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 The information provided may be out of date.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.
mur - complex permeability - magnetic losses
Posted Mar 11, 2011, 7:07 p.m. EST Low-Frequency Electromagnetics Version 4.0a 6 Replies
Please login with a confirmed email address before reporting spam
I see in Comsol 4.0a there is now the ability to define the magnetic constitutive relation as "Magnetic Losses" which takes as input u' and u" from the complex mur = u' - iu"
I have the data for mur from the ferrite manufacturer, it is frequency dependent. I have defined an Interpolation Function from a file, after being imported becomes from a table. The table is labelled columns of t and f(t). I named it threeC90mu1 (tried to name it 3C90mu1 but that didn't work).
How do I call the function? Is it by name? With round brackets taking the parameter? The closest thing I can find in the documentation is under COMSOL Multiphysics > Global and Local Definitions > Global and Local Functions > Defining an Interpolation Function. At the very bottom it says after naming the function tempfun
"You can now use the function tempfun with x and y as input arguments in a 2D model to get the interpolated value for the temperature at any position." So this means call it using tempfun(x,y) ??
My function depends on frequency. Is the variable for frequency "freq" ? So I would call as threeC90mu1(freq) ?
Does anyone know how to do this in Comsol 3.5a where the only options for constitutive relation are B=uH, specify Magnetization, or Remnant Flux ?
I have the data for mur from the ferrite manufacturer, it is frequency dependent. I have defined an Interpolation Function from a file, after being imported becomes from a table. The table is labelled columns of t and f(t). I named it threeC90mu1 (tried to name it 3C90mu1 but that didn't work).
How do I call the function? Is it by name? With round brackets taking the parameter? The closest thing I can find in the documentation is under COMSOL Multiphysics > Global and Local Definitions > Global and Local Functions > Defining an Interpolation Function. At the very bottom it says after naming the function tempfun
"You can now use the function tempfun with x and y as input arguments in a 2D model to get the interpolated value for the temperature at any position." So this means call it using tempfun(x,y) ??
My function depends on frequency. Is the variable for frequency "freq" ? So I would call as threeC90mu1(freq) ?
Does anyone know how to do this in Comsol 3.5a where the only options for constitutive relation are B=uH, specify Magnetization, or Remnant Flux ?
6 Replies Last Post Mar 18, 2011, 2:20 a.m. EDT
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
My function depends on frequency. Is the variable for frequency "freq" ? So I would call as threeC90mu1(freq) ?
I am assuming that is the name because that's what is says in the Help under
COMSOL Multiphysics > Global and Local Definitions > Variables > Summary of Common Predefined Variables
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
thendefault name for the frequency for frequency sweeps is "freq" so if you have data for a varianble you define an interpolation function/list depending on a variable and then you call this in your material definition
Thee are numerous materials defined like that in the material data base (DB)
--
Good luck
Ivar
thendefault name for the frequency for frequency sweeps is "freq" so if you have data for a varianble you define an interpolation function/list depending on a variable and then you call this in your material definition
Thee are numerous materials defined like that in the material data base (DB)
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
thendefault name for the frequency for frequency sweeps is "freq" so if you have data for a varianble you define an interpolation function/list depending on a variable and then you call this in your material definition
Thee are numerous materials defined like that in the material data base (DB)
Hi Ivar, thanks for your reply.
So functions are called using functionName(freq)
Do you have an example of such a material defined in this way?
Perhaps this is a new question topic, but in Comsol 4.0+ can you explain the different between setting frequencies at "> Step 1: Frequency Domain 1" OR "Solver Configurations > Solver 1 > Stationary Solver 1 > Parametric 1" as shown in this image img607.imageshack.us/img607/1904/freqsettingquestion.jpg
I have noted that even if frequencies are specified in the first location, the values used are actually from the "Parametric" settings.
Thank you!
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
you can upload your images to the forum, then its easier to follow in 3-4 months for other too.
Indeed a frequency sweep is "only" a parametric sweep of freq. BUT it also selects a harmonic development of your physical equations so there are more settings behind a stationary, transient or a frequency sweep selection, even if all depend on a "parametric" node in the solver sequence
--
Good luck
Ivar
you can upload your images to the forum, then its easier to follow in 3-4 months for other too.
Indeed a frequency sweep is "only" a parametric sweep of freq. BUT it also selects a harmonic development of your physical equations so there are more settings behind a stationary, transient or a frequency sweep selection, even if all depend on a "parametric" node in the solver sequence
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
1)
So what is the difference between specifying frequencies in the Frequency Domain node or the Parametric node? Empirically, I've noted that Comsol will only use the values in Parametric node, causing my question of their relationship.
2)
Do you know any particular material that is defined using interpolation function as you mentioned?
3)
Any idea why the solution would only take 200s with mur defined as a constant, but solution does not converge after 12+ hours when mur is defined as an interpolation function? The function is mur(freq) and only one freq parameter was entered, so how is that any different from a constant ?
Thanks
So what is the difference between specifying frequencies in the Frequency Domain node or the Parametric node? Empirically, I've noted that Comsol will only use the values in Parametric node, causing my question of their relationship.
2)
Do you know any particular material that is defined using interpolation function as you mentioned?
3)
Any idea why the solution would only take 200s with mur defined as a constant, but solution does not converge after 12+ hours when mur is defined as an interpolation function? The function is mur(freq) and only one freq parameter was entered, so how is that any different from a constant ?
Thanks
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
PLS note that when you select the different solvers, COMSOL arranges its internal equations (see the equation tab of your physics), so if you want a frequency scan, you need the harmonic development of the PDE, this is only activated by selecting the frequency domain solvers (except if you have taken fully control of the PDE writing ;).
So use the frequency parameter field and not an external parametric sweep for that. Exception, if you have a very special custom solver sequence, BUT then you need to copy your external "freq" defined variable into the frequency domain solver "freq" field to get it to work i.e. single "freq" analysis. Hope I'm clear
run through the material DB you can look at them, probably you can read the material definition file too (could in older version, was slightly cryptic, but understandable)
not converging, is often due to a starting point (initial conditions) that are too ar away from the solution, and you enter a smooth "side-blind" wally of the optimisation process
--
Good luck
Ivar
PLS note that when you select the different solvers, COMSOL arranges its internal equations (see the equation tab of your physics), so if you want a frequency scan, you need the harmonic development of the PDE, this is only activated by selecting the frequency domain solvers (except if you have taken fully control of the PDE writing ;).
So use the frequency parameter field and not an external parametric sweep for that. Exception, if you have a very special custom solver sequence, BUT then you need to copy your external "freq" defined variable into the frequency domain solver "freq" field to get it to work i.e. single "freq" analysis. Hope I'm clear
run through the material DB you can look at them, probably you can read the material definition file too (could in older version, was slightly cryptic, but understandable)
not converging, is often due to a starting point (initial conditions) that are too ar away from the solution, and you enter a smooth "side-blind" wally of the optimisation process
--
Good luck
Ivar