Drag Coefficient VS Average Drag Coefficient

[keager]

Can someone explain the difference between the 2 Drag results in Flow? Drag coefficent VS Average drag coefficient. I can understand that it's a transient flow, but if I need to tell someone what the CD is for their conceptual form, which result should I be telling them? I assume I'm to be communicating the 'Average drag'.

Thanks.

[keager]

*Update:
The form is a vehicle and I've had it running for 2 hours now. Resolution is at 190, windspeed 28m/s. It stabilized after about 45 minutes and had an average drag of 0.17. I let it run for another 30 minutes and it appears to be fluctuating from 0.17 up to 0.21. Why does the average fluctuate, is the average of that supposed to be between 0.17 - 0.21 (0.19)? The drag coefficient # appears to have remained the same at 0.13 though. Any explanation of this above post is appreciated. Thanks.

[hartogj]

Flow Design solves transiently with every run. This captures some behavior you don't normally see when assuming steady state flow, for example vortices and other disturbances that are inherently time varying. It is calculating behavior at sucessive timesteps and the Cd that is reported is the drag coeff at that particular timestep - sort of an instantaneous measure. The average Cd looks back over a lot of recent timesteps and takes an average of all the instantaneous Cd's, smoothing over the peaks and valleys for you.

Note that, for most models, when things first kick off the Cd is pretty high. This is when air first starts being pushed through the tunnel. You'll see the instantaneous Cd settle down over the first few iterations. You typically see that reflected in the moving average later on once the simulation has been run long enough.

After many timesteps, the average Cd should be right in the middle of your most recent instantaneous Cds. Did you let this run out further to get to that point?

Jon

[keager]

Yes, I let it run overnight a few times. I appreciate the explanation and it's sort of as I assumed. Only thing, using the exact same model it seems like it can vary up to +/- 0.02 on each run even after running several hours (even over night) and the model is a pretty basic car shape. That's not a lot, but when we're talking 0.15 VS 0.19 it makes a difference. I've got the res at 190. Perhaps it's because I'm having to set the floor height each time, but I'm zooming right in and placing the floor level right at the base of the tires so that difference seems high. It would be nice if the default floor height was right at the base of the model instead of floating below.

I understand this is for conceptual use and it's served it's purpose well. I'm also not claiming to be an aero engineer, so I'll go with what I know and leave the detail to the experts. Thanks for your reply, it did help!

Cheers.