• DocumentCode
    1320192
  • Title

    Multiobjective Sequential Optimization Method for the Design of Industrial Electromagnetic Devices

  • Author

    Lei, Gang ; Chen, X.M. ; Zhu, J.G. ; Guo, Y.G. ; Xu, Wei ; Shao, K.R.

  • Author_Institution
    Fac. of Eng. & Inf. Technol., Univ. of Technol., Sydney, NSW, Australia
  • Volume
    48
  • Issue
    11
  • fYear
    2012
  • Firstpage
    4538
  • Lastpage
    4541
  • Abstract
    A multiobjective sequential optimization method (MSOM) is presented to deal with practical design problems of industrial electromagnetic devices. MSOM consists of a sequential optimization strategy of multiobjective optimization model and a modified central composite design (CCD) sampling method. To improve the optimization efficiency, Kriging model is employed to construct the approximate multiobjective optimization models. Then a modified CCD sampling method is presented to update the sample sets with the obtained Pareto optimal points and Kriging models. Thereafter, by investigating a test function and a three-dimensional permanent magnet transverse flux machine, it can be found that the proposed method is efficient, and the computation cost of finite element analysis can be saved remarkably.
  • Keywords
    Pareto optimisation; approximation theory; electromagnetic devices; finite element analysis; permanent magnet machines; sampling methods; Kriging model; MSOM; Pareto optimal point; approximate multiobjective optimization model; finite element analysis; industrial electromagnetic device; modified CCD sampling method; modified central composite design sampling method; multiobjective sequential optimization method; three-dimensional permanent magnet transverse flux machine; Computational modeling; Electromagnetic devices; Finite element methods; Pareto optimization; Sampling methods; Central composite design; Kriging model; PM transverse flux machine; multiobjective optimization;
  • fLanguage
    English
  • Journal_Title
    Magnetics, IEEE Transactions on
  • Publisher
    ieee
  • ISSN
    0018-9464
  • Type

    jour

  • DOI
    10.1109/TMAG.2012.2202311
  • Filename
    6332590