hello,

i'm trying to simulate 2d pitching/heaving airfoil. I've verified my code with oscillating cylinder and it works.

next i tried to verify my code against some published results on pitching/heaving airfoil. the problem is that although the cl compares relatively well, the cd does not. some are close but others are quite far off by 50-100%.

i wonder what could be the problem. i've tried to reduce the time step as well as refine my grid but the result does not change much.

at this moment, i wonder how i should proceed. can anyone give me some advice?

thank you.

How are you computing drag? Are you comparing total drag, pressure drag, or viscous drag? Is the flow laminar or turbulent? If turbulent, what turbulence model are you using? Are you sure the turbulence model (if any) is suitable for massively separated flows? How many grid points do you use to capture the boundary layer at the airfoil surface?

thank you for your response.

i'm computing drag/lift using numerical differentiation/integration, 2nd order ie get du/dy,dv/dx etc at the surface of the wall and sum them up. i can also compute using tecplot and the answers are similar.

both pressure/viscous drag are compared and they're both different from the published result.

Re=1100 and it's laminar flow. i used the "grid spacing calculator" to find an estimate. at Re=1100, spacing=0.01 units. My fine grid is around 0.005 from wall.

i hope the information is sufficient. thank you.

When you say that Cl compares reasonably well, what kind of agreement are you actually getting? Your lift is going to be dominated by the pressure distribution, which is also repsonsible for the pressure drag that doesn't match. Have you compared your pressure distributions to published results? Remember that matching lift and drag doesn't mean a whole lot - since they are integrated quantities there are an infinite number of pressure and shear stress distibutions that can generate the same Cl and Cd. I would try to find a test case that provides pressure data as well as lift and drag and look at more detailed comparisons to see if there may still be a problem with your code.

tks for pointing it out, ag!

i'm also wondering too. in fact, for both lift/drag, they are dominated by the pressure, instead of viscosity in my simulation case. the contribution of viscous force is relatively smaller. hence if lift compares well (less than +-5%), drag should too since they are just components of pressure.

the published result only 've the average cp for some cases. maybe i'll try to look for more data.

but i just can't find the reason...