VSPAero Panel Method

[Jason Fugate]

Hello,

I am currently developing Matlab functions capable of manually triangulating a geometry and then running the panel method solver through VSPAero. I have a few questions about how this procedure works through OpenVSP:

1) How does OpenVSP/VSPAero handle intersections when triangulating? I have attached screenshots of a wing entering the fuselage for the manual triangulation program that I have developed as well as the triangulation generated for an equivalent aircraft built in OpenVSP. It seems as if OpenVSP simply connects the closest "outside" wing section to the fuselage, whereas I manually generate an intersecting section at the surface boundary and instead connect that to the fuselage. Ultimately, OpenVSP's method (at least in this case) leads to strong spikes in the CP distribution and a divergence from expected lift/drag results.

2) I was wondering if there will be a release in the future where the panel method results files include data similar to the data included when running VLM? I would like to be able to determine span-wise lift distributions, and so I was hoping that, at the very least, CP could be included in the outputs so that manual post-processing could be performed.

3) Are there any ways of incorporating control surfaces when using the panel method? My understanding is that VSPAero ignores the DegenGeom file in favor of the Tri file to perform the panel method. Is it possible to use control surface definitions from within the DegenGeom file in the panel method analysis?

Thank you!

Jason

[Rob McDonald]

Good questions Jason,

OpenVSP executes CompGeom and writes out a *.tri file before it runs
VSPAERO. CompGeom starts with the wireframe you see on-screen. It
drops in diagonals to make triangles and then it performs a CSG-style
intersection of the components.

This intersection has the exact wetted surface of the faceted
representation of the two components. However, as you have noted, it
can result in low-quality 'sliver' triangles. near the intersection.

VSPAERO takes triangles as input, but it actually performs
computations on arbitrary polygon vortex rings. These rings can have
many sides.

VSPAERO has an algorithm to improve the input mesh. It has a number
of heuristics and approaches that it uses. For example, it will often
remove the diagonal and merge two neighboring triangles into a
quadrilateral. Similarly, it will detect neighboring sliver triangles
and merge them into a higher quality polygon.

I've attached two images from the VSPAERO viewer. One showing the
surface triangulation and the other showing the computational mesh.
When viewing the computational mesh, you can also use the +- keys to
view the coarsened versions of the computational mesh used in the
agglomeration scheme. In that scheme, groups of panels are merged
into a larger ring vortex for distant calculations.

This quality improvement step tries to reduce the kinds of pressure
anomalies you see caused by the sliver triangles.


Adding a spanwise load distribution output to the panel method is
certainly a good idea. I'm not sure how soon it will get done, but
you aren't the only one who would like to see it.

You might find utility in the 'vspslicer' tool that was introduced a
few versions back. It is totally undocumented (search the mailing
list for an example input file), but it will produce an x,y,z slice of
your geometry and output the Cp along that slice.


Control surfaces aren't yet supported in panel mode. I hope that will
be included in the next version -- very soon.

The subsurface geometry will be impressed on the surface geometry via
the subsurface geometry in CompGeom.

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.
> For more options, visit https://groups.google.com/d/optout.

[Jason Fugate]

Thanks a bunch, Rob!

[Jose Daniel Pérez Muñoz]

Hello,

Are Control surfaces fully supported in panel method in the version 3.20.1?. I am modeling an airplane with more than one control surface. The panel method removes the surface where I implement more than one CS. However, if I place only one CS, it works.

Thank you,

Jose Daniel Perez

> email to ope...@googlegroups.com.

[Rob McDonald]

It shouldn't remove the surface -- except for bad luck in the CompGeom process.

That said, the thick-surface panel method does not yet know how to deflect control surfaces.  So, they aren't really supported.

Rob

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/16a88084-232b-4106-be03-ad93009f32e4%40googlegroups.com.