• DocumentCode
    3572612
  • Title

    The application of intelligent systems in power transformer design

  • Author

    Geromel, H. Luiz ; Souza, R. Carlos

  • Author_Institution
    CEFET/PR, Sao Paulo, Brazil
  • Volume
    1
  • fYear
    2002
  • fDate
    6/24/1905 12:00:00 AM
  • Firstpage
    285
  • Abstract
    The purpose of this work is to present a novel power transformer design methodology using intelligent systems. The methodology described in this paper allows the application of the artificial neural networks in some specific stages of the design. As it is well known, designing power transformers may be a time-consuming task owing to the large number of variables to be considered and their multiple interrelations, which are not all completely known. This novel methodology is, therefore, an important tool, not only for optimizing the projects, but also for minimizing substantially the necessary time for their execution, as it simplifies the design procedure. A database that contains the test results of several hundreds of actual power transformers is used for the validation of the method The design methodology is presented in detail and the results of its application are discussed in this work.
  • Keywords
    eddy current losses; magnetic hysteresis; neural nets; power engineering computing; power transformers; transformer cores; transformer windings; artificial neural networks; core geometry; core losses; database; eddy current losses; hysteresis losses; intelligent systems; power transformer design methodology; transformer temperature; winding impedance; winding losses; Artificial intelligence; Artificial neural networks; Costs; Design methodology; Design optimization; Intelligent systems; Power transformer insulation; Power transformers; Production; Temperature;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Electrical and Computer Engineering, 2002. IEEE CCECE 2002. Canadian Conference on
  • ISSN
    0840-7789
  • Print_ISBN
    0-7803-7514-9
  • Type

    conf

  • DOI
    10.1109/CCECE.2002.1015233
  • Filename
    1015233