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.
Using unlinear AC/DC-material
Posted Mar 1, 2011, 5:30 a.m. EST Low-Frequency Electromagnetics, Materials 7 Replies
Please login with a confirmed email address before reporting spam
Hi all, Im trying to use the unlinear material "Soft Iron" from the AC/DC database. No matter what I do I get the error message;
Attempt to evaluate negative power of zero.
- Function: ^
Failed to evaluate temporary symbolic derivative variable.
- Variable: mod1.mef.murzz@VDN$mod1.curlAz
- Defined as: (dmod1.mat4.def.MUR_dx1((sqrt(((mod1.mef.Bz^2)+((mod1.mef.By^2)+(mod1.mef.Bx^2))))/unit_T_cf))*(((2*mod1.mef.Bz)*(0.5*(((mod1.mef.Bz^2)+((mod1.mef.By^2)+(mod1.mef.Bx^2)))^(-0.5))))*(1/unit_T_cf)))
Failed to evaluate temporary symbolic derivative variable.
- Variable: mod1.mef.murinvzz@VDN$mod1.curlAz
- Defined as: (((-1)/(mod1.mef.murzz^2))*((dmod1.mat4.def.MUR_dx1((sqrt(((mod1.mef.Bz^2)+((mod1.mef.By^2)+(mod1.mef.Bx^2))))/unit_T_cf))*(((2*mod1.mef.Bz)*(0.5*(((mod1.mef.Bz^2)+((mod1.mef.By^2)+(mod1.mef.Bx^2)))^(-0.5))))*(1/unit_T_cf)))))
Failed to evaluate temporary symbolic derivative variable.
- Variable: mod1.mef.Hz@VDN$mod1.curlAz
- Defined as: ((mod1.mef.murinvzz+(mod1.mef.Bz*((((-1)/(mod1.mef.murzz^2))*((dmod1.mat4.def.MUR_dx1((sqrt(((mod1.mef.Bz^2)+((mod1.mef.By^2)+(mod1.mef.Bx^2))))/unit_T_cf))*(((2*mod1.mef.Bz)*(0.5*(((mod1.mef.Bz^2)+((mod1.mef.By^2)+(mod1.mef.Bx^2)))^(-0.5))))*(1/unit_T_cf))))))))*(1/mu0_const))
Failed to evaluate expression.
- Expression: d(d((-mod1.mef.Hx*test(mod1.curlAx)-mod1.mef.Hy*test(mod1.curlAy)-mod1.mef.Hz*test(mod1.curlAz)+mod1.mef.Jx*test(mod1.Ax)+mod1.mef.Jy*test(mod1.Ay)+mod1.mef.Jz*test(mod1.Az))*(dvol),{test@3}),mod1.curlAz)
I have tried to refine the mesh and lowering the tolerance but nothing seems to work. The model Im experimenting with is the "power_inductor" from the model library. I just open the model and change the core material to "Soft Iron(without losses)". What do I do wrong?
Best regards Niklas K
Attempt to evaluate negative power of zero.
- Function: ^
Failed to evaluate temporary symbolic derivative variable.
- Variable: mod1.mef.murzz@VDN$mod1.curlAz
- Defined as: (dmod1.mat4.def.MUR_dx1((sqrt(((mod1.mef.Bz^2)+((mod1.mef.By^2)+(mod1.mef.Bx^2))))/unit_T_cf))*(((2*mod1.mef.Bz)*(0.5*(((mod1.mef.Bz^2)+((mod1.mef.By^2)+(mod1.mef.Bx^2)))^(-0.5))))*(1/unit_T_cf)))
Failed to evaluate temporary symbolic derivative variable.
- Variable: mod1.mef.murinvzz@VDN$mod1.curlAz
- Defined as: (((-1)/(mod1.mef.murzz^2))*((dmod1.mat4.def.MUR_dx1((sqrt(((mod1.mef.Bz^2)+((mod1.mef.By^2)+(mod1.mef.Bx^2))))/unit_T_cf))*(((2*mod1.mef.Bz)*(0.5*(((mod1.mef.Bz^2)+((mod1.mef.By^2)+(mod1.mef.Bx^2)))^(-0.5))))*(1/unit_T_cf)))))
Failed to evaluate temporary symbolic derivative variable.
- Variable: mod1.mef.Hz@VDN$mod1.curlAz
- Defined as: ((mod1.mef.murinvzz+(mod1.mef.Bz*((((-1)/(mod1.mef.murzz^2))*((dmod1.mat4.def.MUR_dx1((sqrt(((mod1.mef.Bz^2)+((mod1.mef.By^2)+(mod1.mef.Bx^2))))/unit_T_cf))*(((2*mod1.mef.Bz)*(0.5*(((mod1.mef.Bz^2)+((mod1.mef.By^2)+(mod1.mef.Bx^2)))^(-0.5))))*(1/unit_T_cf))))))))*(1/mu0_const))
Failed to evaluate expression.
- Expression: d(d((-mod1.mef.Hx*test(mod1.curlAx)-mod1.mef.Hy*test(mod1.curlAy)-mod1.mef.Hz*test(mod1.curlAz)+mod1.mef.Jx*test(mod1.Ax)+mod1.mef.Jy*test(mod1.Ay)+mod1.mef.Jz*test(mod1.Az))*(dvol),{test@3}),mod1.curlAz)
I have tried to refine the mesh and lowering the tolerance but nothing seems to work. The model Im experimenting with is the "power_inductor" from the model library. I just open the model and change the core material to "Soft Iron(without losses)". What do I do wrong?
Best regards Niklas K
7 Replies Last Post Mar 2, 2011, 5:17 a.m. EST
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
are you sure you should not use MF instead of MEF if you do only magnetics then MF should hold, you do nt need to solve for V ?
2) in MEF you need to set some conductivity for the air too (by default the material properties propose sigma=0 S/m
3) have you defined a new Ampere Law 2 for the iron ? and set the Magnetic field to |H|=f(|B|) ?
Often its better to enter the material data by hand one get a better feeling of what is happening and the ways COMSOL changes the demands depending on the way we define the physics
--
Good luck
Ivar
are you sure you should not use MF instead of MEF if you do only magnetics then MF should hold, you do nt need to solve for V ?
2) in MEF you need to set some conductivity for the air too (by default the material properties propose sigma=0 S/m
3) have you defined a new Ampere Law 2 for the iron ? and set the Magnetic field to |H|=f(|B|) ?
Often its better to enter the material data by hand one get a better feeling of what is happening and the ways COMSOL changes the demands depending on the way we define the physics
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi!
You should add a second "Ampere's Law", select your iron part and change "Magnetic Field" from "relative permeability" to "HB curve".
Ralf
You should add a second "Ampere's Law", select your iron part and change "Magnetic Field" from "relative permeability" to "HB curve".
Ralf
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi, thanks for a quick reply
1. I specify a current through the coil, dont I need to use MEF then?
2. I read about that trick, otherwise it woudent solve for linear materials either.
3. The third point did the trick, wasent aware that you had to change that. I do run into nother error though;
For mesh-case 0 there are 10161 void equations (empty rows in matrix) for the variable mod1.V
at coordinates: (-0.0707496,-0.0482953,-0.0100317), (-0.0726887,-0.0478022,-0.0104545), (-0.0721955,-0.0482953,-0.00931775), (-0.0545397,-0.0485191,-0.00719249), (-0.0553267,-0.0470382,-0.00678248), ...
and similarly for the degrees of freedom (empty columns in matrix)
Returned solution has not converged.
Any ideas?
Best regards Niklas K
1. I specify a current through the coil, dont I need to use MEF then?
2. I read about that trick, otherwise it woudent solve for linear materials either.
3. The third point did the trick, wasent aware that you had to change that. I do run into nother error though;
For mesh-case 0 there are 10161 void equations (empty rows in matrix) for the variable mod1.V
at coordinates: (-0.0707496,-0.0482953,-0.0100317), (-0.0726887,-0.0478022,-0.0104545), (-0.0721955,-0.0482953,-0.00931775), (-0.0545397,-0.0485191,-0.00719249), (-0.0553267,-0.0470382,-0.00678248), ...
and similarly for the degrees of freedom (empty columns in matrix)
Returned solution has not converged.
Any ideas?
Best regards Niklas K
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi, thanks for your help, did the trick but only to take me to the next error:-), see above
Best regards Niklas
Best regards Niklas
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
Yes MF can solve for currents, you do not get the voltage field as ES thats all, and in MF you get the lumped parameter of the voltage drop in your coils with the port/terminals BC's
and MF is far simpler and quicker to solve. Always check the dependent variables and as yourself where you need them, simpler the model the quicker you get to the goal ;)
--
Good luck
Ivar
Yes MF can solve for currents, you do not get the voltage field as ES thats all, and in MF you get the lumped parameter of the voltage drop in your coils with the port/terminals BC's
and MF is far simpler and quicker to solve. Always check the dependent variables and as yourself where you need them, simpler the model the quicker you get to the goal ;)
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi, I had a look at the MF but dont see how I should add the current. As I wrote previously, my model looks like the "power inductor" from the model library. I see how the current could be added as "external current density" if a idealized coil is used as in "helmholtx coil" from the same place. If I use MF, where do I find the terminal and ground?
Best regards Niklas K
Best regards Niklas K
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
its all dependent if you are in 2D or 3D and for which vector component(s) you solve
In MF you enter the current via current densities, either surface values or domains
in 2D (in plane or out of plane respectively) or in 3D
A frequency response solver will then modulate the current density
It's correct that port terminal are only in MEF as you need to solve for V, but with the current and the section + conductivity and possibly temperature you should be able to extract voltage drop too, no ?
--
Good luck
Ivar
its all dependent if you are in 2D or 3D and for which vector component(s) you solve
In MF you enter the current via current densities, either surface values or domains
in 2D (in plane or out of plane respectively) or in 3D
A frequency response solver will then modulate the current density
It's correct that port terminal are only in MEF as you need to solve for V, but with the current and the section + conductivity and possibly temperature you should be able to extract voltage drop too, no ?
--
Good luck
Ivar