Research On Application Of High Voltage Circuit Breaker Status Evaluation Method Based On Chaotic Neural Networks

Electrical equipment on-line condition monitoring and diagnosis is an important content of intelligent transmission in real time.High voltage circuit breaker as the key equipment of the control and protection of electric power system,monitoring,evaluation and early diagnosis for it can effectively protect the power grid,improve the reliability of power system.In recent years,people use a variety of theoretical knowledge to analysis of high voltage circuit breaker status,have made the practicability of the high voltage circuit breaker on-line monitoring system,can be a very good real-time monitoring the condition and judging fault.This paper adopted a kind of chaos theory and neural network method of combining the state of high voltage circuit breaker mechanical performance for evaluation.The article analyze the condition assessment parameters,the mechanical properties and the parameters of the chaotic characteristics of high voltage circuit breaker,and build the state information table,which can evaluate the status of high voltage circuit breaker qualitatively.Chaotic neuron model was studied,and then build the chaotic neural network.On the basis of BP network,we established a hybrid chaotic neural network based on Logistic mapping,and studied the network learning algorithm and the training process.By introducing chaos system,the hybrid chaotic neural network can overcome BP neural network the shortcoming of easily trapped in local minimum,and the nuances of the model has better recognition effect.In order to verify the effectiveness of the method,this article choose the data samples of close coil current extraction and dynamic contact part of the eigenvalues from the database of laboratory research and normalizd the data.Combining with Matlab R2014 a,the BP neural network and mixed chaotic neural network respectively to the simulation and analysis,with multiple sets of feature vector as the two kind of network training and testing input,its corresponding fault status as the output of the network.Simulation results show that the hybrid chaotic neural networks can effectively identify the state of high voltage circuit breaker,and is now widely used BP neural network simulation results contrast,high accuracy,fast,small error,high practicability and reliability,embodies the hybrid chaotic neural network is applied to the advantages of high voltage circuit breaker status evaluation.This method can be extended to other equipment condition assessment.