[COMSOL_Users] Error: Undefined value found in the equation residual vector

[Mazyar]

Hi,

I am doing a structural modeling and I am getting this error:
------------------------------
Undefined value found.
- Detail: Undefined value found in the equation residual vector.
There are 628 degrees of freedom giving NaN/Inf in the vector for the variable mod1.solid.pw
at coordinates: (0.00452,0), (0.00443445,0), ...
------------------------------

Sounds like those points stated in the error are on my boundaries.

My material is hyperelastic and I get the above error after I apply my
user-defined strain energy function. When I use incompressible material
assumption the above message in invoked, but when I try to see how
compressible material behaves I get this error:
------------------------------
Failed to find a solution.
System matrix is zero.
Returned solution is not converged.
------------------------------
I have added a few variables, and I believe that my strain energy
function should work fine for the entire domain. Also, I am using
exactly the same boundary conditions that I used with a built-in strain
energy function and I could get correct solution. So I expect that my
BC's should work fine with the new function.

I also doubt that my user-defined strain energy function is causing the
problem, because when I use a constant value for the strain energy
function, I still get exactly the same error.

I changed the initial values for displacement in the Initial Values from
zero to another small value, but had no luck to get rid of the error.
Manually entering a Scaling factor for the variable mod1.solid.pw did
not help either.

I appreciate your help suggesting what the problem can be from.

Jessica

[marjan zakerin]

Hi Dear Mazyar,

 In hyperplastic part, select almost compressible material (deselect almost incompressible material). I think I solved the problem like that. In fact polymers, specially cross linked polymers which use neo hookian formalism are compressible.

Yours,
Marjan


________________________________
From: Mazyar <***@yahoo.com>
To: "***@yahoogroups.com" <***@yahoogroups.com>
Sent: Friday, August 19, 2011 5:56 PM
Subject: [COMSOL_Users] Error: Undefined value found in the equation residual vector


 
Hi,

I am doing a structural modeling and I am getting this error:
------------------------------
Undefined value found.
- Detail: Undefined value found in the equation residual vector.
There are 628 degrees of freedom giving NaN/Inf in the vector for the variable mod1.solid.pw
at coordinates: (0.00452,0), (0.00443445,0), ...
------------------------------

Sounds like those points stated in the error are on my boundaries.

My material is hyperelastic and I get the above error after I apply my
user-defined strain energy function. When I use incompressible material
assumption the above message in invoked, but when I try to see how
compressible material behaves I get this error:
------------------------------
Failed to find a solution.
System matrix is zero.
Returned solution is not converged.
------------------------------
I have added a few variables, and I believe that my strain energy
function should work fine for the entire domain. Also, I am using
exactly the same boundary conditions that I used with a built-in strain
energy function and I could get correct solution. So I expect that my
BC's should work fine with the new function.

I also doubt that my user-defined strain energy function is causing the
problem, because when I use a constant value for the strain energy
function, I still get exactly the same error.

I changed the initial values for displacement in the Initial Values from
zero to another small value, but had no luck to get rid of the error.
Manually entering a Scaling factor for the variable mod1.solid.pw did
not help either.

I appreciate your help suggesting what the problem can be from.

Jessica

[Mazyar]

Hi Marjan,

Thank you very much for your reply.
As I had mentioned in my posting, unfortunately I get different errors for both incompressible and compressible materials. Also, I am using a user-defined strain energy function. When I use Neo-hookean, I do not run into any problem.
Your further input is appreciated.

Bests,
Mazyar



________________________________
From: marjan zakerin <***@yahoo.com>
To: "***@yahoogroups.com" <***@yahoogroups.com>
Sent: Monday, August 22, 2011 3:00 PM
Subject: Re: [COMSOL_Users] Error: Undefined value found in the equation residual vector


 
Hi Dear Mazyar,

 In hyperplastic part, select almost compressible material (deselect almost incompressible material). I think I solved the problem like that. In fact polymers, specially cross linked polymers which use neo hookian formalism are compressible.

Yours,
Marjan


________________________________
From: Mazyar <***@yahoo.com>
To: "***@yahoogroups.com" <***@yahoogroups.com>
Sent: Friday, August 19, 2011 5:56 PM
Subject: [COMSOL_Users] Error: Undefined value found in the equation residual vector


 
Hi,

I am doing a structural modeling and I am getting this error:
------------------------------
Undefined value found.
- Detail: Undefined value found in the equation residual vector.
There are 628 degrees of freedom giving NaN/Inf in the vector for the variable mod1.solid.pw
at coordinates: (0.00452,0), (0.00443445,0), ...
------------------------------

Sounds like those points stated in the error are on my boundaries.

My material is hyperelastic and I get the above error after I apply my
user-defined strain energy function. When I use incompressible material
assumption the above message in invoked, but when I try to see how
compressible material behaves I get this error:
------------------------------
Failed to find a solution.
System matrix is zero.
Returned solution is not converged.
------------------------------
I have added a few variables, and I believe that my strain energy
function should work fine for the entire domain. Also, I am using
exactly the same boundary conditions that I used with a built-in strain
energy function and I could get correct solution. So I expect that my
BC's should work fine with the new function.

I also doubt that my user-defined strain energy function is causing the
problem, because when I use a constant value for the strain energy
function, I still get exactly the same error.

I changed the initial values for displacement in the Initial Values from
zero to another small value, but had no luck to get rid of the error.
Manually entering a Scaling factor for the variable mod1.solid.pw did
not help either.

I appreciate your help suggesting what the problem can be from.

Mazyar

[marjan zakerin]

Dear Mazyar,

 In order to be sure that your errors arise from hyperelastic part, please first try to choose linear elastic properties for these parts that are supposed to be hyperelastic and try to optimize your system as much as possible. When there is no error, change the material to hyperelastic. In the next step enter user defined strain stress. That is the best way.   

Apart from that, Just look at your structure and consider the amount of deformation that you have with respect to the size of your slave geometry. If the deformation of slave (deforming part) is more that 20%, then you should change from Neo kookean to Moony-Rivlin solid which is complex. Therefore, I suggest that you consider a very small deformation to be able to stick to Neo hookean. On the other hand, be careful about Poisson Numer. It should not be 0.5 although you suppose that it is a rubber. 

 Yours,
Marjan


________________________________
From: Mazyar <***@yahoo.com>
To: "***@yahoogroups.com" <***@yahoogroups.com>
Sent: Monday, August 22, 2011 11:13 PM
Subject: Re: [COMSOL_Users] Error: Undefined value found in the equation residual vector


 
Hi Marjan,

Thank you very much for your reply.
As I had mentioned in my posting, unfortunately I get different errors for both incompressible and compressible materials. Also, I am using a user-defined strain energy function. When I use Neo-hookean, I do not run into any problem.
Your further input is appreciated.

Bests,
Mazyar


________________________________
From: marjan zakerin <***@yahoo.com>
To: "***@yahoogroups.com" <***@yahoogroups.com>
Sent: Monday, August 22, 2011 3:00 PM
Subject: Re: [COMSOL_Users] Error: Undefined value found in the equation residual vector


 
Hi Dear Mazyar,

 In hyperplastic part, select almost compressible material (deselect almost incompressible material). I think I solved the problem like that. In fact polymers, specially cross linked polymers which use neo hookian formalism are compressible.

Yours,
Marjan


________________________________
From: Mazyar <***@yahoo.com>
To: "***@yahoogroups.com" <***@yahoogroups.com>
Sent: Friday, August 19, 2011 5:56 PM
Subject: [COMSOL_Users] Error: Undefined value found in the equation residual vector


 
Hi,

I am doing a structural modeling and I am getting this error:
------------------------------
Undefined value found.
- Detail: Undefined value found in the equation residual vector.
There are 628 degrees of freedom giving NaN/Inf in the vector for the variable mod1.solid.pw
at coordinates: (0.00452,0), (0.00443445,0), ...
------------------------------

Sounds like those points stated in the error are on my boundaries.

My material is hyperelastic and I get the above error after I apply my
user-defined strain energy function. When I use incompressible material
assumption the above message in invoked, but when I try to see how
compressible material behaves I get this error:
------------------------------
Failed to find a solution.
System matrix is zero.
Returned solution is not converged.
------------------------------
I have added a few variables, and I believe that my strain energy
function should work fine for the entire domain. Also, I am using
exactly the same boundary conditions that I used with a built-in strain
energy function and I could get correct solution. So I expect that my
BC's should work fine with the new function.

I also doubt that my user-defined strain energy function is causing the
problem, because when I use a constant value for the strain energy
function, I still get exactly the same error.

I changed the initial values for displacement in the Initial Values from
zero to another small value, but had no luck to get rid of the error.
Manually entering a Scaling factor for the variable mod1.solid.pw did
not help either.

I appreciate your help suggesting what the problem can be from.

Mazyar

[Mazyar]

Hi Marjan,

Thanks for your reply.
I agree with you and like that approach, however, the error that I get is invoked with implementation of a strain energy function that is specific to hyperelastic materials; and these materials cannot be correctly modeled by an elastic linear model.
Regarding the usage of Neo-Hookean, again I agree with you that that model is limited to 20-30% deformations. In my case, I did not have problem with Neo-hookean and experimental data pretty much validate that. At this point, instead of using Moony-Rivilin, I am interested in using a Fung-type model, which is also giving me the error.

Thanks,
Mazyar



________________________________
From: marjan zakerin <***@yahoo.com>
To: "***@yahoogroups.com" <***@yahoogroups.com>
Sent: Tuesday, August 23, 2011 2:54 AM
Subject: Re: [COMSOL_Users] Error: Undefined value found in the equation residual vector


 
Dear Mazyar,

 In order to be sure that your errors arise from hyperelastic part, please first try to choose linear elastic properties for these parts that are supposed to be hyperelastic and try to optimize your system as much as possible. When there is no error, change the material to hyperelastic. In the next step enter user defined strain stress. That is the best way.   

Apart from that, Just look at your structure and consider the amount of deformation that you have with respect to the size of your slave geometry. If the deformation of slave (deforming part) is more that 20%, then you should change from Neo kookean to Moony-Rivlin solid which is complex. Therefore, I suggest that you consider a very small deformation to be able to stick to Neo hookean. On the other hand, be careful about Poisson Numer. It should not be 0.5 although you suppose that it is a rubber. 

 Yours,
Marjan


________________________________
From: Mazyar <***@yahoo.com>
To: "***@yahoogroups.com" <***@yahoogroups.com>
Sent: Monday, August 22, 2011 11:13 PM
Subject: Re: [COMSOL_Users] Error: Undefined value found in the equation residual vector


 
Hi Marjan,

Thank you very much for your reply.
As I had mentioned in my posting, unfortunately I get different errors for both incompressible and compressible materials. Also, I am using a user-defined strain energy function. When I use Neo-hookean, I do not run into any problem.
Your further input is appreciated.

Bests,
Mazyar


________________________________
From: marjan zakerin <***@yahoo.com>
To: "***@yahoogroups.com" <***@yahoogroups.com>
Sent: Monday, August 22, 2011 3:00 PM
Subject: Re: [COMSOL_Users] Error: Undefined value found in the equation residual vector


 
Hi Dear Mazyar,

 In hyperplastic part, select almost compressible material (deselect almost incompressible material). I think I solved the problem like that. In fact polymers, specially cross linked polymers which use neo hookian formalism are compressible.

Yours,
Marjan


________________________________
From: Mazyar <***@yahoo.com>
To: "***@yahoogroups.com" <***@yahoogroups.com>
Sent: Friday, August 19, 2011 5:56 PM
Subject: [COMSOL_Users] Error: Undefined value found in the equation residual vector


 
Hi,

I am doing a structural modeling and I am getting this error:
------------------------------
Undefined value found.
- Detail: Undefined value found in the equation residual vector.
There are 628 degrees of freedom giving NaN/Inf in the vector for the variable mod1.solid.pw
at coordinates: (0.00452,0), (0.00443445,0), ...
------------------------------

Sounds like those points stated in the error are on my boundaries.

My material is hyperelastic and I get the above error after I apply my
user-defined strain energy function. When I use incompressible material
assumption the above message in invoked, but when I try to see how
compressible material behaves I get this error:
------------------------------
Failed to find a solution.
System matrix is zero.
Returned solution is not converged.
------------------------------
I have added a few variables, and I believe that my strain energy
function should work fine for the entire domain. Also, I am using
exactly the same boundary conditions that I used with a built-in strain
energy function and I could get correct solution. So I expect that my
BC's should work fine with the new function.

I also doubt that my user-defined strain energy function is causing the
problem, because when I use a constant value for the strain energy
function, I still get exactly the same error.

I changed the initial values for displacement in the Initial Values from
zero to another small value, but had no luck to get rid of the error.
Manually entering a Scaling factor for the variable mod1.solid.pw did
not help either.

I appreciate your help suggesting what the problem can be from.

Mazyar