Research On Simulation And Improved Fuzzy Petri Net Identification Method Of Magnetizing Inrush Current In Transformer

In transformer differential protection,how to identify inrush current accurately and quickly is very important to improve the correct operation rate of transformer.Therefore,this thesis presents a method of transformer magnetizing inrush current identification based on improved fuzzy Petri net.The simulation models of single-phase transformers and three-phase transformers are deeply studied in combination with the generation mechanism and transient process of transformer magnetizing inrush current.By summarizing the waveform of inrush current and internal fault current,the difference between the two is obtained,which lays the foundation for the proposed method.Through the comparative analysis of inrush current identification methods,the feature extraction method based on EMD-SVD,wavelet packet decomposition method and harmonic analysis method are combined to extract feature quantities in this thesis.Three characteristic vectors of singular spectral entropy,wavelet energy,and harmonic components are integrated as an input eigenvector in the improved fuzzy Petri net model,to make it more stable performance.The improved fuzzy Petri net model is used for identification and prediction in transformer magnetizing inrush current.Aiming at the defects of fuzzy Petri net parameters,the BP neural network is used to optimize the parameters of weight,threshold and credibility.The final convergence value is used as the parameter value to establish the fuzzy Petri net,so as to realize the optimization of the fuzzy Petri net.In order to verify the superiority of the improved fuzzy Petri net,five control groups are set up to simulate the same sample under the same conditions.The approximation effect and error distribution of the actual output and expected value of the sample are analyzed comprehensively.Simulation experiments show that proposed method of extracting the inrush current eigenvector is stable and has a high degree of discrimination.At the same time,the improved fuzzy Petri nets has higher recognition accuracy,so it can effectively identify excitation inrush current,which improves the recognition accuracy of intermal fault currents.