Abstract:
Aiming at the insufficient study of partial discharge(PD) characteristics of MV circuit breakers subjected to a wide range of pressure and voltage stresses in a mixed gas insulation environment, a time-frequency clustering-based study of PD in the circuit of a MV circuit breaker filled with SF
6/N
2 mixed gas is proposed. The partial discharge characteristics of a 22 kV MV circuit breaker prototype when filled with SF
6/N
2 mixed gas were analyzed using time-frequency mapping(T-F mapping) and cluster analysis techniques. The PD test was carried out in the pressure range of 1~5 bar and under different gas ratios, focusing on key parameters such as PD onset voltage, phase-resolved PD(PRPD) pattern and discharge repetition rate. Through the signal clustering analysis of time-frequency mapping, three types of discharge pulse clusters with significant differences were successfully identified: short-duration pulse clusters, long-duration pulse clusters and repetitive pulse clusters. Real-time monitoring of the density changes of each cluster can provide a dynamic assessment basis for the insulation degradation of circuit breakers. The study further shows that, under the premise of ensuring insulation performance, when the SF
6 gas percentage is not less than 10%, a gas mixture configuration of 10% SF
6 and 90% N
2 at a pressure of 3 bar can be used to achieve excellent PD suppression in 22 kV medium-voltage circuit breakers.