Inventor & Nastran, Using Nastran changes .ipt

[NigelHay]

Not sure if I should post here or in the Nastran forum but I'll start here. My colleague has started using Nastran for modal analysis of jigs used for vibration testing of our products. The object is to ensure that the jig does not introduce any unwanted resonance within the test vibration spectrum. Something is happening to the .ipt when he does this analysis which makes the .ipt change to a dumb solid. I've attached a simple jig to illustrate, this was created in Inv2023 as a rectangular extrusion, chamfers added then 2 sets of holes added. If you look at the part now it just consists of 'Solid1'. The 'Hole1' & 'Hole2' I have added afterwards because the drawing I made of the part would not recognise the existing holes for dimensioning. These holes show with issues because adding them did not remove any material, the holes were already there.

Can anyone suggest what is happening? My colleague is not aware of anything he is doing to change the part. For now, he will make copies of any parts he is looking at to avoid this problem.

[CGBenner]

@John_Holtz 

John, have you ever seen anything like this happening with Nastran?

[John_Holtz]

Hi @NigelHay 

I am not sure I fully understand your description.

1. Is the .ipt file you provided after the analysis when some change has occurred?
2. What is the change? Is it that the entry in the Model tree named "Solid1" used to be something else? What was it before it became Solid1?
3. You said that Nastran is being used. However, there was no setup for a Nastran analysis. There was a setup for Inventor Stress Analysis (which is not Nastran).

Also, the way I read it, you are saying that the original Inventor model did not work properly because you had to add fake holes to get something to work. It sounds like some type of problem with the original model.

You should be able to go back to the "OldVersions" and retrieve the model before it changed.

Whatever is occurring appears to be unrelated to Nastran.

John

[NigelHay]

John, thanks for your reply.

1. Yes, the .ipt is how it was after the analysis.

2. The 'Solid1' used to be what you would expect for a part created in Inventor. The browser would have shown an extrusion plus chamfer & 2 sets of holes.

3. I may have misunderstood which stress analysis was being used, it could have been Inventor.

The fake holes were added to allow me to produce a drawing. The drawing hole command did not recognise either the tapped or the c/bored holes in the Solid1.

I looked at the old version folder but the only part in there was the same as the one I attached to the post with a 'Solid1' in the browser.

For reference, the same has now happened to another part which has has an analysis done.

[pcrawley]

I recreated the pre-analysis version (attached) using your sketch as the starting block.  

I haven't been able to make it revert to a dumb solid with any of the analysis tools (yet).

Since it has happened again, can you find out which analysis package was used (Nastran or the built-in "Stress Analysis") - and what steps were taken before it broke?

And just checking versions: Inventor 2023.1.1 and (if applicable) Inventor Nastran v 2023.1.2.48 (Nastran 2023 version 17.1.2.33)

[NigelHay]

I opened your version of the part & that is pretty much as my part would have been originally.

My colleague has been using the Inventor analysis & Nastran. I can't say at what point the change to a dumb solid occurred, I only discovered this when I tried to create a drawing of the part some time after the analysis had been done. We have used the Inventor analysis many times in the past with other items & not encountered this problem.

We are using Inventor 2023.1.1, we have Nastran 2023 but my colleague has been using 2022 as he found that '23 was a bit buggy.

[pcrawley · Accepted solution]

That might be the problem - you're using 2023 but the analyst is using 2022. 

I can't remember the specifics, but when using an older version of Inventor to open newer version models, the older Inventor uses the AnyCAD functionality - basically rendering the new feature tree as a base solid.  Perhaps, during part of the analysis process, your colleague managed to save the "dumbed down" variant over the original.  I've no idea of the workflow that would require, but it feels possible (but it is also a guess!).  Maybe one of legendary Autodeskers can advise.

[NigelHay]

That certainly sounds like a possible & plausible explanation.