Study Of High-voltage Circuit Breaker Fault Diagnosis Based On RBFNN

In the electrical power system, equipment failures not only have an effect on its normal operation and economics, but also probably lead the overall power system to collapse, especially while the severe failure of equipment happens. Therefore, the earlier prediction and diagnosis on the failures of equipments especially the critical equipment, which play an importmant role to guarantee the safity of the power system and improve the reliability of the electrical power system. In recent years, a growing attention has been paid to studying it. The high-voltage circuit breaker as a kind of electric automation equipment is the link between electricity generations and supply of power. In the thesis, the artificial neural network technology is studied to diagnose the mechanical fault of high-voltage circuit breaker .Firstly, the model and the characteristics of RBF neural network as well as the the generalization ability of RBF are discussed in detail. Simulations of the prediction and indentification on the system pamater based on the RBF neural network as well as the ones of fault diagnosis are given, and the simulating results show the method is useful.Secondly, as the core of this paper, the structures of the RBF neural network and learning algorithms are studied. A combinative method of RBF neural network based on the orthogonal and the gradient algorithm is presented, the simulation results using Matlab indicates that the proposed method has characteristics of fast training, high accuracy and the error concentrations, compared to the traditional BP(Back Propagation algorithm,BP) neural network and the RBF neural network based on the clustered learning algorithm.Finally, with the research above and the collection of training samples, through the analysis on really scenes of the vacuum circuit-breaker operating system and the Trip Coil Profiling curve characteristics of its electromagnetic coil, a kind of mechanical fault diagnosis model for the vacuum circuit-breaker is proposed. The experimental result indicates that the proposed method has higher accuracy and more practical value.