CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   STAR-CCM+ (http://www.cfd-online.com/Forums/star-ccm/)
-   -   Airfoil problem (http://www.cfd-online.com/Forums/star-ccm/71456-airfoil-problem.html)

AirWolf January 3, 2010 23:05

Airfoil problem
 
Greetings people.
I had some problem with my simulation.
FYI, I need to find the drag & lift coefficient for NACA 0063.
The velocity of the wind is 3 m/s while the chord of the airfoil is 0.15m. So, I use laminar in this simulation
During the simulation, the graph for lift coefficient is not converged even when I performed at 10000 iterations.
Therefore, I need suggestions to solve this problem.
I hope you guys would share some experience & knowledge about airfoil simulation to me..
thanks

ericnutsch January 5, 2010 18:00

In my experience it usually has to do with not a small enough resolution in grid size or time step.

Is the solution steady state? What is the resolution of your mesh? Is you mesh structured or unstructured?

AirWolf January 6, 2010 00:36

I assume the solution is steady state.
what do u mean by structured or unstructured mesh?
I suspect my mesh is not good enough.
Therefore, I need to redo my work.
What do u think the suitable solution mesh for this airfoil?
The chord is 0.15m and the span is 0.3m.

ericnutsch January 6, 2010 12:40

Structured would be squares or hexahedrals, and unstructured would be triangles or tetrahedrals.

The right resolution is usually the largest number of cells that your hardware can reasonably handle :)

The lower your resolution, the better you have to be at everything else( mesh quality, error analysis, models, solver variables, etc) to get reasonable results. So for starting out, I recommend putting the resolution to the max( 1GB ram = 1million cells max, aprox).

If your solution is 3D I recommend switching to 2D if the assumption can be made. It allow you much more resolution.

Best of Luck!

abdul099 January 7, 2010 08:14

Hi

your simulation has got a very low reynolds number. Maybe there is a problem with laminar separation bubbles? I'm doing simulations on a E603 on a reynolds number of 1e6, and there are still laminar separation bubbles, so it is not really a stationary case.

But i've made the experience, the lift and drag coefficient hits the experimental data very well, even if it seems, the simulation is not converged. Residuals stay the same from approx. 200 iterations, but the forces are meeting the experimental data from approx. 1000 iterations.
I don't know the exact reason, but i'm assuming, it has to do with the normalisation of the residuals...

greetings

kazakage January 11, 2010 23:37

Of course our simulation has lower reynolds number since the velocity only about 3 m/s.
i've some questions that play in my mind, should i running iterations go further until the drag and lift coefficients converge..or should i make some alteration regarding prism layer thickness and meshing settings...thanks


All times are GMT -4. The time now is 03:42.