Hysteresis Loop Parameter Identification Based on the Preisach Model under Different Excitations

Introduction/Objective: The classic Preisach hysteresis model is limited by the difficulty of determining its distribution function and the inherent numerical errors during computational implementation. This study aims to overcome these limitations by developing an improved modeling and identification framework to enhance the accuracy and efficiency of simulating hysteretic characteristics. Methods: As a basis for the creation of an upgraded model, hyperbolic cosine functions are utilized in conjunction with the reconstructed distribution function, which generates the analytical representation of the Everett function. Following this, a Preisach Analytical Model is developed with clearly defined mathematical expressions, significantly reducing the required computational overhead. Additionally, parameter identification is performed more efficiently using particle swarm optimization, differential evolution, and an upgraded adaptive differential evolution algorithm. Results: The validation process conducted through experimental data has confirmed an accurate overall fit of the proposed improved hysteresis model to all the static and dynamic hysteresis characteristics of both silicon steel and nanocrystalline materials when subjected to both symmetric and asymmetric excitation. Additionally, the validation results provide evidence that the proposed model is capable of matching the local hysteresis loop data from complex energetic input signals with greater accuracy than previous models. Discussion: The explicit analytical formulation effectively handles reversible magnetization at loop tips, and integrating chaotic mapping with adaptive algorithms physically stabilizes the parameterextraction process. Conclusion: The study provides theoretical justification for the use of an optimally effective methodology to accurately model the hysteresis characteristics of materials.