Power System Transmission Line Fault Classification Based On Wavelet Analysis And Support Vector Machines

As one of the mainly support industries, the electric power system is developing faster and faster with the rapid development of Chinese economy and the construction of the infrastructure. The scale of electric power system is enlarging. The transmission power and the voltage grade of transmission lines become higher than before. The degree of the transmission network’s complexity is increasing. When facing problems, rapid fault judging and precise fault locating are needed for the electric power system. It is necessary to restore the faults in time to avoid the enlargement of the range of the faults, ensuring the safety and stability of the system.The paper focuses on common electric transmission line faults research and makes a deep analysis on the comparison of different categories of faults in both China and foreign countries. On account of these, one way of classifying faults based on wavelet analysis and support vector machine is developed and its validity has been proved by simulation experiment and empirical analysis.Large amounts of method work have been done for the research. A model based on the combination of wavelet energy entropy and support vector machine for faults analysis is presented. Four-dimensional eigenvectors include the extract and analysis of the information of the faults with wavelet analysis, the decomposition for the fault of transmission signals with the theory of wavelet energy entropy, the calculation of energy entropy of different currents and the construction of the category for the faults combined with the zero-sequence current. This model includes linear classification module and nonlinear classification module. The linear classification module operates preliminary classification with using the setting of the threshold parameter of zero sequence current and the two layers of classification for the data sample. Then select the support vector machine to operate nonlinear classification according to the characteristics of the faults of the small sample of the transmission line.For simulation and empirical validation, the use of MATLAB simulation model proves the effectiveness of the measure mentioned on classifying the lines faults. Influences on the results of the classification of the parameter setting for the model are analyzed in this paper. An example for an actual power system transmission line faults data from Linyi, Shandong is used for proving the practicability of the measure mentioned in this paper.