FEA-Static Stress Simulation results creating stress

[mbostonsprint]

I can't consider myself an FEA expert (by a long shot), but I am having some difficulty duplicating a result using exactly the same part, simulation setup, etc. I am getting some results (after the first apparently lucky shot) that are completely wacko.

Is there an Autodesk Fusion employee expert to whom I can send a couple of examples? These are simple glass disks with 3 contact pads, gravity loaded. Thanks --

[AndrewSears]

You can add the files to this post or email them to me.  Can we discuss the details of your simulation and model in the forum?

Andy

andrew.sears@autodesk.com

[mbostonsprint]

Andrew,

Thanks for the response. I have sent you an email containing a couple of screen shots of (more-or-less) back to back trials of same part and same settings (I hope). Please refresh my memory about posting of models here (or for email) for further analysis.

A couple of notes, some of which is in my email:

- This is a wedged glass disc, supported on 3 pads, set up as frictionless constraints, with gravity loading.

- I would expect the deformation values to be in the low nanometer range;, i have gotten some repetitive results that indicate almost 0.9 mm(!).

- One screen shot shows ballpark expectation, the other shows approximately 0.5 mm deformation.

- These were all solved "in the cloud".

[mbostonsprint]

A quick follow-up note:

- I saw that there is some sort of rotational torque that is applied in the solving of the analysis; I don't know why that occurs. As noted, the part is wedged, on the unconstrained side, so I'm speculating that somehow the solution is using the wedge in a strange fashion.

[AndrewSears]

Can you add the files so I can look at your setup?

Thanks,

Andy 

[mbostonsprint]

Hi, Andrew --

Files attached. Let me know the error of my ways.

[AndrewSears]

Thanks for the models and it is odd.  I am not sure why they are producing different results.  I am still looking into it.

Andy 

[AndrewSears]

The problem is your setup is statically unstable so that can lead to unusual results.  That is the best explanation I can give.  Fortunately I have a simple workflow that will produce stable results.

When working on linear static models it is best to make sure you limit the amount of available movement.  This can be done by holding the part with other parts that will restrict the motion in certain directions caused by the loads or you can use the fixed constraints.  I assumed in your design that the glass part is not rigid but more held in place by something clamping it in place.

Here are the results from Part1.

Now for Part2.

I replaced you small offsets with a new component to act as a support for the glass.  Attached are both models.

Review the models and let me know what you think.

Andy 

[mbostonsprint]

Andrew,

Thanks -- I'll take a look and go from there. My intent is to have the glass essentially unconstrained by any hard clamping, hence the selection of frictionless constraints. I assume that rigid constraints would have larger effects in the adjacent areas, which I would like to avoid for analysis purposes, at least.

[AndrewSears]

The glass is sitting on the extra parts.  After creating the contacts I edited them from the default Bonded and selected Separation.  These are more real world conditions.  

Andy