Research On High-voltage Circuit Breaker Intelligent Fault Diagnosis

Along with the rapid growth of the national economy, grid security and stability operation have been got the more and more attention of people, the high voltage circuit breaker has been put forward higher request becase it play two role of protection and control in grid, the high voltage circuit breaker failure not only may lead to grid of an accident, but also may cause the collapse of the entire grid. Therefore, the study of high voltage circuit breaker intelligent fault diagnosis methods, is not only beneficial to analyze the working state of the high voltage circuit breaker, improve grid security and stability, but also to avoid unnecessary maintenance work, improve the economic benefits of grid.This paper first introduced the importance of electric power system in the national economy, statistics at home and abroad in recent years has undergone significant grid accident and the proportion of high voltage circuit breaker in the power grid accident, summarizes the high voltage circuit breaker intelligent fault diagnosis method the latest results, analyzes of the problems of the existing high voltage circuit breaker intelligent fault diagnosis method. According to Conference international des Grands Reseaux Electriques and China Electric Power Research Institute survey of the high voltage circuit breaker in working condition, we can see the high voltage circuit breaker is existing fault type and the reasons..Secondly, using the theory of quotient space realized the feature extraction of high voltage circuit breaker, the quotient space theory can be achieved from different granularity levels to analyze huge amounts of data, in the case of ensure accuracy problem solving, be able to choose the most appropriate granularity space, and can also reduce the complexity of problem solving. By analyzing the high voltage circuit breaker points, closing coil current signal in the role of intelligent fault diagnosis, combining with the quotient space of properties and projection method to realize high voltage circuit breaker points,closing coil current historical data feature extraction for neural network training the feature set.Finally, the high voltage circuit breaker intelligent fault diagnosis model is studied from the perspective of the neural network. By using the theory of quotient space to extract the high voltage circuit breaker points, closing coil current signal characteristics of set, which is ensure the feature set and fault type of correlation, and also reduces data redundancy to prevent neural network running into the infinite cycle. Quotient space to extracted the feature set to build based on BP network, RBF neural network and probabilistic neural network of high voltage circuit breaker intelligent fault diagnosis model and use the high-voltage circuit breaker monitoring points, the closing coil current signal to verified the model, analysis and comparison of the three models in high voltage circuit breaker intelligent fault diagnosis of the advantages and disadvantages, probabilistic neural network fault diagnosis effect better than the other two kinds of network, its error and running time is minimum, and this model convergence rate quick, good fault tolerance and high stability, and have strong ability of additional, can better diagnose the fault type and position of the high voltage circuit breaker, timely make maintenance for electric power department to provide some practical guidance basis.