Fault Classification And Location Of Transmission Line Based On BP Neural Network Improved By PSO

As the skeleton of the power system, high voltage transmission line can not only transmit the electric energy to a long distance, but also has low loss. However, if the high voltage transmission line has malfunction which is not eliminated in time, it will have a great impact on the power users, the security and reliability of power supply can also not be met. It is very important to determine the fault location for the safe and stable operation of the power grid after classifying the fault types of transmission lines.The method of traveling wave distance measurement, single terminal distance measurement and double terminal location are the main methods used in power system. After analyzing and comparing these methods, the two terminal asynchronous algorithm is chose as the fault location algorithm in this paper for its advantages of less investment and high accuracy of fault location. In this paper, the PSCAD/EMTDC software can be used to obtain the simulation waveforms of different positions, different types and different fault resistances.The analysis of above simulation waveforms can get the obvious changes of current, so wavelet algorithm can be used to extract the fault current as feature vector of fault classification. And then apply the particle swarm optimizing BP neural network for transmission line fault classification. The experimental results show that the optimized BP neural network can improve the accuracy of fault classification and greatly improve the computational speed.The line parameters are changing values in the process of power grid operation, so it's necessary to estimate parameters of transmission line. Therefore, Newton iterative method is adopted to estimate the line parameters, but the convergence of Newton method is slow, and the improper initial value can lead to non convergence. So the Newton iteration method is improved, and the results calculated by the least square method are used as the initial value of the improved Newton iteration method to determine the fault location. The simulation results show that the improved Newton iterative method is more accurate.