Issues with Assembly FEA Analysis

[cristobal_swinburn]

Hello everyone! 

I´m currently doing my first FEA simulations with assemblies. I used to do a lot of FEA analysis of parts that were relative simple and the results were great, with beautiful convergence plots and low errors.

However, with the finite element analysis of assemblies, I strugle a lot to achive good and representative results. When the convergence rate is good (< 5%), the error is still really high

(15-20%). I would like to know what you guys recommend to achieve better results in this kind of analysis. In the next section, I'll include some pictures and context about the part that i´m working on so you can understand a bit better what i´m dealing with.

For a bit of context, my team and I are currently designing an upright and all its components for an FSAE car. The part that I highlighted is the caliper mounting peace of the rear wheel. according to our calculations, each of the 2 holes will be subjected to 600 N forces aprox. 



I then I proceeded with the FEA: I defined the material, constraints, forces, the contacts and the convergence settings.
The results are strange because the values change a lot from one simulation to another, always resulting in high error in the convergence plot as you can see in this images: 

So, I would like to know what you recommend when you have really high percentage error in the convergence plot. As I said, this is my first FEA assembly analysis, so I´m kind of lost to be honest. Any help is more than welcome: tutorials, links, general advice, etc.

Thank you very much for taking the time to read this.

Best regards,
Cristóbal 

[John_Holtz]

Hi Cristobal,

I'm not sure if anyone knows what "Error" means or how it is calculated. Without knowing the details, the only thing that can be said is a smaller value is better than a larger value. 

It may be a good idea to do the mesh refinement by hand to get a better sense of

• where the stress is changing
• Is the change in stress real or a singularity. (If a singularity, the stress is generally confined to 1 node and quickly dissipates by the time it get to the next node moving away. Also, the stress at a singularity will continue to rise as the mesh size gets smaller, but the high stress remains in the smaller volume instead of occurring in the same volume that has more elements with the smaller mesh size.)
• Look at the result "Other > Solid mesh convergence" to see what it reports.
• Since you have an assembly, there must be contact. The results at the contact faces are an approximation at best, so I would not put much faith in the values on the contact face.

John