Applied Thermal Engineering 219 (2023) 119431

Contents lists available at ScienceDirect

Applied Thermal Engineering

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

Research Paper

Flow and heat transfer model for turbulent-laminar/turbulent gas-liquid

annular flows
Shaoting Jia b, Chuanshuai Dong a, *
a
Key Laboratory of Enhanced Heat Transfer and Energy Conservation of Education Ministry, School of Chemistry and Chemical Engineering, South China University of
Technology, Guangzhou 510640, China
b
China Oil and Gas Pipeline Network Corporation, Guangzhou 510620, China

A R T I C L E I N F O A B S T R A C T

Keywords: Two-phase gas–liquid annular flows are characterized by the liquid flowing in the annular-shaped channels as
Gas-liquid two-phase flow liquid film and gas flowing in the center as gas core. Adequate understanding of the flow and heat transfer
Heat transfer coefficient mechanism is important. This paper aims at developing a model for flow and heat transfer in two-phase annular
Annular flow
flows. First, the physical process of flow and heat transfer for two-phase annular flow is mathematically
Void fraction
formulated and a new heat transfer model is developed based on two-fluid concept. Then, the new model is well
validated using the experimental void fraction, liquid film thickness, and heat transfer coefficient collected from
various sources. Third, the effect of the orientations and flow parameters on the heat transfer of two-phase
annular flows are comprehensively investigated. Finally, the dependence of the two-phase heat transfer multi­
pliers on the void fractions and pressure multipliers is identified quantitatively and a simple heat transfer cor­
relation is developed based on Chilton & Colburn analogy.

homogenous flow model is that gas and liquid are considered as ho­
1. Introduction mogenous singe pseudo-fluids [6,35]. Thus, the homogenous flow is not
suitable in two-phase annular flows because of its distinct interface and
Two-phase gas liquid flows are widely found in many natural and huge difference of velocities between liquid and gas phases. The drift-
engineering fields, such as chemical columns and petroleum pipelines. flux model is characterized by the consideration of relative movement
Due to the deformable interface between liquid and gas phases, between gas and liquid using a constitutive relation [15]. In two-fluid
depending on physical properties, gravity, flow velocities, and orienta­ model, gas and liquid are considered as continuous fully-separated
tions of the flow channel, etc., two-phase flow is very complicated flows [46,8]. To consider the deformable interface and complicated
[51,55,9]. The phenomenon occuring in two-phase flows are too com­ interaction between gas and liquid, a flow regime-based model is pro­
plex to be amenable to a rigorous mathematical solution [44,23,33]. posed for the specific flow regime.
Amongst the main flow regimes, two-phase annular flows are widely The flow behavior of annular liquid film is of key importance in two-
encountered in many fields, such as evaporators, condensers, and phase annular flows [31,24]. He et al. [22,21] investigated the local
boiling-water reactors [38]. Two-phase annular flows are characterized liquid film behavior of annular flows on rod-bundle orientations and
by the continuous annular liquid film flowing on the channel wall and proposed a mechanistic model for calculating the local film thickness
gas travelling through in the gas core. The annular liquid film could be and spatial distribution of gas flow. Fan et al. [16] performed both
entrained into the gas core and disintegrated into small liquid droplets at experimental and numerical investigation on the formation and devel­
high gas velocities [54]. Meanwhile, the liquid droplet in the gas core opment of disturbance waves in annular flows. Ju et al. [25] and Wang
might deposit on the liquid film [28]. et al. [48] investigated the wave characteristics of vertical annular flows
Considering the importance of two-phase flow, tremendous efforts and developed models for calculating the wave velocity, and wave
are made in understanding its flow mechanism [7,1,27,22,21]. Several height.
models of two-phase flow are proposed. Homogeneous flow models, Owing to the flow complexity of annular flows, the heat transfer
drift-flux models, two-fluid models, and flow regime-based models are process is much more complicated. Keniar and Gerimella (2021) pro­
often used in two-phase flow simulation. The central assumption in posed a mechanistic model for calculating the condensation heat

* Corresponding author.
E-mail address: [email protected] (C. Dong).

https://doi.org/10.1016/j.applthermaleng.2022.119431
Received 18 May 2022; Received in revised form 3 September 2022; Accepted 29 September 2022
Available online 5 October 2022
1359-4311/© 2022 Elsevier Ltd. All rights reserved.
S. Jia and C. Dong Applied Thermal Engineering 219 (2023) 119431

Nomenclature θ Inclination angle [◦ ]

μ Dynamic viscosity [Pa⋅s]
A Area [m2] ρ Density [kg/m3]
Cp Specific heat capacity [kJ/(kg⋅K)] τ Shear stress [N/m]
d Diameter [m] Φ2f Two-phase pressure drop multiplier [-]
D Hydraulic diameter [m] Φ2h Two-phase heat transfer multiplier [-]
f Friction factor [-]
g Gravitational acceleration [m/s2] Subscripts
h Hear transfer coefficient [W/(m2 ⋅K)] C Gas core [-]
ks Surface roughness [m] D Liquid droplet [-]
L Length [m] F Liquid film [-]
Lt Thermal entrance length [m] G Gas [-]
m Mass flow rate [kg/s] i Inner [-]
Nu Nusselt number [-] I Interface [-]
P Pressure [Pa] L Liquid [-]
Pr Prandtl number [-] m Gas-liquid mixture [-]
q Heat flux [W/m2] MCD Mixed convection [-]
Re Reynolds number [-] o Outer [-]
S Perimeter [m] s Surrounding [-]
T Temperature [℃] sim Simulated [-]
V Velocity [m/s] SL Superficial liquid [-]
SG Superficial gas [-]
Greek Symbols W Wall [-]
α Void fraction [-] 1Φ Single phase [-]
δ Liquid film thickness [m] 2Φ Two phase [-]
ε Entrainment rate [-]

transfer coefficients of intermittent and annular flow regimes. The unclear mechanism of heat transfer enhancement by the injection of gas
model was developed using the laminar assumption and applicable to into liquid, this study aims at developing a model of flow and heat
arbitrarily shaped micro- or mini-channels. Thome et al. [26,43] transfer for two-phase annular flows to investigate the mechanism of the
developed a flow-regime based models for two-phase flows. Zhang et al. heat transfer enhancement comprehensively. The physical process of
[52] proposed a unified heat transfer model for various flow regimes but flow and heat transfer is formulated mathematically.
did not investigated the influence of flow parameters and orientations
on the heat transfer process. 2. Hydrodynamic model
Recently, we have developed a series of mechanistic models of flow
and heat transfer for gas–liquid two-phase slug flow and stratified flow As shown in Fig. 1, the annular liquid film flows continuously on the
[12,13,14]. Considering the periodic flow behaviors of two-phase slug annulus-shaped channel wall, while the gas flows in the gas core, car­
flows in horizontal and vertical pipes, its mechanistic models were rying varying amounts of liquid droplets. The two regions are separated
developed based on the slug unit cell concept. For two-phase stratified by a distinct interface. The annular liquid film becomes wavy and part of
flows, the mechanistic model was developed based on the two-fluid the liquid might be entrained in the gas core at high gas velocity.
model because of the distinct interface between gas and liquid. Both Because of the gravity effect, the annular liquid film in the top is usually
the gas and liquid contacts with the pipe wall in two-phase stratified thinner than that in the bottom in horizontal and inclined pipes. Such
flow. However, as the velocity of the gas increases, the liquid is blown by difference becomes small and thus uniform film thickness around the
the gas and finally forms a continuous annular film on the pipe wall. pipe periphery is adopted in this model [44]. The formulation of the
Thus, there exists huge difference in the flow and heat transfer between hydrodynamic model is developed based on continuity and momentum
two-phase annular flow and stratified flow. equations of both liquid and gas.
Considering the importance of two-phase annular flow and its

Fig. 1. Schematic of the two-phase annular flows.

2
S. Jia and C. Dong Applied Thermal Engineering 219 (2023) 119431

2.1. Continuity equations the pressure drop. SF and SI are the wet perimeters. τF and τI are the
shear stresses. θ is the inclination angle from the horizontal orientations,
The liquid phase flows in the annular liquid film region as continuous where θ = 0o is assumed for flow in horizontal pipes, θ > 0o is
liquid film and in the gas core region as entrained liquid droplet. Thus, assumed for flow in upward inclined pipes, and θ < 0o is assumed for
the continuity equation of the liquid phase is given by: flow in downward inclined pipes.
Varying amount of liquid droplets are entrained in the gas core. To
m = ρL VSL A = mF + mD = ρL VF AF + ρL VD AD (1)
simply the model, the gas core is assumed as a homogeneous single
wherem,mF , and mD are the mass flow rates of the liquid in the whole pseudo-fluid. Therefore, the momentum equation of the gas core is given
flow channel, annular liquid film region, and gas core, respectively. ρL is by:
the density of the liquid. VSL is the superficial liquid velocity.A,AF , and dP
− AC − τI SI − ρC AC gsinθ = 0 (12)
AD are the cross-sectional areas of the whole channel, the annular liquid dz
film region, and the entrained liquid droplet, respectively. VF and VD are
the flow velocities of the annular liquid film and liquid droplet. where ρC is the mixture density, estimated by:
AF is given by: ρC = αC ρG + (1 − αC )ρL (13)
π[ ]
AF = d2 − (d − 2δ)2 (2)
4 where αC is the void fraction of the gas core. αC = 1 means that no
liquid droplets are contained in the gas core. Owing to the equal pressure
where δ is the liquid film thickness. Due to the rather small thickness of drops in the gas core and annular liquid film, Equation (11) and (12) are
the liquid film, AF can also be approximately estimated by: combined by eliminating dP dz as:
AF = πdδ (3) ( )
τ F SF 1 1
− + τI SI + − (ρL − ρG )gsinθ = 0 (14)
AF AC AF
AD is given by:
AD = AC (1 − αC ) (4) where τF andτI , are estimated by:

where αC refers to the void fraction of the gas core and AC refers to the ρF VF2
τF = fF (15)
cross-sectional area. 2

1 ρC (VG − VF )|VG − VF |
AC = π(d − 2δ)2 (5) τI = fI (16)
4 2

The void fraction in the gas core,αC , is defined by: where fF is the wall fraction factor of the liquid and fI is the interfacial
fraction factor. fF is estimated as follows [36]:
AD εmL /ρL /VD εVSL /VD
αC = 1 − = 1− = 1− (6) For laminar flow (Re < 2300),
AC A − πdδ 1 − d4 δ
64
f = (17)
where VD is the velocity of the liquid droplet. Due to the much smaller ReF
size of the liquid droplet, the velocity of the liquid droplet is assumed to
be equal to the actual gas velocity. ε is a measure of the fraction of liquid For turbulent flow (Re ≥ 2300),
that flows in the gas core, calculated by Wallis correlation [47]. 1
(
ks 2.51
)
√̅̅̅ = - 2log + √̅̅̅ (18)
ε = 1 − e− 0.125(ϕ− 1.5)
(7) f 3.70d ReF f

where ϕ is defined as: where ks represents the surface roughness of the pipe. ReF represents the
Reynolds number at real liquid velocity in the annular region.
( )0.5
VSG μG ρG Several correlations were developed for predicting the interfacial
ϕ = 104 (8)
σ ρL fraction factor. Among them, Willis (1969) proposed a simple model to
predict the interfacial friction factor using the film thickness.
Due to the small film thickness and quite low flow velocity, little gas
δ
bubbles are contained in the annular liquid film and therefore the gas fI = 0.005[1 + 300 ] (19)
d
phase only flows at the gas core region. Thus, the continuity equation of
the gas phase is given by: Or
ρG VSG A = ρG VG AC αC (9) fI = 0.005[1 + 75(1 − α)] (20)

Or where α is the overall void fraction of the annular flow. Equation (20)
VSG A = VG AC αC (10) works as long as the entrained liquid droplets could be ignorable
(αC →1).
where VSG is the superficial velocity of gas and VG is the actual gas ve­ By now, a hydrodynamic model of annular flows has been estab­
locity in the gas core. lished. Compared with most of the existing models of two-phase annular
flows which requires complicated numerical simulation, only a simple
2.2. Momentum equations iteration in terms of the liquid film thickness is used in current model.
The liquid film thickness is firstly assumed and substituted into the
The momentum equation of the annular liquid film is formulated by: model to estimate the specific flow parameters, such as the pressure
drop, overall void fraction, and actual flow velocities of liquid film and
− AF
dP
− τF SF + τI SI − ρF AF gsinθ = 0 (11) gas core, etc. Then, the assumed liquid film thickness will change and
dz approach to the true value at each iteration step. When the calculated
liquid film thickness approaches infinitely to the assumed value
where AF is the cross-sectional area of the annular liquid film region. dP
dz is

3
S. Jia and C. Dong Applied Thermal Engineering 219 (2023) 119431

(δcal − δass, i ≤ 10− 6 ) at the ith step, δass, i is considered as the true value of
K(qF SF + qI SI ) = − qI SI (28)
the liquid film thickness. Then, the other flow flow parameters, such
asVG ,VD ,VF , andαC , can be estimated using the calculated value of the
where
liquid film thickness.
CPG VG AG ρG
K= (29)
3. Heat transfer model CPF VF AF ρF

The schematic of the temperature field of the annular flow is shown Replacing qF and qI by Equations (23) and (24) in Equation (27) yields:
in Fig. 2. The inlet temperature in annular film and gas core are TFi K[hF (TF − TS )SF + hI (TF − TC )SI ] = hI (TF − TC )SI (30)
andTCi , and the outlet temperature are TFo andTCo , respectively. The
surrounding temperature isTS . The overall heat transfer flux is given as:
As shown in Fig. 2, the heat transfer occurs at the pipe wall between
the annular liquid film and the pipe wall and at the interface between q2Φ = h2Φ (Tm − TS )πd (31)
the liquid film and the gas core. The contact area is A andAC , respec­
Or
tively. Therefore, the energy balance equation of the liquid film in the
control volume with the length of dL is formulated by: h2Φ =
q2Φ
(32)
(Tm − TS )πd
(TFo − TFi )CPF VF AF ρF = − qF SF dL − qI SI dL (21)
where Tm is the mass flow-based bulk temperature of the gas–liquid
Or mixture.
∂TF qF SF + qI SI Therefore, the next is to identify the local heat transfer coefficients at
= − (22) the pipe wall and gas–liquid interface. Gnielinski [20] proposed a cor­
∂L CPF VF AF ρF
relation for predicting the heat transfer coefficients of turbulent flows
where qF and qI are the heat transfer fluxes between the annular liquid (Re ≥ 1 × 104 ) in concentric annular ducts. For annular two-phase flow,
film and the pipe wall and between the annular liquid film and the gas the liquid film flows in the annular region and thus Gnielinski correla­
core, respectively. tion (2009) is just used to calculate the heat transfer coefficient of the
annular liquid film. The expression of Gnielinski correlation is given by:
qF = hF (TF − TS ) (23)
(fann /8)RePr dH
qI = hI (TF − TC ) (24) Nu = √̅̅̅̅̅̅̅̅̅̅̅̅ [1 + ( )2/3 ]Fann k (33)
k1 + 12.7 fann /8(Pr2/3 − 1) L

where hF and hI are the local heat transfer coefficients between the where k1 and k are the parameters.
annular liquid film and the pipe wall and between the annular liquid
film and the gas core, respectively. k1 = 1.07 +
900
−
0.63
(34)
The energy conservation equation of the gas core is formulated by: Re 1 + 10Pr

(TCo − TCi )CPC VG AC ρC = qI SI dL (25) (

PrB
)0.11
K= (35)
PrW
Or
∂TC qI SI The friction factor of the annulus flow,fann , is given as:
= (26)
∂L CPC VG AC ρC fann = (1.8log10 Re* − 1.5)− 2
(36)
Considering the equal temperature gradients of the annular liquid film [ ( )2 ] [ ( )2 ]
and gas core in the control volume, combing Equations (22) and (26) by 1 + ddoi ln ddoi + 1 − ddoi
eliminating ∂∂TL yields: Re* = Re [ ( )2 ] (37)
1 − ddoi ln ddoi
CPC VG AC ρC
(qF SF + qI SI ) = − qI SI (27)
CPF VF AF ρF
where di and do are the diameters of the inner and outer tubes.
or Fann is the factor to take into account the dependence onddoi . Fann is
estimated as follows. [20]:

Fig. 2. Schematic of the temperature field in the control volume of annular flow.

4
S. Jia and C. Dong Applied Thermal Engineering 219 (2023) 119431

( )0.6
di cient at the same liquid flow rates, is introduced.
Fann = 0.9 − 0.15 (38)
do h2Φ
Φ2h = (45)
h1Φ
For the laminar liquid film flow in concentric annular ducts, the heat
transfer coefficients only depend onddoi . Table 1 presents the variation of At this point, a new model of flow and heat transfer for non-boiling two-
Nusselt number on ddoi for laminar falling film flows in concentric annular phase annular flow is mathematically formulated. The influence of ori­
entations and flow parameters, such as the pipe inclination angles, void
ducts [39]. Therefore, the heat transfer coefficients of laminar annular
fraction, superficial liquid and gas velocities, and pressure drop, on the
liquid film are determined using the interpolation method in terms ofddoi .
heat transfer coefficients will be well examined in the next section.
What’s more, the interpolation method is also applied in identifying the
heat transfer coefficients of transitional liquid film flow (2.3 × 103 < 4. Results and discussion
ReF < 1.0 × 104 ) in terms ofReF .
As for the heat transfer coefficient at the interface, Zhang et al., [52] 4.1. Model validation
indicated that the interfacial heat transfer coefficients were equal to the
heat transfer coefficients of the gas core. To evaluate the predictive performance of the new model, the
NuC λC comparison between the experimental and calculated void fractions,
hI = hC = (39) liquid film thicknesses, and two-phase heat transfer coefficients is con­
dC
ducted in this section. Two statistical indices, namely, mean relative
where hI is the heat transfer coefficient at the interface and hC is the heat deviation,mrel , and mean absolute relative deviation,mrel, ab , are intro­
transfer coefficient of the gas core.NuC ,λC , and dC are the Nusselt num­ duced to evaluate the predictive performance of the new model quan­
ber, thermal conductivity, and equivalent diameter of the gas core, titively. The definitions of these statistical indices are given as follows.
respectively. Recently, Meyer et al. [30] developed a well-validated heat
1 N jcal − jexp
transfer correlation of laminar, transitional, quasi-turbulent, and tur­ mrel = Σ × 100% (46)
N i = 1 jexp
bulent flows in circular tubes based on a large amount of experimental
data. Thus, Meyer et al. correlation (2019) is adopted to predict the heat ⃒ ⃒
1 N ⃒jcal − jexp ⃒
transfer coefficient of the gas core. The expression of Meyer et al. cor­ mrel = Σ × 100% (47)
relation (2019) is given as follows. N i= 1 jexp
For laminar flow.
where N is the number of the samples and j is the parameters, such as
NuC = 4.36 + NuC1 + NuC2 (40) void fraction, liquid film thickness, and two-phase heat transfer
coefficient.
where The void fraction is one of the key parameters in annular flows. The
1[ ] accurate calculation of void fraction is very important to the analysis of
NuC1 = − 0.84PrC− 0.2 LtMCD + 0.72(ReC dC )0.54 PrC0.34 LtMCD
0.46
(41) flow mechanism. Fig. 3 presents the comparison between the simulated
L
and experimental void fractions of annular flows. The experimental data
1( ) were taken from Spedding and Nguyen [41], Mukherjee [32], Bhagwat
NuC2 = 0.207GrC0.305 − 1.19 PrC0.42 (ReC dC )− 0.08
(L − LtMCD ) (42)
L and Chajar [3], Luo et al. [29], and Bhagwat and Chajar (2017). The
pipe inclination angles ranges from − 90◦ to + 90◦ . The pipe diameter
where LtMCD is the mixed convection thermal entrance length and L is the
ranges from 0.0125 m to 0.0600 m. Both laminar and turbulent flows are
pipe length.
involved. Fig. 3 indicates that all the void fractions are predicted within
⎧ 0.6 ± 10 % errors without any systematic bias. The mean relative deviation
⎨ 2.4ReC PrC dC for L > LtMCD
⎪
LtMCD = 0.57
GrC (43) is − 1.50 % and mean absolute relative deviation is 3.33 %.
⎪
⎩ The liquid flows in the annular-shaped channel as thin liquid film,
L for L < LtMCD
forced by the gas core. Thus, the liquid film thickness is critical to the
flow and heat transfer process. Fig. 4 presents the comparison of the
For quasi-turbulent and turbulent flow
experimental and simulated mean thickness of the liquid film. The
( )0.11
PrC experimental data were taken from Shedd and Newell [40], Wang et al.
NuC = 0.018Re−C 0.25 (ReC − 500)1.07 PrC0.42 (44)
PrC,w [49], and Uche [45]. Both laminar and turbulent flows are involved.
Fig. 4 indicates that most of the mean thickness of the liquid film are
whereReC ,PrC , and GrC are the Reynolds number, Prandtl number, and predicted within ± 20 % errors without any systematic bias. The mean
Grashof number of the gas core, respectively. relative deviation is − 3.35 % and the mean absolute relative deviation is
To clearly demonstrate the heat transfer difference between two- and 8.39 %.
single-phase flows, a heat transfer multiplier,Φ2h , defined as the ratio of Apart from the flow parameters, the accurate prediction of the heat
two-phase heat transfer coefficient to single-phase heat transfer coeffi­ transfer parameters is also very important. Fig. 5 presents the compar­
ison of the experimental and simulated heat transfer coefficients of
Table 1 annular flows. The experimental data were collected from Rezkallah and
Summary of the Nu number of laminar falling film in concentric annular ducts Sims [37], Ghajar and Tang [18], Bhagwat and Ghajar [2], Nada [34],
[39]. and Bhagwat and Ghajar [4]. Both laminar and turbulent flows are
involved. 85.2 % of the heat transfer coefficients are predicted within ±
di /do Nui (Heat insulation of external Nuo (Heat insulation of internal
surface) surface) 20 % errors without any systematic bias. The mean relative deviation is
− 5.92 % and the mean absolute relative deviation is 10.3 %.
0 / 3.66
0.05 17.46 4.06 The small discrepancy between simulated and experimental param­
0.10 11.56 4.11 eters results from several reasons, such as the adoption of empirical
0.25 7.37 4.23 correlations for predicting entrainment rates and heat transfer co­
0.50 5.74 4.53 efficients, and the measurement uncertainties of the experimental data.
1.00 4.86 4.86
However, most of the parameters are accurately predicted around the

5
S. Jia and C. Dong Applied Thermal Engineering 219 (2023) 119431

Fig. 3. Comparison of the simulated and experimental void fractions.

Fig. 4. Comparison of the simulated and experimental thickness of the liquid film.

experimental data without any systematic bias. In conclusion, the Bhagwat and Ghajar [4] flow regime map. The selected superficial liquid
newly-developed model of flow and heat transfer for two-phase annular velocity ranges from 0.03 m/s to 2.2 m/s and the superficial liquid
flow is well validated. The pipe orientations ranges from − 90◦ to + 90◦ . Reynolds number ranges from 4.70 × 102 to3.45 × 104 . Both laminar
What’s more, since no mass transfer between the liquid film and gas is and turbulent liquid flows are involved. The selected superficial gas
considered, the new model is only suitable in predicting the flow and velocity ranges from 17 m/s to 50 m/s and the superficial gas Reynolds
heat transfer performance of non-boiling two-phase flows. number ranges from 1.47 × 104 to4.33 × 104 . To investigate the effect of
pipe inclination angles on the heat transfer coefficients, θ = 0◦ , 5◦ , 10◦ ,
4.2. Effects of orientations and flow parameters on heat transfer 20◦ , 30◦ , 45◦ , 60◦ , 70◦ , 80◦ , and 90◦ are simulated. Table 2 summarizes
coefficient part of the simulated flow conditions (θ = 90o ).

1200 flow conditions are simulated to analyze the effect of the local 4.2.1. Superficial liquid Reynolds number
and overall orientations and flow parameters on the heat transfer pro­ Fig. 7 presents the variation of the heat transfer coefficients,h2Φ , and
cess using the newly-developed model. The pipe diameter is selected as two-phase heat transfer multiplier,Φh , with the superficial liquid Rey­
0.0125 m. As stated by Bhagwat and Ghajar [4], two-phase annular flow nolds number in horizontal pipes. The superficial liquid Reynolds
occurs at high flow rates of gas and low-to-moderate flow rates of liquid. number ranged from 4.70 × 102 to3.45 × 104 . Fig. 7 indicates that the
Fig. 6 shows part of the selected flow conditions (θ = 0o ) located in influencing mechanism of the superficial liquid Reynolds number on the

6
S. Jia and C. Dong Applied Thermal Engineering 219 (2023) 119431

Fig. 5. Comparison of the simulated and experimental heat transfer coefficients.

annular flow.
In Region II, the two-phase heat transfer multipliers decrease, while
the two-phase heat transfer coefficients increase rapidly with the su­
perficial liquid Reynolds number. The reason is as ReSL increases, the
intrinsic flow turbulence of liquid flow increases and the heat transfer
process is intensified but the enhancing effect of gas on the heat transfer
process is reduced compared with the two-phase flow in Region I. Re­
gion II is considered as the turbulent-transitional two-phase annular
flow.
In Region III, the two-phase heat transfer multiplier continues
decreasing, indicating that the enhancing effect of gas on the heat
transfer process became less prominent. The reason is in Region III, the
liquid flow is fully turbulent flow and the role of gas in improving the
flow turbulence is reduced. However, all the two-phase heat transfer
multipliers are higher than 1.0, indicating that the gas is still effective in
enhancing the heat transfer process in Region III. Region III is consid­
ered as the turbulent-turbulent two-phase annular flow.
Fig. 8 presents the effect of the superficial liquid Reynolds number on
the heat transfer of two-phase flows with different pipe inclination an­
gles. The pipe inclination angle ranges from 0◦ to 90◦ and
ReSG = 14738 is fixed. Although the heat transfer are similar in
different pipe inclination angles, the heat transfer coefficient decreases
Fig. 6. Flow conditions in the flow regime map proposed by Bhagwat and slightly as the pipe inclination angles increase from 0◦ to 90◦ . The reason
Ghajar [4]. is due to the gravity effect, the velocity of the liquid film decreases with
the increase of the pipe inclination angles, and the flow turbulence and
heat transfer process is quite complicated. To clearly express its influ­ heat transfer coefficient decrease as well.
encing mechanism, the entire region is divided into three sub-regions.
In Region I, the two-phase heat transfer coefficient decreases with 4.2.2. Superficial gas Reynolds number
the increase of the superficial liquid velocity. The reason is the liquid Fig. 9 shows the effect of superficial gas Reynolds number on the heat
film is laminar due to the low liquid flow rate in Region I and the heat transfer of two-phase annular flows. The superficial gas Reynolds
transfer performance of the liquid film in the pipes with uniform cross number ranges from 1.47 × 104 to4.33 × 104 . The influence of the su­
section depends little on the ReF [42] and only depends on di /do at perficial gas Reynolds number on the heat transfer varies in different
laminar liquid film. As ReSL increases, the liquid film thickness increases regions. When ReSL ≤ 10000 (laminar and transitional flows), the flow is
and the heat transfer performance is deteriorated, which is consistent not fully developed and the introduction of gas could significantly
with the expectation of higher heat transfer across a thin film. What’s improve the flow turbulence. Thus, the two-phase heat transfer multi­
more, as shown in Fig. 7(b), the two-phase heat transfer multiplier is pliers are as high as 6.59 and 3.64 at ReSL = 470 andReSL = 9408,
highest in the entire region. It means the enhancing effect of gas on the respectively, with the first introduction of gas (ReSG = 14738), as shown
heat transfer process is most effective in Region I. The reason is in Fig. 9(b). The reason is when ReSL is low, with a certain value of the
whenReSL < 2300, the flow is laminar flow and the injection of gas could liquid, more momentum is transferred from the gas to the liquid, and the
effectively enhance the flow turbulence of the liquid film, rapidly reduce enhancing effect is more pronounced at a low flow rate of the liquid.
the resistance to heat transfer, and increase the heat transfer coefficient However, due to the lowReSL , the first introduction of gas (ReSG =
accordingly. Region I is considered as the turbulent-laminar two-phase 14738) could improve the flow turbulence significantly and further

7
S. Jia and C. Dong Applied Thermal Engineering 219 (2023) 119431

Table 2
Summary of the simulated flow conditions atθ = 90o .
( )] ( )] ( )]
VSL [m/s] VSG h2Φ [W/ m2 ⋅K VSL [m/s] VSG h2Φ [W/ m2 ⋅K VSL [m/s] VSG h2Φ [W/ m2 ⋅K
[m/s] [m/s] [m/s]
0.03 17 1706.6 0.03 20 1898.0 0.03 23 2059.5
0.06 17 1839.9 0.06 20 2037.3 0.06 23 2205.1
0.09 17 1925.2 0.09 20 2125.9 0.09 23 2297.3
0.12 17 1988.9 0.12 20 2191.7 0.12 23 2365.6
0.20 17 2492.8 0.2 20 2669.0 0.2 23 2801.9
0.40 17 5796.8 0.4 20 6153.6 0.4 23 6428.0
0.60 17 10353.3 0.6 20 10883.7 0.6 23 11295.2
0.80 17 12579.2 0.8 20 13626.9 0.8 23 14608.2
1. 17 13857.1 1 20 14904.8 1 23 15891.0
1.2 17 15046.4 1.2 20 16090.3 1.2 23 17076.9
1.4 17 16170.6 1.4 20 17207.9 1.4 23 18191.9
1.6 17 17244.6 1.6 20 18273.2 1.6 23 19252.1
1.8 17 18278.6 1.8 20 19296.6 1.8 23 20268.6
2.0 17 19279.6 2 20 20285.6 2 23 21249.2
2.2 17 20253.1 2.2 20 21245.7 2.2 23 22199.6
0.03 28 2289.2 0.03 32 2449.8 0.03 36 2596.3
0.06 28 2444.1 0.06 32 2611.1 0.06 36 2763.1
0.09 28 2542.0 0.09 32 2713.0 0.09 36 2868.5
0.12 28 2614.5 0.12 32 2788.5 0.12 36 2946.7
0.2 28 2953.7 0.2 32 3029.0 0.2 36 3074.0
0.4 28 6741.2 0.4 32 6890.5 0.4 36 6970.0
0.6 28 11764.9 0.6 32 11984.5 0.6 36 12093.9
0.8 28 16116.8 0.8 32 17227.4 0.8 36 18100.2
1 28 17414.0 1 32 18539.1 1 36 19593.2
1.2 28 18606.5 1.2 32 19739.6 1.2 36 20803.0
1.4 28 19722.4 1.4 32 20859.1 1.4 36 21927.6
1.6 28 20779.5 1.6 32 21916.6 1.6 36 22986.8
1.8 28 21789.7 1.8 32 22924.6 1.8 36 23994.2
2 28 22761.3 2 32 23892.0 2 36 24959.0
2.2 28 23700.5 2.2 32 24825.4 2.2 36 25888.2
0.03 40 2731.9 1.4 40 22935.4 0.4 50 6869.2
0.06 40 2903.4 1.6 40 23997.3 0.6 50 11867.6
0.09 40 3012.0 1.8 40 25005.0 0.8 50 17682.3
0.12 40 3092.5 2 40 25968.3 1 50 22841.0
0.2 40 3095.4 2.2 40 26894.4 1.2 50 24084.6
0.4 40 6993.1 0.03 50 3036.5 1.4 50 25229.7
0.6 40 12113.3 0.06 50 3217.2 1.6 50 26299.6
0.8 40 18096.4 0.09 50 3332.1 1.8 50 27310.1
1 40 20585.6 0.12 50 3417.5 2 50 28271.9
1.2 40 21805.1 0.2 50 3576.5 2.2 50 29193.1

Fig. 7. Effect of superficial liquid Reynolds number on heat transfer of two-phase annular flow in horizontal pipes.

increase in gas flow rate could not result in pronounced increase in the in the absolute two-phase heat transfer coefficient, as shown in Fig. 9(a).
absolute two-phase heat transfer coefficient, as shown in Fig. 9(a).
WhenReSL > 10000, the flow is fully developed and higher gas flow rate 4.2.3. Reynolds number and thickness of liquid film
is required to improve its flow turbulence. Thus, due to the highReSL , In two-phase annular flow, the liquid is assumed to flow in the
further increase in the gas flow rate could result in pronounced increase annular-shaped channel as liquid film. Thus, the local flow parameters

8
S. Jia and C. Dong Applied Thermal Engineering 219 (2023) 119431

Fig. 8. Effect of superficial liquid Reynolds number on heat transfer of two-phase annular flow in different pipe inclination angles.

Fig. 9. Effect of the superficial gas Reynolds number on the two-phase heat transfer coefficients.

of the liquid film, such as the Reynolds number and thickness, are
important to the heat transfer process. Due to the interrelated rela­
tionship between the film thickness, two-phase heat transfer coefficient,
and film flow rate, 3-D plots in terms of two-phase heat transfer coef­
ficient/multiplier, film thickness, and Reynolds number of liquid film is
adopted. Fig. 10 presents the dependence of two-phase heat transfer
multiplier on the liquid film thickness and Reynolds number of the
liquid film. The two-phase heat transfer multipliers first decrease rapidly
then slightly with both the Reynolds number and liquid film thickness.
This is because as ReF increases, the intrinsic flow turbulence increases
and the enhancing effect of gas becomes less prominent. What’s more, as
δF increases, the gas flow only accelerates the interfacial liquid instead of
the interior liquid close to the pipe wall, which also attenuate the
enhancing effect.

4.2.4. Void fraction and pressure drop

Fig. 11 shows the variation of the two-phase heat transfer multipliers
with the void fractions. The liquid holdup, 1-α, is used in the figure.
Apparently, the two-phase heat transfer multiplier depends heavily on
the void fraction. As the void fraction decreases (liquid holdup in­
creases), the two-phase heat transfer multiplier decreases rapidly. Fig. 10. Effect of the Reynolds number and thickness of liquid film on two-
phase heat transfer multiplier.

9
S. Jia and C. Dong Applied Thermal Engineering 219 (2023) 119431

using the Chilton-Colburn analogies. The formulation is expressed as:

( )n
Φ2h = (1 − α)m Φ2f (48)

where m and n are the exponents. Here, m and n are determined using
the simulated data. Thus, the relationship between the heat transfer
multiplier and the pressure drop multiplier is quantitatively expressed
as:
( )0.400
Φ2h = (1 − α)− 0.0652 Φ2f (49)

Fig. 13 evaluates the predictive performance of the new correlation. The

calculated heat transfer multiplier refers to the value calculated by the
new correlation, Equation (49), and the simulated heat transfer multi­
plier refers to the value calculated by the model, as expressed in Sections
2 and 3. As shown in Fig. 13, although a very simple formulation is used,
most of the data fall within ± 20 % errors without any systematic bias.
What’s more, Equation (47) also achieve the aptitude that when Φ2f
approaches 1.0 and α approaches 0, Φ2h could approach 1.0.

Fig. 11. Effect of the void fraction on two-phase heat transfer multiplier. 5. Conclusions

This paper developed a model of flow and heat transfer for turbulent-
What’s more, Φ2h approaches 1.0, as α approaches 0 infinitely. It is
laminar/turbulent gas–liquid two-phase annular flows. The flow and
physically correct, since α→0 means the two-phase flow approaches
heat transfer behaviors of two-phase annular flow are investigated. The
single-phase flow and the two-phase heat transfer coefficient approaches
effect of orientations and flow parameters, such as the superficial Rey­
single-phase heat transfer coefficient.
nolds numbers of the liquid and gas, Reynolds number and thickness of
The pressure drop multiplier,Φ2f , defined as the ratio of the pressure liquid film, void fraction, pressure drop, and inclination angles, on the
drop of two-phase flow to that of single-phase flow, is used in the heat transfer coefficients were investigated extensively. The achieve­
analysis. Fig. 12 shows that the two-phase heat transfer multipliers ments are as follows:
depend heavily on the pressure drop multipliers. The heat transfer
multipliers increase rapidly with the pressure drop multipliers. What’s (1) The physical process of flow and heat transfer for annular flow
more, Φ2h approaches 1.0, as Φ2f approaches 1.0 infinitely. It is physically was mathematically formulated. A model of flow and heat
correct, since Φ2f →1.0 means the two-phase flow approaches single- transfer for turbulent-laminar/turbulent two-phase annular flows
phase flow and the two-phase heat transfer coefficient approaches was developed. The new model achieves both simplicity and high
single-phase heat transfer coefficient. accuracy.
Chilton & Colburn analogy indicates that the heat transfer is analo­ (2) The heat transfer enhancement by the injection of gas into liquid
gous to the momentum transfer [53,50]. Recently, Dong and Hibiki was investigated comprehensively. The two-phase heat transfer
[10,11] derived a formulation to express the relationship between the multiplier decreased with the superficial liquid Reynolds number,
heat transfer multipliers, void fractions and pressure drop multipliers but increased with the superficial gas Reynolds number.
(3) The relationships between the two-phase heat transfer multipliers
and the void fractions and pressure drop multipliers were

Fig. 12. Effect of the pressure drop multiplier on two-phase heat trans­ Fig. 13. Comparison between the simulated and calculated two-phase heat
fer multiplier. transfer multipliers.

10
S. Jia and C. Dong Applied Thermal Engineering 219 (2023) 119431

identified quantitatively. A simple correlation of heat transfer [21] H. He, L. Pan, Q. Ren, T. Ye, D. Zhang, Local liquid film behavior of annular two-
phase flow on rod-bundle geometry-II. Modeling and verification, Int. J. Heat Mass
multiplier was developed based on Chilton & Colburn analogy
Tran. 143 (2019), 118533.
and a large amount of data, acting as a rough but quick prediction [22] H. He, L. Pan, Q. Ren, T. Ye, D. Zhang, Local liquid film behavior of annular two-
of the heat transfer multipliers of two-phase annular flows. phase flow on rod-bundle geometry-I. Experimental phenomenon and analysis, Int.
J. Heat Mass Tran. 141 (2019) 58–70.
[23] M. Ishii, T. Hibiki (Eds.), Thermo-Fluid Dynamics of Two-Phase Flow, Springer
Data availability New York, New York, NY, 2010.
[24] D. Ito, P. Papadopoulos, H.M. Prasser, Liquid film dynamics of two-phase annular
flow in square and tight lattice subchannels, Nucl. Eng. Des. 300 (2016) 467–474.
The data that support the findings of this study are available from the [25] P. Ju, Y. Liu, X. Yang, M. Ishii, Wave characteristics of vertical upward adiabatic
corresponding author upon reasonable request. annular flow in pipes, Int. J. Heat Mass Tran. 145 (2019), 118701.
[26] N. Kattan, J.R. Thome, D. Favrat, Flow boiling in horizontal tubes. Part 1:
Development of adiabatic two-phase flow pattern map, J. Heat Transfer 120 (1998)
140–147.
Declaration of Competing Interest [27] K. Keniar, S. Garimella, Modeling of annular and intermittent flow condensation in
micro- and mini- channels with experimental validation, Int. J. Heat Mass Tran.
176 (2021), 121507.
The authors declare that they have no known competing financial [28] P. Kumar, A.K. Das, S.K. Mitra, Physical understanding of gas-liquid annular flow
interests or personal relationships that could have appeared to influence and its transition to dispersed droplets, Phys. Fluids 28 (2016), 072101.
the work reported in this paper. [29] W. Luo, Y. Li, Q.H. Wang, J.L. Li, R.Q. Liao, Z.L. Liu, Experimental study of gas-
liquid two-phase flow for high velocity in inclined medium size tube and
verification of pressure calculation methods, Int. J. Heat Technol. 34 (3) (2016)
Data availability 455–464.
[30] J.P. Meyer, M. Everts, N. Coetzee, K. Grote, M. Steyn, Heat transfer coefficients of
laminar, transitional, quasi-turbulent and turbulent flow in circular tubes, Int.
Data will be made available on request.
Commun. Heat Mass Transf. 105 (2019) 84–106.
[31] T.A. Moreira, R.W. Morse, K.M. Dressler, G. Ribatski, A. Berson, Liquid-film
thickness and disturbance-wave characterization in a vertical, upward, two-phase
Acknowledgments
annular flow of saturated R245fa inside a rectangular channel, Int. J. Multiph.
Flow 132 (2020) 103412.
The project is supported by the National Natural Science Foundation [32] H. Mukherjee, An experimental study of inclined two-phase flow, The University of
of China (NSFC), No. 51906071 and No. 51936005; the Natural Science Tulsa, Tulsa, USA, 1979.
[33] S. Mukhopadhyay, G. Nimbalkar, Fundamental study on chaotic transition of two-
Foundation of Guangdong Province, No. 2019A1515011536; Young phase flow regime and free surface instability in gas deaeration process, Exp.
Talent Support Project of Guangzhou Association for Science and Comput. Multiph. Flow 3 (4) (2021) 258–288.
Technology. [34] S.A. Nada, Experimental investigation and empirical correlations of heat transfer in
different regimes of air-water two-phase flow in a horizontal tube, J. Therm. Sci.
Eng. Appl. 9 (2017), 021004-1-7.
References [35] N. Panicker, A. Passalacqua, R.O. Fox, Computational study of buoyancy driven
turbulence in statistically homogeneous bubbly flows, Chem. Eng. Sci. 216 (2020),
115546.
[1] A.M. Aliyu, Y.D. Baba, L. Lao, H. Yeung, K. Kim, Interfacial friction in upward
[36] M.C. Potter, D.C. Wiggert, B.H. Ramadan, Engineering Fluids Mechanics, 4th ed.,
annular gas-liquid two-phase flow in pipes, Exp. Therm. Fluid Sci. 84 (2017)
Cengage Learning, 2012.
90–109.
[37] K.S. Rezkallah, G.E. Sims, New data on two-phase, two-component heat transfer
[2] S.M. Bhagwat, A.J. Ghajar, Similarities and differences in the flow patterns and
and hydrodynamics in a vertical tube, J. Thermophys. 3 (2) (1989) 213–219.
void fraction in vertical upward and downward two phase flow, Exp. Therm. Fluid
[38] A. Saxena, H.M. Prasser, A study of two-phase annular flow using unsteady
Sci. 39 (2012) 213–227.
numerical computations, Int. J. Multiph. Flow 126 (2020), 103037.
[3] S.M. Bhagwat, A.J. Ghajar, Experimental investigation of non-boiling gas liquid
[39] R.K. Shah, A.L. London, Laminar flow forced convection in ducts, Academic Press,
two phase flow in upward inclined pipes, Exp. Therm. Fluid Sci. 79 (2016)
1978.
301–318.
[40] T. Shedd, T.A. Newell, Characteristics of the liquid film and pressure drop in
[4] S.M. Bhagwat, A.J. Ghajar, Experimental investigation of non-boiling gas-liquid
horizontal, annular, two-phase flow through round, square and triangular tubes,
two phase flow in downward inclined pipes, Exp. Therm. Fluid Sci. 89 (2017)
J. Fluid Eng. 126 (2004) 807–817.
219–237.
[41] L. Spedding, V.T. Nguyen, Data on holdup, pressure loss and flow patterns for two-
[6] C. Chalons, M. Girardin, S. Kokh, An all-regime Lagrange-Projection like scheme
phase air water flow in an inclined pipe, University of Auckland, Auckland New
for 2D homogeneous models for two-phase flows on unstructured meshes,
Zealand Report Eng, 1976, p. 122.
J. Comput. Phys. 335 (2017) 885–904.
[42] W.Q. Tao, Heat Transfer Engineering. Higher Education Press, 5th Edition., 2019.
[7] A. Cioncolini, J.R. Thome, Void fraction prediction in annular two-phase flow, Int.
[43] J.R. Thome, Update on advances in flow pattern based two-phase heat transfer
J. Multiph. Flow 43 (2012) 72–84.
models, Exp. Therm. Fluid Sci. 29 (2005) 341–349.
[8] A. Clausse, M.L. Bertodano, Natural modes of the two-fluid model of two-phase
[44] J.R. Thome, Encyclopedia of two-phase heat transfer and flow I, World Scientific,
flow, Phys. Fluids 33 (2021), 033324.
2016.
[9] Deendarlianto, A. Rahmandhika, A. Widyatama, O. Dinaryanto, A. Widyaparaga,
[45] O. Uche, Evaluation of liquid film thickness in gas-liquid annular flow in horizontal
Indarto, Experimental study on the hydrodynamic behavior of gas-liquid air-water
pipes using three methods, Int. J. Energy Environ. Sci. 5 (4) (2020) 57–65.
two-phase flow near the transition to slug flow in horizontal pipes, Int. J. Heat
[46] K. Ueyama, A study of two-fluid model equations, J. Fluid Mech. 690 (2012)
Mass Tran. 130 (2019) 187–203.
474–498.
[10] C. Dong, T. Hibiki, Correlation of heat transfer coefficient for two-component
[47] G.B. Wallis, One dimensional two-phase flow, McGraw-Hill, New York, 1969.
two-phase slug flow in a vertical pipe, Int. J. Multiphase Flow 108 (2018)
[48] G. Wang, Z. Deng, M. Ishii, Wave structure and velocity in vertical upward annular
124–139.
two-phase flow, Exp. Therm. Fluid Sci. 120 (2021), 110205.
[11] C. Dong, T. Hibiki, Heat transfer correlation for two-component two-phase slug
[49] M. Wang, D.D. Zheng, Y. Xu, Z.Q. Cui, Liquid film thickness measurement for gas-
flow in horizontal pipes, Appl. Therm. Eng. 141 (2018) 866–876.
liquid two phase flow using ultrasound. 2019 IEEE International Instrumentation
[12] C. Dong, R. Qi, L. Zhang, Flow and heat transfer for a two-phase slug flow in
and Measurement Technology Conference (I2MTC), 2019.
horizontal pipes: A mechanistic model, Phys. Fluids 33 (2021), 034117.
[50] C. Wenzel, T. Gibis, M. Kloker, U. Rist, Reynolds analogy factor in self-similar
[13] C. Dong, R. Qi, L. Zhang, Mechanistic modeling of flow and heat transfer in
compressible turbulent boundary layers with pressure gradients, J. Fluid Mech.
turbulent-laminar/turbulent gas-liquid stratified flow, Phys. Fluids 33 (7) (2021)
907 (2021) R4.
073313.
[51] F. Xie, X. Zheng, M.S. Triantafyllou, Y. Constantinides, Y. Zheng, G.E. Karniadakis,
[14] C. Dong, R. Qi, L. Zhang, Mechanistic modelling of flow and heat transfer in
Direct numerical simulations of two-phase flow in an inclined pipe, J. Fluid Mech.
vertical upward two-phase slug flows, Phys. Fluids 34 (2022), 013309.
825 (2017) 189–207.
[15] C. Dong, S. Rassame, L. Zhang, T. Hibiki, Drift-flux correlation for upward two-
[52] H. Zhang, Q. Wang, C. Sarica, J.P. Brill, Unified model of heat transfer in gas-liquid
phase flow in inclined pipes, Chem. Eng. Sci. 213 (2020), 115395.
pipe flow, SPE Prod. Oper. (2006) 114–122.
[16] W. Fan, C. Li, L. Li, X. Zhang, M. Du, X. Yin, Hydrodynamics of gas/shear-thinning
liquid two-phase flow in a co-flow mini-channel: Flow pattern and bubble length,
Phys. Fluids 32 (2020), 092004.
[18] A.J. Ghajar, C.C. Tang, Heat transfer measurements, flow pattern maps and flow
visualization for non-boiling two-phase flow in horizontal and slightly inclined
pipes, Heat Transf. Eng. 28 (6) (2007) 525–540.
[20] V. Gnielinski, Heat transfer coefficients for turbulent flow in concentric annular
ducts, Heat Transf. Eng. 30 (6) (2009) 431–436.

11
S. Jia and C. Dong Applied Thermal Engineering 219 (2023) 119431

[53] Y. Zhang, W. Bi, F. Hussain, Z. She, A generalized Reynolds analogy for [55] A.H. Zitouni, A. Arabi, Y. Salhi, Y. Zenati, E.K. Si-Ahmed, J. Legrand, Slug length
compressible wall-bounded turbulent flows, J. Fluid Mech. 739 (2014) 392–420. and frequency upstream a sudden expansion in gas-liquid intermittent flow, Exp.
[54] Z. Zhang, Z. Wang, H. Liu, Y. Gao, B. Sun, Experimental study on entrained Comput. Multiph. Flow 3 (2) (2021) 124–130.
droplets in vertical two-phase churn and annular flows, Int. J. Heat Mass Tran. 138
(2019) 1346–1358.

12