Is this what NASA actually use?

[Nathan Van Ymeren]

Neat tool!

Is this what NASA actually use to generate models for analysis?

Nathan

[Rob McDonald]

NASA is a big place, with a lot of different engineers using different
tools for different jobs.

VSP's direction and development was guided by some of the conceptual
design groups in NASA. Those groups use VSP extensively. As VSP has
gained more engineering capability, more groups inside and outside
NASA have begun adopting it for their needs.

Rob

[Nathan Van Ymeren]

I guess I should have been more specific.

I haven't had a chance to play with it yet: Does VSP have the capability to create meshes with enough fidelity to run CFD/FEA routines?

[Rob McDonald]

That depends on what you need to do.

For FEA, see Armand Chaput from UT Austin's paper at last week's ASM
conference. He and his team are developing tools based on the
structural modeling capability of VSP. They are going to FEA. Their
work is supported by NASA.

For CFD, the answer is certainly yes, but with some conditions.

A number of groups at NASA are using VSP to feed CFD in the form of
Cart3D. Search for papers recently by Irian Ordaz, Wu Li, and Andy
Hahn, all at NASA Langley

One group is working on low-boom design for supersonic aircraft. The
other on engine integration for transonic transport aircraft.

VSP does not create volume meshes, so any CFD workflow must be able to
handle that outside of VSP. We have two solutions right now.

Cart3D is a Cartesian mesh code which generates its own volume mesh.
Its surface meshing requirements are somewhat unique -- and VSP is
very well suited to them. VSP can export three logically different
files in native Cart3D format. If you're really interested in Cart3D,
I can get more specific. Cart3D is available from NASA if you're US
government or Desktop Aeronautics for everyone else. VSP/Cart3D has
proven to be a very productive tool set.

VSP can write out files which can be used by TetGen
(http://tetgen.berlios.de/) to generate volume meshes. If you start
with a watertight mesh from VSP (usually not a problem) then TetGen
seems to work very well. TetGen does not seem suited for generating
layers in the boundary layer suitable for a NS mesh. I do not think
anyone has used this toolchain to drive an Euler code, but it should
be very doable.

VSP could write out files in NASCART format. NASCART is a NS
Cartesian code developed by Steve Ruffin at Georgia Tech. I have not
used NASCART in about six years, so I am not sure if things have
changed which would prevent that from working.

VSP has also been used to feed more traditional CFD toolchains by
writing out Bezier surfaces to Rhino and then feeding them through the
traditional mesh generation and CFD process. Usually at NASA Langley,
this has been done with VGrid/Usm3D as the targets.

With all that said, VSP is the starting point for only a tiny fraction
of the CFD work that happens at NASA. Most of NASA's CFD work starts
with traditional CAD.

Rob

On Fri, Jan 13, 2012 at 9:20 AM, Nathan Van Ymeren

[Nathan Van Ymeren]

Great response; thanks for taking the time to type it.

Sounds like vsp is more than meets the eye. I look forward to seeing it in action!