FEA Material Nightmare

[lkrenzler]

I set up a simple stress analysis to try to figure out the Shear Modulus problem I put in an earlier post.  I found a whole new problem.

If I run a simple simulation with the material "As Defined" I get a max stress of about 500ksi (see screen captures).

If I manually select the material (notice it's exactly the same material!!) I get a max stress of about 800ksi.  This is pretty different results with what should be the very same material.  I've run it with all kinds of meshing variations but with similar results.  

[lkrenzler]

Here's the last screen cap.  Seems I can't attach more than 3 to one post...

[lkrenzler]

Here's the files if anyone wants to try it.  Any thoughts would be much appreciated.

I'm still in the dark about the very whacky shear modulus issue posted here as well:

http://forums.autodesk.com/t5/Inventor-General/Custom-Material-Properties-Shear-Modulus-in-particula...

[blair]

I haven't looked at your files but make sure you install all the SP's and Hot-Fixes for your product. There is at least one major patch for the FEA in IV2014.

[lkrenzler]

Thanks.  I did that already.  This is with 2014 SP2 (the latest SP).

[JDMather]

I didn't take the time to try to figure out how you got where you did -

but when I open the part file the material definition is incomplete and incorrect.

[lkrenzler]

Hmm.  Maybe that was from the Pack and Go I used or something.  When I open the part file here it is complete, just like the image on the right.  I can't find any incomplete materials here.

I've attached one that includes all the libraries and styles.  It was large but compessed better than I expected.

[LT.Rusty]

Something else to think about is that FEA results take a lot of interpretation.  They are not something that you can just bang out and have instant success with.

On a lot of my models, particularly ones with welds, I'll get tiny little pinpoints of ridiculously high stress, which are absolutely not replicated by real-world testing, and it's quite possible that you're seeing the same thing here.

One way that I make things easier to read and interpret is to make some changes to the view setup.  The first thing I do is set the displacement to actual instead of Adjusted x1, then I set the colors to Color Bands instead of Smooth Shading.  Finally, I change the settings on the color bar so that the max value shown is the max value that I want to see.  (Usually I'll set this to be the yield strength of the steel - I'm designing for a zero margin on the safety factor, basically, because my test load already incorporates a large safety margin.)

Setting things up this way will make it a little easier for you to see what's going on in your part.

(All this of course assumes that your material is set up correctly in the first place - if it's not, then it doesn't matter how you set up the screen, it won't give you usable results.) 

[lkrenzler]

Thank you very much for the insights.  I'll set up an even simpler test with no welds.  The part that worried me the most was that a single change from one material to the very same material made such a difference with nothing else in the simulation being changed.

Will do some more testing today.

[raviburla · Accepted solution]

Hi,

With respect to "no change in results when shear modulus is changed", here is why that is happening:

For linear isotropic materials only two quantities are independent of the three – Young's Modulus, Poisson’s ratio and Shear Modulus. Inside the FEA-solver we choose to use Youngs moduls and poisson’s ratio as independent values and compute Shear modulus from them. This is the reason why any changes to Shear modulus will not cause change in results.

Please let us know if you have more questions/concerns.

Thanks,

Ravi Burla (Autodesk)

[lkrenzler]

Thank you very much Ravi.  That does answer one of the two important questions.

The other question was why I would see such significant changes to the results when the material is manually specified to what should be the exact same material and absolutely nothing else has been changed.  See the screenshots and sample above.

The sample I posted earlier should be able to reproduce this for you.

Thanks for your help! - Len

[raviburla · Accepted solution]

Hi,

I looked into the data you had posted. While I was not able to reproduce the exact behavior that you had seen, but I do see few discrepancies.

The assembly is created with two different parts.

1. Part1: 000007936.ipt has a material definition for G40.21-44W having the material properties as shown in the attached image (7936.png)
2. Part2: 000007937.ipt has a material definition for G40.21-44W having the material properties as shown in the attached image (7937.png)

Notice that both the materials have same names while different properties. So I am thinking  the following:

When “As Defined is used”, the material data is being picked from the corresponding part’s material definition.

When “Override is chosen”, the material data is being picked from “000007936.ipt” definition. Since these two lead to different material settings, the results are different.

Can you update the material property in both parts to have same values and check if you still see the different results?

Thanks,

Ravi Burla (Autodesk)

[lkrenzler]

Thank you very much Ravi.  I'll check.  Looks like user error on my part.  Thanks again.