International Journal of Heat and Mass Transfer 122 (2018) 602–613

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer

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

Numerical investigation on heat transfer performance and flow

characteristics in enhanced tube with dimples and protrusions
Shuai Xie a,⇑, Zheng Liang a, Liang Zhang a, Yulin Wang b, Hu Ding c, Jie Zhang a
a
School of Mechatronic Engineering, Southwest Petroleum University, Chengdu, Sichuan 610500, PR China
b
Si Chuan Chuan Guo Ketaida Energy Technology Co., Ltd, Chengdu, Sichuan 610500, PR China
c
Chen Du Zheng Mao Material Co. Ltd, Chengdu, Sichuan 610500, PR China

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

Article history: In the present investigation, a new design of enhanced tube aiming to improve heat transfer by employ-
Received 12 December 2017 ing dimples and protrusions was put forward. The special enhanced tube with dimples and protrusions is
Received in revised form 5 January 2018 obtained by extruded the conventional plain tube. The objective is to present details of flow field char-
Accepted 23 January 2018
acteristics and heat transfer mechanisms for the ETDP, then effects of protrusion depth, pitch and radius
on thermal-hydraulic performance also being discussed. The operating Reynolds number ranged from
5000 to 30,000 and the validated realizable k-e turbulence model was employed on the numerical sim-
Keywords:
ulations. The local streamlines, velocity contour, temperature contour and Nusselt number were pre-
Dimpled/protruded tube
Heat transfer enhancement
sented to illustrate the heat transfer enhancement mechanisms. From this investigation, it is found
Thermal-hydraulic performance that the varying geometric parameters of ETDP play an important role in thermal-hydraulic characterises.
The main findings are that the ETDP have an advantage for augmented heat transfer rate and PEC com-
pared with the plain tube, due to improved flow mixing, interrupted the boundary layer, formed periodic
jet flows and swirl flows induced by dimples and protrusions. The Nusselt number and friction factor
increase and PEC decrease with an increasing protrusion depth. Among the investigated different protru-
sion pitch, it is found that the friction factor first decrease and then increase with the increase of protru-
sion pitch. For varying the protrusion radius, the ETDP with R = 3 mm have the largest PEC at the most Re.
Under operating condition and geometric parameters considered, the ETDP with D = 3 mm, P = 30 mm
and R = 4 mm offers the largest PEC value of about 1.65.
Ó 2018 Elsevier Ltd. All rights reserved.

1. Introduction In passive technologies families, fins, ribs, twisted taps, wire

In many industrial applications, such as cooling towers, aero-
space industries, oil and gas flow in reservoirs and chemical pro-
cessing, heat exchangers play an important role in heating and
cooling. However, to reduce the energy consumption and increase
economic benefit, the traditional heat exchangers with plain tube
are always not competent enough due to they have low and ineffi-
cient thermal performance. In order to overcome this problem, the
technologies of heat transfer enhancement were put forward.
According to whether it needs external additional power, the heat
transfer enhancement technologies can be classified into active
and passive. Generally, due to reliability in operation, low energy
input and inexpensive cost in maintenance, the passive technolo-
gies have more wide applications and potential promise [1].
⇑ Corresponding author.
E-mail address: xieshuai_1@126.com (S. Xie).
coils, insert device, dimples, etc. are commonly used to enhance
the heat transfer rate. The protrusions/dimples surface can
improve the heat transfer performance with relatively low pres-
sure loss penalty. Therefore, during a few decades, heat transfer
enhancement technologies utilize protrusions/dimples surface
have received much attention for enhancing heat transfer. Xie
et al. [2] and Leontiev et al. [3] studied the flat with dimples, Zheng
et al. [4] and Zahid et al. [5] investigated the plain tube with dim-
ples. Chen et al. [6] and Vicente et al. [7] investigated heat transfer
performance and friction factors of protruded tube by experiment.
They proposed correlations for predicting the Nusselt number and
friction factors. Thianpong et al. [8] conducted experimental work
on dimpled tube with twisted tape inserts. The results show that
the pitch ratio and twist ratio have the significant influence on heat
transfer performance. Wang et al. [9,10] and Liang et al. [11] inves-
tigated the spherical dimpled tube and ellipsoidal dimpled tube by
experiment and simulation. They reported that the ellipsoidal dim-
pled tube provides a better performance than spherical dimpled

https://doi.org/10.1016/j.ijheatmasstransfer.2018.01.106
0017-9310/Ó 2018 Elsevier Ltd. All rights reserved.
 S. Xie et al. / International Journal of Heat and Mass Transfer 122 (2018) 602–613
Nomenclature
A heat transfer area, m2
cp special heat, J kg
1 K
1
D protrusion depth, mm
Dh equivalent diameter, mm
f friction factor
P protrusion pitch, mm
Pr Prandtl number
p pressure, Pa
Dp pressure drop, Pa
m mass flow rate, kg s
1
Nu Nusselt number
R protrusion radius, mm
Re Reynolds number
T temperature, K
u velocity, m s
1
u⁄ friction velocity
tubes (Fig. 1a, b). Chang et al. [12] experimentally studied the heat
transfer performance of plain tube and hexagonal duct with dim-
ples (Fig. 1c). They suggested that the hexagonal duct with dimples
provided the higher flow momentum compared to rectangular
dimpled channels. Heo et al. [13,14] numerically studied the man-
ufacturing process and thermal performance of flat with dimples.
Garcia et al. [15] experimentally studied three types of enhance-
ment techniques, consisting of inclined corrugate tubes, dimpled
tubes and wire coils. Their study provided some recommendations
for choosing an enhancement tube at different Reynolds numbers.
Kumar et al. [16] experimentally studied the enhanced tube with
the protruded surface (Fig. 1d). The results indicated that the pro-
truded surface tube showed a significant enhancement in heat
transfer rate and friction factor over conventional tube. Kukulka
et al. [17] conducted experiments on dimpled tube which have
been named EHT (Fig. 1e). They found that the EHT can enhance-
ment the heat transfer rate, evaporation and condensation heat
transfer coefficient. Li et al. [18,19] conducted experiment and sim-
ulation works on enhanced tube with dimples (Fig. 1f). The results
indicated that the dimples could disturb and mix the boundary
layer and generate secondary flows that improve the turbulence
level. Shafaee et al. [20] and Mashouf et al. [21] experimentally
studied thermal-hydraulic performance of helically dimpled tube
(Fig. 1g). The experiment shows that the heat transfer coefficients
of dimpled tube are 1.29–2 times larger than plain tube, and the
Fig. 1. The application of dimples on enhanced tubes: (a) dimpled and protruded tube [9]; (b) ellipsoidal dimpled tube [11]; (c) hexagonal ducts with dimples [12]; (d)
protruded tube by Kumar et al.[16]; (e) EHT tubes [17]; (f) enhanced dimpled tube by Li et al.[18]; (g) helically dimpled tube by Shafaee [20]; (h) dimpled tube by Aroonrat Li
et al. [22].
603
y+ mesh resolution indicator
Greek symbols
q fluid density, kg m
3
l dynamic viscosity, Pa s
lt turbulent viscosity
k thermal conductivity, W m
1 K
1
U total heat rate, W
H temperature gradient
Subscripts
i inside
max maximum
s plain tube
ref reference
dimples have significantly impacts on the two-phase flow pattern.
Aroonrat et al. [22] experimentally studied investigates the heat
transfer and pressure drop of R-134a during condensation inside
a dimpled tube. They proposed correlations for predicting heat
transfer and pressure drop in the two-phase condition (Fig. 1h).
Even though previous research work have shown that the addi-
tion of dimples or protrusions on tube can provide a realizable heat
transfer enhancement with relatively low pressure loss penalty.
However, only a few researchers pay attention to the ETDP
(enhanced tube with both dimples and protrusions) at the same
time. The main objective of the present study is to present a
numerical investigation of thermal-hydraulic performance of ETDP
at Re (Reynolds number) ranged from 5000 to 30,000. Flow field
characteristics and heat transfer mechanisms of the ETDP are
investigated under the single phase condition. Effects of protrusion
depth, pitchs and radius on thermal-hydraulic performance also
being discussed. Generally, this study may provide some theoreti-
cal guidelines and suggestions in potential application of the ETDP.
2. Physical model
Fig. 2 shows the research processing for constructed the ETDP
geometric model. The geometric model of the ETDP is too compli-
cate to construct by 3D (Three-dimensional) software, such as Por/
E, UG, Solidwrork. Therefore, plastic forming method is employed
604 S. Xie et al. / International Journal of Heat and Mass Transfer 122 (2018) 602–613
Fig. 2. The research process of the ETDP.
to construct the geometric model of ETDP. Fig. 3 presents finite ele-
ment model of the plain tube before the extrude forming. Fig. 4
presents the geometric model of the ETDP after the extruded form-
ing. From the figure, the protrusions dented and the dimples pro-
truded from the plain tube, respectively. Diameter of the plain
tube is 20 mm, wall thickness is 0.5 mm. The main deign parame-
ters contain protrusion depth D, protrusion pitch P and protrusion
radius R. Considering the potential promise in heat exchanger
design and manufacture, the primary parameter is D = 4 mm, P =
30 mm and R = 4 mm, while various arrangements for different
cases are presented in Table 1.
3. Computational details
3.1. Overview
A realizable k-e turbulence model is employed in the present
study to investigate the thermal-hydraulic performance for the
ETDP. The finite modelling is accomplished using the commercial
CFD (Computational fluid dynamics) software FLUENT version
17.2. The SIMPLE (Semi-Implicit Method for Pressure Linked Equa-
tion Consistent) algorithm is adopted for handling the couple
between pressure and velocity fields. The second-order scheme
was employed for discretization of the energy equation and the
turbulent kinetic energy equation. The minimum convergence cri-
Y
Z 20m
Fig. 3. Finite element model of the plain tube and the extrusion heads (only two distances shown in this figure).
Pro/E
Generate smooth tube
ABAQUS
Generate geometry
model of extrued tube
ICEM
Generate mesh
Fluent
Thermal-hydraulic
performanc
CFD-Post
Flow visualization
terion for the continuity, momentum equations, k and e equations
is 10
6, while 10
8 is utilized for the energy equation. Moreover,
the commercial software package ICEM version 17.2 is used to gen-
erate the grid system for plain tubes and ETDPs.
3.2. Selection of turbulence model
It is greatly important to select turbulence model for a specific
simulation. The selected of turbulence model is based upon com-
putational demands and decently accurate predicted flow phe-
nomenon. Overall, four different turbulence models: RNG k-e
model, realizable k-e model, SST k-x model, and the RS (Reynolds
stress) model have been initially considered to compute the Nus-
selt number. The agreement between the predicted Nusselt num-
ber and experimental data is the criterion for selecting a
turbulence model. These experiments are carried out by Wang
et al. [10].
Fig. 5 shows comparison from different turbulence models. It
can be seen that the turbulence models provide similar predicted
results except for the SST k-x model, and realizable k-e model pro-
vided a closer Nusselt number to experiment data than other tur-
bulence models. The deviation of the realizable k-e model from
experiment data is 7.2%, such deviations are small, thus, accept-
able. Overall, the realizable k-e model is employed for simulation
in the present investigation.
60
°
m δ
 S. Xie et al. / International Journal of Heat and Mass Transfer 122 (2018) 602–613 605

Dimple

(a) Protusion

10
10

D
Outer surface of ETDP
Inner surface of ETDP 5
Outer surface of ETDP

Y/(mm)
Inner surface of ETDP
Y/(mm)

Outer surface of plain tube

0 Inner surface of plain tube
0
P
-5
R

-10 -10

0 15 30 -10 -5 0 5 10
Z/(mm) (b) X/(mm)

Fig. 4. Schematic diagram of plain tube after extruded: (a) the 3D view of the ETDP; (b) the plane view of ETDP.

Table 1 100
Geometrical parameters of the protruded tube.
Experiment data[10]
Case Extruded Extruded pitch Extruded Realizable k-ε
depth D (mm) P (mm) radius R (mm) RNG k-ε
80
Case 1 4 30 4 SST k-ω
Case 2 2–5 30 4 RS
Case 3 3 25–40 4
Case 4 3 20 2–5
Nu

60

3.3. Governing equation

40
The numerical simulation involves solving basic conservation
equations. The associated governing equations as follows:
Mass conservation or continuity equation: 20
@ 1 2 3
ðqui Þ ¼ 0 ð1Þ Re/(104 )
@xi
Momentum conservation equation: Fig. 5. Validation of turbulence models by comparison the Nusselt number with

 experiment data [10].
@ @p @ @ui @uj
ðqui uj Þ ¼
þ ð l þ lt Þ þ ð2Þ
@xj @xi @xj @xj @xi 
 
@ @ lt @k
ðqkuj Þ ¼ lþ þ C
qe ð4Þ
Energy conservation equation: @xj @xj rk @xj

 
@ @ l lt @T 
 
ðui TÞ ¼ þ ð3Þ @ @ lt @ e e2
@xi @xi Pr Prt @xi ðqeuj Þ ¼ lþ þ C 1 Ce
C 2 pffiffiffiffiffi ð5Þ
@xj @xj re @xj k þ te
where q is fluid density, u is velocity, p is pressure, l is dynamic vis-
cosity, lt is turbulent viscosity, T is temperature, Pr is the Prandtl where u can be expressed as :


number. @ui lt @ui @uj @ui
The realizable k-e turbulence model have two equations for tur- C ¼
ui uj ¼ þ ð6Þ
@uj q @xj @xi @xj
bulence kinetic energy k and turbulence dissipation rate e:
606 S. Xie et al. / International Journal of Heat and Mass Transfer 122 (2018) 602–613

2
k In order to set the periodic boundary condition in the simulation,
lt ¼ q C l ð7Þ
e the numbers of nodes on the inlet and outlet surface must be equal.
Furthermore, no-slip velocity conditions are applied at all channel
 
lt surface.
C 1 ¼ max 0:43 ; C 2 ¼ 1:0; rk ¼ 1:0; re ¼ 1:2 ð8Þ
lt þ 5 Simulations are conducted using air as the working fluid, the
density of air is 1.128 kg/m3, specific heat is 1015 J/(kgK), thermal
conductivity is 0.02756 W/(kgK), viscosity is 1.91  10
5kg/(ms).
3.4. Data reduction Uniform mass flow rate and uniform temperature of 300 K are uti-
lized at the channel inlet. The specified mass flow rate of the fluid
The y+ is a dimensionless parameter which determines the size changes in such a way that the Re ranging from 5000 to 30,000. The
of the wall-adjacent cell. temperature of the wall assumed to be constant at 350 K, the prop-
uy erties of air are assumed to be constant and incomprehensible.
yþ ¼ ð9Þ
l=q
where y is the height of the wall to wall-adjacent cell and u⁄ is the 3.6. Grid details and independent test
friction velocity.
rffiffiffiffiffi Numerical calculation accuracy and computational time mainly
s0 depend on the rationality of the algorithm and grid, therefore, the
u ¼ ð10Þ
q grid system need repeated debugging to obtain results with
acceptable levels of accuracy and computational efficiency. The
where s0 is the shear stress of wall.
schematic of the grid system of the computational domain pre-
The mass flow rate is defined as:
sented in Fig. 6(b) and (c). At the near wall regions, the mesh are
pDh very dense, where the highest gradients of temperature and veloc-
m ¼ Re  l ð11Þ
4 ity are expected to occur. In general, the height of the wall-
where m is mass flow rate, Dh is the equivalent diameter, Re is the adjacent cell can be calculated by Eqs. (9) and (10), and the grids
Reynolds number. extend to a y+ value around or below 1.0. As such, enhanced wall
The Reynolds number is a non-dimensional parameter, which functions are employed in realizable k-e turbulence model for
stands for the ratio of inertial force to viscous force. The Reynolds the predictions.
number is expressed as: In order to test the grid independence, numerical solution is
carried out by utilized the grid systems with 1,781,126 cell,
qDh u 1,902,082 cell, 2,139,723 cell, and 2,461,381 cell for Case 1 at a
Re ¼ ð12Þ
l Reynold number of 30,000. The Nusselt number Nu and friction
The f (friction factor) described by Fanning friction factor is factor f from these predictions is listed in Table 2. The Nu and f
employed, which is defined as: gap between grid systems of 2.4 M and 2.1 M cells are much smal-
ler than the gap between 1.7 M and 1.9 M. Thus, to save computa-
2Dp Dh tion resources and balance computational economy and prediction
f ¼ ð13Þ
qu2i L accuracy, the grid system of 2.1 M cells is chosen for Case 1.
where Dp is the pressure drop for the test section and L is the length
of the test section. 3.7. Validation
The Nu (Nusselt number) indicates the intensity of convective
heat transfer. The Nu is defined as: The numerical method to simulate turbulent flows in ETDP
should be validated first to ensure the accuracy of present numer-
hDh ical results. Accordingly, the numerical results for ETDP are com-
Nu ¼ ð14Þ
k pared with experimental data of Ref. [10] and the experiment
The heat transfer coefficient h is defined as: conducted by Wang et al. Fig. 7(b) and (c) present comparison
between the numerical and experimental results at different Rey-
U
h¼ ð15Þ nolds numbers. It can be seen that the numerical results reason-
AðT wall
T ref Þ
ably agree well within the range of maximum uncertainties of
where U is total heat rate, A is the heat exchange area, Twall is the the Nu and f. The decent agreement with numerical and experi-
local temperature of wall, and Tref is the mass-weight average tem- mental results confirmed the reliability of the simulation method
perature of the fluid. adopted for this study.
A PEC (performance evaluation criteria) considering both the
heat transfer enhancement and friction loss penalty, which pro- 4. Study on flow field characteristics and heat transfer
posed by Gee and Webb. The PEC is defined as follows: characteristics
Nu=Nus
PEC ¼ 1=3
ð16Þ 4.1. Flow structures
ðf =f s Þ
where fs is the baseline friction factor, and Nus is the baseline Nus- Fig. 8(a) and (b) presents the velocity magnitude distribution in
selt number, fs and Nus are obtained from the simulations work X = 0 and Y = 0 plane for different Reynolds number, respectively.
with plain tube. The velocity presents with periodic increase and decrease in the
flow-channel for different Reynolds number, and the maximum
3.5. Boundary conditions and fluid properties velocity occurs at the centre of flow-channel with dimples/protru-
sions. This phenomenon is due to that as the fluid flows through
Fig. 6(a) shows fluid computational domain and boundary the flow-channel with protrusions/dimples, the flow-channel
conditions. In order to obtain the nearly fully-developed flow situ- shrink at the protrusion windward surface which make the veloc-
ation, the periodic boundary is applied in the present investigation. ity increases. After the fluid flow out of the centre of the protru-
 S. Xie et al. / International Journal of Heat and Mass Transfer 122 (2018) 602–613 607

Wall Period2(Outlet)
mm Period1(Inlet)
P=30

R=
4m
m

(a) (b) (c)

Fig. 6. Flow domain and grid system of ETDP: (a) geometrical structures of flow domain; (b) 3D view of the grid system; (c) right view of the grid system.

Table 2
Mesh independence test (Re = 30,000).

Grids Number of cells Nu Diff,% f Diff, %

Grid1 1,781,126 123.11 8.4 0.1839 5.7
Grid2 1,902,082 126.73 6.3 0.1865 4.2
Grid3 2,139,723 131.83 1.5 0.1919 1.3
Grid4 2,461,381 133.45 Baseline 0.1944 Baseline

0.06
80 Smooth tube Smooth tube
Experience,Wang et al.[10] Experience,Wang et al.[10]
Simulation Simulation
0.05
60
Nu

0.04
f

40

0.03
20

0.02
1 2 3 1 2 3
Re/(10 ) 4
Re/(104)
(a) (b)
Fig. 7. Validation of simulation results: (a) comparison of predicted Nusselt number with experimental results; (b) comparison of predicted friction factor with experimental
results.

sions/dimples, the flow-channel expansion at the protrusion lee-
ward surface which makes the velocity decreases. With the peri-
odic variation of velocity, the momentum exchange is improved,
the fluid flow mixing between the near wall zone and main stream
zone is enhanced.
Fig. 9(a) and (b) shows velocity magnitude distributions for dif-
ferent Z planes when Re = 15,000 and Re = 30,000. The relatively
high velocity is present at the protrusions adjacent region, the rel-
atively low velocity are present at the dimples adjacent area. The
maximum average velocity is found at Z = 0 mm plane compared
with Z = 5 and Z = 10 mm plane, this phenomenon is mainly due
to the cross-sectional area of Z = 0 mm is smaller than other planes.
Furthermore, the periodic variation of velocity also can seem from
these figures. The shape of protrusions and dimples caused
impingement effect and block effect, which induced the flow fluid
flushing on the tube wall and leaving from the tube wall. This phe-
nomenon results in interrupted the boundary layer, increased the
turbulent kinetic energy and turbulent intensity, hence improved
the heat transfer performance.
In order to clearly shows the induced recirculation flows by pro-
trusions/dimples, Fig. 10(a) presents the streamlines on the X = 0
plane for different Reynolds number. As expected, the recirculation

Flow direction Flow direction

Re=5000 Re=15000 Re=30000 Re=5000 Re=15000 Re=30000

Y
Z X
Z

(a) (b)
Fig. 8. Velocity distribution along the flow direction for Re = 5000–30,000, H = 4 mm, P = 30 mm, R = 4 mm: (a) velocity magnitude on X = 0 planes; (b) velocity magnitude on
Y = 0 planes.
608 S. Xie et al. / International Journal of Heat and Mass Transfer 122 (2018) 602–613
Protrusions
Z=0mm Z=5mm Z=10mm
Dimples
Y
X Z=0mm Z=5mm Z=10mm
Fig. 9. Velocity distribution for different Z planes at H = 4 mm, P = 30 mm, R = 4 mm: (a) Re = 15,000; (b) Re = 30,000.
flows are usually found at the downstream boundary of the protru-
sion and the adjacent area. However, when the fluid flows through
the dimples region, the recirculation flows are usually found at the
upstream boundary of the dimples and the adjacent area, which is
consistent with the observations from the study of Ref. [23]. Addi-
tionally, the strength and extent of the recirculation flow are dra-
matically decreased with increasing of the Reynolds number.
Fig. 10(b) shows the 3D streamlines on the flow domain. When
the fluid flow into the protrusions surface, the protrusions wind-
ward side blockage the flow-channel which makes the flow
impinged the protrusion windward surface, thus results in the
interrupt the boundary layer and enhance the heat transfer. As
the fluid flow into both sides of the protrusions, the protrusions
makes the fluid separated and flows around the protrusions, these
can improve the fluid flow mixing and turbulent intensity. When
the fluid flows into the dimples channel, the streamlines present
with swirl shape along the flow direction. The phenomenon is
interesting due to the dimples surface can induce the fluid flows,
and this is the mainly different from previous dimpled tubes.
Moreover, with the help of these swirl flows, the boundary layer
thickness is reduced and the flow mixing is improved, hence
improving the heat transfer performance.
4.2. Heat transfer characteristics
Temperature distributions on the X = 0 plane and Y = 0 plane
along the flow direction in different Reynolds number is shown
Re=5000
Y
Z
Protrusions
X
Y Dimples
Re=5000
Fig. 10. Streamlines for different for Re = 5000–30,000, D = 4 mm, P = 30 mm, R = 4 mm: (a) streamlines on X = 0 planes; (b) three-dimensional streamlines on flow domain.
Z=15mm Z=20mm Z=25mm Z=30mm
(a) Re=15,000
Z=15mm Z=20mm Z=25mm Z=30mm
(b) Re=30,000
in Fig. 11. From the figure, the average temperature of the fluid
increases with an increase of Reynolds number. In addition, the fig-
ures also show that the boundary layer is compressed and the tem-
perature gradient in the near wall zone is augmented,
consequently, the heat transfer rate can be dramatically enhanced
in compared to that of the plain tube.
Fig. 12 presents the temperature distribution at different Z
planes for Re = 15,000 and Re = 30,000. It can be observed that
due to the cold fluid flowing through the heated tube wall, the
average temperature of the fluid increased along the flow direc-
tion. Generally, it is can be seen that under the influence of dimples
and protrusions, the temperature fields are changed and distorted
in comparison with the well know the temperature field of concen-
tric circles in plain tube. With further observation, the temperature
gradient at protrusions and the adjacent area is larger than
dimples.
Fig. 13(a) shows the Nusselt number distribution on the tube
wall for different Reynolds number. There are three relatively high
Nusselt number area in tube wall and the largest Nusselt number
are associated with area one. For area one, the high Nusselt num-
ber area are usually found at the protrusion windward surface,
these consistent with observations from Ref. [11]. This phe-
nomenon is caused by the flow impinging and flushing on the pro-
trusion windward surface. For area two, the high Nusselt number
area is located at the protrusion leeward surface. This phenomenon
is mainly due to the flow reattachment at the protrusion leeward
surface. For area three, the high Nusselt distribution present with
Re=15000 Re=30000
(a)
Re=15000 (b) Re=30000
 S. Xie et al. / International Journal of Heat and Mass Transfer 122 (2018) 602–613 609

Flow direction
Flow direction
Re=5000 Re=15000 Re=30000
T/(K) Re=5000 Re=15000 Re=30000
X
Z
Z

(a) X=0 (b) Y=0

Fig. 11. Temperature distribution along the flow direction for Re = 5000–30,000: (a) temperature distribution on X = 0 plane; (b) temperature distribution on Y = 0 plane.

Protrusions
T/(K)

Z=0mm Z=5mm Z=10mm Z=15mm Z=20mm Z=25mm Z=30mm

(a) Re=15,000
Dimples

Y
X Z=0mm Z=5mm Z=10mm Z=15mm Z=20mm Z=25mm Z=30mm
(b) Re=30,000

Fig. 12. Temperature distribution at different Z planes for different Re: (a) temperature distribution when Re = 5000; (b) temperature distribution when Re = 30,000.

‘‘helically” shape along the dimples, this phenomenon is mainly
attributed to the helically dimples generated the swirl flows. How-
ever, there is only one relatively low Nusselt number area in tube
wall. The low Nusselt number area present with ‘‘Y” shape and
rightly corresponding to locations where the recirculation zone
(seen in Fig. 10) formed and the flow-channel expanded. The recir-
culation flow and expansion flow which results in decrease of the
velocity and turbulent energy, thus reduced the heat transfer
performance.
Fig. 13(b) presents the local Nusselt number on the path for dif-
ferent Reynolds number (the predefined path shown in Fig. 13(a)).
There are two peaks at the each paths. For the peak1, which caused
by the cold fluid directly impingement the protrusion. As to the
peak2, this is caused by the protrusion-induced advection flow
and the make the flows reattachments. In order to presents heat
transfer enhancement by the ETDP, the average of Nusselt number
for plain tube at different Reynolds number also shown in Fig. 13
(b). When Re = 30,000, the maximum and the average of the local
Nusselt number on path2 for ETDP has been increased to 5.2 and
3.6 times as compared with plain tube, respectively.
5. Effect of structural parameters on flow structure and heat
transfer performance
5.1. Case 1: Effects of depth
Fig. 14(a) presents the velocity distribution for three different
protrusion depth on the X = 0 planes and Y = 0 planes when Re =
30,000. With increasing of protrusion depth, the velocity magni-
tude increases. Protrusion volume is defined as the volume occu-
pied by the protrusions in the tube side. Obviously, the
protrusion volume is proportional to the protrusion depth. Based
on the conservation of the mass, lager protrusion volume generates
higher velocity, this result to the velocity increase with the increas-
ing protrusion depth. The pressure distributions are shown in
Fig. 14(b). The high pressure appeared on the protrusion windward
surface and dimple leeward surface, the maximum pressure
located at top of the protrusion. Moreover, the pressure also pre-
sent with periodic changes, these phenomenon can improving
the heat transfer performance. Fig. 14(c) shows the temperature
distribution on the X = 0 planes and Y = 0 planes for protrusion

Re=30000 Re=15000 Re=30000 300

Re=15000 Path1(Re=15000) Smooth tube(Re=15000)
Path2(Re=30000) Smooth tube(Re=30000)
250
Flow direction

200
2
Peak1

3 Peak2
Nu

150
1
100
Average value
Z Z
50
X Y
Path1 Path2 0 10 20 30
(a) Z/(mm)
(b)
Fig. 13. Nusselt number distribution for different Re: (a) Nusselt number distribution at wall when Re = 15,000 and Re = 30,000; (b) Nusselt number distribution on paths.
610 S. Xie et al. / International Journal of Heat and Mass Transfer 122 (2018) 602–613

Flow direction Flow direction

V/(m/s) D=2mm D=3mm D=5mm D=2mm D=3mm D=5mm

Y X
Z Z

(a)
p/(Pa) D=2mm D=3mm D=5mm D=2mm D=3mm D=5mm
Y X
Z Z

(b)
T/(K) D=2mm D=3mm D=5mm D=2mm D=3mm D=5mm

Y X
Z Z

(c)
Fig. 14. Flow field characteristics and temperature field characteristics for different depth: (a) velocity distribution; (b) pressure distribution; (c) temperature distribution.

depth. The temperature of fluid is increased with an increasing
depth, because that larger protrusion volume generates higher
flow mixing and boundary layers separation compared with smal-
ler protrusion volume.
Fig. 15(a) shows the averaged Nusselt number with varying
Reynolds number for different protrusion depth. These are normal-
ized using the plain tube Nusselt numbers at the same Reynolds
number, corresponding to fully developed turbulent flow. The
value for Nu0 of fully developed channel flow is obtained by the
Gnielinski correlation [24]. From the figure, the averaged Nu/Nu0
values decrease with an increase of the Reynolds number for differ-
ent protrusion depth. The largest normalized Nusselt number are
obtained by D = 5 mm and the smallest Nusselt number are associ-
ated with D = 2 mm with Reynolds number ranging from 5000 to
30,000. The protrusion can blockage of the flows-channel, for D =
5 mm, which generates the largest protrusion volume in the Case2.
Therefore, the larger protrusion generates higher turbulence level
and flow separation, which caused the D = 5 mm provided the lar-
gest normalized Nusselt number in Case 2. In addition, the heat
transfer increase rate becomes smaller when the Reynolds number
is larger than 20,000, and it reaches an asymptotic value of about
1.80 for D = 5 mm.
Fig. 15(b) gives the friction factor with varying Reynolds num-
ber for different protrusion depth. The value for f0 is obtained by
the Petukhov correlation [25]. The overall friction factor of all three
protrusion depth, compared to the fully developed plain tube val-
ues is 2.40–1.90 (D = 2 mm), 4.65–3.85 (D = 3 mm), 15.43–13.32
(D = 5 mm) with Reynolds number ranging from 5000 to 30,000.
When D = 2 mm and D = 3 mm, the friction factor decreases lin-
early with the variation of Reynolds number. However, when D =
5 mm, the friction factor is the great difference and first it
decreases and then increases slightly with the increase of Reynolds
number. These behaviour are mainly attributed to the larger pro-
trusion volume leads to serious pressure drop penalty.
It has been found that the heat transfer enhancement caused by
the dimples/protrusion is followed by a considerable pressure
drop. The performance evaluation criteria (PEC, seen in Eq. (16))
which are employed to evaluate the thermal-hydraulic perfor-
mance for ETDP. Fig. 13 presents the PEC for different protrusion
depth. From the figure, the PEC decrease with an increase of the
Reynolds number, these phenomena are mainly caused by the fric-
tion factor grows much faster than the Nusselt number with the
increases of Reynolds number. The highest PEC is obtained by D
= 3 mm when the Re = 5000. However, when Re > 5000, the highest

16 1.7
D =2mm D =2mm
3.5 D =3mm D =3mm
D =5mm 1.6
12 D =2mm D =5mm
3.0 D =3mm
D =5mm 1.5
Nu/Nu0

PEC
f/f0

8
2.5 1.4

2.0 4 1.3

1.5 1.2
5 10 15 20 25 30 5 10 15 20 25 30 5 10 15 20 25 30
Re/(103) Re/(103) Re/(103)
(a) (b) (c)
Fig. 15. Effect of different depths on thermal–hydraulic performance at different Reynolds number from 5000 to 30,000 for Case2: (a) Nusselt number versus Reynolds
number; (b) friction factor versus Reynolds number; (c) PEC versus Reynolds number.
 S. Xie et al. / International Journal of Heat and Mass Transfer 122 (2018) 602–613 611

Flow direction Flow direction

V/(m/s) P=25mm P=35mm P=40mm P=25mm P=35mm P=40mm
Y X
Z Z

(a)
T/(K) P=25mm P=35mm P=40mm P=25mm P=35mm P=40mm

Y X
Z Z

(b)
Fig. 16. Velocity and temperature distribution for different pitchs when Re = 30,000: (a) velocity distribution; (b) temperature distribution.

300
Flow direction Path3(P=25mm)
P=25mm P=35mm P=40mm Path3 Path4(P=35mm)
Path3 Path4 Path5 Path5(P=40mm)
250
Flow direction

Nu 200

150

Z
X 100
Nu
(a) Y 0 10 20 30 40
Z/(mm)
(b)
Fig. 17. Nusselt number distribution for different pitch at Re = 30,000: (a) average Nusselt number distribution at wall; (b) average Nusselt number distribution on paths.

3.5 10
P=25mm P=25mm
P=35mm
1.6
P=35mm
P=40mm 8 P=40mm
3.0
1.5
Nu/Nu0

PEC
f/f0

6 P=25mm
P=35mm
1.4
2.5 P=40mm
4
1.3
2.0 2
5 10 15 20 25 30 5 10 15 20 25 30 5 10 15 20 25 30
Re/(103) Re/(103) Re/(103)
(a) (b) (c)
Fig. 18. Effect of different pitch on thermal-hydraulic performance at different Reynolds number from 5000 to 30,000 for Case 3: (a) Nusselt number versus Reynolds
number; (b) friction factor versus Reynolds number; (c) PEC versus Reynolds number.

Flow direction
V/(m/s) T/(K)
R=3mm R=3.5mm R=5mm R=3mm R=3.5mm R=5mm
Y Y
Z Z

(a) (b)
Fig. 19. Velocity and temperature distribution for different radius when Re = 30,000: (a) velocity distribution; (b) temperature distribution.
612 S. Xie et al. / International Journal of Heat and Mass Transfer 122 (2018) 602–613

9 1.6
R=3mm R=3mm R=3mm
3.2 R=3.5mm R=3.5mm R=3.5mm
R=5mm R=5mm R=5mm
1.5
3.0
Nu/Nu 0

PEC
f/f0
1.4
2.8

7
2.6 1.3

5 10 15 20 25 30 5 10 15 20 25 30 5 10 15 20 25 30

Re/(103) Re/(103) Re/(103)

(a) (b) (c)
Fig. 20. Effect of different radius on thermal-hydraulic performance at different Reynolds number from 5000 to 30,000 for Case 4: (a) Nusselt number versus Reynolds
number; (b) friction factor versus Reynolds number; (c) PEC versus Reynolds number.

PEC is associated with the D = 2 mm and the smallest PEC obtained the velocity and temperature increase slightly with an increase of
by D = 5 mm for all the considered performance. protrusion radius.
Fig. 20(a) presents the averaged Nusselt number for different
5.2. Case2: Effects of pitch protrusion radius. The Nusselt number decreases with an increase
of Reynolds number. The highest and smallest Nusselt number is
Fig. 16(a) shows the velocity distribution for three protrusions obtained by R = 5 mm and R = 3 mm for all the considered perfor-
pitches. The smaller pitch indicates the protrusions occupied lager mance, respectively. Fig. 20(b) shows the friction factor for differ-
volume in per unit length of tube. Based on the conservation of the ent protrusion radius. The trend of friction factor first decreases
mass, lager protrusion volume generates higher velocity. As and then increases. The PEC is provided in Fig. 20(c). From the fig-
expected, the velocity decreases with an increase of the pitch. ure, the PEC decrease with an increase of protrusion radius, how-
Fig. 16(b) presents the temperature distribution for different ever, there is no significant difference in PEC for three different
pitches. The temperature also decreases with an increase of the protrusion radius.
pitch.
Fig. 17(a) presents the average Nusselt number distribution on
the tube wall for different pitch at Re = 30,000. For the protrusions
6. Summary and conclusions
region, the high Nusselt number area is located at the protrusion
windward surface. Whereas for the dimples region, the high Nus-
The complicated geometric model of ETDP constructed by the
selt number area appears on near the dimple centerline. Further-
plastic forming method. The local streamlines, velocity contour,
more, the high Nusselt number area and the maximum Nusselt
temperature contour and Nusselt number were carried out to
number value decrease with an increase of the pitch. Fig. 17(b)
describe heat transfer mechanisms and flow field characteristics.
shows the the local Nusselt number distribution at the path for dif-
The effects of protrusion depth, pitch and radius on thermal-
ferent Reynolds number (the predefined path shown in Fig. 16(a)).
hydraulic performance were also provided and analyzed. The main
From the figure, the local Nusselt number distribution present with
conclusions are drawn as follows.
‘‘U” shape on at the path for different pitchs, the high Nusselt num-
ber area and low high Nusselt number area which corresponding to
(1) The ETDP performs better heat transfer rate than the plain
the protrusion region and dimple region, respectively. In addition,
tube due to the dimpled and protruded surface improved
average of Nusselt number at paths for P = 25 mm, P = 35 mm and
the flow mixing, interrupted the boundary layer, formed
P = 40 mm has been increased to 3.29, 3.8, 3.6 times as compared
periodic jet flows and swirl flows. In comparison with plain
with plain tube, respectively.
tube, the ETDP hold completive potential promise in heat
Fig. 18(a) presents the averaged Nusselt number with varying
exchanger enhancement applications.
Reynolds number for different protrusion pitch. The Nu/Nu0 values
(2) The fluid velocity and pressure distribution periodic chang-
decrease with an increase of pitch or Reynolds number. In addition,
ing which enhanced the fluid disturbance. The maximum
the Nu/Nu0 values decrease rate becomes smaller when the Re >
Nusselt number area located on the protrusion windward
20,000, and it reaches an asymptotic value of about 2.78 for P =
surface. Whereas the minimum Nusselt number area associ-
25 mm, 2.52 for P = 35 mm, 2.15 for P = 40 mm, respectively.
ated with the dimples centreline.
Fig. 18(b) shows the friction factor for different protrusion pitch.
(3) The geometric parameter of protrusion depth and pitch have
From the figure, the friction factor first decreases and then
the significant effect on the heat and transfer characteristic,
increases, and all the curves have a turning point at Re = 15,000.
however, protrusion radius have relatively small effects on
Fig. 18(b) provides the PEC for different protrusion pitch. The
it. Among considered geometric parameters and operating
smallest PEC are obtained by P = 35 mm for all the considered per-
condition, the ETDP with D = 3 m, P = 30 mm, R = 4 mm and
formances. The highest PEC is obtained by P = 25 mm when the Re
Re = 5000 has the largest value of about 1.65 in PEC.
 10,000. However, when Re > 10,000, the highest PEC is associ-
ated with the D = 40 mm.

5.3. Case 4: Effects of radius Conflict of interest

Fig. 19(a) and (b) presents velocity and temperature distribu- We wish to confirm that there are no known conflicts of interest
tion for three different protrusion radius. It is clearly seen that associated with this publication.
 S. Xie et al. / International Journal of Heat and Mass Transfer 122 (2018) 602–613
Acknowledgements
This research work was supported by the Graduate Innovation
Foundation of School of Mechatronic Engineering of SWPU
(CX2014SY30).
References
[1] M. Sheikholeslami, M. Gorji-Bandpy, D.D. Ganji, Review of heat transfer
enhancement methods: focus on passive methods using swirl flow devices,
Renew. Sust. Energy Rev. 49 (2015) 444–469.
[2] G. Xie, J. Liu, W. Zhang, G. Lorenzini, C. Biserni, Numerical prediction of flow
structure and heat transfer in square channels with dimples combined with
secondary half-size dimples/protrusions, Numer. Heat Transfer Part A – Appl.
65 (2014) 327–356.
[3] A.I. Leontiev, N.A. Kiselev, S.A. Burtsev, M.M. Strongin, Y.A. Vinogradov,
Experimental investigation of heat transfer and drag on surfaces with
spherical dimples, Exp. Therm. Fluid Sci. 79 (2016) 74–84.
[4] L. Zheng, D. Zhang, Y. Xie, G. Xie, Thermal performance of dimpled/protruded
circular and annular microchannel tube heat sink, J. Taiwan Inst. Chem. Eng. 60
(2016) 342–351.
[5] Z.H. Ayub, A.H. Ayub, G. Ribatski, Two-phase pressure drop and flow boiling
heat transfer in an enhanced dimpled tube with a solid round rod insert, Int. J.
Refrig. 75 (2017) 1–13.
[6] J. Chen, H. Müller-Steinhagen, G.G. Duffy, Heat transfer enhancement in
dimpled tubes, Appl. Therm. Eng. 21 (2001) 535–547.
[7] P.G. Vicente, A. Garc, A. Viedma, Heat transfer and pressure drop for low
Reynolds turbulent flow in helically dimpled tubes, Int. J. Heat Mass Transfer
45 (2002) 543–553.
[8] C. Thianpong, P. Eiamsa-ard, K. Wongcharee, S. Eiamsa-ard, Compound heat
transfer enhancement of a dimpled tube with a twisted tape swirl generator,
Int. Commun. Heat Mass Transfer 36 (2009) 698–704.
[9] Y. Wang, Y. He, Y. Lei, R. Li, Heat transfer and friction characteristics for
turbulent flow of dimpled tubes, Chem. Eng. Technol. 32 (2009) 956–963.
[10] Y. Wang, Y. He, Y. Lei, J. Zhang, Heat transfer and hydrodynamics analysis of a
novel dimpled tube, Exp. Therm. Fluid Sci. 34 (2010) 1273–1281.
[11] Z. Liang, S. Xie, L. Zhang, J. Zhang, Y. Wang, Y. Yin, Influence of geometric
parameters on the thermal hydraulic performance of an ellipsoidal protruded
enhanced tube, Numer. Heat Transfer, Part A: Appl. 72 (2017) 153–170.
613
[12] S.W. Chang, K.F. Chiang, T.C. Chou, Heat transfer and pressure drop in
hexagonal ducts with surface dimples, Exp. Therm. Fluid Sci. 34 (2010) 1172–
1181.
[13] S. Heo, Y. Seo, T. Ku, B. Kang, Formability evaluation of dimple forming process
based on numerical and experimental approach, J. Mech. Sci. Technol. 25
(2011) 429–439.
[14] Y. Seo, H. Lee, S. Jeon, T. Ku, B. Kang, J. Kim, Homogenization of dimpled tube
and its application to structural integrity evaluation for a dimple-type EGR
cooler using FEM, Int. J. Precis. Eng. Manuf. 13 (2012) 183–191.
[15] A. Garcia, J.P. Solano, P.G. Vicente, A. Viedma, The influence of artificial
roughness shape on heat transfer enhancement: corrugated tubes, dimpled
tubes and wire coils, Appl. Therm. Eng. 35 (2012) 196–201.
[16] P. Kumar, A. Kumar, S. Chamoli, M. Kumar, Experimental investigation of heat
transfer enhancement and fluid flow characteristics in a protruded surface
heat exchanger tube, Exp. Therm. Fluid Sci. 71 (2016) 42–51.
[17] D.J. Kukulka, R. Smith, W. Li, Comparison of condensation and evaporation
heat transfer on the outside of smooth and enhanced 1EHT tubes, Appl. Therm.
Eng. 105 (2016) 913–922.
[18] M. Li, T.S. Khan, E. Al Hajri, Z.H. Ayub, Geometric optimization for thermal-
hydraulic performance of dimpled enhanced tubes for single phase flow, Appl.
Therm. Eng. 103 (2016) 639–650.
[19] M. Li, T.S. Khan, E. Al-Hajri, Z.H. Ayub, Single phase heat transfer and pressure
drop analysis of a dimpled enhanced tube, Appl. Therm. Eng. 101 (2016) 38–
46.
[20] M. Shafaee, H. Mashouf, A. Sarmadian, S.G. Mohseni, Evaporation heat transfer
and pressure drop characteristics of R-600a in horizontal smooth and helically
dimpled tubes, Appl. Therm. Eng. 107 (2016) 28–36.
[21] H. Mashouf, M. Shafaee, A. Sarmadian, S.G. Mohseni, Visual study of flow
patterns during evaporation and condensation of R-600a inside horizontal
smooth and helically dimpled tubes, Appl. Therm. Eng. (2017) 1392–1400.
[22] K. Aroonrat, S. Wongwises, Experimental study on two-phase condensation
heat transfer and pressure drop of R-134a flowing in a dimpled tube, Int. J.
Heat Mass Transfer 106 (2017) 437–448.
[23] J. Liu, Y. Song, G. Xie, B. Sunden, Numerical modeling flow and heat transfer in
dimpled cooling channels with secondary hemispherical protrusions, Energy
79 (2015) 1–19.
[24] V. Gnielinski, New equations for heat and mass-transfer in turbulent pipe and
channel flow, Int. Chem. Eng. 16 (1976) 359–368.
[25] B. Petukhov, Heat transfer and friction in turbulent pipe flow with variable
physical properties, Adv. Heat Transfer 6 (1970) i565.