Hi All, I need help! I asked my friend CFX5.5 user to analyse a racecar model. He did the solution with fine mech of "half car" (model is simmetrical). Problem is that numbers we received are unreal: 18000Nm of downforce and 700Nm of drag at 100mph! I wish it was true but it can't be so. I have his command file, images prodused and results. May somebody could take a look and find out what's wrong? Also I'd like to know how do you find Pitch moment in CFX?

Thank you Ted

Hi Ted,

For the pitch moment, see another post of mine:

http://www.cfd-online.com/Forum/cfx.cgi?read=7015

As for what is wrong with the simulation, there could be many things wrong. Was adequate convergence acheived? Was it steady state of transient? Was a grid refinement study done? Were the boundary conditions properly applied?

Regards, Glenn

Thank you Glen, I understand that there could be many things wrong. Sergey - the guy hwo works on it says that everything should be fine i really don't know what to do.... Can I ask you permition to send you his command file or better yet ALL the file from Build? This way you can look for error and may be point us in right direction. I'm really interested in proper results as a lot of other disign aspects (chassis, suspension geometry etc) are stuck because of this. What I'm trying to design is a car with underbody tunnels which soopose to give less pitch/roll/rideheight sensitive downforce than conventional flat bottom with diffusers design.

Thank you Ted

Hi Ted,

I suggest you mention to Sergey that if the result is wrong there's a fair chance there is something wrong in the simulation.

The best way for people on this forum to try to pick what is wrong is for you to post your output file onto the web. If it is short then put it here on CFD online, but if it is long either put it on another site and put a link to it, or just post the significant parts (the initial CCL & CEL, the convergence, any error messages etc). Also of use are pictures of the important parts of the mesh.

Regards, Glenn

Ted did you sort out the problem ? How many elements did you use for the simulation ? As Glenn said you need to do a mesh refinement study, adding more and more elements to the model (usually doubling the total number), to check to see if the results do not change. When the results do not change then the model can be said to be mesh independent. For a complete car, I would imagine that you are looking at a mesh greater than 10 million elements. This may be on the low side as I know a lot of the car manufacturers run models on large clusters of PC's. Bob