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.

Measure for deformation

Please login with a confirmed email address before reporting spam

Hi,

I am looking at the deformation of microwells in an elastic structure. The model is working, but I am having some difficulties extracting useful data from the results.

I would like to know how much wells are deformed after a certain applied stress, and what the resulting stress is within the wells (they are filled with a certain material).

I have tried a few different things, including volume maxima/averages for displacement and von mises stress within a certain well, but the results don't make much sense. For example, displacement is not a measure of the deformation of the well. I've also considered calculating the ellipticity of the wells after deformation from the visual representation, but I'm not entirely sure that the representation is completely to scale.

Specifically, if anyone can offer any help concerning:
- extracting information on obtaining deformation and stress data from a substructure in a model.
- if the graphical results are to scale, i.e. if there is a displacement of 100 µm on total length of 1 mm, this shows up accordingly. Or else, how to set COMSOL to do this.

Thank you,
Valerie

2 Replies Last Post 09.01.2016, 22:10 GMT-5
Jeff Hiller COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 9 years ago 08.01.2016, 08:06 GMT-5
Default deformation plots are scaled so you can see the displacements (Displacements are frequently extremely small relative to the size of the structure and would not be visible without that scaling). This has been discussed a few times in the Discussion Forum, see for instance www.comsol.com/community/forums/general/thread/95741 , www.comsol.com/community/forums/general/thread/92481, www.comsol.com/community/forums/general/thread/39873, www.comsol.com/community/forums/general/thread/98331 , etc...
Best,
Jeff
Default deformation plots are scaled so you can see the displacements (Displacements are frequently extremely small relative to the size of the structure and would not be visible without that scaling). This has been discussed a few times in the Discussion Forum, see for instance http://www.comsol.com/community/forums/general/thread/95741 , http://www.comsol.com/community/forums/general/thread/92481, http://www.comsol.com/community/forums/general/thread/39873, http://www.comsol.com/community/forums/general/thread/98331 , etc... Best, Jeff

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: 9 years ago 09.01.2016, 22:10 GMT-5
Hi Valerie
I read many mixed questions in your request, and I believe you might also have several issues that can give "strange results"

As your model seem slightly complex and not 100% clear to me from your few words about it (and I fully understand why not to give all details of proprietary studies on a public Forum ;). I'll also try to be rather generic in my replies (but I might also miss the "nail, be aware)

First the easy one: scaling of deformations in COMSOL structural:
1) if you are in linear geometrical mode (the default) you will have an arbitrary graphical scaling adapted to give "visible" results, but you can always go into the Deformation sub node of your Plot node and adapt the scale to something you prefer "1" or "1000" or whatever (or even a function of space and time, I believe, I have never tried this as not really thought of an application for it, but it could be a way to differentiate the solved model deformation from i.e. some theoretical results you have from another study means ?)
2) If you are doing non-linear geometrical deformation, (which turns on for certain type of physics and physics combination) then by default COMSOL uses a Plot Deformation of "1", but again it's up to you to change that to something more suitable. Note this only applies to the plot results, It's not changing the Solver Data Set. so if you are populating the Derived Values node you are extracting original data from the solver results, no scaling applies, other than what you do with the "[Units]" conversions on your equations.

Now you are looking for "deformation" versus "stress". So I believe you must better define "deformation" ie. the volume change, the ellipticity of a cylinder, the elongation ? ....
and for the "stress" you must define where you are looking? at the "well interfaces, average over the volume min or max ? and which stress Mises, 1st Piola ... a given tensor value or combination thereof ?
the latter min/max are a bit tricky, as a max stress can easily be driven by the mesh density and even mathematical singularities, in such cases a mesh sensitivity must be performed (should always be done anyhow) and you might decide to exclude some local region to avoid singularities. But that is based on your engineering judgment on the detailed results.

Also min max calculations tend to be long to calculate if you have a detailed model, you can speed up by reducing the interpolation level (check the tabs of the min max node or Derived Values node) at least until you have debugged your model. Then turn on back to default 2nd or 4th order.

As you want to plot one versus the other "deformation" versus "stress" you should use the Derived Values to calculate a table of the two values (for your different load or parameter cases) and plot that table (there is an icon to click and the plot is made up for you ready to be tailored ;)

Last comment: the "disp" variable in COMSOL is really sqrt(u^2+v^2+w^2) so it's a only positive value it might not be the most optimum parameter for you, just as the von Mises stress is a complex sqrt() average value it might not be the best indicator in your case, both due to their loss of the sign, but full vector or tensor dependencies are often tricky to interpret.

--
Good luck
Ivar
Hi Valerie I read many mixed questions in your request, and I believe you might also have several issues that can give "strange results" As your model seem slightly complex and not 100% clear to me from your few words about it (and I fully understand why not to give all details of proprietary studies on a public Forum ;). I'll also try to be rather generic in my replies (but I might also miss the "nail, be aware) First the easy one: scaling of deformations in COMSOL structural: 1) if you are in linear geometrical mode (the default) you will have an arbitrary graphical scaling adapted to give "visible" results, but you can always go into the Deformation sub node of your Plot node and adapt the scale to something you prefer "1" or "1000" or whatever (or even a function of space and time, I believe, I have never tried this as not really thought of an application for it, but it could be a way to differentiate the solved model deformation from i.e. some theoretical results you have from another study means ?) 2) If you are doing non-linear geometrical deformation, (which turns on for certain type of physics and physics combination) then by default COMSOL uses a Plot Deformation of "1", but again it's up to you to change that to something more suitable. Note this only applies to the plot results, It's not changing the Solver Data Set. so if you are populating the Derived Values node you are extracting original data from the solver results, no scaling applies, other than what you do with the "[Units]" conversions on your equations. Now you are looking for "deformation" versus "stress". So I believe you must better define "deformation" ie. the volume change, the ellipticity of a cylinder, the elongation ? .... and for the "stress" you must define where you are looking? at the "well interfaces, average over the volume min or max ? and which stress Mises, 1st Piola ... a given tensor value or combination thereof ? the latter min/max are a bit tricky, as a max stress can easily be driven by the mesh density and even mathematical singularities, in such cases a mesh sensitivity must be performed (should always be done anyhow) and you might decide to exclude some local region to avoid singularities. But that is based on your engineering judgment on the detailed results. Also min max calculations tend to be long to calculate if you have a detailed model, you can speed up by reducing the interpolation level (check the tabs of the min max node or Derived Values node) at least until you have debugged your model. Then turn on back to default 2nd or 4th order. As you want to plot one versus the other "deformation" versus "stress" you should use the Derived Values to calculate a table of the two values (for your different load or parameter cases) and plot that table (there is an icon to click and the plot is made up for you ready to be tailored ;) Last comment: the "disp" variable in COMSOL is really sqrt(u^2+v^2+w^2) so it's a only positive value it might not be the most optimum parameter for you, just as the von Mises stress is a complex sqrt() average value it might not be the best indicator in your case, both due to their loss of the sign, but full vector or tensor dependencies are often tricky to interpret. -- Good luck Ivar

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.