• DocumentCode
    357846
  • Title

    Crosstalk prediction on wire bundles by Kriging approach

  • Author

    Salio, S. ; Canavero, F. ; Lecointe, D. ; Tabbara, W.

  • Author_Institution
    Dept. of Electron., Politecnico di Torino, Italy
  • Volume
    1
  • fYear
    2000
  • fDate
    2000
  • Firstpage
    197
  • Abstract
    A wide variety of interconnections among electronic devices are performed by means of randomly twisted bundle of wires, particularly for transport and industrial applications. This work examines the crosstalk analysis of such bundles. A new model based on fractals for describing the position of wires winding throughout a bundle is presented. It allows the wires´ meandering degree to be tuned simply by means of a parameter (the fractal dimension). It also allows for a statistical analysis of crosstalk because it can simulate independent different realizations of the same interconnection. An efficient multifactor prediction method known as “Kriging” is applied in order to describe the crosstalk as a function of the input factors (cable loads, cable height above the ground plane, etc.) of the cable model. It allows for a drastic reduction in the simulation time insofar as it predicts the crosstalk at untried combinations of the input factors from a small base of input combinations. In comparing the crosstalk simulated using the fractal model with the experimental measurements good agreement was found
  • Keywords
    crosstalk; electromagnetic compatibility; fractals; interconnections; statistical analysis; wires (electric); EMC problems; Kriging approach; cable height; cable loads; cable model; crosstalk analysis; crosstalk prediction; electronic devices; experimental measurements; fractals; ground plane; industrial applications; interconnection; interconnections; multifactor prediction method; randomly twisted wire bundles; simulation time reduction; statistical analysis; transport applications; Analytical models; Crosstalk; Electronics industry; Fractals; Industrial electronics; Power cables; Predictive models; Statistical analysis; Transmission line theory; Wire;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Electromagnetic Compatibility, 2000. IEEE International Symposium on
  • Conference_Location
    Washington, DC
  • Print_ISBN
    0-7803-5677-2
  • Type

    conf

  • DOI
    10.1109/ISEMC.2000.875562
  • Filename
    875562