Hello guys ,
I am making a 2D Airfoil simulation Using pimpleFoam solver and the K Kl Omega transient turbulence Model . As i know, in a transient Simulation we have to pay attention to the Co (CFL) and it has to be less than 1 and we should be okk.
I run my simulation 2 times same mesh ,same boundary conditions and same evrything
the only difference is in the
maxCo in the controlDict.
for the first simulation i used maxCo 0.2 and for the second simulation i used maxCo 0.9
surprisingly i got different values of Cd and Cl
simulation1: Cd=0.0583 Cl =0.2747 simulation2: Cd= 0.01329 Cl=0.3424
Why is that ? and what is the reason behind this big diffrence in results?
this is the content for my control dict file:
application pimpleFoam; startFrom latestTime; startTime 0; stopAt endTime; endTime 3; deltaT 0.001; writeControl adjustableRunTime; writeInterval 0.01; purgeWrite 0; writeFormat binary; writePrecision 6; writeCompression off; timeFormat general; timePrecision 6; runTimeModifiable true; adjustTimeStep yes; maxCo 0.2; functions { forces { type forces; libs ("libforces.so"); writeControl timeStep; writeInterval 10; patches (wing); rho rhoInf; log true; rhoInf 1.204; CofR (0 0 0); } forceCoeffs1 { // Mandatory entries type forceCoeffs; libs ("libforces.so"); patches (wing); // Optional entries // Field names p p; U U; rho rhoInf; rhoInf 1.204; // rhoInf 1.204; // Reference pressure [Pa] pRef 0; // Include porosity effects? porosity no; // Store and write volume field representations of forces and moments writeFields yes; // Centre of rotation for moment calculations CofR (0 0 0); // Lift direction liftDir (0 1 0); // Drag direction dragDir (1 0 0); // Pitch axis pitchAxis (0 0 1); // Freestream velocity magnitude [m/s] magUInf 10.679; // Reference length [m] lRef 1; // Reference area [m2] Aref 1; // Spatial data binning // - extents given by the bounds of the input geometry } }
First of all, a change in the CFL means a change in the time or space computational steps. In your case you are increasing the time step, a fact that has an influence on the action of the numerical errors.
Then, do you use a transient solver to compute a steady RANS solution?
Filippo Maria Denaro
I am using OpenFoam 6
the solver am using is pimpleFoam which is also a transient solver.
also for openFoam ,it gives you the ability to control Courant number to get convergence .
also the solver is controlling the time steps as i said in order to control Co and keep it below the value i want which was in the 1s simulation 0.2 and 2nd simulation 0.9.
so basicly the solver was manipulating timesteps but at the same time keeping Courant number
Hello Oussama Brahimi
Based on definition of Courant number, it is relative to velocity, time step and mesh size,
CFL = u*dt/dx
Then when you decrease CFL number, pimpleFoam solver automatically change time step to satisfy modified CFL number and depends on simulation conditions, maybe result could be different. That means perhaps your mesh is not fine enough (in this case specially boundary layer mesh) to calculate correct Cd and Cl.
In addition, if you are using pimple solver with nOuterCorrectors = 1 in FvSolution, then the solver works like Piso solver. The pimple solver normally used for large time steps(CFL > 1). So in order to increase stability you have to add more corrector step (nOuterCorrectors and nCorrectors) which is not the object of your simulation.
My advice is to use K-Omega SST as turbulent model and very fine mesh (y+
Mahdi Rezaei Adariani Thank you for Your Answer . I will try To make the simualtion using the K-OmegaSST turbulence Model .
For the Wall functions Should i remove them for nut omega and epsilon or only for nut ?and how can i remove them?
Thank you
Oussama Brahimi
Yes, you have to modify all of them. A clear answer is here:
https://www.researchgate.net/post/Should_the_RANS_model_kOmegaSST_be_fully_resolved_at_low_Re-number_Is_it_possible_to_use_it_with_a_wall_function
Again pay attention, k-omega SST is too sensitive to mesh size. Near the airfoil y+ could be less than one.
Regards
Mahdi Rezaei Adariani One last Question
My Re =700 000 is this okk with the KOmegaSST turbulence model Or there might be a better turbulence for a relativly hight Re value?
Oussama Brahimi
In my idea, the turbulence model is OK. However, based on this Re number, your mesh near the walls of airfoil could be around 3e-5 or less.
I have asked a relative question a few months ago. You can find useful information in the answers there. Here it is:
https://www.researchgate.net/post/Does_wall_function_has_an_effect_on_the_result_of_aerodynamic_forcesDrag_or_lift_and_separation_point_in_the_simulation_of_flow_around_a_cylinder
Best regards and good luck
- Badges
- Science topic
- Similar topics
- Force