Magnetic hysteresis is the lag between magnetic field and magnetization in ferromagnetic materials. The magnetic field inside the material depends not only on the current applied field but also on the history of the field changes.
Physical Models
Classical Models
- Preisach model for SPS main magnets - Distribution of switching thresholds
- Jiles-Atherton model for hysteresis - Energy-based approach with parameters , , , ,
- Bouc-Wen model for magnetic hysteresis - Used in PhyLSTM architecture
Model Parameters
Each model requires different parameter sets that capture the magnetic material properties of SPS magnets.
Neural Network Integration
Physics-Informed Architectures
- PhyLSTM - Incorporates Bouc-Wen model for magnetic hysteresis
- PETE - Physics-enhanced transformer encoder
- Neural ODE - Continuous dynamics modeling
Data-Driven Approaches
- Temporal Fusion Transformer - Primary production model
- LSTM variants for sequence modeling
Related Concepts
- Hysteresis Compensation - Engineering studies to mitigate hysteresis effects
- SPS Main Dipole Magnets - Primary application target
- Magnetic Field Prediction - Computational objective