Non-invasive diagnosis of SF6 high voltage Selfblast circuit breaker nozzles

The maintenance of high voltage gas circuit breakers requires high personal as well as monetary efforts for the asset operator. Furthermore, a faulty reassembly of a circuit breaker during maintenance can lead to a circuit breaker failure during operation. One possibility to reduce those efforts as well as the failure risk is the application of non-invasive diagnostic techniques. This research work examines such a non-invasive technique for assessing the wear of the insulating nozzle inside the switching chamber of a circuit breaker. The approach applied is based on the measuring of the transient pressure signal at the gas connector of the circuit breaker during a switching operation without electrical load. The pressure signal is investigated regarding characteristically features which yield information for the determination of the condition of the switching chamber. This contribution addresses two different approaches for the evaluation of the measured pressure signals: firstly a comparative analysis of the signals and secondly an evaluation based on machine learning algorithms. Both of them use a dataset containing measurements with different variations of the nozzle shape representing the wear of the nozzle. In the first approach, the characteristics of measurements with different nozzle diameters are compared and assessed with regard to deviations of the characteristics. In the second approach, machine learning algorithms are used to automatically scan for dependencies in the given dataset obtained from the measurement data. Applied on the results of exemplary tests, both approaches provide a clear indication with regard to a prediction of the nozzle wear and show a proof of concept for the new non-invasive diagnostic technique. Taking into account a large database from field tests planned in the future, especially the machine learning algorithms can show a valuable benefit for the application of this test measure during circuit breaker maintenance.