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 The information provided may be out of date.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.

Failed to find a solution. System matrix is zero. Returned solution is not converged.

Please login with a confirmed email address before reporting spam

I am trying to solve 2 coupled ODEs and am using the coefficient form PDE module to solve it. I have zeroed out the coefficients of the terms with time derivative. Btw this is a 1D problem where the variables are only a function of x. I get the error

Failed to find a solution. System matrix is zero. Returned solution is not converged.

After going through the comsol discussion forum I found that it must be a case of BCs. I have 4 boundary conditions for the problem: 3 drichlet boundary conditions and 1 constraint. The constraint contains derivatives of the 2 variables. Is this the problem.

And after some more runs the error changes and I get

Failed to find a solution.
Singular matrix.
There are 28 void equations (empty rows in matrix) for the variable mod1.omg.
at coordinates: (0.4), (0.433333), (0.466667), (0.5), (0.533333), ...
There are 28 void equations (empty rows in matrix) for the variable mod1.u.
at coordinates: (0.4), (0.433333), (0.466667), (0.5), (0.533333), ...
and similarly for the degrees of freedom (empty columns in matrix).
Returned solution is not converged.

There is also a warning :
New constraint force nodes detected: These are not stored.

5 Replies Last Post Sep 10, 2015, 10:28 a.m. EDT

Please login with a confirmed email address before reporting spam

Posted: 9 years ago Sep 8, 2015, 2:53 a.m. EDT
yes, I think the problem is related to the PDE that you defined which is not matching to the ODEs.

Try to change your boundary conditions.

Good luck
Sepideh
yes, I think the problem is related to the PDE that you defined which is not matching to the ODEs. Try to change your boundary conditions. Good luck Sepideh

Please login with a confirmed email address before reporting spam

Posted: 9 years ago Sep 8, 2015, 3:59 a.m. EDT

yes, I think the problem is related to the PDE that you defined which is not matching to the ODEs.

Try to change your boundary conditions.

Good luck
Sepideh


Could you please elaborate. Do you mean that the coefficient form PDE module cannot be used for defining my ODEs. The coefficients for the terms with time derivative ea and da are put to zero. The rest of the coefficients are in such a way that they match my ode. (though all the coefficents are themselves a function of the first derivative of the variables. IS THIS Causing the problems.

Ive tried with a few boundary conditions and all get the same errors
[QUOTE] yes, I think the problem is related to the PDE that you defined which is not matching to the ODEs. Try to change your boundary conditions. Good luck Sepideh [/QUOTE] Could you please elaborate. Do you mean that the coefficient form PDE module cannot be used for defining my ODEs. The coefficients for the terms with time derivative ea and da are put to zero. The rest of the coefficients are in such a way that they match my ode. (though all the coefficents are themselves a function of the first derivative of the variables. IS THIS Causing the problems. Ive tried with a few boundary conditions and all get the same errors

Please login with a confirmed email address before reporting spam

Posted: 9 years ago Sep 8, 2015, 3:59 a.m. EDT

yes, I think the problem is related to the PDE that you defined which is not matching to the ODEs.

Try to change your boundary conditions.

Good luck
Sepideh


Could you please elaborate. Do you mean that the coefficient form PDE module cannot be used for defining my ODEs. The coefficients for the terms with time derivative ea and da are put to zero. The rest of the coefficients are in such a way that they match my ode. (though all the coefficents are themselves a function of the first derivative of the variables. IS THIS Causing the problems.

Ive tried with a few boundary conditions and all get the same errors
[QUOTE] yes, I think the problem is related to the PDE that you defined which is not matching to the ODEs. Try to change your boundary conditions. Good luck Sepideh [/QUOTE] Could you please elaborate. Do you mean that the coefficient form PDE module cannot be used for defining my ODEs. The coefficients for the terms with time derivative ea and da are put to zero. The rest of the coefficients are in such a way that they match my ode. (though all the coefficents are themselves a function of the first derivative of the variables. IS THIS Causing the problems. Ive tried with a few boundary conditions and all get the same errors

Please login with a confirmed email address before reporting spam

Posted: 9 years ago Sep 8, 2015, 9:28 a.m. EDT
I tried to break it down and find the error was using simpler equations.

i have 1 equation:

and 2 BC : and

Here when i use mu as a constant it is readily solved. But I have a problem when i use


where




It doesn't converge.... and then i started tinkering with it and i found that when

it does converge and I get a solution. Do you know the reason for this? And how I can get the earlier equation to converge.
I tried to break it down and find the error was using simpler equations. i have 1 equation: [math] (1/r^2 )* d/dr(r^2 *mu*(du/dr -u/r))=0 [/math] and 2 BC : [math] u(0.3)=1[/math] and [math] u(1.3)=0 [/math] Here when i use mu as a constant it is readily solved. But I have a problem when i use [math] mu= mu_v + mu_s/gamma [/math] where [math] gamma = sqrt((du/dr -u/r)^2) , [/math] [math] mu_v = const, [/math] [math] mu_s=const[/math] It doesn't converge.... and then i started tinkering with it and i found that when [math] mu= mu_v + mu_s*gamma [/math] it does converge and I get a solution. Do you know the reason for this? And how I can get the earlier equation to converge.

Please login with a confirmed email address before reporting spam

Posted: 9 years ago Sep 10, 2015, 10:28 a.m. EDT
Do you have a singularity in gamma^-1? That may take precedence in the computation before the simplification cancels out gamma in the 2nd term in mu, and might explain why when multiplied you get convergence.

Why not do the simplification and set it to (r^2*(mu_v*(du/dr-u/r)+mu_s))?

-Ado
Do you have a singularity in gamma^-1? That may take precedence in the computation before the simplification cancels out gamma in the 2nd term in mu, and might explain why when multiplied you get convergence. Why not do the simplification and set it to (r^2*(mu_v*(du/dr-u/r)+mu_s))? -Ado

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.