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 The information provided may be out of date.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.
HB curve solution not physical
Posted Dec 5, 2011, 9:38 a.m. EST Low-Frequency Electromagnetics Version 4.2 5 Replies
Please login with a confirmed email address before reporting spam
Hi again,
In the attached model I have an Iron core which has an HB-curve trait instead of the regular mu_r. There is a coil with external current density around one of its sides, which has DC in it. When the model is solved, the solution is simply not physical - the Magnetic Flux Density in the middle of the coil is weaker than the one in the other side of the core - which is clearly wrong. Any ideas?
be noted - I have a problem saving at the moment, so the model attached needs to be changed before you try to run it - firstly remove the parametric sweep, we need ust one stationary solution; second - the HB curve has units problem - there's a 1e4 factor there, just insert it in the Iron's material properties. Third - the conductivity of the iron is something*e9 there, where it should be e5. not sure if that matters though.
The numbers don't matter too much, but the solution has to be physical first, only then I'll take care of these.
thank you very much for your time!
Shoval.
In the attached model I have an Iron core which has an HB-curve trait instead of the regular mu_r. There is a coil with external current density around one of its sides, which has DC in it. When the model is solved, the solution is simply not physical - the Magnetic Flux Density in the middle of the coil is weaker than the one in the other side of the core - which is clearly wrong. Any ideas?
be noted - I have a problem saving at the moment, so the model attached needs to be changed before you try to run it - firstly remove the parametric sweep, we need ust one stationary solution; second - the HB curve has units problem - there's a 1e4 factor there, just insert it in the Iron's material properties. Third - the conductivity of the iron is something*e9 there, where it should be e5. not sure if that matters though.
The numbers don't matter too much, but the solution has to be physical first, only then I'll take care of these.
thank you very much for your time!
Shoval.
Attachments:
5 Replies Last Post Dec 13, 2011, 8:52 a.m. EST
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
The model attached to this post is the more updated one, the problem is still the same.
Attachments:
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
have you checked that the current is flowing around in the ring ? for me it's looks like you have a coordinate frame issue
something is wrong, plot the external current and try a the coordinate plot on a data set x-z plane cut
I would suggest that update to latest version and then I would suggest that you remake your model from scratch, your basic approach look correct so far.
--
Good luck
Ivar
have you checked that the current is flowing around in the ring ? for me it's looks like you have a coordinate frame issue
something is wrong, plot the external current and try a the coordinate plot on a data set x-z plane cut
I would suggest that update to latest version and then I would suggest that you remake your model from scratch, your basic approach look correct so far.
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi Ivar and thanks for your answer.
I have already rebuilt the model from scratch, and the model attached above is the rebuilt version. It solved on my weak computer after around 25 hours of computing, and the solution is the same. I tried looking at the current density, both at the norm and at each of the components - and in both regular current density and external. It all seems correctly placed around the coil and nowhere else, as it should be.
I have one question to ask, though, that might be the cause for the problem:
In the model, open the Materials node, Iron, and look at the HB-curve settings window(not the interpolation). In the Magnetic field norm line in the output properties, the value is:
HB(10000*normB[1/T])[A/m].
I understand that HB is the name of the interpolation function defined under the node; and the [A/m] is the units in which the solution is to be presented. but what's inside the ( ), I'm not sure how to describe. I have two guesses -
we take the value of normB, unitless, multiply it by 1/T and then put it in the function to recieve the result;
we take the value of normB, with the Tesla units, and multiply it by 1/T to make it unitless so that the function can work(just like we do with sin(t[1/s]) ), apply the function and then multiply by A/m to have results with units.
both seem logical, sort of. Could the problem be this, when it should be normB[T]?
the second guess seemed more logical to me; however, when I tried adding * between the normB and the [1/T], a syntax error message jumped up, so I'm really lost in this. again, the default is
HB(10000*normB[1/T])[A/m].
(the 10000* is my addition of course, to fix the Tesla-Gauss ratio).
Thank you very much,
Shoval
p.s.
by the latest version you mean 4.2a, or is there a new update to 4.2?
I have already rebuilt the model from scratch, and the model attached above is the rebuilt version. It solved on my weak computer after around 25 hours of computing, and the solution is the same. I tried looking at the current density, both at the norm and at each of the components - and in both regular current density and external. It all seems correctly placed around the coil and nowhere else, as it should be.
I have one question to ask, though, that might be the cause for the problem:
In the model, open the Materials node, Iron, and look at the HB-curve settings window(not the interpolation). In the Magnetic field norm line in the output properties, the value is:
HB(10000*normB[1/T])[A/m].
I understand that HB is the name of the interpolation function defined under the node; and the [A/m] is the units in which the solution is to be presented. but what's inside the ( ), I'm not sure how to describe. I have two guesses -
we take the value of normB, unitless, multiply it by 1/T and then put it in the function to recieve the result;
we take the value of normB, with the Tesla units, and multiply it by 1/T to make it unitless so that the function can work(just like we do with sin(t[1/s]) ), apply the function and then multiply by A/m to have results with units.
both seem logical, sort of. Could the problem be this, when it should be normB[T]?
the second guess seemed more logical to me; however, when I tried adding * between the normB and the [1/T], a syntax error message jumped up, so I'm really lost in this. again, the default is
HB(10000*normB[1/T])[A/m].
(the 10000* is my addition of course, to fix the Tesla-Gauss ratio).
Thank you very much,
Shoval
p.s.
by the latest version you mean 4.2a, or is there a new update to 4.2?
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
you should consider the units in square brackets as separate from the numerics, so your second guess B[1/T] means the numerical value of B (of original units in Tesla) expressed unitless from a base of Tesla (=from the [1/T] expression).
Depending on how you have defined your function the 10'000 might be required or not ;) check carefully !
The result of a function is also unitless hence you multiply it by the units desired [A/m] (by the way this is changing these days, I see COMSOL is adding units on some function definition in/out values)
Latest version is, for me, the one on the main COMSOL web site ("support Product updates", as it changes regularly) I have a patch to my 4.2a (mine says COMSOL 4.2.1.134) today.
--
Good luck
Ivar
you should consider the units in square brackets as separate from the numerics, so your second guess B[1/T] means the numerical value of B (of original units in Tesla) expressed unitless from a base of Tesla (=from the [1/T] expression).
Depending on how you have defined your function the 10'000 might be required or not ;) check carefully !
The result of a function is also unitless hence you multiply it by the units desired [A/m] (by the way this is changing these days, I see COMSOL is adding units on some function definition in/out values)
Latest version is, for me, the one on the main COMSOL web site ("support Product updates", as it changes regularly) I have a patch to my 4.2a (mine says COMSOL 4.2.1.134) today.
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
I checked everything, and I still don't understand what is wrong. I checked all the settings; of the physics, the material's HB curve.. and I ran it on a good mesh. I don't have any idea about what I can do next, I'm completely lost now.
The model hasn't changed since the last time I attached it; the only change could be the [1/T] in the HB curve - I don't remember how it is in the attached model. Please help me find what's wrong.
I will also try to contact COMSOL about it.
Thank you very much for your time and help,
Shoval.
The model hasn't changed since the last time I attached it; the only change could be the [1/T] in the HB curve - I don't remember how it is in the attached model. Please help me find what's wrong.
I will also try to contact COMSOL about it.
Thank you very much for your time and help,
Shoval.