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.
"Include Geometric Nonlinearity" for Membrane
Posted Dec 13, 2016, 8:56 p.m. EST Acoustics & Vibrations, Studies & Solvers, Structural Mechanics Version 5.2a 5 Replies
Please login with a confirmed email address before reporting spam
I read through the manual for Membrane node and found that "Include Geometric Nonlinearity" is necessary to be checked in the Study Step. This is to include the prestress specified in "Initial Stress and Strain" node.
However, after turning it on, the study could not converge as the maximum Newton iterations of 25 has been reached.
The 3D model involves a 0.5 mm thick rubber membrane fixed along its edges for acoustic studies.
Hope someone can advise. Thank you.
Please login with a confirmed email address before reporting spam
As the membrane does not have any bending stiffness you need to pre-stress for it to make any sense when used in an acoustic application. To get a bit of inspiration take a look at one of the condenser microphone models under Acoustics > Electroacoustc Transducers in the Model Library.
Best regards
Mads
Please login with a confirmed email address before reporting spam
I managed to specify pre-stress using the example "Vibrating Membrane". However, it seems that there are convergence issue when a membrane is included in the acoustics study.
--
Regards,
Linus Ang
Please login with a confirmed email address before reporting spam
If you open the documentation for for the model (I think there is one). Then you will notice that you need to manually set the scaling for membrane variables. The reason is that when prestressing only in-plane the membrane is of course not moving out of plane. This makes it difficult for COMSOL to predict appropriate scaling for 0 deformation. Have a look at one of the condenser microphone models.
Best regards
Mads
Please login with a confirmed email address before reporting spam
I don't understand what do you mean by having to manually scale the membrane variables. In the condenser microphone, I don't see a scaling done anywhere. Please advise where can I locate this. Thank you.
--
Regards,
Linus Ang
Please login with a confirmed email address before reporting spam
I assume we are talking about the "dependent variable solver scale values", set by the solver, it is logged in the solver log file and defined under one of the solver sub-nodes dependent variables.
Mostly COMSOL proposes "good* scaling values, but in certain cases it's useful to tweak these by hand.
The idea is to have a numerical value close to "1" (I.e. expected displacements in the 1-10 um range => the variable should be scaled by 500'000. Or a contact pressure expected in the 100-1000 [MPa] scale by 1/500'000'000 ...)
The difficulty is when a value such as i.e. a (contact) pressure goes from 0[Pa] to 1[GPa] or 9 orders of magnitude, this is tough to represent correctly in standard Double Floating Point numerical notations, here one need to test a little, and or split the calculations i.e. for p=0-1[MPa] and then from 1[MPa] to 1 [GPa] with different scaling values via a manual 2 step solving process.
--
Good luck
Ivar