RHINO-MAG: Recursive H-Field Inference based on Observed Magnetic Flux under Dynamic Excitation

Abstract

Driven by the MagNet Challenge 2025 (MC2), increased research interest is directed towards modeling transient magnetic fields within ferrite material. An accurate time-resolved and temperature-aware H-field prediction is essential for optimizing magnetic components in applications with quasi-stationary / non-stationary excitation waveforms. Within the scope of this investigation, a selection of model structures with varying degrees of physically motivated structure are compared. Based on a Pareto investigation, a rather black-box gated recurrent unit (GRU) model structure with a graceful initialization setup is found to offer the most attractive model size vs. model accuracy trade-off, while the physics-inspired models performed worse. For a GRU-based model with only 325 parameters, a sequence relative error of 8.02 % and a normalized energy relative error of 1.07 % averaged across five different materials are achieved on unseen test data. With this excellent parameter efficiency, the proposed model won the first place in the performance category of the MC2.

Figures (9)

• Figure 1: Exemplary prediction task for $H$ based on available $B$.
• Figure 2: Overview of the model structure. 'N' is a normalization operation, while 'D' is the denormalization.
• Figure 3: Detailed view of model warmup and $H$-trajectory estimation for the . The superscript numbers resemble specific indexing of a vector (in Python style).
• Figure 4: Exemplary prediction performance for the and the presented Preisach model variation. Note that within this preliminary evaluation, the normalized magnetic flux density $\tilde{B}$ was the prediction target and the normalized magnetic field $\tilde{H}$ was provided, which is reversed compared to the direction generally targeted within this work and .
• Figure 5: One forward propagation step for the vector field .