Ziqing Guo;Binh H. Nguyen;Hamed Hamzehbahmani;Ruth V. Sabariego
{"title":"Dynamic Hysteresis Model of Grain-Oriented Ferromagnetic Material Using Neural Operators","authors":"Ziqing Guo;Binh H. Nguyen;Hamed Hamzehbahmani;Ruth V. Sabariego","doi":"10.1109/TMAG.2025.3600089","DOIUrl":null,"url":null,"abstract":"Accurately capturing the behavior of grain-oriented (GO) ferromagnetic materials is crucial for modeling electromagnetic devices. In this article, neural operator models, including Fourier neural operator (FNO), U-net combined FNO (U-FNO), and deep operator network (DeepONet), are used to approximate the dynamic hysteresis models of GO steel. Furthermore, two types of data augmentation strategies, including cyclic rolling augmentation and Gaussian data augmentation (GDA), are implemented to enhance the learning ability of models. With the inclusion of these augmentation techniques, the optimized models account for not only the peak values of the magnetic flux density but also the effects of different frequencies and phase shifts. The accuracy of all models is assessed using the <inline-formula> <tex-math>$L2$ </tex-math></inline-formula>-norm of the test data and the mean relative error (MRE) of calculated core losses. Each model performs well in different scenarios, but FNO consistently achieves the best performance across all cases.","PeriodicalId":13405,"journal":{"name":"IEEE Transactions on Magnetics","volume":"61 10","pages":"1-7"},"PeriodicalIF":1.9000,"publicationDate":"2025-08-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"IEEE Transactions on Magnetics","FirstCategoryId":"5","ListUrlMain":"https://ieeexplore.ieee.org/document/11129051/","RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"ENGINEERING, ELECTRICAL & ELECTRONIC","Score":null,"Total":0}
引用次数: 0
Abstract
Accurately capturing the behavior of grain-oriented (GO) ferromagnetic materials is crucial for modeling electromagnetic devices. In this article, neural operator models, including Fourier neural operator (FNO), U-net combined FNO (U-FNO), and deep operator network (DeepONet), are used to approximate the dynamic hysteresis models of GO steel. Furthermore, two types of data augmentation strategies, including cyclic rolling augmentation and Gaussian data augmentation (GDA), are implemented to enhance the learning ability of models. With the inclusion of these augmentation techniques, the optimized models account for not only the peak values of the magnetic flux density but also the effects of different frequencies and phase shifts. The accuracy of all models is assessed using the $L2$ -norm of the test data and the mean relative error (MRE) of calculated core losses. Each model performs well in different scenarios, but FNO consistently achieves the best performance across all cases.
期刊介绍:
Science and technology related to the basic physics and engineering of magnetism, magnetic materials, applied magnetics, magnetic devices, and magnetic data storage. The IEEE Transactions on Magnetics publishes scholarly articles of archival value as well as tutorial expositions and critical reviews of classical subjects and topics of current interest.