ANSYS_Fluent_Aerofoil
Budget: ₹2,000 – ₹2,750 INR
Allow experience to be gained in setting up and running a flow computation with a
commercial CFD package.
Enable an exploration to be carried out on the effects and accuracy of different
convection schemes and grid densities.
To make comparisons between predictions using a simple turbulence model and
experimental measurements.
The laboratory exercise allows one to use the Fluent software to compute the flow around
a conventional airfoil model. The flow is 2-dimensional, and should be computed as both
a low-Reynolds-number laminar case, and a high-Reynolds-number turbulent flow.
Task
Grid-Dependence and Accuracy of Discretization Schemes
For these investigations, a laminar flow should be computed at a Reynolds number of 275
(based on chord length and free-stream velocity).
The flow should be computed at an angle of incidence of 10o
. This is set when defining
the inlet boundary conditions.
Because of time constraints, it is suggested that only three grids are examined during the
lab. The computed lift and drag coefficients from the finest grid are included for reference
in the table below.
A solution should be obtained on the coarsest grid using the first order upwind
scheme for convection. Record the values of lift and drag coefficients, Cl and Cd, in
the table below.
Obtain a corresponding solution using the QUICK convection scheme on the coarsest
grid. Note that it will usually be faster to start this calculation from the converged
1st order upwind scheme one, instead of re-initializing the solution; simply go to
Solution Methods in the navigation tree, select the QUICK scheme, and then go
to Run Calculation, without re-initializing the solution.
Repeat the above calculations using first order upwind and QUICK schemes for the
two medium grids.
From the Cl and Cd results obtained, draw conclusions on the level of grid independence achieved using the two convection schemes on the different grids. Is the
behaviour what one might expect from a first order and a third order discretization
scheme as the grid is refined?
For the upwind scheme results, use Richardson extrapolation from two of the grids
to obtain higher order estimates for Cl and Cd. Do these estimates agree with
conclusions you may have drawn on the numerical accuracy of the results?
From three of the upwind scheme results, use Richardson extrapolation formulae
to estimate the order of accuracy of the scheme. Does it give a result close to 1st
order?
Turbulent Flow Calculations
Experimental data have been reported by Nakayama (1985) for a turbulent flow at
a Reynolds number of 1.2×106 and angle of incidence of 0o
. Compute this flow and
compare the results with experimental data.
For this last part, use the grid provided in the airfoil-turb.cas case file. This is a
220x100 grid, but has a much higher near-wall grid density than the earlier grids, to
capture the much steeper near-wall gradients which will be present in the turbulent flow.
To obtain as accurate numerical results as possible, use the QUICK scheme for convection
2
of momentum, although it may be more stable to use the 1st or 2nd order upwind or
power law schemes for the convection of turbulence quantities (k and ε). To account for
the turbulence, it is suggested that the 2-equation k-ε model be used. When applied on
this grid with the Enhanced Wall Treatment selected the near-wall viscous layer will not
be fully resolved by the simulation, but a wall-function approximation will be employed.
commercial CFD package.
Enable an exploration to be carried out on the effects and accuracy of different
convection schemes and grid densities.
To make comparisons between predictions using a simple turbulence model and
experimental measurements.
The laboratory exercise allows one to use the Fluent software to compute the flow around
a conventional airfoil model. The flow is 2-dimensional, and should be computed as both
a low-Reynolds-number laminar case, and a high-Reynolds-number turbulent flow.
Task
Grid-Dependence and Accuracy of Discretization Schemes
For these investigations, a laminar flow should be computed at a Reynolds number of 275
(based on chord length and free-stream velocity).
The flow should be computed at an angle of incidence of 10o
. This is set when defining
the inlet boundary conditions.
Because of time constraints, it is suggested that only three grids are examined during the
lab. The computed lift and drag coefficients from the finest grid are included for reference
in the table below.
A solution should be obtained on the coarsest grid using the first order upwind
scheme for convection. Record the values of lift and drag coefficients, Cl and Cd, in
the table below.
Obtain a corresponding solution using the QUICK convection scheme on the coarsest
grid. Note that it will usually be faster to start this calculation from the converged
1st order upwind scheme one, instead of re-initializing the solution; simply go to
Solution Methods in the navigation tree, select the QUICK scheme, and then go
to Run Calculation, without re-initializing the solution.
Repeat the above calculations using first order upwind and QUICK schemes for the
two medium grids.
From the Cl and Cd results obtained, draw conclusions on the level of grid independence achieved using the two convection schemes on the different grids. Is the
behaviour what one might expect from a first order and a third order discretization
scheme as the grid is refined?
For the upwind scheme results, use Richardson extrapolation from two of the grids
to obtain higher order estimates for Cl and Cd. Do these estimates agree with
conclusions you may have drawn on the numerical accuracy of the results?
From three of the upwind scheme results, use Richardson extrapolation formulae
to estimate the order of accuracy of the scheme. Does it give a result close to 1st
order?
Turbulent Flow Calculations
Experimental data have been reported by Nakayama (1985) for a turbulent flow at
a Reynolds number of 1.2×106 and angle of incidence of 0o
. Compute this flow and
compare the results with experimental data.
For this last part, use the grid provided in the airfoil-turb.cas case file. This is a
220x100 grid, but has a much higher near-wall grid density than the earlier grids, to
capture the much steeper near-wall gradients which will be present in the turbulent flow.
To obtain as accurate numerical results as possible, use the QUICK scheme for convection
2
of momentum, although it may be more stable to use the 1st or 2nd order upwind or
power law schemes for the convection of turbulence quantities (k and ε). To account for
the turbulence, it is suggested that the 2-equation k-ε model be used. When applied on
this grid with the Enhanced Wall Treatment selected the near-wall viscous layer will not
be fully resolved by the simulation, but a wall-function approximation will be employed.