drag monitore wrong but in post it's ok!
 

I encountered a problem which has baffled me for a long time. I am using CFX11 to simulate an oscilating airfoil with reference frame. I put the airfoil in a circle region which is oscilating and the outer circle aerea is stationery. I monitored the Fx Fy as drag and lift because the incidence is always horizontal from the far field inlet. However the monitored drag is not right and the lift seems reasonable. When I stop the solver to check it up in the post, the drag turned out to be reasonable. how could this happen? is that because the cfx solver doesn't support the reference frame syntax? if so why the lift is ok. I am carrying out unsteady cases, and now I have to save a lot results for a whole cycle to get the drag hysteresis loop and it's not smooth since the points are limited compared with monitored curves. Could anyone help me? Thanks!

Hi,

Both the solver and Post support rotating frames of reference. They should be giving the same answer.

How are you calculating the forces in the solver and Post? If you are getting the force from an integral of the surface pressure then I can see why the lift is OK but the drag is wrong - you are not including the viscous effects (and the fix is to use the force_x/y/z CEL expressions, not area integral expressions in both solver and post)

Glenn

Thanks Glenn for your feedback. I use CLE expressions to monitor the drag and the lift force force_x()@airfoil, force_y()@airfoil,torque_z()@airfoil. And my rotating domain is pictching with the airfoil under a mean aoa of 15, and the different pitching amplitudes, 10, 5, etc. And the pitching motion is sinusoidal. The outer domain is stationery with horizontal inlet wind speed. I also monitored the instantaneous AOA. The lift and moment are alright, but the drag turned out to be wrong,sometimes it was even negative during a whole monitoring cycle. See here http://img242.imageshack.us/img242/6...ficientwm7.pngThis is not reasonable. The AOA is always above 5 degrees. But when I stop the solver to postprocess the drag using the same expression that I defined in the Pre, it turned out to be reasonable.

I have no idea what is going on. If you can show that the forces returned in the solver and CFD-Post with the same CEL function are different then you may have found a bug. Have you tried it in CFX12?

Glenn Horrocks