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.
Number of degrees of freedom solved for: 0.
Posted 23 mar 2011, 11:33 GMT-4 Low-Frequency Electromagnetics, Acoustics & Vibrations, Studies & Solvers Version 4.2a, Version 5.2 21 Replies
Please login with a confirmed email address before reporting spam
Hello all,
When I try to use 4.0, I met with this problem when solving.
Does anyone know how to fix this? I think it might be that I did not add the mesh to the solver, which is different from 3.5a.
Thank you.
When I try to use 4.0, I met with this problem when solving.
Does anyone know how to fix this? I think it might be that I did not add the mesh to the solver, which is different from 3.5a.
Thank you.
21 Replies Last Post 12 apr 2016, 09:11 GMT-4
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Maybe you could upload your model? That would help.
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi,
I have attached the model.
Everything has been set okay, you could try to solve it in 4.0a.
Thank you.
I have attached the model.
Everything has been set okay, you could try to solve it in 4.0a.
Thank you.
Attachments:
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi,
The reason why COSMOL does not compute any degrees of freedom is that the "Solve for this field" check box for the only dependent variable, the magnetic vector potential A, is cleared. The default is to solve for all fields in all physics interfaces in a model, so it is unclear to me why it is not selected.
Select the "Solve for this field" check box in the Settings window for Study 1>Solver Configurations>Solver 1>
Dependent Variables 1>mod1_A (and then re-solve) to get a solution (or delete the solver to use the default solver settings).
Another modeling tip: When you use the "From material" setting for material properties, you only need one Ampère's Law node for all domains where a material is defined. COMSOL Multiphysics picks the material properties for each domain from the materials assigned to the domains in the Materials branch. You only need additional Ampère's Law nodes for user-defined material properties.
Best regards,
Magnus Ringh, COMSOL
The reason why COSMOL does not compute any degrees of freedom is that the "Solve for this field" check box for the only dependent variable, the magnetic vector potential A, is cleared. The default is to solve for all fields in all physics interfaces in a model, so it is unclear to me why it is not selected.
Select the "Solve for this field" check box in the Settings window for Study 1>Solver Configurations>Solver 1>
Dependent Variables 1>mod1_A (and then re-solve) to get a solution (or delete the solver to use the default solver settings).
Another modeling tip: When you use the "From material" setting for material properties, you only need one Ampère's Law node for all domains where a material is defined. COMSOL Multiphysics picks the material properties for each domain from the materials assigned to the domains in the Materials branch. You only need additional Ampère's Law nodes for user-defined material properties.
Best regards,
Magnus Ringh, COMSOL
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hello,
I have kind of the same problem, but for me I do not succeed in ticking "solve for this field". Could you please have a look on the attached file and tell me what is wrong with what I did... Maybe a variable definition problem.
I would really appreciate help from all of you.
Thank you,
Nicolas
I have kind of the same problem, but for me I do not succeed in ticking "solve for this field". Could you please have a look on the attached file and tell me what is wrong with what I did... Maybe a variable definition problem.
I would really appreciate help from all of you.
Thank you,
Nicolas
Attachments:
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi,
I am having the same problem as Nicolas over here. I have opened the default solver and went exactly to the same location, but the box that needs to be checked is whitened, and cannot be ticked. I cannot change it, and the model solves for 0 degrees of freedom, and thus, is pretty useless.
Please help me, the model is attached.
Thanks again,
~Shoval
p.s.(edited)
I forgot to mention that I am using 4.2a.
I am having the same problem as Nicolas over here. I have opened the default solver and went exactly to the same location, but the box that needs to be checked is whitened, and cannot be ticked. I cannot change it, and the model solves for 0 degrees of freedom, and thus, is pretty useless.
Please help me, the model is attached.
Thanks again,
~Shoval
p.s.(edited)
I forgot to mention that I am using 4.2a.
Attachments:
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
I'd like to third this problem. It seems that there is another setting somewhere that's deactivating the "Solve for this field" check box. I'm not sure what this would be, but it's happened to me in several models in which I try to use a custom study rather than the predefined studies. If anybody has an answer to this problem, it would be helpful.
Thanks,
-Matt
Thanks,
-Matt
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi,
uh since I posted this I have made some tests, and it seems that the combination of MEF with time-dependent makes this problem. I changed to MF and it worked perfectly, I have no idea why.
Shoval
uh since I posted this I have made some tests, and it seems that the combination of MEF with time-dependent makes this problem. I changed to MF and it worked perfectly, I have no idea why.
Shoval
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
can you explain what is MEF and MF?
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Add physics --> AC/DC --> Magnetic Field=MF, Mangetic and Electric Field = MEF.
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
the differences between MF and MEF is in the number of dependent variables. MF solves for A the magnetic vector potential, while MEF solves for A and V the Electric Potential, this means many more DoFs (degrees of Freedom) more RAM and time to solve, including more BCs to define But you get more info out.
So it's all depending on what you want to do.
Note: in MEF all materials need some conduction, otherwise you cannot define V and you will get an error
In any case my suggestion, always start simple hence MF, if absolutely required then add complexity: i.e. rewrite the model for MEF.
MEF is used for rotating motors, or for Eddy current simulations, when you induce back electromagnetic forces. most magnetics (including current flow can me solved with MF, while MFNC is magnetics without currents
--
Good luck
Ivar
the differences between MF and MEF is in the number of dependent variables. MF solves for A the magnetic vector potential, while MEF solves for A and V the Electric Potential, this means many more DoFs (degrees of Freedom) more RAM and time to solve, including more BCs to define But you get more info out.
So it's all depending on what you want to do.
Note: in MEF all materials need some conduction, otherwise you cannot define V and you will get an error
In any case my suggestion, always start simple hence MF, if absolutely required then add complexity: i.e. rewrite the model for MEF.
MEF is used for rotating motors, or for Eddy current simulations, when you induce back electromagnetic forces. most magnetics (including current flow can me solved with MF, while MFNC is magnetics without currents
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi again,
I am once again using a model in which a time dependent solver combined with the MEF physics, and once again I have no option to tick the "solve for this field" box for either A or V.
This time MEF is especially required for the model, so I have no other option and need your help. The model is huge - after meshing it has 1.5 million elements - definitely too large to be uploaded here.
Help, anyone? I'm still using 4.2a, has this been fixed in 4.3/4.3a?
Thanks,
Shoval
I am once again using a model in which a time dependent solver combined with the MEF physics, and once again I have no option to tick the "solve for this field" box for either A or V.
This time MEF is especially required for the model, so I have no other option and need your help. The model is huge - after meshing it has 1.5 million elements - definitely too large to be uploaded here.
Help, anyone? I'm still using 4.2a, has this been fixed in 4.3/4.3a?
Thanks,
Shoval
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
if you check the equations you will see that MEF is not set up to be solved in time domain, you can only use it in frequency domain / harmonic development. my understanding is that the ACDC hypothesis imposes that everything is "instantaneous", else you should use RF
Now, if you have several physics and one is time dependent / transient, then you must analyse the MEF for each of these time steps. Indeed in v4.3a the parametric sweeps solver node allowsnow to point into another time dependent solver so this is possible
--
Good luck
Ivar
if you check the equations you will see that MEF is not set up to be solved in time domain, you can only use it in frequency domain / harmonic development. my understanding is that the ACDC hypothesis imposes that everything is "instantaneous", else you should use RF
Now, if you have several physics and one is time dependent / transient, then you must analyse the MEF for each of these time steps. Indeed in v4.3a the parametric sweeps solver node allowsnow to point into another time dependent solver so this is possible
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Thank you Ivar for your insight; but this makes it a hard work for me to extract the needed info from my model.
If I'm still using 4.2a(and probably won't upgrade in the near relevant future), how should I proceed, supposing that:
1) my model consists of 2 materials - one is with magnetic properties that require MF/MEF
2) my model uses three terminals, of which two are time dependent supply of voltage
3) I eventually need to extract both magnetic field results and electric potential results.
Thanks for your help,
Shoval
If I'm still using 4.2a(and probably won't upgrade in the near relevant future), how should I proceed, supposing that:
1) my model consists of 2 materials - one is with magnetic properties that require MF/MEF
2) my model uses three terminals, of which two are time dependent supply of voltage
3) I eventually need to extract both magnetic field results and electric potential results.
Thanks for your help,
Shoval
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
the capacitance is a geometrical and material epsilon_r related issue so if these are static, you can always run a few stationary ES cases to get the different capacitances.
Magnetic material simulation requires some surrounding, as the EM field expands also into air/vucuum and you need this geometrical zone to loop around your filed lines, else your reults would not be "physical".
So you seem to need MEF, well ths is OK add some extra domain nodes for the other material, and in MEF you can have several terminals, it's only that the predefined c11 variables of COMSOL are not defined (in my v4.3a and if I have got this right ;)
--
Good luck
Ivar
the capacitance is a geometrical and material epsilon_r related issue so if these are static, you can always run a few stationary ES cases to get the different capacitances.
Magnetic material simulation requires some surrounding, as the EM field expands also into air/vucuum and you need this geometrical zone to loop around your filed lines, else your reults would not be "physical".
So you seem to need MEF, well ths is OK add some extra domain nodes for the other material, and in MEF you can have several terminals, it's only that the predefined c11 variables of COMSOL are not defined (in my v4.3a and if I have got this right ;)
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi again,
I believe you misunderstood my question.
At the moment, my model looks like this:
I have a core of magnetic material that uses a BH-curve via ampere's law (and current conservation since it's MEF).
I have a sphere of air(with existent conductivity) aurrounding everything.
I have 3 copper wires that peek out of the sphere so that I can apply voltage terminal/ground terminal to them at both ends of each.
The voltage supply is time dependent in two of those wires.
I have to use either MF or MEF in order to use the BH-curve as I want, and I have to choose MEF in order to be able to pull out the information about the Electric Potential and field. But as it turns out the MEF module does not support the time dependency of the terminals.
So how can I solve this model in version 4.2a?
Thank you very much,
Shoval.
I believe you misunderstood my question.
At the moment, my model looks like this:
I have a core of magnetic material that uses a BH-curve via ampere's law (and current conservation since it's MEF).
I have a sphere of air(with existent conductivity) aurrounding everything.
I have 3 copper wires that peek out of the sphere so that I can apply voltage terminal/ground terminal to them at both ends of each.
The voltage supply is time dependent in two of those wires.
I have to use either MF or MEF in order to use the BH-curve as I want, and I have to choose MEF in order to be able to pull out the information about the Electric Potential and field. But as it turns out the MEF module does not support the time dependency of the terminals.
So how can I solve this model in version 4.2a?
Thank you very much,
Shoval.
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
OK MEF solves for A and V, MF only A but still current conservation is respected in MF (MFNC is without current), it's only if you need Eddy currents, Lorentz terms etc that you really need MEF from my understanding, I mostly get around with MF, its far lighter and solves quicker.
If you are in 3D, do not forget that you need to loop the currents on the external shell to conserve current correctly
If you need the full voltage drops, as predefiend variables yes yu have them in MEF, but as you say MEF is not supporting time domain (At and Att are not in the equations) but you have the steady state harmonic development with omega so you can use a frequency domain solver, and fix the frequency.
if you use MEF, do not forget to have some conductivity in all domains 1-100 S/m for air is mostly OK, and do not orget to set the GND (gauge value for V, else the solver has problems to converge.
--
Good luck
Ivar
OK MEF solves for A and V, MF only A but still current conservation is respected in MF (MFNC is without current), it's only if you need Eddy currents, Lorentz terms etc that you really need MEF from my understanding, I mostly get around with MF, its far lighter and solves quicker.
If you are in 3D, do not forget that you need to loop the currents on the external shell to conserve current correctly
If you need the full voltage drops, as predefiend variables yes yu have them in MEF, but as you say MEF is not supporting time domain (At and Att are not in the equations) but you have the steady state harmonic development with omega so you can use a frequency domain solver, and fix the frequency.
if you use MEF, do not forget to have some conductivity in all domains 1-100 S/m for air is mostly OK, and do not orget to set the GND (gauge value for V, else the solver has problems to converge.
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
I am well aware of all those things, thanks.
I would need your help tho, in using the Frequency Domain in the study.
I've never used it, and I do not know how it changes.
The change in equations between Time Dependent and Frequency Domain is using j and w which I do not know how to control.
namely, E is defined as
E=-grad(V)-jwA
(whereas it should be dA/dt, so I guess it assumes A~exp(jwt)
and J is defined:
J=sigma*E+j*w*epsilon0_const*epsilon_r*E + Jext
where I really have no idea where that bolded part came from.
~~
Suppose I change my model to a Frequency Domain Study. In the FD node, under 'frequencies' I put just 20[Hz]. How do I get that frequency into the terminals? how does that take part in my model via the equations?
Thanks for your help, I appreciate it a lot.
Shoval
I would need your help tho, in using the Frequency Domain in the study.
I've never used it, and I do not know how it changes.
The change in equations between Time Dependent and Frequency Domain is using j and w which I do not know how to control.
namely, E is defined as
E=-grad(V)-jwA
(whereas it should be dA/dt, so I guess it assumes A~exp(jwt)
and J is defined:
J=sigma*E+j*w*epsilon0_const*epsilon_r*E + Jext
where I really have no idea where that bolded part came from.
~~
Suppose I change my model to a Frequency Domain Study. In the FD node, under 'frequencies' I put just 20[Hz]. How do I get that frequency into the terminals? how does that take part in my model via the equations?
Thanks for your help, I appreciate it a lot.
Shoval
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
its "simply" the harmonic development of the derivatives (as for signal processing ... d/dt is replaced by omega* and d^2/dt^2 by omega^2*
Often signal processing use s and s^2 as notation, or in the z z^2 for z developments
for the harmonic or frequency domain, you define the frequency in Hz in the solver section and all loads terms are considered as amplitudes (remove any *sin(...*t) as t is no longer defined). the amplitude may be complex to take into account de phasing expressed as a complex phasor.
Check the doc it's all in there
--
Good luck
Ivar
its "simply" the harmonic development of the derivatives (as for signal processing ... d/dt is replaced by omega* and d^2/dt^2 by omega^2*
Often signal processing use s and s^2 as notation, or in the z z^2 for z developments
for the harmonic or frequency domain, you define the frequency in Hz in the solver section and all loads terms are considered as amplitudes (remove any *sin(...*t) as t is no longer defined). the amplitude may be complex to take into account de phasing expressed as a complex phasor.
Check the doc it's all in there
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi again,
So taht makes good sense I think. But one question: Will it turn ALL into harmonics? I mean, one terminal should be sending in only DC, so how do I make it unaffected by the Frequency Domain?
I will look at the doc anyway tho. thanks!
Shoval
So taht makes good sense I think. But one question: Will it turn ALL into harmonics? I mean, one terminal should be sending in only DC, so how do I make it unaffected by the Frequency Domain?
I will look at the doc anyway tho. thanks!
Shoval
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
Hi everyone!
I get the same exact error mentioned in the topic while conducting time-dependent analysis of poroelastic waves. It is also impossible for me to tick the "solve for this field" box, neither for pressure nor for displacement field. As far as I understand after reading the discussion, I can only conduct frequency-domain analysis for this physics module (because of equivalent reasons as MEF, right?). However, I am indeed allowed to select time-dependent study under custom-studies tree.
The thing is, I need to conduct time-dependent analysis of poroelastic waves in order to incorporate non-linear material properties. Is that possible? And how can I proceed to get rid of the aforementioned error? I am using Version 5.2.
Many thanks in advance,
Dimitris
I get the same exact error mentioned in the topic while conducting time-dependent analysis of poroelastic waves. It is also impossible for me to tick the "solve for this field" box, neither for pressure nor for displacement field. As far as I understand after reading the discussion, I can only conduct frequency-domain analysis for this physics module (because of equivalent reasons as MEF, right?). However, I am indeed allowed to select time-dependent study under custom-studies tree.
The thing is, I need to conduct time-dependent analysis of poroelastic waves in order to incorporate non-linear material properties. Is that possible? And how can I proceed to get rid of the aforementioned error? I am using Version 5.2.
Many thanks in advance,
Dimitris
Please login with a confirmed email address before reporting spam
Posted:
9 years ago
Hi Dimitris,
The Poroelastic Waves interface models propagation of waves in the frequency domain as you can read in the documentation of this interface. This means that only equations for the frequency domain are implemented. In the study tree, you see the supported studies under Preset studies. The studies under Custom studies are not supported.
Sometimes there might be a reason to use a custom study. You can, for example, set up a frequency domain calculation using a stationary study. Then you need to specify the equation form and the frequency to use in the Equation section of the Poroelastic Waves node to be able to use the stationary study.
If you want to get help with your modeling, you can always contact support. And if you would like some new functionality to be added, please contact us and provide a reference to the equations you would like to solve and an example of when this functionality can be used.
Best regards,
Elin Svensson
The Poroelastic Waves interface models propagation of waves in the frequency domain as you can read in the documentation of this interface. This means that only equations for the frequency domain are implemented. In the study tree, you see the supported studies under Preset studies. The studies under Custom studies are not supported.
Sometimes there might be a reason to use a custom study. You can, for example, set up a frequency domain calculation using a stationary study. Then you need to specify the equation form and the frequency to use in the Equation section of the Poroelastic Waves node to be able to use the stationary study.
If you want to get help with your modeling, you can always contact support. And if you would like some new functionality to be added, please contact us and provide a reference to the equations you would like to solve and an example of when this functionality can be used.
Best regards,
Elin Svensson