Conference: 2024 IEEE Energy Conversion Congress and Exposition (ECCE)
Authors: Ji-Hyeon Lee, Hyun-Su Kim, Soo-Hwan Park, Jin-Cheol Park, Myung-Seop Lim
DOI: 10.1109/ECCE55643.2024.10861858
In recent times, there has been growing interest in electric motors, particularly induction motors (IMs), which do not require rare-earth materials. IMs efficiency is a critical parameter as it directly influences the overall efficiency of the system. The iron loss is a significant factor in assessing the overall efficiency of IMs. Therefore, it is crucial to make accurate predictions of iron loss during the design of IMs. Analyzing iron loss of IMs using transient analysis offers accuracy but demands high computational costs due to induced current in the rotor bars and the difference between the stator current frequency and slip frequency. Conversely, utilizing the virtual blocked rotor (VBR) for analysis reduces computational cost but decreases accuracy, because this method can match the stator current frequency and slip frequency. Thus, in this paper, we propose a method aimed at enhancing the accuracy of the VBR while retaining its advantage of lower computational cost compared to conventional transient analysis, utilizing both the VBR and transfer learning. The transfer learning is conducted using a large amount of iron loss excluding slot harmonic components derived from the VBR and a small amount of iron loss including slot harmonic components derived from transient analysis. This method enables accurate determination of iron loss including slot harmonic components with a limited dataset, thereby improving the efficiency of the IMs design process.