Journal of Engineering Advancements Vol. 01(02) 2020, pp 43-52 https://doi.org/10.38032/jea.2020.02.

003

Near-wall and Turbulence Behavior of Swirl Flows through an Aerodynamic Nozzle

Md. Tanvir Khan*, Sharif M. Islam, and Zahir U. Ahmed
Department of Mechanical Engineering, Khulna University of Engineering & Technology, Khulna-9203, BANGLADESH
Received: June 04, 2020, Revised: June 21, 2020, Accepted: June 21, 2020, Available Online: June 28, 2020

ABSTRACT

It is often a challenge to achieve uniform flow in turbulent swirl flow and to predict the flow within the nozzle as measurement
diagnostics face difficulty to capture both mean flow and turbulence. The purpose of this study is to numerically investigate the near
wall flow characteristics and turbulent behavior for the effect of different tangential inlet numbers of an incompressible turbulent
swirl air jet. In this regard, axial-plus-tangential flow based swirling nozzle is considered for the simulation using finite volume
method, where turbulence is approximated by the Shear Stress Transport (SST) k-ω model. The results show that axial and tangential
velocity at the wall vicinity response the most. Moreover, the turbulent flow characteristic for no swirl flow is nearly uniform, but
for swirl flow it fluctuates abruptly near the inlet section where the swirl has introduced. The skin friction coefficient for 2TP is the
maximum for swirl flow and for no swirl condition the skin friction coefficient is nearly uniform. Due to the swirl introduction the
pressure drop characteristics near the nozzle center response quickly and near the wall vicinity this property changes slowly. The
magnitude of swirl decay fluctuates before the nozzle converging section however after the nozzle converging section the swirl
decay is nearly constant. The local swirl near the inlet is highly unpredictable although after the nozzle converging section the local
swirl profile is nearly similar for 2TP, 3TP and 4TP.
Keywords: Turbulent; Nozzle; Swirl; Aerodynamic; CFD; Pressure.

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.

1. Introduction Yajnik and Subbaiah [8] experimentally studied the effects

of swirl on internal turbulent flows by conducting experiments
Swirl flows are a classical form of fluid flow and typically
on turbulent pipe flow. They found that the mean velocity
encountered in many engineering applications, such as gas
profiles close to and away from the wall admit similarity
turbine combustor and cyclonic separator, as well as in nature,
representations at sufficiently large Reynolds numbers and
such as tornedo. In swirling flows generally two types of vortex
extended velocity defect law is sensitive to swirl as the wall law
emerge, namely solid body rotation and free vortex flow.
is not sensibly dependent on swirl. Chang and Dhir [9] found a
However, in actual practice, a mixed of these two typically
flow reversal region in axial velocity profile in the central portion
appears whereby azimuthal velocity increases with radius in the
of an acrylic tube and an axial velocity increase near the wall.
core and decreases afterwards. The academic and industrial
Kito and Kato [10] studied the near wall velocity distribution of
research on swirling jets was found to be challenging due to their
turbulent swirling flow in circular pipes and concluded that the
unique characteristics, such as flow reversal, vortex bubble and
flow becomes three-dimensional after transitional swirl intensity.
anisotropic turbulences in both free and impinging jets. The
Later, Kitoh [11] investigated the turbulent behavior of free-
detailed flow physics, instability and above features of swirling
vortex-type swirling flow through a long straight circular pipe
flows are explicitly available in the literature [1]. The axial and
and reported that the swirling component decays downstream as
radial pressure gradient in swirling flows affects downstream
a result of wall friction. Buschmann et al. [12] explored the wall
flow development and wall-bounded flows. Swirl flows are
behavior, the location of the peak Reynolds shear stresses and the
generated in different applications in many ways, such as rotating
three normal stresses of turbulent channel/pipe flows, and stated
pipe, rotating vanes and twisted tapes inside stationary pipe,
that no scaling works equally for all parameters. Ahmed et al.
axial-plus-tangential entry [2]-[7]. The wide variations of
[13] conducted measurements of mean velocity and turbulence
generations can be divided into two broad groups: geometrical
of swirling flows using dual wire CTA probe both in the core and
and aerodynamical. In geometrically generated swirl flows, for
near-wall regions. However, the measurements were confined to
example rotating vanes or twisted tapes inside pipes, the flow is
the immediately above the nozzle exit plane. Effect of viscosity
disturbed by such geometries or obstacles, which exacerbates the
and surface tension of fluids and associated instabilities in
generality of the flow and difficult to draw a summative
annular flow were also analyzed in some studies [14]-[15].
observation. On the other hand, aerodynamically generated swirl
flows, such as rotating pipe or axial-plus-tangential entry flow, Since both DNS and LES typically require sufficiently
have better control of the flow and common flow physics may smaller sized mesh or large mesh quantity to resolve small-scale
establish in different researches. Despite geometrically generated turbulence, Reynolds Averaged Navier-Stokes (RANS) were
swirl flows are widely available in the literature and still widely used in swirling flows by many investigators. Although it
progressing, fundamentals of aerodynamic swirl flows are is believed that among various RANS models the Reynolds
currently being investigated in detail because of its prevalence to Stress Model (RSM) would perform better because of its ability
use recently in number heat treatment applications. to capture anisotropy of turbulent shear stress, this hypothesis

*Corresponding Author Email Address: [email protected] Published by: SciEn Publishing Group
 M.T. Khan et al. /JEA Vol. 01(02) 2020, pp 43-52

was not found to be universal in the past [16]-[17]. In fact, some

studies revealed that two-equation models (based on Boussinesq
approximation) can be applied with acceptable accuracy in
comparison with experimental data for moderate swirl flows -
[18]-[19]. Nouri-Borujerdi and Kebriaee [20] simulated the
turbulent boundary layer of an incompressible viscous swirling
flow through a conical chamber using finite volume method and
reported that the boundary layer thickness is dependent on the
velocity ratios, Reynolds number and nozzle angle. Najafi et al.
[21] investigated turbulent swirling decay in a vertical straight
fixed pipe where swirl in induced by rotating honeycombs. They
showed RSM with two-layer zone model for different near wall
approaches are fairly well to predict the swirling flow but fail to
predict the pressure distribution along the pipe wall. Islam et al.
simulated an aerodynamically generated swirl flow using SST k-
ω model in the nozzle exit plane. The result showed centerline
velocity [22] decay for introducing low levels of swirl into the
impinging jet and a significant reduction in turbulent kinetic
energy at the wall region. More recently, researchers [23]-[27]
performed simulations of aerodynamic swirl flow using SST k-
ω model and validated their results with the literature.
The above discussion revealed that although a significant Fig. 1 (a) CAD view of the aerodynamic swirl nozzle (sliced
amount of research available on swirling flows, they are mostly to show the internal cavity), and (b): different number of
confined to gas turbine combustors, cyclonic separator, or tangential ports to impart azimuthal component.
geometry induced swirl flows. Relevant aerodynamically
swirling flows are either classical in nature (experimental) with 2.2 Governing Equation and Boundary Conditions
outdated experimental facility to resolve near-wall and The governing equations to solve the incompressible,
turbulence behaviors inside the nozzle or tested robustness of steady-state flow characteristics within the aerodynamic swirl
various numerical schemes against experimental data with nozzle are the conservation of mass and momentum as follows:
limited focus on flow behaviors. Thus, a detailed study for the
effect of geometric parameters on near-wall and turbulence   Vˆ  0 (1)

 
behaviors for an aerodynamic swirl nozzle appears to be limited
in the literature. As such, the current numerical study will bridge  Vˆ  Vˆ  Pˆ  2Vˆ
this gap by investigating non-swirling and swirling flows from (2)
an aerodynamic nozzle for the same initial and flow conditions.
ˆ
The paper will examine the effect of number of tangential ports Here, V is the velocity vector, and 𝜌 and 𝜇 are the density
on mean and turbulence flow development along the length of and dynamic viscosity of the fluid, respectively. Since the
the nozzle. current problem is turbulent in nature, RANS approach is
applied to solve the mean and turbulence quantities. In RANS
2. Methodology approximation, each variable is composed of time-averaged
part (steady) and turbulence part, as shown below:
2.1 Problem formulation
An aerodynamically generated swirl nozzle, which is
ˆ      (3)
where  is
capable of seamless transition from non-swirling to highly
a variable used in equations (1) and (2). Upon
swirling jets, is considered in this study. The nozzle is axial-
implementing the equation (3) into the governing equations and
plus-tangential entry type and consists of three tangential
using equation (1) and setting time-average of turbulence
around the nozzle periphery and an axial port at the bottom of
equals to zero, the resulting RANS equations emerge. The
the nozzle. Detailed dimension of this nozzle is available in
RANS equations are similar to the governing equations, except
[4],[28], hence is not repeated here for brevity. The uniform ̅̅̅̅̅̅̅
′ ′
an additional term (𝜌𝑢 𝑖 𝑢𝑗 ) in equation (2). This additional term
flow was ensured by the flow settling chamber with
honeycomb. The aerodynamic swirl flow was generated when is known as turbulent shear stress, which governs the
both the uniform axial and tangential flows from different turbulence characteristics. The shear stress components are
number of circumferentially oriented and inclined ports mix determined via mean velocity gradients by the Boussinesq
together farther downstream. In this case, three variants of hypothesis:
tangential ports: 2-Tangential ports (2TP), 3-Tangential ports
(3TP) and 4-Tangential ports (4TP) are considered. The exit of 2  u u j 
uiuj  kij  t  i  
the nozzle has diameter (D) 40 mm, with a total length of 577
3  x j xi 
mm. A three-dimensional view with the relative orientation of   (4)
tangential ports is shown in Fig. 1.

44
 M.T. Khan et al. /JEA Vol. 01(02) 2020, pp 43-52

Where, μt is the turbulent viscosity and is a function of k and ω,

which are determined via an appropriate turbulence model. In
this research, the SST k-ω model, proposed by Menter and
Egorov [29], is chosen to model turbulence transport quantities:
turbulence kinetic energy, k, and the specific dissipation rate,
ω. SST k-ω model was found to be one of the best performing
RANS models in swirling and wall-bounded flows [16],[30]-
[31].
The boundary conditions considered in this problem are
mass flow inlet at the axial and tangential ports, pressure outlet
at the nozzle exit plane and wall at the nozzle wall. The inlet
conditions are adopted from the study [4] for realistic
predictions. For no-swirl flow, axial flows (only) are provided
from the bottom of the nozzle, with no flows from tangential Fig. 2 Sensitivity of mesh elements for the case 3TP.
ports. Conversely, for the swirling flows, three tangential ports
supply the same amount of mass flow, with no axial flow from
the bottom of the nozzle. This ensures the same Reynolds
number, where the average velocity in Reynolds number is
determined by volume flow rate divided by nozzle exit area.
The pressure outlet condition is applied at the nozzle exit with
atmospheric pressure whereby turbulence is specified by 2%
intensity and hydraulic diameter. Finally, no-slip condition at
the nozzle wall with ambient temperature is used.
2.3 Numerical Settings and Model Validation
The above governing equations (1-4) and two transport
equations (SST k-ω) are solved by finite volume based
commercial software ANSYS Fluent v17. The pressure-based
coupled algorithm is used to simultaneously solve the
governing equations. For the pressure discretization PRESTO
(PREssure STaggering Option) is applied, and the second-order
(a)
central-differenced for diffusion terms and second-order
upwind scheme for convective terms of the transport equations
are used. Hexahedral mapped mesh type was used in multi-
zone meshing with an element size 0.0025m. A typical mesh
independence test is shown in Fig. 2 for 3TP. In this case, four
different grid elements are tested, namely, 974001, 1310326,
1647089 and 1892870 elements. The results are found to be
invariant, except the 974001 elements. As such, a mesh
containing 1647k element is chosen in this study. Another
testing is also done for first layer height on y+ values and wall
shear stress profiles, and is shown in Fig. 3. It appears the first
layer cell height has little influence on those wall
characteristics. As such, a first layer height of 0.05 mm is
chosen in this study. The figure also shows that above settings
ensured y+ values less than 0.2 for the whole domain. The
solution is assumed to be converged when the residuals of the (b)
flow parameters are less than 10-5. Mass conservation is also
checked between inlet ports and the nozzle exit for converged Fig. 3 Effect of first layer cell height on y+ and normalized
solutions and the difference between inlets and outlet is found wall shear stress.
close to zero. Cartesian coordinates are given by (x, y, z) with
sum of flowrates through all axial and tangential inlets.
corresponding velocity components (Ux, Uy, Uz). For
Mathematically,
convenience, results of this 3D simulation are presented by
polar coordinates (x, r, θ) with x-axis coinciding for each Qt Q
Qr   t
system. In this regard, the corresponding axial (U) and Qa  Qt QT (5)
tangential (W) velocity components are derived from the
Cartesian coordinate data using the axes transformation rule. The Reynolds number is defined as,

2.4 Parameter Definition Q D

Re  T
The flow ratio 𝑄𝑟 is the ratio of total mass flow rate A (6)
through tangential inlets to the total flow rates in the nozzle, i.e.
45
 M.T. Khan et al. /JEA Vol. 01(02) 2020, pp 43-52

Where, A is the nozzle area at the exit plane of the nozzle

and ν is the dynamic viscosity of fluid. Two different swirl
number definitions are common as the ratio of axial flux of
tangential momentum to the axial flux of axial momentum:

R
2  r 2UWdr
S 0
R

2 R   r U 2  0.5W 2 dr
0  (7)

R 2
 r
S  0
UWdr
R
R  rU 2 dr
0 (8)
Finally, the local swirl number is defined as, (b)

W
S * (r ) 
U (9)
2.5 Validation
The simulation data of the swirl nozzle is first tested with
different turbulence models and validated by comparing with
experimental data [4] at the nozzle outlet plane with three
tangential ports for the case 𝑄𝑟 = 0 and 1 (Fig. 4). It appears
that SST k-ω model predicts the flow behavior well for both
non-swirling and swirling cases. As such, SST k-ω model will
be used for the data presented in the ensuing results and
discussion section. Fig. 4 also shows a good agreement
between the numerical prediction and the experimental data for
𝑄𝑟 = 0. A slight deviation is observed for 𝑄𝑟 = 1, but
importantly, the numerical data predicts well the profile (c)
behaviors and peak locations. The deviation is attributed to the
possible measurement inaccuracy associated with CTA X-wire Fig. 4 Different turbulence models against experimental data
and experimental flow settings [13]. This results an [4] for (a) U/U_b for Q_r= 0, (b) U/U_b for Q_r= 1 and (c)
overestimation of the mean velocity components than their W/U_b for Q_r= 1.
corresponding true values. The deviations may also be partly
attributed to the inability of RANS approach to accurately 3. Results and Discussion
capture highly swirling flows. This section includes presentation of mean and turbulence
controlling parameters, such as velocity, boundary layer
thickness, pressure drop, wall shear stress and turbulent shear
stress for the effect of different number of tangential ports at
two flow conditions (Qr = 0 and Qr = 1).
Swirl decay along the length is shown in Fig. 5 for the 3TP
case. The swirl number is calculated using the Equation (7) and
Equation (8). It is observed that near the nozzle inlet the swirl
number 𝑆 is highly fluctuating and after the converging section,
the swirl number is nearly constant. Initially, the swirl number
decreases from the inlet. When the tangential port has
introduced the value of swirl number increases and
immediately after the increment the swirl number drops again.
In the nozzle converging section the magnitude of the swirl
number rises again at first and then the value decreases and
finally increases before coming at a constant magnitude. In case
of 𝑆 ′ the values initially increasing from the nozzle inlet come
(a) to a constant magnitude after the nozzle converging section. It
is evident that the swirl number 𝑆 shows a very unpredictable
nature before the nozzle converging section.

46
 M.T. Khan et al. /JEA Vol. 01(02) 2020, pp 43-52

higher near the inlet and towards the nozzle exit, the velocity
magnitude decreases. The axial velocity magnitude is
proportional to the number of tangential ports and the velocity
magnitude is zero at the wall due to no-slip condition. The axial
velocity vectors at 𝑥/𝐷 = 4.75 is symmetric from the
centerline when swirl is induced (Fig. 7b-d), however, the
velocity distribution becomes asymmetrical as the flow
approaches the outlet. The velocity distribution is more
symmetrical along the radial direction for 2TP than 3TP & 4TP
with a slight decrease near the center except for exit plane. At
𝑥/𝐷 = 14.425, velocity reduction near the center is highest for
2TP.

Fig. 5 Swirl decay along the length of the nozzle for 3TP case.

Fig. 6 shows the local swirl numbers along the radial

location of the nozzle at the different axial positions (𝑥/𝐷 =
1.275, 3.1, 5.65, 9.4, and 14.425) of the nozzle for three
different numbers of tangential ports, namely, 2TP, 3TP, and
4TP. It is observed that the profile of local swirl at 𝑥/𝐷 =
5.65, 9.4, and 14.425 are almost the same in nature for all the
cases. The local swirl increases gradually from the center of the
nozzle toward the wall and near the wall, a sudden drop occurs.
However, the local swirl profiles at 𝑥/𝐷 = 3.1 are similar for
2TP and 3TP although for 4TP the local swirl profile is
parabolic from the nozzle center toward the nozzle wall. The
value of local swirl near the inlet at 𝑥/𝐷 = 1.275 is highly
unpredictable due to the swirl induction. The local swirl profile
for 2TP and 3TP are nearly similar but for 4TP the profile is
different. The local swirl near the nozzle wall suddenly
increases and drops immediately after the increase and finally
raise to the wall for both 2TP and 3TP. However, the local swirl
profile at 4TP is like a bell shape; increasing from the nozzle
center it becomes constant at 𝑟/𝐷 = 0.2 to 𝑟/𝐷 = 0.35 and
then decreases again towards the nozzle wall. As the swirl is
introduced at this location the local swirl profile becomes
highly unpredictable.
The axial velocity vectors at five different axial locations
(xD = 4.75, 7.50, 9.75, 12.5 and 14.425) are presented in Fig.
7 for non-swirling (Fig. 7a) and swirling (Fig. 7b-d)
conditions. Velocity is found to be the almost uniform after the
converging section for no swirl flow (𝑄𝑟 = 0), but the
magnitude is the highest near the exit plane, with a reduction
Fig. 6 Local swirl charactersitics along the length of the
towards the wall due to boundary layer formation. In contrast,
nozzle (a) 2TP, (b) 3TP, (c) 4TP.
for swirl flow (𝑄𝑟 = 1), velocity magnitude is found to be the

47
 M.T. Khan et al. /JEA Vol. 01(02) 2020, pp 43-52

Fig. 7 Axial velocity vectors at different axial locations for non-swirling (a), and swirling flows: (b) 2TP, (c) 3TP and (d) 4TP.

48
 M.T. Khan et al. /JEA Vol. 01(02) 2020, pp 43-52

Fig. 8 depicts the boundary layer thicknesses along the observed. In every case of swirl flow the Skin friction
length of the nozzle for both non-swirlign and swirling (2TP, coefficient at the nozzle exit is the maximum. Although the
3TP and 4TP) conditions. It is observed that the magnitude of Skin friction coefficient along the axial location is not
boundary layer thickness at 2TP is pretty low compare to the significantly changing however, near the exit it suddenly
other. For swirl flow conditions the boundary layer thickness at increase. The overall distribution of the Skin friction
the beginning is small and near the exit the boundary layer coefficient is not similar for all the case.
thickness increases. At the middle of the nozzle the boundary
Fig. 11 displays the pressure drop at different radial
layer thickness is nearly constant. However, for no swirl flow
locations for swirling flow (3TP). The inlet pressure is taken at
condition the boundary layer thickness higher and then the
the reference pressure and pressure drop is calculated based on
boundary layer thickness gradually decreases towards nozzle
the pressure of the nozzle inlet. It is observed that the pressure
exit.

Fig. 8 Boundary layer thickness along the length of the

nozzle.
Fig. 9 demonstrates the axial (U/Ub) and tangential (W/Ub)
velocity components near the wall i.e. at 𝑦/𝐷 = 0.01, 0.05
and 0.1 for 𝑄𝑟 = 0 and 𝑄𝑟 = 1. In the figure, 𝑌 represents the
distance from the wall of the swirl nozzle. It is observed that at
for swirl condition the near-wall axial velocity profile is
identical for all the location and it increases from the inlet to
the nozzle converging section. Then, the velocity suddenly
decreases at 𝑌 = 0.01𝐷 immediately after the converging
section. The axial velocity is then nearly constant for all the
axial location of the nozzle. However, the axial velocity
decreases very little after the nozzle convergence at 𝑌 = 0.05𝐷
and 0.1𝐷, and no significant change occurs. For swirl flow
condition the near-wall axial and tangential velocity increases
after the swirl flow introduction. Then after the nozzle
converging section the near-wall axial and tangential velocity
show a very little deviation and along the nozzle axial location
the velocity profile is nearly uniform at 𝑌 = 0.05𝐷 and 𝑌 =
0.1𝐷. However, the near-wall axial and tangential velocity
profile at 𝑌 = 0.01𝐷 is fluctuates highly along the length of the
nozzle, especially the tangential velocity profile shows a totally Fig. 9 Near-wall normalized axial (U/Ub) and tangential
unpredictable nature. The overall near-wall axial and tangential (W/Ub) velocity distrution along the nozzle length: Qr= 0 (a),
velocity distribution at 𝑌 = 0.01𝐷 is low in magnitude than at and Qr= 1 (b and c).
𝑌 = 0.05𝐷 and 𝑌 = 0.1𝐷. It can be observed that the velocity
profile at the wall vicinity (𝑌 = 0.01𝐷) response highly than drop at the center of the nozzle ( 𝑟⁄𝐷 = 0 ) is greater than all
the other position. other radial location in almost every axial position of the
nozzle, however, near the nozzle outlet it suddenly reduces.
The interplay between the number of tangential ports and While pressure drop at every radial location increase after the
the swirl intensity for the effect of skin friction coefficient introduction of the tangential ports, at ( 𝑟⁄𝐷 = 0.4 ) it suddenly
along the length of the nozzle is shown in Fig. 10. It is observed decrease. Pressure drop near the wall is gradually increasing
that for no swirl condition the magnitude of Skin friction after the converging section of nozzle but pressure drop near
coefficient is very low and along the nozzle axial location no the nozzle center is nearly constant. It can be concluded that
change occurs. However, the Skin friction coefficient at high Due to the tangential flow introduction the pressure drop
swirl flow at 2TP is the maximum along the nozzle axial characteristics near the nozzle center response quickly and near
location although for 3TP and 4TP no significant deviation is the wall vicinity this property change slowly.
49
 M.T. Khan et al. /JEA Vol. 01(02) 2020, pp 43-52

Fig. 10 Skin friction coefficient along the nozzle length. Fig. 11 Pressure drop along the nozzle length for swirling
flow.

Fig. 12 Turbulent normal stress profiles at different axial locations.

Fig. 12 presents the radial distribution of normalized at 𝑥 ⁄𝐷 = 1.275 and 𝑥 ⁄𝐷 = 4.45 the overall turbulent flow
turbulent normal stress components at various axial locations. characteristic is very small in magnitude. On the contrary at
It is observed that the turbulent flow characteristic near the wall swirl flow condition (𝑄𝑟 = 1) near the inlet of the nozzle at
is very high at 𝑥 ⁄𝐷 = 9.4 and 𝑥 ⁄𝐷 = 14.425 at no swirl flow 𝑥 ⁄𝐷 = 1.275 the overall turbulent flow characteristic is very
(𝑄𝑟 = 0) although at the center of the nozzle the turbulent flow high. However, the turbulent normal stress near the outlet
characteristic is very low. Moreover, near the inlet of the nozzle section at 𝑥 ⁄𝐷 = 9.4 and 𝑥 ⁄𝐷 = 14.425 is small in these

50
 M.T. Khan et al. /JEA Vol. 01(02) 2020, pp 43-52

locations. While at no swirl condition the turbulent normal introduction of swirl flow. The overall magnitude of the
stress characteristic at the center of nozzle is very low at swirl turbulent normal stress is very small in no swirl flow condition
flow condition, the 𝑢′ 𝑢′ and 𝑤 ′ 𝑤 ′ component is very high at but for swirl condition it is very high.
the nozzle center especially at 𝑥 ⁄𝐷 = 1.275 due to the

Fig. 13 Turbulent shear stress profiles at different axial locations.

Finally, Fig. 13 presents the radial distribution of very high. However, the overall turbulent share stress at 𝑥 ⁄𝐷 =
normalized turbulent shear stress components at various axial 9.4 and 𝑥 ⁄𝐷 = 14.425 is very uniform along the radial
locations. It is observed that for no swirl flow (𝑄𝑟 = 0) the location for swirl flow. The overall magnitude of the turbulent
turbulent shear stress component 𝑢′ 𝑣 ′ and 𝑣 ′ 𝑤 ′ near the wall is flow characteristic is very small in for swirl flow compared to
very high at 𝑥 ⁄𝐷 = 9.4 and 𝑥 ⁄𝐷 = 14.425 although at the the swirl condition.
center of the nozzle the turbulent share stress component is very
low (Fig. 13a and Fig. 13c). However, the 𝑢′ 𝑤 ′ component at 4. Conclusion
𝑥 ⁄𝐷 = 4.45, 9.4 and 14.425 is reducing near the wall of the An incompressible turbulent swirling air jet is investigated
nozzle. Near the inlet of the nozzle at 𝑥 ⁄𝐷 = 1.275 the overall numerically. In this regard, an axial-plus-tangential swirl flow
turbulent share stress is uniform and does not deviate much for is considered for non-swirling ( Q r = 0) and highly swirling
no swirl flow (𝑄𝑟 = 0). On the other hand at swirl flow ( Q r = 1) cases. Finite volume based commercial software
condition (𝑄𝑟 = 1) near the inlet of the nozzle at 𝑥 ⁄𝐷 = 1.275 ANSYS Fluent v17 is used in the simulation to investigate
all the three component of turbulent share stress is fluctuates mean flow and turbulence characteristics. Governing equations
are approximated by the RANS equations and turbulence is
highly and the magnitude of overall turbulent share stress is
51
 M.T. Khan et al. /JEA Vol. 01(02) 2020, pp 43-52

characterized by the SST k–ω model. The study examines the [14] Ahmed, Z.U., Khayat, R.E., Maissa, P. and Mathis, C., 2012.
effect of number of tangential inlets on mean flow behaviors Axisymmetric annular curtain stability. Fluid Dynamics
and turbulent characteristics. The magnitude of swirl decay is Research, 44(3), p.031401:1-23.
fluctuates highly before the nozzle converging section and after [15] Ahmed, Z.U., Khayat, R.E., Maissa, P. and Mathis, C., 2013.
the nozzle converging section the swirl decay is nearly Non-axisymmetric annular curtain stability. Physics of
Fluids, 25(8), p.082104:1-37.
constant. The boundary layer thickness in swirl flow is found
to be the smallest for 2TP. The skin friction coefficient along [16] Lu, P. and Semião, V., 2003. A new second‐moment closure
approach for turbulent swirling confined flows. International
the nozzle axial position at no swirl flow is uniform and the journal for numerical methods in fluids, 41(2), pp.133-150.
magnitude is very small. For swirl flow, the skin friction [17] Tsai, J.H., Lin, C.A. and Lu, C.M., 1995. Modelling dump
coefficient along the nozzle axial position is fluctuating combustor flows with and without swirl at the inlet using
especially for 2TP. Moreover, the skin friction coefficient at the Reynolds stress models. International Journal of Numerical
nozzle exit is the maximum for all the case. Due to the Methods for Heat & Fluid Flow, vol. 5, no. 7, pp. 577–588.
tangential flow introduction, the pressure drop near the nozzle [18] Gorman, J.M., Sparrow, E.M., Abraham, J.P. and Minkowycz,
center response quickly and near the wall vicinity this property W.J., 2016. Evaluation of the efficacy of turbulence models for
changes slowly. The turbulent normal and shear stress for no swirling flows and the effect of turbulence intensity on heat
swirl flow is nearly uniform, but for swirl flow it fluctuates the transfer. Numerical Heat Transfer, Part B: Fundamentals, 70(6),
most near the section where the swirl flow is introduced. pp.485-502.
[19] Saqr, K.M. and Wahid, M.A., 2014. Effects of swirl intensity on
Acknowledgements heat transfer and entropy generation in turbulent decaying swirl
flow. Applied thermal engineering, 70(1), pp.486-493.
The first author acknowledges the Department of [20] Nouri-Borujerdi, A. and Kebriaee, A., 2012. Simulation of
Electronics and Communication Engineering at Khulna turbulent swirling flow in convergent nozzles. Scientia
University of Engineering & Technology, Bangladesh for part Iranica, 19(2), pp.258-265.
of the computation facility provided during the study program. [21] Najafi, A.F., Saidi, M.H., Sadeghipour, M.S. and Souhar, M.,
2005. Numerical analysis of turbulent swirling decay pipe
References flow. International communications in heat and mass
transfer, 32(5), pp.627-638.
[1] Sloan, D.G., Smith, P.J. and Smoot, L.D., 1986. Modeling of [22] Islam, Md. M., Tasnim, S., and Ahmed, Z. U., 2017. Numerical
swirl in turbulent flow systems. Progress in Energy and study of a swirl nozzle at moderate swirl number, International
Combustion Science, 12(3), pp.163-250. Conference on Mechanical Engineering and Renewable Energy,
[2] Rose, W.G., 1962. A swirling round turbulent jet, Journal of 20-22 December, Chittagong, Bangladesh.
Applied Mechanics, vol.29, Trans. ASME, vol. 84, Series E, pp. [23] Islam, S. M, Khan, M. T., Ahmed, Z. U., 2020. Effect of design
616-625. parameters on flow characteristics of an aerodynamic swirl
[3] Toh, K., Honnery, D. and Soria, J., 2010. Axial plus tangential nozzle, Progress in Computational Fluid Dynamics, In press.
entry swirling jet. Experiments in Fluids, 48(2), pp.309-325. [24] Khan, M. H. U., and Ahmed, Z. U., 2019. Fluid flow and heat
transfer characteristics of multiple swirling impinging jets at
[4] Ahmed, Z.U., 2016. An experimental and numerical study of various impingement distances, International Journal of
surface interactions in turbulent swirling jets. PhD Thesis, Edith Thermofluid Science and Technology, vol. 6, no. 4, pp.
Cowan University, Australia. 19060403:1-12.
[5] Gore, R.W. and Ranz, W.E., 1964. Backflows in rotating fluids [25] Ahmed, Z. U., Al-Abdeli, Y. M., 2017. Flow characteristics due
moving axially through expanding cross sections. AIChE to jet impact at low intensity, In: Proceedings of Int. Conf.
Journal, 10(1), pp.83-88. Engineering, Research, Innovation and Education, Sylhet,
Bangladesh, 156, p. 1-6.
[6] Jafari, M., Farhadi, M. and Sedighi, K., 2017. An experimental [26] Debnath, S., Khan, M.H.U. and Ahmed, Z.U., 2020. Turbulent
study on the effects of a new swirl generator on thermal Swirling Impinging Jet Arrays: A Numerical Study on Fluid Flow
performance of a circular tube. International Communications in and Heat Transfer. Thermal Science and Engineering Progress,
Heat and Mass Transfer, 87, pp.277-287. p.100580.
[7] Markal, B., 2018. Experimental investigation of heat transfer [27] Ahmed, Z.U., Khan, M.H.U., Khayat, R.E. and Tasnim, S., 2018,
characteristics and wall pressure distribution of swirling coaxial July. Effect of flow confinement on the hydrodynamics and heat
confined impinging air jets. International Journal of Heat and transfer characteristics of swirling impinging jets. In AIP
Mass Transfer, 124, pp.517-532. Conference Proceedings (Vol. 1980, No. 1, p. 040008). AIP
[8] Yajnik, K.S. and Subbaiah, M.V., 1973. Experiments on swirling Publishing LLC.
turbulent flows. Part 1. Similarity in swirling flows. Journal of [28] Thomas, B.K., Ahmed, Z.U., Al-Abdeli, Y.M. and Matthews,
Fluid Mechanics, 60(4), pp.665-687. M.T., 2013. The optimisation of a turbulent swirl nozzle using
[9] Chang, F. and Dhir, V.K., 1994. Turbulent flow field in CFD. Proceedings of the Australian Combustion Symposium,
tangentially injected swirl flows in tubes. International journal of November 6-8, Perth, Australia.
heat and fluid flow, 15(5), pp.346-356. [29] Menter, F.R. and Egorov, Y., 2010. The scale-adaptive
[10] KITO, O. and KATO, T., 1984. Near wall velocity distribution of simulation method for unsteady turbulent flow predictions. Part
turbulent swirling flow in circular pipe. Bulletin of 1: theory and model description. Flow, Turbulence and
JSME, 27(230), pp.1659-1666. Combustion, 85(1), pp.113-138.
[11] Kitoh, O., 1991. Experimental study of turbulent swirling flow in [30] Ahmed, Z.U., Al-Abdeli, Y.M. and Matthews, M.T., 2015. The
a straight pipe. Journal of Fluid Mechanics, 225, pp.445-479. effect of inflow conditions on the development of non-swirling
[12] Buschmann, M.H., Indinger, T. and Gad-el-Hak, M., 2009. Near- versus swirling impinging turbulent jets. Computers &
wall behavior of turbulent wall-bounded flows. International Fluids, 118, pp.255-273.
Journal of Heat and Fluid Flow, 30(5), pp.993-1006. [31] Ahmed, Z.U., Al-Abdeli, Y.M. and Guzzomi, F.G., 2017. Flow
[13] Ahmed, Z.U., Al-Abdeli, Y.M. and Guzzomi, F.G., 2016. field and thermal behaviour in swirling and non-swirling
Corrections of dual-wire CTA data in turbulent swirling and non- turbulent impinging jets. International Journal of Thermal
swirling jets. Experimental Thermal and Fluid Science, 70, Sciences, 114, pp.241-256.
pp.166-175.

52