Determination of effect of slot form on slot leakage flux at rotating electrical machines by the method of artificial neural networks


Nogay H. S., Akinci T. C., Guseinoviene E.

ENERGY EDUCATION SCIENCE AND TECHNOLOGY PART A-ENERGY SCIENCE AND RESEARCH, vol.29, no.1, pp.451-462, 2012 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 29 Issue: 1
  • Publication Date: 2012
  • Journal Name: ENERGY EDUCATION SCIENCE AND TECHNOLOGY PART A-ENERGY SCIENCE AND RESEARCH
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED)
  • Page Numbers: pp.451-462
  • Keywords: Slot Leakage Self-Induction coefficient, Slot equivalent conductivity, Artificial neural networks, Back propagation, Slot form, PREDICTION, BIODIESEL, MOTORS
  • Kayseri University Affiliated: No

Abstract

In this study, first of all, for the one-layered winding form of the rotating electrical machines, self-induction coefficient of slot leakage has been calculated according to varying forms of slot. Given the differences in the rotating electrical machines, a dataset has been created with the slot size measurements required for each slot form as well as rotor connection forms (with squirrel cage with wound rotor) representing input and with the slot equivalent conductivity representing the output. Then, a dataset has been used to train an ANN forecasting model. Coefficient of slot leakage self-induction so calculated was compared with coefficient of predicted slot leakage self-induction. With the trained ANN model, slot equivalent conductivity of a rotating electrical machine with any size for a possible slot form has been predicted with an accuracy rate of 99.998%.