A first look

From magnetic behavior to a learned core-loss model.

Magnetic waveforms, temperature, and frequency become the inputs to an AI model of volumetric core loss.

Full workshop opens ahead of IEEE APEC 2027

A dark ferrite magnetic core with copper windings, fine field arcs, and abstract connected AI nodes on cream paper. — An illustrative preview of AI core-loss modeling.

The case study

We’ll use MagNet Challenge core-loss measurements to explore how to vibe code an AI model to learn relationships between magnetic excitation and core loss.

InputsMagnetic flux density waveform B(t), temperature, and frequency

ModelFeedforward neural networks and Generalized Steinmetz equation

OutputVolumetric core loss P_v, in W/m³

The hands-on prompts and experiments are coming with the full workshop.

Back to Overview