Algorithm for Automated Assessment of the Condition of a High-Voltage SF₆ Circuit Breaker Based on a Limited Data Set

One of the main tasks of condition-monitored maintenance and repair of electrical equipment is to assess its technical condition. The current regulations provide a procedure for assessing the technical condition of electrical equipment by calculating the health index (HI). To calculate the HI of a 110 kV tank-type SF₆ circuit breaker, it is necessary to evaluate 46 parameters. The value of the HI should be updated and predicted for the next 5 years at least once a year. However, to update about half of the 46 parameters, it is required to disable the circuit breaker. This paper proposes an algorithm for automating the evaluation of the HI of a circuit breaker and determining the HI from a limited set of data obtained from the automated systems installed at the facility. A complete set of retrospective data is used for machine learning model, which allows automatic calculation of the HI of the circuit breaker during operation without the participation of maintenance personnel. Retrospective data on 20% of like equipment are used for training the model. After that, the HI of the remaining equipment can be automatically determined from a limited data set, without human intervention. This will allow us to reduce the resources for calculating the HIs of circuit breakers and to determine the HI of a circuit breaker during operation, without taking it out of operation