliq-290 3.2acvliq-290 3.3

Figure 2.3-51 MOD3.2 and MOD3.3 Broken Loop Accumulator Liquid Volume Comparison for a Small Break Transient in a Typical PWR



0 400 800 1200Time (sec)

0

0.2

0.4

0.6

0.8

1G

as V

oid F

racti

onTypical PWR - Typ1200

voidg-345010000 3.2voidg-345010000 3.3

Figure 2.3-52 MOD3.2 and MOD3.3 Core Outlet Void Fraction Comparison for a Small Break Transient in a Typical PWR

0 400 800 1200Time (sec)

0

1

2

3

4

5

6

7

Pres

sure

(MPa

)

Typical PWR - Typ1200Secondary System Pressure

p-180010000 3.2p-180010000 3.3

Figure 2.3-53 MOD3.2 and MOD3.3 Intact Loop Steam Generator System Pressure (Steam Dome) Comparison for a Small Break Transient in a Typical PWR



2.3-3. E. J. Kee, P. J. Schally, L. Winters, Base Input for LOFT RELAP5 Calculation, EGG-L0FT-5199,July 1980.

2.3-4. P. D. Bayless and J. M. Divine, Experimental Data Report for LOFT Large BreakLass-of-Coolant Experiment L2-5, NUREG/CR-2826, EGG-2210, August 1982.

2.3-5. S. L. Thompson and L. N. Kmetyk, RELAP5 Assessment: LOFT Large Break Loss-of-CoolantExperiment L2-5, NUREG/CR-3608, SAND83-2549, January 1984.

2.3-6. System Design Description for the Mod-A Semiscale System, Addendum 1, "Mod-A Phase IAddendum to Mod-3 System Design Description," EG&G Idaho, Inc., December 1980.

2.3-7. R. A. Dimenna, RELAP5 Analysis of Semiscale Mod-A Single-Loop Single-ComponentSteady-State Natural Circulation Tests, EGG-SEMI-6315, June 1983.

2.3-8. C. D. Fletcher et al., Lose of Offsite Power Scenarios for the Westinghouse Zion-I PressurizedWater Reactor, EGG-CAAP-5156, May 1980.

2.3-9. C. D. Fletcher and M. A. Bolander, Analysis of Instrument Tube Ruptures in Westinghouse4-Loop PWRs, NUREG/CR-4672, EGG-2461, December 1986.

0 400 800 1200Time (sec)

-50

0

50

100

150M

ass

Erro

r (kg

)Typical PWR - Typ1200


Figure 2.3-54 MOD3.2 and MOD3.3 Mass Error Comparison for a Small Break Transient in a Typical PWR


CONCLUSION

3 CONCLUSION

The results of this assessment demonstrate that, with exceptions to be noted, the MOD3.3 version ofRELAP5 produces qualitatively correct and more accurate simulations than earlier versions. In addition,RELAP5/MOD3.3 is more robust than previous versions and requires less user intervention to achieve asatisfactory result.

The exceptions include: (1) simulations using the nearly implicit numerical option, which producesqualitatively incorrect results when used in combination with the default Henry-Fauske choked flow modeland (2) simulation of rapid depressurization processes wherein the pressure recovery is under predicted.In the first case it is clear from results of the Edward’s pipe blowdown problem that the Henry-Fauskechoked flow model is not coded to the nearly implicit numerical scheme. Further investigation will beneeded to find and correct this deficiency. In simulations of the Edward’s pipe problem and the Marvikentests 22 and 24, it was evident that following the initial rapid depressurization the vessel pressure did notrecover as quickly as indicated by the data and by simulations using the MOD3.2 version of the code.Clearly the MOD3.3 changes to the nucleation and vapor generation modeling has produced a slowerresponse to pressure changes. However, this does not appear to produce any adverse effect on the morecritical safety parameter predictions such as peak clad temperature.

3.1 Phenomenological Problems

Without exception, simulations of the ten phenomenological problems using RELAP5/MOD3.3produced results in as good or better agreement with the physics of each situation than has been attainedusing prior versions of the code. The nine volume problem (Section 2.1.1) demonstrated that thegravitational effect and kinematic models for phase continuity provide a qualitatively correct result.Compared to MOD3.2, the results indicate that the interphase friction modeling in MOD3.3 produces lessinterphase drag so that the liquid falls to the bottom of the pipe more quickly. Analytical solutions to thisproblem for specific interphase drag models do not exist and so a quantitative assessment is not possible.For the idealized case of free fall, the liquid phase falls to the bottom of the pipe more quickly thanpredicted by MOD3.3, a qualitatively correct result.

The MOD3.3 simulation of the manometer problem (Section 2.1.2) produced a result having a periodin agreement with the analytical model for this problem, 4.5 seconds. The amplitude damping was verysmall and much smaller than obtained with MOD3.2. This result is in good qualitative agreement withlarge tube liquid-vapor manometer experience.

The branched flow problems (Section 2.1.3 and Section 2.1.4) illustrated that the unphysicaloscillations and mass error experienced with the MOD3.2 code version have been eliminated and MOD3.3produces a qualitatively more correct result. No unphysical recirculation developed in two-dimensionalnodalizations having a liquid-vapor interface.

Results for the three stage turbine simulation (Section 2.1.6) again showed that the improvements inMOD3.3 result in reduced mass error. The predicted pressure and torque at each stage were essentially thesame for the MOD3.3 and MOD3.2 versions of the code.

Phenomenological Problems 153 NUREG/CR-5535/Rev 1-Vol III

CONCLUSION

The results for Workshop Problems 2 and 3 (Section 2.1.7 and Section 2.1.8) showed that while thetransient simulations were comparable, the MOD3.3 results were more stable for both problems and themass flow rates in Problem 3 did not have the erratic flow reversals predicted by MOD3.2. The MOD3.3results are qualitatively more correct for this problem.

The horizontally stratified countercurrent flow problem (Section 2.1.9) showed qualitatively that thehorizontal stratified flow model in MOD3.3 is functioning properly. Although the wave propagationcalculated by the code was lower than the theoretical value (frictionless environment assumed) it wasqualitatively in the right direction as a result of the interphase drag and virtual mass models.

The results for Pryor’s problem (Section 2.1.10) demonstrated that the water packing mitigationscheme works in MOD3.3 as well as in MOD3.2. There was no significant difference in the results andboth are judged to be qualitatively correct. No significant water packing pressure spikes occurred.

3.2 Separate Effects Problems

For the Edward’s pipe problem (Section 2.2.1) the MOD3.3 version predicted a somewhat more rapidblowdown than MOD3.2, apparently due to the Henry-Fauske model predicting a greater mass dischargerate than the choked flow model of MOD3.2. As mentioned earlier, the pressure recovery due to flashingfollowing the rapid depressurization is delayed compared to MOD3.2 results and data. The cause for thischange is not known at this time. However, this characteristic has little or no effect on safety relatedparameters.

When the nearly implicit numerical scheme was used for modeling the Edward’s pipe blowdown,qualitatively incorrect results were obtained. Apparently the Henry-Fauske model is not coded in thenearly implicit numerical scheme. The nearly implicit option was not tried on other problems such as theMarviken blowdowns. This deficiency needs to be corrected.

The RELAP5 simulation of the Dukler Air/Water problem (Section 2.2.2) reasonably calculated theliquid downflow rate trends of the data and showed that the implementation of the CCFL model in the codeis correct. The magnitude of the calculated flow rate was less than the data. Improvements were observedin the calculation when a value of 0.8915 was used for the CCFL gas intercept constant C (original valuewas 0.88) along with a slope of 0.9364.

The Marviken Tests 24 and 22 simulations (Section 2.2.3 and Section 2.2.4) revealed significantdifferences between the MOD3.2 and MOD3.3 code versions. The most significant difference occurredafter 20 seconds of blowdown at which time the discharge flow became two-phase. The difference is dueto the different choked flow models that are used, Trapp-Ransom in MOD3.2 and Henry-Fauske inMOD3.3. MOD3.2 underpredicted the mass discharge rate while MOD3.3 overpredicted the massdischarge rate. A trial modification to the Henry-Fauske model to permit slip at the break produced resultsin good agreement with the data for Marviken Test 24. This approach should be considered for inclusionin future versions of the code.

The simulation of the LOFT L3-1 accumulator blowdown (Section 2.2.5) compared well with thedata. Modeling improvements between the two code versions did not affect the calculated performance ofthe accumulator model.

NUREG/CR-5535/Rev 1-Vol III 154 Separate Effects Problems

CONCLUSION

The CHF correlation in the code resulted in a reasonable comparison with the data (Bennett: Section2.2.6, RIT Tube Test: Section 2.2.7, and ORNL bundle tests: Section 2.2.8). However, the surfacetemperature above the CHF location was under calculated. It is suggested that too much entrainmentprovided too much cooling capability. Additional investigation into the interphase drag model iswarranted. When the PG-CHF correlation was used CHF was noted to occur at a lower elevation, but thecalculated peak surface temperature and temperature trend matched the data better.

The two subcooled boiling tests (Section 2.2.9 and Section 2.2.10) simulated showed reasonableresults compared to data. The improved subcooled boiling model showed a marked improvement in thecalculated response for the Shoukri low pressure subcooled boiling tests.

Simulation of the MIT pressurizer test (Section 2.2.11) showed the ability to reasonably calculate thetrend of the data. It was noted that numerical mixing of the stratified thermal boundary resulted in a fasterdepressurization after the pressurizer inflow was terminated. Inclusion of the thermal front tracking modelshowed a notable difference in the thermal boundary as the pressurizer filled and better response in thepressure when the liquid inflow was terminated.

The code calculated response agreed well with the data from the FLECHT-SEASET force refloodexperiments (Section 2.2.12). Improvements in earlier versions of the code have strengthened MOD3.3’sability to calculate more accurately the thermal/hydraulic phenomenon associated with both low and highreflood rate behavior. Too much interfacial heat transfer was calculated as evidenced by the low vaportemperatures.

The FLECHT-SEASET Boiloff Test - 35658 (Section 2.2.13) measured to calculated comparisonindicated that the interphase drag calculated by the code was too high and the rate of coolant lost out thetop of the bundle was higher than the data. MOD3.3 showed some improvement over MOD3.2 in thisregard.

3.3 INTEGRAL TEST PROBLEMS

The code calculated response showed good agreement with some of the key parameters of both theL3-7 (small break) and L2-5 (large break) tests (Section 2.3.1 and Section 2.3.2). No significantdifferences exist between the MOD3.3 and the MOD3.2 calculated results for LOFT L3-7. However, itwas observed in the L2-5 test, that the cladding temperatures at the higher elevations in the core weresignificantly different. MOD3.3 results show no heatups at the two highest elevations. This is in contrastto the data in which two consecutive heatups and quench are shown at both locations. For the complexityof these types of transients, the code reasonably calculated the transient behavior. Specific differencesbetween the measured and calculated values noted were related to the way the code handles choked flowand interphase drag. It is clear that further improvements are needed in the reflood model.

RELAP5/MOD3.3 simulated the Semiscale natural circulation tests (Section 2.3.3) reasonably wellfor the higher PCS mass inventories. Also, at the higher steam generator mass inventories, the codecalculations are in good agreement with the measured data. At lower mass inventories it appears theinterphase drag model allowed too much liquid to be entrained thus affecting the results of the calculation.Further work to investigate the interphase drag model is warranted.

INTEGRAL TEST PROBLEMS 155 NUREG/CR-5535/Rev 1-Vol III

CONCLUSION

The code to code comparison for the hypothetical Zion-1 small break transient showed significantdifferences (Section 2.3.4). Although the trends were similar, some of the timing of the major events suchas ECCS initiation and accumulator injection initiation were different. Generally MOD3.2 calculatedthese events to occur earlier in time. One of the significant differences between the two calculations is thedefault break flow model. MOD3.2 used the original RELAP5 critical flow model, whereas MOD3.3 usesthe Henry-Fauske critical flow model. Other differences are improvements in the interfacial drag terms inthe sum and difference momentum equations, formerly user option 47, which have been made default inMOD3.3.

NUREG/CR-5535/Rev 1-Vol III 156 INTEGRAL TEST PROBLEMS

relap5/mod3.3 code manual volume iii: … · relap5/mod3.3 code manual volume iii: developmental...

Documents