International Journal of Heat and Mass Transfer 117 (2018) 50–57

Contents lists available at ScienceDirect

International Journal of Heat and Mass Transfer

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

Flow and heat transfer characteristics over a square cylinder with corner
modifications
Tehmina Ambreen, Man-Hoe Kim ⇑
School of Mechanical Engineering, Kyungpook National University, Daegu 41566, South Korea

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

Article history: In the present study, the effect of corner modifications on fluid flow and heat transfer characteristics
Received 16 June 2017 across a square cylinder has been analyzed numerically in an effort to improve thermohydraulic param-
Received in revised form 28 September 2017 eters. Two-dimensional simulations have been carried out for laminar flow across a square cylinder with
Accepted 29 September 2017
sharp, round, chamfered and recessed corners for Reynolds number range 55–200. Corner variations have
Available online 7 October 2017
been made for dimension c/D = 0.125, where c and D indicate corner size and cylinder diameter respec-
tively. When compared with the sharp-cornered cylinder, the results illustrated that corner modifications
Keywords:
lead to significant drag reduction, however, the penalty in terms of Strouhal number increment is com-
Square cylinder
Corner configurations
paratively low. Deflected flow from upstream modified corners promotes flow separation from down-
Forced convection stream corners in contrast to the sharp-cornered cylinder results in narrow wake width with intense
Pressure coefficient fluid circulations, farther vortex shedding location and consequently reduced pressure drag as well as
Nusselt number improved heat transfer coefficients. Recirculating fluid inside the upstream corner cut of recessed corners
additionally contributes towards drag increment as compared to other corner configurations. Results also
indicate that upstream corners modifications are more influential in amelioration of thermohydraulic
characteristics as compared to downstream corners. Moreover, a new correlation for the average
Nusselt number has been developed as a function of Reynolds number.
Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction investigations of Tamura et al. [3,4], elaborated the effects of cor-

Fluid patterns in the premises of a bluff body attracted a great
attention in past years because of diverse fluid flow and heat trans-
fer relevancies e.g. fluid loading on skyscrapers, chimneys, offshore
structures, pipelines, suspension bridges, towers, mast, wires and
heat exchangers. Under cross-flow fluid exposure, flow inertia
impels boundary layer to a bluff body encountering retardation
at structure surface and subsequently results in a large pressure
gradient, drag coefficient and vortex shedding phenomena. This
wake becomes responsible for inciting fluctuating forces on the
structure and thus devastating the structural integrity. For bluff
structures with rectangular cross-section, Sohankar et al. [1] and
Subhankar et al. [2] highlighted the prime contribution of down-
stream (Re
100) and upstream corners (Re > 125) towards flow
disintegration.
Few studies suggested that corner modifications can lead to a
reduction in aerodynamic forces experienced by the square
cross-sectioned structure. In their experimental and numerical
⇑ Corresponding author.
E-mail addresses: tehminaambreen91@gmail.com (T. Ambreen), manhoe.kim@
knu.ac.kr (M.-H. Kim).
ner altercations on wake structures at Reynolds number 104. Anal-
ogous results and additional drag coefficient reduction was
predicted by Dalton and Zheng [5] in their numerical investigation
of cross-flow over the square and diamond cylinders with rounded
corners (c/D = 0.125, where c represents corner dimension and D
cylinder diameter) at Re = 1000. Hu et al. [6] and Mola et al. [7]
investigated fluid flowing over a cylinder with rounded corners
at Reynolds number of 2600–6000 and 105 respectively. Their
results demonstrated that three-dimensional flow instabilities
close to the open end of the cylinder with sharp corners, discon-
certing vortex shedding synchronization behind the cylinder, are
responsible for drag coefficient variations about the cylinder for
turbulent flows. These variations are insignificant when cylinder
corners are rounded. Sajjad and Sohn [8] elucidated the effect of
corner roundedness (c/D = 0.1, 0.2, 0.3, 0.4, 0.5) on aerodynamic
forces experienced by square cylinder at Re ¼ 500 and determined
that corner radius significantly influences flow characteristics. He
et al. [9] investigated the upstream corner cut effects on the drag
at Re ¼ 1035 by employing particle image velocimetry technique.
Their results showed that front corner cut for selected dimensions,
significantly reduced drag coefficient. Tong et al. [10] numerically
investigated the cross flow over a square and octagonal chamfered

https://doi.org/10.1016/j.ijheatmasstransfer.2017.09.132
0017-9310/Ó 2017 Elsevier Ltd. All rights reserved.
 T. Ambreen, M.-H. Kim / International Journal of Heat and Mass Transfer 117 (2018) 50–57 51

Nomenclature

c Corner size [m] U free stream velocity ½m s1 

Cdav g Drag coefficient ð¼ Cdv þ Cdp Þ [-] u non-dimensional stream-wise velocityðu ¼ u =UÞ [-]
Cdv viscous drag [-] u stream-wise velocity ½m s1 
Cdp pressure drag [-] uc average non-dimensional stream-wise velocity[-]
cp specific heat of the fluid [J kg1 k1] v non-dimensional cross stream velocityðv ¼ v  =UÞ [-]
 
Cp

pressure coefficient ¼ qUp 2 [-]
v cross-stream velocity ½m s1 
x stream-wise dimension of coordinates [m]
D cylinder diameter [m] y cross-stream dimension of coordinates [m]
fs Vortex shedding frequency [1/s] x non-dimensional stream-wise dimension of coordinates
G grid size ð¼ x =DÞ
hx convective heat transfer coefficient ½W=ðm2 KÞ y non-dimensional cross stream dimension of coordinates
k thermal conductivity of fluid ½W m1 K1  ð¼ y =DÞ
Lu upstream flow domain [m]
Ld downstream flow domain [m] Greek symbols
Nuav g average Nusselt number [-] b non-dimensional temperature [-]
Pr Prandtl number [-] t non-dimensional time (¼ t UD1 ) [-]
P dynamic pressure [Pa] h non-dimensional temperature
Re Reynolds number ð¼ UD=Þ [-] q density of fluid [kg m3]
Recrit critical Reynolds Number [-]  
St Strouhal Number [-] e blockage ratio HD
T fluid temperature at inlet [K] l viscosity of fluid
Tw constant wall temperature at cylinder surface [K]

cylinder at Re = 2  106 . Recently Sajjad and Sohn [11] studied dimensions are c=D ¼ 1=8 ¼ 0:125. The cylinder has been placed
round cornered square cylinder with flow incidence angle and near the inlet in order to allow sufficient downstream distance to
determined critical angle of 12° for minimum drag and lift forces. capture vortex shedding phenomena. Constant cylinder surface
In the context of the presented literature review, it can be stated temperature T w has been defined slightly higher than incoming
that although some effective geometric configurations have been fluid temperature T to minimize the temperature effects on fluid
suggested by past researchers in order to modify the fluid behavior thermo-physical properties. All geometrical dimensions have been
to achieve desirable characteristics of reduced aerodynamic forces non-dimentionalized with cylinder diameter D, velocities with free
however no one discussed heat transfer characteristics which are stream velocity U, frequencies with U=D, physical times with D=U
essential for optimum cross flow heat exchangers’ designs. There- while pressure coefficients with dynamic pressure, P ¼ p =qU 2 .
fore in the present study, fluid flow and heat transfer parameters of With standardized notation of Cartesian tensors, the flow gov-
fluid flow over a cylinder with sharp and modified corners have erning equations can be formulated as;
been investigated numerically. Fig. 1 shows the corner configura-
@u @ v
tions considered for present study which includes corner rounding, þ ¼0 ð1Þ
@x @y
chamfering and additionally corner chamfering which has been not
been introduced in previous studies. Non-dimensional parameters !
i.e. coefficient of drag ðCdav g Þ, Strouhal number ðStÞ and pressure @u @u2 @vu @P 1 @ 2 u @ 2 @u
þ þ ¼ þ þ ð2Þ
coefficient distribution ðC p Þ over cylinder surface have been ana- @t @x @y @x Re @x2 @y2
lyzed in order to interpret fluid behavior, while local heat transfer !
coefficient (hx ) and average Nusselt number ðNuav g ) have been dis- @ v @uv @ v 2 @P 1 @ 2 v @ 2 v
cussed to elaborate heat transfer characteristics. Moreover instan- þ þ ¼ þ þ ð3Þ
@t @x @y @y Re @x2 @y2
taneous streamlines have been plotted to visualize flow attributes.
!
@b @ub @ v b 1 @2b @2b
2. Computational model þ þ ¼ þ ð4Þ
@t @x @y RePr @x2 @y2

2.1. Physical and mathematical details where non-dimensional variables are;


u v 
x y 
p tU
The physical description along with the applied boundary con- u¼ ;v ¼ ;x ¼ ;y ¼ ;p ¼ ;t ¼ ð5Þ
ditions [1,12] has been shown in the Fig. 2, where modified corner
U U D D qU 2 D

Fig. 1. Cylinder with (a) Sharp, (b) Rounded, (c) Chamfered and (d) Recessed corners.
52 T. Ambreen, M.-H. Kim / International Journal of Heat and Mass Transfer 117 (2018) 50–57

Fig. 2. Flow configuration.

where U is constant fluid velocity in inviscid region away from non-dimensional time step size of the order of Dt = 0.01 has been
cylinder. Non-dimensional Prandtl number, Reynolds number and specified to ensure Courant number less than 1. Convergence crite-
Nusselt number are nominated as; ria for all parameters (continuity, momentum, energy) has been
restricted bellow 106 .
lC p qUD hD
Pr ¼ ; Re ¼ ; Nuavg ¼ ð6Þ In order to attain grid and domain insensitive solution, five
k l k
grids have been tested for all geometric configurations as depicted
Non-dimensional temperature b which is defined as; in Fig. 3. At first, structured non-uniform grids G1, G2 and G3 (finer
grid near cylinder surface) with domain sizes of 15D  25D,
T  T 20D  20D and 30D  50D respectively has been studied. Grid sen-
b¼ ð7Þ
Tw  T  sitive parameters, RMS coefficient of drag and lift at Reynolds num-
ber 200, have been compared to evaluate the domain sizes. Fig. 4
where Tw is wall surface temperature and T  is inlet fluid
shows that G2 and G3 estimated identical results therefore G2 with
temperature.
grid elements 43,800 has been used for grid sensitivity test. Finer
The coefficient of lift and drag has been calculated by the pres-
grid G4 improved the results however results remained insensitive
sure and viscous forces exerted on the cylinder. Consequently, the
to further increment of the grid elements G5 (75,500). Therefore
Strouhal number is described as St ¼ f Us D. Where f s represents vor- G4 with grid elements of 62200 (Y+ <1) has been employed for all
tex shedding frequency. Heat transfer from the cylinder wall to subsequent simulations in order to reduce computation time.
the external fluid flowing has been calculated in terms of a non- To assess the accuracy of numerical methodology, present
dimensional Nusselt number Nuav g . results of Strouhal number (St) and time-averaged drag coefficient
(Cdavg) computed for a sharp cornered cylinder has been compared
2.2. Numerical model and validation with the experimental results of Norberg [14] and numerical
results of Sharma and Eswaran, Sen et al. and Cao et al. [15–17]
Commercial software Ansys Fluent 17.1 [13] has been employed as shown in Fig. 5. The percentage difference of calculated St from
to simulate the present problem by using control volume the experimental results of Norberg [14] at Re = 175 is 3% while
technique for transient simulations. The coupled scheme has with the numerical results of Cao et al. and Sen et al. [15,16] at
been implemented for pressure-velocity coupling whereas convec- Re 150 and 200 is 2.5%. Present Cdavg has a percentage difference
tive terms discretization has been done by Quick schemes. The of <5% with the literature [15–17].

Fig. 3. Close-up view of Grid for cylinder with (a) Sharp, (b) Rounded, (c) Chamfered and (d) Recessed corners.
 T. Ambreen, M.-H. Kim / International Journal of Heat and Mass Transfer 117 (2018) 50–57
3. Results and discussions
To illustrate the influence of corner modifications on the fluid
flow parameters, results have been reported in terms of coefficient
of drag ðCdav g Þ, Strouhal number ðStÞ and pressure coefficient dis-
tribution ðC p Þ in addition to flow streamlines. While Average Nus-
selt numberðNuav g ) and heat transfer coefficient distribution (hx )
along cylinder surface has been plotted to elaborate heat transfer
characteristics.
3.1. Fluid flow characteristics
Cdavg and Strouhal number as a function of Reynolds number for
all cylinder corner configurations have been presented in Figs. 6
and 7, respectively. Results illustrate that Cdavg of sharp-cornered
cylinder observes local minima at Re = 150 followed by an incre-
ment at higher Reynolds number [18]. When modifying the cylin-
der corners, a characteristic decrease in Cdavg has been observed
which becomes more pronounced at higher Reynolds number.
Fig. 4. Grid and Domain Independence study at Re = 200.
Fig. 5. Comparison of present St and Cdavg with the previous result.
53
The trend of drag coefficient can be explicitly analyzed by time-
averaged streamlines as shown in Figs. 8–10. In accordance with
the literature [1,15], present results suggest that for the sharp-
cornered cylinder, flow extricating from the upstream corners of
cylinder re-attaches to the downstream lateral surfaces at
125 6 Re 6 150 and consequently reducing the wake width to a
minimum value. As shown in Fig. 8, the streamwise recirculating
fluid trapped between fluid shear layers and cylinder upstream lat-
eral surfaces deflect the flow towards downstream lateral surfaces.
At higher Reynolds number, wider recirculation region encounters
disruption as a result of back fluid flow thus diminishing the flow
reattachment phenomena. Completely detached flow results in
wider wake, lower wake pressure and consequently higher Cdav g .
The reduction in drag coefficient can be described as corners
responsible for flow separation at cylinder surface are not
predetermined for modified corner cylinder in contrary to the
sharp-cornered cylinder. For round corners, the corner roundness
contributes towards additional shear stress and thus flow follows
Fig. 6. Variation of drag coefficient with Re for sharp and modified corner cylinders.
Fig. 7. Variation of Strouhal number with Re for sharp and modified corner
cylinders.
54 T. Ambreen, M.-H. Kim / International Journal of Heat and Mass Transfer 117 (2018) 50–57

Fig. 8. Streamlines around sharp corner cylinder at Re = 100, 125, 150 and 200.

Fig. 9. Streamlines around round corner cylinder at Re = 175 and 200.

Fig. 10. Streamlines around (a) chamfered and (b) recessed corner cylinder at Re = 200.
 T. Ambreen, M.-H. Kim / International Journal of Heat and Mass Transfer 117 (2018) 50–57
the cylinder contours instead of upstream corner separation.
However, at Re = 200, a little streamwise recirculation appears on
the lateral surface of cylinder indicating a shift of flow separation
towards upstream corners. By chamfering the corners, the fluid
behavior is identical to round corner cylinder but the recirculating
region is absent even at Re = 200.
For chamfered corners, although primary upstream corner initi-
ate flow separation but recirculating fluid inside the upstream cor-
ner cut deflect the flow towards secondary upstream, reducing the
pressure to a minimum value and thus suppressing the upstream
corner flow separation as shown in Fig. 10(b). This upstream fluid
recirculation provides additional pressure increment at cylinder
upstream, causing the pressure drag higher than rounded and
chamfered corners. Downstream flow separation for modified cor-
ner cylinders is additionally accompanied by the farther diversion
of detached shear layers, longer vortex formation wake length, and
narrower wake width and reduction of the pressure drag.
The highest Cdavg for the sharp-cornered cylinder is analogous
to the highest Cdavg at an aspect ratio of 0.67 which is associated
with reduced wake length and vortices generated closer to cylinder
rear surface. As a result, the rolling up of the shear layer becomes
weak and the wake pressure decreases [19,20].
Modification of corners, induces higher vortex shedding
frequency St for obvious reasons, however in comparison to drag
Fig. 11. Pressure distribution along the cylinder with (a) Sharp, (b) Round, (c) recessed and (d) Chamfered corners at Re = 100, 175 and 200.
55
reduction, the penalty in St Increment is not remarkably high i.e.
at Re = 200, percentage decrease of drag coefficient is 14%, 16%
and 19% whereas percentage increment of St 5%, 3% and 6% for
recessed, rounded and chamfered corners respectively.
Pressure coefficient in Fig. 11 explains the pressure recovery
region in upstream corner cut of the recessed corners, more
clearly. Cylinder upstream surface exhibits highest pressure coef-
ficient because of stagnation whereas pressure coefficient is neg-
ative from the upstream corners to cylinder rear surface
indicating the wake formation. For round upstream corners, the
stagnation point is not as distinct as for square cylinder because
of corner curvature. For chamfered corners, C p curve at cylinder
front face shows a little deflection as compared to sharp and
round corners, depicting incomplete flow attachment at cham-
fered corners.
As expected higher Re results in higher negative pressure on the
cylinder rear surface as well as upstream corner cut.
3.2. Heat transfer characteristics
Variation of Nuavg as a function of Reynolds number has been
depicted in Fig. 12 for all geometric configurations under discus-
sion. Results for sharp corner cylinder has also been validated with
the average Nusselt number correlation proposed by [17] as;
56 T. Ambreen, M.-H. Kim / International Journal of Heat and Mass Transfer 117 (2018) 50–57

Fig. 12. Variation of average Nusselt number with Re for sharp and modified corner
cylinders.
pffiffiffiffiffiffi
Nuav g ¼ 0:359 Re þ 0:442
An increase in Nuavg with Reynolds number has been observed
for sharp [21] and modified corners but corner altercation results
in higher Nuavg at specific Re, as wake width reduces to a minimum
value and consequential intense fluctuations of fluid inside wake
are incited. The increment in Nuavg for modified corner cylinders
becomes more significant at higher Reynolds number. At Re = 200
percentage increases of Nuavg when compared to the sharp-
cornered cylinder is 7%, 16% and 20% for recessed, rounded and
chamfered cornered cylinder respectively.
Distribution of convective heat transfer coefficient along cylin-
der surface for all cases is plotted in Fig. 13. For all cases under con-
sideration, hx is almost analogous on top and bottom surfaces of
the cylinder. hx is optimum at front surface followed by rear and
top surfaces. On the lateral surfaces, the hx decreases, because of
the wider wake and consequently the fluid trapped between
detached boundary layer from upstream corner and cylinder sur-
face hinders further convective heat transfer. On the rear corner,
hx again increases as the recirculating fluid at the rear of cylinder
extends up to downstream corners and this recirculating fluid is
cooler as compared to upstream approaching fluid thus convective
heat transfer increases. hx gradually increases with increasing Re

Fig. 13. Time average local Nusselt No. distribution vs Reynolds number for cylinders with (a) Sharp, (b) round (c) Recessed and (d) Chamfered corners.
 T. Ambreen, M.-H. Kim / International Journal of Heat and Mass Transfer 117 (2018) 50–57
on front and rear surfaces and this amplification is more predom-
inant on the rear side.
For modified corners, hx has increasing trend with Re at top and
bottom surfaces unlike sharp cornered cylinder. By rounding the
corners, hx curve follows almost similar trend but a bit curvy at
upstream and downstream corners because of curvature, moreover
hx value at two lateral surfaces of cylinder are slightly higher than
that of sharp corner cylinder as a result of less wider wake.
For recessed corners, distribution of hx along the corner cut is
such that it initially begins with optimum value, followed by an
intense reduction then again a maxima exists at the end and this
trend is significant at the front corners. This is because the fluid
trapped inside the corner cut provides resistance for convective
heat transfer. Chamfered cornered cylinder also has identical but
less intense hx trend at cylinder corners accentuating the state-
ment discussed earlier that fluid is not in complete contact at
chamfered locations. hx results for all cases indicate that upstream
corners modifications are more influential to total heat transfer
rate in contrast to downstream corner modifications.
A generalized correlation of Nuav g as a function of Reynolds
number ðRe 6 200Þ has been proposed for given corner configura-
tions by regression analysis. As;
Nuav g ¼ C n Re0:45
where C n the constant and its value is 0.571, 0.528 and 0.590 for
round, recessed and chamfered corner cylinders. The correlation
predicts the average Nusselt number with the deviation of ±4%.
4. Conclusion
Two-dimensional free stream flow across square cylinder, sub-
jected to constant wall temperature, has been investigated numer-
ically at four different corner dimensions. By using air (P r ¼ 0:7) as
the operating fluid, the parametric range of Reynolds number has
been selected 55 6 Re 6 200 in order to investigate fluid flow
and heat transfer characteristics under forced convection. Strouhal
numberðSt Þ, Drag coefficientðC d Þ, Pressure coefficient ðC p Þ in addi-
tion to convective heat transfer coefficient ðhx Þ and average Nusselt
number ðNuav g Þ and their variation with Reynolds number has
been investigated for each case. The deducted results can be
summarized as;
 Upstream modified corners deflect the flow towards down-
stream corners instead of flow separation thus results in a
reduction of wake width and drag coefficient. Moreover, the
penalty in Strouhal number is not as high as drag reduction
i.e. at Re = 200, the percentage decrease of drag coefficient is
14%, 16%, and 19% whereas percentage increment of Strouhal
number is 5%, 3% and 6% for recessed, rounded and chamfered
corners respectively.
 Streamwise recirculating fluid inside the upstream corner cuts
of recessed corners provides additional thrust at the cylinder
frontal face, making the drag slightly higher than other corner
modifications.
 Corner altercation results in higher heat transfer as corner mod-
ifications result reduced wake width and consequential intense
fluctuations of fluid inside wake are incited. At Re = 200
57
percentage increases of Nuavg when compared to the sharp-
cornered cylinder is 7%, 16% and 20% for recessed, rounded
and chamfered cornered cylinder respectively.
 Upstream corners modifications are more influential to fluid
flow and heat transfer characteristics as compared to down-
stream corners.
Conflict of interest
The authors declared that there is no conflict of interest.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at https://doi.org/10.1016/j.ijheatmasstransfer.
2017.09.132.
References
[1] A. Sohankar, C. Norberg, L. Davidson, Numerical simulation of unsteady low-
Reynolds number flow around rectangular cylinders at incidence,
Doktorsavhandlingar Vid Chalmers Tek. Hogsk. 71 (1998) 189–201.
[2] S. Sen, S. Mittal, Free vibration of a square cylinder at low Reynolds numbers, J.
Fluids Struct. 27 (2011) 875–884.
[3] T. Tamura, T. Miyagi, T. Kitagishi, Numerical prediction of unsteady pressures
on a square cylinder with various corner shapes, J. Wind Eng. Ind. Aerodyn. 74–
76 (1998) 531–542.
[4] T. Tamura, T. Miyagi, The effect of turbulence on aerodynamic forces on a
square cylinder with various corner shapes, J. Wind Eng. Ind. Aerodyn. 83
(1999) 135–145.
[5] C. Dalton, W. Zheng, Numerical solutions of a viscous uniform approach flow
past square and diamond cylinders, J. Fluids Struct. 18 (2003) 455–465.
[6] J.C. Hu, Y. Zhou, C. Dalton, Effects of the corner radius on the near wake of a
square prism, Exp. Fluids 40 (2006) 106–118.
[7] A. Mola, G. Bordonaro, M.R. Hajj, Low-frequency variations of force coefficients
on square cylinders with sharp and rounded corners, J. Struct. Eng. 135 (2009)
828–835.
[8] S. Miran, C.H. Sohn, Numerical study of the rounded corners effect on flow past
a square cylinder, Int. J. Numer. Methods Heat Fluid Flow. 25 (2015) 686–702.
[9] G.S. He, N. Li, J.J. Wang, Drag reduction of square cylinders with cut-corners at
the front edges, Exp. Fluids 55 (2014) 1–11.
[10] J.C.K. Tong, E.M. Sparrow, W.J. Minkowycz, J.P. Abraham, A new archive of heat
transfer coefficients from square and chamfered cylinders at angles of attack in
crossflow, Int. J. Therm. Sci. 105 (2016) 218–223.
[11] S. Miran, C.H. Sohn, Influence of incidence angle on the aerodynamic
characteristics of square cylinders with rounded corners, Int. J. Numer.
Methods Heat Fluid Flow 26 (2016) 269–283.
[12] B. Gera, P.K. Sharma, R.K. Singh, CFD analysis of 2D unsteady flow around a
square cylinder, Int. J. Appl. Eng. Res. 1 (2010) 602–610.
[13] ANSYS FLUENT Documentation, Theory Guide and User guide. Release 17.1, 2016.
[14] C. Norberg, Flow around rectangular cylinders: pressure forces and wake
frequencies, J. Wind Eng. Ind. Aerodyn. 49 (1993) 187–196.
[15] S. Sen, S. Mittal, G. Biswas, Flow past a square cylinder at low Reynolds
numbers, Int. J. Numer. Methods Fluids 67 (2011) 1160–1174.
[16] S. Cao, Y. Ge, Y. Tamura, Shear effects on flow past a square cylinder at
moderate Reynolds numbers, J. Eng. Mech. 138 (2012) 116–123.
[17] A. Sharma, V. Eswaran, Heat and fluid flow across a square cylinder in the two-
dimensional laminar flow regime, Numer. Heat Transf. Part A Appl. 45 (2004)
247–269.
[18] M. Breuer, J. Bernsdorf, T. Zeiser, F. Durst, Accurate computations of the
laminar flow past a square cylinder based on two different methods: lattice-
Boltzmann and finite volume, Int. J. Heat Fluid Flow. 21 (2000) 186–196.
[19] P.W. Bearman, J.M.R. Graham, E.D. Obasaju, G.M. Drossopoulos, The influence
of corner radius on the forces experienced by cylindrical bluff bodies in
oscillatory flow, Appl. Ocean Res. 6 (1984) 83–89.
[20] A. Okajima, Numerical simulation of flow around rectangular cylinders, J.
Wind Eng. Ind. 33 (1990) 171–180.
[21] M. Rahnama, S. Hashemian, M. Farhadi, Forced convection heat transfer from a
rectangular cylinder, in: 16th Int. Symp. Transp. Phenom. (2005) 1–5.