Lift and Drag Coefficient off by two orders of magnitude

Hello!

I am performing an external aerodynamic analysis of a model of an aircraft. It is a 1:11 scale model and is simulated in a small wind tunnel based on the NASA Langley 14x22 where I basically extruded the test section and applied a slip wall condition as described in the actual CFD analysis performed by NASA. All boundary conditions reflect their experiment and I tried to match the Reynolds number by adjusting the kinematic viscosity. It is also also a symmetric analysis and runs on the k-omega SST turbulence model.

After three successful simulations with a rather coarse grid compared to what they used (19million cells as opposed to 114m cells), the results for my lift and drag coefficient seem to be a very close match, with the exception that they’re off by two orders of magnitude, eg. Cl = 0.009 instead of their Cl = 0.84. I suspect it’s something to do with the scaling but the model size is the same as in both their experiments and CFD simulations. The same applies to my drag coefficients.

From my understanding and previous experience, these coefficients shouldn’t be that small anyway, so do you have any ideas what could be the cause?

Thanks

Jakub

Hi Jakub
very interesting results.
Did you make sure that the Reynolds number within your experiment and the date are the same.
Since your simulation is to the scale of 1/10 your velocity has to change as well to match the Reynolds number.
Which could lead to an error, with order of magnitude of two, since velocity is effecting in that order as well to drag.

Best regard Sebastian

Hi Sebastian,

In the most recent analysis I calculated the Reynolds number by hand to make sure it’s the same as in their CFD and adjusted the viscosity to match it but the results are very close to my previous runs which had the default air properties.

The issue I had with their Reynolds number is that they used a dimensional form, namely, Re per inch and the way interpreted it was that they just divided it by the characteristic length. Their CFD Re was 44000/in. It is not specified what characteristic length they used so I assumed it to be the length of the aircraft model. I can also use the wingspan but I doubt it will change the results by a significant amount.

I’m starting to think they’re somehow converting their CFD results to reflect the full-scale aircraft and these are the results in their reports but it is not mentioned anywhere so it’s difficult to say.

Regards,

Jakub

Hi @jpyszka,

One question: how did you determine the reference area used in the calculation of the lift & drag coefficients?

Hi Jakub, please share the paper and I will take a look at it.

Hi,

I calculated the wing area as the reference length and this seemed to be the issue. I found an older paper where the exact reference dimensions were provided and now the results match up in terms of the order of magnitude.

I have a couple more things to tweak in this model but this problem is solved.

Thank you,

Jakub

Ohh, OK, I am happy to hear that.