[kcn]

Hi,

I am simulating flow around an airfoil. For low angles of attack I got results for Cd and Cl that very closely match with experimental values. I used RANS and simpleFoam solver with k-omega SST model.

However for some future work, I need to perform the simulations in transient mode. So I used the same mesh with same parameters that worked well in steady mode and used pisoFoam solver for unsteady simulation.

The flow becomes steady after a certain amount of time, as observed on cd vs time curves. However it converges to a wrong value. Cd is about 50% higher than the correct value obtained in steady simulation.

Can anyone please help me with this?

Thanks

[vatavuk]

Hi kcn,
There is a nice paper by C.L. Rumsey, "Apparent Transition Behavior of Widely-Used Turbulence Models", in which he deals with the possibility that certain turbulence models may supply different solutions if they use different initial conditions or numerical methods. This is caused by the presence of a numerical laminar region that may have a variable size.
This phenomenon may be happening in your computations. To understand what’s happening, I suggest that you try to use the simpleFoam solution as the initial condition for the transient solution and vice versa.
Best Regards,
Paulo

[kcn]

Dear Paulo,

Thanks a lot. I'll go through the paper and try your suggestion also.