The Surface Impedance approach enables simulation of high frequency induction heating in few minutes, without building a huge mesh. Induced heat in a steel strip is accurately calculated in 10 minutes for 300 kHz, while traditional FEM simulation takes about 10 hours.
About the Surface Impedance approach
For induction heating with high frequencies, when EM field has very shallow penetrates into a workpiece, the skin depth may be measured in microns. Traditional FEM calculations, which require resolution of the skin layer by a mesh, takes hours and even days.
However, for such small penetration, induced current can be precisely calculated including a surface impedance to the equations and neglecting magnetic field inside the workpiece. In this case, we can avoid building a super fine mesh for the workpiece but use rough 2D surface mesh only.
The computational gain depends on frequency. While computational time of traditional FEM dramatically increases with frequency, exceeding a week for MHz cases, the Surface Impedance approach always delivers results in several minutes.
Read more about the Surface Impedance approach in the paper by Nerg & Partanen (2001)
Calculation accuracy
The Surface Impedance approach is mostly used to estimate induced current and Joule heat distribution, since it is predicted within 97% of accuracy already at relatively low frequencies. The skin-depth to workpiece-thickness (d/D) is suggested as a non-dimensional criteria here.
While the Surface Impedance approach predicts apparent power in a coil and Joule heat in a workpiece already at d/D < 10%, accurate temperature prediction is possible only at d/D < 0.5%. Heat is defined as a surface source, alike a cooling condition (convection or/and radiation). Therefore, temperature is underestimated with up to 30% inaccuracy for lower frequencies.
So, the Surface Impedance approach, implemented in CENOS Induction Heating simulation software, is a powerful design tool for high frequency installations.
#3D, #strip, #SurfaceImpedance, #highfrequency, #Jouleheat
