International Journal of Heat and Mass Transfer: Stéphane Lips, Josua P. Meyer
International Journal of Heat and Mass Transfer: Stéphane Lips, Josua P. Meyer
International Journal of Heat and Mass Transfer: Stéphane Lips, Josua P. Meyer
a r t i c l e i n f o a b s t r a c t
Article history: This article is the second part of a two-part paper, dealing with an experimental study of convective con-
Available online 12 October 2011 densation of R134a at a saturation temperature of 40 °C in an 8.38 mm inner diameter smooth tube in
inclined orientations. The first part concentrates on the flow pattern and the heat transfer coefficients.
Keywords: This second part presents the pressures drops in the test condenser for different mass fluxes and different
Convective condensation vapour qualities for the whole range of inclination angles (downwards and upwards). Pressures drops in a
Inclined two-phase flow horizontal orientation were compared with correlations available in literature. In a vertical orientation,
Pressure drop
the experimental results were compared with pressure drop correlations associated with void fraction
Void fraction
correlations available in literature. A good agreement was found for vertical upward flows but no corre-
lation predicted correctly the measurements for downward flows. An apparent gravitational pressure
drop and an apparent void fraction were defined in order to study the inclination effect on the flow.
For upward flows, it seems as if the void fraction and the frictional pressure drop are independent of
the inclination angle. Apparent void fractions were successfully compared with correlations in literature.
This was not the case for downward flows. The experimental results for stratified downward flows were
also successfully compared with the model of Taitel and Dukler.
Ó 2011 Elsevier Ltd. All rights reserved.
0017-9310/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijheatmasstransfer.2011.09.034
406 S. Lips, J.P. Meyer / International Journal of Heat and Mass Transfer 55 (2012) 405–412
Nomenclature
Subscripts Superscripts
eq equivalent ⁄ apparent
entrainments for a gas–vapour flow (air-n-heptane) and heat DPmeas, and a correction DPlines, which depends on the inclination
transfer coefficients during condensation of n-heptane in down- angle:
ward flows. They concluded that the inclination angle has no effect
on the pressure drop. Beggs and Brill [18] proposed a correlation to
predict the void fraction for all inclination angles (both downward DPtest ¼ DPmeas þ DPlines ð1Þ
and upward flows). However, data were collected for air-water As the refrigerant is fully vapour in the pressure lines, the pres-
flow in 25 mm and 38 mm diameter tubes and thus the proposed sure drop in the lines can be depicted as:
correlation cannot be extrapolated for condensation of refrigerant
in an 8.38 mm inner diameter tube.
In conclusion, there is a lack of predictive models to determine DPlines ¼ qv gLDp sin b ð2Þ
the pressure drops in inclined tubes during convective condensa-
tion of refrigerant, especially because of the need to know the void
fraction as a function of the inclination angle. The effect of the con- where b is the inclination angle of the test condenser. b is positive
densation process is not clear as most of the correlations presented for upward flows and negative for downward flows.
in the literature were developed for adiabatic gas–liquid flows. The pressure drops in the test condenser were recorded for the
The purpose of this paper is to get a better understanding of the same experimental conditions summarised in Part I (Fig. 2) of this
different phenomena that affect the pressure drops during convec- paper in terms of mass fluxes G and vapour qualities x. In horizon-
tive condensation of refrigerant in inclined tubes. In the first part tal orientation, the conditions correspond to intermittent and
[19] of this two-part paper, an experimental study of flow patterns annular flow patterns at the boundary of the stratified flow regime
and heat transfer coefficients during convective condensation in an of the Thome–El Hajal–Cavallini [20] map. For each data point, the
inclined tube was presented. This second part is dedicated to the measurements were realised for different inclination angles. The
study of the pressure drops and void fractions in the same experi- FP2000 Sensotec differential pressure transducer was calibrated
mental set-up at the same conditions reported in the first part. to an accuracy of 50 Pa. During all the experiments, the saturation
temperature and the heat transfer rate in the test condenser were
2. Data reduction and experimental procedure kept constant at 40 ± 0.5 °C and 200 ± 5 W respectively.
" #1
0:25
x x 1x 1:18ð1 xÞ½g rðql qv Þ
erh ¼ ð1 þ 0:12ð1 xÞÞ þ þ
qv qv ql Gq0:5
l
ð7Þ
For the experimental conditions in the present study, the
momentum pressure drop calculated with this correlation was al-
Fig. 2. Measured pressure drops for different mass fluxes (x = 0.5). ways lower than 10% of the frictional pressure drop, so the choice
of void fraction correlation is not of great importance for the
momentum pressure drop determination. Note that, as stated by
3. Effect of inclination angle on pressure drops and void Dalkilic et al. [15], the choice of the void fraction correlation is of
fractions great importance for determining the gravitational pressure drop.
For the horizontal orientation, the gravitational pressure drop is
3.1. Experimental study of pressure drops equal to zero, whatever the void fraction, so it is possible to deter-
mine the frictional pressure drops. Fig. 3 represents a comparison
The pressure drops measured in the test condenser as a function between the experimental frictional pressure drops in horizontal
of the inclination angle are plotted in Fig. 1 for G = 300 kg/m2 s and orientation and the predictions of several correlations and models
for different vapour qualities. An angle of 90° is for vertical down- published in the literature. The model of Moreno Quibén and
ward flow, 0° is for horizontal flow and +90° is for vertical upward Thome [8] best represents the experimental results. This model
flow. The pressure drops increase when the inclination angle in- is a flow pattern-based correlation and was developed for adiabatic
creases because of the gravitational pressure drop. We can note flows and for convective evaporation in smooth tubes. In this
that the increase is stronger for upward flows than for downward study, it was used with the flow pattern map of El Hajal et al.
flows. Furthermore, the smaller the vapour quality, the stronger [20] to predict the flow pattern during convective condensation.
the increase of the pressure drops with the inclination angle. For The other correlations presented in Fig. 3 were developed for adi-
the horizontal and vertical downward orientations, the pressure abatic flows and mostly for annular flow patterns. The correlation
drops increase when the vapour quality increases. However, for of Friedel [24] and the one of Grönnerud [4] agree well with the
vertical upward orientation, the pressure drops decrease when experiments whereas the correlation of Chisholm [25] gives good
the vapour quality increases. results for high pressure drops only, which correspond to annular
Fig. 2 gives the pressure drops as a function of the inclination flow pattern. The homogeneous model, the correlations of
angle for different mass fluxes and for x = 0.5. The different curves Müller-Steinhagen and Heck [3] and those of Lockart–Martinelli
follow the same trend for different mass fluxes. The pressure drops [26] show a higher discrepancy between the predictions and the
increase when the mass flux increases and for upward flows. How- measurements.
ever, the relative inclination effect on the pressure drop is almost Several correlations were developed to predict the pressure
independent of the mass flow. drops in vertical tubes, especially for upward flows. However, to
It is commonly admitted in the literature that the measured be able to compare the experimental results with these correla-
pressure drops, DPtest , are the sum of three different terms: the tions, we have to determine the gravitational pressure drops and
gravitational pressure drop, DP grav , the momentum pressure drop, thus choose a void fraction correlation. Fig. 4 represents the
DPmom , and the frictional pressure drop, DP fric :
DPtest ¼ DPmom þ DPgrav þ DPfric ð3Þ
The momentum pressure drop depends on the kinetic energy at
the inlet and outlet of the tube and thus on the void fraction as a
function of the vapour quality, which depends on the inclination
angle:
" ! ! #
2 ð1 xÞ2 x2 ð1 xÞ2 x2
DPmom ¼ G þ þ ð4Þ
ql ð1 eÞ qv out
ql ð1 eÞ qv in
The subscripts in and out refer to the inlet and outlet of the tube
respectively. The gravitational pressure drop is directly linked to
the inclination angle b of the tube
DPgrav ¼ qeq gLDp sin b ð5Þ
Fig. 5. Comparison of experimental pressure drops with different correlations for Fig. 6. Apparent gravitational pressure drops for different vapour qualities as a
vertical downward flow. function of inclination angle (G = 300 kg/m2 s).
S. Lips, J.P. Meyer / International Journal of Heat and Mass Transfer 55 (2012) 405–412 409
plotted in Fig. 8 for G = 300 kg/m2 s and for different vapour qual-
ities. For upward flows, the apparent void fraction can be consid-
ered constant, at least for a void fraction higher than 0.25. For
downward flows, the apparent void fraction increases when the
inclination angle increases. For each curve, three markers repre-
senting different correlations are plotted. For b = 0°, the marker
represent the value of the Steiner [22] version of the correlation
of Rouhani and Axelsson [23]. This correlation was developed for
horizontal flows. The void fraction correlation of Chisholm [25],
which is supposed to be independent of the tube orientation, is
plotted for b = 45°. Lastly, the Rouhani and Axelsson correlation
for vertical tubes [23] is plotted for b = 90°. Note that the different
correlations and the apparent void fraction for upward flows fol-
low the same trends. Thus, it would be interesting to further inves-
tigate the link between the apparent and the actual void fraction.
In the same figure is plotted in thick lines the mean apparent void
fraction for upward flows, which is determined by doing a linear
Fig. 7. Effect of the mass flux on the apparent gravitational pressure drop as a regression of the apparent gravitational pressure drop as a function
function of inclination angle. of the sinus of the inclination angle. The range of inclination angles
used for the linear regression is 5–90°, which corresponds to the
range where the curves can be considered as linear. Eqs. (9) and
mass flux. On the contrary, for downward flows, stratified flows oc- (10) are then used to calculate the mean apparent void fraction.
cur: this kind of flow is strongly dependent on the gravitational Fig. 9 represents the same type of curve than that in Fig. 8 but
forces and thus on the inclination angle of the tube. As a conse- for three different mass fluxes and three different vapour qualities.
quence, the slip ratio between the phases, the void fraction and This graph confirms the fact that the mass flux has almost no influ-
the frictional pressure drops strongly depend on the inclination ence on the apparent void fraction as well as on the void fractions
angle. predicted by the correlations.
As for the apparent gravitational pressure drop, different behav-
3.2. Theoretical study of the void fraction iours can be distinguished in terms of apparent void fractions: for
upward flows, the apparent void fraction is almost independent of
From the apparent gravitational pressure drop, it is possible to the inclination angle whereas it depends strongly on the inclina-
determine an apparent void fraction, which is defined as the void tion angle for downward flows and low vapour qualities. For high
fraction that would have led to the apparent gravitational pressure vapour qualities, the apparent void fractions remain almost con-
drop: stant whatever the tube orientation. By means of the observations
ql q presented in Part I of the present article, it is possible to link these
¼ ð9Þ three types of behaviour to the three main types of flow patterns,
ql qv
namely intermittent, stratified and annular flows respectively. The
where q is the apparent density of the flow: strongest variation of the apparent void fraction is encountered for
low mass fluxes when the tube orientation varies from slightly
DPgrav
q ¼ ð10Þ downwards to slightly upwards. It corresponds to the modification
gLDP sin b of the flow pattern from stratified to intermittent (Fig. 4, Part I).
The apparent void fraction is equal to the actual void fraction The experimental mean apparent void fractions for upward
only if the frictional pressure drops for the inclined orientation flows are plotted in Fig. 10 as a function of the vapour quality for
are the same as those for the horizontal orientation. However, G = 300 kg/m2. The void fractions predicted by different correla-
keeping this limitation in mind, it is interesting to study the tions are also presented. A good agreement is observed between
apparent void fraction as a function of the inclination angle. It is the experimental results and the correlations of Friedel [24] and
Fig. 8. Effect of inclination angle on the apparent void fraction and comparison with
different correlations (G = 300 kg/m2 s). The thick horizontal lines represent the Fig. 9. Effect of the mass flux on the apparent void fraction. The thick horizontal
mean apparent void fraction between 5° and 90°. lines represent the mean apparent void fraction between 5° and 90°.
410 S. Lips, J.P. Meyer / International Journal of Heat and Mass Transfer 55 (2012) 405–412
Fig. 10. Apparent void fraction for horizontal and upward flows and comparison Fig. 11. Pressure drops predicted by the model of Taitel and Dukler compared with
with different correlations (G = 300 kg/m2 s). experimental results (G = 200 kg/m2 s; x = 0.25).
Chisholm [25]. Note that the LMTD void fraction correlation [21],
used by El Hajal et al. [20] for the flow pattern and heat transfer
[27] models of convective condensation in horizontal tubes, does
not represent the measurements well.
In conclusion, the apparent void fraction may be a possible esti-
mation of the actual void fraction for upward flows. The linearity of
the apparent gravitational pressure drops as a function of the sinus
of the inclination angle tends to show that the frictional pressure
drops and the void fraction can be considered constant in these
conditions. This is not the case for downward flows where the
inclination angle has a stronger influence on the flow pattern and
thus on the frictional pressure drop and void fraction. Thus, for
downward flows, the apparent void fraction has not really any
physical significance. A specific analysis has to be conducted for
these configurations, especially to understand the inclination effect
on stratified flows.
Fig. 12. Pressure drops predicted by the model of Taitel and Dukler compared with
experimental results (G = 300 kg/m2 s; x = 0.1).
3.3. Specification of stratified flows
The most-used model for stratified flows in inclined tubes is model has not been developed for condensing flows, there is a good
that of Taitel and Dukler [28]. The model assumes a smooth strat- agreement between the model and the experiments for the data
ified flow with a flat liquid–vapour interface. The momentum bal- point situated in the stratified flow regime.
ance on the vapour phase yields: The model of Taitel and Dukler [28] also allows determining the
liquid hold-up and the void fraction in the tube. The liquid hold-up
dP is defined as the ratio between the height of liquid in the tube and
AV ¼ sv w Sv þ si Si þ Av qv g sin b ð11Þ
dL v the tube diameter. A comparison between the flow visualisation
and the liquid hold-up predicted by the model of Taitel and Dukler
and for the liquid phase, it gives: is represented in Fig. 13. The pictures represent an image of the
average height of liquid in the tube and the lines represent the pre-
dP diction of the model between b = 60° and b = 15°. Although the li-
Al ¼ slw Sl þ si Si þ Al ql g sin b ð12Þ quid–vapour interface is wavy, we can note a good agreement
dL l
Taitel and Dukler [28] showed that it is possible to solve these
equations considering that the liquid is immobile compared to
the vapour, which leads to si sv w . This model allows the calculation
of the pressure drop for two-phase flows in slightly inclined tubes.
We can compare the prediction of the model with experimental re-
sults only for stratified flows, i.e. for low vapour qualities and low
mass fluxes. In our experimental database, two sets of conditions
led to stratified flows: G = 200 kg/m2 s with x = 0.25 for 45° 6
b 6 5° and G = 300 kg/m2 s with x = 0.1 for 20° 6 b 6 5°. The
comparison between the experimental pressure drops and the
ones calculated by the Taitel and Dukler model is presented in
Fig. 11 for G = 200 kg/m2 s and x = 0.25 and in Fig. 12 for G = 300
kg/m2 s and x = 0.1. The experimental pressure drop is the sum of
the gravitational and frictional pressure drops, i.e. the measured Fig. 13. Comparison between the flow visualisation and the liquid-hold-up
pressure drop minus the momentum pressure drop. Although the predicted by the Taitel and Dukler model [28].
S. Lips, J.P. Meyer / International Journal of Heat and Mass Transfer 55 (2012) 405–412 411
[18] D.H. Beggs, J.P. Brill, A study of two-phase flow in inclined pipes, J. Pet. Technol. [24] L. Friedel, Improved friction pressure drop correlations for horizontal and
25 (1973). vertical two-phase pipe flow, dans: European Two-phase Flow Group Meeting,
[19] S. Lips, J.P. Meyer, Experimental study of convective condensation in an Paper E, 1979.
inclined smooth tube. Part I: Inclination effect on flow pattern and heat [25] D. Chisholm, Pressure gradients due to friction during the flow of evaporating
transfer coefficient, Int. J. Heat Mass Transfer 55 (2012) 395–404. two-phase mixtures in smooth tubes and channels, Int. J. Heat Mass Transfer
[20] J. El Hajal, J.R. Thome, A. Cavallini, Condensation in horizontal tubes. Part 1: 16 (1973) 347–358.
two-phase flow pattern map, Int. J. Heat Mass Transfer. 46 (2003) 3349–3363. [26] R.W. Lockhart, R.C. Martinelli, Proposed correlation of data for isothermal
[21] J.R. Thome, Engineering Data Book III, Huntsville, AL: Wolverine Tube, Inc. two-phase, two-component flow in pipes, Chem. Eng. Prog. 45 (1949) 39–48.
(2006). [27] J.R. Thome, J. El Hajal, A. Cavallini, Condensation in horizontal tubes. Part 2:
[22] D. Steiner, Heat transfer to boiling saturated liquids, VDI-Wārmeatlas (VDI new heat transfer model based on flow regimes, Int. J. Heat Mass Transfer 46
Heat Atlas), Verein Deutscher Ingenieure, VDI-Gesellschaft Verfahrenstechnik (2003) 3365–3387.
Und Chemieingenieurwesen (GCV), Düsseldorf, Chapter Hbb. (1993). [28] Y. Taitel, A.E. Dukler, A model for predicting flow regime transitions in
[23] S.Z. Rouhani, E. Axelsson, Calculation of void volume fraction in the subcooled horizontal and near horizontal gas–liquid flow, AICHE J. 22 (1976) 47–55.
and quality boiling regions, Int. J. Heat Mass Transfer 13 (1970) 383–393.