Modelling of high voltage SF6 circuit breaker reliability based on Bayesian statistics

Abstract

The biggest drawback in modelling the reliability of high voltage circuit breakers is the lack of access to data on failures in service, due to the very long lifetime of circuit breakers. This paper presents the application of reliability calculation based on Bayesian statistics to a 245 kV SF6 circuit breaker and its operating mechanism. By using the Bayesian theorem, the prior probability density function of failures in circuit breaker components, which is calculated based on data on the circuit breaker and the operating mechanism failures in service, is combined with data on the failures registered during an extensive mechanical development tests. During the tests more than 32000 "CO" (close-open) operations were performed. Based on the posterior probability density function, the reliability of circuit breaker components and the overall reliability of the breaker is estimated. The paper also presents some analysis of the impact of circuit breaker maintenance on its reliability.