A comparable method for benchmarking the reliability performance of electric utilities

This paper describes the results of an assessment of various methods of identifying events whose impact on the reliability of electric service to customers is so severe as to dominate reliability statistics and mask the true design quality of the power delivery system or the true operational performance of the electric utility. Alternatives are compared based on realistic applicability within a utility and statistical results when applied to historical data from two major metropolitan areas in the USA.. A method based on the lognormal probability distribution applied to daily SAIDI is recommended. This work was part of an effort of the IEEE Power Engineering Society´s Task Force on Reliability Indices.