Eddy current - Ansys Maxwell

  • Last Post 06 September 2019
  • Topic Is Solved
Aliosha86 posted this 05 August 2019

Hi guys,

I am trying to solve a problem in Ansys Maxwell.

Basically, I want to calculate the eddy current induced in a thin vessel due to an inductor. The current in my inductor varies in time with a specific time-dependent law. It grows up to 200 kA linearly in 20 ms, then it has the flattop for 40 ms and finally, it goes to zero in 20 ms.

I would like to know whether Ansys Maxwell can be useful for these type of problems or it is better to change approach. I was wondering if the time scale is too rapid that the software cannot follow the right dynamics.

I hope in an answer.


Order By: Standard | Newest | Votes
pblarsen posted this 16 August 2019


Yes, you could use Maxwell Transient if the waveform is just the trapezoidal shape that you described.  The easiest way would be to use a dataset, and to define the excitation expression as "pwl(ds1,time)", where pwl will use the dataset lookup, and time is the simulation time.  20ms is an acceptable time-range for Maxwell, and you would want to use ~2-4ms timesteps so that there are several time-steps up the ramp-up and down the ramp-down.

If the trapezoidal shape is the varying amplitude of a faster sine-wave, then you might consider choosing one amplitude and using the eddy current solver.

Remember to use sufficient skin-depth meshing for the induced currents.

 - Paul

  • Liked by
  • Aliosha86
balajisriram15 posted this 20 August 2019

Hi Allessio,

Magnetostatics does'nt take care of skin depth. It assumes the whole cross section has the same current density. Eddy current meshing on the other hand takes into account the skin depth based on your frequency for meshing.  

  • Liked by
  • Aliosha86
rmagargl posted this 21 August 2019

You could do an FFT on the waveform to see where the current is at the highest harmonics.  In these switching type of waveforms, the highest harmonics come from the edge rate of the signal, not the period, so if you just want a rough number, take the twice the inverse of the rise time as a ballpark characteristic frequency.  You could then use a skin depth mesh operation or a link to an AC eddy current simulation at that frequency.

  • Liked by
  • Aliosha86
Aliosha86 posted this 16 August 2019

Hi Paul, Thank you very much for the advice. Regarding the mesh I linked the mesh I used for the Magnetostatic simulation in order to have an Adaptative solver. Do you think it is good?

Aliosha86 posted this 20 August 2019


I attach you a picture of my current profile. As you can see the problem is that I do not have a frequence for my problem. The profile is just trapezoidal and calculating the skin depth is almost impossible.

I hope this can help you giving me the best advice.

Aliosha86 posted this 20 August 2019

Hi Paul

I attach you below the profile of the current. As you can see I do not have a frequency to evaluate skin depth for eddy currents.


Aliosha86 posted this 06 September 2019

Thank you, but unfortunately I am not able to use the skin depth mesh operation because it does not work for my geometry and I don't know why.

I attach you the section of the geometry. I am selecting the surface and use the command but it does not work.