Negative CDi when moving from 3.23.0 to 3.25.0

[lsol...@gmail.com]

I'm having some issues with negative CDi values when moving from OpenVSP version 3.23.0 to 3.25.0. I created a generic wing to highlight the issue. This wing uses the default wing shape with 'U' increased and the default NACA6 airfoil. The same wing file was run using panel mode in 3.23.0 and 3.25.0 without any changes to the file. There are some slight differences between CL and Cm but CDi is off by quite a bit. I saw a few other people in the forums were having trouble with negative CDi values but using the advice posted in those forums

https://groups.google.com/g/openvsp/c/YbNF7cHNeQY/m/NFceWl4vCAAJ

https://groups.google.com/g/openvsp/c/Mt22QS05udo/m/KSEsl28nAQAJ

did not change the results significantly. The changelog shows that VSPAERO had some changes between 3.23 and 3.25 so I can only assume that this maybe causing the problem. I can change the results drastically (positive CDi) by changing the airfoils but unfortunately the airfoils I need are the ones causing negative CDi. Any ideas what might be causing this issue?

[Rob McDonald]

I'll share this case with Dave, maybe he will have some suggestions.  It looks like the issue is related to the thin cusp-like trailing edge.

I'm not sure how complex your eventual model is, but I would generally recommend you use the VLM model for this -- it doesn't exhibit the kinds of issues that the panel method does.

Rob

--
You received this message because you are subscribed to the Google Groups "OpenVSP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openvsp+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/openvsp/c917f0cf-59dc-4ab5-b495-2515a9b6f423n%40googlegroups.com.

[lsol...@gmail.com]

Thanks Rob, unfortunately fuselage effects are non-negligible and the panel method has shown a much better match with CFD for my case.

[lsol...@gmail.com]

Rob, were you able to get any insight from Dave about the negative CDi issue?

[Rob McDonald]

Sorry for not coming back...  Here is what he said...

"At first look... I don't think it likes the cusped airfoil. This is a typical issue with panel solvers and I've yet to implement a more robust kutta condition that would probably solve this... add it to the list.

Note that the code still writes out a "CDtrefftz" value which is what it used to write out as Cdi... now Cdi is calculated using KJ theorem over the edges.  For wings and such it's worth looking at that, especially for panel solves, to get an idea of how your doing..."

Rob

To view this discussion on the web visit https://groups.google.com/d/msgid/openvsp/a922805e-ec17-4dd1-b67e-42ea7eed7aedn%40googlegroups.com.