TR2023-042

Comparison of Learning-based Surrogate Models for Electric Motors


    •  Xu, Y., Wang, B., Sakamoto, Y., Yamamoto, T., Nishimura, Y., "Comparison of Learning-based Surrogate Models for Electric Motors", Conference on the Computation of Electromagnetic Fields (COMPUMAG), May 2023.
      BibTeX TR2023-042 PDF
      • @inproceedings{Xu2023may,
      • author = {Xu, Yihao and Wang, Bingnan and Sakamoto, Yusuke and Yamamoto, Tatsuya and Nishimura, Yuki},
      • title = {Comparison of Learning-based Surrogate Models for Electric Motors},
      • booktitle = {Conference on the Computation of Electromagnetic Fields (COMPUMAG)},
      • year = 2023,
      • month = may,
      • url = {https://www.merl.com/publications/TR2023-042}
      • }
  • MERL Contact:
  • Research Areas:

    Applied Physics, Artificial Intelligence, Electric Systems, Multi-Physical Modeling

Abstract:

Multi-objective optimization is frequently employed in electric motor design, where iterative numerical simulations are required to evaluate a large number of design candidates. A trial-and-error design methodology like this is very time-consuming. In this paper, we propose learning-based surrogate models that use trained deep neural networks (NNs) to accomplish the rapid evaluation of motor designs. A motor design candidate can be described with either a list of geometrical parameters of the motor design, or a colored image of the motor cross-section. Different deep learning models can be constructed with either parameter-based input or image-based inputs. Our analysis reveals that deep convolutional neural networks (CNNs) utilizing image-based inputs exhibit a higher degree of predictive accuracy for more intricate responses, such as cogging torque, in comparison to models employing parameter-based inputs, albeit at the cost of increased training time.