Closing trailing edge

[Alfredo Escolano Lopez]

Hello !
I am doing a geometry of a propeller defining the profile on different sections. Those sections are defined in a .txt and it is a closed profile. However, when creating the stl, the geometry is not completely closed , as shown on the image. I 'd like to know how can i close the geometry. 
Thank you, 

Alfredo

[Rob McDonald]

Try it with a coarser mesh.

I suspect that MeshLab is merging nearby points -- and is creating the
hole on import. A coarser mesh is a quick way to test this idea. The
edge lengths should be above MeshLab's tolerance.

Past that, please post an example file, we can't experiment with an image.

Rob

On Thu, Jul 23, 2026 at 7:36 AM Alfredo Escolano Lopez
<alfredo...@gmail.com> wrote:
>
> Hello !
> I am doing a geometry of a propeller defining the profile on different sections. Those sections are defined in a .txt and it is a closed profile. However, when creating the stl, the geometry is not completely closed , as shown on the image. I 'd like to know how can i close the geometry.
> Thank you,
> Alfredo
>

> --
> 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/f5209476-aebf-47a2-a611-10ea8afea274n%40googlegroups.com.

[Alfredo Escolano Lopez]

I have tried to do a coarser mesh. However, the problem remains the same. I attach the vsp file i am working on, thank you.

Alfredo

[Rob McDonald]

You're defining the prop with 67 cross sections defined by File Airfoils -- that are themselves badly formed.  This is a disaster waiting to happen.

The blue section here is XSec #43.  Notice that it (and many of the others) does a self-intersecting loop?  That is bad.

This is not how to use OpenVSP.  Use as few XSecs as possible -- let OpenVSP do the blending and interpolating between XSecs.  Use as few points as possible to specify the chord and twist distribution.  If you read them in via a BEM file, then use the approximation capabilities to reduce the number of points specifying the curves.  Fix your airfoil files so they aren't total garbage.

Rob

[Alfredo Escolano Lopez]

Hello another time Rob. 

The files of the profiles are not badly formed, in fact it is a profile that has been written via a matlab script and gives the form attached on the image ( points joined with straight lines) . Furthermore, the profile that can be seen on OpenVSP also does not  have these self-intersecting loops. I suppose it is a problem related with the high number of sections i have used. 

I was just trying to replicate a propeller geometry. and i was checking how it improves the geometry created with an increasing number of sections. I suppose i will keep a lower number of sections. 

Thank you and sorry for the inconvenience ,

Alfredo 

 

[Rob McDonald]

OpenVSP would not produce that loop without bad airfoil points.  Please post the points for that foil and maybe I can help you identify the problem.

If the points are fine, then there is a deeper problem I need to sort out -- and that will still be helped by sharing the points.

Rob

[Alfredo Escolano Lopez]

I have just realized i have a duplicated point that may be the issue of the self-intersecting loop. I am going to fix that to see if that solves the problem . Thank you, 

Alfredo

[Alfredo Escolano Lopez]

I attach the .txt file of the section 43. 

El viernes, 24 de julio de 2026 a las 20:58:42 UTC+2, Rob McDonald escribió: