Flows: Pressure Drops and Void Fractions in Horizontal Two-Phase Potassium
Flows: Pressure Drops and Void Fractions in Horizontal Two-Phase Potassium
Flows: Pressure Drops and Void Fractions in Horizontal Two-Phase Potassium
OF
Potassium
and RICHARD E. BALZHISER
Vapor-liquid flow of a potassium-8% sodium mixture was studied with the use of o boiling
heat transfer test loop. Pressure drop data were obtained from a 36-in. long, 0.495-in. I.D.
unheated test section. Void fractions were measured at the midpoint of the test section with
gamma ray attenuation.
Flows were essentially frictional. Quality, total flow rate, and absolute pressure were the
principal variables affecting the pressure drop results, which were correlated in terms of a
two-phase friction factor. A metallic void fraction correlation was developed from previously
reported data for other metollic fluids together with the potassium data of this study.
The potassium two phase friction factors fall substantially lower than volues predicted by wellknown frictional pressure drop correlations. Metallic void friction values are appreciably lower
than data for other fluid systems. Velocity slip ratios appear to be much higher than for other
fluid systems due to high liquid-to-vapor density ratios and low void fractions.
In recent years metallic fluids have received consideration as possible heat transfer media in space electrical
power generation cycles which involve boiling and condensing of the carrier fluid. The rigorous designs required
in such applications necessitate the ability to make highly
accurate predictions of two-phase flow phenomena. Although a voluminous literature exists on the subject of
two-phase fluid flow, little has been reported on flow of
metallic systems. The two-phase pressure drop along a
tube in general is the summation of losses due to friction,
acceleration effects, and hydrostatic head. Two-phase values usually are greater than those experienced in singlephase flows with comparable fluid throughputs. Frictional
losses are always present and occur for any orientation of
the flow channel and for adiabatic or heated conditions.
Acceleration losses occur in forced-circulation boiling
flows, where the continuous phase chan e causes the mixture density to vary along the tube. Suc losses also occur
in adiabatic flow at low pressure levels. Hydrostatic head
terms are present only for vertical or inclined flow systems.
Knowledge of the mean two-phase mixture density is
necessary in prediction of acceleration losses and hydrostatic head. In vapor-liquid flow, the mean velocities of
the two phases, based on the cross-sectional area of each,
generally are not equal. Because of this slip, the true
fraction of the pipe cross section occupied by either phase
differs from that calculated on the basis of the volumes of
gas and liquid entering the tube. As a result, the mean
mixture density cannot be calculated on the basis of quality (vapor mass fraction) alone, but requires a knowledge
of the void fraction (fraction of the channel cross section
occupied by vapor). Void fraction, then, is an important
parameter in predicting the hydraulics of many two-phase
flow systems, since without it the accelerative and hydrostatic contributions to the pressure drop cannot be evaluated.
In this study, potassium two-phase pressure drop and
void fraction data were obtained from a boiling heat transfer test loop. Two-phase mixtures were generated in a
preheater section, and the pressure drops were measured
over a horizontal 3-ft. length of 0.495-in. 1.D. tube. Mixture qualities at the ends of this test section were obtained
by heat balance. Void fractions were measured at the
middle of the pressure drop section using single-shot
gamma ray attenuation.
The potassium test loop included a heat transfer test
section, a double-pipe type of exchanger, in which condensing sodium on the shell side transferred heat to twophase potassium flowing on the tube-side. Prior to making
experimental runs a leak developed in this test section,
allowing sodium to enter the potassium system. Two-phase
flow runs were not prevented but the data obtained are
for flows of a potassium-sodium mixture containing 8%
sodium by weight. Chemical analysis of fluid samples
showed that this composition was constant over the period
during which the two-phase flow data were obtained. In
this paper the fluid is referred to as potassium. A study of
the data indicates that the results closely approximate the
two-phase flow behavior of pure potassium (1) .
Page 50
PRESSURE DROP
TEST SECTION \T
~-T%i
SCINTILLATON
DETECTOR
UGAMMA
RAY
1$31
SOURCE
HEAT TRANSFER
TEST SECTION
CONDENSER
DIFFERENTIAL
PRESSURE
TRANSMITTER
SUB-COOLER
THROTTLE VALVE
A.1.Ch.E. Journal
January, 1966
EXPERIMENTAL
The Potassium Test Loop
Two-phase flow studies were conducted in a specially instrumented test section incorporated within a forced-circulation
heat transfer loop. The loop was designed for use with
molten alkali metals as the circulating fluids. A flow diagram
of the system is shown in Figure 1. Liquid potassium flowed
from the electromagnetic pump through the throttle valve,
the magnetic flowmeter, and then entered the preheater where
vapor was generated. The resulting two-phase mixture passed
through the heat transfer section and the horizontal pressure
drop test section. Flow continued through the condenser and
into the hot well from which liquid was returned to the pump
via the subcooler. The circuit was equipped with a liquid
metal supply tank as well as valves and auxiliary lines for
bleeding, pressurizing, venting, levelling, and draining.
The pump was a Mine Safety Research Style I1 electromagnetic conduction pump. Liquid flow rate was controlled
by varying the applied voltage to the pump or by regulating
the throttle valve, which was a stainless steel bellows sealed
ball valve. The flowmeter was a Mine Safety Research Style
FM-2 magnetic flow meter. Accuracy of measured flow rates
was found to be better than 2.8% ( 1).
The preheater consisted of three parallel 45-in. lengths of
4-in. O.D. x 3-in. I.D. Haynes-25 pipe heated by electrical
radiant heaters. The total rated power was 30 kw., and values
were measured to within 0.1 kw.
The heat transfer section was a double-pipe heat exchanger
designed for condensation of sodium vapor in the annulus
with two-phase potassium flowing in the central tube. The
details of the mechanical design of this section are furnished
by Balzhiser et al. (2). As mentioned previously, the thin
inner wall of this section developed a leak, preventing experimental use of the section during two-phase flow runs.
The condenser and subcooler each consisted of 10 ft. of
stainless steel finned tube over which air was passed at rates
up to 800 std. cu. ft./min. The subcooler's specific function
was to help prevent pump cavitation by subcooling the liquid
flowing from the hot well. Functions of the hot well were
to assure a liquid level above the pump intake even in the
event of large flow rate instabilities, to act as an accumulating
chamber for system cover gas (argon), and to act as an expansion tank for liquid.
Four thermocouples (TC 25, TC 26, TC 30, TC 31 ) were
vital to the establishment of a loop heat balance. These
platinum-platinum 10% rhodium thermocouples were calibrated against a National Bureau of Standards secondary
standard and provided measurements accurate to within 1"F.
Standardized thermocouple lead wire connected all thermocouples to a potentiometer selector switch, and all were electrically insulated from the loop. About forty other chromelalumel thermocouples were used to monitor loop operation.
All piping was traced with electrical warmup heaters, and
a zirconium chip hot trap was employed to maintain low
oxygen concentraticn. The loop was remotely operated from
a control panel which contained several automatic safety
features in addition to operating controls. Smith ( 1 ) provides
a detailed description of the potassium loop.
For each pressure drop and void fraction value, the fluid
quality was calculated at both pressure taps by a heat balance.
Equilibrium was assumed to exist between liquid and vapor
at both points. Heat loss data, essential for quality calculations, were obtained from a series of all-liquid runs. The heat
losses from the preheater, heat transfer section, and pressure
drop section were essentially independent of flow rate and
were correlated as functions of fluid temperature ( 1 ).
The theoretical basis of the gamma ray attenuation technique for measuring void fractions has been reviewed elsewhere (1, 3 ) . Although the mean two-phase fluid density in
the region traversed by the radiation beam is sought, this
value is not directly obtained in practice due to lack of appropriate absorption coefficients. The usual procedure is to
obtain the void fraction by interpolation between detector signals read for single-phase liquid and vapor flows. For the
probable flow regimes encountered, it can be shown from
Beer's Law that ( 1 , 3 )
In ( N / N z )
a=
(1)
In ( N , / N I )
SOURCE
A.1.Ch.E. Journal
HAYNES -25
TUBING
HIGH VOLTAGE
TO TUBE
PREAMPLIFIER
Page 51
counts/min). The preamplifier was built to match the electronic characteristics of the scaler.
The source was positioned under the flow tube, the radiation
passing upward through the pipe and fluid. The detector was
mounted on top of the pipe, and the two components were
accurately aligned on a common frame, the position of which
was adjusted to align the gamma ray beam with the center
line of the flow tube (1). The beam was collimated to a
diameter of about 1/32-in. at the source. Radiation reached
the detector via a 0.75-in. diameter hole in a 2.75-in. thick
steel detector face shield. The source-to-detector distance was
18.25 in.
Void fractions were measured with a "single-shot'' procedure. To obtain a value, the count rate was obtained from
the collimated beam. Then the beam was "shut off by substituting a 1.5-in. thick steel plug for the collimator, allowing
reading of the background count rate due to scatter. Both
count rates were corrected for coincidence loss ( 1 ) and then
subtracted to give the value of N for Equation ( 1 ). Flow rate
and thermal data were obtained simultaneously with the void
fraction measurements.
RESULTS AND CORRELATION OF DATA
f = f(x)
(2)
where
Pr
f=
n(AP/AL)gc
G"
(3)
Figure 3 presents a plot of the pressure drop data according to the relationship suggested by Equation ( 2 ) .
The least-squares line through the data is given by
= 0,0138 x'."
(4)
226
0.054 to 3.10
108 to 752
0.54 to 15.8
923 to 1428
0.0004 to 0.3788
Average quality
0.0065to 0.3784
-0.0051 to $0.0312
Uncertainty
............
0.05 (maximum)
3. % (maximum )
0.03 (max. at 1,000"F.)
4 (max. at 1,000'F.)
Depends on quality level ( 1 )
0.003 at x = 0.05
Depends on quality level ( I )
0.003 at x = 0.05
Depends on quality level ( 1 )
0.003 at x = 0.05
Void fraction
Page 52
17
131 to 749
879 to 1320
0.0172 to 0.1647
0.153 to 0.860
A.1.Ch.E. Journal
Uncertainty
............
3.% (maximum)
4 (max. at 1,00O0F.)
January, 1966
For flows where the liquid and vapor each exhibit superficial Reynolds numbers in the turbulent range (tt flow
type), X may be approximated by (12)
AVERAGE QUALlTY, x
Fig. 3. Two-phase frictional pressure gradient correlations for potassium and comparison with other correlations.
A.1.Ch.E. Journal
Page 53
For about 60% of the pressure drop data, the temperature decrease across the test section was greater than
15"F., and the mixture quality increased. In many cases
the quaIity increase was between 0.01 and 0.02, representing a significant fraction of the test section average
quality. I t was previously stated that the two-phase pressure gradient correlation, Figure 3, is for frictional pressure losses. This statement demands justification in view
of the possible acceleration components which could accompany the quality increases cited above.
For any flow system in which a change of phase occurs,
the differential pressure loss is usually considered to be
the sum of components due to friction, fluid acceleration,
and hydrostatic head. Since the potassium flows were
horizontal
-dP
= -dP,
+ p -vdv
-
g.
ing
-dp=-@,---
G' d;
( 10)
g c
I .c
MEmLL/C DATA
METALLIC DATA
- LEAST SQUARES
0.6
04
0.2
0.I
0.06
0.04
0.02
i
i
I
I
Tang,
Tang,
K-Hg Amalgam
Hg
I l k
0.0 I
QOOE
-r-
0-004
0.002
0.00I
0.01 0.02 0.04 0.06 0.1
0.2
2.0
20
40 60
100
X
Fig. 4. Metallic liquid fraction correlation and comparison with other data and correlations.
Page 54
A.1.Ch.E. Journal
January, 1966
10
II
xx
from horizontal air-water flows in pipes of various dimensions. Except for the heated runs of Johnson and AbouSabe, and the nearly adiabatic data of Pike, the data in
Figure 5 were obtained from isothermal flows. The data
furnished by Tek ( 2 0 ) are from flows of oils and air or
natural gas in various sized pipes, The data of Reid et al.
were obtained from pipes of large diameter. The data
shown in Figure 5 all fall substantially higher than the
potassium results, in keeping with the comparison with
other correlations (7, 12) which are based on data of the
type plotted in Figure 5.
A few experiments on two-phase flows of metallic SYStems have been reported (21) , but either the flow characteristics involved, the system geometries, or the method
of data presentation precluded comparison with the results of this study. The only metallic pressure gradient
and liquid fraction data for which a valid comparison
could be made are the horizontal isothermal mercurynitrogen results of Koestel ( 2 2 ) . The pressure gradient
data are compared with the potassium correlation in Figure 6 which shows very favorable agreement between the
two fluid systems. This agreement, as compared with disagreement between the potassium and nonmetallic systems, suggests that some fundamental difference exists in
pressure drop behavior between metallic and nonmetallic
two-phase fluids.
An effort was made to determine a parameter which
distinguishes the metallic data of Figures 3 and 6 from
the data of other systems shown in Figure 5. Neither the
vapor-to-liquid viscosity ratio nor density ratio provided
a suitable parametric effect (1) . Of several other parameters examined, the most promising was the liquid vis-
10-1
10-
to-*
. -
.c
&
0
I0
or'
10-
I-
LL
I-
LT
I-
0
r
x
cn
10-2
10'
LL
P
I-
LEAST -SQUARES
LINE THROUGH
POirASSfUM DATi
IO-~
cn
4v
5
10'
.
A
I0
1.0
PIKE, H20-STEAM
ISBIN. ET. AL,
H 2 0 -STEAM
REID, E T AL. AIR- H20
POZ
DA?
4''
. PlPF
. ..
REID, E T AL., AIR-HzO, 6" PIPE
GOVIER, E T AL, AIR - H 2 0
JOHNSON. ET AL. AIR-HgO ISOTH
JOHNSON; E T AL; AlR-H;O
HEAT
TEK, OIL-GAS
0.I
0.0I
MERCURY-NITROGEN
OF KOESTEL (22)
10-E
Fig. 5. Comparison of potassium two-phase pressure gradient correlation with data of other investigations.
lo-'
.tp
0.01
AVERAGE QUALITY, x
I01
Fig. 6. Comparison of potassium two-phase pressure gradient correlation with mercury-nitrogen data of Koestel (22).
A.1.Ch.E. Journal
Page 55
cosity influence number of Ros ( 8 ) which includes surface tension. Indications are that interfacial characteristics
and wettability may be principal factors which Werentiate two-phase flow behavior among various fluid systems,
so surface tension would be expected to have influence in
more general correlations.
Metallic Void Fractions
where
[Partial notation is used in Equation (11) since temperature is also a variable. However, since temperature corrections were made in calculating void fractions ( 1 ) , it is
assumed any contributions to a'( a) from temperature
variance are negligible.] The working beam count rate is
The variance of any individual gamma ray counting measurement is equal to the number of counts obtained. Then
Page 56
VOID FRACTION, a
A. I .Ch.E. Journal
(Y
January, 1966
v,
s =Vt = ( -)
1-x
(+)
(E)
(15)
CONCLUSIONS
1. Although the experiments were run with a sodiumpotassium mixture containing 8 wt. % sodium, it is believed that the two-phase flow correlations are representative of pure potassium behavior.
2. Two-phase frictional pressure gradients for potassium
were correlated in terms of a two-phase friction factor f
defined by Equation ( 3 ) , which accounts for the effects
of total mass flow rate, tube diameter, and system pressure. Quality was the other correlating variable.
3. Well-known, two-phase pressure dro correlations
and a theoretical model, when applied in t e usual manner, predict potassium pressure gradients that are substantially greater than the experimental values.
4. The potassium two- ase frictional pressure gradient
correlation shows remar ably good agreement with previously reported mercury-nitrogen data ( 2 2 ) . These
metallic data are much lower than values reported for
flows of water-air, oil-air, and water-steam systems.
5. The potassium void fraction data, together with
other metallic data reported previously (11, 13), have
yielded a Lockhart-Martinelli type of liquid fraction correlation for single-component metallic fluids.
6. Metallic void fractions are Iower than for watersteam and water-air systems; previously published correlations do not fit the metallic data. The mercury-nitrogen
data ( 2 2 ) , with which the potassium pressure gradients
agreed favorably, showed large divergence with respect
hp
Eh
=
=
=
=
Greek Letters
a
p
p
u
=
=
=
=
Subscripts
A. I.Ch.E. Journal
Page 57
Removal of
Small
A flow system wos used to measure the physical adsorption isotherms of two mixtures of
nitrogen and hydrogen and two mixtures of methane and hydrogen on a synthetic zeolite. The
meosurements were made a t liquid nitrogen temperature and a t pressures of 4 to 85 atm.
Static systems were used to measure the pure component isotherms of nitrogen, methane, and
hydrogen at the some temperature and over the appropriate pressure ranges.
Although many different methods have been proposed for using pure component adsorption
isotherms to predict the mixture isotherm, the only method giving quantitative agreement with
the data of this study is the empirical adsorption enhancement factor (4).
In addition to the adsorption isotherms, the concentration-time or breokthrough curves of
the mixtures were measured a t flow rates ranging from 40 to 300 Ib./(hr.)(sq.ft.). The method
proposed by Eagletan and Bliss (18) was used to correlate these curves.
Page 58
A.1.Ch.E. Journal
January, 1966