FEA Symmetry

[Anonymous]

Hi,

I've been trying to use symmetry to simplify a finite element model. Any ideas why I see different stress results in these two instances.

Simulation 1: Rectangular Hollow Section, capped at both ends, subject to internal pressure

Simulation 2: Half model i.e. RHS split down the middle lengthwise, with a frictionless constraint applied to the split plane.

Actual problem I want to analyse is a bit more complicated and I was hoping to use symmetry to reduce the processing time? Any thoughts appreciated.

Thx.

[JDMather]

---

 Any thoughts appreciated.

---

Uhmmm, here is one.  Attach the file here.  We spend all of this time creating geometry to communicate, let's communicate with geometry.

Can you reproduce the behavior in a simplified problem?

[blair]

Most of my FEA is in Autodesk Algor, but you need to use the symmetry tool to identity with planes symmetry exists. If you haven't done this, it will assume the model is smaller.

[blair]

I don't think that there is a Sym tool in the FEA that ships with Inventor. I use it a lot in the Algor version, but we then have to identity the planes of symmetry. This allows for small models down to 1/4 of the model.

[Daren.Lawrence]

My recommendation for accuracy's sake is to keep the assembly intact and not use the split parts with frictionless constraint. You can probably see more performance gains in loosening up your convergence settings - particularly the number of h refinements and stop criteria. As shown in the attached image, I created a rectangular, hollowed out tube with end caps as you were describing with a maximum of 3 h refinements in my settings and a stop criteria when it gets within 3%. It only used one h refinement to get within the 3% stop criteria I asked it for (the first 3 points are p refinements in the plot).

The default is 10% stop criteria with 0 h refinements, but using these settings can give you more accurate results. The picture assembly used Aluminum 6061 material for the tube and end caps and 100 PSI of pressure on all internal faces and the two end cap ends were used as fixed constraints.

Using the part split with frictionless constraints can skew your results somewhat and when I did it on this part, I was closer to 1.75 ksi Von Mises stress compared to the 1.133 pictured, but the max was one of the sharp internal corners rather than the one of the external faces as expected.

The following pages are great resources on convergence:

http://usa.autodesk.com/adsk/servlet/item?siteID=123112&id=12953291&linkID=9242016

http://usa.autodesk.com/adsk/servlet/ps/dl/item?siteID=123112&id=13458825&linkID=9242018

Please let us know if that helps. Thank you.

[salariua]

Daren,

I had the exact results when playing with symetry and thin shell parts.

Compared with full shell.