1 s2.0 S0735193323005547 Main

International Communications in Heat and Mass Transfer 150 (2024) 107165
Contents lists available at ScienceDirect
International Communications in Heat and Mass Transfer

journal homepage: www.elsevier.com/locate/ichmt
On falling film evaporator – A review of mechanisms and critical

assessment of correlation on a horizontal tube bundle with
updated development
Hao-Yu Lin a, M. Muneeshwaran b, Cheng-Min Yang b, Kashif Nawaz b, Chi-Chuan Wang a, *
a
Department of Mechanical Engineering, National Yang Ming Chiao Tung University, Hsinchu 300, Taiwan
b
Multifunctional Equipment Integration Group, Building Technologies Research and Integration Center (BTRIC), Oak Ridge National Laboratory, Oak Ridge, TN 37831,
United States of America
A R T I C L E I N F O A B S T R A C T
Keywords: Optimization of the energy efficiency and reduction of the greenhouse gas emission for the heat pump system is
Falling film evaporation imperative to meet the net-zero goals at 2050. Chillers with falling film evaporator design not only possess better
Nucleate boiling system performance but also contain less refrigerant inventory. Hence, accurate prediction of the evaporator
Mist flow heat transfer
performance is pivotal especially when charged with low-GWP refrigerant. The present study reviews the cor
Bundle effect
Dry-out
relations for falling film evaporator with a horizontal tube bundle configuration. The major efforts of this study
Enhanced tube include four tasks: (a) literature review of the experimental studies and available empirical correlations; (b)
comprehensive discussion of the falling film evaporation heat transfer mechanism; (c) development of a new
rationally based correlation based on available literature; and (d) comparison of different correlations based on
the existing data. The collected data includes 4114 data points from 8 sources, 6 refrigerants (R-600a, R-290, R-
245fa, R-134a, R-1234ze(E), R-123), 5 types of the tubes (Plain, Turbo-GII-HP, GEWA-B, Low-fin, High-Flux),
liquid Weber number from 2.2 × 10− 6 to 0.7, imposed vapor Weber number from 0 to 37.2, heat flux from 2.5 to
151.5 kW/m2, and film Reynolds number from 1 to 3159.8. The new correlation gives a MAD of 28.3%, and an
R2 of 0.86. Yet, the developed correlation considers various heat transfer mechanisms, including the transition
point from falling film evaporation to the nucleate boiling, local evaporation, dry-out, mist flow, imposed flow,
and enhanced tube effects.
contain less refrigerant inventory. Hence, it is very attractive for heat

pumps installation in both commercial and residential buildings as far as
1. Introduction the energy efficiency and decarbonization are concerned [8]. However,
they differ in terms of heat transfer mechanism and liquid charging
1.1. Background level. In the case of a flooded evaporator, the tube bundle is completely
immersed in liquid, and the primary heat transfer mechanism is nucleate
Reducing the greenhouse gas emission and improving the energy boiling. On the other hand, the liquid film falls from the top of the
efficiency are the critical path to reach the net-zero goals at 2050. One of evaporator and flows alongside the tube bundle in the falling film
the important thermal systems, heat pump, may consume >9% heating evaporator. The main heat transfer mechanisms in the falling film
demand of buildings [7]. Hence, improvement of the system perfor evaporators are boiling and evaporation. Besides, droplet evaporation
mance of the heat pump system is imperative as far as energy efficiency. and forced convection also play a key role to the falling film heat
One of the critical components in the heat pump system is an evaporator. transfer.
Both flooded and falling film evaporators contain a common shell and In typical commercial refrigeration and air-conditioning units that
tube heat exchanger design architecture, as shown in Fig. 1. For larger use flooded evaporators, very large refrigerant charge is needed due to
heat pumps or chiller systems, the evaporator can be of either flooded or significant volume in the shell side of the evaporator. This is quite costly.
falling film configuration. Among these two configurations, the falling Moreover, the current commercial refrigeration units on the market
film evaporator features superior heat transfer characteristics while also
* Corresponding author.
E-mail address: [email protected] (C.-C. Wang).
https://doi.org/10.1016/j.icheatmasstransfer.2023.107165
Available online 6 December 2023

0735-1933/© 2023 Elsevier Ltd. All rights reserved.
H.-Y. Lin et al. International Communications in Heat and Mass Transfer 150 (2024) 107165
Nomenclature Nunb,smooth Nucleated boiling Nusselt number on a smooth tube

Nupartial− wetting Nusselt number with dry-out
A1 , A2 , A3 , B, Ct Coefficients in the Alhusseini et al. [1] research NuTubeBundle Tube bundle Nusselt number
Alam , Awavy , Atur , A Area of laminar, wavy, turbulent, total region, Pcri Critical pressure
respectively. These parameters are proposed by Chun and Pfin Fin pitch of the low-fin tube, m
Seban [2], m2 Pr Prandtl number
ALow− fin− tube Outer surface area of a low fin tube, m Pr Reduced pressure
Asmooth− tube Outer surface area of a smooth tube, m q Heat flux, Wm− 2
a Coefficient for calculating fin efficiency qcri Critical heat flux, Wm− 2
Boff Falling film Boiling number qplateau plateau heat flux (W m− 2), proposed by Roques and Thome
b Unknown parameter set of regression analysis [3,4]
Co , Kff , Kff,plateau , n Coefficients in Roques and Thome [3,4]’s R2 coefficient of determination
research Reff,threshold Dry-out threshold Reynolds number
CP Specific heat capacity, Jkg− 1 K− 1 Reff Reynolds number of falling film
D Outer diameter of tube, m r* Constant for calculating film efficiency
Dbubble Bubble departure diameter, m
S Suppression factor
Dmin Minimum gap between neighbor tubes, m Tsat Saturated temperature, K
E Enhancement factor
tp Tube pitch, m
Emist Enhancement factor due to mist flow
tpmin Minimum tube pitch in the study of Roques and Thome
F Regression formula
[3,4], m
Fnb HTC ratio of boiling on between smooth tube and
Tr Reduced temperature
enhanced tube
Weff Falling film Weber number
FVap,imposed Correction factor of the imposed vapor flow
Wej,Liq Liquid superficial velocity Weber number
g Gravity constant, ms− 2
Wej,Vap Vapor superficial velocity Weber number
Gabubble Galilei number, which characteristic length is referred to
X independent variables
bubble departure diameter
Y dependent variable
Gt ,F,π0 Coefficients proposed by Marcel and Thome [5,6]
y True value
hcv Falling film evaporation HTC, Wm− 2 K− 1
y
̂ Predicted value
hcv,lam Laminar falling film evaporation HTC, Wm− 2 K− 1
y Averaged true value
hcv,tur Turbulent falling film evaporation HTC, Wm− 2 K− 1
hff Falling film HTC, Wm− 2 K− 1 Greek symbols
hnb Nucleated boiling HTC, Wm− 2 K− 1 Γ Mass flow rate, kgs− 1 m− 1
hsmooth HTC on a smooth tube, Wm− 2 K− 1 ΓCri Critical mass flow rate, kgs− 1 m− 1
ilv Latent heat, Jkg− 1 ΓEvap Evaporated mass flow rate, kgs− 1 m− 1
Js Evaporated mass flux, kgs− 1 m− 2 ΓLiq Liquid mass flow rate, kgs− 1 m− 1
jLiq Liquid superficial velocity, ms− 1 ΓLiq,k Liquid mass flow rate of k-th tube in tube bundle, kgs− 1
jVap Vapor superficial velocity, ms− 1 m− 1
Kaff Kapitza number ΓVap Vapor mass flow rate, kgs− 1 m− 1
kl Liquid thermal conductivity, Wm− 1 K− 1 ΓVap,s Total mass flow rate s-th tube in tube bundle, kgs− 1 m− 1
kfin Fin thermal conductivity, Wm− 1 K− 1 ΓVap,acc,k Accumulated mass flow rate of evaporated vapor k-th tube
le Length constant for calculating fin efficiency, m in tube bundle, kgs− 1 m− 1
lfin Fin height of the low-fin tube, m ΓVap,imposed Imposed vapor mass flow rate, kg s− 1 m− 1
M Molecular weight, kgkmol− 1 ηLow− fin− tube Fin efficiency of a low-fin tube
m Square root of the ratio between convective and Φ Constant for calculating fin efficiency
conduction heat transfer ϕr remained percentage CO2 compared to the initial state
N Number of samples ϕx remained percentage refrigerants compared to the initial
NTube,Vertical Number of tubes in a vertical tube bundle state
Nucv Falling film Nusselt number ΔT Temperature difference, K
Nucv,D Falling film evaporation Nusselt number, which δ Film thickness, m
characteristic length is referred to outer diameter of tube δfin Fin thickness of the low-fin tube, m
Nucv,δ Falling film evaporation Nusselt number, which δ+ Dimensionless falling film thickness in Alhusseini et al.
characteristic length is referred to film thickness [1]’s research
Nucv,Low− fin Falling film evaporation Nusselt number on a low-fin μl Liquid dynamic viscosity, Pa s
tube ν Kinematic viscosity, ms− 1
Nulam Laminar falling film evaporation Nusselt number ρl Liquid density, kgm− 3
Nulam,δ Laminar falling film evaporation Nusselt number, which ρv Vapor density, kgm− 3
characteristic length is referred to film thickness σ Surface tension coefficient, Nm− 1
Nutur Turbulent falling film evaporation Nusselt number
Nutur,δ Turbulent falling film evaporation Nusselt number, which Acronym
characteristic length is referred to film thickness AD Absolute Deviation
Nufull− wetting Nusselt number without dry-out CFC Fluorocarbon
Nunb Nucleated boiling Nusselt number GWP Global Warming Potential
Nunb,enhanced Nucleated boiling Nusselt number on an enhanced tube HCFC Hydro Chloro Fluoro Carbon
Nunb,Low− fin Nucleated boiling Nusselt number on a low-fin tube HFC Hydro Fluoro Carbon
2
HFO Hydro Fluoro Olefins SSTO Sum of squared total

2 1
HTC Heat transfer coefficient, Wm− K− SSE Sum of squared error
MAD Mean Absolute Deviation
employ high-GWP refrigerants, such as R-134a, in their systems. Using the Rohsenow-Nusselt expression to axisymmetric bodies, including
these high-GWP refrigerants in the flooded evaporators that demand cylindrical and rectangular geometries. They included the variable
huge refrigerant inventories may lead to serious concerns in global gravity effects in terms of effective gravity and the influence of the body
warming issues. Hence, the industry is recently moving toward compact shape in their expression, which made the modified Rohsenow-Nusselt
evaporators that can minimize the refrigerant inventory [9] in associa expression more suitable to different body shapes, including horizon
tion with the low-GWP refrigerants [10–13] to achieve net-zero goals in tal cylinders, vertical plates, rotating plates, cones, and spheres.
2050. The falling film evaporators are regarded as a potential replace Similarly, semi-analytical models were developed in the past to
ment for the flooded evaporators because of their reduced refrigerant predict the heat transfer performance of falling film evaporators. Lorenz
charge inventory and superior heat transfer performance. In fact, the and Yung [21] introduced a model that included the effects of thermally
studies showed that the falling film evaporators could minimize the developing and fully developed regions. However, it was found that the
refrigerant charge by around 30% as compared with the flooded evap length of the thermally developing region was miscalculated in their
orators [14]. Moreover, in addition to the applications of refrigeration model. Similarly, Cerza and Sernas [22] introduced an analytical model
system, water is also widely used in the falling film evaporators are for the thermally developing region of falling film evaporation on a
widely utilized in other applications such as nucleate reactor cooling vertical plate. Note that these models do not account for the heat
[15], and desalination [16]. However, the aim of this study stresses on transfer enhancements caused by the jet impingement of liquid film.
the literature regarding refrigeration applications only. Parken and Fletcher’s [23] model included the effects of heat transfer
Although falling film evaporators offer superior heat transfer per augmentation by the jet impingement of liquid film around the apex of
formance than that of flooded evaporators, the liquid film flow and heat the tube. Additionally, Sabin and Poppendiek [24] and Nakazatomi and
transfer behavior are much more complex with the falling film evapo Bergles [25] proposed a falling film evaporation model in the thermally
rators. Moreover, the falling film evaporators are susceptible to dry-out fully developed region based on Nusselt’s theory.
issues if they are not properly designed. Mechanism-based design tools Chyu and Bergles [26] found that the above models could lead to an
(i.e., empirical correlations in this case) are imperative upon designing erroneous result because they were developed without or partially
the performance of falling film evaporators. However, the correlations considering the effects of: (1) convection heat transfer at higher film
available in the open literature are derived mainly based on specific flow rates; (2) jet impingement at the apex of the tube; and (3) a ther
design conditions, such as tube type (smooth or enhanced), tube size, mally developing region and a fully developed region. Moreover, most
operational condition, and type of refrigerant. Extrapolation of the of the models were derived based on the assumption of pure conduction
correlations outside its database is uncertain. The subsequent review of along the liquid film, which is true only for thin films but not for rela
the existing correlations about their predictive ability will be addressed tively thicker films. Chyu and Bergles [26] developed a model that ac
thoroughly. And it will be shown shortly that the applicability of these counts for the effects of stagnation, jet impingement, thermally
correlations are quite limited. Yet, all the available correlations lack the developing region, and fully developed region. They derived two
database of low-GWP refrigerants (e.g., hydrofluoroolefins (HFO)). Note models, and the first one used the assumption of the pure conduction
that according to the Paris agreement, HFO refrigerants are the major based on the Nusselt’s theory and the second model adopted the Chun
potential replacement for hydrofluorocarbons (HFC), as shown in Fig. 2. and Seban [27] correlation in a fully developed region. Their results
In essence, the major efforts of this study include four tasks: (a) showed that the second model results were well matched with the
literature review of the experimental studies and assessment of the experimental results, whereas the first model, based on the pure con
available empirical correlations; (b) comprehensive discussion of the duction assumption, underpredicted the heat transfer coefficients.
falling film evaporation heat transfer mechanism; (c) development of a
new rationally based correlation based on available literature; and (d)
comparison of different correlations based on the existing data. 1.3. Falling film heat transfer mechanism
1.3.1. Falling film evaporation and nucleate boiling

1.2. Analytical models The falling film heat transfer is the combination of film evaporation
and the nucleate boiling. At the earliest stage of development, re
In a falling film evaporator, heat transfer occurs through film evap searchers studied the falling film subject to the low heat flux which is
oration, nucleate boiling, and mist evaporation. A complex heat transfer lower than the onset condition of nucleated boiling. Therefore, most the
mechanism associated with the falling film evaporator makes it difficult historical researches assumed the dominant heat transfer is the liquid
to develop an analytical solution. However, the early research on falling film evaporation. As expected, the boiling mechanism may become more
film evaporators had developed a simplified analytical model based on and more important with the rise of heat flux.
several assumptions. Nusselt [17] studied falling film evaporation and The general assumption of a smooth film interface was considered in
condensation on a vertical wall with constant heat flux. He derived an most of the analytical models, but this may not be true in the real sce
expression to predict the heat transfer by solving the Navier-Stokes nario and could lead to erroneous predictions. Typically, a wavy liquid
equation. However, it was derived based on several simplified assump film interface is observed in the falling film evaporators, and this phe
tions that may not be applicable to the complex falling film evaporators, nomenon still prevails even at small Reynolds numbers. The wavy film
which could lead to erroneous predictions. Subsequently, researchers could induce flow disturbance and enhance heat transfer, and incorpo
attempted to improve Nusselt’s development, which can be used for rating the wavy film effects into an analytical model can be more
different geometries his theory turns out and flow conditions in falling complex. In this regard, the researchers experimentally studied the
film evaporators. For instance, Rohsenow [18] and Sparrow and Gregg falling film flow and heat transfer behavior of various tubes and re
[19] corrected the linear temperature profile assumption and relaxed frigerants under different operating conditions. Based on the experi
the negligence of the inertial term in the momentum equation in con mental results, empirical correlations were developed to predict the heat
ventional Nusselt theory. Dhir and Lienhard [20] attempted to extend transfer coefficients, which account for the complex flow and heat
3
transfer mechanisms associated with the falling film heat transfer. The falling film heat transfer correlations for full wetting and partial dry-out
correlations related to falling film heat transfer are listed in Table 1. conditions, which were constructed using different refrigerant data such
Chen and Seban [28] were the first to introduce the heat transfer cor as R-134a, R-1234ze(E), R-32, R-600a, and R-290. It should be
relations for vertically falling film. They developed the correlation in emphasized that the above developed correlations were based on their
terms of the Reynolds number and the Prandtl number. They also own experimental data. Recently, based on the above studies experi
observed a sudden reduction in wall temperature at the bottom of the mental data bank, Shah [37–39] constructed a general correlation that is
tube under certain operating conditions, indicating a nucleate boiling applicable for a wider range of operating conditions and different re
phenomenon in falling film evaporation. The correlation developed by frigerants, such as isobutane, propane, ammonia, water, R-11, R-123, R-
Chun and Seban [28] may not be suitable for fluids with a Prandtl 1234ze(E), R-134a, R-245fa, and R-32.
number >5. Hence, Alhusseini et al. [1] developed a correlation that can
cover the Prandtl number range of 1.73–46.6. Their correlation for wavy 1.3.2. Droplet entrainment effect
laminar film contained the Reynolds number and Kapitza number, while As aforementioned, most of the correlations for falling film heat
they employed the semi-empirical correlation approach for turbulent transfer were developed based on the nucleate boiling or falling film
regions. evaporation heat transfer mechanisms, which are considered to be the
Similar to Chun and Seban [28], Lorenz and Yung [21] observed a primary heat transfer mechanisms in falling film heat transfer. However,
nucleate boiling phenomenon in addition to the falling film evaporation; in addition to nucleate boiling and film evaporation, the droplet
hence, they added the nucleate boiling term (in addition to the falling entrainment effect could also contribute to the enhancement of falling
film evaporation) into their model. Roques and Roques and Thome [3,4] film heat transfer. The liquid film can be superheated under certain
observed that there is barely any variation of heat transfer coefficients operating conditions when it flows along the tubes. When it is super
when Reynolds number exceeds certain threshold value (referred to as a heated, it may detach from the tube and form a group of tiny droplets.
threshold Reynolds number), indicating an intensive nucleate boiling in Then, evaporation can occur on the surface of those superheated tiny
falling film heat transfer. Upon well above the threshold Reynolds liquid droplets. If the detached tiny droplets are not fully evaporated,
number, the heat transfer coefficients dramatically reduced due to wall they may rejoin the liquid film when their temperature reaches satura
dry-out. This observation by Roques and Thome [3,4] confirmed the tion. The cyclic detachment, tiny droplet formation, and rejoining
impacts of nucleate boiling on falling film heat transfer. Consequently, phenomena is referred to as the droplet entrainment effect. With the
Chien and Cheng [29] used the superposition model to determine the droplet entrainment, the falling film heat transfer can be significantly
heat transfer coefficients that combine the influence of nucleate boiling enhanced due to the evaporation of a considerable number of tiny
and falling film evaporation heat transfer, which is basically derived droplets.
from the superposition principle used by Chen [30] for convective Zadrazil et al. [40] conducted a flow visualization experiment to
boiling heat transfer coefficients. Following the superposition principle, examine the droplet entrainment effect in the vertical tube. The co-
Chien and Tsai [31] and Chien and Chen [32] proposed the empirical current flow of water and air on a vertical tube was observed using
correlations for R-254fa and R-134a, respectively. Similar method is also particle tracking velocimetry and laser-induced fluorescence tech
suggested in Lin et al.’s [33] correlation, their correlation is available for niques. The smooth liquid film turned into a wavy liquid film when the
R-245fa, R-1234ze(E), R-123, R-32, R-410 A, R-290, R-600a. Alterna liquid velocity was higher than that of the gas velocity. The transition of
tively, Jige et al. [34] used the power law to combine the convective and smooth film into wavy film is mainly caused by interfacial shear stress
nucleate boiling heat transfer in falling film. Zhao et al. [35] developed and flow instability. The wavy films can cause the droplet entrainment
the correlations based on R-134a data, and the correlation can satis effect as shown in Fig. 3. Additionally, the bubble (i.e., gas phase) that
factorily predict the falling film heat transfer under both full wetting and was trapped in the liquid film may rupture the liquid interface and burst
partial dry-out conditions. Similarly, Jin et al. [36] introduced the into the vapor phase due to the flow instability and the pressure
Fig. 1. Schematic of (a) flooded evaporator and (b) falling film evaporator.
4
Fig. 2. Global warming potential requirement under different protocol.
difference between the inside and outside of the bubble. This bubble stationary dry patch in the falling film. Followed Based on the similar
burst phenomenon could also lead to a droplet entrainment effect in the concepts of Zuber and Staub [49], He et al. [50] had accounted for the
falling film. Fujita et al. [41–43] studied the droplet entrainment effect curvature effect. They developed a two-dimensional model to predict
and concluded it was caused by bubble bursting. They showed that the the growth of the dry patch in the falling liquid film over the horizon
droplet entrainment effect strongly depends on the evaporation rate, tally heated tubes.
vapor flow and the film thickness. Gogonin [44] extended Fujita et al. In addition to the analytical approach, various empirical correlations
[41–43] ‘s experimental work, and he bridged the heat transfer were proposed to predict the dry-out of falling film evaporation. The
enhancement to the entrainment of liquid by the Weber number and the criteria were either based on the morphological change through the
vapor-to-liquid density ratio. Gao and Li [45] showed that liquid visualization or the transition point of heat transfer coefficient. Roques
splashing could occur at the apex of the tube due to liquid crown for and Thome [3,4] experimentally investigated the falling film heat
mation. Guo et al. [46] and Bao et al. [47] conducted the numerical and transfer of R-134a on a vertical array for four types of horizontal tubes
experimental study regarding the droplets impact on a thin liquid film including plain, Turbo-BII HP, GEWA-B, and High-Flux tubes. They
which contains bubbles via boiling. They showed the droplets would proposed using a threshold Reynolds number to determine the onset of
disturb the collision of the bubbles and further enhance the liquid crown dry patch which is a strongly related to heat flux. However, the empir
formation. Inoue and Maeda [46] theoretically analyzed the droplet ical constants in the formula vary with the test fluid and tube, which
entrainment from the flat liquid film and also attributed the entrainment limits the application of this method. Habert and Thome [51,52]
is originated from the shear stress. Zhao et al. [47] reviewed the vapor developed a quite general method for prediction of the onset of dryout
flow effect on falling film evaporation under various operating on plain and enhanced tubes subject to falling film conditions. They
conditions. further defined a single set of dimensionless number to predict the
threshold Reynolds number, which reduces the empirical constants
1.3.3. Dry-out required for the model. Marcel and Thome [5,6] based on the above
Falling film evaporation is able to keep the tube wall at low tem method and proposed an additional geometric factor to consider the
perature while the wall is covered with liquid film. As the heat flux in effect of enhanced tubes. Shah [39] comprehensively reviewed the dry-
creases or film flow rate decreases to a critical level, the film starts out of falling film on a horizontal plain tube in open literature and
breaking up and leads to a dry area. The heat transfer within the dry developed a correlation that is applicable for a wide range of data. Seven
patch relies on the vapor-phase convection, and thus the overall heat fluids including R-11, R-123, R-134a, R-236fa, R-410 A, isobutane, and
transfer coefficient significantly decreases. In essence, dry-out signifi propane from eight sources were used for validation. Table 2 summa
cantly jeopardizes the heat transfer process of the falling film evapora rizes the empirical correlations for predicting the dry-out of the falling
tion and should be averted. Hence, reliable prediction of the occurrence film.
of dry out is essential.
To predict and further avoid the formation of the dry patch in the 2. Methodology
falling film, many researchers had conducted in-depth studies and pro
posed various theoretical models. Hartley and Murgatroyd [48] From the literature review, it is understood that a comprehensive
analyzed a thin liquid film flowing over a solid surface and derive the universal correlation for falling film heat transfer in a horizontal tube
criteria for the occurrence of dry patch. It was assumed that a continuous bundle is still not available. Therefore, this study aimed to develop an
liquid stream will breakdown into individual rivulets if the flow rate is updated rationally based correlation which includes various heat
reduced to a critical value. The force balances between gravity, liquid transfer mechanisms for falling film heat transfer on tube bundles. The
viscosity, and surface tension in association with the minimum energy correlation was developed using a much larger data bank available in
analysis were conducted to predict the minimum wetting rate and film the open literature, and the following assumptions were considered for
thickness. Zuber and Staub [49] extended the work of Hartley and the development of correlation:
Murgatroyd [48] and considered the thermal effects due to the thermo-
capillary force and evaporation thrust for the heated wall. The derived (1) A shell-and-tube evaporator configuration with a horizontal tube
criteria can predict the minimum film thickness, the minimum flow rate arrangement was considered.
and minimum surface shear that can be maintained without forming a
5
Table 1
Empirical correlation for falling film evaporation HTC.
Authors Refrigerant Reff Heat flux, Tube type Correlation
kWm-2
Alhusseini et al. Water, Propylene 34- 7.8-19.5 Single; Smooth ( )1/5

Nucv = Nu5cv,lam + Nu5cv,tur
[1] glycol 15,600 (vertical)
0.158
Nucv,lam = 2.65Reff − Kaff 0.0563
δ+1/3 Pr
Nucv,tur = )
(A1 Pr + A2 Pr + A3 Pr1/4 + Ct + Kaff 1/2 Pr1/2 B
3/4 1/2
( )
Roques and Thome R-134a <3000 19.1-73.9 Tube bundle; hff Reff
Kff = = Kff,plateau
[3,4] Smooth, hnb Reff,threshold
Turbo-BII HP, hff = Co qn
Gewa-B, ( )[
tp ( q ) ( q )2 ]
High-Flux Kff,plateau = 1 + b1 b2 + b3 + b4
tpmin CHF CHF
0.25
qcrit,nb = 0.131ρ0.5
v ilv (gσ(ρl − ρv ) )
( )
Reff,threshold = 2 cqplateau + d
( ) ( )
q q0 2
qplateau 2 − 2c qplateau − 2d =0
Reff Reff
Marcel and Thome R-236fa, 0-12,000 20-80 Tube bundle; hff D
= 9.623 × 104 (π0 )0.0328 G1.2449 F
[5,6] R-134a Turbo-Bii HP, kl t
Gewa-B, q2 D(ρl − ρv )
High-Flux, π0 =
ilv 5/2 μl
Turbo-EDE2, Gewa- ( )
B4, Reff
F = min 1,
Turbo-B5, Reff,threshold
Gewa-B5 ( )
qD 1.04 0.175
Reff,threshold = 20.721 Gt
ilv μl
( )
Chien and Cheng R-11, 157- 2-100 Single, Smooth; hff = 0.4 + 3737.648Weff 0.5896 Reff − 0.20327 Boff − 0.7433
hnb + hcv
[29] R-123, 2500 Turbo-B (
hcv = hlam 5 + htur 5
)1/5
R-134a,
R-141b,
R-22
( )
Chien and Tsai R-245fa 115-372 4.5-48.5 Single; Smooth hff = 0.0152Weff 0.2833 Reff 1.2536 Boff 1.1789 hnb + hcv
[31] [( )1/3 ]
ν2
hcv = 0.0386Reff 0.09 Pr0.986 / /k
g
hnb = 537P0.12−
r
0.2log10 (Roughness)
( − log10 (Pr ) )0.55 M− 0.5 q0.67
( )
Chien and Chen R-134a 165-750 4.5-48.5 Single; Smooth hff = 56.13Weff 0.5878 Reff 0.2457 Boff − 0.1798 hnb + hcv
[32] ( ) 1/5
hcv = hlam 5 + htur 5
hnb = 537P0.12−
r
0.2log10 (Roughness)
( − log10 (Pr ) )0.55 M− 0.5 0.67
q
Zhao et al. [35] R-134a 579- 10-168 Single; Smooth Nufull− wetting = 0.00464Reff Boff 0.43 Pr0.15 Weff − 0.45
1.51
2702 Nupartial− wetting = 3.58 × 10− 9 Reff 2.89 Boff 0.37 Pr0.2 Weff − 1.13
Reff,threshold = 53600Boff 0.0045 Pr− 0.52 Weff 0.5

Jige et al. [34] R-1234ze(E), 100- 2.5-40 Single; Smooth (
hff = h5cv + h5nb
)1/5
R-245fa 1000 ( )1/5
hcv = h5cv,lam + h5cv,tur
Jin et al. [36] R-134a, 198- 5-150 Single; Smooth Nufull− wetting = 23.3Reff 0.817 Boff 0.6331 Pr− 0.0864
R-290, 3410 Nupartial− wetting = 11.7Reff 0.8931 Boff 0.5278 Pr− 0.0287
R-600a,
R-32,
R-1234ze(E)
( )
Shah [37] Water, 19- 1-208 Single; Smooth hff = max hcv,lam , hcv,tur + hnb
ammonia, 10,734 {
55Pr 0.12
( − 0.4343lnPr )− 0.55 M− 0.5 q0.67 , HC − refrigerent
R-11, hnb = ( )
R-32, 3.596 × 10 q Pcri 0.69 1.8Pr 0.17 + 4Pr 1.2 + 10Pr 10 , other − refrigerent
− 5 0.7
R-123,
R-134a,
R-245fa,
R-1234ze,
propane,
isobutane
Lin et al. [33] R-245fa, 83–3500 2.5–168 Single; Smooth Nucomb = SNunb + ENucv
R-1234ze(E), ( ρ )-1 ( q )0.99
S = Prl 0.474 Reff 0.968 2Boff Kaff 0.565 Gabubble 1 Pr -0.037 π0 -0.883 l
( )1
R-123, ρv qcri
R-32, ( ρ )1 ( q )− 1
Kaff − 0.242 Gabubble -1 Pr -0.253 π0 0.418 l
( )0.46
R-410 A, E = Prl − 0.465 Reff 0.642 2Boff
ρv qcri
R-290,
R-600a
6
(2) A liquid distributor was assumed to be at the top of the shell The appropriate independent variables X and the relationship F were
chamber, which distributes a uniform and steady liquid film flow selected based on the trend of experiment data. MAD (mean absolute
alongside the tube bundle film without splashing. The whole deviation) and R2 [54] values are used in the present study to evaluate
liquid film from the distributor must fall on the top tube. the agreement between the experimental data and predictions. MAD
(3) The vapor exhaust was considered to be at either the top or represents the average of the percentage difference between the exper
bottom of the shell chamber, and the uniform and stable vapor imental data and predicted values, as shown in Eq. (2).
flow prevailed.
1∑
(4) A liquid level controller was considered to be at the bottom of the MAD = AD (2)
N
shell chamber to avoid flooding.
(5) The tube bundle arrangement can be either inline or staggered. ⃒ ⃒
⃒y − ̂
y ⃒⃒
(6) Uniform heat flux was considered on the tube surface in both the AD = ⃒⃒ (3)
y ⃒
circumferential and longitudinal directions.
where Nis the number of samples, y and ̂ y represent the experiment and
predicted value, respectively. The coefficient of determination, R2, is
2.1. Regression method employed for further assessment, which establishes a relationship be
tween the SSE (sum of squared error) and the SSTO (sum of squared
Regression analysis is a process for establishing the mathematical total), as shown in Eqs. (4)–(6). An R2 value of 1 represents the perfect
relationship between a dependent variable and one or more independent prediction, while a value lower than 1 indicates the deviation between
variables. The following iterative process was developed to establish the the prediction and experimental data. It should be noted that R2 does not
HTC correlation: necessarily have to be a positive value. A negative value of R2 indicates
that the prediction is simply worse than the mean value of all data.
(1) Collect the dependent variable Y (i.e., from the available exper
SSE 2
imental HTC). R2 = 1 − ,R ≤ 1 (4)
(2) Choose the independent variables X. SSTO
(3) Assume a formula F for describing the relationship between the where
independent variables X, the dependent variable Y, and the un ∑
known parameters b. SSE = y )2
(y − ̂ (5)
Y = F(X, b) (1) ∑
SSTO = (y − y)2 (6)
(4) Find the unknown parameter set b using the non-linear least
2.2. Collection and processing of experimental data
squares method.
(5) Conduct the analysis to ensure the prediction agrees with the
2.2.1. Range of the experimental data
experimental data (i.e., HTC). If the predictive deviation is out of
A total of 4114 experimental data were collected and details of the
the tolerance, iterative analysis repeats from step (2).
test data are listed in Table 3. The test data covers a wide range of design
and operating conditions, including 6 refrigerants (R-600a, R-290, R-
245fa, R-134a, R-1234ze(E), and R-123), 5 types of tubes (plain, Turbo-
GII-HP, GEWA-B, Low-fin, and High-Flux), liquid Weber number
ranging from 2.2 × 10− 6 to 0.7, imposed vapor Weber number of 0–37.2,
heat flux in the range of 2.5–151.5 kWm− 2, and film Reynolds number of
1–3159.8. The large data bank ensures that the proposed correlation
incorporates the complex heat transfer and the flow behavior effects in
the falling film evaporators.
Upon organizing the collected data, the self-designed enhanced tubes
were also classified within the commercial tube category. For instance,
the self-designed 3D structured tube used by Zhao et al. [43,44] was
treated as GEWA-B tube in this study.
2.2.2. Data processing

This section provides an overview of the calculation process used to
derive the parameters from the experimental data listed in Table 2. The
thermophysical properties of the refrigerants were obtained from an
open library, CoolProp [57]. The parameters derived from the experi
mental data are presented below:
(1) Estimate the flowrate (single side local mass flow rate per unit
length) for both liquid ΓLiq and vapor ΓVap :
The liquid and vapor flowrate interact through the evaporated
mass, ΓEvap , which is calculated by Eq. (7). Note that only half of
the surface area is considered because only one side of the tube is
taken into consideration due to symmetry.
( )
q πD
ΓEvap = (7)
ilv 2
Fig. 3. Droplet entrainment caused by breaking wavy film and bursting bub
bles [40].
7
Table 2
Correlations for estimation of onset of dry-out.
Authors Refrigerant Tube type Equations
Hartley and Murgatroyd [48] Not specified Smooth (vertical plate) (μ ρ )1/5
Γcri = 1.69[σ(1 − cosθ) ]3/5 l l
g
Zuber and Staub [49] Not specified Smooth (vertical plate) [
ρl g(ρl − ρv )
]2 [
3Γcri μl
]4/3
σ(1 − cosθ) ∂σ q
=[ ]1/3 + cosθ
15 μl g(ρl − ρv )ρl 3Γcri μl ∂T kl
g(ρl − ρv )ρl
[ ]1/5
q (ρl − ρv ) 2
+ρl cos θ
ρl ilv ρl
Braines et al. [53] Water, Smooth (vertical plate) ( 6c )1/3 ( σ )1/3 ( ρ )2/3
2 l l
R-113 CHF = ρv ilv Ul
c c c ρl Ul L ρv
(1 3 4 )
Roques and Thome [3,4] R-134a Smooth, Reff,threshold = 2 cqplateau + d
Turbo-BII HP,
( ) ( )
Gewa-B, q0 q0 2
qplateau 2 − 2c qplateau − 2d =0
High-Flux Rff Reff
Smooth,
R-134a, Gewa-C + LW, ( qD )b
Habert [51,52] Reff,threshold = a
R-236fa Turbo-EDE2, μl ilv
Gewa-B2
Zhao et al. [35] R-134a Smooth Reff,threshold = 53600Boff 0.0045 Pr− 0.52 Weff 0.5
( )
He et al. [50] Water Smooth ∂δ ρg ∂δ
= l 3δ2 sinθ + δ3 cosθ − Js
∂t 2μl R ∂θ
R-134a, ))
( ( qD
Shah [39] R-236fa, Smooth Reff,threshold = 938 1 − exp − 0.055
μl ilv
R-410 A,
Fig. 4 illustrates the interaction among liquid flow, evaporated The minimum distance between tubes, Dmin , is adopted for the
vapor flow and accumulated vapor flow in a tube bundle. The characteristic length, as shown in Fig. 7. The superficial velocities
mass flowrate of liquid (ΓLiq ) decreases in the downstream due to of liquid and vapor phases are expressed in Eqs. (11) and (12),
the evaporation, and it can be expressed from Eq. (8). respectively.
ΓLiq,k = ΓLiq,k− 1 − ΓEvap,k− (8) ΓLiq
1
jLiq = (11)
ρl Dmin
ΓVap
jVap = (12)
Note that the vapor flowrate increases at the downstream of vapor ρv Dmin
flow direction as a result of vapor accumulation, and the direction of
vapor flow depends on the location of the vapor outlet. The vapor outlet
can be installed in three ways: at the top only, at the bottom only, or at
the both top and bottom of the evaporator, as illustrated in Fig. 5. The (3) Weber number:
cumulative vapor flowrate for each tube can be calculated using Eq. (9). Weber number is a crucial dimensionless number to assess the
droplet entrainment phenomenon in two-phase flow, which
∑ represents the ratio of the inertial force to cohesion force. The
s=ks
ΓVap,acc,k = Γevap,s (9)
s=0 role of the Weber number on the droplet entrainment will be
discussed in the Section 2.3.3 (Mist flow effect). In Eqs. (13) and
In the equation, the index k represents the tube index alongside the
(14), the superficial velocity is employed as the characteristic
vapor flow direction (i.e., based on the vapor outlet location), and the
velocity for calculating the Weber number.
cumulative vapor flowrate for the kth tube is obtained by summing the
vapor flow rate from the beginning of the vapor stream (s = 0) to the kth ρl j2Liq Dmin
Wej,Liq = (13)
tube (s = ks). σ
Besides the evaporated vapor flow, additional vapor flow can be
imposed through liquid distributor in an actual falling film evaporator, ρv j2Vap Dmin
Wej,Vap = (14)
as shown in Fig. 6. The reason causes the imposed vapor flow will be σ
illustrated in Section 2.3.4 (Imposed vapor flow effect). Thus, the
imposed vapor flow ΓVap,imposed should be also considered when calcu
(4) Film Reynolds number:
lating the effective vapor flow ΓVap , as expressed in Eq. (10).
Reynolds number is a dimensionless number that characterizes
ΓVap,k = ΓVap,acc,k + ΓVap,imposed (10) the ratio of inertia force to viscous force. Film Reynolds number is
used to determine the liquid film thickness upon calculating the
HTC of falling film.
(2) Superficial velocity:
Superficial velocity is a commonly used parameter in two- 4ΓLiq
Reff = (15)
phase flow analysis, it averages the volumetric flow rate of the μl
liquid phase and vapor phase by the total cross-section area of the
channel. Zadrazil et al. [40] pointed out that the high liquid and
vapor superficial velocity leads to the onset of the droplet (5) Boiling number:
entrainment. Superficial velocity is obtained by dividing the Boiling number is a dimensionless number that quantifies the
volumetric flowrate of each phase by the total cross-section area.
8
ratio of heating power to the latent heat of the liquid flow, which (7) Nucleate boiling Nusselt number for smooth tube:
plays a key role in determining the onset of dry-out. The nucleate boiling Nusselt number for the smooth tube
( ) Nunb,smooth is calculated by Jung et al. [59] correlation, as shown
q˝ πD
Boff = (16) in Eqs. (22) to (25). Jung et al. [59] correlation was developed
ΓLiq ilv 2
based on the horizontal tube of 19 mm outer diameter and the
different refrigerants data, including R-123, R-11, R-142b, R-
134a, R-12, R-22, R-125, and R-32. The experiment data con
(6) Falling film evaporation Nusselt number:
ducted by Jung et al. [59] are close to the present collected falling
The falling film evaporation Nusselt number Nucv is the com
film heat transfer data, especially the tube diameter and
bination of laminar film Nusselt number (Nulam,δ ) and turbulent
refrigerants.
film Nusselt number (Nutur,δ ) [58] with the fifth-order power law
suggested by Alhusseini et al. [1], as shown in Eqs. (17)–(19). hnb D
Nunb = (22)
( )1/5 kl
Nucv,δ = Nu5lam,δ + Nu5tur,δ (17) [ ]a
kl qDbubble
hnb,smooth = 10 Pr 0.1 (1 − Tr )− 1.4
(Prl )− 0.25
(23)
where Dbubble kl Tsat
( )0.309
Nulam,δ = 2.65Reff − 0.158
Kaff 0.0563 (18) ρ
a = 0.855 v Pr − 0.437
(24)
ρl
Nutur,δ = 0.03Reff 0.2
Pr 0.7
(19)
In order to have a reasonable comparison between the falling

film evaporation Nusselt number and nucleate boiling Nusselt [ ]0.5
number, the characteristic length of Nucv is converted from film 2σ
Dbubble = 0.511 (25)
thickness δ to the tube diameter D, as shown in Eqs. (20) and (21). g(ρl − ρv )
D
Nucv,D = Nucv,δ (20)
δ (8) Nucleate boiling Nusselt number for enhanced tube:
The nucleate boiling Nusselt number for the enhanced tube is
( 2 )1/3
ν calculated by the empirical correlation proposed by Roques and
δ= (21)
g Thome [3,4], as shown in Eqs. (26)–(27) and Table 4. Note that
the smooth tube experiment conducted by Roques and Thome
[3,4] is also calculated by Eqs. (26) to (27), rather than from the
Jung et al. [59] correlation. This is because the tube roughness in
Table 3
The conditions of collected experiment data.
Authors and Fluid Tube type Tube Channel Heat flux, Reff Wevap Wevap,imposed Tsat , Uncertainty of
year diameter, mm diameter, mm kWm-2 C HTC, %
◦
Roques and R-134a Smooth, 18.84–18.91 25.04–25.08 19.1–73.9 1.4–2989.2 0.0018–0.687 0 4.85 ±22%
Thome, GEWA-BII-
2007 [3] HP, GEWA-B,
High Flux
Chien and R-123, Smooth, 19 23 5.6–48.3 160.5–392 3.3 × 10− 4- 0 2–20 Smooth:
Tsai, 2011 R-245fa Low-fin 0.304 ±14%;
[31] Low-fin:
±44.3%
Chien and R-134a Smooth, 19 23 4.2–48.5 94.8–668.5 9.06 × 10− 5- 0 2–26 Smooth: ±8%;
Chen, Low-fin 0.0984 Low-fin:
2012 [32] ±25%
Zhao et al., R-134a Smooth 16–25.35 21.23–21.7 10.2–168.6 577.9–2702.9 6.12 × 10− 6- 0 6–16 ±20%
2016 [35] 0.0167
Zhao et al., R-134a GEWA-B 19.05 3.4 20–40 42.1–1569.6 0.01–60.7 0–37.2 6 ±28%
2018 [55]
Zhao et al., R-134a GEWA-B 19.05 3.4 1.2–39.3 1.7–3159.8 0.013–11.489 0 6–16 ±25%
2018 [56]
− 6
Jin et al., R-600a, Smooth 19.05 215 5–151.5 189.9–2136.5 6.8 × 10 - 0 6–10 ±25%
2019 [36] R-134a, 0.012
R-290
Jige et al., R-245fa, Smooth 19.05 60.48 2.5–20 107.4–897.3 1.61 × 10− 5- 0 10–20 ±20%
2019 [34] R-1234ze 0.0015
(E)
Total R-245fa, Smooth, 18.84–19.05 3.4–215 1.2–73.9 1.4–3159.8 6.8 × 10− 6- 0–37.2 2–26 ±14–44.3%
R-1234ze Low-fin, 60.7
(E), GEWA-BII-
R-123, HP, GEWA-B,
R-134a, High Flux
R-290,
R-600a
9
the experiment conducted by Roques and Thome [3,4] is higher ⎧

than that the tests carried out by Jung et al. [59]. ⎪
⎪ − b * −
⎨ a(mle ) , if Φ > 0.6 + 2.257(r )
0.445
ηLow− fin− tube = tanh(Φ)

hnb,enhanced D ⎪ , if Φ ≤ 0.6 + 2.257(r* )− 0.445
Nunb,enhanced = (26) ⎪
⎩
Φ
kl
a = (r* )− 0.246
where
Φ = mle (r* )n
hnb,enhanced = C(q)n (27)
n = exp[0.13ln(mle ) − 1.3863 ] (31)
The low-fin tube HTC is considered as the smooth tube HTC with the {
effect of the additional fin surface area and the fin efficiency of the 0.9107 + 0.0893r* , if r* ≤ 2
b=
circular fin, as shown in Eqs. (28) or (29). 0.9706 + 0.17125r* , if r* > 2
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
ALow− fin− tube
hnb,low− fin = hnb,smooth η (28) m=
2hsmooth
Asmooth− tube Low− fin− tube
kfin δfin
ALow− fin− tube δfin

hcv,low− fin = hcv,smooth η (29) le = lfin +
Asmooth− tube Low− fin− tube 2
where
(9) Boiling HTC enhancement ratio:
⎧ ( ) ⎫ It represents the ratio of enhanced tube boiling HTC to the
⎪
⎪ πD Pfin − δfin ⎪
⎪
ALow− fin− tube
⎪
1 ⎨ +π( D + lfin )δfin
⎪
⎬ smooth tube boiling HTC, as expressed in Eq. (32).
= (30)
Asmooth− tube πDPfin ⎪
⎪ [( ] ⎪
⎪ hnb,enhanced
⎪ )
⎩ +2 × π D + lfin 2 − D2 ⎪
⎭ Fnb = (32)
4 hnb,smooth
where Pfin , δfin , and lfin represents fin pitch, fin thickness, and fin height,
respectively. Note that D denotes for the outer diameter at the fin base.
Fin efficiency [60] is expressed in Eq. (31). 2.3. The developed correlation
In this section, the experimental data will be analyzed using the

aforementioned parameters in the previous section. Based on those pa
rameters, an empirical correlation will be developed. Falling film heat
Fig. 4. Schematic of the liquid flowrate reduction along the tube bundle.
10
Fig. 5. Vapor flow affected by exhauster location. (a) Top, (b) Bottom, (c) Top and bottom.
transfer involves various effects, and a schematic showing the resultant (1) Falling film HTC is higher than that of the nucleate boiling HTC in
HTC subject to various influential parameters is illustrated in their the low heat flux region, which occurs when falling film evapo
relationship is depicted in Fig. 8. Falling film heat transfer comprises of ration dominates over the nucleate boiling.
falling film evaporation and nucleate boiling heat transfer, and these (2) Falling film HTC is higher than the nucleate boiling HTC in the
two mechanisms are influenced by the bundle effect. The bundle effect high heat flux regime, which is caused by the mist flow.
can cause local dry-out issue and mist flow heat transfer. Both the dry- (3) Falling film HTC is lower than that of nucleate boiling, which is
out and mist flow phenomena are strongly influenced by the liquid due to the dry-out phenomenon.
and vapor flowrates. In addition to these effects, the enhanced tubes also
significantly influence the nucleate boiling behavior. These deviations provide insights into the complex interaction be
Fig. 9 depicts the measured falling film HTCs in comparison with the tween the falling film evaporation, nucleate boiling, mist flow heat
calculation using the nucleate boiling correlation. The majority of the transfer, and dry-out phenomena. Subsequent sections will provide
data points that comes closer with the dash line, indicating the falling further insights into the phenomena.
film HTC can be described by the nucleate boiling HTC. However, there
are three types of departures of the falling film HTC with nucleate 2.3.1. Dominant heat transfer mechanism
boiling HTC that are clearly seen and can be categorized as: Falling film heat transfer is dominated by nucleate boiling at high
11
Fig. 6. Difference between evaporated vapor flow and imposed vapor flow.
heat flux while falling film evaporation takes control at low heat fluxes. (1) Beyond the critical heat flux, the liquid film could detach from
This is because the nucleate boiling HTC generally increases with the the heated surface [61] and lead to a dry-out. The effect of liquid
rise of heat flux until it attains the critical heat flux (CHF) [59]. In flowrate is negligible where the dry-out is caused by the critical
contrast, the falling film evaporation HTC is relatively independent of heat flux.
the heat flux [1]. Therefore, there exists a transition heat flux at which
the heat transfer mechanism is dominated by nucleate boiling. The
Table 4
transition heat flux of 10 kWm− 2 was observed by Jige et al. [34], as
Nucleate boiling HTC coefficients for enhanced tube [3,4].
depicted in Fig. 10.
Type of tube C n
2.3.2. Dry-out effect Smooth 170.8 0.3758

As shown in Fig. 11, the underlying causes of dry-out is associated Turbo-GII-HP 2.17 × 106 − 0.4316
GEWA-B 3.062 × 104 − 0.04228
with two mechanisms:
High-Flux 3.917 × 107 − 0.6228
Fig. 7. Characteristic diameter of superficial velocity.
12
(2) When the film Reynolds number is below the threshold film
Reynolds number, the uniform liquid film could split into a group
of rivulet liquid film which cannot cover the whole heated surface
[48] and results in a dry-out patch. he dry-out effects caused by
the liquid flowrate (i.e., film Reynolds number) are minimally
impacted by heat flux.
When both the Reynolds number and heat flux reach the dry-out
condition, the dominant factor leading to the dry-out is the detach
ment of liquid film for reaching close to the critical heat flux. The impact
of dry-out on the HTC is presented in Fig. 12, and the boiling number is
adopted to compare the influence of the heat flux and Reynolds number.
The deviation between falling film HTC and nucleate boiling HTC will be
presented as HTC ratio in the following analysis. For the film Reynolds
number of 800 and 1600, a significant decline of HTC is observed when
heat flux exceeds 109 kW m− 2. For heat flux ranging from 40 to 60
kWm− 2, the decline of HTC is observed at the film Reynolds number of
700. Moreover, it can be observed that Reynolds number influences the
HTC ratio even when the dry-out does not occur. This is because the
liquid entrainment rate is proportional to the liquid film flowrate [43],
and liquid entrainment is the cause of the mist flow heat transfer. Mist
flow heat transfer mechanism will be discussed in the subsequent
section.
2.3.3. Mist flow effect

Mist flow can be triggered by two mechanisms as shown in Fig. 3
[40], and is explained as:
(1) As shown in upper part of Fig. 3, the strong vapor flow or liquid
flow forces turns the smooth liquid film into the solitary wavy
film which subsequently ruptures into a group of suspension
droplets. The suspension droplets finally form a mist flow pattern.
(2) In lower part of Fig. 3, the bursting nucleation bubbles would
break the liquid film into a group of suspension droplets. The
suspension droplets form a mist flow pattern.
The tiny droplets in the mist flow results in an increased effective

heat transfer area which enhance the evaporation heat transfer rate
appreciably. The enhanced evaporation heat transfer on the droplets
surface is further augmented with the rise of the superheat of the
droplets. The continuous colliding of droplets on the tube walls and
liquid films, as shown in Fig. 13, potentially promotes the droplets
superheating. The smaller tube pitch could boost the colliding process
and further enhance the heat transfer. Furthermore, appropriate vapor
flow can also enhance the interfacial instability and further form a mist
flow pattern if the vapor flowrate reaches a threshold [40]. However,
some adverse effects could occur if the vapor flow blows away the
droplets.
Fig. 9. Comparison between experiment falling film HTC and nucleate boiling
HTC, grouped by (a) type of tube and (b) working fluid.
To the authors’ knowledge, there are no available experimental data

conducted with the liquid entrainment rate applicable for refrigerant
falling film on the horizontal tube bundle. Thus, the above mechanism
will be incorporated by the heat transfer regression analysis.
2.3.4. Bundle effect

Bundle effect is the impact of different tube bundle configurations on
Fig. 8. Structure of present issue. the dry-out effect and the mist flow effect. The accumulated effect amid
13
tube bundle leads to departure in HTC prediction when compared to a

single tube. This section will discuss the interaction among tube bundle
configuration, dry-out effect, and mist flow effect. The deviation of
liquid flowrate and vapor flowrate along the tube bundles due to the
evaporation will be addressed accordingly.
Fig. 14 and Fig. 15 illustrate the bundle effect, tests are conducted at
similar conditions except the vapor outlet location. Fig. 14 shows the
experimental data of Zhao et al. [56] whose vapor outlet is located at the
bottom of the evaporator as shown in Fig. 5(b). The arrangement forces
the evaporated vapor flow downward. Fig. 15 shows the experimental
data by Roques and Thome [3]. The figure shows only the data of odd
tubes because the data of even tubes was not shown in their research.
Their vapor outlet is located at both the top and bottom of the evapo
rator as shown in Fig. 5(c). In this arrangement, the evaporated vapor
near tubes no.1 and 2 may flow upward and the vapor near tubes no. 7
and 9 may flow downward.
The HTC ratio is compared to the local boiling number which
calculated by the local liquid flow rate, and the local vapor Weber
number by the local vapor superficial velocity. Fig. 14(b) and Fig. 15(b)
indicate that HTC decreases at the downstream vapor flow rather than
downstream liquid flow. From the observation, it is clear that the in
fluence of vapor flow is more pronounced than the liquid flow as long as
dry-out does not occur. The results suggest that using Wevap is essential
in developing the correlation. On the other hand, Fig. 14(a) and Fig. 15
(a) show that the influence of boiling number on the HTC ratio is
comparatively small. In conclusion, evaporated vapor flow dominates Fig. 11. Mechanisms of dry-out.
the bundle effect provided that the dry-out does not present.
[55]. They artificially imposed a downward vapor flow on a vertical row
2.3.5. Imposed vapor flow effect of tubes and studied the influence of imposed vapor flow on the heat
The presence of an additional vapor flow, known as the imposed transfer. As shown in Fig. 16(a), without the imposed vapor (i.e., at
vapor flow may be encountered in addition to the vapor being generated WeImposed = 0), HTC ratio decreases in the downstream tubes. Whereas
through evaporation. This imposed vapor flow may come from the the HTC ratio increased up to a certain value of WeImposed , followed by a
expansion device upon the throttling process. It is important to note that decline when the imposed vapor flow velocity exceeds WeImposed . The
the evaporated and imposed vapor flows are circulated through natural increase in HTC ratio at smaller WeImposed can be attributed to the
convection and forced convection, respectively. Consequently, these
stronger mist flow induced by the imposed vapor flow. Whereas the
two types of vapor flows affect the heat transfer.
decrease in HTC ratio at higher WeImposed is due to the blow away of the
The imposed flow effect is observed in the experiment by Zhao et al.
mist at higher imposed vapor velocities. Furthermore, a significant
reduction in HTC ratio is observed only in the tubes with the boiling
number exceeding the onset condition of dry-out (i.e., nearly 0.2), as
shown in Fig. 16(b).
2.3.6. Enhanced tube effect

Normally the enhanced tubes play the key role in the nucleate boiling
heat transfer. However, as depicted in Fig. 17, all type of tubes except
low-fin tube do not enhance the HTC ratio This is probably due to the
difference in enhancement mechanisms. GEWA, turbo B, or high flux
incorporated the heat transfer mechanism with more nucleate sites to
promote heat transfer. This advantage is roughly the same with flooded
type evaporator. However, the liquid film on the low-fin tube is driven
by surface tension under film coverage. The droplets may be detached
from the fin tip to form entrainment flow. This eventually promotes the
heat transfer under the falling film configuration. Note that this occurs
only in falling film situation, the conventional flooded type evaporator
with low fin tube does not have this is totally different from the flooded
type condition where the additional entrainment flow does not exist.
2.3.7. New correlation equation

Based on the above discussion, a new correlation for the falling film
heat transfer on the horizontal tube bundle is proposed. The following
effects are considered in this correlation:
(a) Dominant heat transfer mechanisms (falling film evaporation and

pool boiling).
Fig. 10. Comparison between experiment total falling film HTC and nucleate (b) Dry-out effect caused by the low liquid flowrate (The dry-out
boiling HTC (Heat flux:2.5–20 kWm− 2, Reff :135–815, experiment: Jige et al., induced by the high heat flux is not considered due to the
2019 [34]).
14
Fig. 12. Comparison between experiment HTC ratio and Boiling number (Heat flux: 5–160 kWm− 2, Reff : 200–2200, experiment: Jin et al. [36]).
Fig. 13. Collision of the droplets.
15
Fig. 14. Comparison between experiment HTC ratio with (a) Boiling number Fig. 15. Comparison between experiment HTC ratio with (a) Boiling number
and (b) vapor Weber number (Heat flux: 20 kWm− 2, Reff : 2–2684, experiment: and (b) vapor Weber number (Heat flux: 20.3 kWm− 2, Reff : 1–2822, experi
Zhao et al. [56]). ment: Roques and Thome [3]).
16
limited data available in the open literature, which is insufficient

for regression.).
(c) Mist flow heat transfer effect.
(d) Enhanced tube effect.
(e) Downward or upward evaporated vapor flow effect.
(f) Downward imposed vapor flow effect.
(g) Local liquid flowrate caused by the evaporation.
To derive the correlation, the first step is the selection of the domi
nant heat transfer as expressed in Eq. (33). The calculation chooses the
higher HTC between falling film evaporation and nucleate boiling, and
times a mist flow enhancement factor.
NuTubeBundle = Max(Nunb , Nucv )Emist (33)
The falling film and boiling Nusselt number are calculated by the Eq.
(17) and (22), respectively. Eq. (34) represents the effective mist flow
enhancement factor which include four terms to represent the liquid
flow effect, dry-out effect, evaporated vapor flow effect, and saturated
temperature effect, respectively. The imposed vapor flow effect in
fluences the dry-out and evaporated vapor flow effects.
⎡ [ ( ) ] ⎤
a0 Ntube tanh ) Liq − a2
[ a1( ln We ]
⎢ +a3 tanh a4 ln Boff − FVap,imposed,1 + a5 ⎥
Emist = exp⎢ [ ( ) ]⎥
⎣ ×tanh a6 ln WeVap,evap − FVap,imposed,2 + a7 ⎦ (34)
+(a8 Treduced + a9 )
By the regression analysis, the mist flow enhancement factor Emist can
be obtained. The coefficients of the mist flow enhancement factor
depend on the dominant heat transfer mechanism, as shown in Eqs. (35)
to (37).
{
Emist,nb , Nu = Nunb
Emist = (35)
Emist,cv , Nu = Nucv
⎡ [ ( ) ] ⎤
− 0.03Ntube[tanh 5.06ln
( )WeLiq − 4.96 ]
⎢ +1.63tanh 0.88ln Boff − FVap,imposed,1 − 0.18 ⎥
Emist,nb = exp⎢ [ ( ) ]⎥
⎣ ×tanh 0.23ln WeVap,evap − FVap,imposed,2 − 0.51 ⎦ (36)
+( − 1.92Treduced − 0.1)
⎡ [ ( ) ] ⎤
0.1Ntube tanh
[ − 2.87ln
( )WeLiq − 0.43 ]
⎢ +0.26tanh 2.16ln Boff − FVap,imposed,1 + 3.64 ⎥
Emist,cv = exp⎢ [ ( )
⎣ ×tanh 8.51ln WeVap,evap − FVap,imposed,2 − 0.5
] ⎥
⎦ (37)
+( − 1.86Treduced + 0.93)
The imposed downward vapor flow effect FVap,imposed,1 and FVap,imposed,2
are expressed by the same formula as shown in Eq. (38). The coefficients
depend on the dominant heat transfer mechanism as shown in Eqs. (39)
to (42).
⎧ ( )
⎨ c0 + c1 ln WeVap,imposed − c2 , WeVap,imposed > 0
⎪
FVap,imposed = 1 + e− [c1 ln(WeVap,imposed )− c2 ] (38)
⎪
⎩
0, WeVap,imposed = 0
⎧ ( )
⎨ − 3.97 + − 0.22ln WeVap,imposed + 3.98, WeVap,imposed > 0
⎪
FVap,imposed,1,nb =
⎪ 1 + e− [− 0.22ln(WeVap,imposed )+3.98 ]
⎩
(39)
⎧ ( )
⎨ 1.19 + − 0.37ln WeVap,imposed + 1.06, WeVap,imposed > 0
⎪ Fig. 16. Experimental HTC ratio subject to (a) imposed vapor flow Weber
FVap,imposed,1,cv = 1 + e− [− 0.37ln(WeVap,imposed )+1.06 ] number and (b) Boiling number (Heat flux: 30 kWm− 2, Reff : 117–566,
⎪
⎩
0, WeVap,imposed = 0 Wevap,imposed : 0–37.2, experiment: Zhao et al. [55]).
(40)
⎧ ( )
⎧ ( ) ⎨ − 0.36 + − 0.21ln WeVap,imposed − 0.15, WeVap,imposed > 0
⎪
⎨ − 2.81 + 0.03ln WeVap,imposed + 3.17, WeVap,imposed > 0
⎪
FVap,imposed,2,cv = 1 + e− [− 0.21ln(WeVap,imposed )− 0.15 ]
FVap,imposed,2,nb = 1 + e− [0.03ln(WeVap,imposed )+3.17 ] ⎪
⎩
⎪
⎩ 0, WeVap,imposed = 0
(42)
(41)
17
Fig. 17. Experimental HTC ratio for various kinds of enhanced tube subject to Fnb. (includes experimental data from [55] [56] [36] [31] [3] [32] [34]).
not take into account the effects listed in Table 5, which yielded a large
3. Results and discussion
deviation against the experimental data. The present correlation in
corporates most of the important heat transfer effects in falling film
3.1. Comparison between different correlations
evaporator, featuring a MAD of 28.3% and an R2 of 0.86.
Fig. 18 compare the predicted HTC by different correlations listed in
Based on the present collected data, an overall comparison across
Table 1 against the experimental data. Further comparisons in terms of
several correlations is shown in Fig. 18 and Fig. 19. The inclusion of
refrigerants and tube types are shown in appendix (Fig. 28 and Fig. 29).
various effects into the correlation provides the least deviation against
the experimental results. However, most of the previous correlations did
Fig. 18. Overall comparison between experiment data and prediction by the proposed correlation.
18
Fig. 19. Overall comparison between the experimental data and prediction by (a) Shah [37] (b) Jin et al. [36] (c) Jige et al. [34] (d) Chien and Chen [32] (e) Roques
and Thome [3,4] (f) Lin et al. [33].
19
Fig. 19. (continued).
20
Table 5
The mechanism being included in the developed correlation with single tube experiment data.
Authors MAD R2 Liquid flow effect Boiling effect Dry-out correction Enhanced tube correction Vapor flow effect Bundle effect
Roques and Thome [3,4] 116.1% − 0.29 No Yes Yes Yes No Yes
Chien and Chen [32] 349.9% − 27.2 Yes Yes No No No No
Jige et al. [34] 52.7% − 0.49 Yes Yes No No No No
Jin et al. [36] 57.9% − 0.61 Yes Yes Yes No No No
Shah [37] 52.8% − 0.46 Yes Yes No No No Yes
Lin et al. [33] 40.2% − 0.01 Yes Yes Yes No No No
This study 28.3% 0.86 Yes Yes Yes Yes Yes Yes
Fig. 20. Comparison between predicted and experimental falling film HTC and Fig. 22. Comparison between predicted and experimental HTC ratio vs. boiling
nucleate boiling HTC (Heat flux:2.5–20 kWm− 2, Reff :135–815, experimental number (Heat flux: 5–160 kWm− 2, Reff : 200–2200, experiment: Jin et al. [36]).
data: Jige et al., [34]).
3.2. Comparison between the experiment data and the proposed
correlation
In the Section 3.2, the effects discussed in Fig. 10 to Fig. 17 will be

used to examine the predictive ability. The comparison between
experimental data and the prediction will be shown in Fig. 20 to Fig. 25.
3.2.1. Dominant heat transfer mechanism

Fig. 20 and Fig. 21 show that the predicted transition point from
falling film evaporation to the pool boiling, which is almost same to the
experiment.
3.2.2. Dry-out effect

Fig. 22 shows that the correlation is able to predict the onset dry-out
caused by the low liquid flowrate. However, the onset boiling number of
dry-out caused by the high heat flux is underpredicted.
3.2.3. Bundle effect

For the bundle effect, Fig. 23 and Fig. 24 shows that both experi
mental and predicted HTC decrease in the downstream of vapor flow
stream. However, deviation of the HTC ratio along the tube bundle is
under-predicted as shown in Fig. 23 and is well-predicted in Fig. 24. This
difference implies that probably there is a parameter that affects the
bundle effect has not been included in the present study.
Fig. 21. Comparison between predicted and experimental HTC ratio vs. Rey
nolds number (Heat flux: 2.5–20 kWm− 2, Reff : 135–815, experimental data:
Jige et al., [34]).
3.2.4. Imposed vapor flow effect
Fig. 25 shows the correlation gives similar results as the experiments:
With the rise of the imposed vapor flowrate, both experimental and
21
Fig. 23. Comparison between predicted and experimental HTC ratio subject to Fig. 24. Comparison between predicted and experimental HTC ratio subject to
(a) Boiling number (b) vapor Weber number (Heat flux: 20 kWm− 2, Reff : (a) Boiling number (b) vapor Weber number (Heat flux: 20.3 kWm− 2, Reff :
2–2684, experiment: Zhao et al. [56]). 1–2822, experiment: Roques and Thome [3]).
predicted HTC shows an appreciable increase initially, followed by a 3.2.5. Enhanced tube effect
moderate decline, and finally the HTC drops more sharply on the bottom Compared to Fig. 17, the prediction also shows only the HTC of
tubes. smooth and low-fin tubes is significantly higher than other enhanced
tubes, as presented in Fig. 26. This phenomenon is explained earlier due
to the nucleation sites are less effective upon falling film situation. For
22
Fig. 26. Predicted HTC ratio compared to the boiling enhancement ratio (in
cludes experiment data from [55] [56] [36] [31] [3] [32] [34]).
transfer augmentation when compared to nucleate boiling only as

depicted in Fig. 26. This argument can be schematically illustrated in
Fig. 27(a) with the supporting evidence from Benther et al. [62]. For low
fin tube, as explained earlier, the surface tension alongside the fin tip can
assist the film flow to drain into droplets for effective entrainment flow
as shown in Fig. 27(a) and Fig. 27(b), the basic mechanism has been
elaborated by Webb et al. [63] for condensate removal. Conversely, the
enhanced tube features re-entrant liquid tunnel that help to stabilize the
liquid film as seen in Fig. 27(a). Although the nucleate boiling mecha
nisms is still preserved as the main heat transfer mechanisms for
enhanced tubes such as GEWA-B, it is found that the enhancement for
falling film evaporation relative to nucleate boiling for flooded evapo
rator is comparatively small. In fact, as depicted in Fig. 26, the HTC ratio
in many cases for enhanced tubes are less unity, meaning a decline in
overall HTC even with the help of evaporation mechanism. To explain
this difference, one can see a typical schematic of nucleate boiling for
flooded evaporator where the enhanced tubes are well submerged below
the liquid level as shown in Fig. 27(c). In this regard, a static pressure
difference exists between the enhanced tube and the liquid level. This
eventually drives the detached bubbles toward the liquid level more
smoothly when compared to the enhanced tubes that were operated
under falling film configuration. In summary, the nucleate boiling
flooded and yields a higher HTC than the falling film configuration if
evaporation is not in control.
4. Conclusions
The present study reviews the correlations for falling film evaporator
with a horizontal tube bundle configuration. Detailed literature review
of the experimental studies and available empirical correlations are
made available and comprehensive discussion on the heat transfer
mechanism is also conducted. Based on the discussed mechanisms, the
authors propose a new rationally based correlation based on the existing
Fig. 25. Predicted and experimental HTC ratio compared to the (a) imposed data and elaborate comparison of different correlations and the pro
vapor flow Weber number and (b) Boiling number (Heat flux: 30 kWm− 2, Reff : posed correlation are also addressed. The collected data includes 4114
117–566, Wevap,imposed : 0–37.2, experiment: Zhao et al. [55]). data points from 8 sources, 6 refrigerants (R600a, R290, R245fa, R134a,
R1234ze(E), R123), 5 types of the tubes (Smooth, Turbo-GII-HP, GEWA-
B, Low-fin, High-Flux), liquid Weber number from 2.2 × 10− 6 to 0.7,
smooth tube, when compared to enhanced tube and integral tube, the imposed vapor Weber number from 0 to37.2, heat flux from 2.5 to 151.5
falling film onto the tube surface may incur more pronounced splashing kW/m2, and film Reynolds number from 1 to 3159.8. Based on the
phenomenon and produced more entrainment droplets to facilitate more foregoing discussions, the following conclusions are made.
evaporation. In this regard, falling film imposes significantly heat
23
Fig. 27. Heat transfer mechanism (a) Liquid film impacts on the tubes (b) Liquid film detached from the tubes (c) Static pressure effect of the nucleate boiling.
24
(1) A new correlation is proposed to predict the falling film evapo Authorship statement
ration. The correlation yield a MAD of 28.3% and R2 of 0.86. It
covers all the essential mechanisms, including the transition Hao-Yu Lin: Conceptualization, conducting numerical works, and
point from falling film evaporation to the nucleate boiling, local writing the original drafts.
evaporation effect, dry-out effect, mist flow effect, imposed flow M Muneeshwaran: Conceptualization, supervision, reviewing and
effect, and enhanced tube effect. revising.
(2) The onset of dry-out is referred to two mechanisms: non-uniform Cheng-Min Yang: Conceptualization, supervision, reviewing and
liquid film caused by low flowrate and detached liquid film revising.
caused by the high heat flux. If the dry-out is not reached, the Kashif Nawaz: Conceptualization, supervision, reviewing and
evaporated vapor flow effect dominates tube bundle effect and revising.
the imposed vapor flow enhances the heat transfer. Chi-Chuan Wang: Conceptualization, supervision, reviewing and
(3) The major heat transfer enhancement in falling film evaporation revising.
is attributed to the mist flow heat transfer. The liquid entrainment
rate is crucial, and it can be estimated by superficial vapor Declaration of Competing Interest
velocity.
(4) The smooth, enhanced, and low fin tube reveals quite distinct None.
heat transfer performance in response to falling film evaporation.
Upon falling film, more pronounced splashing phenomenon may Data availability
occur for smooth tube and additional surface tension force may
assist the formation liquid entrainment for the low fin tube, Data will be made available on request.
thereby improving the heat transfer performance significantly.
For conventional enhanced tube, nucleate boiling is still the main Acknowledgements
heat transfer mechanism for its structure offers the effective
liquid reentrance via cavity. The authors gratefully acknowledge the financial support from Na
tional Science and Technology Council Taiwan under the contract NSTC
112-2221-E-A49-026.
Appendix A. Appendix
Fig. 28. Overall comparison between experiment and prediction by present equation.
25
26
Fig. 29. Overall comparison between experiment and prediction by (a) Shah [37] (b) Jin et al. [36] (c) Jige et al. [34] (d) Chien and Chen [32] (e) Roques and
Thome [3,4] (f) Lin et al. [33].
Fig. 29. (continued).

References [12] C. Mateu-Royo, C. Arpagaus, A. Mota-Babiloni, J. Navarro-Esbrí, S.S. Bertsch,
Advanced high temperature heat pump configurations using low GWP refrigerants
for industrial waste heat recovery: a comprehensive study, Energy Convers. Manag.
[1] A.A. Alhusseini, K. Tuzla, J.C. Chen, Falling film evaporation of single component
229 (2021), 113752, https://doi.org/10.1016/j.enconman.2020.113752.
liquids, Int. J. Heat Mass Transf. 41 (12) (1998) 1623–1632, https://doi.org/
[13] M.O. McLinden, J.S. Brown, R. Brignoli, A.F. Kazakov, P.A. Domanski, Limited
10.1016/S0017-9310(97)00308-6.
options for low-global-warming-potential refrigerants, Nat. Commun. 8 (1) (2017)
[2] K.R. Chun, R.A. Seban, Heat transfer to evaporating liquid films, J. Heat Transf. 93
14476, https://doi.org/10.1038/ncomms14476.
(4) (1971) 391–396, https://doi.org/10.1115/1.3449836.
[14] G. Censi, L. Giancotti, A. Padovan, Experimental investigation of a new low-
[3] J.F. Roques, J.R. Thome, Falling films on arrays of horizontal tubes with R-134a,
approach evaporator with reduced refrigerant charge, in: The International
Part I: boiling heat transfer results for four types of tubes, Heat Transfer Eng. 28 (5)
Refrigeration and Air Conditioning Conference, 2018.
(2007) 398–414, https://doi.org/10.1080/01457630601163629.
[15] P. Zhang, R. Chen, G.H. Su, W. Tian, S. Qiu, Experimental investigation and
[4] J.F. Roques, J.R. Thome, Falling films on arrays of horizontal tubes with R-134a,
modeling of falling film heat transfer on partial dry-out condition, Appl. Therm.
Part II: flow visualization, onset of dryout, and heat transfer predictions, Heat
Eng. 229 (2023), 120550, https://doi.org/10.1016/j.
Transf. Eng. 28 (5) (2007) 415–434, https://doi.org/10.1080/
applthermaleng.2023.120550.
01457630601163736.
[16] Y. Guo, M. Bao, L. Gong, S. Shen, K. Zhang, Numerical investigation of the falling
[5] M. Christians, J.R. Thome, Falling film evaporation on enhanced tubes, Part 2:
film thickness and heat transfer characteristics over horizontal round tube, Int. J.
prediction methods and visualization, Int. J. Refrig. 35 (2) (2012) 313–324,
Multiphase Flow 149 (2022), 103977, https://doi.org/10.1016/j.
https://doi.org/10.1016/j.ijrefrig.2011.10.019.
ijmultiphaseflow.2022.103977.
[6] M. Christians, J.R. Thome, Falling film evaporation on enhanced tubes, Part 1:
[17] W. Nusselt, Die oberflachenkondensation des wasserdamphes, VDI-Zs 60 (1916)
experimental results for pool boiling, onset-of-dryout and falling film evaporation,
541.
Int. J. Refrig. 35 (2) (2012) 300–312, https://doi.org/10.1016/j.
[18] W.M. Rohsenow, Heat transfer and temperature distribution in laminar-film
ijrefrig.2011.10.020.
condensation, Trans. Am. Soc. Mech. Eng. 78 (8) (1956) 1645–1647.
[7] The Future of Heat Pumps, International Energy Agency, 2022. Paris.
[19] E.M. Sparrow, J.L. Gregg, A boundary-layer treatment of laminar-film
[8] J. Jiang, B. Hu, R.Z. Wang, N. Deng, F. Cao, C.-C. Wang, A review and perspective
condensation, J. Heat Transf. 81 (1) (1959) 13–18.
on industry high-temperature heat pumps, Renew. Sust. Energ. Rev. 161 (2022),
[20] V. Dhir, J. Lienhard, Laminar film condensation on plane and axisymmetric bodies
112106, https://doi.org/10.1016/j.rser.2022.112106.
in nonuniform gravity, 1971.
[9] J. Jiang, B. Hu, R.Z. Wang, H. Liu, Z. Zhang, H. Li, Theoretical performance
[21] J. Lorenz, D. Yung, A note on combined boiling and evaporation of liquid films on
assessment of low-GWP refrigerant R1233zd(E) applied in high temperature heat
horizontal tubes, 1979.
pump system, Int. J. Refrig. 131 (2021) 897–908, https://doi.org/10.1016/j.
[22] M. Cerza, V. Sernas, Nucleate boiling heat transfer in developing laminar falling
ijrefrig.2021.03.026.
water films, in: Proc. ASME-JSME Thermal Engineering Joint Conference
[10] D. Wu, B. Hu, R.Z. Wang, Vapor compression heat pumps with pure low-GWP
Proceedings, 2023.
refrigerants, Renew. Sust. Energ. Rev. 138 (2021), 110571, https://doi.org/
[23] W. Parken, L.S. Fletcher, Heat transfer in thin liquid films flowing over horizontal
10.1016/j.rser.2020.110571.
tubes, in: Proc. International Heat Transfer Conference Digital Library, Begel House
[11] N. Abas, A.R. Kalair, N. Khan, A. Haider, Z. Saleem, M.S. Saleem, Natural and
Inc., 2023.
synthetic refrigerants, global warming: a review, Renew. Sust. Energ. Rev. 90
(2018) 557–569, https://doi.org/10.1016/j.rser.2018.03.099.
27
[24] S. Sabin, H. Poppendiek, Film evaporation of ammonia over horizontal round [45] G. Xuan, L. Ri, Spray impingement cooling: the state of the art, in: S.M.S. Murshed
tubes, in: Proc. Proceedings of the Fifth Ocean Thermal Energy Conversion (Ed.), Advanced Cooling Technologies and Applications, IntechOpen, Rijeka, 2018.
Conference, Miami, FL, 2023. Ch. 3.
[25] M. Nakazatomi, A. Bergles, Film evaporation model on horizontal tubes, Res. [46] C. Inoue, I. Maeda, On the droplet entrainment from gas-sheared liquid film, Phys.
Reprint Ube Techn. College 27 (1981). Fluids 33 (1) (2021), 011705, https://doi.org/10.1063/5.0038399.
[26] M.-C. Chyu, A. Bergles, An analytical and experimental study of falling-film [47] C.-Y. Zhao, L.-W. Liang, D. Qi, W.-T. Ji, W.-Q. Tao, The effect of gas streams on the
evaporation on a horizontal tube, 1987. hydrodynamics, heat and mass transfer in falling film evaporation, absorption,
[27] K.R. Chun, R. Seban, Performance prediction of falling-film evaporators, 1972. cooling and dehumidification: a comprehensive review, Build. Environ. 219
[28] K. Chun, R. Seban, Heat transfer to evaporating liquid films, 1971. (2022), 109183, https://doi.org/10.1016/j.buildenv.2022.109183.
[29] L.-H. Chien, C.-H. Cheng, A predictive model of falling film evaporation with [48] D.E. Hartley, W. Murgatroyd, Criteria for the break-up of thin liquid layers flowing
bubble nucleation on horizontal tubes, HVAC&R Res. 12 (1) (2006) 69–87, https:// isothermally over solid surfaces, Int. J. Heat Mass Transf. 7 (9) (1964) 1003–1015,
doi.org/10.1080/10789669.2006.10391168. https://doi.org/10.1016/0017-9310(64)90042-0.
[30] J.C. Chen, Correlation for boiling heat transfer to saturated fluids in convective [49] N. Zuber, F.W. Staub, Stability of dry patches forming in liquid films flowing over
flow, Industr. Eng. Chem. Proc. Design Developm. 5 (3) (1966) 322–329, https:// heated surfaces, Int. J. Heat Mass Transf. 9 (9) (1966) 897–905, https://doi.org/
doi.org/10.1021/i260019a023. 10.1016/0017-9310(66)90064-0.
[31] L.-H. Chien, Y.-L. Tsai, An experimental study of pool boiling and falling film [50] Y. He, K. Xiong, E. Xia, Growth of dry patches in an evaporating film flowing
vaporization on horizontal tubes in R-245fa, Appl. Therm. Eng. 31 (17) (2011) around a horizontal circular tube, Int. Commun. Heat Mass Transf. 125 (2021),
4044–4054, https://doi.org/10.1016/j.applthermaleng.2011.08.007. 105291, https://doi.org/10.1016/j.icheatmasstransfer.2021.105291.
[32] L.H. Chien, R.H. Chen, An experimental study of falling film evaporation on [51] M. Habert, J.R. Thome, Falling-film evaporation on tube bundle with plain and
horizontal tubes using R-134a, J. Mech. 28 (2) (2012) 319–327, https://doi.org/ enhanced tubes—Part II: new prediction methods, Exp. Heat Transf. 23 (4) (2010)
10.1017/jmech.2012.35. 281–297, https://doi.org/10.1080/08916152.2010.502046.
[33] H.-Y. Lin, M. Muneeshwaran, C.-M. Yang, K. Nawaz, C.-C. Wang, Universal [52] M. Habert, J.R. Thome, Falling-film evaporation on tube bundle with plain and
correlation for falling film evaporation on a horizontal plain tube, Int. J. Refrig. enhanced tubes—Part I: experimental results, Exp. Heat Transf. 23 (4) (2010)
(2022), https://doi.org/10.1016/j.ijrefrig.2022.11.006. 259–280, https://doi.org/10.1080/08916152.2010.502045.
[34] D. Jige, H. Miyata, N. Inoue, Falling film evaporation of R1234ze(E) and R245fa on [53] R.P. Baines, M.A. El Masri, W.M. Rohsenow, Critical heat flux in flowing liquid
a horizontal smooth tube, Exp. Thermal Fluid Sci. 105 (2019) 58–66, https://doi. films, Int. J. Heat Mass Transf. 27 (9) (1984) 1623–1629, https://doi.org/10.1016/
org/10.1016/j.expthermflusci.2019.03.012. 0017-9310(84)90274-6.
[35] C.-Y. Zhao, W.-T. Ji, P.-H. Jin, W.-Q. Tao, Heat transfer correlation of the falling [54] R. Peck, C. Olsen, J.L. Devore, Introduction to Statistics and Data Analysis, Cengage
film evaporation on a single horizontal smooth tube, Appl. Therm. Eng. 103 (2016) Learning, 2011.
177–186, https://doi.org/10.1016/j.applthermaleng.2016.02.090. [55] C.-Y. Zhao, W.-T. Ji, P.-H. Jin, S. Yoshioka, W.-Q. Tao, Effect of downward vapor
[36] P.-H. Jin, Z. Zhang, I. Mostafa, C.-Y. Zhao, W.-T. Ji, W.-Q. Tao, Heat transfer stream on falling film evaporation of R134a in a tube bundle, Int. J. Refrig. 89
correlations of refrigerant falling film evaporation on a single horizontal smooth (2018) 112–121, https://doi.org/10.1016/j.ijrefrig.2018.02.027.
tube, Int. J. Heat Mass Transf. 133 (2019) 96–106, https://doi.org/10.1016/j. [56] C.-Y. Zhao, W.-T. Ji, P.-H. Jin, Y.-J. Zhong, W.-Q. Tao, Experimental study of the
ijheatmasstransfer.2018.12.053. local and average falling film evaporation coefficients in a horizontal enhanced
[37] M.M. Shah, A general correlation for heat transfer during evaporation of falling tube bundle using R134a, Appl. Therm. Eng. 129 (2018) 502–511, https://doi.org/
films on single horizontal plain tubes, Int. J. Refrig. 130 (2021) 424–433, https:// 10.1016/j.applthermaleng.2017.09.135.
doi.org/10.1016/j.ijrefrig.2021.04.025. [57] I.H. Bell, J. Wronski, S. Quoilin, V. Lemort, Pure and pseudo-pure fluid
[38] M.M. Shah, Prediction of heat transfer in evaporation of saturated falling films on thermophysical property evaluation and the open-source thermophysical property
bundles of horizontal tubes, Int. J. Refrig. 131 (2021) 416–425, https://doi.org/ library CoolProp, Ind. Eng. Chem. Res. 53 (6) (2014) 2498–2508, https://doi.org/
10.1016/j.ijrefrig.2021.07.004. 10.1021/ie4033999.
[39] M.M. Shah, Prediction of dryout in evaporation of falling films on horizontal plain [58] Y. Fujita, M. Tsutsui, Experimental and analytical study of evaporation heat
tubes, Chem. Eng. Res. Des. 179 (2022) 527–534, https://doi.org/10.1016/j. transfer in falling films on horizontal tubes, in: Proc. International Heat Transfer
cherd.2022.01.039. Conference Digital Library, Begel House Inc., 1994.
[40] I. Zadrazil, C.N. Markides, O.K. Matar, L. Naraigh, G.F. Hewitt, Characterisation of [59] D. Jung, Y. Kim, Y. Ko, K. Song, Nucleate boiling heat transfer coefficients of pure
downwards co-current gas-liquid annular flows, in: Proc. THMT-12. Proceedings of halogenated refrigerants, Int. J. Refrig. 26 (2) (2003) 240–248, https://doi.org/
the Seventh International Symposium On Turbulence Heat and Mass Transfer, 10.1016/S0140-7007(02)00040-3.
Begel House Inc., 2012. [60] W.M. Rohsenow, J.P. Hartnett, Y.I. Cho, Handbook of heat Transfer, Mcgraw-Hill
[41] T. Ueda, M. Inoue, S. Nagatome, Critical heat flux and droplet entrainment rate in New York, 1998.
boiling of falling liquid films, Int. J. Heat Mass Transf. 24 (7) (1981) 1257–1266, [61] I.A. Mudawwar, T.A. Incropera, F.P. Incropera, Boiling heat transfer and critical
https://doi.org/10.1016/0017-9310(81)90175-7. heat flux in liquid films falling on vertically-mounted heat sources, Int. J. Heat
[42] T. Fujita, T. Ueda, Heat transfer to falling liquid films and film breakdown—I: Mass Transf. 30 (10) (1987) 2083–2095, https://doi.org/10.1016/0017-9310(87)
subcooled liquid films, Int. J. Heat Mass Transf. 21 (2) (1978) 97–108, https://doi. 90088-3.
org/10.1016/0017-9310(78)90212-0. [62] J.D. Benther, J.D. Pelaez-Restrepo, C. Stanley, G. Rosengarten, Heat transfer during
[43] T. Fujita, T. Ueda, Heat transfer to falling liquid films and film breakdown—II: multiple droplet impingement and spray cooling: review and prospects for
saturated liquid films with nucleate boiling, Int. J. Heat Mass Transf. 21 (2) (1978) enhanced surfaces, Int. J. Heat Mass Transf. 178 (2021), 121587, https://doi.org/
109–118, https://doi.org/10.1016/0017-9310(78)90213-2. 10.1016/j.ijheatmasstransfer.2021.121587.
[44] N.I. Pecherkin, A.N. Pavlenko, O.A. Volodin, Heat transfer and critical heat flux at [63] L.-H. Chien, R.L. Webb, A nucleate boiling model for structured enhanced surfaces,
evaporation and boiling in refrigerant mixture films falling down the tube with Int. J. Heat Mass Transf. 41 (14) (1998) 2183–2195, https://doi.org/10.1016/
structured surfaces, Int. J. Heat Mass Transf. 90 (2015) 149–158, https://doi.org/ S0017-9310(97)00302-5.
10.1016/j.ijheatmasstransfer.2015.06.050.
28

1 s2.0 S0735193323005547 Main

Uploaded by

Copyright:

Available Formats

1 s2.0 S0735193323005547 Main

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1 s2.0 S0735193323005547 Main

Uploaded by

Copyright:

Available Formats

International Communications in Heat and Mass Transfer 150 (2024) 107165

Contents lists available at ScienceDirect

International Communications in Heat and Mass Transfer

On falling film evaporator – A review of mechanisms and critical

contain less refrigerant inventory. Hence, it is very attractive for heat

Available online 6 December 2023

Nomenclature Nunb,smooth Nucleated boiling Nusselt number on a smooth tube

HFO Hydro Fluoro Olefins SSTO Sum of squared total

1.3.1. Falling film evaporation and nucleate boiling

Fig. 2. Global warming potential requirement under different protocol.

Alhusseini et al. Water, Propylene 34- 7.8-19.5 Single; Smooth ( )1/5

Reff,threshold = 53600Boff 0.0045 Pr− 0.52 Weff 0.5

2.2.2. Data processing

In order to have a reasonable comparison between the falling

the experiment conducted by Roques and Thome [3,4] is higher ⎧

ηLow− fin− tube = tanh(Φ)

ALow− fin− tube δfin

In this section, the experimental data will be analyzed using the

2.3.2. Dry-out effect Smooth 170.8 0.3758

Fig. 7. Characteristic diameter of superficial velocity.

2.3.3. Mist flow effect

The tiny droplets in the mist flow results in an increased effective

To the authors’ knowledge, there are no available experimental data

2.3.4. Bundle effect

tube bundle leads to departure in HTC prediction when compared to a

2.3.6. Enhanced tube effect

2.3.7. New correlation equation

(a) Dominant heat transfer mechanisms (falling film evaporation and

Fig. 13. Collision of the droplets.

limited data available in the open literature, which is insufficient

Fig. 19. (continued).

In the Section 3.2, the effects discussed in Fig. 10 to Fig. 17 will be

3.2.1. Dominant heat transfer mechanism

3.2.2. Dry-out effect

3.2.3. Bundle effect

transfer augmentation when compared to nucleate boiling only as

Fig. 29. (continued).

You might also like