Induction heating modeling in Ansys

[Zeekar]

Hi

I'm trying to preform a simple model of a ferromagnetic bar which has a coil wrapped around it. The bar has a 4×4 cm base and is 15 cm tall, the coil has an inside radius of 29 mm and outside radius 30 mm and a height of 10 cm. The goal is to find the eddy currents in the bar, generated joule heating, temperature distribution (convection is taken in account at the edge of the bar) in the bar and in the end stresses in the bar because of the joule heating.
The model is very similar to the example done in the low-frequency electromagnetic analysis guide and if I follow it I do get results which are logical.
Now my problem lies in this:

- the example in the guide takes in accord symmetries and while my own model is also symmetric I would like to preform it without taking in account symmetries. How do I apply current excitation in the coil? If you take in account symmetries you have 2 symmetric planes at which you can set the boundary conditions appropriate for a current fed solid conductor.
I modeled the coil as a closed hollow cylinder and according to the ansys you are supposed to apply current amount at one side with the coupled volt DOF over all nodes at that area and at the other area you set VOLT DOF to zero. Since I don't have 2 areas in a closed cylinder that is a bit difficult.

- the flux-parallel boundary condition is also causing me problems since I haven't done electromagnetism in ages and its causing me problems. In the guide the boundary conditions are set on the exterior surface of the air enclosure, the az degree of freedom is set to zero. I'm assuming that is to hold the magnetic field flux inside the air enclosure (to be as a limit to the magnetic field). Am I right or wrong about this?

If you need more details I can attach a picture of the model, or if something is unclear please ask me since my wording may be a bit off (English isn't my primary language sadly so I apologize for any ill formulated sentences ).

 

[Enthalpy]

Induction heating does NOT rely on eddy current losses but on hysteresis losses in the ferromagnetic material. Hence copper or aluminium pans don't work.

 

[Mech_Engineer]

Somehow I missed this post, it's an interesting problem!

Induction heating does NOT rely on eddy current losses but on hysteresis losses in the ferromagnetic material. Hence copper or aluminium pans don't work.

Inductive heating will work on any conductive material, but it's more efficient on some materials than others. It depends mainly on the current and frequency the inductive heater is operating at.

- the example in the guide takes in accord symmetries and while my own model is also symmetric I would like to preform it without taking in account symmetries. How do I apply current excitation in the coil? If you take in account symmetries you have 2 symmetric planes at which you can set the boundary conditions appropriate for a current fed solid conductor.
I modeled the coil as a closed hollow cylinder and according to the ansys you are supposed to apply current amount at one side with the coupled volt DOF over all nodes at that area and at the other area you set VOLT DOF to zero. Since I don't have 2 areas in a closed cylinder that is a bit difficult.

I'm having trouble imagining your geometry, but you should be able to apply an excitation to an area, which if you're doing a 3-D model will suffice.

Keep in mind however that depending on the frequency of your inductive heater, skin depth will play a dominant role in your model (http://en.wikipedia.org/wiki/Skin_effect). For accuracy your model will need about 2-3 elements across the surface of the part to properly rsolve the skin depth's current distribution, so if your heater is running at 100khz or more you're looking at a very small skin depth (<1mm) which may be your limiting factor for proper modeling.

This basically shows what the fundamental difference is between ANSYS "low-frequency emag" and "high frequency emag"; HF emag works in the MHz range where the skin depth cannot be resolved through mesh density. Inductive heating tends to play in the awkward inner space between LF and HF unfortunately...

- the flux-parallel boundary condition is also causing me problems since I haven't done electromagnetism in ages and its causing me problems. In the guide the boundary conditions are set on the exterior surface of the air enclosure, the az degree of freedom is set to zero. I'm assuming that is to hold the magnetic field flux inside the air enclosure (to be as a limit to the magnetic field). Am I right or wrong about this?

For an electromagnetic problem, you need to define a flux-parallel condition as a way of "grounding" the model. This condition should be set in an airspace with enough distance from the model to approximate a far-field condition.

 

[Enthalpy]

My mistake. I had only an induction cooker in mind. Ovens for steel do rely on ohmic losses.