Research On Fault Prediction Method Of High Voltage Circuit Breaker

High-voltage circuit breakers are widely installed in various parts of the power system.They are important switching devices in power systems and one of the most complex power devices.Due to its relatively frequent movements,the probability of failure is relatively large.If it fails,it will not be able to control and protect the safe and stable operation of the entire power system,causing huge economic losses.The importance of fault diagnosis and prediction of high-voltage circuit breakers is self-evident.According to statistical analysis,most of the types of faults that occur in high-voltage circuit breakers are mechanical faults.In this paper,a 500kV porcelain column SF₆ high voltage circuit breaker is taken as the research object,and a set of high voltage circuit breaker mechanical fault prediction system is designed.The main research contents are as follows:（1）Design of signal acquisition system for high voltage circuit breaker operation characteristics.Combined with the Internet of Things technology,the online monitoring system for the collection and closing of the high-voltage circuit breaker,the three-phase contact displacement signal,the temperature signal and the spring mechanism energy storage signal are built.The part includs the overall framework design of the system,the design of the hardware platform,the selection of devices,the construction of the server,and so on.（2）Research on fault diagnosis method of high voltage circuit breaker.In this part,the quantum neural network is studied,and the fault diagnosis model of high voltage circuit breaker based on improved quantum neural network is constructed.The characteristic signal is taken as input and the model is built under the TensorFlow deep learning framework.The results show that the model has a significant improvement in the number of iterations of the training and the diagnostic error.（3）Research on fault prediction methods for high voltage circuit breakers.This part combines the LSTM（Long Short Term Memory）cyclic neural network to solve the long-term dependence of data.The LSTM cyclic neural network model is constructed by taking the time-series of the closing coil current of the high-voltage circuit breaker as the characteristic input.CAS（Chaotic Ant Colony Algorithm）is used as a training optimization algorithm.The fault prediction method of high voltage circuit breaker based on CAS optimized LSTM cyclic neural network is proposed and compared with the training optimizer provided by TensorFlow deep learning framework.The simulation results show that the method has higher prediction accuracy and shorter training steps.（4）Interactive platform design.This part uses LabVIEW as a platform to build a high voltage circuit breaker fault prediction system,which effectively links the characteristic signal acquisition system,the high voltage circuit breaker fault diagnosis model and the fault prediction model.The part completed the tasks of system framework and design,user login and prediction interface design,server connection and so on.Finally,the circuit breaker fault prediction platform is run,and the results show that the fault prediction function of the high voltage circuit breaker can be basically realized.