Transient Induction Heating Simulation with BH Curve Material

Please login with a confirmed email address before reporting spam

Dear respected experts,

I am encountering difficulties in simulating an induction heating process and would appreciate your guidance. The configuration consists of a U-shaped single-turn coil carrying AC current to heat a soft magnetic material plate via induction.

The key technical issue is: When defining the soft magnetic material's constitutive relation using a BH curve, frequency-domain analysis becomes unavailable. However, when switching to transient calculation, I cannot apply current excitation to the single-turn coil. Attempting voltage excitation results in severe convergence issues.

Core requirements:

Maintain BH curve definition (necessary for accurate hysteresis loss calculation)

Perform transient analysis with AC current in U-shaped coil

Constraints observed:

Relative permeability models cannot replicate measured hysteresis losses

Experimental data confirms hysteresis loss dominates over eddy current loss in this heating process

Has anyone successfully implemented such a combination (BH curve material + transient AC current excitation for single-turn coils)? Any suggestions on solver settings/excitation implementation would be highly valuable.



1 Reply Last Post Mar 28, 2025, 9:49 a.m. EDT
Magnus Olsson COMSOL Employee

Please login with a confirmed email address before reporting spam

Posted: 3 days ago Mar 28, 2025, 9:49 a.m. EDT

Hi,

Note that the built-in material models for B-H curve (stationary and transient) and Effective B-H curve (frequency domain) only account for the magnetic saturation effect. To account for hysteresis, you need to use the Hysteresis Jiles-Atherton model in the time domain or use a measured lookup table for the losses as a function of magnetic flux density magnitude and temperature in the frequency domain. In general, I would expect it to be challenging to measure/extract Jiles-Atherton parameters as a function of temperature, especially in the higher temperature range. The current driven solid conductor coil typically results in numerical difficulties in the time domain, especially if you try to ramp up the current too fast from zero - or (unphysical) try to start from a nonzero coil current without first doing a stationary initialization. That is, you cannot start from initial conditions representing zero magnetic energy and go directly to a state with finite magnetic energy in zero time as that requires infinite input power (in the lab as well as in numerical modeling).

Given also the vastly different thermal and electric time scales, I would go for a frequency-transient approach like in the attached example and introduce the magnetic losses by using the Magnetic losses constitutive relation with mur'(|B|,T) and mur''(|B|,T) implemented as lookup tables/interpolation functions. The attached example uses functions to define the apparent mur (B/H) as a function of B and T and neglects the hysteresis losses.

-------------------
Magnus
Hi, Note that the built-in material models for B-H curve (stationary and transient) and Effective B-H curve (frequency domain) only account for the magnetic saturation effect. To account for hysteresis, you need to use the Hysteresis Jiles-Atherton model in the time domain or use a measured lookup table for the losses as a function of magnetic flux density magnitude and temperature in the frequency domain. In general, I would expect it to be challenging to measure/extract Jiles-Atherton parameters as a function of temperature, especially in the higher temperature range. The current driven solid conductor coil typically results in numerical difficulties in the time domain, especially if you try to ramp up the current too fast from zero - or (unphysical) try to start from a nonzero coil current without first doing a stationary initialization. That is, you cannot start from initial conditions representing zero magnetic energy and go directly to a state with finite magnetic energy in zero time as that requires infinite input power (in the lab as well as in numerical modeling). Given also the vastly different thermal and electric time scales, I would go for a frequency-transient approach like in the attached example and introduce the magnetic losses by using the Magnetic losses constitutive relation with mur'(|B|,T) and mur''(|B|,T) implemented as lookup tables/interpolation functions. The attached example uses functions to define the apparent mur (B/H) as a function of B and T and neglects the hysteresis losses.

Reply

Please read the discussion forum rules before posting.

Please log in to post a reply.

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.