Evaluation of arresters' failure probability using ANN

  • Authors:

  • C. A. Christodoulou;I. A. Stathopulos

  • Affiliations:

  • National Technical University of Athens, School of Electrical and Computer Engineering, High Voltage Laboratory, Athens, Greece;National Technical University of Athens, School of Electrical and Computer Engineering, High Voltage Laboratory, Athens, Greece

  • Venue:
  • ISTASC'08 Proceedings of the 8th conference on Systems theory and scientific computation
  • Year:
  • 2008

Quantified Score

Hi-index 0.00

Visualization

Abstract

Lightning is a major cause of faults on overhead transmission high voltage lines. In order to improve the lightning performance of the lines and to reduce the failure rate, surge arresters are installed between each phase and earth. The failure probability of the arresters can be evaluated using appropriate methodologies and software tools. In the current work an ANN model is used for this purpose. The ANN model, which is trained using the simulation results of Simulink®7.1 package, has five inputs (the tower footing resistance, the peak lightning current, the nominal discharge current of the arrester, the arrester's energy withstand capabilities and the arrester's interval) and one output (the arrester's failure probability). The ANN model's results are compared with the simulation results showing satisfactory agreement.