|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
EXPERIMENTS INT CRYOGENIC TWO PHASE FLOW
CHRISTOPHER JAMES VELAT
A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
UNIVERSITY OF FLORIDA
Christopher James Velat
I would like to express my appreciation to all of the individuals who have assisted
me in my educational development and in the completion of my thesis. My greatest
gratitude is extended to my supervisory committee chair, Dr. James Klausner. Dr.
Klausner's boundless patience, constant encouragement, friendly demeanor, and
professional expertise have been critical to both my research and education. He shares an
outstanding level of professionalism and respect for all students. He is an inspirational
professor and regarded highly. Dr. Renwei Mei also deserves recognition for his patience
and astonishing technical expertise. I would like to further thank Dr. David Hahn for
kindly serving on my supervisory committee.
I would like to additionally recognize my fellow graduate associates Jelliffe
Jackson, Yusen Qi, Jun Liao, and Yi Li for their friendship and technical assistance.
Their diverse cultural background and character have provided an enlightening and
Additional acknowledgment must be addressed to Mrs. Becky Hoover for her
continual administrative support and guidance. Mrs. Hoover has been an invaluable
resource and has continually offered support and encouragement.
I would like to further acknowledge the Florida Space Grant Consortium for its
funding and the opportunity to tour the Kennedy Space Center. This research was also
funded by NASA Glenn Research Center under contract NAG3-2750.
Finally I would like to recognize my parents and family for their continual support
and encouragement. I would like to thank my mom and dad for instilling the values that
have afforded me success in my continual endeavors.
TABLE OF CONTENTS
ACKNOWLEDGMENT S ............ ...... .__ .............. iii...
LIST OF FIGURES .............. ....................vii
NOMENCLATURE .............. .................... xi
AB STRAC T ......__................ ........_._ ........xi
1 INTRODUCTION ................. ...............1.......... ......
1.1 Cryogenic Chill down Overview............... ...............2
1.1.1 Horizontal Flow Regimes ................. ...............3................
1.1.2 Heat Transfer Mechanisms ................. ...............5..............
1.2 Outline of Current Investigation............... ..............
2 EXPERIMENTAL FACILITY .............. ...............8.....
2.1 System Overview................. ..... .............
2.2 Liquid Nitrogen Storage and Supply ................. ...............10...............
2.3 Test Section Design ................ ...............11................
2.4 Instrumentation and Calibration ............__......___....._ ............1
2.4.1 Static Pressure Transducers ................. ...............13................
2.4.2 Test Section Pressure Drop ............... .. ...... ..............1
2.4.3 Flow Meter Calibration and Velocity Correlation............... ..............1
2.4.4 Temperature Measurements .............. ...............19....
2.5 Heat Transfer Section Design ................. ...............21........... ...
2.6 Vapor Volume Fraction ................ ...............22................
2.7 Vapor Quality Estimation ................. ...............24................
2.8 Data Acquisition System .............. ...............26....
2.9 Digital Imaging Facility ................. ...............28................
3 CRYOGENIC CHILLDOWN VISUALIZATIONS IN HORIZONTAL FLOW .....29
3.1 Introduction and Literature Survey ................. ...............29........... ...
3.2 Visualizations of Cryogenic Chilldown............... ...............31
3.2.1 Pre-Quenching Front Visualizations .............. ...............34....
3.2.2 Quenching Front Visualizations ................. ...............................37
3.2.3 Post Quenching Front Visualizations ................. .... ............... ....._.42
3.3 Vapor Volume Fraction and Quality Measurement Results ............... ... ............48
3.4 Discussion of Results............... ...............5
3.5 Sources of Experimental Error ................. ...............55........... ...
4 WALL TEMPERATURE PROFILES AND CHILLDOWN TIME INT
CRYOGENIC HORIZONTAL FLOW ................. ...............57........... ....
4.1 Introduction and Literature Survey ................. ...............57........... ...
4.2 Wall Temperature Profile and History .............. ...............58....
4.3 Discussion ................. ...............71........... ....
4.4 Sources of Error............... ...............74
5 CONCLUSIONS AND RECOMMENDATIONS .............. ...............75....
5.1 Accomplishments and Findings............... ...............75
5.2 Recommendations for Future Research ................. .............. ......... .....76
A PROPERTIES OF NITROGEN .............. ...............77....
B EXPERIMENTAL CRYOGENIC CHILLDOWN VISUALIZATIONS ..................78
LIST OF REFERENCES ................. ...............109................
BIOGRAPHICAL SKETCH ................. ...............112......... ......
LIST OF FIGURES
1-1 Typical flow regimes observed in horizontal two-phase flow. ............. ..................5
2-1 Schematic diagram of the cryogenic flow facility. ................ ....__ ...............9
2-2 Schematic overview of a typical vacuum jacketed cryogenic cylinder ............1 1
2-3 Wall thickness selection chart used to design the visual test section [l l]. ..............12
2-4 Schematic of the flange assembly. ................ ....__. ...... ........._.....13
2-5 Calibration plot of the inlet pressure transducer. ......... ................ ...............14
2-6 Calibration plot of the exit pressure transducer. ................ ................. ....... 14
2-7 Calibration plot of the test section differential pressure transducer. ................... .....15
2-8 Calibration plot of the venturi differential pressure transducer. ............. ...... ........._18
2-9 Calibration plot of the actual velocity versus the ideal velocity. ............. ...... ..........19
2-10 Schematic of the thermocouple arrangement on the steel transfer line. ........._........20
2-11 Thermocouple arrangements along the pipe wall of the heat transfer section. ........21
2-12 Thermocouple arrangements along the insulation of the heat transfer section. .......22
2-13 Diagram of the various flow regimes observed. ........._._. ......__ ..............23
2-14 Diagram of the model used for the stratified flow volume fraction computation....23
2-15 Diagram of the model used for the annular flow volume fraction computation......24
3-1 Experimentally observed flow structure transitions during cryogenic chilldown
for horizontal flow. ................. ...............33....... .....
3-2 Photographs of the unstable vapor flow that occurs at the beginning of
chilldown. (a) G = 25 kg/m2-S, T, = 8.3 o C, (b) G = 25 kg/m2-S,
T = -0.2" o ................. ...............35.._._. ....
3-3 Photographs of the unstable vapor flow that occurs at the beginning of
chilldown. (a) G = 64 kg/m2-S, T, = -12.4 o C, (b) G = 62 kg/m2-S,
T, =-14.8" o ........._.._ ..._.._. ...............36.
3-4 Photographs of the unstable vapor flow that occurs at the beginning of
chilldown. (a) G= 104 kg/m2-S, Tw 7.5 o C, (b) G= 115 kg/m2-S,
Tw = -2.3 o C................. ...............37....__ ...
3-5 Sequential photographs of the fi1m boiling front as it crosses through the visual
test section. (a) G 24 kg/m2-S, Tw -10.0 o C,
(b) G = 24 kg/m2-S, Tw = -10.7 o C, (c) G = 26 kg/m2-S, Tw = -11.7 o C. ........._.....38
3-6 Sequential photographs of the fi1m boiling front as it crosses through
the visual test section. (a) G = 72 kg/m2-S, Tw = -28.9 o C,
(b) G = 72 kg/m2-S, Tw = -30.2 o C, (c) G = 72 kg/m2-S, Tw = -31.4 o C. ........._.....40
3-7 Sequential photographs of the fi1m boiling front as it crosses through
the visual test section. (a) G = 154 kg/m2-S, T, = -26.7 o C,
(b) G= 143 kg/m2-S, T, = -28.0 o C............... ...............41...
3-8 Photographs of the various flow structures observed during the chilldown
process following the passing of the film boiling front.
(a) G 23 kg/m2-S, Tw -18.7 o C, (b) G 72 kg/m2-S,
Tw = -84.3 o C, (c) G = 40 kg/m2-S, Tw= -170.6 o C............... ...................4
3-9 Photographs of the flow structures during the chilldown process following the
passing of the fi1m boiling front. (a) G = 84 kg/m2-s, T- = -59.80 C,
(b) G= 104 kg/m2-s, =-120.90 C, (c) G= 112 kg/m2-s, = 161.6 o C,
(d) G= 126 kg/m2-s, =-175.6 o C, (e) G= 150 kg/m2-s, =-177.2 o C.........44
3-10 Photographs of the various flow structures observed during the chilldown
process following the passing of the film boiling front. (a) G 176 kg/m2-S,
Tw = 39.7o C, (b) G = 234 kg/m2-S, Tw = -51.80 C, (c) G = 211 kg/m2-S,
T, = -62.5o C,(d) G = 338 kg/m2-S, T, = -178.4 o C. ............. .....................4
3-11 Graph of the vapor volume fraction and quality for experiment # 10......................51
3-12 Graph of the vapor volume fraction and quality for experiment # 2. ................... ....5 1
3-13 Graph of the vapor volume fraction and quality for experiment # 1........................52
3-14 Graph of the vapor volume fraction data for experiment # 1 1 ................ ...............54
4-1 Schematic representation of the thermocouple arrangement on the
stainless steel transfer line. .............. ...............60....
4-2 Temperature history during chilldown for experiment # 1. ............. ...................61
4-3 Mass flux history for experiment # 1. ............. ...............61.....
4-4 Temperature history during chilldown for experiment # 2. ............. ...................62
4-5 Mass flux history for experiment # 2. ............. ...............62.....
4-6 Temperature history during chilldown for experiment # 3. ................ ................. 63
4-7 Mass flux history for experiment # 3 .......... .......................... ...............63
4-8 Temperature history during chilldown for experiment # 4. ............. ...................64
4-9 Mass flux history for experiment # 4. ............. ...............64.....
4-10 Temperature history during chilldown for experiment # 5. ............. ...................65
4-11 Mass flux history for experiment # 5. ............. ...............65.....
4-12 Temperature history during chilldown for experiment # 6. ............. ...................66
4-13 Mass flux history for experiment # 6. ............. ...............66.....
4-14 Temperature history during chilldown for experiment # 7. ............. ...................67
4-15 Mass flux history for experiment # 7. ............. ...............67.....
4-16 Temperature history during chilldown for experiment # 8. ............. ...................68
4-17 Mass flux history for experiment # 8. ............. ...............68.....
4-18 Temperature history during chilldown for experiment # 9. ............. ...................69
4-19 Mass flux history for experiment # 9. ............. ...............69.....
4-20 Temperature history during chilldown for experiment # 10. ............. ..................70
4-21 Mass flux history for experiment # 10. ............. ...............70.....
B-1 Chilldown images from experiment # 1. ............. ...............79.....
B-2 Chilldown images from experiment # 2. .............. ...............81....
B-3 Chilldown images from experiment # 3 ................ ................ ......... ...._84
B-4 Chilldown images from experiment # 4. .............. ...............87....
B-5 Chilldown images from experiment # 5. ............. ...............89.....
B-6 Chilldown images from experiment # 6. .............. ...............92....
B-7 Chilldown images from experiment # 7. .............. ...............94....
B-8 Chilldown images from experiment # 8. ............. ...............97.....
B-9 Chilldown images from experiment # 9. ......___ .... ... ._._ ... .._.... ........0
B-10 Chilldown images from experiment # 10. ....._____ ..... ..___ .......__ .......0
A cross sectional area (m2)
Co empirical constant
C, specific heat (kJ/kg/K)
C, critical wave velocity (m/s)
C,, inviscid wave velocity (m/s)
f frequency (Hz)
g gravitational acceleration (9.81 m/s2)
G mass flux (kg/m2-S)
h liquid level height (m)
hiv latent heat of vaporization (J/kg)
k thermal conductivity (W/mK)
Kt2 Vortex flow meter constant
ma measured amperage (milliamps)
P pressure (kPa or psi)
Q volumetric flow rate (cfm or L/min)
r radius (m)
S slip velocity, ratio of vapor and liquid velocities
t time (s)
T temperature (oC or K)
T, average wall temperature (oC)
u axial velocity (m/sec)
U average axial phase velocity (m/s)
V voltage (volts)
X vapor quality
a volume fraction
/7 angle of inclination from the horizon (degree)
3 liquid film thickness (m)
pu dynamic viscosity (Ns/m2)
p density (kg/m3)
o liquid/vapor surface tension (N/m)
Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfdllment of the
Requirements for the Degree of Master of Science
EXPERIMENTS IN CRYOGENIC TWO PHASE FLOW
Christopher James Velat
Chair: James F. Klausner
Major Department: Mechanical and Aerospace Engineering
A cryogenic chilldown facility has been designed, fabricated, and tested to
experimentally study the flow structure and heat transfer processes that occur in
horizontal two-phase flow during chilldown. The facility incorporates a Pyrex viewing
section from which the flow structure and heat transfer transitions may be observed and
recorded using a high resolution CCD camera system. Temperature and pressure
measurements are recorded from sensors throughout the facility to monitor the flow
conditions and the external transfer line wall temperature.
Visualizations of the flow structure have revealed the presence of several
transitions during the chilldown phase. The general chilldown process begins with purely
vapor flow. The vapor flow results from the rapid evaporation of the cryogenic liquid
when it encounters the hot transfer line. The vapor surges through the system and was
observed to choke at the throttling ball valve. As chilldown progresses droplets of liquid
begin to accumulate on the bottom of transfer line. Film boiling is evident at this phase
and the droplets travel at high speed on top of a thin vapor layer. A more observable
stream continues to develop until a quenching front is observed. The quenching front
marks the transition from film boiling to nucleate boiling. The nucleate boiling phase
typically assumes a stratified/wavy flow structure, which gradually transitions into slug
and plug flow near the end of chilldown. At high flow rates the nucleate boiling phase is
suppressed and occurs for a shorter duration. Following the nucleate boiling phase the
dominant heat transfer mechanism observed is two-phase bulk turbulent convection, with
evaporation at the piqued/vapor interface. Annular flow was observed in several
experiments but is not specifically addressed in the current research.
Temperature profiles were produced from 6 thermocouples that were placed on
the external wall surface of the transfer line and insulated with Melamine foam
insulation. Chilldown times were computed from temperature profies and an
approximate Leidenfrost temperature of -980C was observed. Large circumferential
temperature gradients were observed during chilldown and moderate axial gradients were
also observed during the passing of the quenching front. The mass flux of the
experiments had a considerable impact on the chilldown time and magnitude of the
The experimental facility and the current data will be used in continued research
to develop a comprehensive and fundamentally based numerical chilldown model.
Cryogenic fluid transfer is a complex physical phenomenon concerning the
transport of low temperature liquids through transfer lines that are initially a couple of
hundred degrees Celsius higher in temperature. When the cryogenic fluid contacts the
comparatively hot pipe wall, voracious evaporation ensues resulting in large unstable
pressure fluctuations and rapid contraction of the transfer system components.
To develop a comprehensive model of the chilldown process an elaborate
investigation into the flow structure, heat transfer, pressure fluctuations, and temperature
variation is required. Early examinations of chilldown conducted by Burke et al. 
resulted in a crude model for estimating the chilldown time in pressurized lines. The
model was constructed from a lumped energy and mass balance analysis of a control
volume and disregarded the detailed transient phenomena of current interest. Bronson et
al.  further investigated chilldown and attributed the existence of large temperature
gradients between the top and bottom of transfer lines attributed to flow stratification.
Additionally Bronson recognized the importance of flow regimes and suggested there
exists good correlation between the empirical oil and water flow regime maps of Baker
 and the two-phase flow of hydrogen. Hedayatpour et al. , and Steward et al. 
have proposed more detailed numerical based chilldown models. Hedayatpour developed
an advanced two-phase flow model for chilldown in vertical transfer lines. Their
predictions of the wall temperature history and chilldown time are in good agreement
with their experiments.
These pioneering investigations on chilldown have led to the development of
functional cryogenic transfer systems. Yet, the detailed momentum and heat transfer
processes occurring during chilldown are still not well understood, which limits the
development of more advanced hydrodynamic and thermal models. The focus of this
research is to provide detailed data on the two-phase flow characteristics of liquid
nitrogen flowing through a horizontal transfer line during chilldown. The data will be
used in future research to evaluate existing flow regime and heat transfer coefficient
models and, when necessary, to develop new models for cryogenic two-phase flows.
Advanced understanding of the chilldown stage of cryogenic transfer will ultimately
improve safety, reduce fuel loss and time associated with the filling of reusable space
vehicles, and encourage the development of advanced cryogenic systems.
1.1 Cryogenic Chilldown Overview
Cryogenic chilldown refers to the process by which the temperature of the transfer
line is lowered to the saturation temperature of the cryogen. This process is highly
unstable and characterized by large pressure fluctuations accompanied by transitory
boiling heat transfer. As noted by Hedayatpour the specific two-phase flow pattern
observed during chilldown is dependent on the pipeline orientation, cryogenic fluid
temperature, vapor quality, and the flow rate. In the current work, visualizations of the
chilldown process have revealed the presence of a quenching front at the onset of
chilldown, which is followed by a sustained stratified liquid growth with evaporation at
the liquid vapor interface. Nucleate boiling has been observed in all of the experiments
and is most prominent at low mass flux.
1.1.1 Horizontal Flow Regimes
Flow regime identification is critical to the development of a comprehensive model
for chilldown. The heat transfer and pressure drop of a system can vary significantly
depending on the flow regime condition. Parameters that affect the observed flow regime
include the pipe orientation, flow rate, and vapor quality. Flow regimes for vertical flow
vary from those observed for horizontal flow. As noted by Carey  horizontal flow has
a propensity for stratifieation, resulting from the tendency of the vapor to migrate to the
upper portion of the tube. As described by Warren and Klausner  in low quality flows
the existing vapor bubbles of bubbly flow tend to agglomerate at the top of the tube due
to buoyant stratifieation in the horizontal flow orientation.
The relative velocity between the vapor and the liquid also influences the observed
flow regimes. Barnea and Taitel  noted that transitions from stratified flow to either
wavy or in more extreme cases slug flow result from the instability generally attributed to
the viscous Kelvin-Helmholtz instability. The Kelvin-Helmholtz instability is caused
primarily from the Bernoulli effect by which the pressure decreases over the wave crest
due to the velocity acceleration . The work by Barnea and Taitel on interfacial and
structural stability of separated flows identified the importance of the wall shear stress in
the stability analysis for transitions to wavy and slug flows using the viscous Kelvin-
Helmholtz neutral stability criterion,
\P; P, A
(U -U)< pa, +pga) "gcosp (1.1) ,
where h is the liquid level, P is the angle of inclination from the horizon, A is the cross
sectional area, U is the axial average phase velocity, a is the ratio of the phase area to the
total cross sectional area, and K is a factor described below. For the inviscid case K = 1
and for the viscous case K = Kv where
(C C, )2
K, = 1 v z (1.2)
P; P, A
gg cos f
P2~ 1 la g g~, 13
Cv CIv is the difference between the critical wave velocity at the inception of instability
and the wave velocity for the inviscid Kelvin-Helmholtz analysis. This term accounts for
the effect of the wall shear stress, which noted by Barnea and Taitel, tends to amplify
disturbances of the film thickness. The viscous and inviscid criteria differ in the
computational difficulty and predictive capabilities, with the latter offering an analytical
solution but tending to over predict stability. In addition to tube orientation and relative
phase velocity, heat transfer also influences the prevailing flow regimes due to the
transverse momentum flux created by evaporation.
The flow structure of a system is affected by the quality and occurrence of
destabilizing heat transfer mechanisms. The introduction of a cryogenic fluid to a
transfer line results in rapid evaporation. This evaporation cools the system until
thermodynamic conditions prevail capable of supporting a liquid film. Destabilizing
boiling heat transfer continues during this initial phase, producing film boiling. As the
system cools further the quenching front transverses downstream, and the liquid level
Figure 1-1. Typical flow regimes observed in horizontal two-phase flow.
1.1.2 Heat Transfer Mechanisms
An understanding of the prevailing heat transfer mechanisms is required to
develop a comprehensive chilldown model. During chilldown three main heat transfer
regimes were typically observed. In the beginning of chilldown the cryogenic fluid is
introduced to the transfer line of the system, which is several hundred degrees Celsius
higher in temperature. The cryogenic fluid undergoes rapid evaporation and the small
liquid slugs present reside on a thin layer of vapor that is formed at the bottom of the
transfer line. This stage, commonly referred to as film boiling, persists until the system
chills down to the Leidenfrost temperature of the fluid. Heat is transferred during film
continues to rise. If the mass flux is low enough, nucleate boiling will be persistent with
high wall superheats. As the rate of heat transfer slows in the later stages of chilldown
vapor plugs and bubbles are observed. The various flow regimes typically observed in
horizontal two-phase flow are illustrated in Figure 1-1.
~c' "' IBubbly Flow
-~ ------ _--- Plug Flow
boiling from the pipe wall to the fluid primarily by conduction through the vapor film and
by radiation from the hot pipe wall. Following the passing of the quenching front
nucleate boiling was observed in all of the experiments; however, sustained nucleate
boiling was generally only observed in experiments with low mass flux. When the mass
flux exceeds a threshold, nucleate boiling is suppressed and evaporation occurs at the
liquid/vapor interface of the liquid film; ebullition is not observed at the wall. Typically
as the system neared chilldown, the test section became filled with liquid, resulting in
single-phase convective heat transfer that persisted until the experiment was shutdown.
Knowledge of the prevailing heat transfer regimes is critical in selecting an
appropriate model for the entire chilldown process. Past researchers Bronson et al. ,
Burke et al. , and Chi  generally assumed average heat transfer properties out of the
need for simplification. Observations of the chilldown process as well as the
experimental data clearly illustrate the importance of considering spatial variations and
transients when developing an accurate and fundamentally based chilldown model.
1.2 Outline of Current Investigation
An experimental examination of the two-phase flow characteristics of liquid nitrogen
flowing through a horizontal transfer line during chilldown has been preformed to clarify
the flow structure and heat transfer mechanisms as well as compile a database on two-
phase flow transport. These data will be used in future research to develop advanced
models capable of predicting the flow regime, pressure drop, and heat transfer coefficient
of cryogenic two-phase flows. An experimental facility, which is described in great
detail in chapter 2, has been fabricated that allows visualization of the flow structure and
heat transfer mechanisms. The facility is capable of operating at pressures as high as
1400 kPa and mass flux ranges of 30-1000 kg/m2-Sec. The cryogenic facility is
instrumented with thermocouples and pressure sensors for measuring the relevant heat
transfer and flow properties. Visualizations of the flow regimes and heat transfer
mechanisms are captured using a high resolution CCD camera system. All
instrumentation in the facility has been fully calibrated and the tested measurement
accuracy has been assessed.
In chapter 3 the visualizations from three experiments are presented. The images are
arranged by increasing mass flux and separated by reference to the transitional boiling
front passage. The visualizations prior to the transitional boiling front passage illustrate
the stratified/wavy flow structure present during the initial phase of chilldown.
Subsequent images of the growth and movement of the transitional boiling front from the
three experiments are shown. The visualization section concludes with images that
capture the nucleate boiling heat transfer and slug and plug flow structure transitions,
which occurred after the quenching front passage. In addition to the visualizations, the
vapor volume fraction data and computed quality for the three experiments are also
presented and followed by a general discussion.
Chapter 4 presents the temperature data of the ten experiments. Demarcations
indicating the approximate Leidenfrost temperature and visually observed transitional
boiling front passage are included on the temperature profiles. The mass flux range as
well as chilldown time data for each of the experiments is also reported.
Chapter 5 concludes the research with a summary of the overall work and discussion
of the unresolved issues. Lastly, final recommendations are made for future research
regarding annular flow; heat transfer coefficient, pressure drop measurements, and the
development flow regime maps for cryogenic two-phase flows.
2.1 System Overview
The experimental work presented herein was preformed in a cryogenic facility that
was designed specifically for this investigation. The experimental flow diagram is
displayed schematically in Figure 2-1. A single pressurized liquid nitrogen cylinder
feeds the facility. The cylinder, which is supplied by Praxair at 1587 kPa (230 psi),
provides the driving potential for the liquid nitrogen fluid. Following discharge from the
main cylinder the nitrogen is immediately directed through a shell and tube heat
exchanger device. The heat exchanger coolant is supplied by another pressurized
nitrogen cylinder, which is throttled and expanded to the atmosphere to achieve a
combined Joule-Thomson and evaporative cooling effect. The additional cooling is used
to keep the high-pressure nitrogen at subcooled conditions after being discharged from
the storage cylinder. The working fluid is then directed through a 12.7 mm inner
diameter 15.9 mm outer diameter 304-stainless steel tube that is instrumented with a
series of external type E thermocouples, a single internal type E thermocouple, and a port
leading to a static and differential Validyne pressure transducer. These instruments record
the inlet flow conditions as well as outer wall temperature profile. After passing through
the length of a stainless steel tube the fluid then proceeds through the visual test section,
where high-resolution images of the flow structure are sequentially captured by the CCD
camera system. The working fluid then passes through another length of instrumented
tubing containing a single internal type E thermocouple and port leading to another static
and common differential Validyne pressure transducer. These instruments record the exit
conditions of the working fluid. Positioned immediately after the aforementioned
instruments is a heat transfer section containing internal inlet and exit thermocouple
probes as well as a series of symmetrically spaced thermocouples secured both on the
pipe wall and exterior of the insulation. The heat transfer section provides information
required for analysis of the unsteady heat transfer coefficient. A cryogenic ball valve is
located downstream of the instrumentation, and is used as a flow throttling device. After
passing through the valve, the fluid is heated through a heating section to ensure only
vapor is passed through the venturi flow meter. The nitrogen vapor is collected in a
storage tank and vented to the atmosphere.
Figure 2-1. Schematic diagram of the cryogenic flow facility.
The facility is operated using a combination of manual control and computer
monitoring. Manual control is required to open the nitrogen cylinders, set the proper
positions of the ball valves, and activate the various computer programs prior to testing.
Monitoring and recording of all temperature and pressure measurements as well as digital
images was achieved via a high performance combination data acquisition image capture
PC-style computer system. The Softwire programming platform is used to process the
collected data. The program designed for this facility is capable of recording and
displaying the data in near real time, allowing for immediate experiment feedback, and
determination of the completed chilldown.
Nitrogen was selected as the cryogen working fluid because it is chemically inert,
colorless, odorless, non-corrosive, nonflammable, relatively inexpensive, readily
available, and poses no significant environmental hazards. Additionally, the properties of
nitrogen are well documented, and may be referenced in Appendix A.
2.2 Liquid Nitrogen Storage and Supply
The test facility requires two vacuum-j acketed pressurized cylinders, schematically
illustrated in Figure 2-2. The cylinders, which are provided and maintained by a local
gas company, are available with a low pressure rating of 207 kPa (30 psi), or a high
pressure rating 1587 kPa (230 psi). The high-pressure tanks are used to achieve high
mass fluxes. To prepare these tanks for use, the pressure building circuit is actuated until
the tank pressure gauge registers the appropriate level. Once the proper pressure level is
reached, the associated liquid valve is opened and the chilldown experiment is
LIquid Level Vem Pressure Rupturre Vemr Relief Pressure Rupture
Gauge .Valve Ga a Doisk Value vorlvr Gauge Disk
Liquide Prsue Prsue Ecnmzr ge a
MI~l:~ Vav uidn Bidng Rglao er av
Llool [.09Reulto Vlv
2.3a Testure Sectione Designle !qud
A crucial component to the cryogenic facility is the visual test section. The visual
test section must be capable of withstanding the dynamic high pressure and low
temperature conditions associated with chilldown, seamlessly attach with the steel tubing,
and provide an excellent view of the working fluid. The success of the experimental
facility is contingent upon meeting these demands while maintaining leak-free operation.
A Pyrex Glass tube was selected for the visual test section. Pyrex was chosen
because of its toughness and superior thermal expansion (32.5 x 107 cm/cm/deg C) as
compared to soda lime and other borosilicate glasses. To prevent condensation on the
outer portion of the visual test section, a vacuum j acketed design was adapted. The wall
thickness of the inner tube was selected to preclude catastrophic failure under the desired
operating conditions. To select the appropriate wall thickness an industrial selection
chart provided by the Bibby Sterilin Company was used and is included as Figure 2-3
55p ~ ~ ~ uti- Dia ate 10
Figure) 2-3 Wal thcns seeto chr use to deig th viua tes seto [l]
Based 0 r on a maiu deig pesre of 140 k raa 19 1mm D1. mI ls
tubea70'Y wa seece for use Th extr thcns of th glstb wssletdoprvn
hazardous~~ exlso ftegls eto swll st ihtn h opesv ocso
the~~ ste flne necesar for lekfe oeaion A noe flng aseml wa
cosrce osemesyscreadsah gls tes secio inpaeTefag
assembly cosite of a seie of the flne fatee wit mahieols.A ea-fe
seal414 wa aciee thog th copeso of a Telo -in anTfonppewapa
depicte sceaial in Figr 2-4.I1IcII ~r
................. ..... ... .......... ........Section
........: ~ #triii ....... ...... .... ............... ..... ....1/ /1/ / Te lo
Figure 2-4. Schematic of the flange assembly.
2.4 Instrumentation and Calibration
2.4.1 Static Pressure Transducers
Two Validyne P2-200V pressure transducers were installed in line on either side of
the visual test section. The first transducer was used to measure the inlet pressure of the
flow while exit pressure transducer measured the pressure just prior to the venturi flow
meter. Each transducer was rated to 1380 kPa and had been independently calibrated
using a mercury monometer. The calibration plots are referenced in Figures 2-5 and 2-6.
The linear curve fits for the inlet and exit pressure transducers are
;ae, = -27.712 + 274.937 V (2.1)
P,,l -30.709 + 276.479 V (2.2)
respectively, where Pinlte and Pexit are given as gage pressure in kPa and V is in volts. The
calibrations are accurate to .25% of their respective full scale.
0.0 0.2 0.4 0.6 0.8 1 .0 1 .2 1 .4
Figure 2-5. Calibration plot of the inlet pressure transducer.
0.0 0.2 0.4 0.6 0.8
Figure 2-6. Calibration plot of the exit pressure transducer.
1.0 1.2 1.4
2.4.2 Test Section Pressure Drop
The pressure drop across the test section is measured using a Validyne model
DP215 variable reluctance differential pressure transducer equipped with a dash-30
diaphragm. A carrier demodulator device is used to convert the transducer signal to an
analog voltage. The carrier demodulator device is capable of conditioning the span and
zero of the signal and was calibrated specifically with a designated transducer.
Calibration of the transducer was accomplished using a monometer containing R 827
monometer oil. The resulting calibration curve is graphically displayed in Figure 2-7,
depicting a linear relationship between pressure and voltage. The resulting calibration
AP= 1.2788 (2.3)
where AP is given in kPa and V is in volts. The standard deviation of the calibration was
0.12%, which is within the 0.25% full-scale accuracy claimed by the manufacturer.
0 2 4 6 8
Figure 2-7. Calibration plot of the test section differential pressure transducer.
2.4.3 Flow Meter Calibration and Velocity Correlation
The flow rate of the nitrogen is measured using a Preso classical venturi flow meter
with an inner diameter of 13.9 mm and a throat diameter of 8.73 mm. The flow meter is
located downstream of the test section just past two large 1-kilowatt coil heaters. The
coil heaters are used to ensure that only vapor enters the venturi during operation. The
pressure ports on the venturi meter are coupled to another Validyne variable reluctance
DPl5 differential pressure transducer. A dash-40 diaphragm was selected for the
transducer to maximize sensor output response while providing moderate overload
protection. Similar to the pressure drop setup discussed in section 2.4.2, the differential
pressure transducer is coupled to a carrier demodulator device. The transducer and
demodulator device were calibrated together using a mercury monometer, Figure 2-8.
The standard deviation of the differential pressure transducer calibration was 0. 17%,
which was within the 0.25% full-scale accuracy claimed by the manufacturer. After the
pressure transducer was calibrated, the venturi device was calibrated with an Omega
vortex flow meter using compressed air. The vortex flow meter provided frequency
information that was conditioned into an analog voltage by measuring the voltage drop
across a known resistor. The voltage measured was then related back to frequency
through the relationship below
f = fm.. *I (2.4)
lxb 20 4
where fmax = 914 Hz is the maximum frequency and ma is the measured amperage in
milliamps, which was related to the measured voltage through the equation
ma = ~-*10 (2.5)
where V is the measured voltage drop across the resistor. The analog voltage was then
related to the volumetric flow rate with the equation below
f = K,, Q (2.6)
where Kt2 = 32.443 Hz/CFM is a constant derived from fluid properties and operating
conditions and Q is the volumetric flow rate measured in CFM.
Once the volumetric flow rate was obtained from the vortex flow meter it was
then used to calibrate the venturi by plotting the actual velocity measured with the vortex
flow meter versus the ideal velocity, Figure 2-9. The ideal velocity was computed using
a modified Bernoulli relation , which accounted for compressibility,
pa = p+ pV +-Ma + M a4 +...i (2.7)
The modified Bernoulli relation in equation 2.7 was solved iteratively. A polynomial
curve, shown in equation 2.8, was then fit to the calibration data and used to correlate the
ideal velocity with the actual velocity.
actual = (1*10-')u4 -(5*10-')*uI + 0.0076 *u12 + 0.5653 *ul (2.8)
The ideal velocities shown in Figure 2-10 are computed based on the compressible and
incompressible form of the Bernoulli equation. The differences between the
compressible and incompressible ideal velocities do not visually appear to be significant
in Figure 2-10. However, an analysis using cryogenic flow experimental data revealed a
difference of about 5-10% at high velocities between the incompressible and
compressible flow ideal. Therefore, for all tests Equation 2.7 is used to compile the ideal
velocity and the actual velocity is determined using Equation 2.8.
Additionally noted in Figure 2-9 is the lack of low velocity measurements on the
calibration plot. A trade off between resolution and operable pressure drop range was
inherent to the diaphragm selection. To prevent possible damage to the instrument a
diaphragm was selected that would safely accommodate the range of anticipated
experimental differential pressures. Unfortunately the pressure range limitations on the
diaphragm of differential pressure transducer prevented accurate measurement of the low
velocity flows during the calibration. This unavoidable compromise is of minimal
impact, as the vast maj ority of flow velocities that were in the high range.
0 2 4 6 8 10 12
Figure 2-8. Calibration plot of the venturi differential pressure transducer.
$ j I ~5 1 Compressible Velocity
~1 00 O Incompressible Velocity-
0 50 100 150 200 250 300
Ideal Velocity (m/sec)
Figure 2-9. Calibration plot of the actual velocity versus the ideal velocity.
2.4.4 Temperature Measurements
Each cryogenic experiment required the acquisition of numerous temperature
measurements. The temperature at the inlet and exit of the test section provided
information that was used to monitor the chilldown process as well as compute the
approximate liquid density. Additionally, a series of six thermocouples placed along the
top and bottom of the 304-stainless steel tube prior to the test section (Figure 2-10) as
well as 8 thermocouples placed circumferentially around the transfer line one meter down
from the visual test section (Figure 2-11) supplied essential information on the
progression of chilldown and flow regime transitions. The 304-stainless steel tube has an
approximate thermal conductivity of 16.3 W/m-oC and a specific heat of 0.46 kJ/kg-oC.
To compile the thermodynamic information for the computation of the heat transfer
coefficient the system was outfitted with an additional series of thermocouples that were
located after the test section (described in Section 2.5). A single thermocouple was also
placed inline and just prior to the venturi to correct for the effects of temperature
variation on the Mach number and vapor density.
Temp 3 Temp 2 Temp l
Temp 6 Temp 5 Temp 4
Figure 2-10. Schematic of the thermocouple arrangement on the steel transfer line.
All of the temperature measurements were made with type E (Chromel-Constantan)
thermocouples that were either manufactured in the laboratory or purchased from Omega
Engineering. The thermocouples that were manufactured in the laboratory were prepared
from 36-gauge wire and welded in the laboratory with a bead size of approximately 0.5
mm. These thermocouples were used exclusively on the exterior of the pipe wall and
insulation to measure the respective temperature distribution throughout chilldown
process. 1/16-inch thermocouple probes were purchased from Omega Engineering to
measure the initial fluid temperature. The probes were placed through precision-drilled
holes in line with the fluid and sealed using a combination of brass compression fittings.
The voltages of all of the thermocouples were recorded via the data acquisition system
and converted to temperature using the supporting software.
2.5 Heat Transfer Section Design
The heat transfer segment of the facility was positioned downstream of the visual
test section and contained 16 thermocouples that were symmetrically positioned around
the exterior of the pipe wall and insulation as shown in Figures 2-11 and 2-12. The
symmetrical groups of thermocouples were carefully positioned so that the exterior
thermocouples that were secured to the insulation were exactly at the same angular
positions as the interior thermocouples that were secured to the pipe wall. Each set of
thermocouples were separated by an axial distance of approximately 9.0 cm and secured
to either the insulation or pipe wall using 0.25-inch Teflon tape. Inlet and exit
thermocouple probes, separated by an axial distance of 35.5 cm, were used to evaluate
the change in fluid temperature. Currently this design provides sufficient temperature
information for future analysis of the unsteady heat transfer coefficient.
Figure 2-1 1. Thermocouple arrangements along the pipe wall of the heat transfer
Figure 2-12. Thermocouple arrangements along the insulation of the heat transfer
2.6 Vapor Volume Fraction
The vapor volume fraction in the facility was computed from digital images of the
flow structure obtained during the experiment. The high resolution digital imaging
system discussed in section 2.5 was used to sequentially record images of the flow
through out the test section. These images were stored and later analyzed using the
Global Imaging Lab software to determine the flow structure and liquid height. The flow
structure was analyzed initially to select the appropriate model to determine the vapor
volume fraction. The flow structures observed during the experiments are illustrated in
Figure 2-13. Previous researchers have used various conventions to describe the names
of various flow structures observed in horizontal two-phase flows. The convention for
naming the flow regimes used in this study follows that suggested by Carey  for
horizontal two-phase flow structures.
- -. _~ Slug.'ntennittent Flow
Figure 2-13. Diagram of the various flow regimes observed.
If stratified flow is prevalent then the liquid height is measured as depicted in
Figure 2-14. Diagram of the model used for the stratified flow volume fraction
The vapor volume fraction is computed from the measured liquid height from,
a, = 1- 2 (-)hIUhh
The same relation is used to compute the vapor volume fraction for intermittent and
wavy flow using the instantaneous liquid height. When annular flow is present, the flow
structure is approximated as depicted in Figure 2-15. As shown in Figure 2-15 the vapor
core is approximated as circular in shape.
I:n: l, I
Figure 2-15. Diagram of the model used for the annular flow volume fraction
The liquid film thickness at the top and bottom of the tube are measured, and the vapor
volume fraction is computed from equation 2. 10.
= bttm o (2.10)
2.7 Vapor Quality Estimation
Direct computation of the quality using an energy conservation analysis is not
feasible because of the non-equilibrium conditions and unknown heat transfer rate from
the pipe walls to the fluid. To obtain an approximation for the vapor quality a correlation
between the vapor volume fraction and vapor quality is used. One method of relating the
vapor volume fraction and vapor quality is through knowledge of the slip velocity and
density ratio. The slip velocity is defined as the ratio of vapor velocity to liquid velocity.
The vapor and liquid phase velocities may be expressed as
G~x Gd (1 -x)
u, =ad u1 (2.11)
where G is the mass flux, u, is the vapor velocity, u, is the liquid velocity, a is the
vapor volume fraction, and x is the quality.
Substituting these relations into the slip velocity definition yields the recognizable
slip velocity expression,
Vapor volume fraction and vapor quality may be expressed in terms of the slip velocity
A limitation of the slip velocity model for estimating the vapor quality is that the
slip velocity needs to remain relatively constant. The slip velocity does not typically
remain constant over a wide range of quality and vapor volume fractions, especially
during transient flow. Consequentially another method for estimating the vapor quality
x 1- a pl
S = .
1 x a, p,
x + S (1 x
Zuber and Findlay  developed a model commonly used to correlate the vapor
volume fraction and quality. Zuber and Findlay fundamentally derived the model by
considering the local volumetric fluxes of the vapor and liquid phases and the continuity
equations of the two phases. The volume fraction may be expressed in terms of the vapor
1= c 1- x pV p~ ~216
where Co and V,, are empirical constants. The first constant Co is the distribution
parameter and accounts for non-uniform flow and concentration profiles, while the
second constant V, accounts for the effect of the local relative velocity. Klausner 
demonstrated in his horizontal flow experiments that the difference in the measured and
calculated vapor volume fraction using the Zuber and Findlay model with empirical
constants for Co and V,, of 1.0 and 0.6 respectively was within +/- 5% for over 90 % of
the data. Due to the excellent agreement with the horizontal flow experiments of
Klausner using R113 as the boiling medium equation 2. 16 is used to estimate the vapor
quality from the measured vapor volume fraction.
2.8 Data Acquisition System
A digital data acquisition system consisting of a personal computer, data
acquisition hardware, and supporting software was developed to record and condition the
analog outputs of each instrument. The personal computer was built using an AMD
Athlon XP 2200 MHz processor board combination supported with 15 Gigabits of
random access memory. A PCIM-DAS 1602/16 8-channel analog to digital board with
16-bit resolution was installed in the personal computer to acquire the data. The analog
to digital processor board was coupled to two CIO-EXP32 multiplexer boards. The
multiplexer boards interfaced with the analog to digital board allowing the system to
monitor up to 64 differential channels (32 channels per board). Each of the multiplexer
boards was constructed of two banks of 16 channels, from which a given gain setting was
selected by positioning a series of switches located on the board. As a result of this
design the banks of channels on each board were separated based on the anticipated
signal voltage and an appropriate gain setting was selected. The gain used for the
thermocouples was 100 while the gain selected for the pressure measurements was 1. An
on board junction temperature that was input to the analog to digital board through a
prescribed channel served as a reference temperature for the thermocouple
measurements. The channel selected for the junction temperature was carefully chosen to
avoid a conflict with the analog to digital board. As per manufacturer' s suggestion the
pads on the back of each channel were soldered into the circuit. The soldering pads
afforded each channel a 7Hz filtered input, open thermocouple detect, and a reference to
ground through a 100k-ohm resistor. The analog to digital board and each of the
multiplexer boards were calibrated to the manufacturer's specifications using the supplied
A computer program was developed to operate the data acquisition hardware,
condition the analog signals, and continuously monitor the system in near real time.
Using the Softwire software, a computer code was written that analyzed each of the
occupied channels on the multiplexer boards. The program sequentially sampled each of
the channels 50 times per second, storing each value in a distinct array. At the
completion of the 50th Sample, the values were time averaged and sent to an excel file for
further manipulation. The time averaging technique was employed to combat the
inherently unstable conditions prevailing in two-phase flows as well as to remove
additional unwanted noise. Prior to being transferred to excel the analog data were
converted to temperatures and pressures in the Softwire program using the
aforementioned calibration relations of Section 2.4 and the multiplexer board reference
temperatures. The processed data were then simultaneously presented through a
graphical user interface and sent to an excel file. The graphical user interface presented
all of the temperature and pressure measurements for the facility and was updated every
second. The excel file further processed the temperature and pressure data to obtain the
mass flux, mach number, vapor and liquid velocity, vapor and liquid density, and vapor
and liquid Reynolds number.
2.9 Digital Imaging Facility
An imaging facility was used to capture images of the flow regime, liquid height,
and boiling phenomena through the visual test section. The imaging facility consisted of
a Pulnix TM-1400CL progressive scan CCD camera, Data Translation DT-3145 frame
grabber card, and supporting Global Imaging Lab software. The CCD camera was
capable of image resolution of 1392 x 1040 pixels and was equipped with a Nikon 50 mm
macro lens for high magnification. The camera was connected via a Camera Link cable
to the DT-3145 frame grabber, which was installed in the abovementioned personal
computer. Image acquisition and analysis was conducted using the Global Imaging Lab
software. The Global Imaging Lab software was used to calibrate each image, from
which accurate liquid height measurements and flow regimes were determined.
CRYOGENIC CHILLDOWN VISUALIZATIONS INT HORIZONTAL FLOW
3.1 Introduction and Literature Survey
A detailed knowledge of the flow and heat transfer regimes is necessary to develop
advanced numerical models capable of predicting the flow behavior of cryogenic fluids
during chilldown. The identification and prediction of these regimes and their respective
transitions is critical to the development of a fundamental numeric chilldown model. In
general crude success has been achieved by previous researchers, which have typically
neglected flow regime transition during chilldown in their numerical analysis.
Burke et al. , Graham et al. , Bronson et al. , and Chi and Vetere 
conducted early notable chilldown experiments. The experimental chilldown studies
preformed by Burke produced a simple chill-down model. The model was based on one-
dimensional heat transfer through a pipe wall and assumed an infinite heat transfer rate
from the pipe wall to the cryogenic fluid. A crude model to calculate the chilldown time
of a pressurized cryogenic transfer line was also proposed that neglected the affects of
flow and heat transfer regime transitions. Subsequent research by Graham examined the
heat transfer and pressure drop of film boiling liquid hydrogen through a heated pipe.
They used the Martinelli parameter to correlate the frictional pressure drop and heat
transfer coefficient; however their analysis did not give special consideration to flow
regime transitions. In 1962 Bronson experimentally observed the presence of stratified
flow in cryogenic chilldown systems and recognized the importance of the
circumferential temperature gradients in the facility design. Bronson further suggested
there exists a good correlation between the early empirical oil-gas mixture flow regime
maps of Baker  to that of boiling two-phase flow of hydrogen. Chi and Vetere
experimentally studied the flow regime transition of hydrogen in a horizontal pipe. They
recorded a temperature trace from thermocouples fixed to the center outside wall of a
glass viewing section. The visual observations were correlated to measured temperature
traces and used to identify flow regime transitions. Their visual observations were in
general agreement with the sequence of flow regime transitions previously observed by
Bronson. These early experiments provided a fundamental understanding of the
chilldown process and have been useful in developing crude chilldown models.
Steward et al.  suggested an analytical chilldown model derived from the basic
conservation equations. The conservation equations were simplified by assuming a
homogeneous two-phase mixture, and one-dimensional flow. Steward considered film
boiling, nucleate boiling, and convective heat transfer regimes in his model, and used the
Kutateladze  correlation to identify the transition from nucleate to film boiling. He
also addressed the presence of surge pressures resulting from the vapor produced from
the rapid evaporation of the cryogenic fluid. Subsequent to Steward's work, Hedayatpour
and Antar  proposed a chilldown model for a vertical line with liquid nitrogen.
Hedayatpour and Antar employed a two-fluid model and analyzed four distinct flow
regimes; complete liquid flow, inverted annular film boiling flow (where a liquid core is
separated from the comparatively hot wall by a vapor film), dispersed flow, and complete
vapor flow. Their model was derived from the one-dimensional form of the single-phase
conservation equations, and a volume-averaged approach was used to address the two-
phase regimes. The results of the model compared well with the work of Kawaji and
Banerj ee [1 8,19], with the maj or deficiency being the requirement of flow regime
knowledge and quenching front speed. Additional attempts to model the chilldown of a
transfer line have been made by Cross et al. . They have attempted a similar
approach to that of Steward that incorporates a homogeneous two-phase mixture model.
The model uses a Dittus-Boelter type correction for film boiling flow and is deficient due
to oversimplifications regarding the heat transfer from the wall to the liquid.
Despite the aforementioned achievements, a more rigorous investigation of flow
regime transition is necessary to accurately forecast the chilldown process. It is the intent
of this portion of work to provide a comprehensive experimental database of visual
images and flow information for the future development of fundamental predictive flow
regime transition, pressure drop, and heat transfer relations. These relations will provide
the necessary basis to complete a comprehensive fundamental chilldown model.
3.2 Visualizations of Cryogenic Chilldown
Visualizations of the chilldown process were conducted using the camera system
described in Section 2.9. The images were recorded at one-second intervals and
subsequently analyzed to determine an approximate chilldown sequence and vapor
volume fraction. A typical chilldown experiment begins with purely vapor flow. The
vapor genesis occurs in the supply tank and facility piping prior to the visual test section.
As the liquid nitrogen contacts the hot pipe wall rapid evaporation ensues which produces
this vapor. The vapor surges through the system at high velocity and is believed to choke
at the throttling ball valve or flow meter. As the vapor continues to flow through the
facility liquid is carried with it. The liquid is observed as a stream of droplets that are
separated from the bottom of the visual tube by a thin layer of vapor. The layer of vapor
at the bottom of the tube is a result of the film boiling heat transfer that occurs in the
initial stage of chilldown. The droplets move at a high velocity and evaporate quickly as
they jet downstream. Eventually the droplets begin to coalesce and the stream of droplets
is more visible. As chilldown continues the stream of droplets is replaced by a moving
quenching front. The quenching front demarks the transition from film boiling to
nucleate boiling. The quenching front typically moves in the direction of the flow but has
been commonly observed to originate at multiple points downstream and flow
backwards. After the quenching front transverses the test section an unstable pulsating
stratified-wavy layer is observed. The liquid height fluctuates with the pulses in the
system and continues to increase as the system cools further. Small vapor bubbles
resulting from nucleate boiling are generally visible in this layer when the flow rate is not
too high. A transition to slug flow is occasionally observed following moderate growth
of the stratified-wavy layer. As the system nears chilldown the tube generally fills with
liquid and transitions to plug flow, which typically persists until shutdown is initiated. A
schematic drawing of the chilldown process is presented in Figure 3-1 and is followed by
experimental images illustrating the various phases formerly discussed. The images
presented in this document are of inferior resolution compared to the computer monitor
versions and have been enhanced with lines denoting the liquid vapor interface as
necessary. It should be noted that at high mass fluxes (generally larger than 600 kg/m2-
sec) annular flow was encountered. When annular flow was encountered, a similar
chilldown process was observed except the annular flow structure persisted throughout
chilldown. Due to the difficulty in determining the liquid height and consequentially the
vapor volume fraction, the majority of experiments were conducted in a manner to avoid
encountering annular flow. Additional changes will be made in the future to the facility
and camera system to analyze this regime and are discussed in greater detail in Section
Vapor Flowv '
Stratified Dmoplet Flow
Boiling Front Crowthl
~ i IncrLEeasing
-- lu Flow~
Fig tre3-1. Experimentally observed flow structure transitions during cryogenic
chilldown for horizontal flow.
3.2.1 Pre-Quenching Front Visualizations
The series of pictures shown in Figures 3-2 through 3-10 of this chapter were
selected from three experiments and have been arranged sequentially by increasing the
computed mass flux to illustrate the general chilldown process as well as examine the
influence of mass flux on the flow structure during chilldown. The images were recorded
during each experiment at one-second intervals and were discriminately selected based
on image quality and representation of typically observed phenomena. The instantaneous
mass flux for each image as well as the average external wall temperature is included in
the figure caption for each series of images.
The chilldown process, as previously outlined in this chapter, consists of several
distinct phases. To help distinguish the individual transitions that occur during
chilldown, a boundary has been drawn around the passing of the quenching front. The
following photographs reveal the events occurring prior to the quenching front and are
the topic of the subsequent discussion.
Prior to the passing of the quenching front a high velocity vapor flow was
observed. The vapor was generated from the film boiling of the cryogenic liquid. The
vapor flow contained droplets of liquid located at the bottom of the test section. The
liquid droplets rapidly traveled through the visual test section on top of a vapor film in
both continuous and pulsating waves. Close examination of the images in Figures 3-2
through 3-4 exemplify this phenomenon and exhibit the development of the liquid layer
on top of a very thin unsteady vapor film at the bottom of the test section. Arrows that
identify the vapor layer have been added to Figures 3-2 b, 3-3 b, and 3-4 b. The
developing liquid layer persisted during chilldown until the passing of the quenching
front. Comparison of the figures demonstrates an increasing thickness of the liquid layer
at the bottom of the test section with increasing mass flux.
,.;.*.E .I p o. :10 ,,.*..,.. ..... ;G = 2 5 kg 'm -sec I = I e
Inner H allI Flow Dir~ectioni
SExp. No. 10. G = 25 kg.'m -see t = 5' sec
Vapor' Filml Laye~r
Figure 3-2. Photographs of the unstable vapor flow that occurs at the beginning of
chilldown. (a) G = 25 kg/m2-S, T, = 8.3 o C, (b) G = 25 kg/m2-S, T,
Figure 3-3. Photographs of the unstable vapor flow that occurs at the beginning of
chilldown. (a) G = 64 kg/m2-S, T, = -12.4 o C, (b) G= 62 kg/m2-S, T, -
Figure 3-4. Photographs of the unstable vapor flow that occurs at the beginning of
chilldown. (a) G 104 kg/m2-S, Tw 7.5 o C, (b) G 115 kg/m2-S, Tw -
3.2.2 Quenching Front Visualizations
The images in Figures 3-5 through 3-7 display the progression of the quenching
front across the visual test section. The images have been enhanced with a dotted line
designating the contour of the quenching front. The quenching front was observed in all
of the experiments and varied in transverse speed, height, and incidence time according
to experimental conditions. Experiments that were run at lower mass flux rates typically
had lower transverse speeds, smaller quenching front heights, and occurred later in time
than for experiments run at higher mass flux rates. This phenomenon is easily confirmed
through visual comparison of the images of Figure 3-5 to that of the images of Figure 3-6
and 3.7. The image of Figure 3-7(a) from experiment No. 1 suggests a reversal of the
direction of the quenching front. This phenomenon was commonly observed in several
experiments and it is speculated that it is a result of liquid overshoot and system pressure
fluctuations. The general shape of the quenching front is evident in the images; however
it should be noted that a degree of distortion is present due to the refraction of the glass
and the double walled construction of the vacuum-sealed visual test section.
Exp. No. 0. G = 4 kig/inr-see t = 74 see
Innr HallFlowr Directioni
Figure 3-5. Sequential photographs of the film boiling front as it crosses through the
visual test section. (a) G = 24 kg/m2-S, Tw = -10.0 o C, (b) G = 24 kg/m2-S,
T, = -10.7 o C, (c) G = 26 kg/m2-S, T, = -11.7 o C.
Exp. No.l 10 = -1~ kg/m-'-sec I. = "5 see
Innr 1allFlow Dir~ectioni
,,,.;.... ...Exp. No. 10-..;;.... .. .. .. G = 26 kig,'m--s~e r = '6 sec
Figure 3-5. Continued.
Figure 3-6. Sequential photographs of the film boiling front as it crosses through the
visual test section. (a) G= 72 kg/m2-S, Tw -28.9 o C, (b) G 72 kg/m2-S,
Tw = -30.2 o C, (c) G = 72 kg/m2-S, Tw = -31.4 o C.
Figure 3-7. Sequential photographs of the film boiling front as it crosses through the
visual test section. (a) G= 154 kg/m2-S, Tw -26.7 o C, (b) G 143 kg/m2-
s, Tw = -28.0 o C.
Inn~er Wall Flow Directionl
.l lIhidiliilikiil sidiss~~h i~n'Lr I.:i & rai ;l: **..ta*i. .,.
Figure 3-7. Continued.
3.2.3 Post Quenching Front Visualizations
The images shown in Figures 3-8 through 3-10 have captured typical events that
occur following the quenching front. Following the passing of the quenching front there
is a noticeable change in the flow structure and liquid height. Immediately after the
quenching front had passed, nucleate boiling was observed in the growing liquid layer.
In general the higher the mass flux, the shorter the nucleate boiling regime persisted.
After the nucleate boiling had ceased a typical stratified and wavy flow structure
occurred. A clearly stratified or wavy flow structure was evidenced in several individual
photographs (Figures 3-8(a & b), 3-9(a & b), and 3-10(b)), however due to the rapid
fluctuations between the traditionally defined regimes a combined stratified/wavy regime
was generally applied to describe the flow structure. At this point in the chilldown
process additional flow structure transitions were observed in most of the experiments.
As the chilldown process progressed the liquid height in the visual section continued to
grow and the flow structure transitioned (in most cases) to slug flow followed by plug
flow. In general as the mass flux was increased the time between these transitions was
decreased. In the case of some low mass flux experiments not all of the flow regime
transitions occurred and the flow structure remained stratified/wavy for the remainder of
the chilldown process as shown in Figure 3-8.
: .: :. xp. o. .~ G = 23 kig.'nr'-sec ( = 88 sec
Trier Yal Flow Direction
xpE,~~~,.LY; ori ._7 imi-sec r = 173 see
': TInner ` H a .i;:ll :r'.I
Figure 3-8. Photographs of the various flow structures observed during the chilldown
process following the passing of the film boiling front. (a) G 23 kg/m2-S,
Tw -18.7 o C, (b) G 72 kg/m2-S, Tw -84.3 o C, (c) G 40 kg/m2-S,
Tw= -170.6 oC.
xp.~bC o. :`` G = O g n-see t = 2*4see
Inne 11al Flowr Dir~ecrion
Figure 3-9. Photographs of the various flow structures observed during the chilldown
process following the passing of the film boiling front. (a) G = 84 kg/m2-s,
=-59.8o C, (b) G = 104 kg/m2-s, =-120.90 C, (c) G = 112 kg/m2-s,
=-161.6 o C, (d) G= 126 kg/m2-s, =-175.6 o C, (e) G= 150 kg/m2-s,
"g Flow Directionl
Figure 3-9. Continued.
Figure 3-9. Continued.
-i;;;; b;r `. ;:."?l?~.i~*r."".:i~i~?~~ sli~;i~i~~Ye;~if~t~
:1 n;::Fi-irli;;;;-.i;: i: iib'~;~
Photographs of the various flow structures observed during the chilldown
process following the passing of the film boiling front. (a) G 176 kg/m2-
s, T, = -39.7o C, (b) G = 234 kg/m2-S, T, = -51.80 C, (c) G = 211 kg/m2-S,
T, = -62.5o C, (d) G= 338 kg/m2-S, T, = -178.4 o C.
3.3 Vapor Volume Fraction and Quality Measurement Results
In addition to providing a visual account of the chilldown process the images were
also used to compute the vapor volume fraction of the cryogenic two-phase flow. The
vapor volume fraction is an important measurement that is essential for estimating the
vapor quality and other flow parameters of interest. Previous research has suggested
several methods for measuring the vapor volume fraction in cryogenic two-phase flows.
The methods range from a crude approach proposed by Chi and Vetere  to
capacitance sensor studies by Willis and Smith , Killian and Simpson , and
Khalil et al.  to an elaborate radio frequency based senor approach suggested by
The approach suggested by Chi and Vetere  incorporated a rough average of the
measured temperature data from a thermocouple placed inline with the transfer fluid.
They proposed a crude relationship relating vapor volume fraction and patterned
temperature variations associated with the various chilldown phases. An improvement
over this approach was that of the capacitance sensor.
In the capacitance sensor approach capacitance probes were placed in line with the
fluid and used to accurately measure fluctuations in the dielectric constant. The changes
in the dielectric constant were then related to the void fraction and average density of the
two-phase cryogenic flows. Typically the flow regime must be known to accurately use
this approach. The limitation of this method resides in the small difference of the
dielectric constants of the two phases. The sensor is less sensitive with small differences
in the dielectric constant. Willis and Smith  and Killian and Simpson  have
successfully applied this method to measure the average density of two-phase hydrogen.
Khalil et al  has additionally reported success in measuring the average density of
two-phase helium by using an ultra sensitive capacitance circuit.
Filippov's  method incorporated a radio frequency based sensor that was used to
compute the dielectric constant of the two-phase flow. This method is similar to that of
the capacitance sensor approach with the one difference being the use of a radio
frequency sensor. The radio frequency sensor has greater sensitivity and was used to
address the very small difference of dielectric constants between the two phases. The
method is limited to stable flow structures due to the non-uniform electric field energy
distribution associated with flow structure variations. For the present study, the visual
method is used to measure the volume fraction due to its simplicity.
The results of the vapor volume fraction measurements and the vapor quality for the
previously discussed experiments are presented in Figures 3-11 through 3-13. The
figures display the raw transient vapor volume fraction measurements; a smoothed fit for
the vapor volume fraction, and the computed vapor quality. The raw vapor volume
fraction data illustrates the unsteady and highly varied flow structure during chilldown.
To capture the general chilldown trend, a smooth curve was fit to the raw vapor volume
fraction data. As discussed in Section 2.7 the Zuber and Findlay  relation is used to
estimate the vapor quality from the smoothed volume fraction data. The wavy contour of
the quality curve results from the unsteady mass flux during chilldown, which is
attributed to the considerable fluctuations in the vapor velocity measured at the venturi.
In addition to the vapor volume fraction and quality data, the figures were also
demarcated with lines indicating where the observed flow structure transitions occurred.
1.2 ~Experiment No. 101.
Vapor Flbw Stratified/Wavy Flow
1.0 0pV-~ 1.0
t~0.4 I 0.4
> 0- Vapor Volume Fraction Experimental Data
-Vapor Volume Fraction Curve FIt
0 50 100 150 200 250 300
Figure 3-11i. Graph of the vapor volume fraction and quality for experiment # 10.
Experiment No. 2
-0 Vapor Volume fraction Experimental Data
1.0 Filli~illR - -- Vapor Volume Fraction Curve Fit 1.0
0.8 - -~ 11 IIII 0.8
E Slug Fow
S0.6 -0.6 2
St atifie / Wav Flow
0 20 40 60 80 100 120 140 160 180
Figure 3-12. Graph of the vapor volume fraction and quality for experiment # 2.
Experiment No. 1
-0 Vapor Volume Fraction Experimental Data
Vapolt Flow~ --- Vapor Volume Fraction Curve Fit
1.0 3 -E~faf Qualiy Slug Flow Plug Flow 1.0
'a 0.8 I UY6lU(t-I -111 1 -0 -1= -l1 I 0.8
0.4 ~~Stratife- -0.
> Wavy low
0.2 1111 11 1 1 1 1 0.2
0.0 ... I .. .. m m r. ..E: 0.0
0 20 40 60 80 100 120 140
Figure 3-13. Graph of the vapor volume fraction and quality for experiment # 1.
3.4 Discussion of Results
The visual observations of the cryogenic chilldown experiments have confirmed and
improved upon the previous work of Bronson et al.  and Chi and Vetere  regarding
the general flow structure and heat transfer sequence. As noted in the literature review
section of this chapter, Bronson and Chi and Vetere both conducted flow visualization
studies of cryogenic chilldown. Bronson's facility was operated at pressures as high as
6900 kPa with mass fluxes up to 3100 kg/m2-Sec, while Chi and Vetere, who did not
report an operating pressure, examined mass fluxes of 0.25 kg/m2-Sec. Bronson reported
the chilldown process to begin with gas being vented from the system, followed by the
presence of droplets of liquid cryogen that were transported through the length of the
visual section before the line was cooled down. Pulses of liquid, which flashed into gas,
then reportedly followed. Subsequently an advancing front of stratified flow that rapidly
transitioned into a slug flow phase emerged. The slug flow phase persisted until the slugs
gradually coalesced filling the visual section with liquid. Chi and Vetere reported a
chilldown sequence similar to that of Bronson with additional consideration for the
prevailing heat transfer regimes. Chi and Vetere reported the existence of film boiling
that persisted through the mist flow and slug flow regimes. The film boiling was
reportedly followed by a short transitional boiling phase, which was then followed by
nucleate boiling. The results of the current research differ moderately with the general
descriptions of the chilldown events offered by Bronson and Chi and Vetere. The current
research offers much greater detailed description of the chilldown sequence and
occurrence of heat transfer regimes. In the current research, the early phase where pulses
of liquid, which flashed into gas, described by Bronson were not specifically observed.
Additionally, the current research did not observe the film boiling heat transfer
phenomena to continue throughout the slug flow regime as indicated by Chi and Vetere.
Instead film boiling was observed to occur in the initial stages of chilldown and subsided
following the passing of the quenching front. Furthermore no transitional boiling stage
The vapor volume fraction and quality curves of the experiments appear to be in
general agreement with the visually observed chilldown trend. The unstable nature of the
chilldown process is also evidenced in the plots of the vapor volume fraction data. The
data reveal the wide fluctuations and instability of the chilldown process. Despite these
wide fluctuations a general trend was observed and smoothed. The graphs also indicate
an approximate area where the flow regime transitions were visually observed. The
observed flow regime transitions appear to occur earlier in time and are typically shorter
in duration as the mass flux of the experiments increased. This general trend was
observed in the majority of experiments with a single noteworthy exception. In one
particular experiment flow instability occurred, which resulted in a deviation from the
general chilldown trend. A plot of the vapor volume fraction data from this experiment is
presented in Figure 3-14 for the purpose of comparison.
Experiment No. 11
0 50 100 150 200 250
Figure 3-14. Graph of the vapor volume fraction data for experiment # 1 1.
It is believed that this digression from the typical chilldown sequence is a result of a
Ledinegg-instability. When a Ledinegg-instability occurs the flow can exhibit multiple
states and is subj ect to an oscillatory instability referred to as a density wave oscillation.
The density wave oscillation is caused by the delay in the system due to the fluid
transport time. This experiment was the only record of such an occurrence and future
investigations into the frequency and general behavior of such an event have yet to be
3.5 Sources of Experimental Error
The image analysis procedure is subj ect to moderate uncertainty due to the highly
unstable nature of the flow structure during chilldown, the image capture rate of the
camera system, and qualitative and subjective human bias. During the chilldown
experiments the liquid height in the visual test section fluctuates considerably. This
constant variation in the liquid height throughout the visual test section limits the
accuracy of the liquid height measurements. In addition to the variation in the liquid
height, the actual flow structure also can inhibit the accuracy of the measurements.
Highly unstable flow structures may contain entrained vapor and/or a wavy interface
within the image. In these situations an approximate interface line was assumed from
which several liquid height measurements were averaged. Two people, one of whom had
limited knowledge of the experiment, analyzed the images from several experiments.
The results of each analysis were compared in an attempt to quantify this error. It was
found that on average the selected liquid heights identified by each examiner were fairly
similar with deviations typically around 1.0 millimeter or less. The largest variations in
measurements came during the beginning of the experiments where the interface was not
very clear. The largest deviation reported during the initial part of chilldown was
approximately 1 mm while during the later stages the error was much less than 0.5 mm.
As result of the squaring of the height term in the vapor volume fraction computation the
larger error in the beginning of the experiment resulted in an uncertainty of
approximately 1.5 % while the error in the later stages of chilldown where the interface
resolution was greater produced an approximate error of 7 %. This experimental
uncertainty is reasonable and compares to the reported uncertainty from the previously
mentioned methods for determining the vapor volume fraction. Additional uncertainty
arose from the sequencing of the camera rate and the data acquisition program.
During the chilldown experiment the data acquisition system was initiated first and
followed by the camera system. The two independent programs were both selected to
record data at one-second intervals however; a varied lag often persisted between the two
systems. The lag was on the order of one second and could not effectively be accounted
for. In addition to the lag during startup it was found that the camera system often
recorded the images at a rate slightly faster than one second. The total variation in time
resulting from this deviation was typically less than a second for an entire experiment and
thus it was considered negligible. Exhaustive efforts that were made to remedy these
variations were met with limited success, and it was concluded that the resultant error
was comparatively small and tolerable. In addition to the camera sequencing error it was
also observed that the transitioning flow structures did not precisely occur. For example,
on occasion it was observed that wavy flow may briefly transition to slug flow and
quickly appear to transition back to wavy flow. To account for the wavering
transitioning behavior during chilldown as well as any error in precisely identifying a
single transitioning event a general transition range was applied in regards to flow
structure changes. The transitioning range varied by experiment and typically spanned
several seconds in duration. It is also recognized that the identification of flow regimes is
subj ect to the qualitative assessment of the images.
The image analysis procedure was a time consuming process by which an individual
must continually make an educated judgment concerning the precise identification of the
liquid level height and flow
WALL TEMPERATURE PROFILES AND CHILLDOWN TIME IN CRYOGENIC
4.1 Introduction and Literature Survey
The ability to accurately predict the chilldown time and temperature profile of
cryogenic transfer lines during the initial cooldown phase is of significant design and
operational importance to the cryogenic industry. Knowledge of the temperature
gradients, temperature history, and chilldown time will allow practitioners to design safer
transfer systems optimized for efficiency. A number of previous studies have been
conducted regarding the chilldown process, which have produced several general
chilldown models. These efforts have been collectively addressed in the introductory
chapter as well as Section 3.1, and thus only a brief overview of their contributions
pertaining to the thermal aspects of chilldown is discussed.
An early approach suggested by Burke et al.  applied an energy and mass balance
to a defined control volume that included the outer wall of the transfer line. The model
assumed an infinite heat transfer coefficient between the transfer line wall and the fluid
and consequentially underestimated the chilldown time of the system.
Subsequent work conducted by Bronson et al.  was done to investigate general
chilldown concerns relating to the development of nuclear powered rocket engines. The
research focus included studies of circumferential temperature gradients and chilldown
times. Bronson assumed a uniform temperature difference between the pipe wall and
fluid temperature of 100K. Solving a heat balance the pipe temperature was determined
through trial and error until convergence was reached with the known exit gas
Steward et al.  proposed another simplified model for chilldown time. Their
model identified the resistance to flow of the vaporized liquid as a critical factor in
An investigation was also conducted by Chi . In his examination of hydrogen
chilldown it was observed that a film boiling front was present when the wall to fluid
temperature difference exceeded 18.50F. From this observation he reasoned that the heat
transfer from the pipe to the two-phase fluid must first be transferred through the vapor of
the film boiling front and therefore a pseudo single-phase gas heat transfer relation was
assumed. Combining a relation for the heat balance of the fluid with a relation for the
local heat flux of the gas phase, Chi derived a relationship for the chilldown time. Chi's
model was an improvement over the previous models but still lacked consideration of the
two-phase nature of the fluid, wall temperature gradients, fluctuations in the flow rate,
and heat transfer from the surroundings.
It is recognized that these additional considerations dramatically increase the
complexity of the problem and are not amenable to closed-form analytical solutions. It is
the intent of the current research to examine these considerations and provide a sufficient
experimental database of chilldown information for the development of advanced
chilldown numerical models.
4.2 Wall Temperature Profile and History
To measure the wall temperature profile and history a section of the transfer pipe was
externally instrumented with a series of eight thermocouples as shown in Figure 4-1. The
thermocouples were arranged circumferentially in two groups of four around the external
surface of the transfer line approximately one meter down from the center of the visual
test section. The thermocouple groups were spaced approximately 8 cm apart and
insulated with Melamine foam insulation. Measurements were recorded by the data
acquisition system at a frequency of 50 Hz and averaged into one-second increments.
The chilldown times, experimental mass flux ranges, quenching front observation times,
quenching front speeds, and the approximate Leidenfrost temperatures are tabulated in
Table 4-1 and followed by the individual temperature and mass flux histories for the first
10 experiments, shown in Figures 4-2 through 4-21. Demarcations on the temperature
histories have been added to identify flow structure transitions, the approximate
Leidenfrost point, and the visually observed transition from film boiling to nucleate
boiling. The visually observed transition from film boiling to nucleate boiling is
designated with an oval on each of the temperature histories. The Leidenfrost
temperature was identified on the temperature plots by the rapid drop in temperature
associated with the transitioning from film-boiling to nucleate boiling. The Leidenfrost
temperature is denoted with a dotted line and the approximate transition point was
selected based on the temperature history of the thermocouples located on the bottom of
the transfer line.
Table 4-1. Summary of the chilldown data from the temperature histories.
Experiment Mass Flux Approximate Approximate Quenching Approximate
Number Range Quenching Chilldown Front Leidenfrost
(kg/m2-Sec) Front Time ** Observation Temperature
Velocity (sec) Time ** (og)
1 125-250 0.052 71 32 -146
2 50-90 0.021 110 43 -144
3 40-120 0.016 145 43 -146
4 200-260 0.062 46 27 -144
5 200-400 0.047 51 24 -148
6 200-350 0.056 54 28 -152
7 125-250 0.047 75 28 -152
8 150-270 0.032 95 32 -148
9 30-60 0.016 197 38 -156
10 20-50 0.019 245 50 -144
** The reported times are referenced to ti, the time at which chilldown was initiated.
Schematic representation of the thermocouple arrangement on the stainless
steel transfer line.
Experiment # 1
Stratified/ I Slug Plug
ti' -*--6 I I I Temp 1 (Top)
-A- Temp 2 (Bottom)
-m Temp 3 (Side)
-* Temp 4 (Side)
-v- Temp 6 (Top)
-0-- Temp 7 (Bottom)
-0- Temp 8 (Side)
SApprox Leidenfro it Tem~p
Experiment # 1
Stratified/ I 31ug Plug
0 20 40 60 80 100
Figure 4-2. Temperature history during chilldown for experiment # 1.
0 20 40 60 80
Figure 4-3. Mass flux history for experiment # 1.
Experiment # 2
Stratified/ Wa y IS~lug
-*- emp 1 (Top)
t _- Temp 2 (Bottom)
-m Temp 3 (Side)
-*- Temp 4 (Side)
-o- Temp 5 (Side)
-v Temp 6 (Top)
-0-- Temp 7 (Bottom)
-0- Temp 8 (Side)
Appro~x Leident ost fTemp
Experiment # 2
S ratified / Wak II Slug
0 20 40 60 80 100 120 140 160 180
0 20 40 60 80 100 120 140 160 180
Figure 4-4. Temperature history during chilldown for experiment # 2.
Figure 4-5. Mass flux history for experiment # 2.
Experiment # 3
0 50 100 150
Figure 4-6. Temperature history during chilldown for experiment # 3.
Experiment # 3
0 50 100
Figure 4-7. Mass flux history for experiment # 3.
Experiment # 4
I tratified /I Wavy I`l Pu
0 20 40 60 80 100 120
Experiment # 4
0 20 40 60 80 100
Figure 4-8. Temperature history during chilldown for experiment # 4.
Figure 4-9. Mass flux history for experiment # 4.
Experiment # 5
Satisfied / Wavyl Slug I Plug
0 20 40 60 80 100 120 140 160
Mass flux history for experiment # 5.
Experiment # 5
0 20 40 60 80 100 120 140 160
Figure 4-10. Temperature history during chilldown for experiment # 5.
Experiment # 6
0 20 40 60 80 100 120
Temperature history during chilldown for
experiment # 6.
Experiment # 6
0 20 40 60 80 100 120 140
Mass flux history for experiment # 6.
Temperature history during chilldown for experiment # 7.
Experiment # 7
Stratifi~ec / Wavy Slug
Experiment # 7
0 50 100 150
Mass flux history for experiment # 7.
Experiment # 8
St ~atified /Wavy I Slugl Plug
-*- Temp 1 (Top)
ti "Ela Temp 2 (Bottom)
-m Temp 3 (Side)
-*- Temp 4 (Side)
-o- Temp 5 (Side)
-v Temp 6 (Top)
-0-- Temp 7 (Bottom)
-0- Temp 8 (Side)
Approx Leiden~frost Temp'
Temperature history during chilldown for experiment # 8.
Experiment # 8
0 50 100
Mass flux history for experiment # 8.
Experiment # 9
0 50 100 150 200 250
Temperature history during chilldown for experiment # 9.
Experiment # 9
Str utified / Wvy
0 50 100 150 200 250 3C
Mass flux history for experiment # 9.
Experiment # 10
0 50 100 150 200 250
Temperature history during chilldown for experiment # 10.
Experiment # 10
Stratified / W avy
0 50 100 150 200 250 3C
Mass flux history for experiment # 10.
The outside wall temperature profies have provided information pertaining to the
chilldown behavior of the facility. The profiles provide specific information regarding
the chilldown time, approximate Leidenfrost temperature, heat transfer transitions, and
the circumferential temperature gradients in the transfer pipe.
The chilldown time for each of the experiments was determined from the
temperature plots. Chilldown was defined to be complete when the wall temperature
measurements leveled out at the approximate saturation temperature of the cryogenic
nitrogen fluid. The chilldown time for each of the 10 experiments as well as the mass
flux range is tabulated in Table 4-1. Results from the experiments support the intuitive
relationship between an increased mass flux and lower chilldown time. Experiments 9
and 10, shown in Figures 4-18 through 4-21, clearly illustrate the effects of low mass
flux. Figures 4-18 and 4-20 illustrate the more gradual and extended temperature history
as well as the lack of flow structure transition beyond the stratified/wavy regime. The
chilldown data from these 10 experiments will be used in future research to compute the
transient heat transfer coefficient. In addition to the chilldown time, the temperature
histories also provided information about the approximate Leidenfrost temperature.
The very high temperature difference, to the fluid from the pipe walls has been
observed to cause film boiling of the liquid cryogen during the initial phases of
chilldown. During film boiling, heat is transferred to the liquid by conduction through
the vapor film and by radiation from the hot pipe wall. The film boiling flow structure
prevails until the wall temperature of the transfer lines has chilled to the Leidenfrost
temperature. Once the pipe has chilled below this temperature, a moving quenching front
is observed briefly and is followed by nucleate flow boiling. This transition is identified
on the temperature profies by a sharp decrease in temperature. This sharp drop in
temperature is used to estimate the approximate Leidenfrost temperature. A diffusion
time of approximately one second, corresponding to the thermal diffusion through the
pipe wall, was computed for the steel transfer line at standard conditions. This time is
relatively small and thus the thermocouple measurements should have good response to
the heat transfer transients, and consequently the identification time for the Leidenfrost
temperature should be reasonably accurate. The approximate Leidenfrost temperatures
for the 10 experiments are tabulated in Table 4-1, and an average of the results indicate
an average Leidenfrost temperature of approximately -1480C. It should be noted that a
large discrepancy in time exists between the instant the quenching front was visually
observed and the sharp decrease in the temperature identified in the plots. This
discrepancy is largely attributed to the slow movement of the quenching front and the
large separation of the thermocouples from the visual test section. For example, with a
quenching front speed of 0.05 m/s, the delay time for the quenching front to pass from the
visual section to the heat transfer section is on the order of 20 seconds.
In addition to the approximate Leidenfrost temperature the general shape of the
temperature histories reveal three distinct heat transfer modes. In the beginning of the
experiments the fi1m-boiling heat transfer regime was evident as previously discussed.
Following the fi1m-boiling regime the nucleate flow-boiling regime is observed, and
followed by evaporative convective heat transfer. In each of theses regimes the dominant
heat transfer mechanisms are varied. To develop precise predictive models these heat
transfer transitions must be accounted for. It is interesting to note that during the
majority of the chilldown, the heat transfer regime is film boiling. This is why the crude
chilldown models have been able to grossly predict the chilldown time. The results show
that experiments conducted at high mass fluxes had faster transitions between each of
these regimes, indicating that the chilldown process is strongly influenced by the
convective fluid velocity. In addition to knowledge regarding the heat transfer regimes
during chilldown, the temperature measurements also have provided quantitative
knowledge of the temperature gradients in the transfer line during chilldown.
Bronson et al  previously reported moderate circumferential temperature
gradients. Their work recognized the presence of large gradients within horizontal
transfer lines due to the stratification of the cryogenic fluid. The results of the 10
experiments confirm not only the presence of large circumferential temperature gradients
but also the existence of small axial temperature gradients during the quenching front
phase. The magnitudes of the circumferential temperature gradients are primarily
influenced by surface wetting, and thus highly influenced by the flow structure. The
largest temperature gradients were observed in experiments that had both a
stratified/wavy flow structure and low liquid level height. In these experiments the low
liquid height was not sufficient to wet the surface in the vicinity of side-placed
thermocouples and thus the heat transfer at these regions was smaller then at bottom of
the transfer line. This corresponds to a more distinguished circumferential gradient.
Noticeable examples of significant circumferential temperature gradients were evidenced
in experiments 1, 2, 3, 4, and 8, shown in Figures 4-2, 4-4, 4-6, 4-8, and 4-16. The
general observations regarding the influence of the flow structure as well as the measured
results pertaining to the temperature gradients will be incorporated into the future
development of numerical heat transfer models for cryogenic transfer systems.
Overall the general shape of the chilldown profies for each of the tests are in
good agreement with the experimental results previously reported by Bronson et al ,
Steward et al , and Chi . The temperature profies have revealed a strong
association between higher mass flux and lower chilldown time, identified an
approximate Leidenfrost temperature of -148oC, identified three distinct heat transfer
regime transitions, and confirmed the existence of large circumferential temperature
gradients as well as small axial gradients during the transition from film boiling to
nucleate flow boiling. This data will be useful in developing and confirming future
numerical chilldown models.
4.4 Sources of Error
As evidenced in all of the transient temperature profiles, the visual observations of
the film-boiling fronts did not coincide with the transition to nucleate boiling displayed in
the temperature profiles. The disparity in time between the visually observed boiling
front passage and the Leidenfrost point suggested by the temperature profiles occurs due
to the separation of the thermocouples from the visual test section. This suggests that
there is a possibility that the flow structure transitions occurring in the steel transfer line
may vary from what is visually observed in the glass viewing section. Future
investigation is required to evaluate the significance of the error.
CONCLUSIONS AND RECO1V1VENDATIONS
In this work, an experimental investigation was conducted to gather detailed
information regarding the flow structure, flow properties, and heat transfer mechanisms
associated with cryogenic chilldown. An experimental facility was constructed and used
to gather visual images of the flow structure and heat transfer mechanisms as well as
temperature and pressure data. This chapter lists the significant accomplishments and
Endings of this experimental investigation and concludes with suggestions for future
5.1 Accomplishments and Findings
The research presented in this document has enhanced the fundamental
understanding of cryogenic chilldown as follows:
* 1. Visualizations of the entire chilldown process were obtained throughout a range
of mass fluxes. The visualization revealed the detailed flow structure and heat
transfer transitions that occur during chilldown. Images captured the transition
from film boiling to nucleate boiling, revealing the quenching front shape, height,
and approximate transverse speed. The detailed knowledge gained of the flow
structure and heat transfer transitions will be incorporated in the future
development of an advanced fundamental numerical chilldown model.
* 2. Wall temperature profies and chilldown time data were also obtained. The wall
temperature profiles helped verify the existence and magnitude of circumferential
and small axial temperature gradients in the transfer line during the various phases
of chilldown, as well as determine the approximate Leidenfrost temperature of the
nitrogen. This information will be useful in analyzing current and developing
future numerical chilldown models capable of predicting the chilldown time and
temperature gradients of transfer systems.
5.2 Recommendations for Future Research
The current research was conducted to gather a detailed understanding of the
mechanisms of cryogenic chilldown with the ultimate goal of developing an advanced
mechanistic based chilldown model. Several important issues remain unresolved and are
suggested for future research:
* 1. During the experiments annular flow was observed at sufficiently high mass
flux rates (typically greater than 600 kg/m2-Sec); however due to current limitations
of the camera system and visual test section, a distinguishable liquid/vapor
interface at the upper portion of the test section could not be easily identified.
Capturing the liquid-vapor interface in annular flow was particularly difficult
because of the rapid flow rate and the distributed flow around the inner wall of the
tube. The distributed flow added additional distortion and concealed the internal
liquid-vapor interface. To capture these images the camera system needs to be
enhanced by changing the lighting setup, lens configuration, and shutter time.
Consideration of annual flow is critical to developing a comprehensive
understanding of chilldown and this issue should be initially pursued in future
* 2. Due to an oversight in the data acquisition setup, the measured pressure drop
across the test section was not accurate. The oversight has been corrected, however
future experiments must now be run to compile a database for a range of flow
conditions. The data should then be compared to existing pressure drop models
and if necessary a new model should be derived.
* 3. The flow structure has a large impact on the thermal behavior of the system
during chilldown and is of considerable importance. To consider the flow structure
in a numerical model for chilldown, a method for predicting flow structure
transition is required. Future research should address the development of a flow
regime map and a method for predicting flow regime transitions. The current data
should be initially compared with the flow regime maps of Steiner , Kattan-
Thome-Favrat [26,27,28], Baker , and Taitel and Dukler .
* 4. Additional research also needs to be conducted to identify the heat transfer
coefficient for the transfer line and develop a method of computing the inside wall
temperature. An inverse heat transfer approach should be useful for this purpose.
PROPERTIES OF NITROGEN
The relations presented in this appendix were used to calculate the relevant physical
and thermodynamic properties of both liquid and gaseous nitrogen. All of the properties
except vapor density were generated from the NIST fluid thermophysical properties
database . The saturation properties are valid for -210 < T < -147 oC. All
temperature and pressure values are reported in oC and kPa respectively.
A.1 Property Relations
1. Saturation Temperature, Ts, (P):
T ,t(OC) = 1* 10-' P3 + 0.0082 *P2 +1.7091*~P +119.51 (A.1)
2. Saturation Pressure, Ps, (T):
Ps, (kPa) =1*10-o' T3 + 0.0075* YT2 +1.5793 T +111.44 (A.2)
3. Density, liquid, p,
p, (kg/m3) = 0.0787 T2 33.907 T 2820.6 (A.3)
4. Density, vapor, pg:
p, (kg/m3) = (Pexit + 101)/ (.2968 (Texit + 273)) (A.4)
5. Liquid Dynamic Viscosity, #u:
ilL (Pa~s)= (-4*10-09)* T3- 2*1 0-06)* T2 -0.0004* T -0.0233 (A.5)
6. Vapor Dynamic Viscosity, pu,:
pR (Pa*s) = (5*10-aS)* T + 210-o (A.6)
EXPERIMENTAL CRYOGENIC CHILLDOWN VISUALIZATIONS
A digital copy of all of the images and experimental data may be attained by request
Dr. James F. Klausner
237 Mechanical and Aerospace Engineering Building B
PO Box 116300
Gainesville, FI 32611-6300
SExperiment# 1 (05-06-04)
t = 56 sec
I = of sec
* = 'Ih Ein
t =36 sec
1 = oo sec
I= I sec
Figure B-1. Chilldown images from experiment # 1.
t = 81sec
t = 106 sec
t 121 sec
t = 116 sec
Figure B-1. Continued.
SExperiment # 2(04-1 5-04)
t = 37 ~pr
t = 37 sec
I t = 42 sec
St = 62 sec IIt = 67 sec
Figure B-2. Chilldown images from experiment # 2.
t = 72 sec
SExperiment #2 (04-15-04)
t= 102 sec
t =117 sec
L -= 1ZZ sc
Figure B-2. Continued.
SExperiment # 2(04-1 5-04)
Figure B-2. Continued.
SExperiment #3 (04-26-04)
t =40 sec
t = 55 sec
t = 70 sec
t = 85 sec
t = 65 sec
t = 80 sec
t = 95 sec
t = 75 sec
t = 90 sec
t =45 sec
t = 60 sec
Chilldown images from experiment # 3.
SExperiment #3 (04-26-04)
t = 100 sec
t =115 sec
t= 130 sec
t =145 sec
t =110 sec
t =125 sec
t =105 sec
t =120 sec
SExperimnent #3 (04-26-04)
t = 170 sec
t =185 sec
Figure B-3. Continued.