International Journal of Thermal Sciences 76 (2014) 137e146

Contents lists available at ScienceDirect

International Journal of Thermal Sciences

journal homepage: www.elsevier.com/locate/ijts

Near-zero shear stress experiments with heat flux effects on falling

film evaporation inside a vertical tube
T. Storch*, Ch. Philipp, A. Doeg, U. Gross
Institute of Thermal Engineering, Technische Universität Bergakademie Freiberg, Gustav-Zeuner-Strasse 7, D-09596 Freiberg, Germany

a r t i c l e i n f o a b s t r a c t

Article history: Local heat transfer coefficients for falling film evaporation of isopropanol were experimentally measured
Received 28 December 2012 ((Uh, 95%/h)max ¼ 3.5%) inside a vertical brass tube (dinside ¼ 28 mm) at near-zero shear stress. Zero shear
Received in revised form stress experiments with falling film evaporation inside a tube are extremely difficult to realize e not to
28 August 2013
say they are impossible. In literature respective heat transfer investigations and visual flow observations
Accepted 5 September 2013
were mostly done at outside tube surfaces surrounded by an extended volume. For the present in-
Available online 8 October 2013
vestigations at near-zero shear stress a special vapour flow sensor was developed and installed in the
experimental tube to locate the level of near-zero vapour velocity. The measurements included film Re
Keywords:
Heat transfer measurement
numbers up to 100, inner wall heat flux up to 12,500 Wm2 without bubble formation in the super-
Zero shear stress heated liquid, and vapour temperatures ranging from 8.5  C to 36  C (Prliquid, freesurface ¼ 14.5e20.8). The
Falling film heat transfer measurements were focussed on Re number and heat flux effects on falling liquid film
Surface wave evaporation. The basic intension of this paper is a comparison of the evaporation heat transfer mea-
Evaporation surements with the characteristics in the various Re number ranges, and also a comparison with the
Measurement technique authors’ earlier condensation results as reported in Refs. [1,2]. Thereby influences of Re, Pr and Ka
numbers on heat transfer are discussed, correlated (within deviations of 2.7% from measured data) and
compared with the literature.
Ó 2013 Elsevier Masson SAS. All rights reserved.

1. Introduction Alekseenko et al. (1994) [17]. Actually, heat transfer was found to be
enhanced by wave formation and transition to turbulence. The
Falling film evaporation and condensation inside vertical tubes structure of falling liquid films is usually classified to be laminar,
are found in many industrial applications like heat exchangers and laminar-wavy, and locally or fully turbulent depending on Rey-
chemical process columns where heat transfer is always affected by nolds, Kapitza and Prandtl numbers. Falling film condensation and/
the vapour-side shear stress. This acts cocurrent or countercurrent or evaporation heat transfer measurements have been presented
to the liquid film depending on vapour flow direction. The open and correlated by many scientists, e.g. Refs. [12e14,25,26], mostly
literature provides lots of heat transfer measurements, both for for vertical walls and the outside of tubes.
evaporation and condensation, and also studies of the hydrody- The present authors performed extended reflux-condensation
namics of falling liquid films, see, e.g. Refs. [3e10], regarding falling heat transfer measurements with water, ethanol, and isopropanol
film evaporation [1,2,11e14], for condensation, and [15e22] for the (2.6  Pr  55) inside a vertical tube in the countercurrent flow sit-
hydrodynamics. For literature concerning wave shape, film struc- uation with the focus on shear stress effects [1,14,27,28]. Excellent
ture and frequency see also Philipp et al. (2006) [21] and Gross et al. reproducibility (1% for water, 0.5% for isopropanol) was obtained
(2009) [2]. allowing extrapolation of the measured data to zero shear stress [1].
Nusselt (1916) [23] presented the first analytical solution of The results have been plotted in terms of the Nusselt number, see
velocity profile and heat transfer across a smooth laminar film with Fig. 1, and they have been referred to Nusselt’s solution [23]. For
and without the influences of vapour-side shear stress. These re- getting the wave enhancement factor Fwave ¼ Nuexperimental/NuNusselt
sults have been confirmed as the lower limit for condensation and (1916) as shown in Fig. 2 and [1]. Furthermore the condensation heat
evaporation heat transfer, e.g. by Brauer (1956) [24] and transfer data have been correlated for the laminar-wavy range done
for three sub-ranges (Eqs. (1)e(3)) corresponding to different wave
characteristics as obtained from visual observations [2] (for the defi-
* Corresponding author.
E-mail address: [email protected] (T. Storch).
nitions, see Eqs. (4)e(8) below):

1290-0729/$ e see front matter Ó 2013 Elsevier Masson SAS. All rights reserved.
http://dx.doi.org/10.1016/j.ijthermalsci.2013.09.003
138 T. Storch et al. / International Journal of Thermal Sciences 76 (2014) 137e146

restricted the wave factor calculation on Re number effects similar

to Kutateladze’s (1963) [38] approach.
Based on additional visual observations of surface waves, the
present authors evaluated the effective wave frequency as the
leading effect on heat transfer enhancement from slow motion
videos for Re < 100 [21]. Various characteristic ranges have been
obtained from both studies with clear criteria for the transition
from a first range (occasional small waves with a limited heat
transfer enhancement of about 2.8%, Eq. (1)) to a second (two-
dimensional waves with low effective wave frequency, Eq. (2)) and
finally a third range (three-dimensional waves with increased
effective wave frequency, Eq. (3)) bringing enhancement factors up
to 34.5% and 37.5%, respectively ([2,41]).
Heat transfer measurements inside vertical tubes (afflicted by
shear stress) are focussed on thermosyphon applica-
tions [25,42,43]. Zero shear stress experiments with falling film
evaporation inside a tube are extremely difficult to realize e not to
Fig. 1. Condensation heat transfer for the limiting case of zero shear stress in a vertical
say they are impossible. Respective heat transfer investigations and
tube (see also Gross and Philipp (2006) [1]). visual flow observations at outside tube surfaces surrounded by an
extended volume are easier to handle. Struve (1969) [44], Chun and
Seban (1971) [3], Alhusseini et al. (1998) [26], Leuthner (1999) [45]

 and Leuthner et al. (1999) [6] reported such local heat transfer
Fwave ¼ 1:028 Re Ka0:09 < 0:8 (1) experiments.
Realization of evaporation heat transfer measurements with

 near-zero shear stress inside a vertical tube was one aim of the
Fwave ¼ 1:096Re0:141 Ka0:0126 0:8  Re Ka0:09 < 4:2 (2) present study. Second aim was to investigate the relation between
changes in local heat transfer and film structure depending on Re,

 Pr and Ka numbers as well as on heat flux effects. The test setup, the
Fwave ¼ 1:270Re0:040 Ka0:0036 4:2  Re Ka0:09 < 7:3 (3) vapour flow sensor, the Reynolds number and heat flux effects will
now be explained and discussed in detail.
where
2. Methods
Gliquid _ liquid
m
Re ¼ ¼ (4) 2.1. Experimental setup
mliquid pdpipeinside mliquid
An experimental setup which previously has been used for local
and reflux condensation heat transfer measurements [1,14,46] has been
modified for the present evaporation measurements. These modi-
g m4liquid fications enabling both cocurrent and countercurrent vapour
Ka ¼ : (5)
rliquid s3liquid flow [41,47].
The experimental plant (Fig. 3) consists of a vertical brass tube
The Kapitza number (Ka) has been established for classification with a total length and inner diameter of 4.2 m and 28 mm,
of the wave activity in falling liquid films. In literature the Ka respectively. Three cycles are established, one for the liquid and two
number was used first with Re ¼ 0.6075Ka1/11 to compile the for vapour, which can be operated independent from each other.
lower limit for the development of sinusoidal waves (see Kapitza Liquid (e.g. isopropanol) is pumped by an infinitely variable gear
(1948) [29]). Three years earlier, Grimley (1945) [30] suggested pump through a mass flow measuring device at a constant tem-
Re ¼ 0.291Ka1/8 as the smallest Re number for getting wavy liquid perature (controlled by a thermostat) to the top of the tube where it
films. The literature contains various Ka number definitions, and penetrates a porous sintered steel section. A uniform liquid film is
Eq. (5) is used e.g. by Chun and Seban (1971) [3], Alhusseini et al. created which flows downward at the inner wall surface with a
(1998) [26] and VDI heat atlas (2010) [31]. There are, however, total hydrodynamic entrance length of about 1580 mm before
different ones like Ka*¼ 1/Ka used by Alekseenko et al. (1994) [17], entering the measuring level. The uppermost part of the tube is
Al-Sibai (2004) [32], and Weise and Scholl (2007) [10]. Further surrounded by a water jacket for heating (heating system). The
modifications are Ka** ¼ (1/Ka)1/3 and Ka** ¼ (3/Ka)1/5, see heating water can be fed into three different inlets of the evapo-
Dietze and Kneer (2010) [33] and Sofrata (1980) [34] respectively. ration section, permitting three different lengths of the heating
Thereby mostly all researchers are using different reference zone. The measuring level for local heat transfer coefficients is
temperatures. located in the lower-most section. The present experiments have
McAdams (1954) [35] and Blangetti (1979) [36] introduced been done with the two lower heating lengths and mainly with the
averaged wave factors, Fwave ¼ 1.28 and 1.15, whereas Labuntsov thermal entrance length of 740 mm.
(1957) [37], Zazuli (see Kutateladze (1963) [38]) and Uehara et al. For measuring the heat transfer in the measuring level two thin
(1983) [39] found the wave factors to increase with the Re number calibrated thermocouples (0.5 mm in diameter) are situated
introducing Fwave f Re0.04, Fwave f Re0.11 and Fwave f Re0.083 directly underneath the in- and outside surfaces of the thick-walled
respectively. Sofrata (1980) [34], Uehara (1983) [39] and Mitrovic brass tube each at three points over the circumference (every 120 ).
(1990) [40] refined the wave factor calculation by including Ka Therefore the thermocouples were inserted in axially blind-end
number effects. In case of laminar wavy film evaporation, Alhuseini bores which were manufactured by spark erosion (150 mm in
et al. (1998) [26] did the same, whereas Chun and Seban (1971) [3] length and 0.6 mm in diameter). The remaining gap between
 T. Storch et al. / International Journal of Thermal Sciences 76 (2014) 137e146 139

Fig. 2. Wave factor (¼Nuexperimental/NuNusselt (1916)) vs. Re number of reflux condensation heat transfer data inside a vertical tube (see Philipp et al. (2006) [21] and Gross et al. (2009)
[2]).

thermocouples and bores was filled with tin solder. To determine Furthermore at the lower end of the vapour flow sensor a
the heat flux at the inner wall surface we use the measured radial miniature camera is positioned inside to observe the liquid-film
temperature difference (thin thermocouples next to in- and outside hydrodynamics, see detail (A) in Fig. 3. Results of these observa-
surface), the radial distances between the thermocouples and the tions are published in Philipp et al. (2006) [21]. The particular
thermal conductivity of the tube. With the measured temperature mass flow rates were measured after liquefaction by means of
difference between inner tube wall and saturation temperature of Coriolis flow meters with an accuracy of about 0.15% on the
the vapour (measured in the centre of the tube by three 0.25 mm measured value. In addition the vapour and liquid temperatures
thick thermocouples) the local heat transfer coefficient is calcu- were measured highly precise at various positions in the respec-
lated. For more detailed information concerning design see also tive cycles, see also [14].
Thumm et al. (2000, 2001) [14,46].
Vapour is created in the evaporation section and it can be 2.2. Vapour flow sensor
forced to flow upwards, i.e. countercurrent to the liquid film
(vapour cycle I), and/or downwards, cocurrent to the liquid film 2.2.1. Design
(vapour cycle II). This is managed by the respective operation of For identification of the zero vapour velocity level a special
two condensers. After a separate measurement of the condensates measuring probe (vapour flow sensor) has been designed, installed
mass flow rate it is finally mixed with the excess, i.e. non- and tested inside the vertical brass tube, see Figs. 3 and 4. Its length
vaporized, liquid coming from the separation device at the bot- and outer diameter of are about 140 mm and 14 mm, respectively.
tom of the tube. Both condensers can be controlled in a way that The flow sensor consists of 8 pairs of micro platinum resistance
the limiting case of near-zero shear stress (i.e. vapour velocity) is thermometers (PT 1000; no 1e8). The PT 1000 used are thin film
obtained in the measuring level. This is done by using a new elements which consist of a ceramic substrate (thickness,
vapour flow sensor; see Figs. 3 and 4. Combined with a very fine S ¼ 0.4 mm) with a thin layer of platinum (S ¼ 0.85 mm). The
condenser setting the ratio between both of the two vapour flow platinum layer is coated with a glass layer (S ¼ 15 mm) to protect it.
rates (upward and downward) can be controlled precisely for The dimensions of one PT 1000 are 2 mm  2 mm  about
shifting the location of near-zero velocity (vapour) exactly to the 0.416 mm (width  length  thickness). The PT 1000 are arranged
heat transfer measuring level. This condition is comparable to in pairs which are fitted together on their backside, see Fig. 3A.
experiments outside a tube where the vapour volume is big These pairs are mounted at equal distances of 20 mm and situated
enough to be considered at rest. helically along the tube axis. The sensor’s justification is centred in
140 T. Storch et al. / International Journal of Thermal Sciences 76 (2014) 137e146

Fig. 3. Schematic drawing of the experimental test plant with two vapour and one film cycles, a special vapour flow sensor on the right; (A) detail zoom of the sensors lower end
including one pair of micro resistance PT and a miniature camera.

relation to the level of local heat transfer measurement, between PT resulting up- and/or downward vapour flow cools the various
4 and PT 5, see Fig. 4b. heated pairs of micro resistance thermometers with different in-
One PT 1000 of a pair is self-heated by direct current and it is tensity. There is a direct correspondence between the resulting
used for the temperature measurements. The heating of the several temperature distribution along the tube axis and the local vapour
PT’s is due to the manufacturing not exactly the same, thus the velocities due to interrelated cooling effect. E.g. the level of
temperatures differ. Therefore we used an adapted offset to level maximum temperature is obtained in a zone of zero velocity level
the measured temperature differences of the PT’s with a high (no vapour flow) and hence without any cooling effect.
temperature resolution. The resulting temperature difference To locate the level of zero velocity several issues have to be taken
measured between the sensor and the surrounding vapour de- into account. For interpretation of the measurements, all 8 sensors
pends on the velocity of the latter one. Thereby only the level of are divided into corresponding pairs of PT’s having the same dis-
near-zero vapour velocity is localized by this sensor without tance from the level of heat transfer measurement, e.g. PT 1 and PT
measuring the real vapour mass flow. The second PT serves as an 8; PT 2 and PT 7, etc. (see Fig. 4b). In Fig. 5, the readings of six
additional heat source in cases of strong cooling. All temperature sensors (PT 2e7) next to the level of local heat transfer measure-
measurements are done in a special measuring chain, and a very ment are plotted versus time. There are dashed and solid lines
high resolution up to 1/1000 K is obtained. representing temperature graphs measured above (PT 2 to PT 4)
and below (PT 5 to PT 7) this level, respectively. The upper diagram
2.2.2. Principle of operation of Fig. 5 shows smooth curves representing the measured data. For
The 8 PT’s have been levelled out against each other prior to the better explanation these data have been time averaged during
experiments. As soon as falling film evaporation is started, the (00:00 to 01:10 min) and (02:30 to 05:00 min) and plotted in the

Fig. 4. (a) schematic drawing with dimensions of vapour flow sensor; (b) set-up of the platinum thermometers 1000 (PT) from PT 1 to PT 8 in relation to heat transfer measuring
level; (c)e(d): exemplary shifting of the level of zero velocity (¼vapour separation level) from position next to PT 4 (c) into the level of heat transfer measurement (d).
 T. Storch et al. / International Journal of Thermal Sciences 76 (2014) 137e146 141

lower diagram. As mentioned, the lower the measured temperature former one will be applied in this paper for the measuring level
with the PT’s, the higher the vapour velocity is (due to cooling ef- with the liquid mass flow rate as measured in the film cycle and
fect). There are two typical time series which are divided from each corrected for the liquid evaporated above the measuring level
other by the dashed bold vertical line representing different (vapour cycle).
condenser settings. The vapour separation level is located in the Measured heat transfer coefficients are expressed as the Nusselt
region of two neighbouring pairs of PT’s which show an opposite number
direction behaviour in the temperature trend after changing the

1=3
condensers setting. E.g. during the first minute the presented
h n2liquid =g
temperature graphs show strong and weak cooling for PT 2/PT 7, Nu ¼ (6)
and PT 4/PT 5, respectively. lliquid
With the measurement of this high resolute difference in tem-
perature it is possible to align the condenser settings for condenser with
I and II very fine (see Figs. 4c and 5 at time ¼ 01:10 min). Conse-
quently, with a short time delay, i.e. approximately one minute h ¼ q_ wall;inside =DTfilm (7)
after changing the condenser setting, the vapour separation level
(level of near-zero velocity) is shifted. And PT 4; PT 3; PT 2 show a and they typically exceed those predicted by Nusselt’s theory [23]
decrease and PT 5; PT 6; PT 7 an increase in temperature. This can for both, condensation and evaporation due to wave formation
be explained by the lower and higher cooling, respectively, of the and turbulence. In the range of laminar-wavy film flow the
PT’s by the vapour. In fact for this example (Fig. 5) the shifted improvement is usually expressed by the wave factor
vapour separation level is situated directly between PT 4 and PT 5. Fwave ¼ Nuexperimental/NuNusselt (1916).
Further, the local heat transfer measurement at near-zero shear Furthermore Prandtl number (Pr) is evaluated from
stress can be realised. With every change in one of the parameters,
cp;liquid mliquid
e.g. mass flow rate, heat flux, temperature, etc., the whole proce- Prliquid ¼ (8)
dure of locating and shifting of the level of zero velocity has to be
lliquid
performed again. For the present evaluation the liquid properties are taken at
 
2.3. Evaluation of data Treference ¼ Twall;inside  freference Twall;inside  Tvapour (9)

Two different Reynolds number definitions are common prac- with freference ¼ 0.25 for the liquid viscosity due to the big influence
tice, namelyRefilm ¼ Gliquid/mliquid, with Gliquid ¼ m
_ liquid =ðpdinside Þ, of the velocity profile next to the wall surface, which is proposed by

based on film thickness and velocity, andRefilm ¼ 4Refilm , derived Rohsenow et al. (1998) [48] and Fujii (1991) [49] for laminar falling
condensate films at low shear stress. For all other freference ¼ 0.5 (see
from tube diameter and superficial velocity of the liquid. The
also [48]) with Prliquid, freesurface as the only exception where
freference ¼ 1 is used. Extended experiments have been carried out at
zero velocity conditions. Local heat transfer coefficients have been
measured for falling film evaporation of isopropanol covering film
Re numbers up to 100, inner wall heat flux up to 12,500 Wm2, and
vapour temperatures ranging from 8.5  C to 36  C (Prliquid,
freesurface ¼ 14.5e20.8).

2.3.1. Uncertainty analysis

For our evaporation measurements an uncertainty analysis
based on Kim et al. (1993) [50] with Eqs. (10)e(12), was done. U is
the uncertainty, B the bias limit and P the precision limit. The
precision limit couldn’t be calculated with Eq. (12) because the
repeatability of every measured parameter is not known. Never-
theless a very good repeatability of 0.5% for the resulting heat
transfer coefficient (h) could be obtained, so we use 0.5% for the
precision limit. For the bias limit the errors in measurement of all
measured parameters used to calculate the heat transfer coefficient
are considered (see eq. (11) and also Thumm 2000 [46]).
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
U ¼ B2 þ P 2 (10)

with
!2
Xn
vf
B2f ¼ B2xj (11)
j¼1
vx j

and
Fig. 5. Example of locating and shifting the level of zero shear stress into the heat
transfer measuring level e next to the highest temperature of two PT’s; upper dia- !2
gram: interpolated data; lower diagram: averaged data (dotted lines: graphs of PT’s Xn
vf
above the level of heat transfer measurement; straight lines: below measuring level, Pf2 ¼ Px2j (12)
compare Fig. 4b). j¼1
vx j
142 T. Storch et al. / International Journal of Thermal Sciences 76 (2014) 137e146

Hence the uncertainty for 95% of the measured heat transfer only moderate additional enhancement of heat transfer is detect-
data presented in this paper (w670 points) was calculated with (Uh, able. Finally the slope is found to be re-increased for about Re > 40
95%/h)min ¼ 1.2% and (Uh, 95%/h)max ¼ 3.5%. with the transition to developed turbulence in range (e). Basically
the new evaporation results verify the authors’ earlier condensa-
3. Results and discussion tion heat transfer data whereby for Re ¼ 25e70 the former ones
were found to be about 2e3% higher.
3.1. Reynolds number effects
3.2. Heat flux and temperature effects
Heat transfer results are plotted in Fig. 6 in terms of Fwave versus
Re number. Most of the experiments have been carried out for 6550 The general Re number effect on Fwave is found to be superposed
(150) Wm2 and 36  C as heat flux and vapour temperature, by a certain influence of the wall heat flux. This is clearly seen in
respectively. In general, the wave factor Fwave is found to increase Fig. 6 for Re > 12 (in range (c)) showing some increase of Fwave
along with the Re number as indicated in Fig. 6 by the middle trend when the heat flux is raised from 2000 Wm2 to 11,500 Wm2. This
line corresponding with general statements in the literature, e.g. influence seems to be even stronger in range (d), i.e. in presence of
Ref. [31]. developing turbulence. Here an additional temperature effect be-
Based on own visual investigations of vertically falling liquid comes visible showing a decrease of Fwave when the temperature is
films, as reported in Philipp et al. (2006) [21] and Gross et al. (2009) raised from 14.7  C to 36  C (as measured for 5000 Wm2). This
[2], five main characteristic Re number ranges are obtained in Fig. 6 strange behaviour asks for analysis and discussion.
with the limits as indicated by bold vertical lines: To show the influence of film temperature and heat flux on local
Nu numbers three different Re numbers have been selected, as
(a) Re < 2.5: Fwave is nearly constant (obtained for heat fluxes indicated in Fig. 6 by bold arrows. They represent the ranges of
7500e9500 Wm2); two-dimensional waves (Re ¼ 12.3), the beginning of transition
(b) 2.5 < Re < 4: The wave factor starts to rise moderately; from three-dimensional waves to local turbulence (Re ¼ 42.2), and
(c) 4 < Re < 25: The increase becomes stronger . finally transition to developed turbulence (Re ¼ 68.0) see also
(d) 25 < Re < 70: . and weaker again . Refs. [2,21,47]. Fig. 7 shows a plot of Nu number versus heat flux for
(e) Re > 70: . and finally a re-intensification is obtained with a these three Re numbers. Characteristic rising curves are obtained
significantly stronger slope of the middle trend line. with a weak superposed vapour-temperature effect as already seen
in Fig. 6. Both the temperature effect and the heat flux effect, i.e.
The various Re number ranges correspond to respective ranges temperature-difference effect, are assumed to be caused by
for the wavy-film flow characteristics: Only occasional, so-called respective thermophysical property variations in the evaporating
“very small” and first small waves are found in (a), both of them liquid film. The combined effects of surface tension, liquid density
being two-dimensional i.e. symmetrically distributed along the and viscosity variations bring the Kapitza number to decrease with
inner perimeter with horizontal wave crests. Range (b) represents the temperature. An intensification of the wave activity is
transition to two-dimensional waves with rising mightiness. Range observed, see Ref. [2], and subsequently the Nu number is forced to
(c) is characterized by increasing wave frequency and heat transfer rise.
enhancement, whereby a transition to two-dimensional-inclined Increasing heat flux at a constant vapour temperature brings
waves with non-horizontal wave crests and finally three dimen- proportionally increasing temperature differences across the
sional wave shapes are found. Range (d) with fully three- evaporating liquid film. The latter one becomes increasingly su-
dimensional waves brings first indications for the transition to perheated next to the wall surface. The following effects have to
local turbulence with further increased wave frequency. However, been taken into account:

Fig. 6. Wave factor (¼Nu experimental/NuNusselt (1916)) vs. Re number for a wide range of heat fluxes and several vapour temperatures (evaporation heat transfer data; liquid:
isopropanol).
 T. Storch et al. / International Journal of Thermal Sciences 76 (2014) 137e146 143

Fig. 7. Nu number enhancement by increasing heat flux (three temperature ranges) for three different Re numbers and film structures (liquid: isopropanol).

 Superheated liquid is in a metastable thermodynamic state with situations, and based on their suggestion Hufschmidt et al. (1966)
the latent chance for the formation of vapour bubbles next to the [61] found (Prliquid/Prliquid, wall)0.11 as an optimized correction term
heating surface. If the excess temperature is large enough, which finally is suggested in the VDI heat atlas (section G1) [31] for
transition to nucleate boiling is obtained with strongly both laminar and turbulent internal convection. This concept will
increased heat transfer coefficients on one side and an now be applied to consideration of heat flux effects on falling film
increasing danger of dry-patch formation at a critical heat flux evaporation heat transfer.
on the other (e.g. for vertical wall [51], outside tube surface [52],
and inside thermosyphons [25,53,54]). These phenomena are 3.3. Correlation
not the scope of the present investigation which is pure related
to evaporation from the falling film surface. All experimental heat transfer data, in terms of Fwave, have been
 Evaporating films show characteristic temperature profiles with correlated with Re and Ka numbers based on the following power
the gradient orthogonal to the wall surface and respective law equation:
variations of the temperature-dependent thermophysical
properties. Temperature profiles in falling liquid films were !0:11
a b Prliquid;freesurface
studied by several scientists (e.g. Ganchev and Trishin [55], Lyu Fwave ¼ CRe Ka (13)
and Mudawar [56], and Mudawar and Houpt [16]) who Prliquid;wall
concluded that the profile depends highly on film thickness
variations due to waves. They presented two main conclusions: This has been done separately for the various Re number ranges
First, the temperature is highest in regions between big wave see Table 1. In range (a) the wave factor has been obtained to be
crests and lowest inside the wave itself (for Re  1250 [56]). The constant at about Fwave ¼ 1.06. For the fully turbulent liquid film in
maximum in difference will increase by raising the heat flux. range (e), the enhancement factor concept based on Nusselt’s so-
Second, the velocity of the big waves is much higher than the lution loses its sense.
residual thin liquid film. Thus only an internal circulation inside The strongest Re and Ka number effects are clearly found for
a big wave as an isolated lump is possible. With the increase of range (c), i.e. for the developed two-dimensional wavy film flow
Re number the area of thin liquid film at higher temperature situation with continuous transition to two-dimensional-
increases due to the raising effective wave frequency (see Ref. inclined and further to three dimensional wave shape. One
[2]). Thus the heat transfer (or resulting wave factor, Fig. 6) possible explanation for this strong increase of Fwave is found in
increases too. the strong increase in effective wave frequency which is more
 Our heat transfer experiments have been evaluated by taking than doubled in this range, see also Gross et al. (2009) [2]. The
mean values of the properties, Eq. (8), ignoring their variations exponents of Re and Ka numbers determined in our correlations
across the liquid film. The liquid viscosity as the most sensitive of the evaporation heat transfer data in range (c) and (d) are
property exhibits increasingly big differences between the wall equal to those of the condensation ones (see Eqs. (1)e(3)).
surface (where it is low) and the vapour-side surface (where it is Hence for low Re the equation used to correlate the data set (Eq.
large) when the heat flux is being raised. (13)) represents the most important properties and relationships
of fluid dynamics and heat transfer for falling film condensation
This phenomenon is not generally new. Heat flux effects on and evaporation.
single-phase convection heat transfer in tubes has already
been studied by Sieder and Tate (1936) [57] who introduced (mliquid/ 3.3.1. Remaining deviations
mliquid, wall)0.14. This concept has been modified by M.A. Mikheev The remaining deviations between correlated and experimental
[58] and by V.V. Yakovlev [59,60] in the 1950’s for turbulent flow data are plotted in Fig. 8 versus Re (left) and Ka numbers (right). The
144 T. Storch et al. / International Journal of Thermal Sciences 76 (2014) 137e146

Table 1
Results of data correlation following Eq. (13).

0:158
Nu ¼ 2:65 Re* ðKaÞ0:0563 (14)
Range Re number range Constant Re number Ka number
exponent a exponent b
which may be rewritten for the wave factor as
(a) Re < 2.5 1.060 e e
(b) 2.5  Re < 4 1.035 0.0467 0.0021
0:175
(c) 4  Re < 25 1.096 0.1410 0.0126 Fwave;Alhusseini et alð1998Þ f Re* ðKaÞ0:0563 (15)
(d) 25  Re < 70 1.296 0.04 0.0036
The Re and Ka number effects are obviously much stronger than
in the present investigations.
Weise and Scholl (2007) [65] presented falling film evaporation
correlation proves to be extremely good with deviations typically results for liquids in the high Prandtl number range with
within 2.7%. This is, of course, due to the limited width of the data Prliquid ¼ 40e130. Overall heat transfer measurements have been
set regarding Re number and also Ka number ranges. carried out with the heat flux, i.e. the driving overall temperature
difference, as one of the parameters. Only a few of the results are
3.4. Comparison with the literature found for the laminar range with no significant heat flux effects. In
the turbulent range, however, there is an enhancement of heat
The open literature contains only a few hints regarding the heat transfer in the order of 20% when the driving temperature differ-
flux effect. For conditions without bubble formation in the super- ence is increased from Dt ¼ 10 Ke20 K. This effect has been found
heated liquid film, only a fistful of earlier studies is found. It should for Pr ¼ 40 and 58 and it seems to vanish for larger Prandtl
be noted that there are big differences between measurements numbers, i.e. Pr 78.
inside and outside of a tube and also plates due to geometrical
influences. 4. Conclusions
Struve (1967) [62] and Elle (1970) [63] investigated falling film
evaporation heat transfer of refrigerant R11 outside a vertical tube. Based on highly accurate local condensation heat transfer
No heat flux effect has been found by them below the limit of first measurements the authors’ interest was focused on performing
bubble formation (q_ wall;outside < 4500 Wm2 ). Shah and Darby local evaporation heat transfer measurements of falling films
(1973) [64] studied surfactant effects on evaporation heat transfer (Re < 100). The measurements were performed inside a vertical
in a liquid film flowing downward a vertical flat plate. Among many tube (dinside ¼ 28 mm) at near-zero shear stress, and under condi-
other findings they obtained results for the limiting case of pure tions without bubble formation in the superheated liquid film. For
water at Re numbers in the range of developing turbulence. There this purpose a great expense in test setup modification was done
are clear signs of heat transfer promotion for increasing heat flux and a new vapour flow sensor was developed, successfully tested
which is found to be moderate in case of surface evaporation. and used in the experimental setup for measurements. For iso-
Beyond the onset of bubble nucleation, at about Dt ¼ 5 K, the heat propanol, the authors evaluated local heat transfer measurements
flux effect becomes much stronger. of falling film evaporation ((Uh, 95%/h)max ¼ 3.5%) at near-zero
Alhusseini et al. (1998) [26] measured mean heat transfer co- shear stress with systematic variations of Re number, heat flux
efficients for evaporation of water and propylene glycol outside a and vapour temperature. In general, the wave factor Fwave
vertical tube and they provided the following correlation (14) valid (¼Nuexperimental/NuNusselt (1916)) was found to increase along with
for the laminar-wavy region: the Re number, which is in good agreement with literature results.

Fig. 8. Deviation of evaporation heat transfer data correlation (Eq. (13) for range bed, see Table 1) vs. film Re number (left) and vs. Ka number (right) (liquid: isopropanol).
 T. Storch et al. / International Journal of Thermal Sciences 76 (2014) 137e146 145

Five characteristic Re number ranges (Re < 100) were found which [22] R.V. Craster, O.K. Matar, Dynamics and stability of thin liquid films, Rev. Mod.
Phys. 81 (2009) 1131e1198.
correspond to respective ranges for the wavy-film flow character-
[23] W. Nusselt, Surface condensation of steam (Die Oberflächenkondensation des
istics, see also [2,21,47]. Basically the new results verify the authors’ Wasserdampfes), Z. Ver. Dtsch. Ing. 60 (1916) 541e545, 569e575 (in German).
earlier condensation heat transfer results. Nevertheless, a deviation [24] H. Brauer, Hydrodynamics and Heat Transfer of Falling Films (Strömung und
in the range of transition from three-dimensional wave shape to Wärmeübergang bei Rieselfilmen) (in German), VDI-Verlag GmbH, Düssel-
dorf, 1956.
turbulence (Re ¼ 25e70) was measured. Thereby the evaporation [25] M.S. El-Genk, H.H. Saber, Heat transfer correlations for liquid film in the
heat transfer data (converted into Fwave) were found to be 2e3% evaporator of enclosed, gravity-assisted thermosyphons, J. Heat Transf. Trans.
higher. Both showed similar Re number effects due to changes in ASME 120 (1998) 477e484.
[26] A.A. Alhusseini, K. Tuzla, J.C. Chen, Falling film evaporation of single compo-
wave structures with rising Re numbers. Furthermore heat flux nent liquids, Int. J. Heat Mass Transf. 41 (1998) 1623e1632.
effects have been detected and correlated with the ratio of Pr [27] U. Gross, Ch. Philipp, S. Thumm, Effect of countercurrent vapor flow on
numbers at vapour temperature and wall temperature to the power condensation heat transfer inside a vertical tube, in: Proc. 12th Int. Heat
Transfer Conf., Grenoble, 2002, pp. 923e928.
of 0.11, as known from pipe flow correlations. The remaining de- [28] U. Gross, Shear stress effects on counter current film condensation inside
viations from the measured data are typically within 2.7%. tubes: a review of recent developments 90 years after Nusselt, in: Proc. Int.
Heat Transfer Conf., 2006.
[29] P.L. Kapitza, Wave flow of thin layers of a viscous fluid: I. Free flow, Zh. Eksp.
Acknowledgements Teor. Fiz. 18 (1948) 3e28 (in Russian). (Engl. translated in: Collected Papers of
P.L. Kapitza, ed. D.T. Haar, Pergamon, 1965, pp. 662e689).
[30] S.S. Grimley, Liquid flow conditions in packed towers, Trans. Inst. Chem. Eng.
The authors like to thank M. Schemmel and A. Wahl as well as 23 (1945) 228e235.
the numerous students for their contributions. The support by the [31] VDI Heat Atlas, second. ed., Springer, New York, 2010.
[32] F. Al-Sibai, Experimentelle Untersuchung der Strömungscharakteristik und
Deutsche Forschungsgemeinschaft (DFG) is greatly appreciated. des Wärmeübergangs bei welligen Rieselfilmen. PhD thesis, RWTH Aachen,
2004.
[33] G. Dietze, R. Kneer, Capillary flow separation in 2- and 3-dimensional laminar
References falling liquid films, in: Proc. 14th Int. Heat Transfer Conf.: IHTC14, 2010.
[34] H. Sofrata, Theoretical study of film wise condensation considering wave
[1] U. Gross, Ch. Philipp, Conjugated shear stress and Prandtl number effects on initiation, Wärme Stoffübertrag. 14 (1980) 201e210.
reflux condensation heat transfer inside a vertical tube, Int. J. Heat Mass [35] W.H. McAdams, Heat Transmission, third ed., McGraw Hill, London, 1954.
Transf. 49 (2006) 144e153. [36] F. Blangetti, Lokaler Wärmeübergang bei der Kondensation mit überlagerter
[2] U. Gross, T. Storch, Ch. Philipp, A. Doeg, Wave frequency of falling liquid films Konvektion im vertikalen Rohr. PhD thesis, Universität Karlsruhe, 1979.
and the effect on reflux condensation in vertical tubes, Int. J. Multiphase Flow [37] D. Labuntsov, Heat transfer in film condensation of pure vapours on vertical
35 (2009) 398e409. surfaces and horizontal tubes, Teploenergetika (1957) 72e80.
[3] K.R. Chun, R.A. Seban, Heat transfer to evaporating liquid films, J. Heat Transf. [38] S.S. Kutateladze, Fundamentals of Heat Transfer, Arnold, London, 1963.
Trans. ASME 93 (1971) 391e396. [39] H. Uehara, H. Kusuda, T. Nakaoka, M. Yamada, Experimental study of filmwise
[4] G. Schnabel, Bestimmung des örtlichen Wärmeüberganges bei der Fall- condensation on a vertical surface, Trans. Jpn. Soc. Mech. Eng. B 49 (1983)
filmverdampfung und Kondensation an gewellten Oberflächen zur Auslegung 666e675.
von Hochleistungsverdampfern. PhD thesis, Universität Fridericiana Karls- [40] J. Mitrovic, Filmkondensation ruhender Dämpfe an senkrechten Flächen, BWK
ruhe, 1980. 42 (1990) 598e609.
[5] H. Schulze, Fallfilmverdampfung und Kondensation an glatten und läng- [41] T. Storch, Ch. Philipp, A. Doeg, U. Gross, Heat flux effects on falling film
sprofilierten Rohren. PhD thesis, Ruhr-Universität Bochum, 1984. evaporation inside a vertical tube, in: 7th World Conf. Exp. Heat T., Fluid
[6] S. Leuthner, A. Harun Maun, S. Fiedler, H. Auracher, Heat and mass transfer in Mech. Thermodynamics. EXHFT-7, 2009.
wavy falling films of binary mixtures, Int. J. Therm. Sci. 38 (1999) 937e943. [42] U. Gross, Kondensation und Verdampfung im geschlossenen Thermosyphon.
[7] A. Lehnberger, Wärmeübergang im Fallfilmverdampfer mit glatten und pro- habilitation treatise, VDI-Verlag, 1991.
filierten Rohren bei kleinen Temperaturgefällen. PhD thesis, TU Braunschweig [43] U. Gross, Falling film evaporation inside a closed thermosyphon, in: Proc. 8th
(Forschrittberichte der VDI 3), 2002. Int. Heat Pipe Conf., 1992, pp. 83e88.
[8] E. Sieder, G. Tate, Heat transfer and pressure drop of liquids in tubes, Ind. Eng. [44] H. Struve, Der Wärmeübergang an einem verdampfenden Rieselfilm, VDI-
Chem. 28 (1936) 1429e1435. Verlag, Düsseldorf, 1969.
[9] F. Weise, Experimentelle Untersuchungen zur Strömung und Wärmeüber- [45] S. Leuthner, Messungen und Modellierungen zum Energie- und Stofftransport
tragung von Reinstoff-Fallfilmen mit hohen Prandtl-Zahlen. PhD thesis, TU in Fallfilmen, in: Fortsch.Ber.VDI, vol. 3, 1999, pp. 1e165.
Braunschweig (ICTV-Schriftenreihe Nr. 3), 2007. [46] S. Thumm, Filmkondensation im senkrechten Rohr bei Gegenstrom von
[10] F. Weise, S. Scholl, Evaporation of pure liquids with increased viscosity in a Dampf und Flüssigkeit. PhD thesis, TU Bergakademie, Freiberg, 2000.
falling film evaporator, Heat Mass Transf. 45 (2007) 1037e1046. [47] T. Storch, Ch. Philipp, U. Gross, Zero shear-stress experiments of falling film
[11] R. Bellinghausen, U. Renz, Heat transfer and film thickness during conden- evaporation inside a vertical tube, in: 5th European Thermal-sciences Conf.
sation of steam flowing at high velocity in a vertical pipe, Int. J. Heat Mass Eurotherm 2008, 2008.
Transf. 35 (1992) 683e689. [48] W.M. Rohsenow, J.P. Hartnett, Y.I. Cho, Handbook of Heat Transfer, third ed.,
[12] J. Müller, Wärmeübergang bei der Filmkondensation und seine Einordnung in McGraw-Hill, 1998.
Wärme- und Stoffübertragungsvorgänge bei Filmströmungen, in: Fortschritt- [49] T. Fujii, Theory of Laminar Film Condensation, Springer-Verlag, New York, 1991.
Berichte VDI Reihe 3: Verfahrenstechnik, 1992. [50] J.H. Kim, T.W. Simon, R. Viskanta, Journal of heat transfer policy on reporting un-
[13] J. Müller, Heat transfer during film condensation e uniform description of certainties in experimental measurements and results, J. Heat Transf. 115 (1993) 5.
heat and mass transfer of film flows and related processes, Chem. Eng. Pro- [51] S.H. Chan, W. Zhang, Rewetting theory and the dryout heat flux of smooth and
cess. 33 (1994) 277e283. grooved plates with a uniform heating, J. Heat Transf. Trans. ASME 116 (1994)
[14] S. Thumm, Ch. Philipp, U. Gross, Film condensation of water in a vertical tube 173e179.
with countercurrent vapour flow, Int. J. Heat Mass Transf. 44 (2001) 4245e [52] B.X. Wang, J.T. Zhang, X.F. Peng, Experimental study on the dryout heat flux of
4256. falling liquid film, Int. J. Heat Mass Transf. (1999) 1897e1903.
[15] P.V. Deo, D.R. Webb, An experimental investigation of the effect of waves on [53] F.E. Andros, Heat Transfer Characteristics of the Two-phase Closed Thermo-
vapour side heat and mass transfer in filmwise condensation inside a vertical syphon (Wickless Heat Pipe) Including Direct Flow Observation. PhD thesis,
tube, Int. J. Multiphase Flow 9 (1983) 37e48. Arizona State University, Tempe, AZ, 1980.
[16] I. Mudawar, R.A. Houpt, Measurement of mass and momentum transport in [54] K. Negishi, T. Sawada, Heat transfer performance of an inclined two-phase
wavy-laminar falling liquid films, Int. J. Heat Mass Transf. 36 (1993) 4151e closed thermosyphon, Int. J. Heat Mass Transf. 26 (1983) 1207e1213.
4162. [55] B.G. Ganchev, V.V. Trishin, Fluctuation of wall temperature in film cooling,
[17] S.V. Alekseenko, V.E. Nakoryakov, B.G. Pokusaev, T. Fukano, Wave Flow of Fluid Mech. Sov. Res. 16 (1987) 17e23.
Liquid Films, Begell House, New York, 1994. [56] T.H. Lyu, I. Mudawar, Statistical investigation of the relationship between
[18] P. Adomeit, Experimentelle Untersuchung der Strömung laminar-welliger interfacial waviness and sensible heat transfer to a falling liquid film, Int. J.
Rieselfilme. PhD thesis, RWTH Aachen, 1996. Heat Mass Transf. 34 (1991) 1451e1464.
[19] G. Karimi, M. Kawaji, An experimental study of freely falling films in a vertical [57] E. Sieder, G. Tate, Heat transfer and pressure drop of liquids in tubes, Ind. Eng.
tube, Chem. Eng. Sci. 53 (1998) 3501e3512. Chem. 28 (1936) 1429e1435.
[20] P. Adomeit, U. Renz, Hydrodynamics of three-dimensional waves in laminar [58] M.A. Mikheev, Izv. Akad. Nauk SSSR 10 (1952) 1448.
falling films, Int. J. Multiphase Flow 26 (2000) 1083e1208. [59] V.V. Yakovlev, Heat transfer in nonboiling water at high thermal loads, Sov. J.
[21] Ch. Philipp, A. Doeg, S. Kufaas Tellefsen, U. Gross, Wave characteristics of At. Energy 2 (1957) 205e207.
liquid films: correlation of heat transfer data with visual observations, in: [60] V.V. Yakovlev, Local and mean heat-transfer for a turbulent flow of nonboiling
Proc. Int. Heat Transfer Conf., 2006. water in a tube with high heat loads, Sov. J. At. Energy 8 (1961) 221e224.
146 T. Storch et al. / International Journal of Thermal Sciences 76 (2014) 137e146

[61] W. Hufschmidt, E. Burck, W. Riebold, Determination of local and averaged Kältemittels R11 und des Kältemittel-Öl-Gemisches R11-51KM33) (in
heat transfer coefficients inside tubes at high heat flux (Die Bestimmung German), TU Dresden, Germany, 1970. PhD thesis.
örtlicher und mittlerer Wärmeübergangskoeffizienten in Rohren bei hohen [64] B.H. Shah, R. Darby, The effect of surfactant on evaporative heat transfer in
Wärmestromdichten), Int. J. Heat Mass Transf. 9 (1966) 539e565 (in German). vertical film flow, Int. J. Heat Mass Transf. 16 (1973) 1889e1903.
[62] H. Struve, Heat transfer to an evaporating falling refrigerant film, in: Proc. [65] F. Weise, S. Scholl, Fluiddynamik und Wärmeübergang bei der Fall-
12th Congress of the Int. Inst. of Refrigeration, 1967, pp. 635e640. filmverdampfung von Reinstoffen mit hohen Prandtl-Zahlen, Chem. Ing. Tech.
[63] C. Elle, Falling film evaporation of refrigerant R11 and refrigerant-oil-mixture 79 (2007) 1145e1153.
R11-51KM33 (Der Wärmeübergang bei der Rieselfilmverdampfung des