International Journal of Heat and Fluid Flow 105 (2024) 109272

Contents lists available at ScienceDirect

International Journal of Heat and Fluid Flow

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

Influence of an upstream attached flat plate on the Newtonian flow past a

square cylinder
Hamid Rahman a, *, Waqas Sarwar Abbasi b, Saba Ismail b, Sumaira Nadeem b
a
Department of Mathematics & Statistics, Women University Swabi, Swabi 23430, Pakistan
b
Department of Mathematics, Air University, Islamabad 44000, Pakistan

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

Keywords: In this study, the lattice Boltzmann method is used to analyze the flow around a single square cylinder with a flat
Cylinder plate attached at front side of the cylinder. In order to perceive the impact of the attached flat plate on flow
Flow Control characteristics of the square cylinder, the dimensionless plate lengths are varied from 0.1 to 6, while the Rey­
Flat Plate
nolds number is fixed at 150. Corresponding to all plate lengths the well-known von Karman vortex street is
Suppression
Vorticity
observed in the cylinder’s wake region. However, the increase in plate’s length results in reducing recirculation
regions, enlarging the wake length and thus stabilizing the flow. The fluid forces and velocity field are signifi­
cantly affected with increment in plate length. A maximum of 65%, 77%, 62% and 11% reduction is observed in
the mean drag coefficient (CDm), root-mean-square values of drag coefficient (CDrms) and lift coefficient (CLrms)
and amplitude of lift coefficient (CLAmp), respectively, as compared to single cylinder without a flat plate.
Pressure is also found to be significantly reduced due to presence of plate at front surface of cylinder.

1. Introduction coefficient (CD) on small Re and also identified the maximum value of
Strouhal number (St) at Re ≈ 150. Okajima (Okajima, 1982) experi­
Numerous experimental and numerical investigations have been mentally determined the vortex shedding frequencies of rectangular
performed for flow behind a single cylinder. Generally, when fluid in­ cylinders over range 70 ≤ Re ≤ 2 × 104. He observed that at low Re, flow
teracts with a single object the vortices shed from sides of the object, separation occurred at trailing edge due to the immediate reattachment
alternately. This formation of alternate vortices in the wake region of the of flow at leading edge and as Re increased, the flow fully separated from
cylinder is known as Von Karman vortex street. Resonance, structural the leading edge without reattachment to side surface. Norberg (Nor­
vibration, acoustic noise may occur from alternate shedding of vortices berg, 1993) conducted an experimental study in a wind tunnel for flow
in the near wake. The structural vibrations emerge due to strong fluc­ past a rectangular cylinder at 5 × 103 ≤ Re ≤ 1.3 × 104 for different
tuating pressure forces in the transverse direction of the flow and may angle of attack (θ) in the range from 0 to 90. He found that at θ = 0,
cause sometime the failure of structure (Billah and Scanlan, 1991; multiple wake frequencies appeared. He also observed that with incre­
Yichang Lin et al., 2022). This phenomenon is termed as the vortex ment in the angle of attack, St increased, and CD and lift coefficient (CL)
induced vibrations and mainly happens due to the resemblance of the decreased. Samani and Bergstrom (Samani and Bergstrom, 2015)
shedding frequency of vortices with natural frequency of structures investigated the influence of a wall on the flow characteristics of an
(Singh and Mittal, 2005). infinite square cross-section cylinder for three different gap ratios (g = 0,
In fluid dynamics analysis field, a prototype of flows around civil 0.5 and 1) at Re = 500. It was observed that time-average values of the
structures is considered as flow past a single cylinder. Some experi­ CD increased and CL decreased when distance of cylinder from wall
mental and numerical studies regarding the flow past a single cylinder increased. At g = 0.5 the vortex shedding frequency decreased and the
are discussed here. Breuer et al. (Breuer et al., 2000) numerically secondary separation zone was more prominent. Porteous et al. (Por­
simulated 2D laminar flow past a square cylinder for Reynolds number teous et al., 2019) studied the influence of the incoming wall boundary
(Re) range 0.5 ≤ Re ≤ 300 using the lattice Boltzmann method (LBM) layer thickness on the noise production from a square finite wall-
and finite volume method (FVM). They reported the dependence of drag mounted cylinder. They showed that the height of the boundary layer

* Corresponding author.
E-mail addresses: hamidrahman@wus.edu.pk, rhamidmath@gmail.com (H. Rahman).

https://doi.org/10.1016/j.ijheatfluidflow.2023.109272
Received 3 October 2023; Received in revised form 15 December 2023; Accepted 26 December 2023
Available online 31 December 2023
0142-727X/© 2023 Elsevier Inc. All rights reserved.
H. Rahman et al. International Journal of Heat and Fluid Flow 105 (2024) 109272

delayed the production of sound. Manzoor et al. (Manzoor et al., 2013) experimentally investigated the flow mechanism of a circular cylinder
conducted experimental and numerical simulations based on the dam­ with a flexible flat plate. Their research indicated that the peak value of
ped free-end conditions to study the lock-in peak amplitudes of a square St declined with each increase in the flat plate’s length. Turki (Turki,
cylinder. Sohankar et al. (Sohankar et al., 2015) numerically analyzed 2008) conducted a 2D numerical investigation using FEM to examine the
the effects of uniform blowing and suction to find optimum condition for vortex shedding behind a square cylinder with a flat plate for Re =
vortex shedding suppression around a square cylinder for Re = 70–150 110–200. The study revealed that for Re < 150, the Strouhal number
and Pr = 0.7. They examined the effect of suction or blowing on front, decreased when the length of the flat plate was increased. However, for
back, and top/bottom surfaces individually. On top and bottom sides, St Re ≥ 150, the St increased with increasing plate lengths. It was also
and CD were decreased while in rear and front sides, St and CD were observed that the presence of the flat plate had a negligible effect on CD.
increased. Barman and Bhattacharyya (Barman and Bhattacharyya, 2015) attached
In order to reduce the fluid forces and subsequently regularize the thin flat plates to a square cylinder in order to control the vortex shed­
shed vortices, different controlling techniques have been investigated in ding and reduce the drag. They concluded that the presence of upstream
past. Active and passive techniques are the well-known controlling flat plate reduced drag but had negligible impact on lift while the
techniques, that can be used to reduce the aerodynamic drag on a bluff downstream flat plate dampened the vortex shedding frequency.
body and regulate the shed vortices. Active control methods involve Based on the preceding discussion, it can be concluded that previous
introducing external energy to the flow through techniques like oscil­ studies have primarily focused on flow control by placing control de­
latory motions of the bluff body or jet blowing. These methods often vices in the near-wake regions of a cylinder. However, limited research
require complex mechanical devices and additional external power. On has been conducted on flow suppression from front face of a square
the other hand, passive control techniques utilize modifications to the cylinder by attaching a flat plate at upstream position. So in order to
shape of the bluff body or the addition of devices in the flow stream to address this shortcoming, the current study is conducted for develop­
suppress or weaken vortex shedding. Passive control methods are ad­ ment of a more effective and practical flow control mechanism using a
vantageous as they do not require additional energy and are relatively flat plate attached at the upstream side of a square cylinder. In this work,
simpler to implement. Among passive control methods, the flat plate has we will conduct numerical simulations to investigate the impact of the
been recognized as a highly effective tool for controlling vortex shed­ flat plate on vortex shedding and the resulting flow-induced forces. By
ding. Numerous studies in the open literature have focused on flow placing the plate in the path of the incoming fluid prior to its interaction
control using various control devices. In the following discussion, with the cylinder, our focus will extend beyond the analysis of forces and
selected papers addressing this topic will be highlighted. flow characteristics. We aim to determine the optimal conditions for
Zhou et al. (Zhou et al., 2005) controlled the vortex shedding behind achieving maximum flow control, expressed in terms of percentage
a square cylinder by placing a control plate at upstream position at Re = differences. This research will provide valuable insights into enhancing
250 for different heights of control plate. They reported the reduction in flow control strategies using the flat plate configuration and contribute
drag and lift of the cylinder in presence of an upstream control plate. to the advancement of practical flow control mechanisms.
Chen and Shao (Chen and Shao, 2013) used small elements of circular, Paper organization: Section 2 contains details of numerical
square, triangular and thin-strip cross sections in order to suppress the method, in Section 3 the geometry of problem is explained, Section 4 is
vortex shedding from a cylinder at Re = 70 to 130. They determined that about discussion of results and Section 5 contains conclusions of current
square elements had wider effective zones than triangle elements by study.
comparison at Re = 110. Moreover, as Re increased (Re > 130), the zone
of the square element shrank. Saha and Shrivastava (Saha and Shriv­ 2. Numerical technique
astava, 2015) numerically investigated the behavior of vortex shedding
for flow past a square cylinder at Re = 100 by using direct numerical The lattice Boltzmann method is used for simulations of the proposed
simulation. It was observed that CD was increased with increasing problem. Among many salient features of this method, the second order
blowing velocity and the twin bubble disappeared at a blowing velocity accuracy (Mohammad, 2011), getting rid of Poisson equation solution
of 0.87 and flow became steady. Taherian et al. (Taherian et al., 2017) for pressure calculation (Wolf Gladrow, 2005) and quasi non-linearity
used the PIV technique in order to visualize flow patterns across a thick are a few features to be mentioned. In this section, the LBM is briefly
blunt trailing edge airfoil with base cavity at low Re. It was observed that introduced and for more details the readers are referred to our published
increasing Re increased the wake length and reduced the wake behind papers (Islam et al., 2014a,b). The LBM models the fluid flow by
the trailing edge, resulting reduction in CD. Han et al. (Han et al., 2019) following the evolution of fluid particles, where the physical space is
investigated the effect of a triangular wedge on fluid flow characteristics discretized into a number of square regular lattices and at each time
of a circular cylinder placed at upstream position for low Re. It was step, the particles move and collide according to certain rules instead of
observed that between the wedge and the main cylinder, two completely solving the typical continuum equations for the conservation laws to be
separated flow patterns occurred. They also observed that at g = 2.8, the obeyed in fluid fields. In the current investigation, the Bhatnagar-Gross-
mean drag coefficient (CDm) was decreased up to 60.1 %, and the fluc­ Krook (BGK) collision model (Bhatnagar et al., 1954) in the traditional
tuating lift was reduced by 71.9 %. Igbalajobi et al. (Igbalajobi et al., Boltzmann equation is employed together with a two-dimensional nine-
2013) conducted experiments to study the flow characteristics around a velocity particles (D2Q9, where D and Q are the dimensions and number
surface-mounted finite-height circular cylinder with a wake-mounted of particles, respectively) model (Guo et al., 2000). The evolution
flat plate at Re = 7.4 × 104. They found 12 % reduction in CDm for equation of the density distribution function for the fluid particles can be
case of finite cylinder. Zhu and Wang (Zhu and Wang, 2019) numerically described as:
examined the flow past a circular cylinder with C-shaped plate at Re =
1
100. They studied vortex induced vibration suppression mechanism gi (x + ei Δt, t + Δt) − gi (x, t) = − (gi (x, t) − geq
i (x, t) ) (1)
τ
under the influence of g = 1.5, 3, 4.5 and 6, induced velocity (Vr = 1–12),
θ and Re. It was observed in their study that by positioning the C-shaped where gi and gieq denote the distribution and equilibrium distribution
plate behind cylinder at g = 1.5, best suppression can be obtained. functions, respectively, and τ is relaxation time which for stability must
Soumya and Prakash (Soumya and Prakash, 2017) numerically inves­ be greater than 0.5. The equilibrium distribution function gieq is given by
tigated the fluid flow characteristics using a flat plate on a triangular
the relation:
cylinder. They reported suppression of vortex shedding mechanism at
Re = 50 to 150 and found 9 % to 57 % reduction in CD of the cylinder
with attached flat plate. Teksin and Yayla (Teksin and Yayla, 2016)

2
H. Rahman et al. International Journal of Heat and Fluid Flow 105 (2024) 109272

[ ]
3 2 9 lattice distribution is also recommended by Islam et al. (Islam et al.,
geq
i = ρωi 1 + 3(ei •u)− u + (ei • u)2 (2)
2 4 2017) and Guo et al. (Guo et al., 2008) to ensure grid independence. The
height of the computational domain is denoted as H, while the length is
where ωi are the weighting coefficients. These weighting coefficients, denoted as l. The length of the flat plate, denoted as L, varies within the
fulfilling the lattice isotropy conditions (Wolf Gladrow, 2005), have range of 0.1D to 6D. Sohankar et al. (Sohankar et al., 1998) recom­
different values in different lattice models used in LBM. For the present mended that in order to effectively reduce the influence of the outlet
study the weighting coefficients are: boundary on flow and wake characteristics, the downstream length (LD)
⎧ should be greater than 15 times the side length of the square cylinder.
⎪ 4
⎪
⎪
⎪ i=0 Keeping in view this, in the current investigation LD (the distance from
⎪ 9
⎪
⎪
⎨ back surface of cylinder to outlet of channel) is taken as 20D while LU
1
ωi = i = 1, 2, 3, 4 (3) (the distance from inlet to front surface of cylinder) is set at 10D. The
⎪ 9
⎪
⎪
⎪ height of the computational domain in fixed at 21D. The length l of the
⎪
⎪
⎪ 1
⎩ i = 5, 6, 7, 8 channel in this study is adjusted at 31D. Note that here D is considered to
36 be the characteristic length and all other lengths are made dimensionless
using D.
The macroscopic quantities like density and velocity are computed by
Parabolic velocity is incorporated at inlet while at the upper and
the following relations:
lower walls of the channel and cylinder surface no-slip boundary con­
∑
8
1∑
8 dition is applied using the bounce back rule that means the fluid parti­
ρ= gi u = ei gi (4) cles hitting with the solid walls bounce back in the opposite directions
ρ
(Mohammad, 2011). When investigating flows with unsteady wake or
i=0 i=1

The artificial speed of sound and pressure are given as: vortices, the convective boundary condition is very important that is
often used at the outlet boundary of the channel to ensure that the fluid
e
cs = √̅̅̅, p = ρcs 2 wheree =
Δx
(5) particles can move easily through the outlet without causing significant
3 Δt changes in the internal flow domain. Similar boundary conditions were
also suggested by Islam et al. (Islam et al., 2018). Fluid forces like drag
The kinematic viscosity is associated with relaxation time parameter as: (FD) and lift (FL) are calculated using the momentum exchange method
( )
1 2 (Chapman and Cowling, 1970).
ν = τ− c Δt (6) The following important parameters are related to the current
2 s
problem:
Although, the lattice Boltzmann method is becoming increasingly pop­ Reynold’s Number:
ular among researchers for simulating complex fluid flow systems. Uin D
However, this approach has certain limitations. For instance, simulating Re =
ν
high-Mach number flows in aerodynamics using LBM is challenging due
to compressibility constraints. Another drawback is the absence of a where Uin is the inflow velocity, D is the side length of the cylinder and ν
thermo-hydrodynamic scheme in LBM. Additionally, LBM requires a is kinematic viscosity.
significant amount of memory for highly refined grids. In some cases, Strouhal Number:
handling boundary conditions can also be difficult because LBM ex­
fs D
presses them in terms of the distribution function rather than macro­ St =
Uin
scopic quantities.
where fs is the shedding frequency of vortices.
3. Statement of the problem and boundary conditions
Lift Coefficient:

Fig. 1 presents the geometry of the problem under consideration. A CL =

2FL
square cylinder having side length D (also termed as the size of cylinder), ρUin2 D
with a flat plate attached at front surface, is placed inside a channel.
Drag Coefficient:
Each surface of the cylinder is divided into 20 lattices (D = 20) for all
calculations in order to ensure grid independence for better results. This

Fig. 1. Geometry of flow past a square cylinder with attached flat plate.

3
H. Rahman et al. International Journal of Heat and Fluid Flow 105 (2024) 109272

2FD the square cylinder vary under the influence of increasing length of plate
CD =
ρUin2 D in the range from 0.1 to 6 at Re = 150. Fig. 2(a-c) presents the vorticity-
streamline contours, illustrating the flow structure mechanism for some
where FL is the lift force, FD is the drag force, ρ is fluid density. selected cases of plate lengths. It is to be noted that the blue color in­
Root Mean Square of Lift Coefficient: dicates the vortices which rotate in clockwise direction while the red
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
∑ color indicates the anticlockwise rotating vortices. All these cases in
(CL − CLm )2 Fig. . 2 (from a to c) as well as those in Fig. 3 are presented at the similar
CLrms =
N iteration step which is the 5 x 105th time step. The color bar indicating
the intensity of vortices in the domain shows negative values for
Root Mean Square of Drag Coefficient:
clockwise rotating vortices and positive ones for anticlockwise rotating
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
∑ vortices. From vorticity variations, it can be observed that the flow is
(CD − CDm )2
CDrms = unsteady in the down wake region of cylinder for all chosen plate
N
lengths with formation of alternate vortices from either side of cylinder.
The incoming fluid first interacts with the front surface of plate resulting
where CDm and CLm represent the average of CD and CL while N is the
in generation of shear layers. These shear layers after detaching from
total number of time steps.
plate corners, engulf the cylinder completely and rollup to form vortices
In order to validate our code, we have computed the flow charac­
in the wake of cylinder. At smaller plate lengths the shear layers
teristics behind a single square cylinder and compared our results with
immediately roll up to form vortices in the region adjacent to back
published data available in literature. For the sake of bravity we have
surface of cylinder. With gradual increment in length of plate, the shear
not given the details here and readers are referred to our similar pub­
layers engulf the cylinder instead of attaching to its surface and the roll
lished results in (Abbasi et al., 2021; Abbasi et al., 2020; Islam et al.,
up point of shear layers, in the wake region, moves gradually away from
2016).
cylinder and thus enlarging the steadiness duration in flow. This phe­
nomenon was also reported by Barman and Bhattacharyya (Barman and
4. Results and discussions
Bhattacharyya, 2015) for flow behind a square cylinder in presence of an
upstream flat plate. According to them, the upstream attached plate has
This section presents a numerical analysis of the impacts on the flow
little impact on the vortex shedding from square cylinder for plate
characteristics at Re = 150 of a flat plate attached with a square cylinder
lengths beyond 1.5 times the size of cylinder.
at the front side. The main aim of this study is to control the unsteady
The streamlines variations show that the inner flow separates and
flow mode by gradually varying the plate lengths and to find out a
reattach with either the upper or the lower surface of the cylinder
suitable length which suppresses the vortex shedding phenomena and
forming a recirculation zone at upper back corner as well as in the wake
reduce the fluid forces on the cylinder.
region of cylinder at L = 0.1 (Fig. 2(a)). Such recirculating region is also
The current simulations reveal that the fluid flow characteristics of

Fig. 2. Vorticity and Streamlines graphs for unsteady flow corresponding to different plate lengths.

4
H. Rahman et al. International Journal of Heat and Fluid Flow 105 (2024) 109272

Fig. 3. Pressure contours for unsteady flow corresponding to different plate lengths.

termed as corner or tip vortex. Further, it can also be noted that as the cylinder due to increasing plate lengths and maximum reduction can be
plate length varies the formation mechanism of recirculating zone observed at L = 6 among all cases considered in current study. Due to
changes, e.g. at L = 0.1 and 6 this zone can be tilted as downwards while reduction in pressure the drag force also reduces which will be discussed
at L = 3 it is in upward direction (Fig. 2(a-c)). The recirculation zone latter. Barman and Bhattacharyya (Barman and Bhattacharyya, 2015)
actually indicates the merging region of shear layers which results in observed that the pressure distribution on the front face of the square
vortex shedding formation in the down wake region. The shape and size cylinder in presence of attached flat plate changes due to the variation of
of vortices alters as plate length changes. In particular, with the incre­ plate length. According to their observations, the pressure at front sur­
ment in the length of plate, the separated shear layers reattach with face of cylinder is positive and have low values at corners of cylinder.
cylinder and rollup in quasi-steady manner. This reattachment and Hwang and Yang (Hwang and Yang, 2007), concluded in their work that
rollup are completely unsteady at smaller plate lengths as evidenced at L the pressure substantially reduces in the vicinity of the front stagnation
= 0.1. In this particular case, the plate does not completely suppress point of the circular cylinder by placing a flat plate in the upstream
vortex shedding behind the square cylinder, as previously reported in region of cylinder.
studies investigating downstream attached plates (Barman and Bhatta­ In Figs. 4(a-c) and 5(a-c), the CD and CL graphs are presented for
charyya, 2015). plate lengths L = 0.1 to 6. The graphs of CD and CL show periodicity due
Fig. 3(a–c) shows the pressure variations around and in the wake of to vortex formation seen in vorticity graphs. The magnitude of both
cylinder corresponding to different plate lengths at Re = 150. It should forces decreases as the length of the plate increases. It can be seen that
be noted that the variations in fluid forces like CD and CL are directly the CD show periodic variation with cycles of larger amplitude followed
related with the pressure distribution. At L = 0.1, the pressure is positive by smaller amplitude cycles (Fig. 4(a-c)). At smaller plate length, un­
and have higher values at front side of cylinder and lower values at steadiness is more severe, with higher amplitude cycles, as compared to
corners due to detaching shear layers (Fig. 3(a)). It can be seen that the the larger plate lengths (see zoomed view of CD). As the length of the
distribution of pressure is not uniform in the computational domain, and plate increases, the amplitude as well as magnitude of CD decreases. The
it varies randomly in different portions of computational domain. The CL starts with minor oscillations initially which can be attributed to
stagnation point can be seen near the midpoint of cylinder at L = 0.1. parabolic incoming flow. After some time, the periodic behavior can be
With increment in plate length the stagnation point changes its position evidenced for all lengths of plate (Fig. 5(a-c)). The periodicity in CL can
from near front surface of cylinder to front surface of plate away from be attributed to vortex shedding behind the cylinder (Fig. 2). Also, as
cylinder which serves as a shield against pressure (Fig. 3(b, c)). It can plate length increases, the amplitude of CL decreases which is due to
also be seen that the pressure falls to negative values in the wake of controlling effect of plate. Zhou et al. (Zhou et al., 2005) also observed

5
H. Rahman et al. International Journal of Heat and Fluid Flow 105 (2024) 109272

Fig. 4. Variation of CD corresponding to different plate lengths.

Fig. 5. Variation of CL corresponding to different plate lengths.

the reduction in drag and fluctuating lift for flow around a square cyl­ position of channel that are x1 = 1 and x5 = 621. Also, three lines are
inder in the presence of an upstream flat plate. drawn at upstream, center, and downstream sides of main cylinder that
The power spectrum of CL is presented in Fig. 6(a-c) at L = 0.1, 3 and are equally spaced apart, which is x2 = 201, x3 = 211 and x4 = 221.
6. A single dominant peak with no secondary peak appears at L = 0.1 These graphs show that the stream wise velocity significantly varies at
which ensures the alternate movement of vortices generated from either different locations of computational domain. The wake region velocity
side of cylinder in the computational domain (Fig. 6(a)). A secondary profiles do not exhibit any systematic pattern. At entrance and exit a
peak overlapping the y-axis can also be observed in Fig. 6b and 6c. These perfect parabolic velocity can be observed which indicates the fully
peaks correspond to the secondary cylinder interaction frequencies and developed flow in domain. The velocity profile near the locations of the
generally appear due to the impact of cylinder interactions on the flow cylinder (x2, x3, and x4) is influenced by both the presence of the cyl­
structure. And in the present case, this can be attributed to the inder and the presence of the plate. Higher values of velocity can be
increasing length of plate. Because of the fact that with increase in plate observed at small plate length (L = 0) which gradually drop when plate
length the incoming flow doesn’t interact completely with the cylinder length increases. Further, it can be noted that the velocity rises quickly
as compared to unaltered cylinder. Some of the shear layers detaching from the inlet and in upstream locations flow velocity varies smoothly
from plate overshoot the cylinder and some inner layers interact with with minimal oscillations.
front surface of cylinder. These interactions result in disturbances in
flow induced forces (mainly the drag force in case of square cylinder 4.1. Statistical analysis of physical parameters
however lift force is also impacted) as seen from the graphs of drag
coefficient for these cases. These graphs indicate that at L = 3 and 6 the In this section, the effects of flat plate length on physical parameters
amplitudes of consecutive cycles of CD vary (Fig. . 4(b) and 4(c)). But such as CDm, St, CDrms, CLrms and CLAmp are discussed. Also, the per­
since these peaks correspond to zero value of Strouhal number, that’s centage differences of physical parameters from solo cylinder for
why are ignored due to the fact that these indicate the zero shedding different plate lengths are presented.
frequency and less impact on CL. Further, in Fig. 6(a-c) the St value,
shown at peak in each graph, first increases with plate length L ≤ 1.5, 4.1.1. Variation of physical parameters under the effect of upstream flat
then decreases until L = 6.0 (detailed discussion is in section 4.1). Also plate lengths
the spectrum energy decreases as the plate length increases from 0.1 to The variations in CDm against the varying plate lengths (L = 0.1 to 6)
6. The reason is that different vortex shedding patterns are directly depicted in Fig. 8(a) at Re = 150. It can be seen that the CDm decreases
associated with qualitative changes in spectrum peak. The decrement in sharply as plate length increases up to L = 3. The highest value of CDm
the spectrum energy indicates the weaker vortex shedding from cylinder can be found at smallest plate length L = 0.1, i.e., 1.0540, where the flow
with increment in plate length, which confirms the effect of plate length. mode is unsteady. However, for L ≥ 3 the decrement in CDm becomes
The velocity field’s orientation at five different axial positions (x = 1, slow as compared to L ≤ 3 and at L = 6, it reaches its minimum value.
201, 211, 221 and 621) and flow field characteristics are presented in Zhou et al. (Zhou et al., 2005) observed that as the height of upstream
Fig. 7(a-c). For this purpose, two lines are drawn at inlet and outlet flat plate increases the drag of square cylinder decreases. Rajkumar et al.

6
H. Rahman et al. International Journal of Heat and Fluid Flow 105 (2024) 109272

Fig. 6. Power Spectrum of CL corresponding to different plate lengths.

(Rajkumar et al., 2014) found that introducing an upstream detached in Fig. 8(c). It can be observed that between plate lengths 0.1 and 0.3,
flat plate decreased the CDm experienced by a circular cylinder. They the CDrms values decrease, then show increasing behavior, and at L = 0.4
reported that CDm decreases as the length of the plate increases for Re = the highest value of CDrms i.e., 0.2791 is attained. After that the values of
20 and 155. CDrms show almost constant behavior from L = 0.5 to 6. This indicates
The St variations against different lengths of plate are shown in Fig. 8 the increment in length of plate does not lead major changes but minor
(b). For smaller plate lengths, the St sharply increases in the range 0.1 ≤ variations in CDrms. The CLrms values also decrease with increase in plate
L ≤ 1 unlike the CDm. The reason for increase in St at smaller plate lengths (Fig. 8(d)) as was observed in case of amplitude of lift cycles (see
lengths can be credited to the formation of elongated vortical structures Fig. 5). It has maximum value 0.2435 at L = 0.1 and decreases with
in the wake as observed in vorticity graphs (see Fig. . 2). In the range of increasing L. At these values of L, flat variations in the lift force were
1.1 ≤ L ≤ 1.5, the Strouhal number exhibits a constant behavior. This is observed without modulations that’s why CLrms reduces with increment
attributed to the fact that the plate has begun to regulate the recircu­ in L. The minimum value of CLrms can be observed at L = 6.
lation of shear layers. As the effect of the plate becomes more dominant Fig. 8(e) shows the alteration of CLAmp for flow around a square
in determining the formation mechanism of vortex shedding, the St cylinder with upstream plate against L = 0.1 to 6. It is important to
values monotonically decrease until L = 6. Within the range of 5 ≤ L ≤ 6, mention here that compared with CLAmp of a single square cylinder,
the St values reach a local minimum. Barman and Bhattacharyya CLAmp of cylinder with attached plate drops slightly. The graph shows
(Barman and Bhattacharyya, 2015) observed that the presence of an that initially CLAmp decreases and then rises sharply with increasing L up
upstream flat plate had minimal impact on the vortex shedding fre­ to 0.5 with maximum value 0.5476. The rapid increase in CLAmp is due to
quency of a square cylinder. However, they noted a significant reduction periodic vortex shedding in the down wake region. After L = 0.5, the
in the mean drag coefficient with the introduction of the upstream flat values of CLAmp smoothly vary with almost constant value for L = 0.6 to
plate. Our results align with these findings, demonstrating a similar 6 which indicates the effect of attached plate.
trend.
The CDrms variations with respect to different values of L are shown

7
H. Rahman et al. International Journal of Heat and Fluid Flow 105 (2024) 109272

Fig. 7. Streamwise velocity at different plate lengths.

4.1.2. Analysis of reduction in physical parameters vortices shedding from cylinder as compared to single cylinder without
The values of physical parameters in terms of percentage difference plate. Fig. 9(c) shows the percentage differences in CDrms from single
from single cylinder without plate are computed in order to see how the cylinder. The percentage values are high at shorter plate lengths, and at
attached flat plate at upstream location affects the CDm, St, CDrms, CLrms L = 0.4, the maximum reduction of 76.74% can be noticed. The per­
and CLAmp. The percentage difference of parameters corresponding to centage difference values do not grow significantly for the remaining
different flat plate lengths are plotted in Fig. 9(a–e). plate lengths in the range L = 0.5–6. When compared to CDm values,
Based on the computed data, all percentage values for physical pa­ CDrms data shows higher reduction at smaller plate lengths while less
rameters are positive, indicating a reduction in the values of these pa­ reduction at higher plate lengths. The variations in CLrms in terms of
rameters due to the attachment of the plate except for St and CLAmp. In percentage differences are shown in Fig. 9(d). It can be seen from the
the range 0.1 ≤ L ≤ 6, the percentage difference in CDm increase with figure that the behavior of CLrms is almost same as observed in CDm but
increasing plate lengths (Fig. 9(a)). Additionally, the least reduction, with slightly less reduction as compared to CDm. That is, by increasing
with a value of 20.2%, is observed at the minimal length L = 0.1 while the plate length the CLrms decreases as compared to single cylinder
the maximum reduction, 64.7%, happens at the longest plate length L = without plate. The minimum 3.7% reduction can be observed at L = 0.1,
6. According to Barman and Bhattacharyya (Barman and Bhattacharyya, whereas maximum reduction occurs at L = 6 with the value 61.7%.
2015) the reduction in average drag is due to reduction in pressure at Fig. 9(e) shows that the behavior of CLAmp in term of percentage dif­
front surface of cylinder. ferences from L = 0.1 to 6. It can be seen that the percentage difference
The variations in St in terms of percentage differences against plate in values of CLAmp is maximum at plate length L = 0.2, with highest
lengths are shown in Fig. 9(b). At L = 0.1, maximum reduction, 1.4% reduction of 10.6%. When L varied from 0.4 to 6.0, the opposite trend
occurs and after that for all plate lengths opposite trend can be observed. can be observed in CLAmp with minimum reduction is -73.2%. Similar to
The negative value means that at these plate lengths, accession in St St this is due to larger recirculation zones and wider wake width which
happens by plate attachment rather than reduction. For the range 0.2 ≤ results in high amplitude lift force as compared to single cylinder
L ≤ 6, the minimum percentage difference value is − 11.18% and without plate.
maximum % value is 1% in St. This indicates that the wake width and
recirculation region increase which results in increase in frequency of

8
H. Rahman et al. International Journal of Heat and Fluid Flow 105 (2024) 109272

Fig. 8. Variations in CDm, St, CDrms, CLrms and CLAmp with increasing plate lengths.

5. Conclusion CDrms, CLrms, St and CLAmp was also investigated. Furthermore, the per­
centage difference in these values was calculated by comparing the
The flow around a single square cylinder in presence of an attached values of these parameters to those of a single square cylinder without
upstream flat plate was investigated at Re = 150 using the LBM. The flat plate. The following are key findings of the current study:
plate length was systematically varied in the range 0.1 ≤ L ≤ 6. The
effects of plate on flow characteristics were examined in terms of 1. At all chosen plate lengths, the wake was found to be unsteady with
vorticity graphs, streamline visualization, variation of time dependent formation of vortices owing to the Von Karman vortex street. The
force coefficients, energy spectrum and streamwise velocity. The influ­ shape and size of vortices as well as the wake formation region was
ence of varying plate lengths on major flow parameters such as CDm, significantly altered due to presence of plate.

9
H. Rahman et al. International Journal of Heat and Fluid Flow 105 (2024) 109272

Fig. 9. Percentage differences in CDm, St, CDrms, CLrms and CLAmp from single cylinder without plate at different plate lengths.

2. The length of the plate had a significant impact on the CD and CL. 4. With increasing plate length, flow parameters such as St, CLrms and
Both of these coefficients appeared to be unsteady at less time in case CLAmp demonstrated steady behavior after decreasing initially, but
of smaller plate lengths, but as plate length increased, the unstead­ CDm and CLrms were observed to decrease at all plate lengths.
iness phenomena slowed down. Additionally, pressure at surface of 5. It was observed that when plate was attached at upstream of the
cylinder significantly reduced due to presence of plate. cylinder, the CDm, St, CDrms, CLrms and CLAmp values reduced up to
3. The primary peak value in the power spectrum graphs, for all plate 64.7%, 1.4%, 76.74%, 61.7% and 10.6%, respectively. At some
lengths, indicated the dominance of vortex shedding frequency. plate lengths the St and Clamp were found to be increasing as
compared to single cylinder without plate values.

10
H. Rahman et al. International Journal of Heat and Fluid Flow 105 (2024) 109272

CRediT authorship contribution statement Islam, S.U., Rahman, H., Abbasi, W.S., 2014a. Grid independence study of flow past a
square cylinder using the multi-relaxation-time lattice boltzmann method. Int. J.
Math., Comput., Phys. Quantum Eng. 8 (7), 1019–1029.
Hamid Rahman: . Waqas Sarwar Abbasi: Conceptualization, Islam, S.U., Rahman, H., Abbasi, W.S., Noreen, U., Khan, A., 2014b. Suppression of fluid
Investigation, Methodology, Supervision, Visualization, Writing – orig­ forces on flow past a square cylinder with a detached flat plate at low Reynolds
inal draft, Writing – review & editing. Saba Ismail: Data curation, number for various spacing ratios. J. Mech. Sci. Technol. 28 (12), 4969–4978.
Islam, S.U., Abbasi, W.S., Ying, Z.C., 2016. Transitions in the unsteady wakes and
Investigation, Writing – original draft. Sumaira Nadeem: Writing – aerodynamic characteristics of the flow past three square cylinders aligned inline.
review & editing. Aerosp. Sci. Technol. 50, 96–111.
Islam, S.U., Manzoor, R., Ying, Z.C., Rashdi, M.M., Khan, A., 2017. Numerical
investigation of fluid flow past a square cylinder using upstream, downstream and
Declaration of competing interest dual splitter plates. J. Mech. Sci. Technol. 31 (2), 669–687.
Islam, S.U., Manzoor, R., Khan, U., Nazeer, G., Hassan, S., 2018. Drag reduction on a
The authors declare that they have no known competing financial square cylinder using multiple detached control cylinders. KSCE J. Civ. Eng. 22 (5),
2023–2034.
interests or personal relationships that could have appeared to influence Manzoor, S., Khawar, J., Sheikh, N.A., 2013. Vortex-induced vibrations of a square
the work reported in this paper. cylinder with damped free-end conditions. Adv. Mech. Eng. 2013, 1–12.
Mohammad, A.A., 2011. Lattice Boltzmann method: Fundamentals and Engineering
Applications with Computer Codes. Springer.
Data availability Norberg, C., 1993. Flow around rectangular cylinder: pressure forces and wake
frequencies. J. Wind Eng. Ind. Aerodyn. 49, 187–196.
Data will be made available on request. Okajima, A., 1982. Strouhal numbers of rectangular cylinders. J. Fluid Mech. 123,
379–398.
Porteous, R., Moreau, D.J., Doolan, C.J., 2019. The effect of the incoming boundary layer
References thickness on the aeroacoustics of finite wall-mounted square cylinder. J. Acoust. Soc.
Am.
Abbasi, W.S., Mahmood, R., Naheed, A., 2020. On the wake interference effects for flow Rajkumar, M.R., Rijas, A.S., Venugopal, G., 2014. Flow past a cylinder with upstream
around tandem bodies. J. Braz. Soc. Mech. Sci. Eng. 42 (53), 1–20. splitter plate. In: 10th International Conference on Heat Transfer, Fluid Mechanics
Abbasi, W.S., Naheed, A., Islam, S.U., Rahman, H., 2021. Investigation of optimum and Thermodynamics, pp. 133–139.
conditions for flow control around two inline square cylinders. Arab. J. Sci. Eng. 46, Saha, A.K., Shrivastava, A., 2015. Suppression of vortex shedding around a square
2845–2864. cylinder using blowing. Sadhana 40 (3), 769–785.
Barman, B., Bhattacharyya, S., 2015. Control of vortex shedding and drag reduction Samani, M., Bergstrom, D.J., 2015. Effect of a wall on the wake dynamics of an infinite
through dual splitter plates attached to a square cylinder. J. Mar. Sci. Appl. 14 (2), square cylinder. Int. J. Heat Fluid Flow 55, 158–166.
138–145. Singh, S.P., Mittal, S., 2005. Vortex-induced oscillations at low Reynolds numbers:
Bhatnagar, P., Gross, E.P., Krook, M.K., 1954. A model for collision processes in gases. I. Hysteresis and vortex-shedding modes. J. Fluids Struct. 20 (8), 1085–1104.
small amplitude processes in charged and neutral one-component systems. Phys. Sohankar, A., Norberg, C., Davidson, L., 1998. Low Reynolds number flow around a
Rev. 94 (3), 511–525. square cylinder at incidence: Study of blockage, onset of vortex shedding and outlet
Billah, K.Y., Scanlan, R.H., 1991. “Resonance, tacoma narrows bridge failure and boundary condition. Int Journal for Numerical Methods in Fluids 26, 39–56.
undergraduate physics textbooks.”. Am. J. Phys. 592, 118–124. Sohankar, A., Khodadadi, M., Rangraz, E., 2015. Control of fluid flow and heat transfer
Breuer, M., Bernsdorf, J., Zeiser, T., Durst, F., 2000. Accurate computations of the around a square cylinder by uniform suction and blowing at low Reynolds numbers.
laminar flow past a square cylinder based on two different methods: lattice Comput. Fluids 109, 155–167.
Boltzmann and finite-volume. Int. J. Heat Fluid Flow 21, 186–196. Soumya, S., Prakash, K.A., 2017. Effect of splitter plate on fluid flow characteristics past
Chapman, S., Cowling, T.G., 1970. Mathematical theory of non-uniform gases, Third a triangular cylinder. J. Phys. no, 822.
edition. Cambridge University Press. Taherian, G., Ahmadabadi, M.N., Karimi, M.H., Tavakoli, M.R., 2017. Flow visualization
Chen, Y.J., Shao, C.P., 2013. Suppression of vortex shedding from a rectangular cylinder over a thick blunt trailing-edge airfoil with base cavity at low Reynolds numbers
at low Reynolds numbers. J. Fluids Struct. 43, 15–27. using PIV technique. J. Vis. 20, 695–710.
Guo, Z., Shi, B., Wang, N., 2000. Lattice BGK Model for Incompressible Navier-Stokes Teksin, S., Yayla, S., 2016. Effects of flexible splitter plate in the wake of a cylindrical
Equation. J. Comput. Phys. 165, 288–298. body. Journal of Applied Fluid Mechanics 9 (6), 3053–3059.
Guo, Z., Liu, H., Luo, L.-S., Xu, K., 2008. A comparative study of the LBE and GKS Turki, S., 2008. Numerical simulation of passive control on vortex shedding behind
methods for 2D near incompressible laminar flows. Computational Physics 227, square cylinder using splitter plate. Eng. Appl. Computational Fluid Mech. 2 (4),
4955–4976. 514–524.
Han, X., Wang, J., Zhou, J., Zhang, G., Tan, S.K., 2019. Numerical Simulation of flow Wolf Gladrow, D.A., 2005. “Lattice-gas cellular automata and lattice boltzmann models.
control around a circular cylinder by installing a wedge-shaped device upstream. An Introduction” Springer.
J. Marine Sci. Eng. 7 (422), 1–11. Yichang Lin, Y., Xie, X., Yan, L., Zhou, Y., 2022. Research on the collapse of Tacoma
Hwang, J.Y., Yang, K.S., 2007. Drag reduction on a circular cylinder using dual detached narrows bridge under the finite element application ANSYS of computational
splitter plates. J. Wind Eng. Ind. Aerodyn. 95, 551–564. mechanics. J. Phys. Conf. Ser. 2230 (2022), 012018.
Igbalajobi, A., Mcclean, J.F., Sumner, D., Bergstrom, D.J., 2013. The effect of awake- Zhou, L., Cheng, M., Hung, K., 2005. Suppression of fluid force on a square cylinder by
mounted splitter plate on the flow around a surface-mounted finite-height circular flow control. J. Fluids Struct. 21 (2), 151–167.
cylinder. J. Fluids Struct. 37, 185–200. Zhu, H., Wang, K., 2019. Wake adjustment and vortex-induced vibration of a circular
cylinder with a C-shaped plate at a low Reynolds number of 100. Phys. Fluids 31,
1–13.

11