You are on page 1of 170

• ~._. _~ . . _._ • • • ~;.,..~ ... ~ ___ ...... ~_ ,.--' _ _ _ _ _ .. "'-'r. _ _ _ ....

~
-
____ _

NUREGCR1864V2
NUREG/CR·1864
BNL·NUREG·51317 1111111111111111111111111111111111111111111111111
VOL. 2 of 3

ASTUDY OF NONEQUILIBRIUM FLASHING OF WATER


IN A CONVERGING-DIVERGING NOZZLE
VOLUME 2 . MODELING

B.J.e. Wu. N. Abuaf. and P. Saha

Date Published: June 1981

EXPERIMENTAL MODELING GROUP


DEPARTMENT OF NUCLEAR ENERGY, BROOKHAVEN NATIONAL LABORATORY
UPTON, NEW YORK 11973

Prepared for the U,S, Nuclear Regulatory Commission


Office of Nuclear Regulatory Research
Washington, D,C. 20555
. '-,
v-·
~~ .', ",'
',', ro"
<'-. I ),:: f.~·:-: f":-. ;"c'

REPRODUCED BY, ~
U.S. Department or Commerce
National Technh:allnfonnalion Service
Springfield, Virginia 22161
NRC FORM 335 1 HI 1'0fn NlJMBE R (AIJ,,,n"rlIJv DOC!
U.S. NUCLEAR REGULATORY COMMISSION
(7·771 NUREG/CR-1864, Vol. 2
BIBLIOGRAPHIC DATA SHEET BNL -NUREG-51317, Vol.2 I
4. TITLE AND SUBTITLE (Add Volume No., If apprO(Jflare' '] IL ,."." hlan.' of J
A STUDY OF NONEQUILIBRIUM FLASHING OF WATER
IN A CONVERGING-DIVERGING NOZZLE 3 RECIPIENT'S ACCESSION NO.
VOLUME 2 - MODELING
7. AUTHOR IS) 5. DATE REPORT COMPLE TED
B.J.C. Wu N. Abuaf P. Saha MON1H I YE AR
June 1981
9. PERFORMING ORGANIZATION NAME AND MAILING ADDRESS (Include Z,p Code) DATE REPORT ISSUED
Department of Nuclear Energy Ma~lch TVEf982
Brookhaven National Laboratory
Upton, New York 11973 6. (Leave blank'

8. (Leave blank'

12. SPONSORING ORGANIZATiON NAME ANO MAl LING ADDRESS (Include Z,p Code)
10. PROJECT/TASK/WORK UNIT NO.
U.S. Nuclear Regulatory Commission
Office of Nuclear Regulatory Research 11. CONTRACT NO.
Washington, D. C. 20555 FIN No. A3045

13. TYPE OF REPORT I PE R I OD COV ERE D (InclUSive dares)

15. SUPPLEMENTARY NOTES 114. (LeBve blank)

16. ABSTRACT (200 words or less)

A steady water loop with well controlled flow and thermodynamic conditions
was designed, built, and made operational for the measurement of net vapor
generation rates under nonequilibrium conditions. The test section consists of
a converging-diverging nozzle with 49 pressure taps and two observation windows
at the exit. Pressure distributions, photographic observations, diamet rical
averaged centerline void fraction distributions, detailed transverse distri-
butions of the chordal averaged void fractions at 27 axial locations, and area
averaged void fraction distributions along the nozzle were recorded under var-
ious flashing conditions. The effects of the various parameters such as inlet
pressure (140 < Pin < 766 kPa), inlet temperature (100 0 < Tin < 149°C), mass
flux (1000 < Gin < 6720 kg/m 2 s), and back pressure on the pressure and void
distributions were investigated and are reported here. Since no information on
the phase velocities was recorded during the present experiments, the calcula-
tion of vapor generation rates from the available experimental data involved the
assumption of a slip model between the two phases.

17. KEY WORDS AND DOCUMENT ANALYSIS 17a. DESCRIPTORS

17b. IDENTIFIERS/OPEN·ENDED TERMS

18. AVAILABILITY STATEMENT 19. SE5UR~TY Cl.../'r~ (Tef'! report) 21. NO. OF PAGES
nc aSSl le
Unlimited 20. SE6UR,TY CLAfS (~/S page) 22. PRICE
nc aSSl le .s
NRC FORM 33517·77)
NUREG/CR·1864
BNL·NUREG·51317
VOl. 2 of 3
AN,R2

ASTUDY OF NONEQUILIBRIUM FLASHING OF WATER


IN ACONVERGING-DIVERGING NOZZLE
VOLUME 1 - EXPERIMENTAL
VOLUME 2 - MODELING
VOLUME 3 - DATA (MICROFICHE)

N. Abuaf, B.J.C. Wu, G.A. Zimmer, and P. Saha

Manuscript Completed: December 19aO


Date Published: June 19a1

Prepared for the


UNITED STATES NUCLEAR REGULATORY COMMISSION
OFFICE OF NUCLEAR REGULATORY RESEARCH
CONTRACT NO. DE·AC02·76CH00016
FIN NO. A-3045

((J-.;
DISCLAIMER

This report was prepared as an account of work sponsored by an agency of the


United States Government. Neither the United States Government nor any agency
thereof. nor any of their employees, nor any of their contractors. subcontractors. or
their employees, makes any warranty, express or implied. or assumes any legal
liability or responsibility for the accuracy. completeness, or usefulness of any
information, apparatus, product. or process disclosed. or represents that its use
would not infringe privately owned rights. Reference herein to any specific com-
mercial product, process. or service by trade name, trademark, manufacturer. or
otherwise, does not necessarily constitute or imply its endorsement. recommenda-
tion, or favoring by the United States Government or any agency, contractor or
subcontractor thereof. The views and opinions of authors expressed.herein do not
necessarily state or reflect those of the United States Government or any agency.
contractor or subcontractor thereof.

Printed in the United States of America


Available from
National Technical Information Service
U.S. Department of Commerce
5285 Pan Royal Road
Springfield, VA 22161

NTIS price codes:


Printed Copy: A08; Microfiche Copy: AOI

JI
FOREWORD

The authors would like to acknowledge Dr. o. C. Jones, Jr. as initiator of


the program and the original principal investigator who conceptualized the re-
search and experiment, supervised the design and construction of the flow loop
and defined the measurement techniques to obtain the necessary data. He also
put the Experimental Modeling and the Systems, Control and Data Acquisition
Groups together to perform the res~arch. The authors would also like to express
their appreciation to Dr. Jones, Jr. for his technical contributions to the
project both in the experiment and the analytical modeling areas and for fruit-
ful discussions that the authors had wi th him during his involvement wi th the
program.

The final series of experimental runs (Runs 254-397), the required data re-
duction and analysis, and the development of the general void growth model were
performed by the present authors. They also assume responsibility for the
organization and content of this final report.

During this research several reports and papers have been published and a
detailed list of all of them is herewith presented. Two specific reports,
BNL-NUREG-26003 and BNL-NUREG-27138 have been transmitted as preliminary data
analysis reports. The present final report summarizes all the data obtained and
the analytical work performed.

- iii -
PUBLICATIONS

The following is a list of publications during the total period of this

activity:

ABUAF, N., JONES, O. C., Jr. and ZIMMER, G. A., "Optical Probe for Local Void
Fraction and Interface Velocity Measurements," BNL-NUREG-5079l, March 1978;
also Rev. Sci Instrum. 49(8), pp. 1090-1094, 1978.

ABUAF, N., JONES, O. C., Jr., ZIMMER, G. A., LEONHARDT, W. J., and SARA, P.,
"BNL Flashing Experiments: Test Facility and Measurement Techniques,"
BNL-NUREG-24336, 1978; also Proc. of the CSMI Spec. Meeting on Transient
Two-Phase Flow, Paris, France, June, 1978 (in press).

ABUAF, N., JONES, o. C., Jr., and ZIMMER, G. A., "Response Characteristics of
Optical Probes," ASME paper 78-WA/HT-3, presented at the Winter Annual
Meeting, San Francisco, Calif., December 1978.

ABUAF, N., ZIMMER, G. A., and JONES, O.C., Jr., "BNL Instrumentation Research
Program," BNL-NUREG-2605l, April 1979; also NUREG/CP/0006, May 1979.

ABUAF, N., FEIERABEND, T. P., ZIMMER, G. A., and JONES, O. C., Jr., "Radio
Frequency (R-F) Probe for Bubble Size and Velocity Measurements,"
NUREG/CR-0769, BNL-NUREG-50997, March 1979; also Rev. Sci. Instrum., 50(10),
pp. 1260-1263, 1979.

ABUAF, N., WV, B. J. C., ZIMMER, G. A., and JONES, O. C., Jr., "Preliminary Data
Analysis Report; Vol. I, I I and III," BNL-NUREG-27138, Jan. 1980.

ABUAF, N., JONES, O. C., Jr., and WU, B. J. C., "Critical Flashing Flows in
Nozzles with Subcooled Inlet Conditions," BNL-NUREG-275l2, March 1980; also
"Polyphase Flow and Transport Technology, R. A. Bajura ed., ASME, Aug. 1980.

ABUAF, N., WV, B. J. C., ZIMMER, G. A., SARA, P., and JONES, O. C., Jr.,
"Nonequilibrium Vapor Generation Rates of Flashing Water Flows," Proc. of the
ANS/ASME International Topical Meeting on Nuclear Reactor Thermal Hydraulics,
Oct. 1980.

JONES, O. C., Jr., and SAHA, P., "Nonequilibrium Aspects of Water Reactor
Safety," BNL-NUREG-23143, July 1977; also in Symp. on the Thermal and
Hydraulics Aspects of Nuclear Reactor Safety, Vol. 1: Light Water Reactors, O.
C. Jones, Jr. and S. G. Bankoff ed. ASME, 1977.

JONES, o. C., Jr., ABUAF, N., ZIMMER, G. A., and FEIERABEND, T. P., "Void
Fluctuation Dynamics and Measurement Techniques," BNL-NUREG-26466, June 1979;
also presented at the Joint U.S.-Japan Inf. Exch. on Two-Phase Flow Dynamics,
Japan, 1979.

- iv -
JONES, o. C., Jr., "Inception and Development of Voids in Flashing Liquids,"
BNL-NUREG-26464, June 1979; also presented at the Joint U.S.-Japan Inf. Exch.
on Two-Phase Flow Dynamics, Japan 1979.

JONES, O. C., Jr., "Flashing Inception in Flowing Liquids," BNL-NUREG-26134,


1979.

SARA, P., "A Review of Two-Phase Steam-Water Critical Flow Models with Emphasis
on Thermal Nonequilibrium," BNL-NUREG-50907, Sept. 1978.

WU, B. J. C., SAHA, P., ABUAF, N., and JONES, o. C., Jr., "A One-Dimensional
Model of Vapor Generation in Steady Flashing Flow," BNL-NUREG-25709, March
1979; also ANS Trans., 32, pp. 490-491, 1979.

ZIMMER, G. A., WU, B. J. C., LEONHARD, W. J., ABUAF, N., and JONES, O. C., Jr.,
"Pressure and Void Distributions in a Converging-Diverging Nozzle with
Nonequilibrium Water Vapor Generation," BNL-NUREG-26003, April 1979.

ZIMMER, G. A., WU, B. J. C., LEONHARD, W. J., ABUAF, N., and JONES, O. C., Jr.,
"Experimental Investigations of Nonequilibrium Flashing of Water in a
Converging-Diverging Nozzle," BNL-NUREG-2S716 , Aug. 1979; also "Nonequilibrium
Interfacial Transport Processes," J. C. Chen and S. G. Bankoff eds., ASME,
1979.

Water Reactor Safety Research Division,


Quarterly Progess Report, Jan.-March 1977 BNL-NUREG-5066l, May 1977 •
" " April-June 1977 BNL-NUREG-S0683, Aug. 1977 .
July-Sept. 1977 BNL-NUREG-S0747, Dec. 1977 .
Oct.-Dec. 1977 BNL-NUREG-5078S, Feb. 1978.
Jan.-March 1978 BNL-NUREG-S0820, May 1978.
April-June 1978 BNL-NUREG-S0883, Aug. 1978.
July-Sept. 1978 BNL-NUREG-S093l, Nov. 1978.
Oct .-Dec. 1978 BNL-NUREG-S0978, Mar. 1979.
Jan.-March 1979 BNL-NUREG-S101S, May 1979.
April-June 1979 BNL-NUREG-S108l, Sept.1979.
July-Sept. 1979 BNL-NUREG-Sl13l, Jan. 1980.
Oct.-Dec. 1979 BNL-NUREG-Sl178, Apr. 1980.
Jan.-March 1980 BNL-NUREG-S12l8, June 1980.
April-June 1980 BNL-NUREG-S122S, Aug. 1980.
July-Sept. 1980 BNL-NUREG-S1298, Nov. 1980

- v -
ABSTRACT

A steady water loop with well controlled flow and thermodynamic conditions
was designed, built, and made operational for the measurement of net vapor
generation rates under nonequilibrium conditions. The test section consists of
a converging-diverging nozzle with 49 pressure taps and two observation windows
at the exit. Pressure distributions, photographic observations, diametrical
averaged centerline void fraction distributions, ~etailed transverse distri-
butions of the chordal averaged void fractions at 27 axial locations, and area
averaged void fraction distributions along the nozzle were recorded under var-
ious flashing conditions. The effects of the various parameters such as inlet
pressure (140 < Pin < 766 kPa), inlet temperature (100° < Tin < 149°C), mass
flux (1000 < Gin < 6720 kg/m 2 s), and back pressure on the pressure and void
distributions were investigated and are reported here. Since no information on
the phase velocities was recorded during the present experiments, the calcula-
tion of vapor generation rates from the available experimental data involved the
assumption of a slip model between the two phases.

The development of voids in nonequilibrium flashing flows is shown through


the Oswatitsch integral to be dependent on three major factors of the void
inception point which determines the initial and subsequent liquid superheats
and must be accurately described; of the interfacial mass transfer rates, which
depend on the local superheat and must be specified; and the local interfacial
area density where the mass transfer occurs. The flashing onset correlation of
Alamgir and Lienhard (1979) was extended to predict flashing inception in pipe
and nozzle flows with subcooled inlet conditions. A void development model for
bubbly flows (a< 0.30) was based on a simple concept for interfacial area de-
nsity in conjunction with a conduction-controlled bubble growth law. A general
model of vapor generation following flashing inception was also developed. In
this model, bubbly flow, bubbly-slug flow, a transitional flow comprising the
annular and annular-mist regimes and finally fully dispersed droplet flow were
assumed to occur at successively higher void fraction ranges.

On the basis that flashing inception occurred at the throat in nozzle flows
with subcooled inlet conditions, and that the pressure undershoot can be
calculated from the Alamgir-Lienhard correlation, a method of calculating the
critical mass flow rates through nozzles was proposed, and it was checked with
existing data.

Comparison of the BNL experiments with TRAC-PIA predictions revealed that,


although the code gave a good qualitative description of the flow, it was
inadequate in predicting the flashing inception point. This failure led to
significant quantitative discrepancies in the predicted and measured flow para-
meters for a number of test runs. The inclusion of a nucleation model should
improve the predictive capabilities of the code.

- vi -
CONTENTS

. VOLUME I - EXPERIMENTAL

FOREWORD •• iii

PUBL IC AT IONS iv

ABSTRACT . • vi

LIST OF FIGURES. xi

LIST OF TABLES . . .. xxiv

NOMENCLATURE • xxvi

1 INTRODUCTION 1

2 REVIEW OF THE LITERATURE • 3

3 EXPERIMENTAL FACILITY AND TECHNIQUES. 7

3.1 Flow Loop •• 7

3.2 Test Section. 9

3.3 Loop Operation Conditions and Instrumentation. 14

4 DATA ACQUIS ITION • • • . • • • • 15

4.1 General Data Acquisition System •• 15

4.2 Static Pressure Measurement Set Up. 15

4.3 y-Densitometer for Void Fraction Heasurements 17

4.3.1 Low Activi ty Single-Beam Densitometer for Axial


Distributions of Centerline Void Fractions. 20

4.3.2 Multibeam y-Densitometer for Transverse and


Axial Void Distributions ••...• 20

4.3.3 High Activity Single-Beam y-Densitometer for


Transverse and Axial Void Distributions •• 22

5 EXPERIMENTAL RESULTS • • • 27

5.1 Pressure Distributions and Photographic Observations. 27

5.1.1 Single-Phase Pressure Calibration . • • • . • 27

- vii -
5.1.2 Pressure Distributions Under Flashing Conditions.. 27

A. Reproducibility Studies •• 32

B. Operational Effects (Effect of Back Pressure) 32

C. Parametr ic Effects. 38

D. Flashing Upstream of the Throat • 45

5.2 Measurement of the Axial Distribution of the Centerline


Diametrical Averaged Void Fraction by a Low Activity Single-
Beam Gamma Densitometer. • • • • • • • •• ••• • • 45

5.2.1 Gamma Densitometer Calibration. 45

5.2.2 Axial Distributions of the Centerline Diametrical


Averaged Void Fractions for Flashing Close to
the Throat. . . . . . . . . . . . . . . . . . . • . 60

5.2.3 Axial Distributions of the Centerline Diametrical


Averaged Void Fractions for Flashing Upstream
from the Throat • • • • • • 66

5.3 Axial Distributions of Area Averaged Void Fractions


Obtained by Means of the Five-Beam Gamma Densitometer •• 66.

5.3.1 Calibrations •• 66

5.3.2 Transverse Void Distributions and Area Averaged


Void Fractions •• 72

5.4 Axial Distributions of Area Averaged Void Fractions


Obtained by Means of the High Activity Single Beam
Gamma Densitometer ... .... 84

5.4.1 Cal ibrations.


····· 84

5.4.2 Transverse Void Distributions and Area Averaged


Void Fract ions. .... ······· 86

5.5 Summary of Exper imental Results: Pressure Distributions


and Area Averaged Void Profiles. 86

5.5.1 149 C Inlet Temperature Runs.


···· 100

5.5.2 121 C Inlet Temperatur e Runs.


····.. 104

5.5.3 100 C Inlet Temperatur e Runs • ..·· 119

5.5.4 Two-Phase Inlet Conditions. .······· 123

5.6 Calculations of Net Vapor generation Rates Under Flashing


Conditions. . . . . . . . . . . . . . . . . .. 123

- viii -
VOLUME II - MODELING

6 ANALYTICAL MODELING AND COMPARISON WITH EXPERIMENTS • 141

6.1 Introduc tion . . . 141

6.2 Flashing Inception. 145

6.2.1 Static Depressurization Results. 145

6.2.2 Flashing Inception in Flowing Systems 147

6.2.3 Pipe Flows . . 151

6.2.4 Nozzle Flows. 153

6.3 Vapor Generation Rate. 162

6.3.1 Bubbly Flow. Low Void Fraction Model (a ~ 0.30).. 162

A. Model. 162

B. Application to Pipe Flows • 170

6.3.2 Gener a1 Model • 177


A. Model. 177

B. Application to BNL Nozzle Data. 196

7 COMPARISON OF TRAC-PIA PREDICTIONS WITH EXPERIMENTAL DATA • 217

7.1 Comparison of TRAC-PIA Predictions with Experimental Data


Consisting of Pressure and Diametrical Averaged Centerline
Vo id Fr ac t ions • • • • • • • • • • 217

7.2 Comparison of TRAC-PIA Predictions with Experimental Data


Consisting of Pressure and Area Averaged Void Profiles 223

7.2.1 Comparison with Runs Per formed at a Nominal


Inlet Temperature of lOODC.
············ 228

7.2.2 Comparison with Runs Performed at a Nominal


Inlet Temperat ureof 121 DC.
············ 228

7.2.3 Comparison with Runs Performed at a Nominal


Inlet Temper a tur e of 149°C.············ 240

- ix -
8 SUHMARY AND CONCLUS IONS • 253

9 ACKNOWLEDGEMENTS. 257

10 REFERENCES. • • • • 258

APPENDICES TO VOLUHE II

Appendix I • • • 264

Appendix II •• 273

VOLUHE III - DATA


(Hicrof iche)

NOTES TO APPENDICES

Appendix A. Single Phase Calibration Data • • • • • • • • • •

Appendix B. Pressure Distribution Data Under Flashing


Conditions • • • • • • •

Appendix C. Pressure and Centerline Diametrical Averaged Void


Fraction Distributions Obtained wi th the Low
Activity Single Beam Densitometer • • • • • • • •

Appendix D. Pressure and Detailed Transverse and Area Averaged


Void Fraction Di stributions Obtained wi th the
l1ulti Beam Gamma Densitometer • • • • • •

Appendix E. Pressure and Detailed Transverse and Area Averaged


Void Fr action Di stributions Obtained wi th the
High Activity Single Beam Gamma Densitometer • • •

- x -
LIST OF FIGURES

3.1 Schematic of BNL Heat Transfer Facility.

3.2 Inside Dimensions of TS-2.

3.3 Deviation From Design of TS-2 Inside Dimensions.

3.4 Schematic Representation of the Test Section.

4.1 Graphical Representation of the Data Acquisition System

4.2 Schematic Representation of y-Densitometer.

4.3 Photograph of the Five Beam Gamma Densitometer.

4.4 Horizontal and Vertical Cross Sections of the Source Holder.

4.5 Schematic Representation of the Detector Holder .

. 5.1 Typical Pressure Distributions Along TS-2 for the Single-Phase Flow
Hydrodynamic Calibration Runs.

5.2 Dimensionless Pressure Distribution for TS-2. Data are Averaged for
All the Hydrodynamic Calibration Runs Performed.

5.3 Typical Representation of an Isothermal Flashing Experiment in the p-T


Diagram.

5.4 Pressure Distributions Under Flashing and Nonflashing Conditions in


TS-2.

5.5 Dimensionless Pressure Distributions in TS-2 Under Flashing Conditions


Compared With Single-Phase Hydrodynamic Calibration Data.

5.6 Comparison of Pressure Distribution in Two Experiments to Show the


Reproducibility of the Results at Low Mass Fluxes, G = 3.03 Mg/m 2 s.

5.7 Comparison of Pressure Distributions in Two Experiments to Show


Reproducibility of the Results at High Mass Flux, G = 4.45 Mg/m 2 s.

5.8 Pressure Distributions Showing the Effect of Condensing Tank Back


Pressure for Identical Nozzle Inlet Conditions.

- xi -
5.9 Photographic Observations for the Experimental Conditions Presented in
Fig. 5.8. In these and all the following photographs, the diameter of
both the front and rear windows is 50 mm.

5.10 Effect of Mass Flux on Pressure Distributions for Identical Nozzle


Inlet Conditions Which are Close to Those For Onset of Flashing in the
Test Section.

5.11 Photographic Observations for the Experimental Conditions Presented


in Fig. 5.10.

5.12 Effect of Mass Flux on the Pressure Distributions in the Test Section.

5.13 Photographic Observations for the Experimental Conditions Presented in


Fig. 5.12.

5.14 Effect of Nozzle Inlet Temperature at Constant Pin-Psat (Tin)


on the Pressure Distribution in the Test Section.

5.15 Photographic Observations for Experimental Conditions Presented in


Fig. 5.14.

5.16 Effect of Nozzle Inlet Temperature at Constant Pin-Psat (Tin)


on the Pressure Distribution in the Nozzle.

5.17 Effect of Nozzle Inlet Temperature at Constant Pin-Psat (Tin)


on the Pressure Distribution in the Nozzle.

5.18 Effect of Nozzle Inlet Temperature at Constant Pin-Psat (Tin)


on the Pressure Distribution in the Nozzle.

5.19 Photographic Observations for Experimental Conditions Presented in


Fig. 5.16 and 5.18.

5.20 Effect of Nozzle Inlet Pressure on the Pressure Distributions in the


Test Section.

5.21 Effect of Nozzle Inlet Pressure on the Pressure Distributions in the


Test Section.

5.22 Effect of Nozzle Inlet Pressure on the Pressure Distributions in the


Test Section.

5.23 Pressure Distributions in the Test Section While The Flashing Onset is
Upstream From the Nozzle Throat.

- xii -
5.24 Nondimensiona1 Pressure Distributions DP* = DP/1/2 Uo in the Test
Section While the Flashing Onset is Upstream From the Nozzle Throat

5.25 Calibration of the test Section Both Empty (Air) and Full of Water as
a Function of Axial Distance.

5.26 Pressure and axial void fraction distributions in the test section.
Plot of the difference between the dimensionless measured pressure
drop and the nondimensiona1 pressure drop measured in the single phase
calibration (DDP = DPi - DP~) as a function of axial distance.

5.27 Pressure and axial void fraction distributions in the test section.
Plot of the difference between the dimensionless measured pressure
drop and the nondimensiona1 pressure drop measured in the single phase
calibration (DDP = DP~ - DP~) as a function of axial distance.

5.28 Pressure and axial void fraction distributions in the test section.
Plot of the difference between the dimensionless measured pressure
drop and the nondimensiona1 pressure drop measured in the single phase
calibration (DDP = DP~ - DP~) as a function of axial distance.

5.29 Pressure and axial void fraction distributions in the test section.
Plot of the difference between the dimensionless measured pressure
drop and the nondimensiona1 pressure drop measured in the single phase
calibration (DDP = DP~ - DP~) as a function of axial distance.

5.30 Pressure and axial void fraction distributions in the test section.
Plot of the difference between the dimensionless measured pressure
drop and the nondimensiona1 pressure drop measured in the single phase
calibration (DDP = DP~ - DP~) as a function of axial distance.

5.31 Pressure and axial void fraction distributions in the test section.
Plot of the difference between the dimensionless measured pressure
drop and the nondimensiona1 pressure drop measured in the single phase
calibration (DDP = DP~ - DP~) as a function of axial distance.

5.32 Pressure and axial void fraction distributions in the test section •
. P1ot of the difference between the dimensionless measured pressure
drop and the nondimensiona1 pressure drop measured in the single phase
calibration (DDP = DPi - DP~) as a function of axial distance.

5.33 Pressure and axial void fraction distributions in the test section
with flashing occurring upstream from the nozzle throat. Plot of the
difference between the dimensionless measured pressure drop and the
nondimensiona1 fressure drop measured in the single phase calibration,
(DDP = DP~ - DP c ) as a function of axial distance.

- xiii -
5.34 Pressure and Axial Void Fraction Distributions in the Test Section
With Flashing Occurring Upstream From The Nozzle Throat. Plot of the
Difference Between the Dimensionless Measured Pressure Drop and the
Nondimensional Pressure Drop Measured in the Single Phase Calibration,
(DDP = DPm*- D~) as a Function of Axial Distance.

5.35 Normalized Calibrations of the Empty Test Section at Five Different


Axial Locations.

5.36 Normalized Calibrations of the Full Test Section at Five Different


Axial Locations.

5.37 Radial Distributions of the Chordal Averaged Void Fractions at Various


Axial Locations Obtained by the Five Beam Gamma Densitometer for
Run 130.

5.38 Radial Distributions of the Chordal Averaged Void Fractions at Various


Axial Locations Obtained by the Five Beam Gamma Densitometer for
Run 130.

5.39 Radial Distributions of the Chordal Averaged Void Fractions at Various


Axial Locations Obtained by the Five Beam Gamma Densitometer for
Run 130.

5.40 Radial Distributions of the Chordal Averaged Void Fractions at Various


Axial Locations Obtained by the Five Beam Gamma Densitometer for
Run 130.

5.41 Radial Distributions of the Chordal Averaged Void Fractions at Various


Axial Locations Obtained by the Five Beam Gamma Densitometer for
Run 130.

5.42 Radial Distributions of the Chordal Averaged Void Fractions at Various


Axial Locations Obtained by the Five Beam Gamma Densitometer for
Run 130.

5.43 Radial Distributions of the Chordal Averaged Void Fractions at Various


Axial Locations Obtained by the Five Beam Gamma Densitometer for
Run 130.

5.44 Axial Distributions of the Area Averaged Void Fractions (A) and of the
Pressure Drop (B) in the Test Section. Runs 129 and 130 were Performed
under the Following Conditions: Pin = 380 kPa~ Tin = 99.9 C,
Pct = 127 kPa at inlet mass flux of 5.97 Mg/m s.

5.45 Normalized Calibrations of the Empty Test Section at Five Different


Axial Locations.

5.46 Normalized Calibrations of the Full Test Section at Five Different


Axial Locations.

- xiv -
5.47 Radial Distributions of the Chordal Averaged Void Fractions at Various
Axial Locations Obtained by the High Activity Single Beam Gamma
Densitometer for Run 310.

5.48 Radial Distributions of the Chordal Averaged Void Fractions at Various


Axial Locations Obtained by the High Activity Single Beam Gamma
Densitometer for Run 310.

5.49 Radial Distributions of the Chordal Averaged Void Fractions at Various


Axial Locations Obtained, by the High Activity Single Beam Gamma
Densitometer for Run 310.

5.50 Radial Distributions of the Chordal Averaged Void Fractions at Various


Axial Locations Obtained by the High Activity Single Beam Gamma
Densitometer for Run 310.

5.51 Radial Distributions of the Chordal Averaged Void Fractions at Various


Axial Locations Obtained by the High Activity Single Beam Gamma
Densitometer for Run 310.

5.52 Radial Distributions of the Chordal Averaged Void Fractions at Various


Axial Locations Obtained by the High Activity Single Beam Gamma
Densitometer for Run 310.

5.53 Axial distributions of the area-averaged and the centerline


diametrical void fractions (A) and of the pressure drop (B) in the
test section. Runs 309 and 310 were performed under the following
conditions: Pin = 555.9 kPa, Tin = 149.1 C, Pct = 378.3 kPa
and at a mass flow rate of 8.8 kg/so

5.54 Axial Distributions of the Pressure (A) and Area Averaged Void Fraction
(B) for Runs Performed at an Inlet Temperature of-149°C and Several
Inlet Mass Fluxes.

5.55 Axial Distributions of Pressure (A) and Area Averaged Void Fraction
(B) for Runs Performed at an Inlet Temperature of 149°C, a Constant
Inlet Mass Flux of 4300 kg/m 2 s, and Decreasing Nozzle Exit or
Condensing Tank, Pressure. '

5.56 Three Dimensional Representation of the Chordal Averaged Void Fraction


Distributions Along the Test Section.

5.57 Three Dimensional Representation of the Chordal Averaged Void Fraction


Distributions Along the Test Section.

5.58 Radial Distributions of the Chordal Averaged Void Fractions at Four


Axial Locations Obtained by the High Activity Single Beam Densitometer
for Run 306.

- xv -
5.59 Radial Distributions of the Chordal Averaged Void Fractions at Four
Axial Locations Obtained by the High Activity Single Beam Densitometer
for Run 310.

5.60 Comparison of Centerline Diametrical Averaged Void Fraction


Distributions with the Area Averaged Void Fractions at the Same Axial
Locations for Runs 306 and 310.

5.61 Axial Distributions of Pressure (A) and Area Averaged Void Fraction (B)
for Five Runs Performed at an Inlet Temperature of 121°C and Increasing
Inlet Mass Flux.

5.62 Comparison of Axial Distributions of Pressure (A) and Area Averaged


Void Fraction (B) for Two Runs Performed Under "Identical" Conditions.
Data in Runs 133-136 Were Recorded with the Five-Beam Gamma
Densitometer with the High Activity Single Beam Densitometer.

5.63 Axial Distributions of Pressure (A) and Area Averaged Void Fraction (B)
for Runs Performed at an Inlet Temperature of 121°C, a Constant Inlet
Mass Flux, and Decreasing Nozzle Exit or Condensing Tank Pressure

5.64 Three Dimensional Representation of the Chordal Averaged Void Fraction


Distributions Along the Test Section.

5.65 Three Dimensional Representation of the Chordal Averaged Void Fraction


Distributions Along the Test Section.

5.66 Three Dimensional Representation of the Chordal Averaged Void Fraction


Distributions Along the Test Section.

5.67 Three Dimensional Representation of the Chordal Averaged Void Fraction


Distributions Along the Test Section.

5.68 Three Dimensional Representation of the Chordal Averaged Void Fraction


Distributions Along the Test Section.

5.69 Axial Distributions of Pressure (A) and Area Averaged Void Fraction (B)
for Runs 149-153, Performed at an Inlet Temperature of 121°C with MgO
Particulates Present in the Water Loop.

5.70 Axial Distributions of Pressure (A) and Area Averaged Void Fraction (B)
for Five Runs Performed at an Inlet Temperature of 100°C and Increasing
Inlet Mass Fluxes (Five Beam Gamma Densitometer).

- xvi -
5.71 Axial Distributions of Pressure (A) and Area Averaged Void Fraction (B)
for Three Runs Performed with the High Activity Single Beam
Densitometer, at an Inlet Temperature of 100°C, and Increasing Inlet
Mass Fluxes.

5.72 Axial Distributions of Pressure (A) and Area Averaged Void Fraction (B)
for Several Runs Performed at an Inlet Temperature of 121°C and Single
and Two-Phase Inlet Conditions.

5.73 Three Dimensional Representation of the Chordal Averaged Void Fraction


Distributions Along the Test Section.

5.74 Three Dimensional Representation of the Chordal Averaged Void Fraction


Distributions Along the Test Section.

5.75 Three Dimensional ~epresentation of the Chordal Averaged Void Fraction


Distributions Along the Test Section.

5.76 Axial Distributions of Pressure (A) and Area Averaged Void Fraction (B)
for Two Runs Performed at Similar Inlet Conditions but Varying
Condensing Tank Pressure.

5.77 Axial Distributions of Pressure (A) and Area Averaged Void Fraction (B)
for Two Runs Performed at an Inlet Temperature of 121°C, Two-Phase
Inlet Conditions, and at Two Inlet Mass Fluxes.

5.78 Top: Measured Pressure (0) and Void Fraction (Q) Distributions in the
Converging Part of the Test Section in Runs 82/821 and the Least Square
Po1ynomina1 Fit to Data. B,ottom: Calculated Net Vapor Generation Rate
Based on the Least Square Fit to the a and p Data.

5.79 Results of Calculation on Data of Runs 80/801. (A) and (B): Pressure
and Void Fraction Distributions;cr, Measured; ---- Cubic Spline Fit;
+, Optimum Knot Locations. (C), (D), and (E): Calculated Value of Rate
of Vapor Generation rv for Three Values of the Distribution Parame.ter
Co·

5.80 Results of Calculations on Data of Runs 141/142. Pin = 240 kPa,


2 .
Tin 121.3 C, Gin 2.97 Mg/m s, Psat (Tin) = 207 kPa,
Pct = 236 kPa, Tct = 121.6 C.

5.81 Results of Calculations on Data of Runs 145/146. Pin 306 kPa,


Tin 121.2 C, Gin 3.70 Mg/m 2 s, Psat (Tin) = 206 kPa,
Pct = 234 kPa, Tct = 121.7 C.

- xvii -
5.82 Results of Calculations on Data of Runs 133/134. Pin = 350 kPa,
Tin 121.2 C, Gin 4.43 Mg/m 2 s, Psat (Tin) = 206 kPa,
Pct = 233 kPa, Tct = 121.7 C.

5.83 Results of Calculations on Data of Runs 140/139. Pin = 465 kPa,


Tin 121.2 C, Gin 5.91 Mg/m 2 s, Psat (Tin) = 206 kPa,
Pct = 234 kPa, Tct = 121.6 C.

6.1 The Flashing Inception Undershoot (Ps - PFio) Predicted by the


Alamgir-Lienhard Correlation.

6.2 Flashing Inception Undershoots and Equivalent Superheats Computed


by the Alamgir-Lienhard Correlation (1979).

6.3 Comparison of the Flashing Inception Data of Reocreux (1974) and of


Seynhaeve, et al (1976) with the Theory Developed by Jones (1979)
Using the Approximate Static Flashing Overexpansion Value of 18 kPa
for the Computation Jones (1979).

6.4 Pressure Distributions for Two Experiments (Run 36, 51) Under 18.3 and
50°C Subcooled Inlet Conditions (Brown, 1961).

6.5 Variation of the Discharge Coefficient with the Reynolds Number for
Data in a Converging-Diverging Nozzle with Subcooled Liquid Inlet
Conditions. Brown (1961), CD = 0.910 + 0.037; Sozzi and Sutherland
(1975), CD=0.918 ~ 0.058; Zimmer, et aT (1979), CD=0.931+0.039.

6.6 Comparison of the Flashing Inception Predicted by Alamgir and


Lienhard (1979) (solid line) with the Locus of the Liquid
Depressurization History (circles connected by dashed line) in Brown's
Nozzle (1961).

6.7 Comparison of the Flashing Inception Predicted by Alamgir and


Lienhard (1979) (solid line) with the Locus of the Depressurization
History in BNL's Nozzle (Runs 76 to 79).

6.8 Comparison of the Flashing Inception Predicted by Alamgir and Lienhard


(1979) (solid line) with the Locus of the Depressurization History in
Reocreux' Pipe Experiments (1974).

6.9 Variation of Surface Area Density With Void Fraction.

6.10 Comparison of Plesset and Zwick (1954) and Forster and Zuber (1954,
1955) Heat Transfer Coefficient with Measured Instantaneous Heat
Transfer Coefficient of Steam-Water During Variable Liquid Super-
heating.

6.11 Comparison of Plesset and Zwick (1954) and Forster and Zuber (1954,
1955) Heat Transfer Coefficient with Measured Instantaneous Heat
Transfer Coefficient of Nitrogen During Variable Liquid Superheating.

- xviii -
6.12 Variation of a with ?r and ZNVG in Present Model and Comparison
with Experimental Data of Reocreux.

6.13 Void Fraction Distribution in a Constant Area Channel as a Function


of Mass Flux G. Tin ~ 116 C. 0, Measurements by Reocreux (1974);
, Calculations Based on the cr
Values Shown.

6.14 Void Fraction Distribution in a Constant Area Channel as a Function of


Mass Flux G. Tin ~ 121 C. 0, Measurements by Reocreux (1974);
----- , Calculations Based on the cr
Values Shown.

6.15 Void Fraction Distribution in a Constant Area Channel as a Function of


Mass Flux G. Tin ~ 126 C. 0, Measurements by Reocreux (1974);
----- , Calculations Based on the cr
Values Shown.

6.16 Effect of Expansion Rate on the Pressure at the Point of Net Vapor
Generation.

6.17 Variation of Point of Net Vapor Generation ZNVG with Mass Flux G and
Initial Temperature Tin in Flashing Flow. Values of ZNVG which
Give "Best Fit" to Reocreux's (1974) Data.

6.18 Variation of Cr with Mass Flux G and Initial Temperature Tin in


Flashing Flow, Values of C Which Give "Best Fit" to Reocreux's (1974)
Data.
r

6.19 The Flow Regimes Map for the General Model.

6.20 Comparison of Correlation of Table 6.2 with Experiment.

6.21 Nomenclature of Bubbly-Slug Flow.

6.22 Surface to Volume Ratios for Cylinders and Spherical Caps.

6.23 Heat and Mass Transfer Rates for Spheres.

6.24 Comparison Between the Void Fraction Data for Run 353 and The "Best-
Fi t" Calculation Using the General Model.

6.25 Comparison Between the Void Fraction Data for Run 358 and the
"Best-Fit" Calculation Using the General Model.

6.26 Comparison Between the Void Fraction Data for Run 362 and the
"Best-Fit" Calculation Using the General Model.

6.27 Comparison Between the Void Fraction Data for Run 145 and the
"Best-Fit" Calculation Using the General Model.

6.28 Comparison Between the Void Fraction Data for Run 137 and the
"Best-Fit" Calculation Using the General Mode.

- xix -
6.29 Comparison Between the Void Fraction Data for Run 137 and the
"Best-Fit" Calculation Using the General Model.

6.30 Comparison Between the Void Fraction Data for Run 344 and the
"Best-Fit" Calculation Using the General Model.

6.31 Comparison Between the Void Fraction Data for Run 291 and the
"Best-Fit" Calculation Using the General Model.

6.32 Comparison Between the Void Fraction Data for Run 284 and the
"Best-Fit" Calculation Using the General Model.

6.33 Comparison Between the Void Fraction Data for Run 273 and the
"Best-Fit" Calculation Using the General Model.

6.34 Comparison Between the Void Fraction Data for Run 278 and the
"Best-Fi t" Calculation Using the General Model.

6.35 Comparison Between the Void Fraction Data for Run 296 and the
"Best-Fit" Calculation Using the General Model.

6.36 Comparison Between the Void Fraction Data for Run 268 and the
"Best-Fit" Calculation Using the General Model.

6.37 Comparison Between the Void Fraction Data for Run 304 and the
"Best-Fit" Calculation Using the General Model.

6.38 Comparison Between the Void Fraction Data for Run 309 and the
"Best-Fit" Calculation Using the General Model.

6.39 The Optimum Bubble Number Density at the Inception Point vs.
the Liquid Superheat at the Inception Point.

6.40 The Interfacial Area Density at the Inception Point vs. the
Liquid Superheat at the Inception Point.

7.1 Comparison of TRAC Predictions with BNL Experimental Results. Run No.
79. Gin = 2270 kg/m 2 s, Pin = 124 kPa, Tin = 99.4°C.

7.2 Comparison of TRAC Predictions with BNL Experimental Results. Run No.
78. Gin = 2610 kg/m 2 s, Pin = 138 kPa, Tin = 99.3°C.

7.3 Comparison of TRAC Predictions and Homogeneous Equilibrium Calculations


with BNL Experimental Results. Run No. 77. Gin = 3060 kg/m 2 s,
Pin = 157 kPa, Tin = 99.4°C.

- xx -
7.4 Comparison of TRAC Predictions and Homogeneous Equilibrium Calculations
with BNLExperimental Results. Run No. 76. Gin = 6040 kg/m 2 s,
Pin = 395 kPa, Tin = 99.3°C.

7.5 Comparison of TRAC Predictions with BNL Experimental 'Results. Run No.
80-801 •. Gin = 4360 kg/m 2s, Pin = 585 kPa, Tin = l48.3°C.

7.6 Comparison of TRAC-PIA Predictions With BNL Experimental Data for


Pressure Distributions and Area Averaged Void Profiles. Pin = 143
kPa, Tin = 99.9 C, Gin = 2.28 Mg/m 2s, Pct = 127 kPa, ~nd
Tct = 100.4 C (mexp =4.6 kg/s, IIlTRAC = 4.8 kg/s).

7.7 Comparison of TRAC-PlA Predictions with BNL Experimental Data for


Pressure Distributions and Area Averaged Void Profiles. Pin = 171
kPa, Tin = 100.2 G, Gin = 3.01 Mg/m 2 s, Pct = 133 kPa, and
Tet = 100.7 C (m exp = 6.1 kg/s, mTRAC = 6.0 ,kg/s).

7.8 Comparison of TRAC-PlA Predictions With BNL Experimental Data for


Pressure Distributions and Area Averaged Void Profiles. Pin = 248
kPa, Tin = 100 C, Gin = 4.49 Mg/m 2s, Pet = 127 kPa, and
Tct = 100.5 C (mexp = 9.1 kg/s, IIlTRAC = 8.6 kg/s).

7.9 Comparison of TRAC-PIA Predictions With BNL Experimental Data for


Pressure Distributions and Area Averaged Void Profiles. Pin = 247
kPa, Tin = 100 C, Gin = 4520 kg/m 2 s, Pet = 113 kPa, and
Tct = 100 C.

7.10 Comparison of Experimentally Measured Pressure Distributions and Area


Averaged Void Profiles, as well as Vapor Generation Rates Calculated
From the Experimental Data With TRAC-PlA Predictions. Pin =241 kPa,
Tin 121.3 C, Gin = 2.97 Mg/m 2 s, Pet = 237 kPa, Tct = 121.7 C.

7.11 Comparison of TRAC-PlA Predictions With BNL Experimental Data for


Pressure Distributions and Area Averaged Void Profiles.' Pin = 305
kPa, Tin = 121.2 C, Gin = 3.7 Mg/m 2 s, Pct = 234 kPa, and
Tct= 121.7 C (m exp = 7.5 kg/s, mTRAC = 7.0 kg/s).

7.12 Comparison of Vapor Generation Rates Calculated From Experimental Data


With TRAC-PIA Predictions for Run 145 Presented in Figure 7.11.

7.13 Comparison of TRAC-PIA Predictions With BNL Experimental Data for


Pressure Distribution and Area Averaged Void Profiles. Pin = 349
Tin 121.2 C, Gin = 4430 kg/m 2 s, Pet = 233 kPa, and
Tct = 121.7 C (mexp = 8.9 kg/s, mTRAC = 8.4 kg/s).

- xxi -
7.14 Comparison of Vapor Generation Rates Calculated From Experimental Data
With TRAC-PIA Predictions for Run 133 Presented in Figure 7.13.

7.15 Comparison of Experimentall Measured Pressure Distributions and Area


Averaged Void Profiles as well as Vapor Generation Rates Calculated
from the Experimental Data with TRAC-PIA Predictions. Pin = 464 kPa,
Tin = 121.6 C, Gin = 5.9 Mg/m 2 s, Pet = 234 kPa, Tct = 121.6 C

7.16 Comparison of TRAC-PIA Predictions With BNL Experimental Data for


Pressure Distributions and Area Averaged Void Profiles for Runs 339-
342. Pin = 320 kPa, Tin 121 C, Gin = 4430 kg/m 2 s, Pet = 240
Tct = 121 C.

7.17 Comparison of TRAC-PIA Predictions With BNL Experimental Data for


Pressure Distributions and Area Averaged Void Profiles for Runs 322-
325. Pin = 341 kPa, Tin 121 C, Gin = 4410 kg/m 2 s, Pet = 200
Tct = 121 C.

7.18 Comparison of TRAC-PIA Predictions With BNL Experimental Data for


Pressure Distributions and Area Averaged Void Profiles for Runs 313-
316. Pin = 341 kPa, Tin = 121 C, Gin = 4430 kg/m 2 s, Pet =
Tct = 119 C.

7.19 Comparison of TRAC-PIA Predictions With BNL Experimental Data for


Pressure Distributions and Area Averaged Void Profiles for Runs 318-
321. Pin 322 kPa, Tin = 121 C, Gin = 4430 kg/m 2 s, Pet = 156
Tct = 113.5 C.

7.20 Comparison of TRAC-PIA Predictions With BNL Experimental Data for


Pressure Distributions and Area Averag~d Void Profiles for Runs 264-
266. Pin = 515 kPa, Tin 149 C, Gin = 2860 kg/m 2 s, Pet = 478
Tct = 149 C.

7.21 Comparison of TRAC-PIA Predictions With BNL Experimetnal Data for


Pressure Distributions and Area Averaged Void Profiles for Runs 291-
295. Pin = 502 kPa, Tin = 148.9 C, Gin = 3170 kg/m 2 s, Pet =
Tct = 148.9 C.

7.22 Comparison of TRAC-PIA Predictions With BNL Experimental Data for


Pressure Distributions and Area Averaged Void Profiles for Runs 284-
288. Pin = 530 kPa, Tin = 149.2 C, Gin = 3580 kg/m 2 s, Pet =
Tct = 149.2 C.

7.23 Comparison of TRAC PIA Predictions With BNL Experimental Data for
Pressure Distributions and Area Averaged Void Profiles for Runs 273-
277. Pin = 573 kPa, Tin = 148.7 C, Gin = 4290 kg/m 2 s, Pet =
Tct = 148.8 C.

7.24 Comparison of TRAC-PIA Predictions With BNL Experimental Data for


Pressure Distributions and Area Averaged Void Profiles for Runs 279-
283. Pin = 688 kPa, Tin = 148.8 C, Gin = 5730 kg/m 2 s, Pet =
Tct = 148.8 C.

- xxii -
7.25 Comparison of TRAC-PIA Predictions With BNL Experimental Data for
Pressure Distributions and Area Averaged Void Profiles for Runs 309-
311. Pin = 556 kPa, Tin = 149.1 C, Gin = 4330 kg/m 2 s, Pct =
Tct = 142 C.

A.l Schematic Representation of a One-Dimensional Nonequilibrium


Vaporization Front.

A.2 Hugoniot Curve of Vaporization Front at Pl 80 kPa in BNL Flashing


Experiment Run 77.

A.3 Comparison of Experimental Mass Flux Gexp and Critical Mass Flux
Gcr at C-J Point in BNL Experiment Run 77.

A.4 Comparison of Experimental Mass Flux Gexp and Critical Mass Flux
Gcr at C-J Point in Run 56 of Schrock, et a1., (1977).

A.5 Comparison of Critical Throat Mass Fluxes Measured by Powell (1961)


with Those Calculated by the Present Method for Different Nozzle Inlet
Pressures and Temperatures.

A.6 Comparison of Calculated with Measured Critical Throat Mass Fluxes


(Powell, 1961) for Various Nozzle Inlet Conditions.

- xxiii -
LIST OF TABLES

3.1 Operational Range of the Facility.

3.2 Test Section Instrumentation.

4.1 Typical Pressure Drop Data.

4.2 Typical Output for the Five-Beam Gamma Densitometer.

4.3 Typical Output for the High Activity Single-Beam Gamma Densitometer.

5.1 Summary of Experimental Conditions. Cold and Hot Single Phase


Pressure Calibrations.

5.2 Summary of Experimental Conditions for Pressure Distribution


Measurements Under Flashing Conditions.

5.3 Summary of Experimental Conditions for Diametrical Averaged Void


Fraction Data with Single Beam Gamma Densitometer.

5.4 Calibration of the Test Section Both Empty (Air), IE' and Full
of Water, IF.

5.5 Distance From Pressure Tap 47 for the 27 Axial Locations (Five Beam
Gamma Densitometer).

5.6 Transverse Radial Locations for Each Y-Beam Center During Axial Scan.

5.7 Summary of Experimental Conditions. Flashing Experiments (Five


Source Gamma Densitometer).

5.8 Distance from Pressure Tap 47 for the 27 Axial Locations (Single Beam
Gamma Densitometer).

5.9 Transverse Radial Locations for y-Beam Center During Axial Scan.

5.10 Summary of Experimental Conditions for Flashing Experiments (Single


Source Gamma Densitometer).

6.1 Conditions in Flashing Flow of Water in a Constant Area Pipe.

6.2 Bubble Rise Velocities.

6.3 Free Rise Velocities of Steam Bubbles in Water.

6.4 Summary of Test Conditions and BNL Model Calculations.

7.1 Comparison of 100 C Inlet Temperature Runs with TRAC-PlA Predictions.

- xxiv -
7.2 Comparison of 121 C Inlet Temperature Runs with TRAC-P1A Predictions.

7.3 Comparison of 149 C Inlet Temperature Runs with TRAC-P1A Predictions.

A.1 Comparison of the Critical Mass Fluxes Calculated by the Jump


Conditions and the Experimental Value Measured by Reocreux
(Run 423)(1974).

- xxv -
NOMENCLATURE

A cross section area

interfacial area density

a thermal diffusivity

discharge coefficient

void distribution parameter

C parameter in Equation (6.15)

constant defined in Equation (6.35)

c static quali ty

specific heat at constant pressure

D diameter

DP pressure differential between the test section inlet (Tap 1) and a


specific Tap location along the nozzle. (This difference does not
include any gravitational head effects.)

DP/l/2 Uo ' dimensionless pressure differential

DPP DPm"'- D~

inverse of cavitation number

f friction coefficient

G mass flux (inlet)

mass flow rate of vapor

expression defined in Table 6.2

g gravitational acceleration

h heat transfer coefficient, or enthalpy

hfg latent heat of vaporization

I number of counts for a specified period of time

number of counts for a specific period of time at a given


location while the test section is empty (full of air)

- xxvi
number of counts for a specific period of time at a given
location while the test section is full of water

number of counts for a specific period of time ata given


location under two-phase conditions

J nucleation rate

j volume tric flux

constant in Equation (6.13)

k thermal conductivity, Boltzmann's constant

L latent heat, or axial length

m(Z,1';) mass of a vapor bubble at Z nucleated at ~

m mass flow rate

N number density

Nu Nusse1t number

p pressure

Pe Peclet number

interfacial heat flux

R radial coordinate, bubble radius

r radius

s surface

T temperature

t time

il' 2 turbulence intensity

test section inlet velocity

u free stream velocity

u velocity

v,V speci fic volume

- xxvii -
vO) bubble free rise velocity

VgQ, relative velocity of the bubble with respect to the liquid

Vgj drift velocity of vapor

v volume

x flow quali ty

Z,z axial coordinate

Greek Symbols

a. void fraction

S defined in Equation (6.15)

fv volumetric rate of vapor mass generation

t:.Ai chordal area subs tended by the gamma beam thickness

!;h perimeter

S dummy variable of integration

f.l Y attentuation coefficient, viscosity

P mass density

a surface tension

L-' depressurization rate

¢ half subs tended angle for spherical caps

w equivalent sphere radius

Subscripts

b bubbles

c single phase calibration

c,cr,crit quantity pertaining to critical size

ct condensing tank (test section discharge)

- xxviii -
d droplets

f saturated liquid

Fio flashing inception under static depressurization

Fi flashing inception

g saturated vapor

i interface

in test section inlet

£, liquid

m measured, or mixture

NVG point of net vapor generation

0 initial point, stagnation condition

R reduced by the critical value

s bubbly-slug flow regime

s or sat saturation

ss steady state

T throat, Taylor bubble

v vapor

Superscripts

* dimensionless

+ throat

. area averaged

Symbol

<> area averaged quantity

- xx ix -
6 ANALYTICAL MODELING AND COMPARISON WITH EXPERIMENTS

6.1 Introduction

In order to predict the critical flow rates of a vapor-liquid two-phase mix-


ture through pipes discharging into a low pressure environment, it is necessary
to understand the characteristics of a flashing two-phase mixture. The trans-
ition from single-phase (liquid) to two-phase flow, due to rapid depressuriza-
tion of the liquid in the flow causing the liquid to vaporize or flash, is par-
ticularly important, because the results of the flashing process may be taken as
the initial conditions for the ensuing two-phase flow.

In addition to those presented in this report, flashing experiments have


been performed by Reocreux (1974) at inlet temperatures of 116° < Tin < 126°C
and pressures of 212 < Pin < 340 kPa. In this section, the nonequilibrium
vapor generation model developed at BNL will be discussed, and also its appli-
cation to the evaluation of experimental data obtained at BNL and elsewhere.

In analogy to the process of nonequilibrium condensation of a supersaturated


vapor during rapid isentropic expansion (see, e.g., Wegener and Wu 1977) and
subcooled boiling (see, e.g., Jones and Saha 1977), the transition from liquid
to vapor-liquid two-phase flow by flashing may be visualized as taking place in
several stages. First, the pressure drop experienced by the initially subcooled
flowing liquid may be sufficient to bring the liquid to a saturated state,
since the temperature change in adiabatic liquid flow is generally negligible.
Vaporization does not occur upon reaching the saturation state, because a finite
liquid-to-vapor temperature difference is required for phase change. Therefore,
the liquid becomes superheated as the pressure decreases further and falls below
the saturation value.

Second, the bubble nucleation, either homogeneously or heterogeneously, be-


gins after an "induction" period when a certain liquid superheat has been at-
tained. The degree of superheat required for nucleation inception to occur in a
given flow system may depend on the flow conditions and the time rate of depres-
surization. Because of the strong dependence of nucleation rate on the ther-
modynamic state of the parent phase, small changes in the liquid superheat may
drastically affect the bubble nucleation rates. Therefore, bubble nucleation
rates will rise quickly to a peak value after the liquid superheat increases
above a "threshold" value. However, the formation of these thermodynamically
stable vapor nuclei and the accompanying loss of latent heat by the liquid to
the vapor is expected to decrease the liquid superheat slightly so as to cause
the nucleation rate to decline to a much lower value. The width of the nucle-
ation zone is a function of the physical properties of the liquid and vapor,

- 141 -
the flow velocity, the depressurization rate, and other flow variables. In the
case of condensation in supersonic flows, it has been found (e.g., Wegener and
Wu 1977), that the nucleation zone is only a few millimeters wide, and the
transport time across this zone is of the order of 10 ~s. It is not clear how
wide the bubble nucleation zone is in flashing flows. However, with the rela-
tively long overall vaporization zone (of the order of 0.1 m) and the relative-
ly low flow speed (of the order of 10 m/s) in observed experiments, nucleation
zones a few cm thick (transport time of a few msecs) may still be considered
narrow. Thus, it is assumed here, for simplicity, that the nucleation zone is a
narrow one.

In addition, Jones and Zuber (1978) found that for the early phases of bub-
ble growth in a variable pressure field, where the pressure varies with time ac-
cording to a power law, e.g., p ~ tn, the bubble radius varies as tn+l/2.
Hence, the rate of bubble volume growth (~t3(n+l/2)) is a strong function of
growth time. For bubbles generated downstream from the nucleation zone, the
growth period is shorter, and the growth rate is always much lower than for
those formed in the nucleation zone. The net vapor generation rate at a loca-
tion downstream from the nucleation zone is dominated by the bubbles generated
in the nucleation zone. Consequently, it is assumed here that, at the end of
the nucleation zone, a certain population of bubbles with an average size deter-
mined by the nucleation mechanism is present, and that additional bubble gener-
ation downstream will be negligible. This is similar to the· assumption used in
earlier research (Edwards, 1968).

Third, vaporization of the liquid, which is still superheated at the sur-


face of these dispersed bubbles, dominates the phase transition. This is the
region of simple bubble growth. Finally, the bubble coagulation becomes ef-
fective at a void fraction of about 30% in substantially altering the bubble
population and size distribution, leading eventually to bubbly-slug or annular-
mist flow.

Unlike the nonequilibrium vapor condensation in supersonic nozzle flow, in


which homogeneous nucleation dominates, flashing as a result of rapid depres-
surization of liquid flows in commercial pipes is most likely initiated by het-
erogeneous nucleation of vapor bubbles in the bulk liquid and/or at crevices or
microcavities along the wall with pre-existing gas phase. Following
Oswatitsch's (1942) treatment of condensation in supersonic nozzles, Zuber et
al. (1966) proposed an expression for the calculation of the mass flow rate of
vapor Gg over a cross section located ~t a point Z along a duct of constant
cross section A:
Z
1
.A f
Z
(6.1)

- 142 -
where ~h is the perimeter of the duct, J( Q is the nucleation rate per unit
wall area at point s along the pipe, m(Z,s) is the mass at Z of a vapor bubble
nucleated at ~, and Zo is a point upstream from the nucleation zone. The
integration effectively sums the vapor mass of all the bubbles nucleated before
point Z. Although Oswatitsch's model has been applied to the study of con-
densation in high speed flows (Wegener and Wu 1977; Wu 1977) with remarkable
success, its extension to f1a~hing flow or subcoo1ed boiling has .been difficult,
mainly because of the lack of understanding of the heterogeneous nucleation
process and of the nucleation rates (J) for flashing flows.

Rohatgi and Reshotko (1975 a,b) carried out such a calculation for flashing
flow of liquid nitrogen and compared their results with the experimental results
of Simoneau (1975). There are two unknown parameters in their heterogeneous
nucleation equations, one bei~g the number of effective nucleation sites per
unit volume, and the other being a function of the contact angle between liquid
nitrogen and the surface of the heterogeneity. They determined these unknown
parameters by "best fits" to the experimental data, but they did not apply the
analysis to any steam-water data.

To circumvent this difficulty, it was decided to treat nucleation and bubble


growth separately. Specifically, an attempt was made to determine empirically
the end of the nucleation zone, or the so-called "net vapor generation" point
ZNVG' while focusing attention on the calculation of vapor generation rate in
the bubble growth region and beyond.

As shown in the following sections, the local vapor generation rate in the
initial bubble growth region can be calculated from conduction limited bubble
growth equations developed by Forster and Zuber (1954) and by P1esset and. Zwick
(1954). The local vapor generation rate depends primarily on three quantities
which are unknown a priori:

a. the onset of flashing or inception pOint, ZNVG' which is usually under


nonequi1ibrium conditions and constitutes the origin of the time
scale t for the bubble growth history;

b. the initial void fraction a o ' at the point of inception; and

c. a quantity C , which is related to the number of bubbles generated at


r
the inception point.

The conduction controlled vapor generation rate proposed was first applied
to the experimental data of Reocreux (1974), and the values of the three para-
meters ZNVG' %' and C r were determined by the "best fit" to the void data.
In reality, they are all related to each other. If the inception superheat is

- 143 -
specified, which determines the onset location ZNVG' the critical bubble
radius at the onset of nucleation is also known. The value of the initial void
fraction, ao ' is then uniquely related to the bubble population at the in-
ception point. It is shown below that the onset of flashing or inception point,
ZNVG' can be fixed from a semi-empirical onset correlation developed by
Alamgir and Lienhard (1979). Their correlation relates the pressure undershoot
below the saturation pressure (at the initial temperature) at the inception
point to the reduced initial temperature and depressurization rate, and can be
used to determine the onset point ZNVG. Once the inception point pressure is
determined, the critical radius at the onset pressure can be calculated, and
following the reasoning of Alamgir and Lienhard, the only free parameter left is
the packing density of bubbles at inception. A value chosen for this free para-
meter will in turn fix ~ and Cr for the specific conditions. With th~ values
of the parameters ZNVG' ao and Cr determined in the above fashion, one can
calculate the vapor generation rates in the nonequilibrium region. This method
was followed for the BNL nozzle experimental results presented in this report.

A second approach to the solution of the same problem involves a study of


the limiting cases. The relationship between conditions across a nonequilibrium
vapor generating zone is the basis of one such method that was tried. The
reasons for seeking such a relationship can be summarized as follows:

The flashing of a superheated liquid under nonequilibrium conditions ap-


peared to be a process analogous to chemical reactions or condensation in high
speed gas flows. Bray (1959) showed that a reacting flow in a nozzle can be
modeled successfully by an equilibrium flow down to a transition point (termed
"sudden freezing point" by Bray) followed by a frozen flow, i.e., no change in
composition. An analogous situation was found in condensing flows in supersonic
nozzles (reviewed by Wegener and Wu 1977). Here the flow was assumed to be
frozen (supersaturated) down to the inception point, where a sudden transition
to equilibrium occurred, and the flow was considered to be in equilibrium and
saturated downstream from the onset point. This approach for condensing flows
is applied successfully except when the .flow is too slow compared with the
characteristic time of condensation.

It was desirable to find out whether an approach similar to the latter could
be applied to flashing flows. The flashing flow with a subcooled inlet can be
considered as a frozen flow (superheated liquid) upstream from the inception
point. The flashing inception generates a given amount of vapor and returns the
flow downstream to the equilibrium conditions. This method of solution, where
the jump conditions are applied across the transition zone at inception, does
not give the details of the transition zone but correlates the upstream and
downstream properties of the flow. When applied to nozzle flows, this method·
requires the additional implicit assumption that the transition region is nar-
row, which was true in reacting and condensing flows in nozzles. The solution
of the jump conditions, which seemed quite simple mathematically, was expected
to give useful information on limiting cases which could then form the basis for
comparison with other methods of prediction. This approach is further discussed
in Appendix I.

- 144 -
6.2 Flashing Inception

The importance of correctly predicting the flashing inception point under


various flow conditions was summarized above. This onset point, which is usual-
ly under varying degrees of nonequilibrium conditions, strongly affects the sub-
sequent vapor generation rate calculations through several factors. The the-
rmodynamic conditions at this inception point determine the local inception
superheat, i.e., the local pressure and thus the critical bubble radius at on-
set. The inception point also constitutes the origin of the time scale for the
bubble growth history, and thus forms the initial point for all the void de-
velopment integration schemes and also fixes the initial superheat or driving
force fo~ the vapor generation and growth. Since most comput~r code calcula-
tions including TRAC-PlA have a bubble growth model, the results predicted de-
pend very strongly on the in~eption point and the initial superheat, in addi-
tion to the heat transfer coefficient and interfacial area densities. This
point is clarified in detail in Section 7, where the present experimental re-
sults are compared with the TRAC-PlA predictions.

In this section, it is shown that the correlation developed by Alamgir and


Lienhard (1979) relating the static pressure undershoot below the saturation
pressure (at inlet temperature) at the inception point to the reduced initial
temperature and local depressurization rate can be applied both to pipe flows
(Jones 1979) and to nozzle flows to determine the flashing inception point.

6.2.1 Static Depressurization Results

Alamgir and Lienhard (1979) developed a correlation to predict- the pressure


undershoot at flashing onset below the saturation pressure during a rapid de-
pressurization of water. Assuming that in practical systems (stainless steel
pipes) heterogeneous nucleation occurs at the surface imperfections of the pipe
walls, they introduced a heterogeneity factor to be determined by experimental
correlation which reduces the critical work of formation of a critical bubble.
An expression for the nucleation rate per unit area and time was presented, and
its integration from the time the local pressure reached the saturation pressure
to the inception time when the depressurization stopped yielded a total density
of nuclei per unit area which was equated to a critical packing density when an
entire surface is covered by bubble nuclei. This derivation provided an expres-
sion for the pressure undershoot as a function of local properties, the hetero-
geneity factor and the depressurization rates. Next, the heterogeneity factor
was assumed to depend on two independent variables, the reduced initial tem-
perature and the depressurization rate, and the functional dependance on each
was evaluated by using published data for rapid depressurization blowdown ex-
periments (Alamgir and Lienhard, 1979). .

- 145 -
The final correlation proposed by Alamgir and Lienhard, 1979 for the pres-
sure undershoot is as follows:

3/2 TR13.76 [1 + 13.25 E,_10~8]O.5


a
Ps - PFio 0.258 (6.2)
AT c
[
1
-:: l
where Ps is the saturation pressure at the initial temperature, PFio is the
pressure at flashing inception, cr is the surface tension, k is Boltzmann's con-
stant, Tc is the critical temperature, TR is the reduced initial tempera-
ture, vf and Vg are the specific volumes of saturated liquid and satura~ed
vapor at the initial temperature in the depressurization process, and i is
the depressurization rate in Mbar/s.

The ranges of the correlation were

0.515 ~ TR ~ 0.935

and

0.004 Mbar/s < t < 1.803 Mbar/s

and the correlation predicted the pressure undershoot at onset with an accuracy
of +10.6%.

Lackme (1979 a, b) on the other hand, while analyzing the critical two-phase
discharge of initially subcooled water from long pipes, proposed a flashing in-
ception (threshold) pressure, PFi' which is affected only by the initial tem-
perature of the liquid,

PFi where k 0.95 (6.3)

while assuming that the mass flow rate (or depressurization rate) plays only a
secondary role. Since Lackme's correlation does not have the expected depen-
dance on depressurization rate and lacks a direct comparison with experimental
data, and since its adequacy was checked with comparisons of critical pressures
and mass flow rates, in the following discussions all the models and consider-
ations are based on the Alamgir-Lienhard correlation, Equation (6.2).

- 146 -
The flashing inception undershoot (ps - PFio) predicted by Equation
(6.2) for water, is plotted in Figure 6.1 as a function of the decompression
rate for several initial temperatures. For a constant, initial temperature the
undershoot is constant for low decompression rates, L < 0.004 Mbar/s, and
then increases monotonically with the decompression rate. The extensions of the
correlation below and above the range of applicability do not
I
correspond to
intuitive expectations. At very low decompression rates ~~O) one would expect
the onset of flashing to occur very close to equilibrium conditions, i.e., close
to the saturation pressure corresponding to the initial temperature (no under-
shoot). At very high or infinite decompression rates, one would expect it to
occur very close to homogeneous nucleation conditions. For low decompression
(L < 0.004 Mbar/s), the experimental data that Alamgir and Lienhard
I
rates
tried to correlate showed an unexplained wide scatter and substantial pressure
fluctuations. On the other hand, the inception undershoots deduced from the
very low depressurization rates encountered in the large vessel during the
Marviken experiments ~ ' - 10- 6 Mbar/s) and evaluated by Abuaf and Wu
(1980), showed a decrease in the inception undershoot with a decrease in the de-
pressurization rates. The undershoot was not constant as extrapolated from the
Alamgir-Lienhard correlation.

In Figure 6.2 the flashing inception undershoot is plotted as a function of


the initial temperature for very low ( i=o Mbar/s) and high ( i= 1.0 Mbar/s)
depressurization rates. As seen, significant inception superheats are predicted
at high decompression rates (-50°C). Even at very low rates, the superheats may
increase to -10°C at high initial temperature, 300°C. As expected, the under-
shoot decreases as the initial temperature gets closer to the critical temper-
ature of water (374.2°C).

The correlation is used in this report even though it does not include the
effects of the differences in water preparation and chemistry, and the present
authors believe that the effects of dissolved gases and particulates as well as
surface roughness should be thoroughly investigated before the results can be
applied universally.

6.2.2 Flashing Inception in Flowing Systems

Seynhaeve, et al. (1976) did critical flow experiments in a long pipe with
inlet temperatures between 111°C and 167°C and mass fluxes between 10 and 20
Mg/m2 s. They found the superheat at flashing inception to vary inversely
with mass flux. Although their data were quite scattered, the superheat ap-
parently decreased to almost zero at the higher mass fluxes and even became
negative in a few cases. In evaluating Reocreux's 1.74 bar data, the present
authors found similar results, presented in more detail in Section 6.3.1 (Wu et
al., 1979). Ardron and Ackerman (1978) conducted critical flow experiments with
subcooled water in a straight pipe, and they also reported liquid superheats at
the inception point (point where the pressure variation departs from a straight
line) which decreased with increasing mass flow rates. The superheat they
measured was smaller by about a factor of four than that measured by Seynhaeve
et al. (1976).

- 147 -
g~--~~rT"nT---.-.-rIlTnn---'-'-TlI\Tn--~-'~IIITn---,--r'-rn~
....;
".

LI ENHARD CORRELAT ION ""


:::: FOR FLASH IN G INCEPTION
0
HATER <T c = 647.3 K) ;
0:
CC
CQ ".
,.
,.
b ,.
a ,.
::::t:
V)
0
....; , 3000 C
~ ,.
UJ
Q
,. 2500 C
z:
=:J :: 2250 C
-
2!
0
I-
c....
UJ
u
::
......
".
2000 C
z: 1750 C
~
2!
......
...
=
....;
... 1500 C
V) ......
:5
LL..
...... 1250 C
...
--
...
- 1000 C
....;
o~~~~~~~--~~~~~~--L-~~~~
0.0001 0.001
- - 0.01
RANGE OF APPLICABILITY - - - -
__-L~-L~~U-__~-L~~~
0.1 1

DECOMPRESSION RATE - MBAR/SEC


Figure 6.1 The Flashing Inception Undershoot (p s - PFio ) Predicted by the
Alamgir-Lienhard Correlation. (BNL Neg. No. 9-823-80)

- 148 -
~~----------------~----~~~------,

t-
o
- - --
==
(/)
0:::
, u.J
~~ ;; 8 BAR/SEC
o
:::::
::::l
z:
o
I-
a... I
u.J
(.,..)
Z I
I
I
(/) I
:5 - PRESSURE-BAR I
w...
- - - TEMPERATURE-OC
.
OL-____-L________~~--------------~
o 200 400
INITIAL TEMPERATURE - °c
~~----------------T---------------~
.-i

t-
o
o
::I:
(/)
0:::
u.J
o
z
:=> ~~ = 1. 0 ~1BAR/ SEC
z
o
o
I- .-i
a...
u.J
(.,..)
z:
c..!j
z:
::I:
(/)
,,
<C
....J
u..
- PRESSURE-BAR
- - - TEMPERATURE-oC
.-i~ ______________ ~~ ____________
,
1
~~

o 200 400
INITIAL TEMPERATURE - °c
Figure 6.2 Flashing Incep~1on Undershoots and Equivalent Superheats
Computed by the Alamgir-Lienhard Correlation (]979).
(BNL Neg. No. 9-822-80)

- 149 -
Concerning two-phase critical flow in nozzles, short tubes, and ori-
fices, Zaloudek (1963) and Henry and Fauske (1971) both observed that with
saturated or subcooled inlet conditions no vapor is formed until the throat of
the nozzle is reached. However, neither provided an explanation for this ob-
servation or developed criteria for predicting the amount of superheat
(metastability) that can be allowed in such nozzles or its dependence on inlet
pressure, decompression rate, or nozzle geometry. Apparently no other ex-
periment has been heretofore undertaken allowing suitable definition for de-
termination of flashing inception superheats.

The boiling inception and onset of net vapor generation in flowing liquids
in the case of heating has been the subject of many studies, including such
well-known works as those of Hsu (1962), and of Saha and Zuber (1974) among
others. Unfortunately, flashing inception does not appear to be characterized
by models applicable to heated liquids, where the superheat is generally con-
fined to the wall layer in bulk subcooled liquids. Instead, bulk superheating
occurs prior to flashing inception while the initial voiding still seems gener-
ally relegated to the wall layer.

The static systems of Edwards and O'Brien (1970) and of Lienhard et al.
(1978) exhibit decompression rates of 0.05 to 1.5 Mbar/s,and the decompression
times are generally less than a millisecond. The flowing system of Reocreux
(1974) and of Seynhaeve et a1. (1976) decompress at rates three orders of mag-
nitude slower, and the decompression times range up to several tens or hundreds
of milliseconds. The only other difference between the static and dynamic
flashing systems seems to be the fluid motion. Of the various factors
influenced by the fluid motion, the turbulent pressure fluctuations were pro-
posed by Jones (1979) to be those most likely to have an effect. Chen (1969)
offered a similar suggestion to explain sodium boiling superheat behavior.

It thus seems that decompressive flashing inception might be characterized


by at least three conditions: initial temperature; decompression rate; and,
under some conditions, degree of liquid turbulence. In order to extend the
, experiments to
Alamgir-Lienhard correlation determined from static decompression
flowing systems, it was proposed that the decompression rate E , should be set
in an Eulerian frame experienced by a fluid particle flowing in a pipe or noz-
zle, and should be evaluated from the following expression:

E' =.iE.= ~ + u ~ (6.4)


dt 3t 3z

where u is the average mixture velocity.

- 150 -
The first term on the right represents the static depressurization rate de-
termined by Alamgir and Lienhard (1979), and the second term represents the ad-
ditional convective depressurization rate experienced by the fluid particle in a
flowing system. In the case of steady flow in a nozzle, only the latter term
exists, but for transient blowdown of a reservoir through a nozzle both terms
may in fact be important. Based on these considerations, attempts were made to
characterize flashing inception in flowing systems to the extent possible in
view of the limited data available, as decribed in the next two sections.

6.2.3 Pipe Flows

Jones (1979) examined the flashing inception superheat for steady flow of
liquids in pipes. By hypothesizing that the inception could be separated into
two parts, one due to decompression and the other to turbulence, he was able to
correlate the latter effects with the existing data for straight pipes by the
dimensionless expression

'PF/= Max [0I - 27


(~~ )
- F.
(6.5)
U2 ~
where ~PFi* is the pressure undershoot at flashing inception in pipe
flows, ~PFi = Psat - PFi' made dimensionless with the static value,
~PFio = Psat-Pfio' which would exist without the effects of turbulence
and convective depressurization and which could be calculated from the Alamgir-
Lienhard correlation, Equation (6.2). The flashing index Fi , given by
Fi = ~U2/2~ PFio' is simply the inverse of the well-known cavitation
number but written in terms of the static decompression pressure undershoot
~PFio' with which the dynamic undershoot is made dimensionless. Note that,
since the interest is in flashing inception at a given location, it is the local
free stream velocity, U, which is of interest and was assumed fullv developed
(Jones, 1979). A value of 0.072 for the turbulent intensity, ~/U , was
found to represent the existing inception data adequately, in agreement with the
single-phase data of Laufer (1950).

The flashing inception data of Reocreux. (1974) and Seynhaeve et al. (1976)
are plotted in Figure 6.3 as a function of the square of the mass flow rate
(Jones 1979). Also shown is the prediction based on a static inception under-
pressure of ~PFio = 18 kPa. The trends observed appear to support the con-
clusions previously stated (Jones, 1979). Note that the lower limit of zero
superheat also appears reasonable and is tentatively supported by the rela-
tively meager amount of data available. For straight pipe flows, the static
flashing inception undershoot would be obtained from Equation (6.2) and would
then be decreased by the turbulence effects given by Equation 6.5.

- 151 -
24

-;;
0..

~ 20 t; o REOCREU)( T1 " 116"C

0: 0 o --.-121"C
'"
<l c:. --,,-126"C
I • SEYNHAEVE TI • III"e
16
z • --,,-123"C
0
!i: 0 0 • - --134"C
'"~
U 0 • - --139"C
12
"zZ
en
<I
...
...J

!;:
'"cr

.
:::l
en
f3 4
cr
0..
cr
'"oZ
:::l o~-------------~--
-


SQUARE OF THE MASS VELOCITY- kg" 1m'. s'

Figure 6.3 Comparison of the Flashing Inception Data of Reocreux (1974)


and of Seynhaeve et a1. (1976) with the Theory Developed by
Jones (1979) Using the Approximate Static Flashing Over-
expansion Value of 18 kPa for the Computation, Jones (1979).
(ENL Neg. No. 3-238-79)

- 152 -
6.2.4 Nozzle Flows

One observation from the BNL flashing experiments reported in Section 5 was
that, in a wide range of experiments performed with subcooled water inlet con-
ditions, the flashing inception occurred very close to the nozzle throat. If
this observation is generally true, the implication is that all the published
experimental data concerning the flow of liquids in converging-diverging nozzles
can be correlated by a single-phase discharge coefficient. The discharge coef-
ficient for an incompressible liquid can be related to the mass flux through the
throat (G T ), the density (p), and the pressure difference between the inlet
and throat (po -PT) where the inlet area is assumed to be large so that the
inlet veloCity is negligibly small:

~=

Published experimental data for converging-diverging nozzle flows with various


subcooled inlet conditions and with flashing in the nozzle were considered, the
data for each experiment including the mass flux, inlet and throat pressures,
and inlet conditions for density determination. The data analyzed included
those of Brown (1961), Schrock et al. (1977),Sozzi and Sutherland (1975), and
the BNL data for water, and, the data of Simoneau (1979) for liquid nitrogen.
Figure 6.4 shows the pressure distributions in the converging-diverging nozzle
as recorded by Brown (1961) for two runs. Run 36 had an inlet pressure of 8.5
x 10 3 kPa and an inlet temperature of 270°C, corresponding to a highly sub-
cooled inlet condition, since the saturation pressure at the inlet temperature
was 5.5 x 10 3 kPa (50°C subcooling (Schrock et al. 1977». Run 51 had an in-
let pressure of 7.14 x 10 3 kPa and an inlet temperature of 286°C, conditions
very close to the saturation line,. since the saturation pressure at the inlet
temperature was 7.03 x 10 3 kPa. Schrock et al. (1977) assumed this run to
have subcooled inlet conditions (18.3°e subcooling), but their own estimated
pressure measurement accuracy of 5 % implies that this run could have had a very
low inlet void fraction. In the calculations below the original investigator's
assertion that this experiment had a subcooled inlet condition is accepted.
These pressure distributions in Figure 6.4 show the large range of subcooling in
Brown's experiments, from lS.3°e to 50°C. The discharge coefficients calculated
for Brown's data (1961), Sozzi and Sutherland's data (1975), and BNL data (Zim-
mer et al. 1979), and for two sets of Simoneau's (1979) liquid nitrogen data
(for his conical nozzle with inlet temperatures of 95°K and 1100K) are plotted
in Figure 6.5 as a function of the throat Reynolds number. The discharge coef-
ficients calculated for each nozzle are scattered around a value of 0.92 ± 0.04,
a value quite close to the expected single-phase discharge coefficient for
converging-diverging nozzles. The slightly larger standard deviation in Sozzi

- 153 -
--------- ----------
8
8ROWN(NOZZLE NO.:3)

o RUN:36 PO=8.48 10 3 KPa,T =270C


O
G = 92.2 103kg/m29
r
RUN 51
o RUN 51 PO· 7.14103 kPa, TO=286C

Gr = 70. 10 3 kg/m 2 9
"'2
..... 5
a
~
ILl
It:
:::J
en
en
ILl
It:
a..

o ~---~~--~------~-----~------~------~------~------~-----~
----0
o 5 9
AXIAL DISTANCE (em)

Figure 6.4 Pressure Distributions for Two Experiments (Run 36, 51) Under 18.3 and
50 C Subcoo1ed Inlet Conditions (Brown, 1961).
(BNL Neg. No. 11-669-79)

- 154 -
and Sutherland's data (Nozzle 1) could be due to the possible existence of inlet
voids for some of the experiments reported here. Their stagnation qualities
were deduced from fluid density in the vessel, p, obtained through hydrostatics,
and the stagnation pressure, Po:
1
p
(6.7)
x
o Vfg(p o )

All the data presented by Sozzi and Sutherland (1975) for Nozzle 1, with a nega-
tive inlet quality from Equation (6.7) implying subcooled inlet conditions, were
included in the results presented in Figure 6.5. This definition (Sozzi and
Sutherland 1975) of subcooling strongly depends on the method of density
measurement in the vessel, since some stratification in the vessel was reported.
Thus, some of the subcooled runs that had a very small negative quality, ~ -2 x
10- 4 , could have had a very low inlet void fraction. In Figure 6.5, the orig-
inal investigator's statement was again accepted and all the negative inlet
quality data were assumed to be subcooled. The small variation of CD over the
wide range of subcooled venturi-nozzle data, implies that the flow up to the
nozzle throat is predominantly single-phase liquid. In addition, it can be con-
cluded that, even if flashing occurred near the throat pressure tap but up-
stream, the density change caused by the flashing inception did not have a
significant effect on the pressure measured at the throat.

The observation mentioned above, that under subcooled inlet conditions,


flashing inception seems to occur very close to the throat, was also sub-
stantiated from a different point of view involving the application of Alamgir
and Lienhard's correlation of the pressure undershoot during rapid static de-
pressurization and flashing of water, summarized in Section 6.2.1. The pressure
undershoot at flashing onset as a function of the depressurization rate
predicted by Alamgir and Lienhard's correlation is presented in Figure 6.6,
along with a plot of the locus of the liquid depressurization history in Brown's
Run 36. The ordinate of each dot was determined by the local pressure measure-
ment, which, when subtracted from the saturation pressure at the inlet tem-
perature, provides the local degree of overpressure or underpressure. The
abscissa, which is a measure of the local depressurization rate, was calculated
from the expression

3
~ = dp dZ
-=
rh dA (6.8)
dt dZ dt 2'"4 . dZ
p A

- 155 -
1.0
-~
0-
I
0-
o ••
Q..
• o 00
N
...... 0.8 o BROWN
C\J~
CD • ZIMMER SINGLE-PHASE CALIBRATION
o ZI MMER FLASH I NG EXP.
110
U • SIMONEAU (CONV. NOZZLE-REVERSE)
o SI MONEAU (CONV. NOZZLE)
-(:( SOZZI 8 SUTHERLAND (NOZZLE No.1)
0.6

Re T =
Figure 6.5 Variation of the Discharge Coefficient with the Reynolds Number for
Data in a Converging-Diverging Nozzle with Subcooled Liquid Inlet
Conditions. Brown (1961), CD = 0.910 + 0.037; Sozzi and Sutherland
(1975), CD = 0.918 i 0.058; Zimmer et.-a1. (1979), CD = 0.931 + 0.039.
(BNL Neg. No. 3-12-80)

- 156 -
where m is the mass flow rate, A the local cross-sectional area, and dA/dZ the
variation of A with axial distance Z, and the effect of wall friction is
neglected.

In Figure 6.6a, the region below the horizontal axis (ps - p ~ 0) re-
presents a subcooled liquid condition; the region between the solid curve pre-
dicted by Alamgir and Lienhard's correlation and the horizontal axis represents
a superheated liquid condition; and the solid curve is the onset of flashing.
Along the lines of the experimental points for Run 36 (Brown, 1961), at pres-
sure Taps 3, 4, and 5 the fluid is seen to be subcooled, and between Taps 5 and
6 (throat), the conditions cross the inception line. A similar plot in Figure
6.6b for Run 51 (Brown, 1961) with To = 286°C, shows that at Taps 1 and 2 the
liquid is subcooled but close to saturation and at Taps 3, 4, and 5, it becomes
superheated, but once again the flashing inception line is crossed between Taps
5 and 6 (throat).

The plot of the locus of a fluid particle moving from inlet to throat in the
converging section of the nozzle shows that, as it gets into smaller cross-
sectional areas, it encounters increasing local dp/dt and enters a region that
can sustain a higher pressure undershoot (higher superheat). At or near the
throat, dA/dZ from geometry and dp/dZ for single-phase flow go from negative to
positive. At the point where dp/dZ > 0 or dA/dZ > 0, the liquid cannot sustain
any superheat and relaxes with flashing inception~ The dp/dt at the throat for
Figure 6.6a and b was calculated from Equation (6.8) with a dA/dZ value
calculated by a second-degree polynomial considering the upstream tap locations
5 and 4.

A similar representation of the pressure undershoot as predicted by Alamgir


and Lienhard's correlation and the locus of a fluid element in the converging
part of the nozzle appears in Figure 6.7 for BNL runs 76 to 79, conducted at an
inlet temperature of 100°C and at various inlet mass fluxes, as reported
previously, with depressurization rates of 10 6 to 10 8 Pals, which are much
lower than those used by ~rown (10 8 to 1011 Pals) and also lower than the
minimum recommended by Alamgir and Lienhard (1979) of (4 x 10 8 Pals). At
these low depressurization rates < 10 8 Pals, the Alamgir-Lienhard correlation
predicts a constant undershoot which depends on the temperature but not on the
depressurization rate. The published experimental data existing in the litera-
ture that Alamgir and Lienhard (1979) tried to correlate in these low ranges had
an unexplained wide scatter. Although their correlation was assumed to apply at
low depressurization rates, some additional static depressurization experiments
are needed to verify its validity in this range. The important observation from
Figure 6.7 is that flashing inception is again predicted to occur within 1 cm of
the throat, between Taps 24 and 25 (throat). These four cases are typical of
all the observations in these series of tests.

- 157 -
o
40 a) Lun 36, T. =270 C TAP 6
l.
I
30 FLASHING I
• BROWN (NOZZLE 3) INCEPTION I
RUN 36 I
II> 20 I
0
I
+ I
0 10 I
Il.
SUPERHEATED LIQUID I
Il. I
I 0
Il.'"
SUBCOOLED LIQUID
-10

-20

10 8 10 9 1012

b) Run 51, T. = 286°C


50 l.
o BROWN (NOZZLE 3) TAP 6
RUN 51 --o~
Q 40 \
..:t \
c I
Q.
Q. 30
I \
(f)

Q. 20 \ r"LASHING
\ INCEPTION
I
10~_---
),TAP 5
SUPERHEATE~S!El!.!.'2-!~1..----
o TAP I TAP 2 TAP 3 SUBCOOLED LIQUID
10
10
dp (PC)
dt s

Figure 6.6 Comparison of the Flashing Inception Predicted by Alamgir and


Lienhard (1979) (solid line) with the Locus of the Liquid
Depressurization History (circles connected by dashed line) in
Brown's Nozzle(1961) (BNL Ne~.!~o. 3-13-80" 11-662-79)

- 158 -
I I
I I
40 f- o BNL RUN 79 -
-425 (THROAT)
o BNL RUN 78
6, BNL RUN
(THROAT) \ \ - 0 2 5 (THROAT)
77 --025
30 f- o BNL RUN 76
TAP 25
\ \ \ -
\ \ \
fALAMGIR
AND
(THROA~" \ \
\ \
" ""
r<l
0 . LIENHARD
-01- 20 f- \, \ \ -

a.
0

TAP
""24J \ ~24 \
)24
\
\
i
a. FLASHINiEPTION
I \
a.
Ul 10 f-
SUPERHEATED LIQUID TAP 2~ I / \
-
I / \
TAP 2~ 023 /

0
TAP 21/ // -9'23 \

SUBCOOLED LIQUID
TAP 200'
.,)J 22d
/ \
22/.
\ l-.i-.l

~(~),
dt s

Figure 6.7 Comparison of the Flashing Inception Predicted by A1amgir and


Lienhard (1979) (solid line) with the Locus of the Depressurization
History in BNL's Nozzle (Runs 76 to 79).
(BNL Neg. No. 11-663-79)

- 159 -
Thus it has been demonstrated from two approaches, the discharge coefficient
and the static depressurization onset, that, in all the experimental data an-
alyzed for converging-diverging nozzles with various degrees of subcooling at
the inlet, flashing inception occurs very close to the throat, and the mass flux
under these conditions can be correlated with the single-phase discharge coef-
ficient concept.

Similar reasoning can be applied to the friction dominated pipe flow ex-
periments (Reocreux, 1974). In this case, the local depressurization rate is
constant for a given mass flux and can be expressed as

(6.9)

where G is the mass flux, f the friction coefficient, the density, and D the
pipe diameter. As seen in Figure 6.8, the depressurization rates of the
Grenoble experiments, 10 4 to 10 6 Pals, are lower than those of the BNL
experiments and are below the range of applicability of Alamgir and Lienhard's
correlation. In Figure 6.8 the pressure undershoots calculated from the
correlation are again independent of the depressurization rate, but they depend
strongly on the inlet temperature. The open circles are the onset points as
determined (Wu et al., 1979) from the void fraction measurements of the Reocreux
data (1974) for experiments done at 116° and 125°C. The vertical dashed lines
present the locus of a fluid element in a pipe flow, from subcooled downstream
conditions to the onset of flashing. The experimentally determined onset points
lie below the onset predictions of Alamgir and Lienhard (1979), and they get
closer to saturation at high mass fluxes or depressurization rates. This
observed behavior was related to the turbulent pressure fluctuations by Jones
(1979) (See Section 6.2.2).

It is well known that the frictional effects in the converging regions of a


nozzle are generally negligible and that the boundary layer thickness decreases
as a result of the convergence. Indeed, the BNL experiments presented in Sec-
tion 5.1.1 showed accelerational pressure drop to be by far the dominant factor
in the converging region of the nozzle, and suggested that the turbulence ef-
fects would be negligible in sharply converging nozzles. Thus, the static incep-
tion criteria of Alamgir and Lienhard (1979) were directly applied to nozzle
flow.

- 160 -
I I I
I

'~25C)FLA5HING
40 I- -

ALAmGIR AND LIENHARD INCEPTION

30 -
, -
ALAmGIR AND LIENHARD(Tin= 116 C) FLASHING INCEPTION

o
Q. 20 I- 0 -
=.c: 0
Q.
I
III
Q.
~
I ) 0

10 -
SUPERHEATED
LIQUID
I
I
I
I
I
I
I
9
I
I
I
r
I
-

I I
I
o I I ! I
I I
I
SUBCOOLED
LIQUID

I I I I I

10 5

.QE.
dt
(£2.)
s
Figure 6.8 Comparison of the Flashing Inception Predicted by Alamgir and Lienhard
(1979) (solid line) with the Locus of the Depressurization History in
Reocreux'. Pipe Experiments (1974) (BNL Neg. No. 11-666-79)

- 161 -
Based on the above two observations, (a) that flashing inception in nozzles
occurs at the throat, and (b) that the underpressure at flashing inception can
be calculated from the Alamgir-Lienhard correlation, a method was proposed for
calculating the critical mass flow rates in nozzle flows with single-phase sub-
cooled or saturated inlet conditions. This approach is further detailed in Ap-
pendix II.

6.3 Vapor Generation Rate

6.3.1 Bubbly Flow, Low Void Fraction Model (a ~ 0.30)


A. Model

After the flashing inception point further vapor generation in a sup~r­


heated, pure liquid is limited mainly by the heat transfer rate to the liquid-
vapor interface and the available interfacial area. In this situation, the lo-
cal equivalent of the Oswatitsch integral shows that the mass of vapor generated
in the liquid per unit volume of the vapor-liquid mixture per unit time, r v '
is given by

A q'!
s ~
L (6.10)

where As is the total area of the vapor-liquid interface per unit volume
of the mixture, qi" is the heat transferred to a unit area of the interface
per unit time, and L is the latent heat of vaporization. Equation (6.10)
expresses the balance, across the interface, of the rate of energy loss due to
vaporization and the rate of energy replenishment by heat transfer from the
bulk. It is obvious from Equation (6.10) that the functional relationships of
both As and qi" to the thermodynamic state and transport parameters of the
two-phase system must be known in order to determine rv. Since, neither
function is known at present, some additional assumptions have to be made.

Interfacial Area Density, As:

After being nucleated at the microcavities along the pipe wall, the bubbles
would leave the surface and move into the core flow when they grow to a certain
size which depends on the flow conditions. Thus, at low void fractions, a bub~
bly two-phase mixture can be expected. As the bubbles continue to grow because
of vaporization of liquid at the interfaces, their number remains unchanged once
nucleation has ceased but before bubble coalescence becomes significant. Since,
in this bubbly flow regime, bubble size differences arise only as a result of
different growth histories, the dispersion of bubble size distribution is

- 162 -
directly related to the length of the nucleatipn zone: The narrower the nucle-
ation zone, the more uniform the bubble sizes. Thus, it is assumed that, in the
bubble growth region, equal-sized spherical* bubbles are to be found at any
cross section. Consequently,

(6.11)

where As,b is the interfacial area density for bubbly flows, Nb is the
number of bubbles per unit volume, a is the void fraction of the mixture, and
rb is the bubble radius.

As vapor generation continues, some of the bubbles may coalesce -- because


of geometrical limitation or differential motion of the bubbles -- to form
larger bubbles which may occupy almost the entire pipe diameter, i.e., Taylor
bubbles, while the volume between the Taylor bubbles is filled with a liquid
slug with smaller bubbles. Thus, at higher void fractions, a bubbly-slug flow
may be expected. The frequency of occurrence and the length of the Taylor bub-
bles may increase with increasing void fraction. The bubbly-slug flow regime
terminates with the transition to annular flow when the Taylor bubbles actually
touch one another to form a continuous void in the core. In annular flow, some
liquid in the form of finely dispersed droplets may be entrained in the core re-
sulting in "annular-mist" flow. Finally, as the void fraction approaches unity,
the flow is expected to change into the dispersed-droplet regime. Assuming
equal-sized droplets, one may write the surface area density for dispersed drop-
let flow in terms of droplet concentration Nd and void fraction a as

(6.12)

During the development of this model, it was not clear how the interfacial
area density could be computed for the bubbly-slug and the annular-mist flow re-
gimes. However, it was noted that, in the bubbly-slug flow regime, the surface
area density may be influenced by two opposing effects: (a) The formation of
large bubbles tends to reduce the surface-to-volume ratio, whereas, (b) de-
viations from sphericity of the bubbles with incr~asing size leading to ir-
regular bubbles tends to increase As' Therefore, th~ variation of As with a

* For the small bubbles (r < 0.8 mm) expected here for a ~ 0.3 and Nb ~ i0 9 ,
m- 3 , surface tension effects dominate.

- 163 -
may be small here, and it is assumed in this work that As remains constant in
the bubbly-slug regime, having the value of As at the transition from bubbly
/
flow, which is assumed to occur at a = 0.3 (Dukler and Taitel 1977). As is
assumed to remain constant until a reaches 0.7, when transition to dispersed
droplet flow occurs and the droplet concentration is arbitrarily taken to be
Nd = Nb · Thus, the variation of As with q is assumed to be symmetrical
about a = 0.5, As being proportional to a 2/3 when a < 0.3, and to
(1 - a)2/3 when a > 0.7 with a "flat top" connecting the two regions, as
indicated in Figure 6.9.

Previously, when the nonequilibrium vapor generation model was being devel-
oped at BNL, Saha (1977) arrived at an expression for As in the bubbly-slug
flow regime, which he assumed to occur when 0.2 < a < 0.7, by making certain as-
sumptions regarding the geometry of the Taylor bubbles and the partitioning of
the vapor phase between the Taylor bubbles and the bubbles in the bubbly mixture
that fills the space between the Taylor bubbles. A similar approach was fol-
lowed to derive an expression for As for annular-mist flow. Saha (1977) then
proceeded to fit a smootfi function As = Cs [a (1 - a)] 2/3 to the values
calculated according to his model covering all three flow regimes. However,
since this model was intended for low void fractions, it was decided, for the
bubbly flow regime, to use directly the As expression, Equation (6.11) rather
than Saha's approximation to this expression. Moreover, Saha's original model
was found to underestimate the vapor generation rate for a > 0.2, presumably be-
cause of underestimation of As for a > 0.2. For these reasons, the
assumption regarding As Was amended to the one presented here~ which yields a
slightly higher As than does the function As = Cs[a(l - a)] 2/~.

This model is intended for application to the low void ( a< 0.3) bubbly flow
regime. A model is proposed in the next section for higher void fractions.

Interfacial Heat Flux; qi":

For high pressure, high speed critical flows with low void fraction, where
the relative velocity between the phases is expected to be of secondary import-
ance, heat is transferred from the liquid to the interface mainly by transient
conduction. Asymptotic conduction-controlled bubble growth laws under constant
liquid superheat condition, as developed by Plesset and Zwick (1954), and by
Forster and Zuber (1954, 1955) are well known. More recently, Jones and Zuber
(1978) have shown how these concepts can be applied to the case where the super-
heat varies in time simultaneously with variation of vapor density due to de-
compression. If the integral effects are neglected in favor of an instantaneous
superheat approximation, the interfacial heat flux according to these models may
be written as

(6.13)

- 164 -
4.0 ~--~--~--~----~--~--~~--~--~----~~

A
/ ""
/ /
/
/
" ""-
/ "
2.0

DROPLET
FLOW

0.0 L-~L- _ _L -_ _L -_ _L -_ _L -_ _L -_ _~~~~~~

0.0 0.5 1.0


a

Figure 6.9 Variation of Surface Area Density With Void Fraction.


(BNL Neg. No. 10-1337-79)

- 165 -
where Ks is a constant incorporating the effect of the moving boundary, and is
either 11 (Plesset and Zwick) or rr/2 (Forster and Zuber), k2 is the thermal con-
ductivity of the liquid, a2 is the thermal diffusivity of the liquid, t is the
time from bubble nucleation, T2 is the liquid bulk temperature, and Ts is the
saturation temperature, which is assumed to represent the vapor temperature. In
terms of a heat transfer coefficient, Equation (6.13) can be cast into
... k
q i £
h .,...
J. 2 -
T
s
= Ks .; 1T a t
(6.14)
2
Although Equations (6.13) and (6.14) are valid strictly for a constant value of
liquid superheat (T 2 - Ts )' Equation (6.14) is compared with the experimental
data obtained for variable superheat conditions in steam-water (Figure 6.10) by
Niino (1975), and in nitrogen (Figure 6.11) by Hewitt and Parker (1968). In
both figures, the experimental values of the instantaneous heat transfer coef-
ficient are calculated from the measured values of the instantaneous bubble
radius and the instantaneous liquid superheat (T 2 - Ts )' which was varying.
It is surprising to see the good agreement between the trends of the data and
Equation (6.14) at least for shorter time (t < 50 ms).

From Equation (6.13) or (6.14) and the expressions for the interfacial area
density as discussed earlier, one can write the expression for the volumetric
vapor generation rate. as

kR,. T2 - T
s
rv Cr L
S (6.15)
rat
2
where S is given by

S = a 2/ 3 when o < ~< 0.3

S 0.3 2 / 3 = 0.4481 0.3 < a < 0.7


S (1 - 0.)2/3 a > 0.7
The quantity C in Eguation (6.15) has the dimension of inverse length and is
proportional tb Nb l/3. The constant coefficients in Equations (6.11), to
(6.13) are all incorporated in C. The effect of relative velocity, which was
neglected in Equation (6.15) is incorporated in the general model discussed in
Section 6.3.2.

- 166 -
I
NIINO'S DATA
o Fig. 67
a Fig. 66
l> Fig. 68
-.....;:: a
U2OO .-......:::::, ~
~
~ 0, 0
---I
.><ON .-....:::::~
---
' 0
elOO
. .~ l> ~ 00 ,
.c. .~ °oj)o
°R B&
FORSTER- 71'/2Kl
/
. '-....::::::: o~
ZUBER:h= '
.,!'7TO'fT ,~

PL ESSET a .j3K.t
ZWICK: h= ~
.,,71'0.',
, (ms)

Figure 6.10 Comparison of P1esset and Zwick (1954) and Forster


and Zuber (1954, 1955) Heat Transfer Coefficient
with Measured Instantaneous Heat Transfer Coefficient
of Steam-Water During Variable Liquid Superheating
(BNL Neg. No. 10-1596-78)

~.
-"';::::'-"';::::,

o -...:::::.
-...;::::
HEWITT-PARKER'Sg .~ o B
DATA 'i:r-.;::::
o Fig. 9 H g'
o Fig. 10
l> Fig II '"",
PLESSET8 ZWICK: '-...;::::,

'../3K
h=~ /
.J1Tolf FORSTER-ZUBER: '

h =71'/2 KJ
./71' 0t'

, (ms)

Figure 6.11 Comparison of P1easset and Zwick (1954) and


Forster and Zuber (1954, 1955) Heat Transfer
Coefficient with Measured Instantaneous Heat
Transfer Coefficient of Nitrogen During Variable
Liquid Superheating. (BNL Neg. No. 10-1596-78)

- 167 -
In steady homogeneous flow, the relation between the transport time and the
coordinate along the centerline of the pipe in the flow direction is given by

Pm
dt G dZ . (6.16)

The origin of the t-scale here is fixed at the point of net vapor generation,
ZNVG' which is specified as an input condition for each calculation. Along a
streamline, combination of continuity of vapor and the two-phase mixture re-

qUirE :hi~ . ~~. :v) = ( (U :: (6.17)


\

where Pm is the local density of the two-phase mixture. Inserting Equation


(6.15) in Equation (6.17), gives
t

f [er
z (T -
dt
x2 - xl =
k, 9- Ts )'] (6.18)
t P L
m
fa9- rt
1
where xl = x(t l ) and x2 = x(t 2 )·

Given values for the area averaged actual vapor quality x and for pressure,
one can write the following expression for the area averaged void fraction:

x
CI. = (6.19)
C [x + (I-x)
o . G
~J
Pf
+ P g
Vgj

where Pg and ~ are the vapor aad liquid phase densities; G is the mass
flux; Co is the void distribution parameter, taken to be unity at this time;
and Vgj is the vapor drift velocity, which is assumed to be given by the ex-
pressIon for bubbly churn-turbulent upflow (Kroeger and Zuber 1968),

vg]. (6.20)

To calculate the local liquid superheat, the assumption of adiabatic flow is


made. In addition, since the main interest is in flows with low x values
(x < 10- 3 ), the contribution to energy balance from kinetic energy change may
be ignored for flows in a constant area duct. Hence, the specific enthalpy of
the flow may be assumed to remain unchanged. * Thus,

~ = ch g + (1 - c)h9., = const (6.21)

* This will be modified later for the general model.

- 168 -
where c, the static quality of the mixture, is given by
o.p
g
c = (6.22)

and hm' h t ' and hg are the enthalpies of the mixture, the liquid phase, and
the vapor phase. Consistent with the assumption that the vapor is at the local
saturation temperature, the vapor enthalpy is taken to be the saturation value
corresponding to the local pressure. Thus, the liquid enthalpy can be
calculated from Equations (6.21) and (6.22). The liquid superheat can then be
computed from

liT (6.23)

where hf = hf(p) is the saturated liquid enthalpy and Cpt is the liquid
specific heat.

The quantity in brackets in Equation (6.18) is not expected to vary drastic-


ally within small step sizes (t2 - tl)' Thus, a first-order integration
formula can be used, yielding

(6.24)

where (> signifies the mean value.

Now all the terms except C rin Equation (6.24) are expressed as functions of
'the local pressure p. Therefore, if pet) is known from experiments, C r can in
principle be solved. In practice, however, an iterative procedure is adopted.

To perform a calculation, the initial conditions, in particular ZNVG, and


the void fraction 0. 0 at this location must be known. Note that, if the vapor
nucleation rate could be calculated, both the location of the point of net .vapor
generation in the pipe and its void fraction could be determined by inte-
grating the nucleation rate. In the absence of an applicable nucleation equa-
tion, values of % and ZNVG must be assumed.

- 169 -
B. Application to Pipe Flows

To determine the value of C from Reocreux's data, the following procedure


was used for each experiment: r

(i) Determine the point where the measured static pressure p is equal to
Psat(T in ) where Tin is the temperature at the test section inlet.

(ii) Set the constant in Equation (6.21) to h in , the enthalpy at the


inlet, which is equal to the enthalpy corresponding to the saturated
liquid at the pressure determined in (i).

(iii) Assume a o ' ZNVG' and C values.


r
(iv) Starting at Z = ZNVG' and a o ' calculate the local pressure by
interpolating the measured pressure distribution, and calculate
the local liquid superheat ~T, etc., from Equations (6.21) to (6.23).

(v) Increase Z by step size ~Z, and estimate ~T, etc. Take average.

(vi) Calculate new x from Equation (6.24) and a from Equation (6.19).

(vii) Evaluate new superheat, etc., and compare with estimates in (v).
Improve until convergence is reached.

(viii) Repeat (v) to (vii) until end test section is reached.

(ix) Compare a(Z) calculated with a(Z) measured; change a o ' ZNVG'
and C values; and repeat (iii) to (ix) until agreement is reached.
r
The point of net vapor generation ZNVG coincides roughly with the end of
the nucleation zone, where a bubble population of Nb , with a certain average
bubble size rb' is generated. Since C depends on Nb , and a o is simply
r
(Nb 4TI~b3/3), it is clear that a o ' ZNVG' and C are related and their
r
values cannot be assigned independently of one another. Fortunately, the
calculated void fraction distribution a(Z) was found not to be sensitive to
changes in the values of a o assumed in this model, and all subsequent
calculations in this section were made with a o = 0.00002 regardless of the C
value. A step size of Z = 0.5 mm was chosen after it passed the usual con-
r
vergence test of the program.

Figure 6.12 shows results of a series of calculations performed on Exp. 421.


Note that the curves of a(z) become steeper as Cr is increased while ZNVG is
fixed, whereas they are translated downstream and steepen slightly as ZNVG is
increased for constant Cr. These effects are to be expected physically since
the vapor generation rate depends directly on Cr , whereas an increase in ZNVG
delays the flashing point, causing the ensuing bubble growth to occur unde~
higher superheat conditions, resulting in a higher rv. Assisted by these ob-
servations, one can make judicious choices of C and ZNVG to achieve the best
r
fit to the experimental data. For this experiment, the combination Cr = 3700
m- l and ZNVG = 1.21 m seems to yield the best agreement with the
measurements.

- 170 -
The results of calculations which give the "best fit" to all experiments
evaluated are summarized in Table 6.1 and Figures 6.13, 6.14, and 6.15. In
general, the model works well for about a < 0.3, but at higher void fractions it
seems to under-predict the vapor generation rate. Maximizing the interfacial
area density for a given Nb did not lead to a sufficient increase in a to
reconcile the difference shown in these figures. Extending the bubbly flow
(Equation 6.11) and droplet flow (Equation 6.12) formulas for As to the point
of a = 0.5 (shown by the broken line in Figure 6.9) gave a calculated value
for one experiment only slightly higher than those presented here. However,
such an assumption on flow regimes is unrealistic (Dukler and Taite1 1977), and
it serves only as an upper bound for estimating possible errors in rv due to
errors in the As assumptions. The discrepancy between the measured and
calculated values of amay be attributed to the inadequacy of the conductive heat
transfer model which is assumed to control the- bubble growth rate and, in turn,
the vapoi generation"rate. A more generalized model is presented in the next
section and applied to BNL nozzle flows. In the transition regime to bubb1y-
slug flow, relative motion of the phases may significantly improve the heat
transfer rate and thus raise the bubble growth rate over that predicted by the
conduction-limited rate. Moreover, the a distribution measured by Reocreux re-
presents diametrically averaged values of a, not the cross-section averaged
values calculated from the present model. For certain radial void distribution
profiles, the diametrical average may be higher than the cross-sectional average
void fraction.

Figure 6.16 shows the relationship between the quantity (Psat - PNVG)
and the mass flux G. Sin~e the liquid superheat at the point of net vap6r
generation is directly related to the pressure difference (Psat - PNVG) , the
latter may be taken as a measure of the maximum superheat sustained by the
liquid before flashing occurs. On the other hand, since the pressure variation
of a liquid with distance in the flow direction in a straight pipe scales as
(1/2) u Z , the time rate of pressure drop is proportional to (1/2)u 3 , or to
G3 (Equation 6.9). Hence, Figure 6.16 may be interpreted as displaying the
effect of expansion rate on the limit of superheat determined from Reocreux's
experiments. It is interesting that higher superheat or (psat -PNVG) was
reached for lower expansion rates and lower superheat for higher expansion
rates. This trend, which has also been observed by Seynhaeve et al., (1976), is
the opposite of that found in supersonic flows with vapor condensation, where
the critical supersaturation, the counterpart of (psat -PNVG) here,
attained is inversely related to the vapor cooling rate, the counterpart of
expansion rate here. This behavior was related to the turbulent pressure
fluctuations by Jones (1979) (see Section 6.2.2).

Finally, the values of ZNVGand Cr determined from these experiments are


presented as functions of the mass flux in Figures 6.17 and 6.18. Again, a
trend opposite to that found in condensing flows is seen in the (ZNVG -
Zsat) curve. The trend of steadily increasing C with G in Figure 6.18 dif-
r
fers from wHat would be expected on the basis of the trend of (psat -PNVG)
vs. G found previously. A low value of (psat - PNVG)' as found in the high
G case, would give rise to a lower nucleation rate leading to a lower C , but a
higher C was found with the current calculation method. r
r .

- 171 -
Table 6.1

CONDITIONS IN FLASIlING FLGI Of WATER IN A CONSTANT AREA PIPE

Measured by Reocreux Deduced From Heasurements Calculated From Present Hodel


I

Exp. G Tin Ptn Psat(T tn) lsa t (ltn - f'sat lNVG Cr PNVG Put - PNVG lNVG - lsat
No. (kg/m 2s) (e) (kPa) (kPa) (m) (kPa) (m) (m -I) (kPa) (kPa) (m)

400 6526 116.1 246.5 118.1 0.511 61.8 1.23 3000 163.6 15.1 0.66 I

401 6465 116.6 246.1 118.1 0.669 68.6 1.23 3000 164.9 13.2 0.56
402 6496 116.1 246.1 118.5 0.622 61.6 1.23 3000 164.5 14.0 0.61
\,

403 4193 116.1 212.3 118.8 0.215 33.5 1.20 3200 163.5 15.3 0.93
404 4165 116.1 211.8 118.8 0.246 33.0 1.22 3100 163.1 15.7 0.91
......
'-.I
N 406 8718 116.3 289.6 116.3 0.903 113.3 1.30 4000 162.5 13.8 0.40
421 6559 121.0 273.2 204.8 0.596 68.4 1.21 3100 190.4 14.4 0.61
423 4383 121. 9 243.3 210.8 0.086 32.5 1.15 3500 192.2 18.6 1.06
430 8529 121.1 313.1 205.3 0.826 108.4 1.22 4300 193.2 12.1 0.39
434 10111 120.9 339.5 204.0 0.963 135.5 1.22 5000 193.9 10.1 0.26
440 6410 125.5 ----- 235.6 0.484 ----- 1.18 4200 219.5 16.1 0.70
448 4210 126.1 ----- 239.9 -0.105 ---- .. 1.12 3100 220.0' 19.9 1.23
451 8520 125.1 ----- 232.7 0.801 .. ---. 1.20 5000 220.5 12.2 0.40
455 10176 125.0 ----- 232.3 0.938 ---- .. 1,22 6300 221.4 10.9 0.28
REOCREUX EXP.421
G'6559 kg/m 2 ,s,Tin ,'21.0 C

SYMBOL

3700 1.21 (BESTFITl


--------- 3500 1.15
0.4 - - - - 3000 1.20
-- - 3500 1.20
-- - -- 4000 1.20
a ---- - 3500 1,25
o
o MEASURED DATA

1,15 1.20 1.25 1.30 1,35


Z (m)

Figure 6.12 Variation of a with Cr and ZNVG in Present Model and Comparison -witr
Experimental Data of Reocreux. (BNL Neg. No. 10-1335-78).

8
0.4 .

a
~1n)!
\~200
EXP 406 2)
G = 8~18 kg/m s
ZSOI- 0.903m
(
0.2 C = 4000 m- I
r
o

o
O.O~~~--~----~~~--~----~---L--~
0.4 0.6 0.8 1.0
Z-ZSOI (m)

Figure 6.13 Void Fraction Distribution in a Constant Area Channel as a function


of Mass Flux G. Tin = 116 C.o,Measurements by Reocreu~;
----- , Calculations Based on the Cr Values Shown.
(BNL Neg. No. 10-501-78)

- 173 -
0
0
Tin""121 C
0.4
~xP 434,kg/m2)s f23)
4383, .
0.086,

1f21)
G=IOIII 6559, 3500.
a
Zsot~O.963~,· ~6~~) 0 . 596 , *
0
3700.
Cr =5000m 0 0.826:
0.2
oI I 4300.
0
.

0.4 0.6 0.8 1.0 1.2


Z-Zsot (m)

Figure 6.14 Void Fraction Distribution in a Constant Area Channel as a Function of


Mass Flux G. Tin ~ 121 C. 0, Measurements by Reocreux;
----- , Calculations Based on the Cr Values Shown.
(BNL Neg. No. 10-504-78)

1In ""'126C

0.4
Exp 455, '
2
G= 10176 kg/m s,
f48~
4210,
-0.105,
0
3700.
Zsot =0.938 m,
a

0.2
Cr =6300m- l.

~451' )
8520,
f40) 6410,
0.484,
o 4200.
I 0.801,
* 5000.
0.0
0.2 0.4 0.6 0.8 1.0 1.2 1.4
Z-Zsot (m)

Figure 6.15 Void Fraction Distribution in a Constant Area Channel as a Function of


Mass Flux G. Tin - 126 C. 0, Measurements by Reocreux
~, 8alcu1ations Based on the Cr Values Shown.
(BNL Neg. No. 10-502-78)

- 174 -
20
SYMBOL T j n (C)
0 0 116
0 121
A 126
0
a... t:.
~ 0
0
C) 0
2
> 15
0.. 0
It- O
<t 0
0..CJ)
0

10 o

o 5x10" 10x10"
G3 [(kg/m 2 .s)3]

Figure 6.16 Effect of Expansion Rate on the Pressure at the Point of Net Vapor
Generation. (BNL Neg. No. 10-1336-78)

- 175 -
SYMBOL Tin (e)
0 116
Cl Cl 121
1.0
I:. 126
0
0
~

E
l-
e:[
en
N I:.
I 0
<.!)
> cP
Z
N 0
0.5
eo

4000 6000 8000 10000

Figure 6.17 Variation of Point of Net Vapor Generation Z G with


Mass Flux G and Initial Temperature T in FIXshing Flow.
Values of ZNVG \Olhich Give "Best Fit II tg
Reocreux' s Data.
(BNL Neg. No. 10-1334-78)

SYMBOL ~
0 116
6000
Cl 121
I:. 126
Cl
I
E
~ Cl
I:.
u 4000 0
cr:. Cl
Cl
0

2000~--~-----L-----L----~----L---~----~----~
4000 6000 8000 10000
G (kg/m 2 .s)

Figure 6.18 Variation of Cl' with Mass Flux G and Initial Temperature
Tin in Flashing- Flow. Values of C Which Give "Best Fit"
to Reocreux's Data. CBNL Neg. No. 10-1333-78)

- 176 -
.The quantity C r incorporates the constant coefficients in Equations (6.11),
(6.12), and (6.13), in addition to Nb l/3. Thus, variations in ~ may re-
flect the need to modify the conductive heat trasnsfer coefficient adopted here.
Contributions from convections, turbulent diffusion, etc., may significantly
alter the heat transfer coefficient and C r '

In the general model some of the simplifying assumptions of the low void
fraction model are relaxed.

6.3.2 General Model

A. Model

It was shown above that the heat trasnfer limited vapor generation rate fol-
lowing flashing inception is

(6.25)
r = A q "/L
v s i

where As is the total area of the liquid-vapor interface per unit volume of
the mixture, ~' is the heat flux to the interface, and L is the latent heat of
vaporization. Both As and ~' are flow regime dependent, and in each flow re-
gime they are functions of t~e thermodynamic state and flow variables. There-
fore, to calculate As and q:' it is necessary to know, a priori, in which flow
regime the system is expect~d to be. In this model, the flow regime is assumed
to be a function only of the local void fraction a. Thus, bubbly flow,
bubbly-slug flow, a transitional flow comprising the annular and annular-mist
regimes, and finally fully dispersed droplet flow are assumed to occur at suc-
cessively higher void fraction ranges, as shown in Figure 6.19. The void
fractions at the transition points are assumed to be

a b max 0.3, as max 0.8, 0.95

These values may be modified according to the comparison with experiments.

The interfacial area densities and the heat transfer coefficients applicable
to interfacial heat transfer for each flow regime shown in Figure 6.19 are dis-
cussed below.

Bubble Nucleation Zone, 0 < a < aD

The bubble nucleation zone is the point of flashing inception, and it serves
as the starting point of the vapor generation calculations. The location of
flashing inception is to be dete.rmined from the flashing inception correlation
discussed in Section 6.2. At the inception point, the vapor is assumed to be in
the form of critical sized bubbles, with radius Rcr:

- 177 -
Inception

"'0
~
~
'"'
!1j
......
......
rz..
::I

~~rz..
01.1
Bubbly Bubb 1y-S1ug QJ

Flow Flow o.
...... 01.1
e-0
!1j III
C::"r"l
~ '"'
o~ "'0
"r"ll QJ
III
"r"l '"'
01.1 !1j
Ill ......
~ g '~"'
III
"r"l
'"' c::
H< ~

Oa a '\, max

Figure 6.19 The Flow Regimes Map for the General Model.
(BNL Neg. No. 1-917-81 ).

- 178 -
20/ CPs (T~) - P ) (6.26)

The critical radius is of the order of a few microns for typical liquid su-
perheat (or pressure undershoot) values found experimentally at flashing incep-
tion. At these sizes the bubbles can certainly be considered to be spherical
and to move with the liquid with no slip. Thus the following expressions can be
written to start the calculation for vapor generation:
41T R3 (6.27)
a 3 cr Nb
o

(6.28)

(6.29)

where Nb is the number of bubbles per unit mixture volume.

Bubbly Flow

For 0.0 < 0.< '1:, max' the flow is assumed to be in the bubbly regime.
The vapor exists in the form .of bubbles of uniform size, although not neces-
sarily spherical. The relative velocity of the bubbles with respect to the
liquid Vg~ and the drift velocity Vgj are given by Wallis (1969) as

Vg ~ = Vg j / (1 - a) v "" (1 - a) n-l , (6.30)

(6.31)

where n and v"" represent an empirical index and the bubble-free rise velocity in
an infinite liquid, with the values given in Table 6.2 for the various bubble
Reynolds number and size ranges. In Region 1 in Table 6.2, the bubbles retain
the spherical shape, and the rise speed is based on the Stokes formula. In re-
gion 5 the bubbles have spherical cap shapes. In between, the shape changes
from spherical to spheroidal to ellipsoidal to spherical cap. The rise
velocities of steam bubbles in water at three temperatures calculated from these
correlations are given in Table 6.3 and compared with data from single bubble
air-water experiments at room temperature in Figure 6.20. The equivalent sphere
radius W is defined such that the bubble volume can be written as

V = 41T riJ/3 (6.32)

The discrepancy at large radii (Region 5 in Table 6.2) may be attributed to


interference of neighboring bubbles (Zuber and Hench as quoted by Wallis 1969).
Note, however, that application of the v"" formulas will be limited mostly to
W below the order of 1 cm, above which transition to slug flow occurs.

- 179 -
TABLE 6.2. BUBBLE RISE VELOCITIES

Region
.= =--
v., Reb range W range n

1/3
2 w2 (PR,-p&) g \I 2
1
9 IIR,
R~ < 2 W <
[i 9.
(1 - p/PS) g J 2

{9\)R,2/[2 (1 - p/PR,)g] ~/3 < W


2 0.33 gO.76 "i o•52 wl •28 2 <: R~ <: 4.02 G -O. 21 "
1 1. 75
< 2.21 ("R, 2/g)
1/3
. G1 -0.0939

1/3
G -0.093! ',<

~
4.02 G - 0.21" <: Reb 2
1 2.21 ("9. / g) • 1 W
3 1. 35 1. 75
PR, W
<' 2.59 G - 0.25
< 0.918 ..J09./(P9. g)
1

1/" R;, > 2.59 G -0.25 0.918 ,jOR,/(PR, g) < W


1
[g oR,(PR,-P )]
4 1.41 il and 1.5
P 2
I.
W < 2 .J OR,/(PR, g) <: 2 ..Jog/(PR, g}

5 Same as 4 w> 2 ..J OR,/(PR. g) W > 2 ,j oR,/(PR, g) 0

11'3
W is the equivalent sphere radius of the bubble = (3 V/4 rr)

"R, is the kinematic viscosity of the liquid = IJR,/P9.'


Reb - 2PR.V.,w/lJR,' G1 =g 119."/(PR, 09. 3
).

- 180 -
TABLE 6.3

FREE RISE VELOCITIES OF STEAM BUBBLES IN WATER

---- ~-
-------

I~egion~ T = 120°C T = 300°C T = 350°C

wb < 30 ~m ~ < 20 ~m ~ < 21 ~m


1
v~ (30 ~m) = 8.4 mm/s v~ (20 ~m) = 6.4 mm/s v 00 (21 ~m) = 6.0 mm/s

30 ~m < wb < 0.64,mm ]9.4 ~m < ~ < 0.39 mm 19 ~m <'~ < 0.28 mm
2 v~ (30 ~m) = 8.2 mm/s va> (19.4 urn) = 6.7 mm/s Voo (19 ~m) = 6.6 mm/s
v~ (0.64mm) = 41.5 em/s va> (0.39 mm) = 31 em/s v (0.28 rom) = 21e~/s
~
f-'
co
f-'
0.66 mm < ~ < 2.2 mm '0.4 mm < Ii1, < 1.3 mm 0.29 mm < ~ < 0.74 mm

3 v~ (0.66 mm) = 40. em/s va> (0.4 mm) = 30 em/s v~ (0.29 ~) = 20. em/s
Voo (2.2 mm) = 21.8 em/s va> (1.3 mm) = 16.7 em/s Voo (0.74 me) = 12.5 em/s

2.2 mm < ~ < 4.9 mm 1.3 mm < ~ < 2.9 mm o. 74 mm < ~ < 1. 6 mm
4
v,., = 21.8 em/s va> .. 16.7 em/s v." = 12 em/s

~ > 4.9 mm Ii1, > 2.9 mm ~ > 1.6 me


5
Same as Region 4
Figllre 6.20 Comparison of Correlation of Table 6.2 with Experiment.
(BNL Neg. No. 1-900-81).
Ca1cu1ations:-----()----- Te 120°C
----Q---- 300°C Uater-Steam
Bubbles
-.-~. 350°C

Experiments: Air-Water from Haberman and Morton (1953)

- 182 -
Because of the complexity of implementing the general form of the bubble
rise velocities, as shown in Table 6.2, a simplified form of the vapor drift
velocity is adopted here:
ga(P£-p ) ] 1/4
Vgj 1.41 [ 2 g i f w > 0.5 (6.33)

For W < 0.5 la/(P£g), a linear interpolation between zero and the above value
is used.

Attention is next focused on the interfacial heat transfer coefficient to be


used for the bubbly flow with relative velocity between the bubbles and the
liquid surrounding them. In Section 6.3.1, a Plesset-Zwick or Forster-Zuber
type of heat transfer coefficient was chosen for the bubbly flow with zero re-
lative velocity. Now that expression, i.e., Equation (6.14), should be modified
in such a way that the convective heat transfer due to the relative velocity is
also accounted for.

One of the requirements for the general expression is that,it should satisfy
the limiting behavior of the heat transfer process at both t+ 0 and t+ ~ . At
the inception point, transient conduction dominates the heat transfer process so
that the general expression should approach the PIes set-Zwick or Forster-Zuber
expression at t + O. However, as the bubbles "age," the convecting heat trans-
fer due to the relative velocity between the bubbles and liquid starts to domin-
ate (Wolfert, 1976). Therefore, at t + 00, the general expression should yield
the coefficient for steady-state convective heat transfer to the bubbles.

The rate of steady-state heat transfer to a spherical bubble moving at con-


stant speed in an infinite liquid is given by the Boussinesque solution (Chao,
1969) which may be expressed as

(6.34)

where the subscript ss stands for steady-state bubble motion, and the bubble
Peclet number is defined as 2V £w/a Q• For a spherical cap, the heat transfer
rate based on the surface areagof the equivalent sphere is estimated (Davenport
et al. 1967) to be about 20% higher than that for the sphere. This estimate
agrees with Calderbank and Lochiel's (1964) model and is in fair agreement with
experiments (Calderbank and Lochiel). Thus, in terms of the heat transfer coef-
ficient based on the equivalent sphere,

where Cl = 2.0 for spheres


2.88 for spherical caps. (6.35)

- 183 -
In the model being developed here, the transition in bubble shape is assumed to
occur entirely in Regions 3 and 4 of Table 6.2. Thus Cl is assumed to vary
linearly with W
from Cl 2.0 at W = 2.tl(u~2/g)1/3Gl-0.0939 (6.36a)

to 2.88 (6.36b)

Two simple expressions are found in the literature which provide a smooth
transition from the Plesset-Zwick expression at t + 0 to the steady-state heat
transfer coefficient at t + 00 These are

C Vg J1. t 11/2
/3.k J1. l
= [ 1 + 3 (6.37a)
h J1.
~
W
J
J1.
and

h ~ =
fikJ1. [
1+
~ -C1
3
V
• ~W
,t ] (6.37b)
l'I1'a~ t I

The first is due to Aleksandrov et al. (1967) as modified by Saha (1977) and the
second to Wolfert (1976). They can also be written, respectively, as

h Q, (6.38a)

and

~
-p -Z
+ h ss (6.38b)

where hp_Z Is the Plesset-Zwick heat transfer coefficient with Ks = 13 in


Equation (6.14). Since both these expressions were based on intuitive physical
arguments, and it is not clear which is more realistic, both should be compared
with experimental data in the hope that one will turn out to be the better cho-
ice. Note, however, that Wolfert's expression (6.37b) always yields a greater
value for h ~ For the present calculations, the modified Aleksandrov expression
(6.38a) with Cl = 2.0 is used.

Since the heat transfer coefficients are defined in terms of the equivalent
sphere, the interfacial area density can be calculated from

3a/w (6.39)

regardless of whether the bubbles are spheres or spherical caps. However, a

- 184 -
method of calculating the equivalent bubble radius, w , and the void fraction,
a , is required.

In the bubbly flow regime, it is assumed that bubble coalescence does not
occur. Thus, the bubble radius changes only as a result of vaporization or con-
densation at the interface,
·11
dw qi
-= --=
dt Lp (6.40)
g

This equation is equivalent to Equation (6.25) written on a basis of unit inter-


facial area rather than unit volume.

To calculate the void frac tion one must first calculate the actual flow
quali ty, x, by using the steady-state vapor continuity equation
dx
rv A ."
sqi A h
s,b t (Tt-T sat )
-= =
dz G GL GL (6.41)

The vapor phase velocity, Vg , is then used to change the independent variable
from z to t to be compatible with Equation (6.40). Once the flow quality is de-
termined, the void fraction can be computed as in Section 6.3.1.

An alternative approach is to utilize an equation for conservation of bubble


number for the bubbly flow with no bubble coalescence or further bubble nucle-
ation (or generation). For steady-state, this equation results in the following
simple equation:

Therefore the bubble density, Nb' can be calculated at any cross-section and
the void fraction can then be calculated from:

a (6.41b)

This approach, along with Equation (6 .. 40), has been used in this report for
numerical advantages.

The. liquid superheat is calculated from the mixture energy equation by as-
suming the vapor phase to be at saturation. This is similar to what was done in

~ + t (~ m
r
Section 6.3.1, but the effect of mixture kinetic energy is now included so that

= constant

This concludes the modeling for the bubbly flow regime.


(6.42)

- 185 -
Bubbly-Slug Flow

As the void fraction increases, bubble coalescence becomes more significant.


Although in an idealized situation with a static, simple cubic configuration of
uniform-sized spherical bubbles, geometric interference of bubbles occurs
only when a > 0.524,* in reality (Dukler and Taitel 1977) bubble coalescence
sets in well below this void fraction value. It is assumed here that, when a
~ max = 0.3, some of the bubbles begin to coagulate to form larger bubbles,
while the others continue to grow by vaporization according to the rate and
mechanism discussed above. Thus, two classes of bubbles coexist in this void
fraction range, the larger bubbles formed by coagulation and the smaller orig-
inal bubbles. As a result of vaporization at the interface, both classes of
bubbles grow, although at different rates. In the meantime, the number of bub-
bles that have undergone coagulation increases and therefore the number of smal-
ler bubbles decreases. The larger bubbles resulting from coagulation are proba-
bly not spherical. They may either approach the spherical cap shape, or, if w
is comparable with or greater than the pipe diameter, acquire an elongated cyl-
inder shape with a rounded head. These latter bullet-shaped bubbles are some-
times called Taylor bubbles, and they are usually separated by liquid (or bubbly
liquid) regions, i.e., slugs.

Since vapor generation may take place on the surface of the Taylor bubbles
as well as the small bubbles, the vapor generation rate or the total heat trans-
fer rate is the sum of the two components:

As , T + %" As , b (6.43a)

or

(6.43b)

where the subscripts T and b designate quantities pertaining to the Taylor bub-
bles and the small bubbles. The total interfacial area density is

As,b + As,T (6.44)

To estimate the interfacial area density, a few assumptions are made con-
cerning the partitioning of the voids between small bubbles and Taylor bubbles.
Figure 6.21, which illustrates the nomenclature for bubbly-slug flow in a cir-
cular pipe shows the Taylor bubbles; the unit cell, in which one Taylor bubble
may be found on the average; and the small bubbles in the bubbly mixture between

* Note that this limit is not a function of bubble number density. For a
given Nb, the maximum diameter of spheres which may be fitted in a simple
cubic lattice is Nb -l/3, which corresponds to a = 0.524.

- 186 -
SMALL BUBBLE

TAYLOR BUBBLE I UNIT CELL


>..
c
D
2X T ~

L
2Rr

r~.,-,:ure 6.21 Nomenclature of Bubbly-·ratt,~ Flow.


(BNLNeg. No. 1-1143-"1)

- 187 -
two Taylor bubbles. The Taylor bubbles are assumed to be cylinders of length
2ArRT and radius RT • Thus, their equivalent sphere radius is

3
4'JTW T
-3- = V
T
or
(6.45)

and the surface to volume ratio of the cylinders is

1 1
2( v;-+ 1)
1) 2( v;- +
ST T
T (6.46)
V
T Rr ( _2_ )1/3
wT 31..
T
For comparison, the surface to volume ratio of spherical caps (subscript c) is

S 2 2
c 3(3-2t,; -t,; ) 3(3-2t,; - t,; ) (6.47)
V
c R (2-3t,; + t,;3) W [2(2-3t,;+ t,;3) ] 2/3
c c
where t,; = cos ~ of a spherical cap, ~ being the half subtended angle, and Rc
is its radius. Plots of (S/V) W for cylinders and spherical caps, as a func-
tion of A and t,; respectively, appear in Figure 6.22. For most spherical cap
bubbles, ~ ~ 50° or t,; ~ 0.6, and the surface to volume ratio is not very dif-
ferent from that for cylinders with a ~ of 4 or 5. Thus, all large bubbles
formed by coalescence of small bubbles are assumed to be cylindrical, with A to
be optimized in comparison with experimental data if necessary.

Next it is assumed that, as the void fraction increases, the Taylor bubbles
absorb (or suck in) the neighboring smaller bubbles and eventually merge with
one another to form a continuous vapor core. This represents the end of the
bubbly-slug regime and the beginning of the annular flow regime, which has been
assumed to occur at a = as max. Therefore, at a = as max'

(6.48)
la s max and

It is also assumed that the length to diameter ratio of the Taylor bubbles re-
mains the same for the entire bubbly-slug region. Therefore,

(6.49)

- 188 -
A • L/ZR

8 ,0 7 6 8 10
"
3

",;>


w

6
Host sPllerfcaA
caps, •• 50

I
t-'
CP
\0

.11

0
90

I
• • 1500 Ito
2L -1 -0.8 -0.6 -0.1' -0.2 0 0.2 O.~ 0.6
t • cos •
0.8 1

Figure 6.22 Surface to Volume Ratios for Cylinders and Spherical


Caps (Sphere: L = 3.0)
(BNL Neg. No. 1-918-81).
Combining Equations (6.48) and (6.49), gives the length of the unit cell as

Lc AT 0 las max (6.50)

Therefore, the void fraction due to the Taylor bubbles alone is given by

aT
V
T
8
~ (6.51)
V 3
c la 0
s max
and the interfacial area density due to the Taylor bubbles can be given by
0. 2 /
3
As,T 4 ( 1 + 2~ T ) T
al /6 0
(6.52)
5 max

Note that, if the length to diameter ratio of the Taylor bubble is 5, the sur-
face area of the two ends of the cylinder is only 10% of the total surface area.
Moreover, the heat transfer at the lateral surface of a Taylor bubble is ex-
pected to be more efficient than that at the two ends because of higher relative
velocity. Therefore, for the calculation of vapor generation, the interfacial
area density due to the Taylor bubbles can be approximated as
2/3
4 aT
As,T = (6.53)
1 6
0. / 0
s max
The advantage of this expression is that it obviates the need for further
adjustment of the parameter, ~.

The average void fraction of bubbly-slug flow should include vapor volumes
in the Taylor bubbles as well as the small bubbles. If the volume of a small
bubble is denoted by Vb, then

a (6.54)

where Ob = Vb/Vc is the void fraction due to the small bubbles. At


a = a b max' the beginning of the bubbly-slug regime, a b = a = a b max and
aT = O. At a = Os max' the end of this flow regime, a b = 0 and
aT = 0.= Os max' Therefore, ata = Cis max' there will be a liquid film
region of volume (1 - Os max)V c with no bubbles. This leads to the idea of
an "effective" volume of the bubbly mixture, which may be written as

Vb,eff = [ Vc - f (1 - Os max) Vc - VT] (6.55)

where f o for a = a b max and f 1 for a = a s max

- 190 -
One possible expression for f is
ct. - ct. b max
f = ( 6.56)
CI. s max -ct. b max

Next, ct.b' is defined as the void fraction due to the small bubbles on the
basis of the "effective" volume of the bubbly mixture:

EV ct.
b b
~~~------~-- (6.57)
Vb,eff
l-f(l-ct. s max)-ct. T
Consistent with the earlier assumption that bubble coalescence starts at
(J, =ct.
b mlRe local void fraction in the bubbly mixture of bubbly-slug flow is as-
sumed never to exceed a b max. Thus, as the average void fraction ct. increases
beyond ab max' some small bubbles are removed from the bubbly region to form
Taylor bubbles, and ct. b I is bounded by ct. b max:

ct. b' ct.b max· (6.58)

Combining Equations (6.54) to (6.57) yields the following expression

ct. = (1 _lct. ) [ct. -ct. max)l _ ~~-ct.s ma~:~ct.-ct.b)max)l ]


T
b. max b I s max max \ (6.59)

Note that the above expression yields the correct limi ts for Cl T at = ct. b max
where ct.T = 0, and at ct. = ct. s max where ct.T = CS max = ct.. Equation (6.59)
together with Equation (6.53) provides the interfacial area density due to the
Taylor bubbles alone.

Regarding the heat transfer coefficients appropriate for Taylor bubbles,


since most of the interface on the Taylor bubble is the lateral surface of the
cylinder, this area is probably responsible for most of the heat transfer. The
heat transfer coefficient to the cylindrical surfaces is approximated by that to
the liquid films. This approximation is the same as that in the TRAC-PlA code
for slug flows (see the TRAC-PlA manual and Rohatgi and Saha 1980). Linehan
(1968) measured the heat transfer coefficient over subcooled liquid films dur-
ing condensing heat transfer of steam and arrived at the following correlation:

Nu£,
St £, constant = 0.0073 , (6.60)
Re£, Pr£,
i.e., hT 0.0073 P£, V £, c p £'. (6.61)

- 191 -
Note that in the TRAC model the Prandtl number dependence in Equation (6.60) is
neglected, and hT becomes hT = 0.0073 P k Vg I ~ . This assumption ob-
viously has limited validity. In additlo~, the ~ffect of vapor shear is not
included. Therefore, the search for a better heat transfer correlation for
Taylor bubbles should continue, and improvements to Equation (6.60) should be
made.

For the small bubbles in the bubbly-liquid mixture between two Taylor bub-
bles, the interfacial area density is given by

As , b (6.62)

and the modified Aleksandrov equation, discussed earlier, continues to be used


for the heat transfer coefficient.

The drift velocity in the slug flow regime has been given by Zuber and
Findlay (1965), among others, as
.I g ( P 9. -P gJD
Vgj,T 0.35 " (6.63)
PQ,
However, in the low void fraction range of the bubbly-slug regime, the flow
field is still dominated by the small bubbles and not by the Taylor bubbles, and
therefore the relative velocity should still be influenced mostly by the rise
speeds of the small bubbles, with no discontinuous change in VgQ, expected.
Thus, the vapor drift velocity in bubbly-slug flow is assumed to be given by the
following volume averaged expressions:

(6.64)

and

(6.65)

where Vgj,b is determined from formulas in Table 6.2 or from Equation (6.33)
and Vgj,T from Equation (6.63). The relative velocity calculated from Equa-
tion (6.65) is to be used in the calculation of the heat transfer coefficients
hb and hT in the bubbly-slug regime.

To summarize, it is assumed that fora b max < a < C5 max' bubbly-slug


flow prevails, where a s max is the maximum void fraction for slug flow to ex-
ist. In the bubbly-slug regime, voids exist in two forms, small bubbles and
Taylor bubbles. The void fraction is partitioned between the two, withab'
assumed to be fixed at a b max' The interfacial area densities of the two
forms of voids are expressed in terms of the void fraction, the length to dia-
meter ratio of the Taylor bubbles, the pipe diameter, and the equivalent radius
of the small bubbles (Equations 6.52 and 6.62). The heat transfer coefficients
are Equation (6.61) for Taylor bubbles and Equation (6.37a) for the small bub-
bles, with the bubble growth rate given by Equation (6.40). For the relative

- 192 -
velocity, Equation (6.64) is adopted. Total heat transfer rate is calculated
from Equations (6.45). The upper limit of the bubbly-slug flow regime is as-
sumed to be as , max = 0.8.

Annular and Annular-Mist Flow

Annular and annular-mist flow take place in the transitional region between
the bubbly-slug and the dispersed droplet regimes. As pointed out earlier, in
the present model the flow pattern changes from the bubbly-slug regime to the
pure annular flow regime at a = as max' For pure annular flow, the
interfacial ar~a density and the heat transfer coefficient are given by

(6.66)

and

h (6.68)

The situation becomes complicated as droplet entrainment begins. The onset


of droplet entrainment can be determined from the correlations of Ishii and
Grolmes (1975). However, there are major uncertainties regarding the rate of
entrainment and the sizes of the droplets entrained in the vapor core. TRAC-P1A
uses the Wallis correlation (Wallis, 1969), which is a modified form of the
Pa1eev and Fi1ippovich correlation (1966), to determine the mass fraction of the
entrained droplets. However, these correlations do not seem to be adequate, and
further investigation is required in this area.

Dispersed Droplet Flow

When a > ad, the liquid is assumed to be fully dispersed as droplets. The
value of ad is assumed to be 0.95. However, the exact value of Cld can also
be calculated from the model for the annular-mist flow as the liquid film at the
wall dries out because of droplet entrainment and film evaporation. In the dis-
persed droplet regime, the droplet size is stabilized by the balance between the
surface tension and aerodynamic forces, and the droplet diameter can be expres-
sed in terms of a critical droplet Weber number,

(6.68)
cr

Wallis (1969) gave the critical Weber number as 12 for droplets in a low viscos-
ity carrier gas, whereas Gyarmathy (1976) found the literature value of Wed in
the range

8 < Wed , cr < 15 •

Wallis's value of 12 appears reasonable, therefore

- 193 -
12 a
d (6.69)
v2
P g g,V,
In the TRAC-PIA manual, a critical Weber number of 5 is quoted, but according to
Rohatgi and Saha (1980) the value used in the TRAC code is actually 2. This
seems too low compared with most of the values quoted in the literature. As-
suming spherical droplets, the interfacial area density can be given by
'6 (l-a)
(6.70)
d
The relative velocity of the droplets is given by Wallis (1969) as

1~4 [g O(:~ - Pgl ] 1/4 (6.71)

which has a form similar to that for Region 5 in Table 6.2 except that Pi is re-
placed by P g in the denominator.

The convective heat transfer t,o liquid droplets has been investigated by a
number of researchers. Most of the data can be correlated by the expression

(6.72)

where the droplet Nusse1t number is

(6.73)

the droplet Reynolds number is


P g Vg i d
Red (6.74)
llg
and the vapor Prandt1 number is
c II
Pr g Eg g (6.75)
k
g
Ranz and Marshall (1952) found C3 = 0.6 for various liquid or solid spheres in
air or other fluids. Lee and Ryley (1968) in experiments on evaporation of
water droplets in superheated steam, found C3 = 0.74 for the Reynolds number
range 64 to 250. Lee and Ryley's correlation is compared with Ranz and Marshal-
l's data and correlation in Figure 6.23. Since the two agree qufte well, and
Lee and Ryley's correlation appears to be valid for droplet Reynolds numbers
up to 40,000, Lee and Ryley's correlation is used here for dispersed droplet
flows.

- 194 -
,
-~

----- CIIlMILAfVI or _ _ !CAy ~ . .,'S P01I - . . ~ 1-.

-. - - - - - - - - - - I'NATOII II IllIG II
... ----.... -- -
- -u
1-1- t-- 1-
.....
~

-.
- .. -
.
-::"':
• ........
---- _1
..,.............
~.--
...............
W"
-- ~I-l-
~.,. .

I~~ ._'r-t1.
. -I--.

-
_ . . . 1_ _ 1
10-

.. - ~'~~n. "I" • all 10 .'1 ..........


lfU\, TO . . U" •
_t _ ';J"'

.. ... It" • ...... 'e I ~V


,"'" 'It . . . .

.!)~.
un
I
I

,!
- ..--
,-I I -
......,
.~~ - ..., ...... !r;:"J'"
~J..O
r
--;-1-

..
"I I
Ii 1_
-l- ~
1

1
• I
• • •••r ,.
.10
..... I - -- - - ..-
""" ..... - ":Oil ....


1
I
.l."l}

Figure 6.23 ~l.eat and :',s.ss Transfer Rates for Spheres. Original
da ta e.nd correlat ion (----) from "[',anz and narshall
(1952). Correlation of Lee an~ ~yley (1968) is
super:uuposed ( - - ) . (PNL Neg. No. 1-916-81).

- 195 -
Summary

The proposed general model of vapor generation following flashing inception


includes the interfacial area density heat transfer coefficients for vaporiza-
tion and the relative velocity in all flow regimes covering the entire void
fraction range. The model for annular-mist flow has not yet been completed,
although a general direction has been specified. Since most of the BNL flashing
data are within a = 0.8, this lack ef a definite model for the annular-mist flow
does not pose any problem in comparing the present model with most of the BNL
data. We shall now proceed to the comparison of the model with the data.

B. Application to BNL Nozzle Data

A computer program was written for the BNL model described above. For a
given run, the input consisted of:

1. The geometry of the BNL nozzle


2. The inlet water condition, i.e., pressure, temperature and
velocity or flow rate
3. The experimental pressure distribution along the length of
the nozzle.

Based on the arguments presented in Section 6.2, and the void fraction data,
it was assumed that the flashing inception began at the throat for all the BNL
runs. Therefore, the bubble radius was determined by use of Equation (6.26). A
bubble number density at the inception point, Nb,o' was then assumed and the
calculation started to march along the axial direction. To alleviate some
numerical problems, Equations (6.41a) and (6.41b) were used instead of Equation
(6.41) during the bubbly flow regime. The axial void fraction profile
calculated by the computer program was then compared with the area-averaged void
fraction data. The free parameter, Nb 0' was then varied until a "best-
fit" between the calculation and the d~ta was found.

A total of 15 runs were simulated by using the BNL model. They can be
grouped under the following four sets:

1. Runs 353, 358 and 362, all with approximately the same inlet water
temperature of 100°C, but with increasing mass flow rates.
2. Runs 145, 133, 137 and 344, all with approximately the same inlet
water temperature of 121°C, but with increasing mass flow rates.
3. Runs 291, 284, 273, 278 and 296, all with approximately the same
inlet water temperature of 149°C, but with increasing mass flow
rates.
4. Runs 268, 304 and 309, all with approximately the same inlet water
temperature of 149°C and the same mass flow rate, but with
decreasing exit pressures.

- 196 -
The results of the calculation with the "optimum" values of bubble number
density at the inception point are shown in Figures 6.24 through 6.38. It can
be noticed the agreement in the bubbly flow regime, i.e., a <0.3, is quite
reasonable. However, there is still room for improvement in the bubbly-slug
regime, i.e., 0.3<a< 0.8. Further examination of the interfacial area and heat
transfer models used in this regime is required.

Table 6.4 provides a summary of the test conditions and the optimum values
of the bubble number density at the flashing inception point. No clear cut re-
lationship between the mass flow rate and the liquid superheating at the flash-
ing inception point can be found. However, when the optimum values for the bub-
ble number density at the inception point, Nb 0' are plotted against the
liquid superheat at the inception point, - ~Tt' , a clear trend can be found as
shown in Figure 6.39. The optimum number of btbbles at the inception point
seems to increase until a liquid superheating of 3°e is reached. Thereafter,
the heat transfer rate to the interface is so large that fewer number of bubbles
are needed to be nucleated as the liquid superheating at the inception point
increases. The phenomenon can be better understood when the interfacial area
density at the inception point, i.e., ai , 0 which is equal to 4TIrb20
,
Nb I 0'
is plotted against the liquid superheating at the inception point. ThlS is
shown in figure 6.40. It is intriguing to note that although the bubble number
density decreased as the liquid superheating dropped below 3°e, the interfacial
area density continued to increase monotonically as the liquid superheating de-
creased even below3°e. This is consistent with the requirement that as the
flow approaches a thermal equilibrium flow, i.e., ~T +0 , the interfacial area
for heat and mass transfer must approach infinity. ft should also be noted that
as ~To + 0, the critical bubble radius rb a -+ro. Therefore, the bubble
~o , +0
number aensity, Nb,o' does not have to approach infinity as ~Tt.o . This
explains the apparent contradiction between Figures 6.39 and 6.40.

Although Figure 6.39 may serve as a guidance for the selection of the op-
timum bubble number density at the inception point once that point is de-
termined, yet it must be realized that the data base-for Figure 6.39 is quite
limited. Only one nozzle size was used, and experiments were conducted only at
low pressures and temperatures. Therefore, the data base must be expanded first
to develop a general correlation for the bubble number density at the inception
point. Furthermore, the model for the bubbly-slug flow regime should be im-
prpved, and the same for the annular-mist flow has to be developed.

- 197 -
RUN353
NO. Of BUBBLES- 1. OO~ 10
10

::0
o

-
0 0 0
0
0 0

/
0

0
o o 0 0
o
o
0 0
- 000 0

rJ

u, i
, I
O.C 100.0 200.0 300.0 100.0 500.0 600.0
ZIMM)
Yigure 6. 2Lc COr.lparison Petween the Void Fraction Data
for Pun 353 and the "Best-Fit" Calculation
Using the Cenera1 Hoc!e1 (R:-;L Neg. No. 1-905-81)

- 198 -
RUN3~8
NO. or BUBBLE~= 1. OO~ 10
10

a ·

r-- -
0
0
0
cr:: 0
0
IV"
CL •
~o
a: I Vo 0


)0
I

0
0
p
0 0 0,
a 0 0
a · - 0
- po.-

a
I
·
0.0 100.0 200.0 300.0 100.0 soo.o 600.0
ZIMM)
Figure 6.25 Cov_parison Between the ·Void F-raction Data
for Run 358 and the "Best-- it" Calculation Using
the ~eneral Hodel (BNL Neg. fo. 1-912-81).

- 199 -
RUN362
NO. or BUBBLES= 1. OO~d 010
o


/" ~
0 0

(0

/
,........
. 0
0

~J

o ·-

0
a 0 p 0 o 00
n
o ~ ~

:.)

100.0 2CO.0 300.0 100,0 500,0 600.0


ZfMMJ
Figure 6.26 Comparison Between the Void Fraction Data
for Run 362 and the "Best-Fit" Calculation
using the General Model (BNL Neg. No. 1-902-8).

- 200 -
RUN145
10
NO. or BUBBLES- 2.00" 10
l";

(n
() - -- -----

( ] - -- ------- .--- r· -- ._----


0
0 0
0 0
(1. 0
, 0
~- or'
r,
" .J C) - . -- --~- . - .. ------.
(J
V---" '.

,. J

- - .-.--_.- ---- /.
{ 00--
'---..... -

- .----

t
0
I
0 0
()
p 0 00
L) (l 0 0 0
.)
r>
( ) - -- -- --- -1" .-------

I J
I
1
. () _0 100.0 200.0 300.0 100.0 ~oo.o 600.0
ZIMMI
~igure 6.27 Comparison Between the Void Fraction Data
for Run 145 and the "Best-Fit" calculation
Using the General Hodel (BNL Neg. NO. 1-901-81).

- 201 -
RUN133
NO. OF BUBBLES= 1.00)( 1010
o

CD
o
.

o
. - -
0
0
0

cr: o 0
I"T'
0....
.....JO

-
cr:
V
f'J
o 0
r:-
0
0
0

0
0 0 0 00
o
o
. o 0 ~ 0 0 ~

.
o,
0.0 100.0 200.0 300.0 100.0 500.0 600.0
ZlMMJ
Figure 6.28 Comparison' Between the Void Fraction Data for
Run 133 and the "Best-Fit" Calculation Using
the General Model (BNL Neg. No. 1-915-81).

- 202 -
RUNJ37
9
NO.OF BUBBLES- 3.00MIO
o

Q.)

o 0

0

~

0
v
0
N
o · 0
0

0
0

0 00 0
o 0 p
0
0 0
p 0
o
o ·

N
oI ·
0.0 100.0 200.0 300.0 400.0 500.0 600.C
Z(MM)
Figure 6.29 Comparison Between ~he Void Fraction Data
for Run 137 and the "Best-Fit" Calculation
Using the General Hodel. (BNL Neg. No. 1-909-81).

- 203 -
RUN344
NO. OF" BUBBL.ES", 5. OO~ 10
10

o ·

0
.- ~-~
o ·
1(0
ro
0

o · J .-

o 0
o · ~ -

oI · ,i I

0.0 100.0 200.0 300.0 100.0 500.0 F:,OO.O


ZIMMJ

Figure 6.30 Comparison Between the Void Fraction Data for


Run 344 and the "Best-Fit" Calculat!1on Using
the General Model. (RNL Neg. No. 1-903-81).

- 204 -
EXP 291
10
NO. or BUB9LES ... 1. or-d 0
o .
I
'0
C ~
o
, --
,
J

,
I

- -- ---

.
1"\1
a
Vo
0

-~
0

0 0
0 "-
0
0

o 0
o -r----
o

0 p po
0 0
" 0 0 0 00

1"\1
oI I
O.r) 10000 20CoO 300.C 10000 500.0 6QOoO
ZlMMi

Figure (,.31 Comparison Between the Void Fraction Data for


;:',un 291 and the "Eest-Fit" Calculation Using
the General Model. (BNL Neg. No. 1-907-81).

- 205 -
[XP 284
10
NO. or BUBBLES = 2.00)( 10
o
I

CD I
o

,
a ·
I
I (. 0
o

r
0 01

o · I
I
0
0
I
o 0 0 0

~

0 0 0

r-.;
I
I I
,..
U.IJ
('

100.0 200.0 300.0 4:;0.0 500.0 b:JG.:


ZIMMI
Figure 6.32 Comparison Between the Void Fraction Data
for Run 284 and the "Best-Fit" Calculation
Using the General Model. (BNL Neg. NO. 1-904-81).

- 206 -
EX?
- 273No.or BUBBLES- S.OO~10'0
o
·

a · --
a · o .J
-
0
0

CI:
I ..
a.... p
-.Jo
CI:

N
a ·
j
a
a · 00 0 I:)

N
aI ·
0.0 100.0 200.0 300.0 ~oo.o 500.0 600.0
Z(MMl

Figure 6.33 Comparison Between the Void Fraction Data for


Run 273 and the "Best-Fit" Calculation Using the
General Model! (ENL Neg. No. 1-908-81).

- 207 -
EXP 278
10
NO. or BUBBLES... 8. OO)E 10

o ·

_0
I- 0

· ~ 0
o

7 0
0 0

I 0
0

N
o

0
o p
o · o 0 p
0
~ ~

0 0 -

N
o
- I
·
0.0 100.0 200.0 300.0 400.0 500.0 600.0
Z(MM)
Figure 6.34 Comparison Bet\'leen the Void Fraction Data for
Run 278 and the "Best-Fit" Calculation Using
the General Model. CBNL Neg. No. 1-906-81).

- 208 -
EXP 296
NO. OF BIJBBLES'" 8. OO)E 1010
CI
·

CD
CI

CI · -~
--- -
""""
a

a:
::r:..,.
CL
.-JCI

If a
a a

a: p

N
CI ·
a
CI
CI · a
00


CI

0.0 100.0 200.0 300.0 100.0 500.0 600.0


Z(MMl

Figure 6.35 Comparison Between the Void Fraction Data for


Run 296 and the "Best-Fit" Calculation Using the
General Model. (BNL Neg. No. 1-9]0-81).

- 209 -
2
G. 4.29 Mg/m
l.n

EXP 268
NO. or BUBBLES= 3. oo~ 10
10

co
o ·

a::
I~
CL •
If 0 00 0

~
~o
a::

N
o · 01
o
o · 0
00
0

N
o
I
·
0.0 100.0 200.0 300.0 100.0 500.0 600.0
Z(MMl

Figure 6.36 Comparison !'etween the Void Fraction Data for


Run 268 and the "Best-Fit" Calculation Using the
General ~!oc~e1. (BNL Neg. No. 1-914-81).

- 210 -
EXP 304
NO. or BUBBLES ..
10
4. L10M 10
a
·

CD
o

0
0 0

o · - - -
~

cr:
I-r
CL
-1
a:
0
• - -

I ~
I
N
o · - /
o
o ·
.
000
p
00 0 p ~ 0
000 p
I
N
oI ·
0.0 luO.O 20a.0 300.0 400.0 500.0 600.0
Z (M~11
~igure 6.37 Comparison Between the Void Fraction Data for
Run 304 and the "Best-Fit" Calculation Using the
General Hodel (BNL Neg. No. 1-913-81).

- 211 -
EXP 309
9
NO.or BUBBLES- 5.0nMIO

co
o
.

V--
I
-t-- / o
0
0

II
0

o 0

o
I 0
c
I
I
("\)

o -~---
I

~
o
- -<>- -.H'T- . o 0 b 0 o
o
0
0 0

II
I
I I
~ ,~
v . 'J iDD.D 200.0 300.0 '100.0 500.0 608.8
ZIMMJ

Figure 6.38 Comparison Between the Void Fraction Data for


Run 309 and the "Best-Fit" calculation Using the
General Hodel. (BNL Neg. No. 1-911-81).

- 212 -
TABLE 6.4 SUMMARY OF TEST CONDITIONS
AND PRESENT MODEL CALCULATIONS*

G p N a.
0 llT9- ,0 b,o rb,o J.,o
Run 0
2 (oC) 3 2 3
No. (kg/s s) (bar) No/rn (\Jm) rn /rn

10
353 18974 0.955 1. 65 1 x 10 20.3 25.5
10
358 25116 0.950 1.80 1 x 10 18.7 23.5
10
362 28314 0.929 2.13 1 x 10 16.3 20.27

0
T. ::::: 121 C
J.n
10
145 15566 1. 742 5.31 2 x 10 3.48 3.043
10
133 18658 1. 634 7.28 1 x 10 2.63 0.865
10
137 24801 1.482 11.82 0.3 x 10 1. 71 0.11
10
344 27935 1.922 2.23 5 x 10 7.88 39.0

0
T. ::::: 149 C
J.n
10
291 13336 4.03 5.0 1 x 10 1.71 0.367
10
284 15061 4.047 5.2 2 x 10 1. 63 0.557
10
273 18048 4.192 3.4 5 x 10 2.46 3.79
10
278 24119 4.257 2.94 8 x 10 2.82 8.04
10
296 27148 4.17 3.7 8 x 10 2.27 4.52
10
2.68 18091 4.057 4.91 3 x 10 1. 73 1.13
10
304 18133 3.997 5.25 4 x 10 1. 64 1. 35
10
309 18217 3.935 6.12 0.5 x 10 1.41 0.125

*The subscript "0" refers to the flashing inception point.

- 213 -
Id l r-------r---.---r-.---.----.-------.--~
LEGEND
o T.
In
= 149°C
6. Tin = 121°C
o T.
In
=100°C

o o
o
.Ii
z
o

109~____~____~~__~~__~______~__~

I 2 3 4 5 7 10 20
6.T (OC)
I ,0

Figure 6.39 The Optimum Bubble Number Density at the


Inception Point vs. the Liquid Superheat
at the Inception Point (BNL Neg. No. 1-1144-81).

- 214 -
50 ~----~--~~~~--~--~-----,-----,

0 Tin = 100°C
t::. T.In = 121°C

0 Tin = 149°C
10

..,
E
.....
C\J
E
0.
c

1.0

0.5

0.1
2 :3 4 5 7 10 20
"T
'-' I ,0
(OC)

Figure 6.40 The Interfacial Area Density at the Inception


Point VB. the Liquid Superheat at the
Inception Point (BNLNeg. No. 1~1142-8l).

- 215 -
7 COMPARISON OF TRAC-P1A PREDICTIONS WITH EXPERIMENTAL DATA

After a computer program was developed for calculating the vapor generation
rates from the experimental data (Section 5) and analytical modeling was done
for flashing inception and void d~velopment (Section 6), typical experimental
results were compared with TRAC-P1A predictions and modeling.

TRAC-P1A (Transient Reactor Analysis Code) developed at Los Alamos


Scientific Laboratory (1979), is an advanced best estimate computer program for
the loss of coolant accident analysis in a pressurized water reactor. The
vessel is modeled by a two fluid, three-dimensional formulation. The balance
equations comprise mixture and vapor mass balance. mixture and vapor energy
balance, and liquid and vapor momentum conservation. These balance equations
require description of relative velocity (V r ) vapor generation rate (rv )'
interfacial drag, interfacial heat and mass transfer, interfacial velocities,
interfacial area, and wall heat transfer for liquid and vapor and corresponding
wetted areas.

The loop components in TRAC are formulated by a one-dimensional drift-flux


model with five conservation equations: mixture mass, vapor mass, mixture momen-
tum, vapor thermal energy, and mixture thermal energy. The additional constitu-
tive relations, which are usually flow regime dependent, specify the relative
velocity between the phases, the volumetric phase change rates •. the vapor-liquid
interfacial heat transfer rates, the wall shear, and the wall heat transfer
rates (Rohatgi and Saha, 1980).

In the computer code calculations presented below, the one-dimensional


steady-state loop component formulation was used, and the converging-diverging
nozzle was modeled with three components:

a. An inlet BREAK where the experimental values of the inlet pressure,


temperature, and void fraction were specified.

b. A PIPE component where the nozzle geometry, number of cells, cell axial
locations, and initial (guess) distributions of pressure, temperature,
void fraction and mixture velocity were specified.

c. An exit BREAK where the. experimental value of the exit pressure was
specified.

The code was run to reach a steady state for the above specified initial and
boundary conditions; and the output consisted of the steady-state flow proper-
ties in each cell, i.e., pressure, void fraction, saturation. liquid and vapor
temperatures, liquid, vapor, and mixture densities, the mixture velocities, the
slip, and the friction factor. The code also provided the steady~state mass
flow rate corresponding to the specified boundary conditions.

7.1 Comparison of TRAC-P1A Predictions With Experimental Data Consisting of


Pressure and Diametrical Averaged Centerline Void Fractions

First the TRAC-P1A code was run for the first series of experiments pre-
sented in Section 5.2, in which only axial distributions of centerline dia-
metrical averaged void fractions were measured. These results are similar to

!!receding page blank] - 217 -


those presented by Reocreux (1974) who also measured only the diametrical
averaged centerline void profiles.

Figure 7.1 presents the comparison of TRAG predictions with the experi-
mental results for Run 79-792, with an inlet mass flux of 2270 kg/m 2 s. The
inlet and exit conditions for the TRAG calculations were Pin = 124 kPa, Tin
= 99.4°G, ain = 0.0, and Pexit = 114 kPa. The steady-state inlet mass flux
predicted by the code for the specified conditions was 1854 kg/m 2 s. The inlet
mass flux and consequently the pressure drop from the inlet to the throat are
lower than the measured values, and these flow conditions correspond to the on-
set of flashing in the test section. The pressure distribution predicted by
TRAG (Figure 7.la) shows a trend similar to the experimental observations but
underpredicts the pressure drop at various locations because of the 18 % under-
prediction of the value of the mass flux. The slight deviation of the pressure
distributions in the diverging section of the nozzle corresponds to the onset
prediction and the vapor generation calculations of the code. The squares de-
pict calculations made with the number of cells in the program doubled from 47
to 94. This change resulted in only slight variations in the pressure distri-
bution.

In comparing the void fraction profiles calculated by TRAG with those


measured during the experiment (Figure 7.1b), note that the void fractions
calculated by TRAG are area averaged whereas the experimental ones for this run
are diametrically averaged. The differences between the two should, however, be
neglibible in this case.

In TRAG-PIA, the mass exchange rate is modeled by

T ) + h iA. (T -T )
s v ~ v s
(7.1)

.where hQi and hvi are the heat transfer coefficients from the liquid and the
vapor to the interface; Ai is the interfacial area density; Ti; Tv, and Ts
are the liquid, vapor, and saturation temperatures; and hfg is the latent heat
of vaporization. According to TRAG-PIA modeling, vapor generation starts as
soon as the liquid reaches the saturation temperature. Thus, no superheating of
the liquid is allowed at the onset of flashing, and the vapor production rate be-
comes positive as the liquid temperature exceeds the local saturation tempera-
ture. Downstream from the inception point, superheat or departure from equili-
brium is allowed, and the degree of liquid superheat depends on the vapor gen-
eration rate and the boundary conditions. From the local steady-state proper-
ties tabulated by TRAG, the vapor generation rates were calculated by computing
the local actual quality and using the steady-state vapor continuity equation,
dx
rv = G dz (7.2)

The local derivative of the quality was evaluated from a second-order poly-
nominal fit to three adjacent calculated values. Figure 7.lc presents the re-
sults forr v for Run 79. The calculation indicates a slight vaporization
starting at Tap 24 followed by condensation starting at Tap 28. A similar trend
is seen in the void fraction profiles in Figure 7.1b.

- 218 -
5 /0 15 20 25 30

,-.. 130

«
Q.;
~

...... 110
UJ
c::: 100 Psa! (Till)
::::I
en
en 0
UJ 90 00
c::: o 0
a...
BO o EXP. 79
0 - TRAC (4B CELLS)
70 c TRAC (96 CELLS) a.

z 1.0
0
I- O.B
u
«
c:::
u... 0.6 o EXP.79
Q - TRAC
0.4
0
>-
0.2
0 0
0 0
0 0 0
b.

UJ
I-
<
a:: 50
z
0 u::
UN 79 TRAC
I-
<C
c::: "'"'~ C
0
= 1.0 and 1.1 (Indistinguishable)
UJ .........
z 0
UJ I!)
(..!;I ~
'-'
e:::
0
a...
<
>- -50
c.

0 50 60

Z ( eM )

Figure 7.1a Comparison of TRAC Predictions with BNL Experimental Results.


Run No. 79. G. = 2270 kg/m 2s , Pin = 124 kPa, Tin = 99.4oC ..
(BNL Neg. No. lnS-282_80).

- 219 -
A method was also developed (see Section 5.6) for calculating the cross-
section averaged net vapor generation rate per unit volume rv from the mea-
sured pressure and void fraction distributions along the test section (Zimmer et
al.,1979). Since these calculations involve the derivatives of the measured
quantities ex and p, they amplify small errors (scatter) in the ex and p data. The
smoothing and interpolation routines of the IMSLX computer program packages were
used to obtain least-square cubic spline curve fits continuous up to the second
derivative to the pressure and void distribution data. The value of r v (see
Section 5.6) was dominated by the first term in Equation (5.13), which re-
presented the variation of ex with z. The calculated values of r v showed a
very weak dependence on the drift velocity (Vgj) but were affected by the
value of the distribution parameter at large void fractions (see Figure 7.5c,
below, for Co = 1.2 and ex> 0.8). Figure 7.1c depicts the comparison between
the vapor -generation rates predicted by TRAC-PIA and those calculated from the
experimental data (7.1a and 7.lb).

Figure 7.2 presents similar comparisons for Run 78. The measured inlet mass
flux is 2610 kg/m 2 s. The measured inlet and exit conditions of Pin = 138
kPa, Tin = 99.3°C, Pexit = 110 kPa, were used for the TRAC calculations.
Under the flow conditions present the onset of flashing occurs between Taps 24
and 25, close to the nozzle throat. The experimental pressure distributions
(Figure 7.2a) show a condensation front, which is also observed in the void
fraction profiles (Figure 7.2b). The steady-state inlet mass flux calculated by
TRAC for these experimental conditions is 2288 kg/m 2 s, which is 12 % lower
than the experimentally measured one. Figure 7.2a shows that the pressure drop
in the converging section is again underpredicted by TRAC because of the lower
inlet mass flux calculated, but the trend in the diverging section agrees
qualitatively with the experimental observations. Figure 7.2b shows that the
void fraction di stributions again agree quali tatively and indicate a full liquid
condition in the converging section, followed by vaporization and a condensation
zone in the diverging section. Figure 7.2c, a plot of the local vapor genera-
tion rates as calculated by TRAC, shows an almost constant vapor generation rate
of - 29 kg/m 3 s in the beginning of the diverging section, followed by a
strong condensation region where r v falls to --55 kg/m 3 s and then goes back
to zero when the all liquid state is reached.

Similar results are seen in Figure 7.3 for Run 77, with a measured inlet
mass flux of 3060 kg/m 2 s. The inlet and exit conditions used for the TRAC
calculations are Pin = 157 kPa, Tin = 99.4°C, and Pexit = 92 kPa. For a
47-cell calculation, the steady state inlet mass flux predicted by TRAC was 2667
kg/m 2s, around 13 % lower than the eXperimental value. The pressure profile
calculated by TRAC (solid line in Figure 7.3a) shows an unexpected variation
near the exit, and so does the void fraction profile (Figure 7.3b); this results
in a large variation in the vapor generation rates close to the nozzle exit
(Figure 7.3c). The dashed lines in Figures 7.3a and 7.3b are plots of the pres-
sure and void fraction profiles calculated for a homogeneous mixture expanding
isentropically, wi th the phases in local dynamic, ki nema tic, and thermal
equilibrium. These were calculated on the BNL computer by using a steam
property calculating subroutine developed by Kroeger (1979) based on the ASTM
formulation. The predictions were calculated for the appropriate inlet mass
flux, Gin = 2370 kg/m 2 s, to reach saturation conditions at the nozzle throat
(Tap 25). Thes~ calculations also underpredict the mass flux, but they
qualitatively predict the variation trends.

- 220 -
IS 20 2S 30 3S 40 45 49
5 10

- <
0..
:..::
o

100 o 00 000
LIJ o 00 0
c::: o
=:I
CI'l
CI'l
LIJ
c::: o EXPo 78
0.. 50 -TRAC

a.

1.0 0 EXP 78
z
0 -TRAC
0.8
I-
u
<
c::: 0.6
u..
Q 0.4
0
>- 0.2

0 0 0 o 0 0 b.
0

_ RUN 78 TRAC
LIJ
I-
<
c::: 50
z ,,-.
0 CI'l

I-
<
c:::
UJ
"'".........
'
::!!:
0
Z
lJJ C,!)
I.!:> ::.::
'-'
c:::
0
a..
< -50 c.
>-
0
Z( eM )

Figure 7.2 Comparison of TRAC Predictions with BNL Experimental Results. Run No.
78. Gin = 2610 kg/m 2 s, Pin = 138 kPa, Tin = 99.3°C.
(BNL Neg. No. 5-280-80).

- 221 -
10 15 20 25

-<
a..
:.=: P.OI (Tin)
......
w 00
00
0000000000000000
0::
:::l
In
In o
W
0:: 50 EXP.77
o
a..
- - TRAC
- - HOM. EO.
a.
QL---+---~--~--~---+----,

z 1.0 RUN 77
0 -TRAC
I- 0.8
u
<
0:: 0.6
u...
Q
0.4
0
:>
0.2 0

0 o 0
o 0 0 0 0 0 00 dt b.
0

400
w
I-
~
z
0
......
In
300 •.•. RUN 77
-TRAC
I- r-r'I
<
0::
:E:
w ....... 200
z
w
l!)

0::
0
Q.
<C
-C)
:.=:

100
:>
c.

AxiAL DlSTANCE 1 L( CM )

Figure 7.3 Comparison of TRAC Predictions and Homogeneous Equilibrium


Calculations with BNL Experimental Results. Run No. 77.
Gin = 3060 kg/m 2 s, Pin = 157 kPa, Tin = 99.4oC.
(BNL Neg. No. 5-281-80).

- 222 -
Figure 7.4 presents the results for Run 76, with an experimental inlet mass
flux of 6040 kg/m 2 s. The inlet and exit conditions for the TRAC calculations
are Pin = 395 kPa, Tin = 99.3°C, Pexit = 85 kPa. The steady-state inlet
mass flux predicted by the TRAC calculations is 6283 kg/m 2 s, which is only 4%
above the experimentally measured value. Since the onset of flashing occurred
very close to the saturation condition for this run, the predictions seem to
agree much better with the experimental results for the pressure distribution
(Figure 7.4a) and the inlet mass flux. The void fraction profiles exhibit some
discrepancy, but a fair comparison would require area averaged void fractions
and the measurements here correspond to diametrical averaged void fractions.
The local vapor generation rates (Figure 7.4c) show a sudden increase at the on-
set and slight fluctuations around a value of 93 kg/m 2 s in the diverging sec-
tion. The homogeneous equi":brium calculations for Run 76 for an inlet mass
flux of 5443 kg/m 2 s are plotted in Figures 7.4a and 7.4b as dashed lines;
these pressure distributions and void fraction profiles are quite close to the
TRAC predictions.

Figure 7.5 presents a comparison between TRAC predictions and experimental


results for Run 80-801, wi th an inlet mass flux of 4360 kg/m 2 s. The inlet
and exit conditions used for the TRAC calculations are Pin = 585 kPa, Tin =
148.3°C, ain = 0, and Pexit = 436 kPa, The 47-cel1 TRAC predictions (solid
lines in Figure 7.5a and 7.5b) show a strange variation in the pressure profile
in the diverging section, though the code seemed to have converged to a steady-
state solution. To investigate whether this variation was due to an input para-
meter, the number of cells was changed from 47 to 94 (results presented as
squares), and the initial pressure in the nozzle, and the exit void fraction
were varied. None of these parameters significantly affected the pressure
variation observed in the diverging section, and the unexpected wavy appearance
was not eliminated. The inlet mass flux predicted by the TRAC calculations was
3930kg/m 2 s, 10 % lower than the experimental value. As seen in Figure 7.5b,
the change from 47 to 94 cells affected the void fraction profiles slightly.
Figure 7.5c, which depicts the local vapor generation rates calculated from the
local properties evaluated by TRAC, shows an increase at the onset which reaches
a maximum, and slight condensation near the exit of the venturi. Also plotted
(dots) are the rv calculations by the computer program from the experimental
pressure and void fraction distributions presented in Section 5.4 for three
values of the distribution parameter, Co = 1.0, 1.1, 1.2. The experimental
results do not show a sudden increase in rv at the point of flashing onset,
but rather, a continuous increase followed by a sudden drop. The latter trend
may be due to the fact that the curve fitted through the experimental points is
a continuous one and does not show a sudden change in the slope of the p and a
curves at the flashing onset. Since the derivatives of these two properties are
involved in the rv calculations, the results are strongly dependent on the
curve fits selected.

7.2 Comparison of TRAC-PIA Predictions With Experimental, Data Consisting of


Pressure and Area Averaged Void Profiles

Typical experimental data obtained during the last series of tests, in-
cluding pressure measurements as well as detailed. transverse voi~ profiles at 27
axial locations which allowed the calculation of area averaged void fractions
were also compared with TRAC-PIA predictions. The results are presented below,
including those for runs with subcoo1ed water inlet at nominal temperatures of
100°, 121°, and 149°C.

- 223 -
5 10 15 20 25 :30 45 49

400 - - - ,-• ....-,-,-,-, i -,,-,-,-, i -,-,,-, i_,-,-,_.1"'-'-'-'-' i -,-,,-, i -,'-'-I i I

o RUN 761
ex: --TRAC
a... 300
:><:: - - EO. CALC.
.......
UJ
0::
:J 200
IJ)
IJ)
UJ
0::
a... L ______ _
100 00000000 0000000000000
o
a.

o RUN 76
5 1.0 - TRAC
- - HOM. EO.
I-
~ 0.8
0::
u.. 0.6
c
00.4
>-
0.2
o 0

b.

UJ
I-
<C - TRAC 76
cr::
Z .......
0 IJ) 100 1..2
I- /'l"\
<
0::
~
LIJ '- .. 1.1
Z
LIJ (!)
l!:) :><:: C 1.0
......... 50 . o
Co = 1.2
jif .

L
0:::
0 = 1.1
~,JC~h
Q.. C
<
>- ,W!:c
r
~~ ~"
!!7
, C =l.a- .~ -' c.
0 o
0 10 20 30 40 50
PXIAL DISTANCE~ Z( CM )

Figure 7.4 Comparison of TRAC Predictions and Homogeneous Equilibrium


Calculations with BNL Experimental Results. Run No. 76.
Gin = 6040 kg/m2 s, Pin = 395 kPa, Tin = 99.30c.
CBNL Neg. No. 5-279-80).

- 224 -
5 10 15 20 25 30 35 40 45

111-1-1-1-111-1-11-1 I-I-11-1-11-1-1-1-11-1-1-1-111-11-1-11-1-1-1-1-1111111
600

~ 500
CJ
Q
"""
a.

- - RUN 80 TRAC
400
o (DOUBLE CELLS) 0
TRAC
o RUN 80 EXP

a.
300~---r---~~----+-----+----+------H

- - TRAC
1.0 o TRAC (DOUBLE CELL)
o EXP. 801
"

0.5

a 00 0 b.
Op-~~o~+o~o~~o~____~~______~____ -+______-H
o

RUN 80 -801

- TRAC
3
10 ••• CALC. FROM EXP. DATA

c.
o 30 40
Z (em)

Figure 7.5 Comparison of TRAC pr_edict i ',rs w t;h _BNI., FX2e~ imenta.1
Results. Run No. 80 801. G1n · dbO "..g/In s, llin ..
585 kPa, Tin = 148.3 0 C. (BNL NRg. ~o. ]-654-79)

- 225 -
150 TA P No. 1.0
z
o
Q...
o
..x: r-
(.)
W «
0::
=:>
~IOO o .6p -RUN 116
-- 0::
LL

o
w • a-RUN 118 o
0::
Q... TRAC - PI A >

50 1---=.-.---..._+-.........__.___-..:::.'--IIr*'..... • •
'-----...!!.!...-.IL-! 0

10 20 30 40
AXIAL DISTANCE (em)
Figure 7.6 Comparison of TRAC-P1A Predictions with BNL Experimental Data for
Pressure Distributions and Area Averaged Void Profiles. Pin = 143
kPa, Tin = 99.9 0 C, Gin = 2.28 ~~/m2 s, Pet = 127 kPa, and
Tct ,; lOO.4 0 C, (m exp = 4.6 kg/s, m = 4.8 kg/s)
TRAC
(BNt Neg. No. 2-200~80)

- 226 -
5 15 25 35 45
TAP No. 1.0
--a..
0 z
-
~

W
150 0
I-
C>
0:: <!
0::

---
:::)
1.L
(f) 0 6p RUN 122
(f) • 0
w • a RUN 124 •• • ••
oooeo
-
0
0:: o ~o 0 0 0 00 0 0 0 0
a.. 100 TRAC- PIA >


•••••• o
10 20 30 40 50 60
AXIAL DISTANCE (em)
Figure 7.7 Comparison of TRAC-PIA Predictions with BNL Experimental Data for
Pressure Distributions and Area Averaged Void Profiles. Pin = 171
kPa,_T in = ~OO.:~G, ~il} = 3.01 ~/m2 s,_Pct = 133 kPa, and
Tct - 1UO.7 C (m - 0.1 kg/s, m TRAC - 6.0 kg/s).
(BNL Neg. No. 2-~g-80)

- 227 -
7.2.1 Comparison with Runs Performed at a Nominal Inlet Temperature of 100°C

Table 7.1 lists the nozzle inlet pressures and temperatures and the nozzle
exit pressures used as boundary conditions in the TRAC-P1A calculations, and
also the comparisons of the experimentally measured mass flow rates with the
TRAC-P1A steady-state predictions.

Figure 7.6 shows experimentally obtained axial pressure distributions and


area averaged void fraction profiles (measured with the five-beam gamma de-
nsitometer system) compared with TRAC-PlA predictions. The pressure and void
fraction profiles predicted by the code calculation agree qualitatively with the
data and indicate the evaporation and the condensation zones, but both profiles
show rates of change much higher than those for the experimentally measured
data.

In Figure 7.7, the experimentally obtained pressure distributions and area


averaged void fractions (measured with the five-beam system) for Runs 122 and
124 are compared with TRAC-PlA predictions. These runs were performed at a noz-
zle inlet pressure of 171 kPa, an inlet temperature of 100.7°C and an inlet mass
flux of 3010 kg/m 2 s. In the experimental data, the slight pressure increase
in the diverging section of the nozzle is reflected as a constant void fraction
region. The TRAC-P1A pressure prediction shows a strange decrease and increase
in the diverging section and close to the exit. The predicted void fractions
are higher than the measured ones, and the predicted flashing inception point,
where vapor starts to be generated, is earlier than the measured one. This is
expected, since the code does not have a built-in nucleation delay. The
predicted mass flow rate is in excellent agreement with the measured value.

Figure 7.8 presents a comparison of experimentally obtained pressures and


area averaged void fractions (measured with the five-beam system) with TRAC-PlA
predictions. Under the high mass flow rate and high inlet subcooling
coknditions of Run 127, where the pressure undershoot below saturation is small
compared with the total pressure drop in the converging section of the nozzle,
the TRAC-PlA predictions agree very well with the experimental data. A slight
pressure decrease is still observed in the diverging section, followed by a fast
recovery close to the nozzle exit in order to satisfy the exit pressure boundary
condition. .

Similar results are presented in Figure 7.9, where experimental data on


pressure distributions and area averaged void fraction profiles are compared
with TRAC-PlA predictions. In this run the detailed experimental void fraction
data and the experimental area averaged void fraction distributions were ob-
tained by means of the high activity single-beam gamma densitometer.

7.2.2 Comparison with Runs Performed at a Nominal Inlet Temperature of 121°C

Table 7.2 presents the inlet pressures, temperatures, and comparisons of ex-
perimentally measured mass flow rates with TRAC predictions for the 121°C in-
let temperature runs.

- 228 -
TABLE 7.1

COMPARISON OF 100°C INLET TEMPERATURE RUNS


WITH TRAC-P1A PREDICTIONS

Exp. Mass TRAC Mass


Flow Rate Flow Rate %
Run No. Pin(kPa) Tin(C) Pexit(kPa) kg/s kg/s Deviation

N
N 116-118 143.0 99.9 121.5 4.6 4.79 4
\0

122-124 171.0 100.2 109.2 6.1 5.97 -2

127-128 248.0 100.0 101.0 9.1 8.58 -6

353- 247.7 100.0 103.6 9.16 8.54 -6.7

356 246.9 99.9 99.8 9.15


250 _LI"1

TAP No. I.O~


-E
-a..
o w
u
~200 z
w «
~
0::: (j)
::::) o ~p RUN 128
(j) o
~ 150 • a RUN 127 ~
0::: TRAC PIA «
0.... X
o «
000000 0 00000000
100 o

• • • • • ••• o
50~~~~~~~~~~~~~~~~~~~~~~

10 20 30 40 50 60
AXIAL DISTANCE (em)
Figure 7.8 Comparison of TRAC-PIA Predictions with BNL Experimental Data for
Pressure Distributions and Area A~eraged Void Profiles. Pin = 248
~a, T = 1000C, G = 4.49 Mg/m s, P t = 127 kPa, and
f i
m
T~ = 50.soc (mex~ ~ 9.1 kg/s, TRAC =c~.6 kg/s).
(BNL Neg. No. 2-197-80)

- 230 -
3 7 II 15 19 23 27 31 35 39 43 47
no.
.0
Q.
oX
TRAC PIA (94cells)
W
0:: o RUN 353
~200
(f)
• RUN 356
w
0::
Q.

100
I. 0 J-I-->-.l....L.l-..LJ....I....J....L1...J,...L..l....l....W'--'--'-w....L..L..L.LJ....L.L..L...I....l....I-..l...1..~..LJ....I....L.!-1--i....W....l...WL........L...L.W....J....I....L.w....y

0.8
z
o TRAC PIA (94 cells)
I- 0.6' o RUN 354
u
<l:
0::: • RUN 355 Iii 0
LL.. 0.4
o
~0.2

Figure 7.9 Comparison of TRAC'-PIA Predictions with BNL Experimental Data for
Pressure Distributions and Area Averaged Void Profiles. Pin = 247
kPa. Tin ~ 100°C, Gin = 4520 kg/Iri 2 s, Pct == 113 kPa, and
Tet = 100°C. (BNL Neg. No. 9-692-80)

- 231 -
TABLE 7.2

COMPARISON OF 121°C INLET TEMPERATL~E RlJNS


WITH TRAC-PIA PREDICTIONS

Exp. Mass TRAC Mass


Flow Rate Flow Rate %
Run No. Pin(kPa) Tin(C) Pexit(kPa) kg/s kg/s Deviation

141- 239.7 + 4.9 121.3 + 0.1 225.3 5.98 4.84 (47 ) -18.9
cells
144 242.5 + 4.8 121. 3 + 0.1 227.7 5.96

145- 306.2 + 0.7 121.2 + 0.1 207.7 7.46 7.04 (94 ) -5.6 *
cells
148 304.1 + 0.6 121. 2 + 0.1 206.0 7.46 7.30 (94 ~ -2.1 **
cells

133- 350.3 + 0.7 121.3 + 0.1 205.5 8.93 8.37 (94 ) -6.4
cells
136 347.9 + 0.6 121.2 + 0.1 202.6 8.95

137- 462.8 + 1.5 121.8 + 0.1 201.3 11.87 11.70 (60 ) -1.6 ***
cells
140 465.2 + 2.2 121.5 + 0.1 202.6 11.93

322- 342.2 + 1.0 121.2 + 0.1 199.9 8.94 8.16 ('94 ) -8.8
cells
325 341. 2 + 0.8 120.9 + 0.1 197.3 8.95

339- 321.0 121.3 250.6 8.97 7.48(94 ) -16.6


cells
342 319.4 121.2 252.4 8.97

313- 341. 2 121.1 194.8 8.99 8.15(94 ) -9.3


cells
316 340.4 120.9 192.7 8.99

318- 323.7 121.4 167.8 8.98 7.56(94 ) -J.5.9


cells
321 320.3 120.8 165.9 8.99

* Surface Tension as calculated by TRAC-P1A.


** Surface Tension as measured by Volyak (1950).
*** 60 cells, fine noding at the throat, annular flow friction factor.

- 232 -
In Figure 7.10 (top), experimental pressure distributions and area averaged
void profiles (measured with the five-beam system) are compared with TRAC-PlA
predictions. Since the TRAC inception criterion for vapor generation is met at
the location where the local liquid temperature corresponding to the local
pressure is equal to the saturation temperature, voids in TRAC tend to be
generated upstream from the throat. In contrast, the experimental area averaged
void profile indicates the existence of single-phase flow in the whole test
section. In Figure 7.10 (bottom) the vapor generation rates (rv = G dx/dz)
calculated from the TRAC-PlA listed cell properties and flow parameters are
compared with those calculated from the experimental data (Co = 1.0) by the
method described in Section 5.6 and the cubic spline fit. Increasing Co to
1.2 does not change the broken line in Figure 7.10 for the vapor generation
rates.

Figure 7.11 presents similar comparisons for Runs 145 to 148 performed with
the five-beam gamma densitometer. The TRAC-PlA predictions (solid lines) in the
diverging section of the nozzle again show pressure recovery close to the
throat, a pressure decrease downstream, and fast recovery close to the exit of
the nozzle. The degree of the oscillation in absolute kPa seems to be more than
that observed in runs with an inlet temperature of 100°C. The area averaged
void profiles, and the vapor generation rates presented in Figure 7.12, reflect
this by the sudden increase and decrease close to the exit of the nozzle. In
Figure 7.12, results are also shown from TRAC-PlA predictions (dashed lines)
with the regular surface tension calculations replaced by the experimental
values measured by Volyak (1950) while the boundary conditions and number of
cells were identical. The change in the surface tension seems to improve the
prediction of the mass flow rate (see Table 7.2) as well as the pressure
profiles close to the nozzle throat.

Similar results are presented in Figures 7.13 and 7.14 for Runs 133 to 136
with the five-beam gamma densitometer. The vapor generation rates derived from
the experimental data recorded are also presented in Figure 7.14 (dashed lines)
for three values of the distribution parameter, Co = 1.0, 1.1, and 1.2.

Figure 7.15 presents similar comparisons between experimental data (with the
five-beam system) and TRAC-PlA predictions for a high mass flow rate (11.9 kg/s)
experiment with an inlet temperature of 121°C. The TRAC-PlA calculations for
pressure and area averaged void profiles (top figure) did not converge to a
steady-state solution with 48 or 96 cells and with a homogeneous flow friction
factor option. The results presented were obtained with 60 cells. with the dis-
tance between cells decreased close to the throat, and with use of the annular
flow friction factor option. The data and the TRAC-PlA calculations agree ex-
ceedingly well for these conditions. In the bottom figure, the experimentally
derived vapor generation rates are compared with the values obtained by hand
calculations based on the TRAC-PlA listed properties and with the values (open
circles) printed out directly by the code itself. The vapor generation rates
calculated 'from the cell properties are almost identical to those printed out
directly by the code calculations. The effect of the distribution parameter
Co is still observed in Figure 7.15, especially for the high void fraction re-
gion in the diverging section of the nozzle.

- 233 -
~eee~~IO~ 20 30 40
eee
o eee z
a.. P s (Tin) e~ o
.::t:.
- - -.. . e t-
w 200 - TRAC (141) e U
<[
a::
::)
o6P RUN 141 e a::
1L.
en
en
• .1P RUN 144 e e a
w
a::
a a RUN 142 ~ e o
a.. • a RUN 143 >
150

TRAC (141)
en 50
I"") EXPERIMENTALLY
E DERIVED (141)
.......
0'1
.::t:.

2
0 7- 0

- - . - . - .

Col.O

a 10 20 30 40 50 60
AXIAL DISTANCE (em)
Figure 7.10 Comparison of Experimentally Measured Pressure Distributions and
Area Averaged Void Profiles, as Well as Vapor Generation Rates
Calculated from the Experimental Data with TRAC-PIA Predictions.
Pin = 241 kPa, Tin 121.3 0 C, Gin = 2.97 Mg/m 2 s, Pet = 237 kPa,
Tet = 121.7 o C. CBNL Neg. No. 5-276-80)

- 234 -
I
300
.-
0
0..
.:s:
TAP No . +.0 z
o
"'-'"
• RUN 148 I-
w o RUN 146 u
a::: o ~p - RUN 145
i <r:
:::J 250 ~
0::
l.L
(f)
(f)
-a RUN 147
W :-:7. TRAC - PI A
o
0::: o
0.. >

a
- o

10 20 30 40
AXIAL DISTANCE (em)
Figure 7.11 Comparisor.. of TRAC-PIA Predictions with BNL Exper:imental Data
for Pressure Distributions and Area Averaged Void Profiles. P in=
305 ~a, Tina = ~21.2~C, Gin = 3.: Mg/m: s, Pet = 234 kPa, and
Tet - 121.7 C (m exp - 7.5 kg/s, mTRAC - 7.0 kg/e).
(BN"L }leg. No. 2-198-80)

- 235 -
400 TRAC (145)
EXPERIMENTALLY
DERIVED .",.
/Co = 1.2
. ."..- ........................ ..
.:;:. , . ."".-_.-. ......
....... ....... .. ......
I .~.~.
""".. . . . . .-'. .......
/~......... ..... .. .,,"'
I - - - - - - - - - - - - - J...j Co = 1.0 ----:-:.:::,-_:: ~-----

AXIAL DISTANCE (em)


Figure 7.12 Comparison of Vapor Generation Rates Calculated from Experimental
Data with TRAC-PIA Predictions for Run 145 Presented in Figure 7.11.
(BNL Neg. No. 5-277 -8 0)

- 236 -
35 45
o 6p RUN 133 1.0
-0
0....
o a RUN 134 z
0
.::It:. • RUN 136
-- 300
W
a::::
--- • RUN 135
I-
u
<{
TRAC PIA iii iii a::::
::::> iii i LL
(/)
(/)
0
W i -
a:::: 0
0.... 200 -
Ps (T i n=12 I .2 C) >

o
o 10 20 30 40 50 60
AXIAL DISTANCE (em)
Figure 7.13 Comparison of TRAC-P1A Predictions with BNL Experimental Data
for Pressure Distribution and Area Avera~ed Void Profiles.
Pin = 34~ Tin =o121:2°C,_G in = 4430 ~g/m :' Pet = 233 kPa,
and Tct - 121.7 C (m exp - 8.9 kg/s, mTRAC - 8.4 kg/s).
(BNL Neg. No. 2-199-80)

- 237 -
200
/C o= 1.2 I
.
TRAC (133) ,.".-.- ............... ....... i
i I I
.i 'I
150 ! ' . ...._...... ..l
. ..... ., .I
EXPERI MENTA LLY i
-CI)
rt')' 100 DERIVED
;' r_.
.j / ...... ,
'.
".
.......... C =I I. .
...../ 0 ./ .I
",". ..........._. _.'" I
I

E
C = 1.0
/' ......
...... -._-_ . . .
.
"'-
O
--
(J)
~ 50
>
~
0

-50

o 10 20 30 40 50 60
AXIAL DISTANCE (em)
Figure 7.14 Comparison of Vapor Generation Rates Calculated from
Experimental Data with TRAC-P1A Predictions for Run 133
Presented in Figure 7.l3. (BNL Neg. No. 5-278-80)

- 238 -
500 0.9
-~ z
-olI:

W
0::: 300 o 6pRUN 137
0
I-
u
<[
::> e6p RUN 140 P. (T. ) 0:::
en u...
en
W
c a RUN 138 5 ~ 0
0::: • a RUN 139 0
a.. >
100
0

+200
-} }CALCULATED
o TRAC (137)
PRINTOUT (137)

+ 100
_.- EXPERIMENTALLY
u; DERIVED
rt)'

E
........
c-
oli:

-100
10 20 30 40 50 60
AXIAL DISTANCE (em)

Figure 7.15 Comparison of Experimentally Measured Pressure Distributions


and Area Ave~aged Void Profiles as Well as Vapor Generation
Rates Calculated from the Experimental Data with TRAC-PlA
Predictions. Pin = 464 kPa, Tin = l2l.6 o C, Gin = 5.9 Mg/m 2 s,
Pct= 234 kPa, Tct = l21.6 o C. .
(BNL Neg. No. 5-275-80)

- 239 -
Comparisons of TRAC-PIA calculations with vapor generation rates derived
from experimental measurements lead to the following conclusions: For low mass
flux (Runs 141 to 142) peak rv predicted by TRAC was more than one order of
magnitude higher than the "experimental" value (Figure 7.10). The TRAC predic-
tion of rv for intermediate mass flux (Runs 145 to 146) was about twice as
high as the value calculated from experimental data (Figure 7.12); and for high
mass flux (Runs 133 to 134 and 140 to 139), the TRAC calculations of rv
roughly agreed with the experimentally derived values assuming Co = 1.1. The
discrepancy at low mass flow rates may be partly attributable to TRAC's in-
ability to account for the metastable liquid conditions at the flashing incep-
tion point.

Similar comparisons are presented in Figures 7.16 to 7.19 for runs with an
inlet temperature of 121°C in which the detailed transverse and area averaged
void fraction data were obtained with the high activity single-source gamma de-
nsitometer. Comparison of the results presented in these figures with the data
in Table 7.2 shows that a slight variation in the inlet pressure (degree of sub-
cooling)strongly affects the deviation of the TRAC-PlA mass flow rate
predictions (16% to 8%) from the experimentally measured values.

7.2.3 Comparison With Runs Performed at a Nominal Inlet Temperature of 149°C

Table 7.3 presents the nozzle inlet pressures and temperature and com-
parisons of the experimentally measured mass flow rates with TRAC-PIA predic-
tions for runs done at an inlet temperature of 149°C. The detailed transverse
and area averaged void fraction data for all these runs were obtained with the
high activity single-beam gamma densitometer.

Figure 7.20 shows the single-phase full calibration of axial pressure dis-
tributions and area averaged void profiles for Runs 264 to 266 at an inlet tem-
perature of 149°C. Since the saturation pressure is reached somewhere in the
converging section of the nozzle, the TRAC-PlA predictions generate some vapor
which is not observed in the experimental data.

Figure 7.21 presents comparisons of experimental data (pressure dis-


trihutions and area averaged void profiles) with TRAC-PIA predictions for Runs
291 to 295. Running the code with'47 and 94 cells resulted in small differences
in the predictions. The plots show a vaporization region downstream from the
throat followed by a condensation zone and pressure recovery.

Comparisons of the results of higher inlet mass flux experiments (Runs 284
to 288, Figure 7.22; Runs 273 to 277, Figures 7.23; and Runs 279 to 283, Figure
7.24) with the TRAC-PlA predictions show that the anomalous pressure variation
predicted by the code, reported earlier, becomes more prominent as the inlet
mass flux increases. ~ Increasing the number of cells from47 to 94 decreases the
fluctuations but not the general variation predicted in the diverging section.
There may be two reasons for this: (1) The code starts generating vapor at the
satuc~tton pressure, and it has no built-in pressure undershoot that would allow
the metastable (nonequilibrium) liquid condition down to the throat of the
nozzle. (2) If the code has a very high vapor generation rate downstream from
the throat (Taps 25 to 33), that will cause the pressure to increase downstream
from the throat, even though the area averaged void fraction distributions

- 240 -
3 7 II 15 19 23 27 31 35 39 43 47

d:..::t:. 300 --08


• iii
i
W -i
0:::
~
(f)
(f)
-TRAG PIA(94 cells) •
o RUN 339
-
w • RUN 342
0:::
a... 200
i
i

1.0

z 0.8
0
I-
u 0.6 - TRAC PIA (94 cells)
<I
0::: Cl RUN 340
I.J....
0 0.4 • RUN341
-
0
> 0.2 i ijj i
I!!I I
Ii

10 20 30 40 50 60
AXIAL DISTANCE (cm)
Figure 7.16 Comparison of TRAC-P1A Predictions with BNl Experimental Data for
Pressure Distributions and Area Averaged Void Profiles for
Runs 339-342. Pin = 320 kPa, Tin 121oC, Gin = 4430 kg/m 2 s,
Pet = 240 kPa, Tct = 121°C. (BNL Neg. No. 9-961-80)

- 241 -
-0300
0...
~

w
cc 8
::l

.. .....
if) -TRAC PIA (94 cells)
if)
o RUN 325
~ 200 ~
0... • RUN 322

1.0

z 0.8

.
0
f-
- TRAC PIA (94 cells)
U 0.6 iii • I!I
o RUN 323 ~
<{
@II
0::
LL 0.4
• RUN 324 iii
0
- •
0 iii
> 0.2 iii

0 10 20 30 40 50 60
AXIAL DISTANCE (em)
Figure 7.17 Comparison of TRAG-plA Predictions with BNL Experimental Data for
Pressure Distributions and Area Averaged Void Profiles for
Runs 322-325. Pin = 341 kPa. Tin 121oC, Gin = 4410 kg/m2 s.
Pct = 200 kPa. Tct = 121 o C. (BNL Neg. No. 9-694-80)

- 242 -
Flow /(U;~;~

no. 9
((((((((/
:r;-p ~, II : : : ~ ~ ~ ~ I I I I I I I r;;;; ~ ;;;;; I I I I I I rI
13 17 21 25 29 33
tI(////LU'

37
r;;;;;::: ~ l
41 45
(

~300
-'"
W
0:
:::>
(/)
(/) - TRAC PIA (94 cells)
• i '
W
o RUN 313
a.. 200
0:
• RUN 316 i

I .0 ~~LLLJLL.Ll...LLl...L.1....l..L.Ll..Ll..LLLLli..l.J...LU...l..Ll.l....Ll..l...LLLl.LL.LU...Lll..l..l.1.l..Ll.l..j

0.8
z
o
I- 0.6 -TRAC PIA (94eells)
u o RUN 314
« o
0:::
lJ... 0.4 • RUN 315
.'
o
o II
§; 0.2

o 10 20 30 40 50 60
AXIAL DISTANCE (em)
Figure 7.1S Comparison of TRAC-P1A Predictions with BNL Experimental Data for
Pressure Distributions and Area Averaged Void Profiles for Runs
313-316. Pin = 341 kPa, Tin = l2l o C, Gin = 4430 kg/m 2 s, Pct =
lS7kPa, Tct = 1l9 0 C. (BNL Neg. No. 9-693-80)

- 243 -
Tap
no.

a: 300
.::t:.

w
n::
~
(f)
(f)
w -TRAC PIA (94 cells) ~
n:: 200 o RUN 318 •
a....
• RUN 321

1.0

0.8
z
0 iii
~ 0.6 TRAC PIA (94 eel ~c:) II fill
U
<{ [J
RUN 319 I
I
n::
u... 0.4 • RUN 320
0
0 [J

> 0.2 •
!!I~
[J

0 iii

40 50 60
AXIAL DISTANCE (em)
Figure 7.19 Comparison of TRAC-PIA Predictions with BNL Experimental Data for
Pressure Distributions and Area Averaged Void Profiles f~r
Runs 318-321. Pin 322 kPa, Ti~ = 121°C, Gin = 4430 kg/m s,
Pet = 156 kPa, Tct = 113.S o C. ~BNL Neg. No. 9-960-80)

- 244 -
TABLE 7.3

COMPARISON OF 149°C INLET TEMPERATURE RUNS


WITH TRAC-P1A PREDICTIONS

Exp. Mass TRAC Mass


Flow Rate Flow Rate %
Run No. Pin(kPa) Tin (C) Pexit(kPa) kg/s kg/s Deviation

264- 514.5+2.5 149.3+0.3 500.4 5.80+0.02 5.46 94 -5.9


cells
266 515.8+9.4 11.8.8+0.1 501.9 5.81+0.03

291- 504.7+2.9 148.9+0.1 472.8 6.43+0.01 5.59 47 -13 .1


cells
295 499.7+3.6 148.8+0.1 469.9 6.43+0.02 5.68 94 -11.7
N
cells
Ln
"'" 284- 530.1+1. ] 149.2+0.1 545.7 7.29+0.02 6.45 47 -11.3
cells
288 530.8+1.1 149.1+0.1 459.1 7.25+0.02 6.66 94 -8.4
cells
273- 573.5+4.5 148.7+0.3 442.4 8.71+0.02 7.77 47 -10.8
cells
277 572.1+2.3 148.7+0.1 441.2 8.71+0.02 7.68 94 -11.8
cells
279- 689.4+2.2 148.8+0.1 433.4 11.64+0.04 10.73 94 -7.7
cells
283 686.7+1. 9 148.8+0.1 431. 7 11. 61+0. 02
309- 555.9+2.2 149.1+0.2 396.5 8.79+0.03 7.43 47 -15.5
cells
311 7.37 94 -16.2
cells
FLOW
----;;.
TAP 3 7 II 15 19 23 27 31 35 39 43 47
NO.
520
0
0...
~
500
W
0::: 0
::::>
(f)
(f)
Ps ("Tin)
~
w
0::: TRAC PIA
0... 450 -(94 CELLS)
• RUN 266 •
0 RUN 264 0

1.0

z 0.8
0
I-
(.) 0.6
<I:
0:::
LL
0.4 - TRAC PIA (94 CELLS)
0
-
0 • RUN 265
> 0.2

10 20 30 40 50 60
AXIAL DISTANCE (em)
Figure 7.20 Comparison of TRAC-piA Predictions .... it!':. BN~ EXt:-:.~::!.illental Data for
Pressure Distributions and Area Averaged Void Profiles for 2
Runs 264-266. Pin = 515 kPa, Ti = 149 0 C, Gi~ = 2860 kg/m s,
Pet = 478 kPa, Tct = 1490C._(BNL~eg. No. 9-641-80)

- 246 -
FLOW~~~~~~~~~~~~~~~~~
TAPNO.~h<~~~~~~~~~~~~~~~~~~~~
500~~~

0.'-:-..,
0
0....
.::.:::.
o. "
o. \
o· \
W
cr
::::J
(f)
Ps (Ti n) ----;;. 0.\
.
o \

o \
\

(f) o \
w - TRAC (94 CELLS) • \J
cr o·
0.... ---TRAC(47 CELLS) 0

• RUN 291
400 o RUN 295
•o
1.0

z 0.8
0 - TRAC (94 CELLS)
f-
0 0.6 --- TRAC (47 CELLS)
<{
0:: o RUN 292
LL
0
0.4 • RUN 293
-
0
> 0.2
.., .., i i Ii
.&0

0 ~--~~~~~~~~~.ii~~ii

10 20 30 40
AX 1AL DI STANCE (em)
Figure 7.21 Comparison of TRAC-P1A Predictions with BNL Experimental Data for
Pressure Distributions and Area Averaged Void Profiles for
Runs 291-295. P. = 502 kPa, Tin = 148.90 C, G . . = 3170 kg/m 2 s,
Pct = 452 kPa, T~~ = 148.9 0 C. (BNL Neg. No. 9-6~8-80)

- 247 ,...
FLOW
~

TAP
NO.
500
a
0...
.x.

W
0::
::J
(j)
(j)
W -TRAC(94 CELLS)
0::
0... --- TRAC (47 CELLS)
o RUN 284
410 • RUN 288
1.0

z 0.8
0 - TRAC (94 CELLS)
I-
0 0.6 ---TRAC (47 CELLS)
<I:
0:: o RUN 285 o 0
LL
0.4 • RUN 286
0
- ••
0
> 0.2

10 20 30 40 50 60
AXIAL DISTANCE (em)
Figure 7.22 Comparison of TRAC-PIA Predictions with BNL Experimental Data
for Pressure Distributions and Area Averaged Void Profiles f~r
Runs 284-288. Pin = 530 kPa, Tin = 149.2oC, Gin = 3580 kg/m s,
Pet = 456 kPa, Tct = 149.2oC. (BNL Neg. No. 9-642-80)

- 248 -
FLOW I
TA P NO ~ l-r-r-r-TT".,...-,-..,..--,-,.--r--rr--r-r"T"""T"'..-r-...--r-r-r-..--.--r-r-r-r-r-r-r-r-..-,r-TITTT-" , : i , ! '

600 35 39 43 47~

o -TRAC (94 CELLS)


0... --- TRAC(47 CELLS)
~

o RUN 273
~500 • RUN 277
::J
<f)
<f)
w
a::
0...

400~
I• 0 I! i ' iii , I i I I I I I I I i I ; 1 I I I I , I I I ;.lLLU j I ! I I I i IUll_Ll.L1-.J'--L.I.-'-Wl....LJ.-'-'-j

~
<.)
0.8

0.6
t
~
-TRAC (94 CELLS)
---TRAC (47 CELLS) .j
C2 ~ o RUN 274 /
I
LL 0.4 t-- • RUN 275
/
Ii
1
/
/

~ ~
/

/Ii 1
> 0.2 --l
. I

o ________________
~ ~~~~M
l
~

II I I I II I II II I I I I I i LI Ii I I I I \I I I Ii! II L.J..ll1 \ \ i \ ! \ I \ "' I \ \ i ,.LLLLlJ


,10 20 30 40 50 60
AXIAL DISTANCE (em)
Figure 7.23 Comparison of TRAC PIA Predictions with BNL Experimental Data
for Pressure Distributions and Area Averaged Void Profiles
for Runs 273-277. Pin = 573 kPa, Tin = l48.7 o C. Gin = 4290 ~
Pet = 452 kPa, Tct = 148.8 o C. CBNL Neg. No. 9-637-80)

- 249 -
FLOW
--;.700
TAP
NO.

0
0...
..:s:.
600
W
0:::
::J - TRAC
CJ)
CJ) (94 CELLS)
W 500 o RUN 279
0:::
0... • RUN 283
Ps (Tj n ) ----';00 i

1.0

z 0.8

.
0
~
<.) 0.6 -TRAC
<J: ~
0::: o RUN 280 II
LL
ii
0.4 • RUN 281
0
-
0
> 0.2

10 30 40 50 60
AXIAL DISTANCE (em)
Figure 7.24 Comparison of TRAC-P1A Predictions with BNL Experimental Data
for Pressure Distributions and Area Averaged Void Profiles for
Runs 279-283. Pin = 688 kPa, Tin = 148.8 0 C, Gi = 5730 kg/m 2 s,
Pet = 452 kPa, Tct = 148.8 0 (BNL Neg. No. 9-64B-80)

- 250 -
are close to.the experimental values. Since the exit pressure is set as a
boundary condition, the code predictions must satisfy the exit pressure; thus,
the calculations will proceed to determine a pressure distribution which ful-
fills the equations of the code with the set boundary condition. The situation
causing this anomalous pressure distribution needs to be investigated and
corrected.

The observations made above are very prominent in Figure 7.25, which com-
pares pressure and area averaged void distributions with the TRAC-PlA predic-
tions for Runs 309 to 311. These runs were done with cooling water sprayed into
the condensing tank to bring the exit as well as the inlet pressure down for a
constant inlet temperature of l49.l DC and an inlet mass flux of 4330 kg/m 2 s.
Run 310 provide typical experimental data, presented in Section 5.5 (Figure
5.60B), shoWing a difference between the centerline diametrical averaged void
fraction and the ~rea averaged void fraction at the same location downstream
from the throat in the diverging section of the nozzle.

It is an important point that, under some transverse void distributions at a


given axial location, the centerline diametrical averaged void fractions are not
representative of the area averaged void fraction values at the same axial loca-
tion. Thus, under specific void distributions, any comparison of the experimen-
tally measured centerline diametrical averaged void fractions (such as those
presented in Section 7.1 and by Reocreux (1974)), with the area averaged void
fraction prediction by one-dimensional models in code predictions can be unfair
and may lead to inaccurate conclusions.

The comparisons of the experimental results with the TRAC-PlA predictions


show that correct prediction of the flashing inception conditions is essen-
tial for accurate predictions of the mass flux and related flow properties
(pressure and void fraction profiles). Since in the TRAC predictions no super-
heating of the liquid is allowed at the onset of flashing, the computer predic-
tions agree well with the experimental results when the inception occurs close
to the saturation pressure or when the pressure undershoot at inception is small
compared with the total pressure drop from the nozzle inlet to the throat. When
the inception superheats are . larger and there is considerable nonequilibrium at
the inception point, large differences are observed between the experimental
results and the TRAC-PlA predictions. As seen in Figure 6.2, significant in-·
ception superheats are predicted at high decompression rates typical of a LOCA.
Even at very low rates the superheats may ap"proach lODC. Thus, for a fair com-
parison of the interfacial mass transfer modeling of computer codes after flash-
ing inception, a well established inception criterion is needed.

- 251 -
FLO W 1L.LLL..Li.'.Li.LL.LJ..i..L..L.I...L.!.LLJ...i.Lt..1..LLLL..'-.LLL.i..L.I..l.1.1.Li..L.ULL1-LlLJ-Li..1.J..LLLi...ULJ..i..LLJ..LLLL.LLLJ.-i.'.L.L..I-LL~
~600hnll""Tlrn~"TI~~"~~~~~~~~~~
TAP 3 7 II 15 19 23 27 31 35 39 43 47
NO.

-TRAC (94 CELLS)


o
0....
.::£
--- TRAC (47 CELLS)
--500 o RUN 309
w
0:::
::J ;:.
if) Ps ("Tin)
if)
W
0::: o
0....
400

0.8 /
z /
o -TRAC (94 CE LLS) I
I

~ iii
G 0.6 --- TRAC (47 CELLS) /
/
I

~

0
iii
.<1: /
0::: 0 RUN 310 /
iii
LL 0.4 •RUN 311
o
o
> 0.2

10 20 30 40 50 60
AX IAL DISTANCE (em)
Figure 7.25 Comparison of TRAC-P1A Predictions with BNL Experimental Data
for Pressure Distributions and Area Averaged Void Profiles f~r
Runs 309-311. Pin = 556 kPa, Tin = 149.1 o C, Gin = 4330 kg/m s,
Pet = 378 kPa, Tct = 142°C. (BNL Neg. No. 9-630~0)

- 252 -
8 SUMMARY AND CONCLUSIONS

In order to determine the vapor generation rates of water under nonequi-


librium conditions, steady flashing flow experiments were conducted in a spe-
cially designed and constructed flow loop with a converging-diverging nozzle
test section. The measurements included static pressure distributions along the
nozzle, photographic records of the flow downstream from the test section, dia-
metrical averaged centerline void fraction distributions along the nozzle, and
detailed transverse (radial) distributions of the chordal averaged void
fractions at 27 axial locations along the test section. The detailed transverse
profiles were used in the calculation of the area averaged void fraction for
each axial location. Since no information on the phase velocities was recorded
during the present experiments, the calculation of the vapor generation rates
from available experimental data involved the assumption of a slip model between
the two phases. The effects of the loop parameters --inlet temperature, inlet
mass flux, and exit pressure --on the flashing phenomenon were investigated, and
the effects of flashing on the measured quantities were recorded under single-
phase subcooled and with low void two-phase nozzle inlet conditions.

Models for flashing inception and void development have been proposed and
applied to the BNL and other experiments. The Alamgir-Lienhard pressure under-
shoot correlation for static depressurization was extended to flowing systems
(pipes and nozzles) to predict the flashing inception pressure with subcooled
inlet conditions. The ability to determine flashing inception was found to be
the key to proper calculation of the critical mass flux through a nozzle with
single-phase inlet conditions. The void growth models agreed reasonably well
with the experiments in the respective ranges of applicability.

The advanced computer code, TRAC-PIA predictions were also compared with the
present experimental data on axial pressure and area averaged void fraction dis-
tributions. Since in the TRAC-PIA code predictions no superheating of the
liquid is allowed at the onset of flashing, the code failed to predict the non-
equilibrium inception point, which might have caused the disagreement between
the calculated mass flow rates and the experimentally measured ones.

The main results and conclusions can be summarized as follows:

a. The pressure distributions and area averaged void profiles


recorded during the experiments show that, for subcooled
inlet conditions, the flashing inception occurs very close
to the nozzle throat.

b. The variation of the exit pressure reveals a choked flow pattern


in the converging section of the nozzle.

c. Depending on the exit pressure, the pressure distributions


in the diverging section either were constant or showed a
constant pressure region "followed by a sudden pressure
recovery (condensation zone) which was also verified by
the area averaged void fraction profiles.

- 253 -
d. The pressure undershoots, below the saturation pressure at
inception, observed at the nozzle throat, increased with
the nozzle inlet temperature. No clear trend was observed
with the nozzle inlet mass flux.

e. Experiments done at inlet temperatures of 149°C and 121°C, with a con-


stant inlet mass flux and decreasing nozzle exit pressures, showed
no significant difference in the area averaged void profiles measured.
However, the detailed transverse (radial) profiles exhibited a drastic
change at axial locations close to the nozzle exit. For lower exit
pressures, the transverse void distributions changed from a nonsym-
metrical configuration to asymmetrical one with a low void core
surrounded by a high void region.

f. With the symmetrical transverse void profile the axial variation of


the area averaged void fractions was found to be different from the ax-
ial variation of the centerline diametrical averaged void fractions.

g. During one experiment in which a large amount of MgO particulates was


present in the facility (due to the failure of a heater), no pressure
undershoot was observed at the throat, and flashing inception occurred
under equilibrium conditions. Comparison of the results with those of
a run with similar inlet conditions but no particulates, in which
a pressure undershoot was observed at inception, did not show
significant differences in the axial pressure and area averaged void
fraction distributions. During the experiment with particulates
present, the transverse (radial) void fraction profiles again showed a
symmetrical configuration at axial locations close to the nozzle exit.

h. In experiments done under low void fraction two-phase inlet condi-


tions, at an inlet temperature of l2loC, a constant inlet mass flux,
and decreasing nozzle inlet pressures, the pressure distributions de-
creased continuously in the converging as well as in the diverging
sections of the nozzle. The area averaged void at the nozzle exit
increased with the decreasing exit pressure and eached values close to
1.0.

i. In the same experiments, although the local pressure decreased con-


tinuously in the converging section (increasing the local superheat),
no significant increases in the area averaged void fractions were ob-
served upstream from pressure Tap 23.

j. The vapor generation rates calculated frqm the experimental data on


axial pressure and area averaged void fraction distributions
required the adoption of a drift flux model. The calculated values of
r v I s were dominated by the variation of area averaged void fraction
with axial distance. The vapor generation rates calculated were not
affected by the value of the drift velocity, but they were very
sensitive to the value of the distribution parameter, especially in the
high void fraction regions (a > 0.6).

- 254 -
k. The pressure undershoot correlation of Alamgir and Lienhard for static
depressurization was extended to flowing systems (pipes and nozzles).
For pipe flows the turbulence effects on the local pressure have to be
considered in order to explain the flashing inception data in the
literature. In nozzle flows, the inception was observed to occur at
the throat, and the inception pressure undershoot can be predicted by
the Alamgir-Lienhard correlation within its limit of applicability.

1. A model of vapor generation following flashing inception has been


proposed including the interfacial area density, heat transfer coef-
ficients for vaporization and the relative velocity in flow regimes cov-
ering the void fraction range O<a<O.8.

m. Comparison of the model predictions witn the experimental area averaged


void fraction distributions showed a reasonable agreement in the bubbly
flow regime. However there is still room for improvement in the
bubbly-slug regime.

n. The optimum bubble number density (for best-fit) at inception, Nb,o'


increased with the liquid superheat at inception, 6T ~ ,until a
liquid superheating of 3C was reached, and Nb,o decre~~ed as the
liquid superheat increased beyond 3C.

o. The interfacial area density calculated from the critical bubble radius
and the optimum bubble number density at inception increased mono-
tonically with decreasing liquid superheat.

p. On the basis that, in nozzle flows with subcooled inlet conditions,


inception occurred at the throat and the pressure undershoot can be
calculated from the Alamgir-Lienhard correlation, a method was proposed
for calculating the critical mass flow rates through nozzles. The
calculated values agreed with the experimental data compared.

q. Representative experimental results were compared with TRAC-PIA steady-


state predictions. The code underpredicts the mass flux when
the inception of flashing occurs under highly nonequilibrium conditions,
and the pressure undershoot below the saturation pressure at incep-
tion, is a significant portion of the total pressure drop from the
nozzle inlet to the throat. The discrepancy diminishes as the onset of
flashing in the nozzle occurs closer to the saturation conditions.

r. The area averaged void fraction distributions calculated from TRAC-PIA


show trends similar to those of the experimentally measured area
averaged profiles, and agree well when onset of flashing is close to
saturation conditions.

- 255 -
s. For some cases, the axial pressure distributions in the diverging
section of the nozzle predicted by the TRAC-PIA code
calculations display a variation not observed experimentally.
This discrepancy increases with higher nozzle inlet temperatures.

t. The vapor generation rates calculated from the experimental data are
also compared with the TRAC-PIA predictions. For a fair comparison,
the phase velocities have to be measured as well as the axial pressure
distributions and axial area averaged void fraction profiles.

- 256 -
9 ACKNOWLEDGEMENTS

The authors would like to thank all the members of the Data Systems and
Operations Group, namely Messrs. James H. Klein; John R. Klages, Carl E.
Schwarz, John J. Barry, and Donald Becker, for their valuable help during the
construction and operation of the experimental facility, and to Mr. Thomas P.
Feierabend for his essential contributions in the electronic instrumentation.
We are indebted to Mr. William L. Leonhardt for the design and construction of
the loop and for his assistance during its shakedown.

Special appreciation is due to Ms. Marisa Canner for typing the manuscript
and for her patience and efforts during preparation of t~ report.

Our thanks to William Massenger of the welding shop for his imagination and
dedication in putting the pieces together and to Dennis Rhodes and Edward McGil-
ley of the electrical shop who ingeniously sorted out the maze of electrical
connections and made the pumps pump. We would also like to express our grati-
tude to Eugene C. Mohlmann for the expeditious manner in which he handled our
rush orders, and especially to Herb Banks for his invaluable help with our
purchases and negotiations with the various vendors. Dr. Hobart W. Kraner's
suggestions, help, and guidance on they -source and y-densitometer development
were also most welcome at critical moments and are well appreciated.

The authors would also like to present their gratitude to Ms. Yako Sanborn
for her help arid patience in running the long and time consuming computing ef-
forts.

We would also like to thank Dr. Ralph J. Cerbone for his advice and con-
structive inputs, and George A. Greene for his help during the final preparation
of this report.

- 257 -
10 REFERENCES

1. ABUAF, N. , and WU, B. J. C., "Comparison of Alamgir-Lienhard Inception


Correlation with Marviken Data," Technical Memo to NRC (March 1980).

2. ALAMGIR, Md., and LIENHARD, H. H., Personal communication (1979); "Cor-


relation of Pressure Undershoot During Hot-Water Depressurization," J. Heat
Transfer, in press.

3. ALEXSANDROV, Y. A., VORONOV, G. S., GORBUNKOW, V. M., DELONE, N. B., and


NECHAYEV, Yu. I., Bubble Chambers, Indiana University Press, Bloomington,
pp. 72-76 (1967).

4. ARDRON, K. H., and ACKERMAN, M. C., "Studies of the Critical Flow of Sub-
cooled Water in a Pipe," Paper presented at the CSNI Specialist Meeting on
Transient Two-Phase Flow, Paris (1978).

5. BAILEY, J. F., "Metastable Flow of Saturated Water," Trans. ASME~,


1109-1116 (1951).

6. BENJ AMIN, M. W., and MILLER, J. C., "The Flow of Saturated Water Through
Throttling Orifices," Trans. ASME~, 419-429 (1941).

7. BOURE, J ., FRITTE, A., GIOT, M. and REOCREUX, M., "Choking Flows and
Propagation of Small Disturbances," Paper No. F-l, European Two-Phase Flow
Group Meeting, Brussels (June 1973).

8. BRAY, K. N. C., J. Fluid Hech~, 1-32 (1959).

9. BROWN, R. A., "Flashing Expansion of Water Through a Converging-Diverging


Nozzle," UCRL-6665-T (October 1961).

10. BURNELL, J. G., "Flow of Boiling Water through Nozzles, Orifices and
Pipes," Engineering 164, 572 (1947).

11. CALDERBANK, P. H., and LOCHIEL, A. C., "Mass Transfer Coefficients,


Velocities and Shapes of Carbon Dioxide Bubbles in Free Rise Through Dis-
tilled Water," Chern. Eng. Sci. ~, 485 (1964).

12. CHAO, B. T., "Transient Heat and Mass Transfer to a Translating Droplet,"
J. Heat Transfer~, 273 (1969).

13. CHEN, J. C., "Effect of Turbulent Flow on Incipient Boiling Superheat," in


Proc. 1st Meeting of Technology Group on Liquid Metal Thermal Sci.,
Brookhaven National Laboratory (May 1969).

14. COURANT, R., and FRIEDRICKS, K. 0., Supersonic Flow and Shock Waves, Wiley,
New York (1948).

15. DAVENPORT, W. G., RICHARDSON, F. D., and BRADSHAW, A. V., "Spherical Cap
Bubbles in Low Density Liquids," Chern. Eng. Sci. ~, 1221 (1967).

- 258 -
REFERENCES (Cont'd.)

16. DAILY, J. W., and JOHNSON, V. E., JR., "Turbulence and Boundary Layer Ef-
fects on Cavitation Inception from Gas Nuclei," Trans. ASHE~, 1695-1706
(1956).

17. DOEBELIN, E. 0., Heasurement Systems: Application and Design, HcGraw Hill,
New York (1966).

18. DUKLER, A. E., and TAITEL, Y., "Flow Regime Transitions for Vertical Upward
Gas-Liquid Flow: A Preliminary Approach Through Physical Hodeling,"
NUREG-0162 (1977).

19. EDWARDS, A. R., and O'BRIEN, T. P., "Studies of Phenomena Connected with
Depressurization of Water Reactors," J. Brit. Nucl. Ener. Soc. 2., 125-135
(April 1970).

20. EDWARDS, A. R., "Conduction Controlled Flashing of a Fluid, and the Predic-
tion of Critical Flow Rates in a One-Dimensional System," U.K. Atomic En-
ergy Authority Report, AHSB (S) R 147 (1968).

21. FAUSKE, H. K., "The Discharge of Saturated Water Through Tubes," Chem. Eng.
Prog. Symp. Ser. ~, 210-216 (1965).

22. FAUSKE, H. K., and HENRY, R. E., "The Two-Phase Critical Flow of One-
Component Hixtures in Nozzle, Orifices and Short Tubes," J. Heat Transfer
464, (1971).

23. FAUSKF., H., "Contribution to the Theory of Two-Phase, One-Component Criti-


cal Flow," ANL-6633, (1962).

24. FORSTER, H. K., and ZUBER, N., "Growth of a Vapor Bubble in a Superheated
Liquid," J. Applied Physics~, 474-478 (1954).

25. FORSTER, H. K., and ZUBER, N.,. "Dynamics of Vapor Bubbles and Boiling Heat
Transfer," AIChEJ. 1,.,531-535 (1955).

26. GYARHATHY, G., in Two Phase Steam Flow in Turbines and Separators, pp.
49-50, M. J. Hoore and C. H. Sieverding, Editors McGraw-Hill, New York
(1976).

27. HABERHAN, W. L., and MORTON, R. K., Trans. Am. Soc. Civ. Eng. 121, 227
(1956).

28. HAYES, W. D., in Fundamentals of Gasdynamics, H. W. Emmons, Editor,


Princeton Univ. Press, Princeton, N. J. (1960).

29. HENDRICKS, R. C., SIHONEAU, R. J., and BURROWS, R. F., "Two-Phase Choked
Flow of Subcoo1ed Oxygen and Nitrogen," NASA TN D-8169 (1976).

30. HENRY, R. E., "Two-Phase Critical Discharge of Initially Saturated or Sub-


cooled Liquid," Nuclear Sci. Eng.~, 336-342 (1970).

31. HENRY, R. E., FAUSKE, H. K., and McCOHAS, S. T., "Two-Phase Critical Flow
at Low Qualities, Part II: Analysis," Nuclear Sci. Eng. ~, 92-98 (1970)

- 259 -
REFERENCES (Cont'd.)

32. HENRY, R. E. and FAUSKE, H. K., "The Two-Phase Critical Flow of One-
Component Mixtures in Nozzles, Orifices and Short Tubes," J. Heat Transfer,
Trans. ASME, Series C, 2l, 179-187 (1971).

33. HEWITT, H. C., and PARKER, J. D., "Bubble Growth and Collapse in Liquid
Nitrogen," J. Heat Transfer, Trans. ASME, Series C, 90, 22-26 (1968).

34. HIRT, C. W., ROMERO, N. C., TORREY, M. D., TRAVIS, J. R., "SOLA-DF: A Solu-
tion Algorithm for Nonequilibrium Two-Phase Flow," NUREG-CR-0690 (1979).

35. HSU, Y. Y., "Review of Critical Flow, Propagation of Pressure Pulse and
Sonic Velocity," NASA TND-6814 (1972).

36. HSU, Y Y., "On the Size Range of Active Nucleation Cavities on a Heating
Surface," J. Heat Transfer 84c (3), 207-216 (1962).

37. ISHII, M., and GROLMES, M. A., "Inception Criteria for Droplet Entrainment
in Two-Phase Concurrent Film Flow," AIChE J. ~, 308-318 (1975).

38. ISBIN, H. S., MOY, J. E., and DaCRUZ, A. J. R., "Two-Phase Steam-Water
Critical Flow," AIChE J. 361, (1957).

39. JAMES, R., "Steam-Water Critical Flow Through Pipes," Proc. Inst. Mech.
Engrs. 176, 741-748 (1962).

40. JONES, O. C., JR., "Flashing Inception in Flowing Liquids," BNL-NUREG-26134


(1979).

41. JONES, o. C., JR., and SARA, P., "Nonequilibrium Aspects of Water Reactor
Safety," in Symposium on the Thermal and Hydraulic Aspects of Nuclear Re-
actor Safety Vol. 1: Light Water Reactors, O. C. Jones, Jr. and S. G. Bank-
off Editors, ASME (1977).

42. JONES, O. C., JR. and ZUBER, N., "Evaporation in Variable Pressure Fields,"
Paper 76-CSME-CSChE-123, National Heat Transfer Conference (1976), and
Trans. ASME, J. Heat'Transfer 100C, 453-459 (1978).

43. KROEGER, P. G. and ZUBER, N., "An Analysis of the Effects of Various Para-
meters on the Average Void Fractions in Subcooled Boiling," Int. J. Heat
Mass Transfer ~, 211-233 (1968).

44. KROEGER, P. G., Personal communication (1979).

45. LACKME, C., "Propagation d'un Front de Vaporisation Dans un Tube Plein
d'Eau Chaude Brusquement Detendue," CEA-R-4986 (1979).

46. LACKME, C., "Autovaporisation Dans une Conduite d'un Liquide Sature ou
Sous-Refroidi a l'Entree," CEA-R-4957 (1979).

47. LACKME, C., "Limitation de l'Autovaporisation d'un Liquide Sursature Liee a


l'Explusion Sonique des Phases Produites," CEA-R-4942 (1978).

- 260 -
REFERENCES (Cont'd.)

48. LAUFER, J., "Investigation of Turbulent Flow in a Two-Dimensional Channel,"


NACA TN 2123 (July 1950).

49. LEE, K., and RYLEY, D. J., "The Evaporation of Water Droplets in Super-
heated Steam," J. Heat Transfer ~, 445 (1968).

50. LEVY, S., "Predictions of Two-Phase Critical Flow Rate," J. Heat Transfer,
Trans. ASME ,Series C, 87, 134-142 (1965).

51. LIENHARD, J. H., ALAMGIR, Md., and TRELA, M., "Early Response of Hot Water
to Sudden Release from High Pressure," Trans. ASME, J. Heat Transfer, 100C,
473-479 (1978).

52. LINEHAN, J. H., "The Interaction of Two-Dimensional, Stratified, Turbulent


Air-Water and Steam:-Water Flows," ANL-7444 (1968).

53. MALNES,D., "Critical Two-Phase Flow Based on Non-Equilibrium Effects," in


ASME Symp Vol Non-Equilibrium Two-Phase Flows, R. T. Lahey and G.- B. Wal-
lis, Editors, pp. 11-17, (1975).

54. MOODY, F. J., "Maximum Flow Rate of a Single-Component Two-Phase Mixture,"


J. Heat Transfer, Trans. ASME, Series C, ~ 134~142 (1965).

55. NIINO, M., "Study of Single Bubble Generation and Growth by Laser Beam,"
PhD Thesis, University of Tohoku, Japan (1975).

56. OSWATITSCH, K., "Kondensationserscheinungen in Uberscha1ldusen," Z. angew.


Math. Mech. ~, 1 (1942).

57. PALEEV, 1. 1., and FILIPPOVICH, B. S., "Phenomena of Liquid Transfer in


Two-Phase Dispersed Annular Flow," Int. J. Heat Mass Transfer ~, 1089,
(1966).

58. PLESSET, M. S., and ZWICK, S.A., "The Growth of Vapor Bubbles in Super-
heated Liquids," J. Aplied Physics £, 493-500 (1954).

59. POWELL, A. W., "Flow of Subcoo1ed Water through Nozzles," .Westinghouse


Electric Corporation, WAPD-PT-(V)-90 (April 1961).

60. RANZ, W. E., and MARSHALL, W. R., "Evaporation from Drops," Chem. Eng.
Prog. 48, 141-146; 173-180 (1952).

61. "Reactor Physics Constants," ANL-5800 (1963).

62. REOCREUX, M., "Contribution a l'Etude des Debits Critiques en Ecou1ement


Diphasique Eau-Vaeur," PhD Thesis, Universite Scientifique. et Medicale de
Grenoble, France (1974).

63. REOCREUX, M. and SEYNHAEVE, J. M., "Ecou1ements Diphasiques Eau-Vapeur; Es-


sais Comparatifs de Debits Critiques," Acta Technical Be1gica.!.2., 3-4;
115-137 (1974).

- 261 -
REFERENCES (Cont'd.)

64. ROHATGI, U. S., and RESHOTKO, E., "Nonequilibrium Two-Dimensional Flow in


Variable Area Channels," In Nonequilibrium Two Phase Flow, ASME (1975a).

65. ROHATGI, U. S., and RESHOTKO, E., "Non-Equilibrium Two-Phase Flow Through
Quasi One-Dimensional Channels," FIAS/TR-75-115 (1975b).

66. ROHATGI, U. S., and SARA, P., "Constitutive Relations in TRAC-P1A,"


NUREG/CR-1651, BNL-NUREG-51258 (1980).

67. SARA, P., "Reactor Safety Programs: Quarterly Progress Report for the Peri-
od April-June 1977," BNL-NUREG-50683, p. 145 (1977).

68. SARA, P., "Review of Two-Phase Steam-Water Critical Flow Models with Em-
phasis on Thermal Nonequilibrium," NUREG/CR-0417, BNL-NUREG-50907 (1978).

69. SARA, P., and ZUBER, N., "Point of Net Vapor Generation and Vapor Void
Fraction in Subcooled Boiling," in Heat Transfer 1974, Proc 5th Int. Heat
Transfer Conf., Vol. 4, 175-179 (1974).

70. SCHROCK, V.' E., STARKMAN, E. S" and BROWN, R. A., "Flashing Flow of Initi-
ally Subcooled Water in Convergent-Divergent Nozzles, J. Heat Transfer.
Trans. ASHE~, 647 (1977).

71. SEYNHAEVE, J. M., "Critical Flow Through Orifices," Paper presented at the
European Two-Phase Flow Group Meeting, Grenoble, France (1977).

72; SEYNHAEVE, J. M., GIOT, M. M., and FRITTE, A. A., "Nonequilibrium Effects
on Critical Flow Rates at Low Qualities," paper presented at the
Speciali'sts' Meeting on Transient Two-Phase Flow, OECD Nuclear Energy Ag-
ency Committee on the Safety of Nuclear Installations, Toronto (August
1976) •

73. SILVER, R. S., "Temperature and Pressure Phenomena in the Flow of Saturated
Liquids," Proc. Roy. Soc. London, Series A, 194, 17 (1948).

74. SIMONEAU, R. J " "Pressure Distribution in a Converging-Diverging Nozzle


During Two-Phase Choked Flow of Subcooled Nitrogen," NASA TMX-71762 (1975).

75. SIMONEAU, R. J., Personal communication (1979)

76. SOZZI, G. L. and SUTHERLAND, W. A., "Critical Flow of Saturated and Sub-
cooled Water at High Pressure," in ASME Symp. Vol. Non-Equilibrium Two-
Phase Flows, pp. 19-25 (1975); General Electric Co. Report NEDO-134l8
(1975).

77. THANG, N. T., and DAVIS, M. R., "The Structure of Bubbly Flow Through
Venturis," Int. J. Multiphase Flow 2., 17-37 (1979).

78. "TRAC-PlA, An Advanced Best-Estimate Computer Program for PWR LOCA An-
alysis," Los Alamos Scientific Laboratory Report LA-7777-MS, NUREG/CR-0065
(May 1979).

- 262 -
REFERENCES (Cont'd.)

79. VOLYAK, L. D., Doklady Akad. Nauk USSR 74 (2),307-310 (1950).

80. WALLIS, G. B., One-Dimensional Two-Phase Flow, McGraw-Hill, New York (1969)

81. WALLIS, G. B., "Critical Two-Phase Flow," Int. J. Multiphase Flow~, 97


(1980)

82. WEGENER, P. P., and WU, B. J. C., "Gas dynamics and Homogeneous Nucle-
ation," in Nucleation Phenomena," A. C. Zettlemoyer, Editor, p. 325,
Elsevier, New York (1977}.

83. WOLFERT, K., "The Simulation of Blowdown Processes with Consideration of


Thermodynamic Nonequilibrium Phenomena," Paper presented at OECD
Specialist's Meeting on Transient Two-Phase Flow, Toronto (1976).

84. WU, B. J. C., "Analysis of Condensation in the Centered Expansion Wave in a


Shock Tube," in Condensation in High-Speed Flows, ASME (1977).

85. WU, B. J. C., SAHA, P" ABUAF, N., and JONES, O. C., JR., "A One-
Dimensional Model of Vapor Generation in Steady Flashing Flow,"
BNL-NUREG-25709 (1979).

86. ZALOUDEK, F. R., "The Critical Flow of Hot Water Through Short Tubes,"
HW-77594 (1963).

87. ZIMMER, G. A., WU, B. J. C., LEONHARDT, W. J., ABUAF, N., and JONES, O. C.,
JR., "Pressure and Void Distributions in a Converging-Diverging Nozzle With
Nonequilibrium Water Vapor Generation," BNL-NUREG-206003 (April 1979).

88. ZUBER, N., STAUB, F. W" and BDWAARD, G., "Vapor Void Fraction in Sub-
cooled Boiling and in Saturated Boiling Systems," in Proc. 3rd Int. Heat
Transfer Conf.,Vol. 5, p. 24 (1966).

89. ZUBER, N., and FINDLAY, J., "Average Volumetric Concentration in Two-Phase
Flow Systems," J. Heat Transfer, 87C, 453 (1965).

- 263 -
APPEr-mIX I

Relation of Thermodynamic States Across a Nonequilibrium Vaporization Region

In order to test the second approach discussed in Section 6.1 as it might


apply to the experimental data reported, it was decided to divide the flashing
flow of an initially subcooled liquid in a converging-diverging nozzle into
three regions. In the first region, the liquid depressurizes, its local pres-
sure falling below the saturation pressure. Although the liquid becomes super-
heated, there is still no vaporization. In this frozen region, it was assumed
that the extrapolated values of the liquid properties from the subcooled region
still apply. Then a certain critical superheat (or "underpressure") is reached,
vapor generation starts under nonequilibrium conditions, and it continues in the
next (second) region, called the transition region, gradually restoring thermo-
dynamic equilibrium. At the end of this region, local thermodynamic equilibrium
is established and a vapor-liquid saturated mixture is assumed to exist. Simi-
lar behavior is also found to occur in straight pipes, where the inception
superheat leads to a low void, nonequilibrium zone gradually tending to equi-
librium. In the last of the three regions, the two-phase mixture undergoes
equilibrum change, and the flow remains in local thermodynamic equilibrium
thereafter. The thermodynamic properties of the mixture are then given by the
equilibrium equation of state, e.g., the steam table.

Figure A.l shows schematically the one-dimensional case of this situation.


The flow in the transition region is governed by the fluid dynamic and transfer
equations, and their solution is not available at this time. On the other hand,
the relationships between the thermodynamic states across the transition region
are governed by the conservation equations and the respective equations of state
in regions 1 and 2 (see Fig. A.l). Thus, for one-dimensional steady flows in a
constant area duct, with a coordinate system in which the transition region is
stationary, one finds

Continuity: P:lYl P2 U 2 G,

2 2
Momentum: PI + Plul P2 + P'2 u 2 F, (A.2)

1 2 1 2
Energy: hI + lUI h2 + I U 2 = H, CA.3)

State: hI f 1(Pl,Pl), (A.4)

h2 = f2~(P2'P2)' (A,5)

where f2 represents the extrapolated liquid equation of state and f2¢ the
equation of state of the saturation mixture, and P,p , h, and u are the pres-
sure, density, enthalpy, and velocity, respectively.

- 264 -
FLOW

0000.00000°0
o
000000 ° 00 00 00 000 0
00 0
°°
0 0 1 0 0 0 0
. 0 0 0 0 00 0 00 0 0 0 0 : 0 0 0 0 0 0
0 0
° 0 00 0 0 0 0
0 0 0 0 0 0 \ 0 0 0 0 0 000 0 0 00 0 0 0 0 0 0 0
\ o 0 0'
0
°
0 0
0 0 0 0 0 0
0 0 0 0 0
00 0 00 • 0 0 0
0 0 000 00
00 0 0 0 0
0 0 0 0 0
0 0
0 ° 0

II
0
o 0 0 0 ,,00\ 00 0 0 0 0 . ' 0 0 0 0 0 0 0 000 0
o 0 0 0 0 0 0 0 0 0 0 0 0 00 0 0 0 0 ~ 0 g o o D
o
0 0 0 0 0
0 0 0 0
0 0 0 0I
0
0 0
0 0 0
• 00 0 0 0
0 0 0 0 0 0 0
0
° 0 0 0 ° 0
00 0 0 0 0 0 0 0

Figure A.l Schematic Representation of a One-Dimensional Nonequilibrium


Vaporization Front. (BNLNeg. No. 10-1202-79)

- 265 -
In region 2, it is assumed that the mixture is homogeneous and there is no
slip between the phases. The constants of motion G, F, and H are the total mass
flux, the thrust per unit area, and the total (stagnation) enthalpy of the flow.
The constancy of F and H is the result of the additional assumption of the ab-
sence of heat loss (or gain) and friction at the walls. Note, however, that
interphase exchanges of energy and/or momentum in the transition region do not
violate this assumption and may be present.

Equations (A.l) to (A.3) are identical to the well-known equations of


normal shock waves in classical gas dynamics, except that for classical shock
waves the gas on either side obeys the same equation of state. The problem here
is more akin to shock waves with chemical reaction, detonation, or combustion
fronts, where two different sets of equations of state are needed. Since vapor
will be present in region 2 but not in region 1, it is expected that P2 <P , It
l
follows from Equation (A.l) that u2 > ul, whence P2 < PI from the
momentum Equation (A.2), and h2 < hI from the energy Equation (A.3). Thus,
in the transition region the fluid is expected to change from an initial
superheated state at (Pl,Pl) to a final saturation state at a lower pressure
and density. In order for this transition to be realized physically, the ad-
ditional requirement on the entropy, s2 > sl, must be satisfied. In con-
trast to shock waves in gas dynamics, as-shown later, the physical requirement
of pressure and density decrease across the transition zone does not necessarily
violate the second law of thermodynamics.

Relationships between the thermodynamic variables in states 1 and 2 may be


obtained by eliminating the velocities ul and u2 in Equations (A. 1) to
(A.3). Thus, from Equation (A.l) and (A.2),

2 2
G G
,...--
PI + P2 + P2
PI

or
PI - P2
-G 2
1 1
- --
PI P2
In terms of the specific volume V - 1/ P,
PI - P2
-G2 (A.6)
VI - V2

This is the well-known Rankine relation, according to which the slope of the
straight line connecting the initial and final states on a p-V diagram must be
negative and is equal to -G2. That the pressure and density must decrease
across the transition region is consistent with real values of G. Next, the
velocities ul and u2 may be elminated between the continuity and energy
Equations (A.l) and (A.3), and the Rankine relation, Equation (A.6), may sub-
sequently be substituted to eliminate the mass flux to yield

- 266 -
(A.7)

With the equations of state (A.4) and (A.S), this equation gives an expres-
sion of P2 in terms of V2 for a prescribed set of PI and VI' This is
the well-known Hugoniot relation, and the curve P2 = P2(V2) on a p-V dia-
gram is called the Hugoniot adiabat.

A particular, limiting case seemingly representative of a few of our ex-


periments will be examined first. If the transition zone is relatively narrow,
then the Rankine-Hugoniot relations may be considered to give the "jump" con-
ditions across a vaporization front, much like a shock wave. For transition
zones of finite thickness, such an approach yields only a crude approximation.
However,. for flashing flow in a constant area duct, it will provide limiting
conditions for the flow and thermodynamic variables. In addition, useful phys-
ical insight may be gained in such a simplified treatment. If the transition
zone is assumed to have vanishing thickness, the above equations may be applied
to flows in a variable area duct. Under this assumption, the flashing flow in a
converging-diverging nozzle may be comnuted in a three-step manner: The liquid
first depressurizes without phase change to a pressure below the saturation
value. Here the flow may be treated by the Bernoulli equation or isentropic ex-
pansion up to the transition point, whereupon flashing begins and the fluid
property "jumps" from a supersatura ... ~d state to an equilibrium two-phase state.
Finally equilibrium expansion once again is used to describe the flow downstream
of the "jump.*" Thus, the jump condition is needed to provide the initial
condition of the flow in the last region of equilibrium expansion. In a few of
the experiments evaluated (see Figures 7.3 and 7.4 in Section 7.1) this indeed
seemed to be close to what was observed. Homogeneous equilibrium seems to be a
good assumption downstream of flashing inception, but in most situations broad.
zones of nonequilibrium were observed, and an extension of this method will be
required.

The Hugoniot relation has been evaluated for the conditions of Run 77.
Assuming PI = 80 kPa (in this experiment, saturation occurred at p ~ 99 kPa)
the Hugoniot curve in Figure A.2 was obtained. Here, in contrast to gas shocks,
the "pre-jump" state (Pl,Vl) does not lie on the Hugoniot curve because
(P2 = Pl> V2 = VI) is not a solution of Equation A.7 since hl'f f2 (Pl'P I ),
A straight line connecting (Pl,Vl) to a point on the intersection of the
Rankine and Hugoniot curves gives a possible solution of Equations (A.l) to
(A.3). The slope of the straight line will give the value of -G2 required to
support such a discontinuity in a stationary configuration. Here the direction
of change is from upper left to lower right in the p-V plane, and the process is
formally analogous to that of a deflagration wave (Courant 1948, Hayes 1960).
Note that there is a point, labeled the Chapman-Jouguet (C-J) point, on the
Hugoniot adiabat which is the point where the adiabat is tangent to the Rankine
line. This is the steepest Rankine line that can be drawn for the Hugoniot

* Schrock et al. (1977) used a similar approach.

- 267 -
80 (~,VI)
RUN 77, BNL
" "" Hugoniot Hugonlot Curve
"" at PI =80 kPa
""
,....
70
Rankine
"/,' - .
0 2
a.. Slope=-G cr "

-
.:s:.
a.
=-(551 kg/m 2 s)2

60 Chapman
Jouguet Point

0.02 0.04 0.06 . 0.08


V(m 3/ kg)
Figure A.2 Hugoniot Curve of Vaporization Front at PI = 80 kPa in BNL Flashing
Experiment Run 77. (BNL Neg. No. 10-1201-79)

- 268 -
curve corresponding to (Pl,Vl)' and it represents the maximum mass flux
allowed by Equations (A.l), to (A.3). Rankine lines for lower mass fluxes will
be less steep and may intersect the Hugoniot at two points, one above and one
below the C-J point, corresponding to the weak and strong deflagration
solutions, respectively. The existence of strong deflagrations is doubtful, but
combustion waves may be described as weak deflagrations (Courant 1948, Hayes
1960). In a weak deflagration, the fluid enters the transition region with a
subsonic velocity, and it leaves the transition again with a subsonic velocity
with respect to a2' the speed of sound in region 2. At the Chapman-Jouguet
point, u2 = a2' The entropy s2 increases along the Hugoniot adiabat from
the point p = PI to the C-J point, reaches a maximum there, and decreases
thereafter. The entropy s2 corresponding to the state downstream from a weak
deflagration transition is greater than sl'

For the present purpose, note that the maximum allowable mass flux while
still maintaining a stable vaporization front satisfying the assumptions laid
down here is given by the slope of the Rankine line at the C-J. point. In Figure
A.2, it is seen that Gmax = Gcr = 551 kg/m 2 s at the point in the nozzle
where PI = 80 kPa for superheated liquid flow. Results of similar
calculations for other values of PI in Run 77 are summarized in Figure A.3.
The curve is a plot of the experimental value of G determined from the measured
nozzle inlet conditions and effective flow area distribution from hydrodynamic
calibration for an isentropic expansion of the liquid to superheated states.*
The circles indicate mass fluxes Gcr at the C-J point. An intersection of the
Gcr and Gexp curves would have provided a solution for the location of the
flashing front where the experimental mass flux may be accomodated and the
flashing front maintained steadily, but such an intersection was not found for
Run 77. Indeed, Gcr is too low by a factor of 10 to 20.

Figure A.4 presents results of similar calculations for an experiment


published by Schrock et. al. (1977). Calculations for the experimental con-
ditions of Reocreux's (1974) Run 423 led to the results in Table A.l.

* It was found that isentropic expansion of the liquid is almost identical to


an isothermal, incompressible flow such as that given by the Bernoulli
equation.

- 269 -
14
..................
..............
.................... RUN 77, BNL
12 ..............
..............

10
-
C\J
(f)

Psat{S in)
E 8
.........
0'1
~
.........
<9
6

60 80 100 120 140 160


P( kPa)
Figure A.3 Comparison of Experimental Mass Flux Gexp and Critical Mass Flux
Gcr at C-J Point in BNL Experiment Run 77.
(BNL Neg. No. 10-1200-79)

- 270 -
60r------.-------.------~------~----__

Schrock I et 01.
RUN 56

40

20

o
o
o
o
o

OL-------L-------~ ______ J __ _ _ _ _ _~_ _~_ _~

2 3 4
P (MPo)
Figure A.4 Comparison of Experimental Mass Flux Gexp and Critical Mass Flux
Gcr at C-J Point in Run 56 of Schrock et a1., (1977).
(BNL Neg. No. 10-1203-79)

- 271 -
TABLE A.l

COMPARISON OF THE CRITICAL MASS FLUXES CALCULATED BY THE JUMP CONDITIONS

AND THE EXPERIMENTAL VALUE MEASURED BY REOCREUX (RUN 423)

Pin 243 kPa Psat(Tin) = 210 kPa


o
121.9 C G = Gin = 4.38 Mg/m 2 s const

175 1.11 4.38

190 1.32 4.38

200 1.51 4.38

Again, no solution was found for either Schrock's or Reocreux's experi-


ment although the discrepancy was smaller. The failure to reach a solution in
these calculations has led to reconsideration of the basic assumption underlying
this first treatment. Perhaps the most serious condition assumed that may not
be realistic is the complete kinematic and thermodynamic equilibrium of the
post-jump condition.

A similar approach was undertaken by Lackme (1978) at Grenoble, who intro-


duced an additional parameter in the conservation equations and the equations of
state. This new parameter allowed a certain fraction of the liquid to remain
superheated (thermal nonequilibrium) while only the remainder" participated in
the vapor generation and reached saturation conditions. This additional para-
meter, which allowed thermal nonequilibrium to be maintained at inception and
downstream, provided the solution to the jump equations. Lackme (1979a, 1979b)
applied his method successfully to the experiments of Reocreux (1974) as well as
to the blowdown exper:iments. It is the authors' belief that a similar approach
may also be undertaken for nozzle flows.

- 272 -
APPENDIX II

Critical Mass Flow Rates in Nozzle Flows With Subcooled Inlet Conditions

The two pieces of information necessary to predict critical mass flow rate
in nozzles with subcooled inlet conditions have now been delineated in the prev-
ious sections:

a. The flashing inception occurring at the throat.

b. The underpressure at flashing inception, which can be calculated from


the Alamgir-Lienhard correlation, Equation (6.2).

The net result of these two criteria is to allow one to consider only ,
single-phase flow up to the throat, and to provide an accurate method of pre-
dicting the throat pressure at flashing.

The critical mass flux at the throat is given directly by the single-phase
relationship

(A.8)

or simply

(A.9)

where ~PFio is given directly by Equation (6.2). Note that the expansion
rate, E', may be given for nozzles from Equation (6.8) by

+.£E.
at (A.IO)

where op/ot is the transient component due to any nonconvective system de-
compression and is equal to zero for steady nozzle flows. Although frictional
effect has been considered negligible in the calculation of L', it is included
in Equations (A.8) and (A.9) in terms of the discharge coefficient. Since
~ PFio is in terms of the critical mass flux itself, Equation (A.9) becomes
transcendental but is easily solved by, say, a simple iteration method.

- 273 -
The method outlined above was applied to Powell's experimental data (Powell,
1961) with subcooled inlet conditions, covering a wide range of inlet pressures
(2,800 to 17,000 kPa) and inlet temperatures (203° to 288°C). The comparisons
of the measured critical mass fluxes at the throat with the calculated values
(solid lines) are presented in Figure A.5 for various inlet pressures and inlet
temperatures. The calculated critical throat mass fluxes are within +5 % of the
experimentally measured ones (Figure A.6) for the entire range of inlet pres-
sures and temperatures reported, and the calculational method also seems to
predict correctly the observed trends of decreasing critical mass flux at a
given inlet pressure with increasing inlet temperatures.

The deviations observed in the predictions of Powell's data may have several
causes. First, since no single-phase calibration data exist, the discharge
coefficient of the nozzle was assumed to be 0.9 in all the calculations.
Second, the Alamgir-Lienhard correlation (Equation 6.2) used for the calculation
of PI has a quoted accuracy of +10 % in predicting the pressure undershoot.
Third, and most important, there is no estimation of experimental measurement
accuracies in Powell's data on the mass fluxes, pressures, and temperatures. In
spite of all these points, the calculational method proposed seems to predict
quite accurately the critical mass flow rates as a function of nozzle inlet con-
ditions (po, To) and nozzle geometry.

- 274 -
15

,~
'0 POWELL
-I' • To =203C
;;;
N • To =230C
E -5%
"- =258C
CI • To
...:
=288C
u • To
(!)

!ci0
a:::
J:
I- 10
I-
<C
X
:J
-'
lL.
Cf)
Cf)
<C
:!:
-'
<C
(J
i=
a:::
(J

5
J.....o

0 5 10 15x 10 3
I NLET PRESSURE (kPo)

Figure A.5 Comparison of Critical Throat Mass Fluxes Measured by Powell


(1961) with Those Calculated by the Present Method for Different
Nozzle Inlet Pressures and Temperatures. (BNL Neg. No. 3-97-80)

- 275 -
~

o
w
~
...J
~
U
...J 10
<t
u
III
N
E
"-
CI
~

u
<.!) 5

5 10
Gc (kg/m 2 5) EXPERIMENTAL

Figure A.6 Comparison of Calculated with Measured Critical Throat Mass


Fluxes (Powell, 1961) for Various Nozzle Inlet Conditions.
(BNL Neg. No. 3-96-80)
* u.s GOVERNMENT PRINTING OFFICE: 714-037#31

- 276 -

You might also like