Numerical investigation on forced

convection heat transfer over a rotating

circular cylinder inside a confined channel
Cite as: AIP Conference Proceedings 2324, 050033 (2021); https://doi.org/10.1063/5.0037673
Published Online: 25 February 2021

Md. Rakib Hossain, Sumon Saha, and Mohammad Mamun

ARTICLES YOU MAY BE INTERESTED IN

Non-Newtonian effect on double diffusive natural convection of nanofluid within a square

cavity
AIP Conference Proceedings 2324, 050030 (2021); https://doi.org/10.1063/5.0037581

Conjugate natural convection in three-dimensional differentially heated cubic partitioned

cavity filled with air and water
AIP Conference Proceedings 2324, 050026 (2021); https://doi.org/10.1063/5.0037525

Pure mixed convection inside a vented square cavity with an isothermally heated rotating
cylinder
AIP Conference Proceedings 2324, 050035 (2021); https://doi.org/10.1063/5.0037533

AIP Conference Proceedings 2324, 050033 (2021); https://doi.org/10.1063/5.0037673 2324, 050033

© 2021 Author(s).
 Numerical Investigation on Forced Convection Heat
Transfer over a Rotating Circular Cylinder inside a
Confined Channel
Md. Rakib Hossaina), Sumon Sahab) and Mohammad Mamunc)

Department of Mechanical Engineering, Bangladesh University of Engineering and Technology, Dhaka-1000,

Bangladesh
a)
Corresponding author: [email protected]
b)
[email protected], c)[email protected]

Abstract. The present study demonstrates a numerical investigation of forced convection heat transfer over a rotating hot
circular cylinder vertically placed at the central axis of a confined horizontal channel. Fully developed laminar flow of air
at ambient temperature is considered at the channel inlet, whereas the surface of the channel is kept insulated. The cylinder
is rotating either clockwise or counter-clockwise direction with a constant angular speed. The governing mass, momentum
and energy equations in non-dimensional form are solved numerically using Galerkin finite element method. Parametric
simulation is carried out over a range of mean flow Reynolds number based on the mean velocity of the working fluid
within the range of 40 ≤ Re ≤ 2000. The variation of blockage ratio is considered as β = 0.05, 0.1 and 0.2 respectively.
Moreover, the rotational Reynolds number of the cylinder based on the peripheral velocity of the cylinder is set at Rec = 4,
10, 20 and 30 respectively. The numerical results are visualized in terms of streamline plots and isosurface plots of thermal
field. The distributions of coefficient of drag, coefficient of lift and average Nusselt number of the hot cylinder surface as
a function of Reynolds number reveal a number of interesting observations from the present findings. It can be concluded
that the changes of Re and β significantly affect both on the aerodynamic and the thermal performance of the confined
channel.

INTRODUCTION

Laminar flow past a cylinder inside a channel has a wide range of applications in various branches like particle
migration inside the veins, combustion including droplet dynamics and vaporization, etc. The flow past a stationary
cylinder inside a channel with the cylinder placed symmetrically at the center of the channel was studied by Chen et
al. [1]. They focused on the nature and occurrence of the bifurcation from steady symmetric flow to the periodic
shedding regime. Kang et al. [2] numerically investigated two-dimensional laminar flow past a circular cylinder
rotating with a constant angular velocity and concluded that the mean lift was increased linearly, whereas the mean
drag was decreased with increasing circumferential speed of the cylinder. Using finite volume method, Stojkovic et
al. [3] solved two-dimensional laminar flow past a circular cylinder rotating with a constant angular velocity and
found that the drag force was decreased with increasing rotational velocity. Later, Mittal and Kumar [4] performed
similar numerical investigation and showed that the results from the stability analysis for the rotating cylinder were in
very good agreement with those from direct numerical simulations. They concluded that the flow became unstable for
speed ratio within 4.3 < α < 4.8. Considering the flow past a circular stationary cylinder in a plane channel,
Chakraborty et al. [5] found that the total drag coefficient was decreased with an increase in the blockage ratio as well
as Reynolds number. Similar numerical investigation was carried out by Sahin and Owens [6] and they identified the
critical Reynolds numbers for different blockage ratios. Besides, Rehimi et al. [7] experimentally showed the effect
of channel confinement of a circular cylinder centered between two parallel walls. Apart from fluid flow, Paramane
and Sharma [8] numerically investigated heat transfer across a rotating circular cylinder maintained at constant

Proceedings of the 13th International Conference on Mechanical Engineering (ICME2019)

AIP Conf. Proc. 2324, 050033-1–050033-7; https://doi.org/10.1063/5.0037673
Published by AIP Publishing. 978-0-7354-4068-5/$30.00

050033-1
temperature and observed flow patterns, Nusselt numbers for different Reynolds numbers and blockage ratios. Later,
Prasad et al. [9] extended the similar works in order to analyze the combined effect of channel-confinement and
cylinder rotation on the flow and heat transfer across a cylinder for various blockage ratio, rotational velocity and
Reynolds number.
From the above literature survey, it can be concluded that few works have been done considering flow and heat
transfer over rotating circular cylinder inside a confined channel. Hence, the main objective of the present study is to
investigate the effect of rotational direction of the circular cylinder, the blockage ratio, the rotational speed of the
cylinder as well as the change of mean velocity of the working fluid on the flow and the thermal characteristics of the
confined channel. Special attention is given to keep the speed ratio (ratio of the peripheral velocity of the rotating
cylinder and the mean velocity of the working fluid) varies with Reynolds numbers, whereas all previous works dealt
with the constant speed ratio. Since constant value of speed ratio makes the rotational speed of the cylinder to be
changed automatically with the change of mean flow Reynolds number which is not always the practical scenario of
this type of problem.

MODELING OF THE PROBLEM

The schematic diagram of the present problem is shown in Fig. 1. The computational domain is presented in
Cartesian coordinate having its origin at the center of the cylinder. The rectangular channel has a width of L = 3D,
where D is the diameter of the channel. The circular cylinder of diameter d is centrally placed inside the channel at a
distance of Lu = 1.2D from the channel inlet. The cylinder is rotating at an angular speed of ω either clockwise (CW)
or counterclockwise (CCW) direction and the surface of the cylinder is kept at a high temperature Th. Three different
diameters of the cylinder are selected and hence, the blockage ratio (β) is varied as β = d/D = 0.05, 0.1 and 0.2
respectively. In the present study, no-slip boundary condition is imposed on the wall surface of the channel, whereas
a parabolic inflow velocity ui for laminar flow of the working fluid maintained at an ambient temperature Tc (< Th) is
considered at the channel inlet. Outflow boundary condition for both flow and thermal fields is maintained at the exit
of the channel.

FIGURE 1. Schematic diagram of the rectangular channel with a rotating circular cylinder in Cartesian coordinate system.

The working fluid in the channel is considered as air and the thermo-physical properties of air are taken to be
constant. The radiation and the buoyancy effects along with viscous dissipation are neglected in this study. The flow
inside the channel is assumed to be two-dimensional, steady, laminar, incompressible and hence, the mass, momentum
and energy conservation equations in dimensionless form are written as follows:

wU wV
 0, (1)
wX wY
wU wU wP 1 § w 2U w 2U ·
U V   ¨  ¸, (2)
wX wY wX Re © wX 2 wY 2 ¹
wV wV wP 1 § w 2V w 2V ·
U V   ¨  ¸, (3)
wX wY wY Re © wX 2 wY 2 ¹

050033-2
 w4 w4 1 § w24 w 24 ·
U V ¨  ¸. (4)
wX wY RePr © wX 2 wY 2 ¹
The following scales are used to get the above non-dimensional governing equations since it is convenient to solve
the above equations in non-dimensional form:
x y u v p T  Tc um D Q
X ,Y ,U ,V ,P ,4 , Re , Pr , (5)
D D um um U um2
Th  Tc Q D
where, x and y are the Cartesian coordinates, u and v are the velocity components in the x- and y-directions respectively,
um is the mean velocity of flow, p is the pressure, T is the temperature, whereas X and Y are the dimensionless Cartesian
coordinates, U and V are the dimensionless velocity components in the X- and Y- directions respectively, P is the
dimensionless pressure, Θ is the dimensionless temperature, ρ, ν and α are mass density, kinematic viscosity and
thermal diffusivity of air respectively, Re and Pr are mean flow Reynolds and Prandtl numbers respectively. The non-
dimensional boundary conditions considered in the present problem are listed in Table 1.
TABLE 1. Boundary conditions of the present model in non-dimensional form.
Boundary Wall(s) Thermal Field Flow Field
Inlet Θ= 0 U = 2(1–4Y2)
Outlet ∂Θ/∂X = 0 P=0
Channel surface ∂Θ/∂Y = 0 U=V=0
Rotating cylinder surface Θ= 1 U = ±2αY, V = 2αX

Here, α = up/um = Rec/(βRe) is called as the speed ratio, u p = ωd/2 is the peripheral velocity of the cylinder, Rec =
upd/ν is the rotational Reynolds number based on the peripheral velocity of the cylinder. The positive (+) sign of U
and the negative (–) sign of V is considered for CW rotation of the cylinder and vice versa for CCW rotation. In order
to evaluate the performance of the rotating object inside the channel, drag coefficient (CD), lift coefficient (CL) and
average Nusselt number (Nu) along the surface of the rotating cylinder are calculated using the following expressions:
FD FL 1 SE § w4 ·
SE ³0 ¨© wN ¸¹ S
CD , CL , Nu dS , (6)
1 2 1 2
U um A U um A
2 2
where, FD is the drag force, FL is the lift force, A is the projected area of the cylinder and S is the arc length of the
cylinder surface.

FIGURE 2. Finite element mesh considered for the present problem, inset: enlarged view of finer triangular mesh elements near
the cylinder surface.

NUMERICAL PROCEDURE
The governing equations (1)-(4) are solved numerically using Galerkin finite element method. The computational
domain is, at first, discretised by applying triangular mesh elements as shown in Fig. 2. In order to maintain reasonable
accuracy of the numerical results, a grid independence test has been accomplished and 77032 elements has been
selected for the current study since there is insignificant change in the average Nusselt number after further refinement
of mesh beyond 77032 elements. Due to lack of available experimental data to verify the present problem, previously
published numerical results have been taken into consideration to validate the mathematical model and the present
numerical scheme. The present numerical results are in good agreement with the published results of Prasad et al. [9]

050033-3
in terms of drag coefficient, lift coefficient and average Nusselt number which are compared and shown in Fig. 3. The
comparison of both results reveals that the present model and the numerical scheme are reasonably good enough to
carry out further simulation of the present problem.

5 -6.5 6
Present model
Prasad et al. [9]
-7.0
4 5

Nu
CD

-7.5

CL
3 4
Present model
-8.0
Prasad et al. [9]

2 -8.5 3
40 60 80 100 40 60 80 100
Re Re
(a) (b)
FIGURE 3. Comparison of the numerical results between the present model and Prasad et al. [9] in terms of (a) CD, CL and (b)
Nu as a function of Re for β = 0.3, α = 2 and CCW direction of rotation.

Re = 40 Re = 500 Re = 2000
CW
CCW

(a)
0 05
β=
β = 0.2

(b)
Rec = 4
Rec = 30

(c)
FIGURE 4. Effects of mean flow Reynolds number and (a) direction of cylinder rotation for Rec = 4 and β = 0.2, (b) blockage
ratio for Rec = 4 and CCW direction of cylinder rotation and (c) rotational Reynolds number of the cylinder for β = 0.05 and CW
direction of cylinder rotation on isosurface plots of temperature.

050033-4
 RESULT AND DISCUSSIONS

Parametric simulation is carried out in the range of 40 ≤ Re ≤ 2000, 4 ≤ Rec ≤ 30 and β = 0.05, 0.1, 0.2. Selected
numerical results of the present problem in terms of streamline plots and isosurface plots of temperature are taken into
consideration to visualize the flow and the thermal fields.
Figure 4 shows the variation of isosurface plots of temperature with the change of mean flow Reynolds number
for different combinations of cylinder rotational direction and speed, and blockage ratio. The top two rows of Fig. 4(a)
reveal the fact that the direction of rotation of the cylinder has little impact on thermal fields. However, with the
increase of mean flow Reynolds number, the shape of the isosurface plots changes sharply since the incoming cold air
blows over the rotating hot cylinder at a faster speed for higher Re. The thermal fields at the downstream of the channel
significantly depend on the blockage ratio and the rotational Reynolds number of the cylinder at low Re. This is clearly
evident in Figs. 4(b) and (c). However, at higher Re, the change of blockage ratio still contributes to affect the thermal
fields, whereas the change of Rec on the region of high temperature inside the channel appears to be insignificant.
Hence, only the blockage ratio of the cylinder plays dominant role on the thermo-fluid characteristics inside the
channel.

Re = 40 Re = 500 Re = 2000
CW
CCW

(a)
β = 0.05
β = 0.2

(b)
Rec = 4
Rec = 30

(c)
FIGURE 5. Effects of mean flow Reynolds number and (a) direction of cylinder rotation for Rec = 4 and β = 0.2, (b) blockage
ratio for Rec = 4 and CCW direction of cylinder rotation and (c) rotational Reynolds number of the cylinder for β = 0.05 and CW
direction of cylinder rotation on streamline plots.

In order to visualize the flow fields inside the channel, the streamline plots for various combinations of geometrical
and flow parameters are presented in Fig. 5. Figure 5(a) shows the variation of streamline plots with the change of
rotational directions (CW and CCW) of the cylinder and mean flow Reynolds number. It is observed that streamline
plots are significantly changed with the increase of incoming air velocity. In the mean free stream, for inviscid flow
over a rotating cylinder, rotation causes recirculation of fluid around the cylinder as also observed by Prasad et al. [9],
which, in turns, results in the formation of the eddies at the downstream of the cylinder. Due to low pressure region at

050033-5
the downstream, some detached eddies are formed. Moreover, at higher mean flow Reynolds number, near wall eddies
are also formed at both top and bottom surfaces of the channel. The fact that the rotational direction does not affect
the flow fields significantly can be visualized from Fig. 5(a). Figure 5(b) demonstrates the effect of cylinder size in
terms of blockage ratio on the streamline plots and it can be concluded that eddies are increased for the larger one as
flow restriction increases with the increasing size of the cylinder. Due to the enhanced effect of Rec, the streamlines
representing the flow fields are shifted further towards the wall of the channel which can be noticed from Fig. 5(c).
Here, it is moving towards the bottom surface of the channel since CW direction of cylinder rotation has been taken
into consideration.
20 15 150
E = 0.05,CCW
18
E = 0.1,CCW
10 125
16 E = 0.2,CCW
E = 0.05,CW
14 E = 0.1,CW 5 100
E = 0.2,CW
12

Nu
0
CL
CD

75
10

8 -5 50

6
-10 25
4

2 1 2 3
-15 1 2 3 0 1 2 3
10 10 10 10 10 10 10 10 10
Re Re Re
(a) (b) (c)
FIGURE 6. Variation of CD, CL and Nu as a function of Re for different β and both CW and CCW rotational direction of the
cylinder at Rec = 4.

20 150 150
Rec= 4 130
Rec= 10 125
15 Rec= 20 110
Rec= 30
100
90
Nu
CD

10 70
CL

75

50
50
5 30
25
10

0 1 2 3 -10 1 2 3 0 1 2 3
10 10 10 10 10 10 10 10 10
Re Re Re
(a) (b) (c)
FIGURE 7. Variation of CD, CL and Nu as a function of Re for different Rec at β = 0.05 and CW rotational direction of the
cylinder.

The distributions of CD, CL and Nu as a function of Re are displayed in Figs. 6(a) and (b) respectively. The
decreasing trend of the variation of CD and the increasing trend of Nu are observed with increasing mean flow Reynolds
number. For a fixed Rec, it is found that both CW and CCW directions of rotation do not alter the values of CD and
Nu. Moreover, the absolute values of CL remain the same, but with the opposite sign due to the change of the rotational
direction of the cylinder. On the other hand, the change of blockage ratio contributes significantly to the change of
both CD and CL at higher Re. Besides, the value of Nu decreases gradually with increasing β. From Fig. 6, it is obvious
that heat transfer is highest for the smallest blockage ratio, β = 0.05. For this reason, Fig. 7 is drawn to analyze the
effect of Rec on CD, CL and Nu at β = 0.05. It can be concluded that the values of CD, CL are almost the same, but a
little discrepancy is shown for higher Rec at low Re. Moreover, with the increase of Rec, CD decreases but CL increases.
However, the value of Nu remains exactly the same for different values Rec. This reveals that heat transfer does not
change with increasing Rec within the selected range.

050033-6
 CONCLUSIONS

The present numerical study analyzes the forced convection heat transfer inside a rectangular channel with a
rotating hot cylinder. Parametric simulation is carried out to observe the influences of Re, Rec, β and direction of
cylinder’s rotation on thermal and flow fields along with the distribution of CD, CL and Nu. From the present findings,
it can be concluded that the value of CD and the absolute value of CL decrease gradually with increasing Re and
similarly, the value of Nu increases with Re. However, the direction of cylinder’s rotation and the change of Rec have
no effect on Nu, even on CD and CL at higher Re. However, the increase of blockage ratio of the cylinder contributes
significantly to decrease CD, CL and Nu.

REFERENCES

1. J. H. Chen, W. G. Pritchard and S. J. Tavener, J. Fluid Mech. 284, 23–41 (1995).

2. S. Kang, H. Choi and S. Lee, Phys. Fluids 11(11), 3312–3321 (1999).
3. D. Stojkovic, M. Breuer and F. Durst, Phys. Fluids 14(9), 3160–3178 (2002).
4. S. Mittal and B. Kumar, J. Fluid Mech. 406, 303–334 (2003).
5. J. Chakraborty, N. Verma and R.P. Chhabra, Chem. Engg. Processing 43, 1529–1537 (2004).
6. M. Sahin and R.G. Owens, Phys. Fluids 16(5), 1305–1320 (2004).
7. F. Rehimi, F. Aloui, S. B. Nasrallah, L. Doubliez and J. Legrand, J. Fluid Struct. 24, 855–882 (2008).
8. S. B. Paramane and A. Sharma, Int. J. Heat Mass Tran. 52, 3205–3216 (2009).
9. K. Prasad, S. B. Paramane, A. Agrawal and A. Sharma, Numer. Heat Transfer, Part A 60, 699–726 (2011).

050033-7