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 The information provided may be out of date.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.

2-D plate model

Please login with a confirmed email address before reporting spam

Hi,

I am trying to build simple model to analyse the eigenfrequencies of a freely vibrating irregularly-shaped 2-D plate. I have selected the plate model in the structural mechanics module. The shape I have imported from a .dxf CAD file. I have around 8000 elements, an amount which I believe to suffice.

When I try to solve my model all I get is an error of undefined value in the stiffness matrix. For a 3-D square plate with the same settings, everything seems to be working fine.

Can someone help me detect what have I forgotten in this model, because it does not work?

Yours,
Henna from snowy Finland






6 Replies Last Post 26 nov 2011, 06:21 GMT-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 12 gen 2011, 10:01 GMT-5
Hi

I have some problems too to run your file too, try to redefine a fresh model.
You must know by now that v4.0 is sensitive to model changes and modifications, you should get hands on the latest 4.1 its much smoother :)

If I extract your geometry and remake a plate model it works OK for me

--
Good luck
Ivar
Hi I have some problems too to run your file too, try to redefine a fresh model. You must know by now that v4.0 is sensitive to model changes and modifications, you should get hands on the latest 4.1 its much smoother :) If I extract your geometry and remake a plate model it works OK for me -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 12 gen 2011, 10:28 GMT-5
Hi,

and thanks for a swift reply. I did not think of redefining the model before. I exported the geometry, and made a new model. Yet, I get the very same problem persisting; an undefined value found (in the stiffness matrix). I have also tried simply restarting COMSOL..Debugging one's models is very difficult in COMSOL.

It might be that I am missing something very simple in my model. There is nothing in particular that I should do to the imported geometry, before meshing? Is it possible for you to post the plate model online, so that I can make a comparison?

However, since you managed to run the model, I think the problem could also be my computer, or my version of the software.

Thanks again.

Henna
Hi, and thanks for a swift reply. I did not think of redefining the model before. I exported the geometry, and made a new model. Yet, I get the very same problem persisting; an undefined value found (in the stiffness matrix). I have also tried simply restarting COMSOL..Debugging one's models is very difficult in COMSOL. It might be that I am missing something very simple in my model. There is nothing in particular that I should do to the imported geometry, before meshing? Is it possible for you to post the plate model online, so that I can make a comparison? However, since you managed to run the model, I think the problem could also be my computer, or my version of the software. Thanks again. Henna

Magnus Ringh COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 12 gen 2011, 10:57 GMT-5
Hi,

The problem is that the model uses a Poisson's ratio that is exactly 0.5, which leads to a division by zero in the computation (and hence an error for Inf in the matrix rows) for a linear elastic material model. That material model uses a factor (1-2*nu), where nu is Poisson's ratio, in the denominator for the elasticity matrix, so a value of 0.5 is not allowed.

Try using a Poisson's ratio that is smaller than 0.5. Then the computation should run just fine.

Best regards,
Magnus Ringh, COMSOL
Hi, The problem is that the model uses a Poisson's ratio that is exactly 0.5, which leads to a division by zero in the computation (and hence an error for Inf in the matrix rows) for a linear elastic material model. That material model uses a factor (1-2*nu), where nu is Poisson's ratio, in the denominator for the elasticity matrix, so a value of 0.5 is not allowed. Try using a Poisson's ratio that is smaller than 0.5. Then the computation should run just fine. Best regards, Magnus Ringh, COMSOL

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 13 gen 2011, 01:45 GMT-5
Hi

I missed that one ;), as I redefined the material

--
Good luck
Ivar
Hi I missed that one ;), as I redefined the material -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 13 gen 2011, 06:47 GMT-5
Oh, yes, indeed. That is the little thing that I missed. I had just selected some of the material properties wrong after reducing the material from orthotropic to isotropic , and the model dimensions from 3D to 2D. One should be careful to remember the actual physics behind!

Thank you very much!
Oh, yes, indeed. That is the little thing that I missed. I had just selected some of the material properties wrong after reducing the material from orthotropic to isotropic , and the model dimensions from 3D to 2D. One should be careful to remember the actual physics behind! Thank you very much!

Muhammad Mohsin Rehman

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 26 nov 2011, 06:21 GMT-5
Hi,

I am also facing the same difficulty for microwaves heating model. Initially, I was simulating a layer of oil lying over a layer of oil and was getting some results, although not convinced.
Then I defined another layer of oil below the layer of water and tried to simulate it but it gave me the following error:


Undefined value found.
- Detail: Undefined value found in the stiffness matrix.
There are 1160 equations giving NaN/Inf in the matrix rows for the variable mod1.E10
at coordinates: (0.0016633,0.00180792), (0.00332659,0.00361583), (0.00382242,0), (0.00714901,0.00361583), (0.00764484,0), ...
There are 290 equations giving NaN/Inf in the matrix rows for the variable mod1.E20
at coordinates: (0.00548572,0.00180792), (0.0016633,0.00514125), (0.00469918,0.00847458), (0.0085216,0.00514125), (0.0191354,0.00333333), ...
There are 290 equations giving NaN/Inf in the matrix rows for the variable mod1.E21
at coordinates: (0.00548572,0.00180792), (0.0016633,0.00514125), (0.00469918,0.00847458), (0.0085216,0.00514125), (0.0191354,0.00333333), ...
There are 871 equations giving NaN/Inf in the matrix rows for the variable mod1.Ez
at coordinates: (0,0), (0.0114673,0), (0.00573363,0), (0.00249494,0.00271187), (0.00498989,0.00542375), ...
and similarly for the degrees of freedom, NaN/Inf in the matrix columns


I have not even changed any other material, BC or property of the model and there is no red flag in the material section either but still I am facing this error, why???

Can you please help!
Hi, I am also facing the same difficulty for microwaves heating model. Initially, I was simulating a layer of oil lying over a layer of oil and was getting some results, although not convinced. Then I defined another layer of oil below the layer of water and tried to simulate it but it gave me the following error: Undefined value found. - Detail: Undefined value found in the stiffness matrix. There are 1160 equations giving NaN/Inf in the matrix rows for the variable mod1.E10 at coordinates: (0.0016633,0.00180792), (0.00332659,0.00361583), (0.00382242,0), (0.00714901,0.00361583), (0.00764484,0), ... There are 290 equations giving NaN/Inf in the matrix rows for the variable mod1.E20 at coordinates: (0.00548572,0.00180792), (0.0016633,0.00514125), (0.00469918,0.00847458), (0.0085216,0.00514125), (0.0191354,0.00333333), ... There are 290 equations giving NaN/Inf in the matrix rows for the variable mod1.E21 at coordinates: (0.00548572,0.00180792), (0.0016633,0.00514125), (0.00469918,0.00847458), (0.0085216,0.00514125), (0.0191354,0.00333333), ... There are 871 equations giving NaN/Inf in the matrix rows for the variable mod1.Ez at coordinates: (0,0), (0.0114673,0), (0.00573363,0), (0.00249494,0.00271187), (0.00498989,0.00542375), ... and similarly for the degrees of freedom, NaN/Inf in the matrix columns I have not even changed any other material, BC or property of the model and there is no red flag in the material section either but still I am facing this error, why??? Can you please help!

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.