Research on Fault Diagnosis Model of SF₆ Circuit Breaker Based on FNN Improved by GA

Abstract

The traditional fault diagnosis methods of SF6 circuit breakers lack systematicity, which are intricate and time-consuming in fault detection and maintenance. A fault diagnosis model of SF6 circuit breaker based on fuzzy neural network improved by genetic algorithm is established to solve the similar problems. In this paper, an online fault diagnosis method for SF6 circuit breaker based on fuzzy neural network improved by genetic algorithm is established. This method adopted the characteristic data of SF6 circuit breakers to build a fault diagnosis model which is trained by inputting parameters, including the decomposition products of SF6 gas in the circuit breaker and the temperature, pressure and density, to diagnose the corresponding faults. Finally, the fault diagnosis model of fuzzy neural network improved by genetic algorithm compared with other diagnosis models in terms of training time and accuracy. The simulation results show that the proposed fault diagnosis model is more efficient and accurate than other available methods. The fault diagnosis model has also been applied to the online monitoring technology of the smart substation with good operations.