Error in calculating mesh?

[Marco Presi]

Hi all, greetings from Italy!

I am a fresh "new" OpneVSP user. I am looking forward to use it to model Radio Controlled gliders.

After reading the most of the Ground School tutorials, I am now modeling a glider.

I started from the fuselage. When calculating its mass properties (I am specifying the density as thin shell), the mesh generates few triangles outside the fuselage body.

I don't think those triangles will influence too much the CG calculation, but I am wondering if it is a bug and if it may affect the result of other calculation (e.g. the drag).

I am attaching here an image showing the error (it shows only the fuselage, the tail boom has not been added yet).

Any hints?

Ciao Ciao,

marco

[Rob McDonald]

That is certainly a bug.

Fortunately, if you are only using the thin shell based density (and volumetric density is zero), those tris will not affect the answer at all.

Those tris are remnants of the slicing procedure and are only used for the volumetric mass properties contribution.

Can you please share the exact input file that is producing that situation?  I am surprised by that behavior -- in particular, your geometry looks quite tame and reasonable.  You are using a large number of slices, but that should be relatively harmless.

If you are doing shell only, you can reduce the number of slices to the minimum and it will not change the answer -- it will run considerably faster.

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 on the web visit https://groups.google.com/d/msgid/openvsp/eeab2a30-f3aa-445b-a2af-72c162ddb7e7n%40googlegroups.com.

[Marco Presi]

Dear Rob,

thank you. I am using thin shell for the fuselage (I plan to make it in carbon fiber) and volume density on the wings (they will be foam-core, with external laminated fiberglass).

Please find attached here the file (I now included wing, tail and tail boom). I see the same bug in the wing tips when calculating the mass.

I get exactly the same output both in Windows 10 and Linux.

On both platforms, openVSP version is 3.31.1 

Best Regards,

Ciao,

Marco

To view this discussion on the web visit https://groups.google.com/d/msgid/openvsp/CAEppYpFJyjm640t6stQoErOi0BZd1VY6fmgVmv0udKefRSb8%3DQ%40mail.gmail.com.

[Rob McDonald]

Thanks for sharing that file.  I can replicate those results here.

I'm not sure when I'll be able to dig into this, but it seems really unusual.  If you can, it would be useful if you could check some old versions of OpenVSP to see if they do the same thing.

Start by going back a long way -- say 3.15.  If it does the same thing, go back further.  If it exhibits different behavior, come forward halfway -- 3.22, try again.  If the behavior ever changed, some bisection to narrow down the version where it changed would be a big help.

Rob

To view this discussion on the web visit https://groups.google.com/d/msgid/openvsp/CANpmRq96iC%2BMq72sT%3DRETM%3DoTz%3DREMXji5J%2BQ1YBUM%2B7%2BUOyJA%40mail.gmail.com.

[Marco Presi]

Dear Rob, I will do it and let you know.

Thanks,

Ciao

Marco

To view this discussion on the web visit https://groups.google.com/d/msgid/openvsp/CAEppYpHDAQhiw9zMRUqDK8msaYYqD%2BapKu9Q7pykzLO-nMWfoQ%40mail.gmail.com.

[Marco Presi]

Dear Rob,

I went back to version 3.15 as you suggested, and then I continued down to version 3.1.2. The bug is always there.

Please let me know if I can help in any other way, and if I should open a bug notification on Github.

Best Regards,

Ciao Ciao

Marco

To view this discussion on the web visit https://groups.google.com/d/msgid/openvsp/CAEppYpHDAQhiw9zMRUqDK8msaYYqD%2BapKu9Q7pykzLO-nMWfoQ%40mail.gmail.com.

[Brandon Litherland]

On a hunch, I downloaded the file to test it out on my end.  I'm assuming that this is in meters.  The fuselage is only 0.87 m long which made me wonder if there was some numerical error somewhere in the computation.  I scaled the model by 10 and ran Mass Prop again and the external tris are eliminated.  Also, your direction at the final XSec point is flipped in the Top direction this is inverting the aft surface somewhat.  You can set them all to -90 or turn off the skinning entirely for that section.
I've tried changing the Superellipse sections to Ellipse, turning off skinning, adjusting surface tessellation, altering the skinning, etc.  I still don't know why that component is unhappy.  Scaling it up seems to fix some issues but it shouldn't be doing this.  The tail boom doesn't seem to exhibit the same behavior.  I also tried creating a new model with a default fuselage scaled down to a length of 0.5 and ran Mass Prop on it.  Everything works just as expected.  No hiccups whatsoever.  Tried recreating the model fuselage with a new component from scratch > Things were fine if no Z offset was applied for the sections.  Apply large Z offset of 0.3 relative to length for every other section and it worked fine.  Recreate the fuselage exactly and it has the extra tris... 

Turning up the Num W to 361 causes the slicer to create a LOT of extra tris.  It's a complete mess.  I'm running out of ideas on how to investigate this for now.  I'll probably come back once I think of something else to try.

[Marco Presi]

Dear Brandon,

thanks for the follow up and for the XSec tips (this is my very first modeling attempt). All in all the mass calculation gives numbers very close to the real ones (we are hobbyists, and every fuselage comes out with some variations). I don't think it is really a critical issue especially if it manifests in particular cases like my one. If you can give me some indication on where to have a look in the code, I may try to debug this.

I find openVSP really interesting. In less than an hour spent on the video tutorials, I was able to make a full design. I think it is amazing. 

I have some open questions on the VSPAERO tool, which I will ask about in the next few days.

Ciao Ciao,

Marco

To view this discussion on the web visit https://groups.google.com/d/msgid/openvsp/650ec250-f224-4dba-a06b-9dec511e93f5n%40googlegroups.com.