-
Filter by Topic
Release
Popular Topics
- List all discussions
Frequency sweep in a magnetic field interface
|
Thread index | Previous thread | Next thread | Start a new discussion |
January 17, 2013 9:34pm UTC
Frequency sweep in a magnetic field interface
Hello everyone,
This should be seemingly siomple problem. I am trying to model the AC magnetic field due to a coil in 3D. I am using a magnetic field interface on a simple cylinder (acting as the coil) kept in air. A current has been given to this cylinder axially using "external current density feature". Now if I run a stationary solver, it is able to solve for a magnetic field without hassle. Even if I give a periodic current to the coil and solve for time-dependent solution at a particular frequency, I am getting an answer. But when I want to do a frequency sweep analysis I am getting the following error:
"Failed to find a solution for the initial parameter.
Matrix has zero on diagonal.
Returned solution is not converged."
Now I know that usually this kind of error is due to inadequate boundary condition. But that does not explain to me how the simulation is successful with a stationary and time domain analysis.
Also if I try to make the coil using a 2D axisymmetric simulation, I am able to do even the frequency sweep giving the exact same boundary conditions that I am using for 3D simulation. I am at loss to see what could be the problem. Any suggestions or advice will be highly appreciated.
I have attached the model file in case anyone wants to check it out.
Venkat
Reply | Reply with Quote | Send private message | Report Abuse
January 17, 2013 10:25pm UTC in response to Venkatram Pepakayala
Re: Frequency sweep in a magnetic field interface
Venkat,
Set air conductivity to small value compared to the conductivity of copper (say, sig_air=10[S/m]) and re-run analysis. This helps solver to be more stable.
Regards,
Sergei
Reply | Reply with Quote | Send private message | Report Abuse
January 18, 2013 3:38am UTC in response to Sergei Yushanov
Re: Frequency sweep in a magnetic field interface
Thanks Sergei! Works like a charm. Makes me wonder why this was not the case in COMSOL 3.5. The conductivity values even if 0 seem to pose no problem in 3.5.
Reply | Reply with Quote | Send private message | Report Abuse
January 18, 2013 1:46pm UTC in response to Venkatram Pepakayala
Re: Frequency sweep in a magnetic field interface
Venkat,
Good question – I am also wondering about the difference between 3.5 and 4.3 in this regard. May be you can submit this question to Comsol Support?
Regards,
Sergei
Reply | Reply with Quote | Send private message | Report Abuse
Rules and guidelines

