spikes at wingtips in panel method solution

[Cindy Chen]

I am running a panel method analysis of a BWB. The case setup and solver parameters are attached. The tip cap is flat. However, there are sharp spikes in the spanwise lift distribution, shown in a screenshot attached. Also attached is the geometry with the mesh overlay.

Has anyone ever encountered this issue?

[Rob McDonald]

I suspect you have not turned on the 'Rotate airfoils to match dihedral' option.  You'll want to turn that on.

Rob

[Cindy Chen]

Hi Rob, I did turn that on in my case.

[Rob McDonald]

Post your *.vsp3 file if you want anyone to be able to help.

Rob

[Cindy Chen]

[Rob McDonald]

Something is wrong with your airfoil files.  It is causing strange behavior around the trailing edge.

I switched all the airfoils to NACA 4-Digit and got a significant improvement.

Then, you have a ton of control surfaces.  These will cause spikes in the load distribution -- but those spikes aren't really important, they are just a bookkeeping thing.  However, to make sure they would clean up, I turned off all the control surfaces and got this result.

You still have a lot of ugly jumps in load distribution across the center body.  This is mostly a post-processing thing getting the load distribution plots (not a problem with the underlying geometry).  You can see that the pressure distribution looks much better.

You could greatly improve the center body stuff by eliminating most of those cross sections and working to get a smoother and cleaner loft.

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 visit https://groups.google.com/d/msgid/openvsp/d5ec2f94-eaed-4bf7-9ac8-e03f92a1b0e9n%40googlegroups.com.

[Cindy Chen]

Hi Rob,

Thanks for looking into this. Did you switch all of the airfoils at the same time or one by one? If one by one, do you remember which airfoil was causing issues? When I plotted the coordinates in MATLAB, they looked fine to me.

Cindy

[Cindy Chen]

Also, to follow up, is there any way to get rid of the spikes in the load distribution completely?

[Rob McDonald]

I did not go one by one, I changed them all.

I don't care what your airfoils look like in Matlab, you aren't using them in Matlab.

If you would actually look at your model, you would see that the trailing edges are hosed.

I suspect you have points wrapping around a blunt trailing edge and this is how OpenVSP is trying to make sense of it.

If you have airfoil points with a blunt TE, you should not have any points down the blunt face.  Your points should start at the lower TE point and end at the upper TE point.

VSPAERO in thick mode might not like having a thick TE.  You may end up sharpening your airfoils in OpenVSP anyway.

Rob

[Rob McDonald]

Start by deleting a bunch of the sections -- your center body is way over determined.

Your chord distribution is lumpy -- because you've over-determined the center body.

Start by getting rid of all the intermediate body cross sections -- just start with the center line and the one at the root of the straight wing section.  Get that working well.  Then, _if_ you need more sections, add them judiciously, one at a time, adjusting the parameters to maintain a smooth shape.

Rob