Fit Model question

[Mike Vivaldi]

I've watched the videos for using the fit model. I seem to be running into an issue with a fit for inserted sections. If I have inserted a section in a fuselage and go to fit the model using one of the new U sections. Can I still lock the U value in the fit to the value of the new section. For example if I insert a section between 1 and 2 at the U = 0.125 position, can I fix U at .125 like I would for the 0,.25,.50,.75,1.00 position? I tried earlier and it didnt look like the distance measurement was referencing the selected section.

Thanks,

Mike

[Rob McDonald]

The U value of an XSec depends on how many XSecs there are.

If your Fuse / Stack has 4 XSecs

[0 .333 .666 1.0]

If you add one more

[0 0.25 0.5 0.75 1.0]

So, my recommendation with Fit Model is to start by manually getting the model reasonable -- make all your topological choices (do I want a rounded rectangle or an ellipse, how many XSecs do I want to use, what skinning parameters do I want to turn on / off, etc).  Then, use FitModel to adjust parameters to best match.

It is a process, but you'll get the hang of it with some practice.

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/13d3b2b6-8c3c-4b0b-b584-9976bd588a8en%40googlegroups.com.

[Mike Vivaldi]

Ah! Thanks Rob. So U in fit model doesnt correspond with X percentage position. Thank you!

Mike

[Rob McDonald]

Correct.

Throughout U, V corresponds to the Bezier surface parameter -- which depends on how the surface was constructed.

Most user-facing uses are U,V on a scale of [0,1].  In the code, W is often used instead of V for $reasons.

Inside the code, U,V runs on a scale of [0,N] where N is the number of cross sections -- and V runs [0,4] in almost all cases.  Inside code, the U variable will often be written U01 if it is in the user-scale.

Somewhat recently, U [0,N] was exposed to the GUI as an option.  That way (in the places that support it), you can set U0N = 4 and it will always correspond to the 4th XSec -- even if XSecs are added / deleted.

There are other component local coordinate systems that are used other places, but these did not exist when FitModel was developed.

R,S,T is a volumetric set of coordinates.  R behaves the same as U.

L,M,N is a volumetric set of coordinates, but L is scaled to be linear along the length of the object.

eta,M,N is a volumetric set of coordinates (only on wings) where eta is the fraction of the span measured perpendicular to the construction planes.

Rob

[Brandon Litherland]

The 2025 OpenVSP Workshop has a great presentation by Rob covering surface representation in detail.  The slides for surface modeling are here and the video is here.

I've included a screenshot of slide 20 where an illustration of an "unwrapped" wing is provided. 

Fit Model is, indeed, a process as Rob mentioned but it's extremely useful and robust _if used properly_.  Our usual mantra of using as few sections as possible to get the shape you want may inadvertently push you to use too few.  For example, if you set a few XSec at your obvious hard fuselage breaks where the angle changes like a corner or edge and then try to fit model a section of cabin where there is an inflection in the surface between those XSec, it will likely be difficult to get what you need.  Adding a section at the inflection helps but you will have to go back to Fit Model and update all of your UW locations.  Personally, I find grabbing targets to be the hardest part of this process but there are some tricks that can help with that.  Also remember that you can save a Fit Model problem and load it back in.