Parameter identification techniques applied to dynamic MOSA models

• Published in: 2014 ICHVE International Conference on High Voltage Engineering and Application pp. 1 - 4
• Main Authors: Xavier, T. S., Lira, G. R. S., Costa, E. G., Maia, M. J. A.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.09.2014
• Subjects: Artificial intelligence; Lighting; Reliability; Surges
• DOI: 10.1109/ICHVE.2014.7035493

The metal-oxide surge arresters (MOSA) are equipment used in power systems protection against several kinds of surges. They decisively contribute for the increase of the reliability, economy and continuity of the system operation, which they protect. Due to the importance of the MOSA for the electrical systems, several models have been proposed to represent the MOSA behavior under overvoltages. The majority of the existing models employ empirical adjustment procedures for determination of models' parameters. So, a comparative study about numerical techniques applied to parameter identification of metal-oxide surge arrester models is presented. These techniques are based on the fitting of the residual voltages measured and obtained from the surge arrester models, from the 10 kA lightning current impulse (wave shape 8/20 µs). Optimization algorithms with characteristics of global convergence were evaluated. The employed procedures have provided good results, showing that it is more accurate to apply optimization algorithms to identify the MOSA models parameters than empirical procedures, thus increasing the reliability of the modeling process.