The variation of temperature modifies the magnetic behavior of ferromagnetic cores which may affect the performance of electrical equipment. Therefore, it is imperative to develop a temperature-dependent hysteresis model to precisely calculate electromagnetic characteristics of electrical equipment. In this paper, a Temporal Convolutional Network (TCN) in combination with the Play operator is developed. The proposed model incorporates the temperature-dependent spontaneous magnetization intensity as the model input to introduce the temperature effect. To enhance the accuracy of model training outcomes, the Bayesian optimization approach for automatically selecting network model parameters is provided. The results show that the proposed model can accurately predict the hysteresis characteristics of materials under varying temperature and frequency conditions.
Read full abstract