Unbelievably large stresses predicted by stress analysis – wrong contacts used?

[splinemadness]

I am modeling an assembly that essentially consists of a pin that holds a rod in a tube (a simplified version is hopefully attached, as well as a jpeg of the stress analysis).  When I try to analyze the stress under a 100 lb load, the forces on the pin skyrocket, even in locations that should not experience any shear.  From pen and paper calculations, a 0.25” diameter pin should experience 2 ksi shear (100lbs/(pi*0.25”^2/4)) and 1.6 ksi bearing load (100lbs/(0.25”*0.125” thick handle walls*2)).  While tinkering with it, the stress analysis has predicted >100 ksi in the center of the pin (!!!)  Anyone know what I am doing wrong?  I tried using a 0.25” hole for the pin to pass through as well as a 0.255” hole, which actually resulted in larger stresses.  Bonded contacts were placed between the pin and rod, and again between the pin and tube.

Also, what is the “separation contact” for - the surfaces freely pass through each other (like two ghosts), allowing the assembly to come apart without any resistance?  It is frictionless and does not appear to prevent penetration of the surfaces with each other, so isn’t that that same as not using any contact? 

Thanks,

John

[splinemadness]

Here's the last file...

[CraigBoyak7585]

Just a guess but what temperatures have you set for environment and reference (for thermal expansion)?  It looks like the high stress is where there is a radial interface between pin and the two other parts.

[JDMather]

---

@CraigBoyak7585 wrote:

...what temperatures have you set for environment and reference (for thermal expansion)? 

---

Can you demonstrate how to do this using Inventor?

[JDMather]

You did not attach pin.ipt  If you attach this file someone can try to figure out what is wrong.

Also I noticed that your sketches are not constrained (something a simple as circles at origin - that Inventor will automatically constrain).
I recommend reading this document and learning to fully constrain sketches before moving on to advanced topics like FEA.  http://home.pct.edu/~jmather/skillsusa%20university.pdf

[splinemadness]

Oops, sorry about forgetting the pin!  I didn’t set any temperature, so I would assume that everything will be at 25°C (based on the listed material properties) but it does look as if the pin is being squeezed – that’s why I was thinking that perhaps it was a result of the bonded contact on a cylindrical part.  I went back and fully constrained all the sketches and re-ran the simulation, but received the same stress result again… it looks like something else is the problem.  Thanks for the suggestions,

John

[CraigBoyak7585]

Setting reference and environment temperatures would be done in the FEA side of Inventor.  I do not use Inventor Professional which would have the FEA module so I cannot provide specific guidance.  But the concept is the same, delta L = alpha x L delta T with delta T being environment - reference.  It's a common mistake, especially if the program assumes Tref = 0.   Set a normal environment T and even at ambient things grow.  A quick check would be to examine a displacement plot with no external loads.

  Craig

[splinemadness]

I tried removing the external load (by placing an equal but opposite force on the other side of the rectangular bar where the load is applied... Inventor requires there to be at least one load in the simulation) and the stress dropped to 0.07 ksi maximum, the displacement has dropped to some 1e-5 value that occurs right at where the loads are applied...  Thanks for the suggestion, but I guess that something other than thermal expansion must be causing it.

-John

[henderh]

Hi John,

  I agree the stress results seem unreasonably high in the pin.  I've logged the issue as 1411189 for Development to look into.

  If we change the manual contact types to "sliding / no separation" we have much better results (assembly attached).  The shear stress (XZ) is near 2 ksi in the pin.

Thank you for bringing this to our attention.

-Hugh

[splinemadness]

Thanks for the help, but when I run it with only sliding/no separation contacts (my assembly or yours) it wants to separate into 3 parts…  It might make Inventor more user-friendly if it could automatically detect and eliminate the intersection of surfaces during stress analysis.  Thanks,

John

[splinemadness]

BTW, it looks as if combining the pin with the rod part prevents the problem (they are one solid).