Flow Around a Circular Cylinder and Its Analysis

Abstract: CFD is widely used for numerous practical application in the field of
aerodynamics. This report investigates the aerodynamin performance of a circular shape. This
2D simulation was performed on a circular airfoil within a rectangular domain. This CFD study
was ANSYS Fluent R2024 environment. SST k-omega turbulence model was used in this
study. The solution is not mesh-sensitive and final mesh had a skewness of 0.56474 and
orthogonal quality of 0.7071, representing good quality of mesh.
Keywords: Fluid Dynamics; Aerodynamics; k-omega model

1 Introduction:
The use of Computational Fluid Dynamics (CFD) is on the rise for examining a broad range of
processes occurring in the lower regions of the atmospheric boundary layer. At the entrance of
the computational domain, the velocity and turbulence distributions are usually fully formed in
a computational fluid dynamics (CFD) simulation. The surface roughness of the land upstream
of the intake plane that is outside of the simulated domain must be appropriately represented
by these distributions[1]. In the computational domain, there are generally three distinct
regions:the central region where the actual obstacles are explicitly modeled with their
geometrical shape.In this case the circular hole is acting as the central region.The upstream and
downstream regions where the actual obstacles are implicitly modeled, meaning their geometry
is not included in the domain but their impact on the flow can be modeled through
roughness.Here, the inlet and outlet section is working as the upstream and downstream region.

FIGURE 1.1. Basic regions for analysis[4]

Wall functions serve as replacements for the actual roughness obstacles but are intended to
have the same overall effect on the flow[2].Turbulent flows are highly affected due to the
presence of wall where the viscosity-affected regions have large gradients in the solution. In
order to fulfill the mechanics in the near wall area,wall functions are equations that have been
analytically constructed. The universal rule of the wall, which essentially asserts that the
velocity distribution very close to a wall matches up for practically all turbulent flows, provides
the foundation for wall functions[3]. When evaluating the suitability of wall functions, one of

1|P age
the most important factors to consider is the dimensionless wall distance, or y+.If (5 < y+ <
30) then it’s the buffer layer Since the region is complicated and the viscous and turbulent loads
are of comparable magnitude,the original wall functions avoid the initial cell center situated in
this area and the velocity profile is poorly defined[3].For (y+ > 30) the turbulence dominates
very slowly in a logarithmic function[3].After considering these factors,the mesh must be done
considering the mesh quality and boundary conditions.

2 Methodology:

2.1 Geometry
A circular cylinder was located at the center a 2D rectangular domain. Diameter of the
cylindrical shape (D1) was taken 25mm, and the length(H2+H3) and height(V4) of the
rectangular domain were taken respectively 750mm and 500mm. The upstream distance from
the inlet to the center of the cylinder (H2) was taken 250mm, which is 10 times greater than
the characteristic length of the circular cylinder.

250 500

R50

500

FIGURE 1.2. Dimension of geometry(in mm)

2.2 Mesh
The domain was discretized using a structured mesh, as illustrated in Figure 3, with edge sizing
applied to the circular shape and face sizing used for the surrounding domain. The edge of the
circular shape was discretized into 200 divisions. A global element size of 10 mm was applied.
Additionally, 15 inflation layers were included near the circular surface to capture the boundary
layer effects accurately. The thickness of the first inflation layer was set to 1.6 × 10⁻⁴ m,
calculated based on the targeted y+ value for proper resolution of the near-wall flow.

2|P age
 Circular Cylinder

Inlet
Outlet

FIGURE 1.3. Mesh

The required 𝑌 + value should be less than 1 [4]. 𝑌 + value was estimated using the following
equation:
ρUτ∆𝑦1
𝑌+ = (1)
μ

Where, 𝜌 is the density of the free stream, ∆𝑦1 is the wall distance, 𝜇 is the dynamic viscosity
of the freestream and 𝑈𝜏 is the frictional velocity or shear stress velocity.

2.2.1 Mesh Sensitivity Analysis

In Figure 4, with a low number of elements (around 2000-3000), the drag coefficient fluctuates,
indicating that the mesh is too coarse to accurately capture the results. As the number of
elements increases (up to around 6000), the drag coefficient decreases significantly, showing
that the mesh is becoming more refined and the solution is converging. Beyond 16,000
elements, the drag coefficient stabilizes, suggesting that the mesh is sufficiently fine and further
refinement will not significantly affect the accuracy of the results.

3|P age
 Figure 1.4: Mesh sensitivity analysis [Platform: Matlab]

2.2.2 Mesh Quality

For an acceptable mesh, the maximum skewness should be less than 0.94, and the orthogonal
quality should be greater than 0.15 [5]. Based on the mesh quality metrics, the maximum
skewness of the generated mesh was 0.6685, and the minimum orthogonal quality was 0.6478.
These values indicate that the mesh is of good quality, with skewness and orthogonal quality
falling within acceptable ranges for accurate numerical simulations.

2.2.3 Boundary Conditions:

Boundary conditions for CFD analysis:
PAREMETER VALUE
Inler air velocity 1 m/s

Viscosity 1.789×10-5 kg/(m s)

Density 1.205 kg/m3

Gauge Pressure 0 Pa

Reynolds number 1.7×103

Cylinder Diamer 25 mm

VISCOUS MODEL 𝑘−𝜔

4|P age
2.3 Mathematical Model
Conservation of momentum: There are two models, namely shear stress transport 𝑘−𝜔
model and SST 𝑘−𝜔 model which are almost identical and work on the following governing
equations, where 𝜔 represents the specific dissipation rate and 𝑘 gives the turbulence kinetic
energy[6]:
𝜕 𝜕 𝜕 𝜕𝑘
(ρk)+ 𝜕𝑥 (ρkui)=𝜕𝑥 [Γ𝑘 ] + Gk – Yk (1)
𝜕𝑡 𝑖 𝑗 𝜕𝑥𝑗

and
𝜕 𝜕 𝜕 𝜕𝜔
(ρω)+ 𝜕𝑥 (ρωui)=𝜕𝑥 [Γ𝜔 ] + Gω - Y ω (2)
𝜕𝑡 𝑖 𝑗 𝜕𝑥𝑗

The SST 𝑘−𝜔 model is entirely dependent both on the standard 𝑘−ε model and the standard
𝑘−𝜔 model. To blend these two models together, the standard 𝑘−𝜔 model has been
transformed into equations based 𝑘 and 𝜔 that leads to the introduction of a cross-difusion
term D𝜔.

3 Results:

3.1 Aerodynamic Performance:

Figure 5 presents the variation of force coefficients with changes in velocity. At a velocity of
1 m/s, the drag and lift coefficients are 0.02215 and 0.000125, respectively. As the velocity
increases to 2 m/s, the drag coefficient rises to 0.0867, while the lift coefficient increases to
5.675 × 10⁻⁴. For a velocity of 3 m/s, the drag and lift coefficients are 0.1937 and 1.319 ×
10⁻³, respectively. The results demonstrate that the drag coefficient increases with velocity,
while the lift coefficient remains nearly constant, approaching zero. This indicates that the
geometrical shape generates no significant lift. Furthermore, the difference between the drag
and lift coefficients widens as the velocity increases.

5|P age
 Figure 1.5: Graph for the different drag and lift coefficient for different velocity

3.2 Velocity Streamlines:

Figure 1.6: Velocity streamline for 1 m/s flow Figure 1.7: Velocity streamline for 2
m/s flow

6|P age
 Figure 1.8: Velocity streamline for 3 m/s flow

Figures 6, 7, and 8 illustrate the velocity streamlines for different velocities. At a velocity of 1
m/s, the maximum velocity near the upper and lower surface of the circular cylinder reaches
1.34 m/s, increasing to 2.70 m/s at 2 m/s, and 4.05 m/s at 3 m/s. In all three cases, no velocity
streamlines are observed on the behind of the cylinder, indicating a flow separation zone behind
the cylinder where the velocity drops significantly.

3.3 Pressure Contour:

At a velocity of 1 m/s, showed in Figure 9, the highest pressure region is observed at the front
of the circular cylinder, with a maximum pressure of approximately 0.65 Pa. The lowest
pressure, around -0.56 Pa, is found behind the cylinder.

Figure 9: Pressure contour and enlarged view for 1m/s

In Figure 10, for a velocity of 2 m/s, the highest pressure is located at the front of the
cylinder, with a maximum value of approximately 3.0 Pa, while the lowest pressure is -2.3

7|P age
Pa, found at the rear. This demonstrates that the pressure gradient increases with velocity, as
compared to the 1 m/s case.

Figure 1.10: Pressure contour and enlarged view for 2m/s

At 3 m/s in Figure 11, the pressure distribution shows a maximum of approximately 5.92 Pa
at the front of the cylinder, while the minimum pressure reaches -4.91 Pa at the rear. This
further confirms the increasing pressure gradient with higher velocities, indicating a stronger
flow separation and wake formation behind the cylinder.

Figure 1.11: Pressure contour and enlarged view for 3m/s

8|P age
4 Conclusion:

This study analyzed the aerodynamic performance of a circular cylinder using CFD simulation
in ANSYS Fluent. The SST k-omega turbulence model was employed for accurate
representation of turbulence effects. Mesh quality was confirmed to be adequate with a
maximum skewness of 0.6685 and minimum orthogonal quality of 0.6478, ensuring reliable
results. Mesh sensitivity ensures that, the result is not mesh sensitive. The study demonstrated
that the drag coefficient, 0.2215 for the velocity of 1 m/s, which increased with velocity, while
the lift coefficient remained nearly zero, indicating negligible lift generation from the circular
geometry. Low pressure region is observed behind the cylinder, with a noticeable increase in
pressure gradients (0.67 Pa to 5.92 Pa) as velocity increased. These results highlight the strong
influence of velocity on the aerodynamic performance of a circular cylinder.

5 References:
[1] T. Senčić, V. Mrzljak, P. Blecich, and I. Bonefačić, “2D CFD simulation of water injection
strategies in a large marine engine,” J Mar Sci Eng, vol. 7, no. 9, Sep. 2019, doi:
10.3390/jmse7090296.
[2] M. J. T. Loutun et al., “2d cfd simulation study on the performance of various naca
airfoils,” CFD Letters, vol. 13, no. 4, pp. 38–50, Apr. 2021, doi: 10.37934/cfdl.13.4.3850.
[3] “Simscale .” Accessed: Sep. 16, 2024. [Online]. Available:
https://www.simscale.com/forum/t/what-is-y-yplus/82394
[4] “Near wall treatment.” Accessed: Oct. 08, 2024. [Online]. Available:
https://innovationspace.ansys.com/forum/forums/topic/near-wall-treatment/
[5] “How to Verify Mesh Quality in ANSYS Workbench.” Accessed: Oct. 08, 2024. [Online].
Available: https://featips.com/2021/05/07/how-to-verify-mesh-quality-in-ansys-
workbench/
[6] M. H. Masud, T. Islam, M. U. H. Joardder, A. A. Ananno, and P. Dabnichki, “CFD analysis of a
tube-in-tube heat exchanger to recover waste heat for food drying,” International Journal
of Energy and Water Resources, vol. 3, no. 3, pp. 169–186, Sep. 2019, doi:
10.1007/s42108-019-00032-w.

9|P age