Statistics of circuit-breaker performance

A wide range of data is presented which shows that the electrical performance of high-voltage a.c. circuit breakers is statistical. The electrical performance cannot be characteristed by a single stress level such as a rated stress; a stress ratio is also required to specify the spread of operating conditions over which both clearances and failures occur fairly often. Present test specifications are inadequate to ensure that a circuit breaker is reliable on the basis of the formal tests alone. Calculations are presented relating the failure rate of a multiunit circuit breaker to that of its separate units. Suggestions are made for improved test strategies in which the present three shots on a multiunit circuit breaker could be supplemented by unit tests to predict a safe rating for the complete circuit breaker.