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 The information provided may be out of date.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.

ACDC mangetic and electric fields app not solving

Please login with a confirmed email address before reporting spam

Hello,
I am using comsol 4.1 to study the magnetic response of a structure.
My 3D model consists of a set of square cross-sectioned wires starting at one of the big environment box domain and ending at the opposite one (so the run through the air box). In the middle there's some magnetic material I would like to study. I set all the conditions as in the "?Integrated Square-Shaped Spiral Inductor" model in the model gallery. I use free triangular mesh and Comsol meshes it with no pbs. When I solve it it gives me this error:

Attempt to evaluate negative power of zero.
- Function: ^
Failed to evaluate temporary symbolic derivative variable.
- Variable: mod2.mef.murzz@VDN$mod2.curlA2z
- Defined as: (dmod2.mat6.def.MUR_dx1((sqrt((((mod2.mef.Bx^2)+(mod2.mef.By^2))+(mod2.mef.Bz^2)))/unit_T_cf))*(((mod2.mef.Bz*2)*(0.5*((((mod2.mef.Bx^2)+(mod2.mef.By^2))+(mod2.mef.Bz^2))^(-0.5))))*(1/unit_T_cf)))
Failed to evaluate temporary symbolic derivative variable.
- Variable: mod2.mef.murinvzz@VDN$mod2.curlA2z
- Defined as: (((dmod2.mat6.def.MUR_dx1((sqrt((((mod2.mef.Bx^2)+(mod2.mef.By^2))+(mod2.mef.Bz^2)))/unit_T_cf))*(((mod2.mef.Bz*2)*(0.5*((((mod2.mef.Bx^2)+(mod2.mef.By^2))+(mod2.mef.Bz^2))^(-0.5))))*(1/unit_T_cf))))*((-1)/(mod2.mef.murzz^2)))
Failed to evaluate temporary symbolic derivative variable.
- Variable: mod2.mef.Hz@VDN$mod2.curlA2z
- Defined as: (((((((dmod2.mat6.def.MUR_dx1((sqrt((((mod2.mef.Bx^2)+(mod2.mef.By^2))+(mod2.mef.Bz^2)))/unit_T_cf))*(((mod2.mef.Bz*2)*(0.5*((((mod2.mef.Bx^2)+(mod2.mef.By^2))+(mod2.mef.Bz^2))^(-0.5))))*(1/unit_T_cf))))*((-1)/(mod2.mef.murzz^2))))*mod2.mef.Bz)+mod2.mef.murinvzz)*(1/mu0_const))
Failed to evaluate expression.
- Expression: d(d((-mod2.mef.Hx*test(mod2.curlA2x)-mod2.mef.Hy*test(mod2.curlA2y)-mod2.mef.Hz*test(mod2.curlA2z)+mod2.mef.Jx*test(mod2.A2x)+mod2.mef.Jy*test(mod2.A2y)+mod2.mef.Jz*test(mod2.A2z))*(dvol),{test@4}),mod2.curlA2z)


I tried resizing the environment domain and the mesh, I set the conductivity of air to 1 as indicated in the "?Integrated Square-Shaped Spiral Inductor" documentation, I tried switching to the other module called magnetic fields (mf), and same stuff.
When I remove the magnetic structure it solves with no problems. The magnetic structure is a collection of rectangular prisms with the material set to soft iron (no losses), without changing anything.

Any suggestions to make it solve INCLUDING the magnetic structure?

Thanks so much!
Giuseppe Schiavone

5 Replies Last Post Jul 12, 2011, 1:59 a.m. EDT
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jan 20, 2011, 9:53 a.m. EST
Hi

if you are in MF I do not belive the conductivity of air (if =0) matters specially, but in MEF yes it does (without I do not believe you manage to solve), so using a conductivity of air =1 in most cases is not a bad idea

Then your erro list looks like a material or BC error and is probably not linked to the mesh. HAve you checked that your material definition does not contain any red warning "STOPS"

Athough with "square" items in MF or MEF you make "singularities" at the right angle edge/vertexes, you should ideally fillet with a small radius your coil and iron corners, but it can become rather tedious to define, for many elements

--
Good luck
Ivar
Hi if you are in MF I do not belive the conductivity of air (if =0) matters specially, but in MEF yes it does (without I do not believe you manage to solve), so using a conductivity of air =1 in most cases is not a bad idea Then your erro list looks like a material or BC error and is probably not linked to the mesh. HAve you checked that your material definition does not contain any red warning "STOPS" Athough with "square" items in MF or MEF you make "singularities" at the right angle edge/vertexes, you should ideally fillet with a small radius your coil and iron corners, but it can become rather tedious to define, for many elements -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jan 20, 2011, 10:07 a.m. EST
Hello Ivar,

thanks for getting back.

I checked the material properties and everything looks ok, there's green ticks on the sides of some of the values.
I tried filletting on the 2D workplane (before extruding), and it doesn't change the result. I reckon you still get sharp edges when you extrude the filleted 2D shapes, so shall I try 3D filleting? That would mean drawing spheres and perform operations etc. right? Anyway again, it works all right without the magnetic structure in the middle, with sharp edged wires, so is this really the problem here?

Thanks again,

Giuseppe
Hello Ivar, thanks for getting back. I checked the material properties and everything looks ok, there's green ticks on the sides of some of the values. I tried filletting on the 2D workplane (before extruding), and it doesn't change the result. I reckon you still get sharp edges when you extrude the filleted 2D shapes, so shall I try 3D filleting? That would mean drawing spheres and perform operations etc. right? Anyway again, it works all right without the magnetic structure in the middle, with sharp edged wires, so is this really the problem here? Thanks again, Giuseppe

Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jan 20, 2011, 10:30 a.m. EST
Hi

as I said rounding off corner is tedious, and It's probably not your main issue here, so long you are aware that the field values in sharp corners are wrong ;)

Then I do not really have any clue.
Apart that a negative power of "0" is a "1/0" and the most common "0" when starting a solver is the default initial values that are set to "0".

Then it's the derivative section (for the Newton gradient estimation) that has a problem. Perhaps check your BC, do you have a temperature T=0[K] instead of T=0[degC] somewhere ? Just a blind guess

--
Good luck
Ivar
Hi as I said rounding off corner is tedious, and It's probably not your main issue here, so long you are aware that the field values in sharp corners are wrong ;) Then I do not really have any clue. Apart that a negative power of "0" is a "1/0" and the most common "0" when starting a solver is the default initial values that are set to "0". Then it's the derivative section (for the Newton gradient estimation) that has a problem. Perhaps check your BC, do you have a temperature T=0[K] instead of T=0[degC] somewhere ? Just a blind guess -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jan 20, 2011, 11:12 a.m. EST
Hello!

I changed the material to Iron from the general library instead of the soft iron from the ACDC module and it works now...

Thanks for your help!

Best,

Giuseppe
Hello! I changed the material to Iron from the general library instead of the soft iron from the ACDC module and it works now... Thanks for your help! Best, Giuseppe

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago Jul 12, 2011, 1:59 a.m. EDT
Hi..

I too faced a similar problem. I modeled a simple coil in 2D axis-symmetry approach.
It works fine when the coil is air core or having a core material with linear BH curve.
But its showing an error similar to mentioned above, when a material with non-linear BH curve (eg- soft iron) is used as core.

Please let me know a solution for this problem.
Hi.. I too faced a similar problem. I modeled a simple coil in 2D axis-symmetry approach. It works fine when the coil is air core or having a core material with linear BH curve. But its showing an error similar to mentioned above, when a material with non-linear BH curve (eg- soft iron) is used as core. Please let me know a solution for this problem.

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.