Transmission Line Faults Identification Method Based On HHT And PNN

In recent years,The rapid growth of the national economy makes the need for more electric energy.High-voltage transmission lines,as a key component of long-distance power transmission in power systems,have been subject to various kinds of determination and random interference factors.The uncertainty and even non-linearity of the faults require higher and higher fault diagnosis capability.Artificial neural network,as an algorithm with good classification ability that has developed rapidly in recent years,has gradually become a new attempt to solve the transmission line fault problem.This paper proposes a fault identification method based on Hilbert-Huang Transform(HHT)and Probabilistic Neural Network(PNN).This method is divided into fault feature extraction and fault classification identification.For fault feature extraction of transmission lines,HHT can fully reflect the characteristics of local transient signals,and EEMD is used to decompose short circuit fault signals of phase currents and zero sequence signals of transmission lines.The HHT transform is performed to obtain the marginal spectrum including frequency and amplitude fault information,and the characteristic energy function values of the marginal spectrum are extracted to form input neural network training and test data sets.For the transmission line fault identification problem,because the classification effect of a pure probability neural network PNN is affected by the smoothing factor ?(smoothing parameter),a genetic algorithm with strong optimization ability is used to optimize the smoothing factor ? of the PNN,and the fault feature is finally determined.The extracted feature data set is input into a neural network that has been trained with a smoothing factor ? to perform testing to complete identification of faults in transmission lines.Using short-circuit fault data obtained from simulation models and real fault recordings,The final input was then decomposed and transformed to the PNN fault classifier for fault classification and identification.The results were compared with the classification results of different algorithms.The results show that this method can effectively identify transmission line faults.In addition,the nonlinear fault signals of transmission lines can be processed well.Classification of faults based on Bayesian decision with minimum classification error rate can ensure the accuracy of fault classification,and it also has good denoising and faults for certain interference noise.