Flow Over Cylinder using Ansys Fluent
OBJECTIVE
PART I
- To run a steady-state and transient simulation of the flow past a cylinder at the Renolds number 100 and calculate the Strouhal number
PART II
- To discuss the effect of Reynolds number on the coefficient of drag by running five different simulation cases with the steady-state solver.
- Reynolds number used 10,100,1000,10000,100000
Overview :
Fluid dynamics is a very interesting and fascinating area of study. Its applications vary from the smallest (single-cell microbes swimming in the water) to the largest (atmospheric boundary layers, not just on earth but on different planets, etc.). One very ubiquitous phenomenon is that of Vortex shedding.
When a flowing fluid interacts with a bluff (or not) body in its path it tends to form a pattern of beautiful looking vortices in its wake. This generally happens at low flow rates, characterized by low Reynolds numbers. For example, when fluid flows past a cylinder, for a Reynolds number > 50 one can start to observe vortex shedding. This happens because the laminar wake flow becomes unstable resulting in separation of the boundary layer. Because of the asymmetric pressure forces in the wake, vortices are shed periodically. Once the flow speed increases the vortices are shed more frequently and forms a repeating pattern this is known as Karman Vortex Street.
The below image shows the simulation to represent how the vortex shedding happens behind the cylinder
GEOMETRY
- A 2d geometry is constructed using the geometry tool in Ansys fluent.
- The geometry is labelled with boundary name
- Using the pull tool surface is created on the whole geometry leaving the circular part
- The circular part represents the cylinder, the blunt object which working object
PART 1
Simulation on Steady and Transient state with Renolds number=100
The geometry used is the above one for both the state.
Mesh Details
- Element Size:0.2
- Nodes:31833
- No of Elements:62714
- Mesh Type: Triangular
2d view of the mesh
As we zoom in to the circular part it can be noticed that the circular part is smooth and has some edge on it, in order to give a smooth finish, follow this
Right-click on Mesh>>Insert>>Edgesizing
No of division:36
And to better capture the physics
Right-click on Mesh>>Insert>>Edgesizing
No of Inflation layers:6
First layer thickness:0.005
Case Setup Details
- Solver Type:
- Fluid: User material
- Boundary Condition
◌. Inlet velocity:1m/s
◌. Outlet: Pressure outlet type, zero gauge pressure
◌. Cylinder walls: Walls type, No-slip condition
◌. Symmetry: Symmetry type
Reynolds number Re=
Details for Steady State
Solution Method: Simple
Solver Time: Steady
Initial Condition: Hybrid initialization
No of iteration:900
Timestep Method: Automatic
Details for Transient State
Solution Method: Simple
Solver Time: Transient
Initial Condition: Standard initialization
Timestep Size:0.1
No of iteration:1600
Results for steady-state
RESIDUAL
It is clearly visible that 350th iteration the solution shows a constant value
VELOCITY PLOT
The velocity of the fluid is calculated at a distance of 8 meters from the cylinder.The point is name is monitor point
The velocity point shows that the fluctuated velocity pattern after 350th iteration becomes constant
By viewing the above plots vertex velocity and Residual plot both profile got constant after 350th iterations
Coefficient of drag
Coefficient of lift
The coefficient drag and lift are 2.5676 and -0.1
Velocity Contour
The vortex shedding starts to shed. A repeated pattern of fluctuation of velocity and velocity distribution is depicted in the contour plot.
The below figure is the simulation video of velocity fluctuation in Von Karman Vortex Street
Pressure Contour
TRANSIENT STATE
Residual
It can be inferred from the graph that the solution of residual showing different upstreams and down making the solutions of equation far aways from the approximation until the good sign is shown at from iteration number 29000 where the pattern becomes constant and thereby converges.
Velocity Plot
The velocity plot portrays the velocity over the flow time in seconds. Right after the 120 iterations, the solution projects a constant value of velocity which concludes that the solution had been converged. As a Transient state has been used in this case, it gives the advantage of providing time accurate results. This paves of an away to find the frequency and Strouhal Number.
The Drag and lift plot
The value of drag and lift are found to be 3.4965 and lift is
Velocity Contour
Above figure showcase, the velocity contour is taken at time step 160 secs 1600 iteration or
The velocity contour display a lot of physical phenomena namely flow distribution, Vertex shedding, Velocity distribution, and eddies movement.
Strouhal Number
Strouhal Number is a dimensionless quantity describing the oscillating flow number mechanism
It is denoted by Sr and the formula is `Sr=(f.l)/U` where f is frequency,l is the characteristic length (radius of the cylinder) and U is the flow velocity
l=2m
U=1m/s
Frequency is obtained by dividing the number of cycles by time taken from the velocity magnitude plot
By reviewing the velocity plot it can be said that stability in the pattern is obtained from 120 flow time to 160 flow time and has 6 cycles
f=6/40=0.15
Sr =0.3
PART II
The geometry used is similar to the of part 1 and the rest of all the steps are the same except that changes are made with a viscosity that would alter a change in Renolds number
The relation is
Here is the list of the velocity incommensurate with the viscosity
No of iteration:900
Results
The plots are arranged in the following manner Case 1>>Case 2>>Case 3>>Case 4>>Case 5
Residual plots
- Clearly the first case took more iteration to finally approach convergence than the rest of all the cases.
- All the other case displayed convergence from the 200-300 iteration and with irregularities in case 5
Velocity Plot
■.The residual plot gave a linkage to velocity as it clearly visible in Case, as it took more iteration to converge so does the velocity
■.Taking about Case 2 and Case 3e solving the great results and among the early one to converge the solution among all, A close look at both plots, a noticeable difference line can ve drawn and that is the repetition of the pattern is way more closer in Case 3 than Case 2 making it form a directly proportional relationship between Renolds number and Vertex velocity
■.The profile of Case 3 shows similar results while in case 5 there a lot of irregularity which shows a sign of turbulence in it
Coefficient of drag
All cases nearly gave a similar profile regarding the coefficient of drag but the oscillation pattern becomes significant after Case 3 and the average of the coefficient of drag keeps in descending from Case I to Cae II.
Coefficient of lift
■.The case I gave no oscillating behaviour while the rest of all the case displayed
■.The only reason contributed towards all case from second to fifth is because of vortex shedding
Velocity Contour
■.As with solid expectation the Case I show no vertex shedding
■.The contour itself provides evidence of presence vertex shedding for all the parameter for all the cases, along with that it can reveal that with an increase in Renolds number the vertex shedding fades. This probably due to irregularities in oscillating behaviours with increasing Renolds number probably indicating giving rise to turbulence in the flow .