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 The information provided may be out of date.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.

Joule Heating Model Wont Solve

Please login with a confirmed email address before reporting spam

Hello,

Our whole class is attemting to develop this model but none of use know how to use comsol well.

I made a basic model with 2 2 blocks end to end. I specified the materials, the heat source as one block, and the heat flux as all surfaces of the other block.

For our more advanced model the geometry includes 3 blocks of two different materials, one as the heat source, all surfaces as the heat flux.

I have tried a few different meshes of different sizes - user defined, physics defined, and swept.

The model will not solve and this is the error:

Failed to find a solution.
Singular matrix.

For mesh-case 0 there are 693 void equations (empty rows in matrix) for the variable mod1.V
at coordinates: (0.0015,-0.0015,0), (0.0015,-0.0015,0.00015), (0.0015,-0.0015,0.0003), (0.0015,-0.0009,0), (0.0015,-0.0012,0), ...
and similarly for the degrees of freedom (empty columns in matrix)
Returned solution has not converged.



Are there some basic steps we can go through to find the different reasons a model will not solve? Its tough just getting the general error message and not knowing if the error lies with the model or the solver settings.

I have attached the model to this post, any advice will be very appreciated by our class.

Thanks.

Sincerely,

Alex


4 Replies Last Post 16 juin 2011, 03:47 UTC−4
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 11 mars 2011, 15:57 UTC−5
Hi

void equations, singular matrices etc typically means that some important material data or insufficient BC have been defined.

so 1) check your material data to see if there are any red signs (I see already one there ;) Always recheck the materials after you have finished with the physics and before you run your model, as while you adapt the physics you might call for new material properties and not always all are defined

Then for model, you need to think physics and start with the default settings and add only subnode as required (and mastered). One thing for HT :you need energy in, but normally you let something out too i.e if you define a fixed temperature wall it means that this wall will provide or sink thermal power/energy over its surface to keep the temperature constant. Normally one starts with everything isolated, one side/boundary at a given temperature and an opposite side with an energy flux (or another temperature)

Once you master HT alone, you can link in Joule heating i.e. Voltage and current drop to generate a body source depending on the impedance of the model

Read carefully the doc and take the exercices in the model library, its worth to start simple (there are also the videos on the main web site)

--
Good luck
Ivar
Hi void equations, singular matrices etc typically means that some important material data or insufficient BC have been defined. so 1) check your material data to see if there are any red signs (I see already one there ;) Always recheck the materials after you have finished with the physics and before you run your model, as while you adapt the physics you might call for new material properties and not always all are defined Then for model, you need to think physics and start with the default settings and add only subnode as required (and mastered). One thing for HT :you need energy in, but normally you let something out too i.e if you define a fixed temperature wall it means that this wall will provide or sink thermal power/energy over its surface to keep the temperature constant. Normally one starts with everything isolated, one side/boundary at a given temperature and an opposite side with an energy flux (or another temperature) Once you master HT alone, you can link in Joule heating i.e. Voltage and current drop to generate a body source depending on the impedance of the model Read carefully the doc and take the exercices in the model library, its worth to start simple (there are also the videos on the main web site) -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 15 juin 2011, 08:26 UTC−4
Hi Ivar,

I found this post google'ing. I have a similar problem, but I know it is not a material problem. The attached model will not solve as it is now, but if blocks 6 & 7 and their derivatives are removed it does solve (with a small obvious change in the applied voltage bc). I can not figure out what is wrong with the added blocks and would very much appreciate any input!

Best,
Robert
Hi Ivar, I found this post google'ing. I have a similar problem, but I know it is not a material problem. The attached model will not solve as it is now, but if blocks 6 & 7 and their derivatives are removed it does solve (with a small obvious change in the applied voltage bc). I can not figure out what is wrong with the added blocks and would very much appreciate any input! Best, Robert


Lechoslaw Krolikowski

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 15 juin 2011, 12:37 UTC−4
Hi Alex,

You are looking for steady state in which the heat inflow must be equal to the heat outflow.
But in your model (new model.mph) to achieve this equality is impossible because
heat inflow = 5000[W/m^3]*V3[m^3] + 5e5[W/m^2]*S[m^2] = const > 0,
heat outflow = 0,
where
V3 - volume of domain 3,
S - external surface of all domains.

Regards,
Andrzej.
Hi Alex, You are looking for steady state in which the heat inflow must be equal to the heat outflow. But in your model (new model.mph) to achieve this equality is impossible because heat inflow = 5000[W/m^3]*V3[m^3] + 5e5[W/m^2]*S[m^2] = const > 0, heat outflow = 0, where V3 - volume of domain 3, S - external surface of all domains. Regards, Andrzej.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 16 juin 2011, 03:47 UTC−4
Hi,

The problem message states that there are void equations for the variable mod1.V, that is your variable Voltage is not being defined thus I suspect that comsol is having issues with the properties you set for Silicon Nitride: you set a zero conductivity which may be real but it forces comsol not to define the variable Voltage.

Furthermore there is a relative permittivity missing from the Platinum properties, which I think should be set to 1.

If you set a small value of conductivity for silicon Nitride you can avoid the error of no equations being defined for the Voltage, however you must check all of your boundary conditions and initial values to make sure you are representing the problem correctly.

Cheers
Hi, The problem message states that there are void equations for the variable mod1.V, that is your variable Voltage is not being defined thus I suspect that comsol is having issues with the properties you set for Silicon Nitride: you set a zero conductivity which may be real but it forces comsol not to define the variable Voltage. Furthermore there is a relative permittivity missing from the Platinum properties, which I think should be set to 1. If you set a small value of conductivity for silicon Nitride you can avoid the error of no equations being defined for the Voltage, however you must check all of your boundary conditions and initial values to make sure you are representing the problem correctly. Cheers

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.