Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Accuracy of Simulation

Please login with a confirmed email address before reporting spam

Dear all,

I am currently working on a complex micro mechanical structure. I simulated this structure in solid, strain-stress module. But I am not sure about accuracy of my simulations. Is there any way to check my results and assure myself that results are reliable and correct?
I appreciate all your kind help in advance.

Best Regards,


2 Replies Last Post 26 févr. 2016, 20:09 UTC−5
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 17 avr. 2013, 14:11 UTC−4
Hi

At least you ar asking the right questins: how to validate a FEM model. Unfortunately I do not know any easy solution. I can give you "my way" which is basically to simplify to the extreme a few parts of a complex model such thta I can use analytical know or dicumented formilas, and then check/validate against those values, when you have curves of parameter sweeps, often one or two can be made to pass close to a known solution.

Other ways are to simplify the FEM model ,and erify, often one physics at the time that we get reasonable results when compared to known alanylitical formulas.

Finally recheck your model (particularly BC settings) , calulate reaction forces, check they are the same as the total boundary loads input, check energy otr fluxy balances in out.

For structural models, what often goes wrong are transfer of material definitions or material properties. For this I always compare my CAD results and FEm results on theemshed inertia for mass, CoG nad inertia tensor values. Unfortunately these are not calculated y default anylonger in COMSOL, you must write out all the formazulas and add them in as "variables" and set them up in "Derived values", quite cumbersome for such a standard consitency check. Probably the COMSOL developpers have never worked on any critical FEM task as siuch for satellites or nuclear plants, so they have never had a "quality manager" n their neck :)
So anyhow its up to the end user (your responsability as engineer) to check that the results are OK, he cannot only rely on the software supplier

--
Good luck
Ivar
Hi At least you ar asking the right questins: how to validate a FEM model. Unfortunately I do not know any easy solution. I can give you "my way" which is basically to simplify to the extreme a few parts of a complex model such thta I can use analytical know or dicumented formilas, and then check/validate against those values, when you have curves of parameter sweeps, often one or two can be made to pass close to a known solution. Other ways are to simplify the FEM model ,and erify, often one physics at the time that we get reasonable results when compared to known alanylitical formulas. Finally recheck your model (particularly BC settings) , calulate reaction forces, check they are the same as the total boundary loads input, check energy otr fluxy balances in out. For structural models, what often goes wrong are transfer of material definitions or material properties. For this I always compare my CAD results and FEm results on theemshed inertia for mass, CoG nad inertia tensor values. Unfortunately these are not calculated y default anylonger in COMSOL, you must write out all the formazulas and add them in as "variables" and set them up in "Derived values", quite cumbersome for such a standard consitency check. Probably the COMSOL developpers have never worked on any critical FEM task as siuch for satellites or nuclear plants, so they have never had a "quality manager" n their neck :) So anyhow its up to the end user (your responsability as engineer) to check that the results are OK, he cannot only rely on the software supplier -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 9 years ago 26 févr. 2016, 20:09 UTC−5
Thank you for your response. It was helpful.
I extracted the theoretical formulas and performed some experiments. The obtained results were very close to simulation results.
Thank you for your response. It was helpful. I extracted the theoretical formulas and performed some experiments. The obtained results were very close to simulation results.

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.