Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.
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.
finding stresses
Posted 21 giu 2012, 13:15 GMT-4 MEMS & Nanotechnology, MEMS & Piezoelectric Devices, Studies & Solvers, Structural Mechanics Version 4.2 3 Replies
Please login with a confirmed email address before reporting spam
I want to first thank the community for all of their help with the meshing issues I had before I was finally able to fix the problem. However, I have since run into another dilema and since I am the primary COMSOL at my university there isnt anyone who can help me. I want to place a small body load on the surface of my device in order to find the highest area of stress on the device. So far the program has crashed 3 times. I have no idea why and suspect I am missing some key parameter but am unsure as to what it could be. Also, I am experiencing a long computation time so I am not sure why, although I suspect it may be due to the small mesh sizes of certain parts of my device. Any help that you all can provide would be wonderful and thanks again in advance. I have attached the file in question so you can see what I have done. I have cleared the mesh results so I can fit the attachment but it is a fairly quick computation. Perhaps a more experienced eye can catch the mistake I am making.
Attachments:
3 Replies Last Post 22 giu 2012, 14:32 GMT-4
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
a few comments to get htere quicker:
1) do add a new "Definitions View", and therein under CAMERA sub node remove the "Preserve aspect ratio" this makes life so much easier in MEMS, you can then switch between the two views easily
2) for such geometries, make a global 3D block and add the layers as "Advanced Layer" (check the doc, give several layers each with its own incremental thickness
3) your mesh density is far to low, you need at least 3-5 mesh elements in the thickness, else you can consider "just" a "thin layer physics BC" i.e. Pt layer ... I would also consider brick or sweep elements
4) add a Domain node "Linear elastic material model" and select all linear elastic materials, at least our Pt, and SiO2, you need a second one for your Si that is normally anisotropic, but not PZT, for your other layers I do not know exactly which structure they have for your case. Then do not forget the order of the tensors, PZT and anisotropic material tensor index orders are different be aware (check the doc)
5) and what about adding some electrodes, a GND perhaps a terminal or voltage distribution layer
--
Good luck
Ivar
a few comments to get htere quicker:
1) do add a new "Definitions View", and therein under CAMERA sub node remove the "Preserve aspect ratio" this makes life so much easier in MEMS, you can then switch between the two views easily
2) for such geometries, make a global 3D block and add the layers as "Advanced Layer" (check the doc, give several layers each with its own incremental thickness
3) your mesh density is far to low, you need at least 3-5 mesh elements in the thickness, else you can consider "just" a "thin layer physics BC" i.e. Pt layer ... I would also consider brick or sweep elements
4) add a Domain node "Linear elastic material model" and select all linear elastic materials, at least our Pt, and SiO2, you need a second one for your Si that is normally anisotropic, but not PZT, for your other layers I do not know exactly which structure they have for your case. Then do not forget the order of the tensors, PZT and anisotropic material tensor index orders are different be aware (check the doc)
5) and what about adding some electrodes, a GND perhaps a terminal or voltage distribution layer
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
4) add a Domain node "Linear elastic material model" and select all linear elastic materials, at least our Pt, and SiO2, you need a second one for your Si that is normally anisotropic, but not PZT, for your other layers I do not know exactly which structure they have for your case. Then do not forget the order of the tensors, PZT and anisotropic material tensor index orders are different be aware (check the doc)
Is there a specific document you are referring to that may help me with finding the differences between the two?
5) and what about adding some electrodes, a GND perhaps a terminal or voltage distribution layer
I want to find the area of the greatest stress and strain first and that will be the area where I will add the electrode.
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
there are 3 sources for the index order, the material node explanations, the help doc for the PZT and material node, and then the main pdf doc for PZT (and structural). The conventions used by COMSOL on tensor index order are the most used ones, but one should notice these index orders are different for PZT and solid
Then for the electrodes, at least ensure you have a GND somewhere or some BC that ensure a fixed point also for the voltage, else there might be no single solution to the PDE and the solver will have issues to converge
Do not forget that in multi-physics, you still need that all dependent variables have enough BCs to ensure a single solution (per physics)
--
Good luck
Ivar
there are 3 sources for the index order, the material node explanations, the help doc for the PZT and material node, and then the main pdf doc for PZT (and structural). The conventions used by COMSOL on tensor index order are the most used ones, but one should notice these index orders are different for PZT and solid
Then for the electrodes, at least ensure you have a GND somewhere or some BC that ensure a fixed point also for the voltage, else there might be no single solution to the PDE and the solver will have issues to converge
Do not forget that in multi-physics, you still need that all dependent variables have enough BCs to ensure a single solution (per physics)
--
Good luck
Ivar