Detect and Eliminate Rigid Body Modes in Static Analysis

[TheCADWhisperer]

Pinging @John_Holtz 

Every year we run this analysis in Fusion 360 and other softwares.

I have a wide flange beam (I-beam) supported on 8" "block walls".

Study 1 - fixed on both ends.  Get expected results.

Study 2 - fixed on one end, frictionless on the other end. Get expected results.

Study 3 - frictionless on both ends (representing simply supported on the "walls"). Have not found a technique to get expected results.  Either excessive displacement or no displacement...

For Study 3 because mesh is not perfectly uniform - we have to add a soft spring (Detect and Eliminate Rigid Body Modes in Inventor) and get the expected results.

In Fusion 360 - if I set that flag I do not get the expected results.

(In the attached I left it off and the Displacement is excessive.)

My memory could be faulty - but I think that years ago we got the expected results in Fusion 360 with Remove Rigid Body Modes.

[TheCADWhisperer]

As a reminder - this was last years discussion, but it was on an assembly rather than a simply supported beam.

https://forums.autodesk.com/t5/fusion-360-design-validate/fea-static-stress-analysis-detect-and-elim...

[TheCADWhisperer]

Bump.

[TheCADWhisperer]

@MikeSmell_ADSK 

In the classroom I run the old ALGOR tutorials (with some modifications) through Fusion 360.

Any comment on this issue?  (See previous Attachment.)

[John_Holtz · Accepted solution]

Hi @TheCADWhisperer 

Sorry for the late reply.

In study 3, the model is not statically stable.

• Inventor stress analysis adds weak springs to make the model stable, and the results are acceptable. (Some of the applied load is "lost" through the weak springs instead of going through the constraints. But if the springs are truly weak, the difference is negligible.)
• Fusion without "remove rigid body modes" gives inaccurate results (if it will run at all) when the model is unstable. That is how math works when the equation 0/0 is solved.
• Using the "remove rigid body modes" in Fusion counters the applied load (gravity) with a uniform load distributed throughout the volume (for example, gravity!) so that the sum of forces is 0. Since one uniform load (gravity) is equal and opposite to the uniform load from the remove rigid body modes at every location, the displacement is zero. As indicated in the older thread and the documentation Guidelines for using Remove rigid body modes, study 3 is not a good model for the use of "remove rigid body modes".

The solution in Fusion is to apply additional constraints to make the model statically stable. For example, constrain one edge (one of the edges parallel to the X axis, but not the one with the frictionless constraint!) on one end of the beam in Ux and Uz.

P.S. I recommend to not use frictionless constraints when a regular constraint is acceptable. In this example, a constraint fixed in the vertical direction (Uy) gives the same behavior as the frictionless constraint, but regular constraints do not introduce side-effects that the frictionless constraint can created.

[TheCADWhisperer]

@John_Holtz 

Thanks.

I understand now.

For some reason I did not understand the explanation a year ago.