FEA - Static Stress Analysis - Detect and Eliminate Rigid Body Modes

[TheCADWhisperer]

Ping @John_Holtz  

I have used this example for more than 10 years in various analysis softwares to demonstrate singularity that designer can ignore where edges meet in top left of this assembly.  We normally add a soft spring (Detect and Eliminate Rigid Body Modes) in other analysis softwares to stabilize the model for analysis.

This is the third lab now that I have had where this does not work as expected in Fusion 360.

I am beginning to think it is a bug, I seem to recall using this technique in Fusion 2 or 3 yrs ago.

Fusion 360 file attached below.

[John_Holtz]

Hi @TheCADWhisperer Sorry for the delay in my response.

According to the requirements in the documentation, your model does not meet the guidelines required by inertial relief. From Remove rigid body modes, 

Are any constraints applied?	Does the sum of applied forces in each constrained direction equal zero?1	Can you use Remove rigid body modes?
Yes	Yes	Yes
Yes	No	No
No	(Not applicable)	Yes

This brings to mind the following questions:

1. Does the software follow these rules now?
2. Did the software follow these rules previously?
3. Do the requirements apply on a part-by-part basis or the entire model? If part A is completely free but some completely unrelated part B has a fixed constraint, does that prevent this option from working on part A?
4. Do the requirements consider contact as a "constraint" (which is constraining the latch bolt in your model in the X and Z directions).
5. Your model would be the second row of the table (yes, no, no) because, technically, your model has constraints but does not have balancing forces in the constrained direction (Z direction). Is this what the documentation means to say?

What I do see from the analysis model that is generated (.nas) is that the parameter InertialRelief is not included. So the analysis will not solve unless the AUTOSPC adds a constraint to the latch bolt to prevent motion in the Y direction. (The analysis should do that.)

I will check with some co-workers to get some answers.

[John_Holtz]

Hi.

I have had some discussion with the developers about this. Although we do not know for certain, there is a possibility that the older version of Fusion used a different solver, and when the solver was changed to the Nastran solver, the method of solving the Remove Rigid Body Mode changed. Maybe that is why it "worked" a few years ago but not now.

What is more certain is that the Remove Rigid Body Modes in its current form is meant for models that are totally unconstrained. 

The documentation is correct about models with constraints. The applied loads need to sum to 0. (It has something to do with the analysis removing the constraints. If the loads are not "balanced", the results are not correct.)

What the documentation does not indicate, and which appears to be important as well, is that the sum of the moments due to the applied loads also needs to be zero. So when the documentation says that "the loads need to balance", I believe it really means the following:

1. Sum of forces in X = 0
2. Sum of forces in Y = 0
3. Sum of forces in Z = 0
4. Sum of moments about X = 0
5. Sum of moments about Y = 0
6. Sum of moments about Z = 0

Let me know if you have any questions.