Issue with Non-Linear heat transfer simulation

[Anonymous]

Hi,

I'm encountering a strange problem with a thermal analysis in In-CAD 2016. I have run this same simulation (a few months ago), in 2015 and it worked perfectly, with temperature stabilizing at around 300K, as expected. But now, without changing anything in the model, though loading it up on a different computer, the temperature climbs continuously, well past 4000K. I have checked units, emmissivity values, time tables, everything I can think of, but can't find the issue.

Do you think it is likely a problem with the .iam?

[AndrewSears]

If you still have the 2015 data, unsaved in 2016, can you supply the model?  The data saved in 2016 will work also if that is what you have.

If you can't attach it to the public forum, email it to [email protected] and we will take a look.

Thanks,

Andy 

[Anonymous]

Thanks for getting back to me Andy.

I do still have the data from the 2015 version, but I can't share it, for proprietary reasons. Is there something in particular you think I should be looking out for?

[AndrewSears]

We make code changes and test forward compatibility with models but sometimes we find something that needs more code work.  What we would be looking for are differences in the generated .NAS file between 2015 and 2016 to make sure everything is being generated correctly from one version to the next.  If you supplied the .NAS files, this will not give us any Inventor CAD data except for material properties but it will give us the nodes and elements (mesh) so we would essentially be able to see what your design looks like so keep that in mind if your designs are confidential.  

Our legal team will provide nondisclosure agreements if that will make your company more comfortable about sharing data with Autodesk.  We do not share customer datasets outside of the company but we understand if you are not able to send this to us.  We will do more testing on our end and see if we can reproduce the issue.  

If you want to check on your own, you can use WinMerge Download or another application and compare the .NAS files.  Let us know if you find something that is not the same between the two .NAS files. 

Andy 

[Anonymous]

I definitely can't send it on to you. And as I'm just an intern, I duobt my company will want to go through any complications regarding non-disclosure agreements and so on. In fact, these simulations aren't vital to our work, they're just for the sake of an additional verification of our model.

But, I will have a look at the .NAS files, and let you know if anything comes up. If you're unable to recreate it, and I can't find any issue with the files, would you recommend just going back to the 2015 version?

[AndrewSears]

Understood.  

I don't know enough about the problem to give you a recommendation about what version of In-CAD to use.  After you compare the .NAS files and let us know what you found, we can investigate and inform later.

Steps:

1. Open the model in 2015 Inventor and enter the In-CAD environment.
2. Generate the .NAS file.
3. Save the .NAS file.
4. Open the model in 2016 Inventor and enter the In-CAD environment.
5. Generate the .NAS file.
6. Save the .NAS file.
7. Change the file type from .NAS to .txt for both files.
8. Open both files with WinMerge and let us know what is different.

Thanks, 

Andy 

[Anonymous]

Hi Andy,

So I contacted my previous employer (where I ran the previous analyses that worked perfectly), and it turns out that I had been using the 2016 version. I must have gotten it confused, as the version was updated while I was working there. So the issue does not appear tp be due to the newer version of In-CAD. I even installed it n my personal laptop and tried it again, and the same issue occurs.

I am cpompletely stumped, and the only thing I can think of is that the .iam files got corrupted or damaged when I copied them onto a hard drive to transfer them to my current job.

[Andrew.Sartorelli]

Hi otoolec2,

When meshing are you using linear or parabolic elements?

Andrew

[Anonymous]

Parabolic

[Andrew.Sartorelli]

Hi otoolec2,

Can you try switching to linear and see if that helps. There is an issue with parabolic elements in some sub-sets of thermal analysis that we are looking at fixing in the next version.

Regards,
Andrew

[Anonymous]

OK, I have tried that, but get the same result.

[AndrewSears]

Try creating a new simulation.  Do you use custom materials where you have modified their physical properties?

It is not likely that the Inventor CAD geometry is causing the problem.  We only store setup information at the part/assembly level.  

Andy 

[Anonymous]

Yes, I do use some custom materials, but they all appear to be unchanged. I had aleady started a new simulation, though will have to wait 'til tomorrow to finish it. I will let you know the results.

[Anonymous]

OK, I have created a new simulation, but am getting the same issue.

[Anonymous]

I think I have found the problem! For some reason, I (or someone else for all I know) changed the specific heat capacity of a material of one of the internal components sometime in the last few months between the previous simulations I was running and the current ones. I honestly don't know why I would have, as the value I had been using was 5 orders of magnitude lower than the correct value. I think the circumstances leading to that are almost more confusing than anything else.

I hope I haven't wasted too much of your time. Thank you for the help! (and I probably never would have thought to check it if you hadn't brought up the custom materials)

[Anonymous]

OK... that didn't work. I'm completely stumped.

[AndrewSears]

Without the data sets, this will be hard to diagnose.  Just keep checking for differences in the setup, materials, etc.