Professional Documents
Culture Documents
The following files must be downloaded from the CFD web pages:
aerofoil.exe (user interface)
streamaero.exe (CFD solver)
gridaero.exe (grid-generator)
Start by double-clicking the graphical user interface. All files will be saved in the folder from
which you run the program.
The program simulates 2-d, incompressible, laminar or turbulent flow around an aerofoil. The
approach-flow velocity U0 is uniform. The angle of attack may be varied.
y
thickness
camber
α x
U0 chord, c
The aerofoil section may be any member of the NACA 4-digit series
NACA mpth
where
m = maximum camber in percentage of chord
p = position of maximum camber in tenths of chord
th = maximum thickness in percentage of chord
1.3 Non-Dimensionalisation
All variables are non-dimensionalised using the approach-flow velocity U0 and the aerofoil
chord c. The Reynolds number is defined as Re = U 0 c/ .
The program outputs to screen the equation residuals and, at both backup and completion:
• the upstream stagnation point and any separation or reattachment points;
• minimum and maximum y+ values (see the lectures on turbulence modelling);
• drag and lift coefficients, based on force components (per unit span) parallel and
perpendicular to the approach flow:
drag lift
cD = 1 2
, cL = 1 2
2 U0 c 2 U0 c
Grid
[Set up laminar] set grid and aerofoil parameters for the default laminar flow
[Set up turbulent] set grid and aerofoil parameters for the default turbulent flow
[Edit parameters] edit grid and aerofoil parameters
[Run] run the grid-generator
Case
[Set up] set default flow, transition and flow-specific plot parameters
[Edit parameters] edit case parameters (including angle of attack)
Solver
[Set up laminar] set parameters for a default laminar-flow calculation (Re = 103)
[Set up turbulent] set parameters for a default turbulent-flow calculation (Re = 106)
[Edit parameters] edit individual solver parameters
[Run] run the CFD solver
Plots
[Set near] set parameters for a default plot focused on the aerofoil
[Set far] set parameters for a default plot showing the whole domain
[Edit parameters] edit general plot parameters
[Grid] plot the grid
[Streamlines] plot streamlines
[Pressure] plot pressure contours
[Turbulent KE] plot turbulent-kinetic-energy contours
[Vectors (all)] plot mean-velocity vectors at all nodes
[Vectors (regular)] plot interpolated mean-velocity vectors on a regular grid (**)
[Profiles] plot streamwise-mean-velocity profiles along the aerofoil (**)
(**) Regular-grid velocities and profiles are only output at the end of a flow calculation.
Graphs
[cp] plot a graph of pressure coefficient
[cf] plot a graph of skin-friction coefficient
Hard copy
[File] save the current plot as a picture file (type .png)
The [Edit parameters] buttons launch menus for grid, case, solver and plot parameters. The
program carries out some checks to exclude unreasonable values, but it is not foolproof.
Grid Parameters
Aerofoil parameters
m maximum camber in percentage of chord
p position of maximum camber in tenths of chord
th maximum thickness in percentage of chord
Size of domain
xMax downstream extent of grid
Radius upstream extent of grid from leading edge
Numbers of cells
NChord number of cells along the chord of the aerofoil
NWake number of cells downstream of the aerofoil
NRadius number of cells in a radial direction
Elliptic-grid-generator parameters
Maximum number of iterations
Tolerance for convergence of the grid generator
Over-relaxation parameter
There are hard-coded limits in the flow solver for the maximum numbers of cells.
Case Parameters
• Angle of attack, .
• Laminar-to-turbulent transition points on the aerofoil surface (as percentage of chord).
• Free-stream turbulence intensity and turbulent-to-molecular viscosity ratio.
• Presentation grid (regular rectangular array for interpolated velocity vectors); specify
the first and last coordinates and the number of points in each directions.
• The length of the lines along which velocity profiles are computed.
• Parameters for the pressure-coefficient (cp) and skin-friction-coefficient (cf) graphs.
Laminar Flow
Set up and generate the default grid for a laminar flow. Include suitable plots of the grid in
your report.
Set up the default case and laminar-flow parameters. Calculate the flow and include plots of
streamlines, shaded pressure contours, velocity profiles (not vectors) and the cp graph in your
report.
For the calculation above record the drag coefficient cD. Repeat grid and solver calculations
for a grid with double the number of grid cells in each direction (i.e., double NChord,
NWake and NRadius and re-run the grid generator and flow solver). Does your solution
give a satisfactorily grid-independent value for drag? Compare your results for drag
coefficient with the Blasius theory for a flat plate of similar length:
1.33
cD = (per side)
Re
Suggest reasons for any differences.
Returning to the default grid and aerofoil section and default laminar flow parameters (with
Re = 1000) compute the flow at angles of attack 3° and 6°. In each case:
• record drag and lift coefficients (as reported by the program);
• record any separation point (as reported by the program);
• plot streamlines, shaded pressure contours, velocity profiles and cp graphs.
Note that the program reports all points of flow reversal. You should exclude any
corresponding to stagnation or reattachment points.
Explain how flow separation can be identified from the skin-friction graph and how lift can
be estimated from the pressure-coefficient graph. Why is this only an estimate?
With the default grid and aerofoil section calculate the flow around the aerofoil at an angle of
attack 6° with Reynolds numbers of 200 and 5000, recording drag and lift coefficients and
any separation point. What effects does the Reynolds number have on flow separation and on
the drag and lift coefficients and why?
Generate the grid with default cell parameters but cambered aerofoil (NACA 2309). Compute
flow with Re = 1000 and angle of attack 3°. Repeat for a thicker aerofoil (NACA 2312).
Record the drag and lift coefficients in each case and, by comparison with your earlier
symmetric-aerofoil calculations, comment on the effects of camber and thickness.
The code produces a data file surface.dat containing columns with the following data:
x/c y/c cp cf x/c y/c
where x and y are cell-face-centre coordinates on the (Outward) areavector
surface of the aerofoil, and ( x, y) is the A = ( Ax , Ay ) = (− y, x )
corresponding surface line segment. cp and cf are
pressure and skin-friction coefficients defined by:
p − p∞
cp = 1 2
, cf = 1 2
2 U0 2 U0
is the shear stress acting tangentially to the surface
in the direction of x, which here corresponds to a
τ
y
clockwise traversal of the aerofoil. For a NACA
x
2312 aerofoil, at Reynolds number 1000 and angle
of attack 3°, show how to use this data to produce AEROFOIL
drag and lift coefficients for the aerofoil, separating
each into pressure and viscous components.
Confirm that the overall drag and lift coefficients yield the values reported by the program.
Turbulent Flow
Set up the default grid parameters for a turbulent flow. Run the grid generator. Set up the
default turbulent flow (Re = 106; Launder-Sharma turbulence model).
For the default symmetric aerofoil (NACA 0009) compute the flow for angles of attack
= 0°, 3°, 6°, 9°, 12°. Plot graphs of lift and drag coefficients against angle of attack,
comparing your lift data with theory for a thin-chord symmetric aerofoil (White, Chapter 8):
c L ≈ 2 sin
Repeat the calculations and graphs in T1 above for the cambered NACA 2309 aerofoil. This
time compare the computed lift coefficient with
2h
c L ≈ 2 sin( + ) (h = maximum camber; angle in radians!)
c
For the cambered NACA 2309 aerofoil and angle of attack 3° try computing the flow with the
following turbulence models (adjusting under-relaxation factors if necessary):
• Linear eddy-viscosity models:
– Launder-Sharma k- (the default);
– Menter’s SST k- model
• Non-linear eddy-viscosity models:
– Craft-Launder-Suga
– Apsley-Leschziner
• Reynolds-stress model
– Hanjalic and Jakirlic
In the last case you will have to reduce the turbulence under-relaxation factor (substantially).
If you wish to see all the equation residuals then you will also have to increase the width of
the command window showing these – right-click on its header bar, then choose “properties”
and the “layout” tab.
For each turbulence model record results for drag and lift coefficients and plot velocity
profiles (not vectors). Comment on how these vary between turbulence models (if at all).
(T4) Transition
Using the aerofoil from T3 above, a 3° angle of attack and the Launder-Sharma model, set the
transition points at 40% chord on the lower (“pressure”) surface and 30% chord on the upper
(“suction”) surface of the aerofoil. Note and explain the effects (if any) on drag and lift.