Research On Fault Location Of UHVDC Transmission Line

There is a reverse distribution of energy concentration and power load demand in China,which is far away.It is necessary to build UHVDC transmission project to achieve high efficiency,large capacity and long distance transmission of electrical energy.The DC line is the central park of electric energy transmission.It is easy to fail because of the long line and complex environment along the corridor,and the fault of DC line is very large proportion in the fault of DC system.Therefore,the accurate and reliable fault location technology of UHVDC transmission line is of great practical significance for the safe and stable operation of the system.The nonlinear relationship between the variational mode energy and the fault distance is analyzed by the relationship between the line fault distance and the natural frequency,and the characteristics of the DC voltage energy concentration frequency band.The variational modal decomposition method can accurately analyze the medium and low frequency band signals,and the energy preservation is high.Therefore,the combination of variational modal energy and least squares support vector regression is used to realize fault location.The least squares support vector regression machine has a strong fitting ability to any nonlinear relation,which it requires small sample data,but the is influenced by the penalty parameters andthe kernel width,so the least squares support direction regression is optimized by the particle swarm optimization,and then fault location is realized.The PSCAD/EMTDC software is used to build a bipolar UHVDC transmission system model,the whole line of different fault location and fault voltage signal transition resistance acquisition,the voltage signal phase transform to extract the line mode voltage components,then the variational mode decomposition method is adopted to parse of common-mode voltage line,get the energy of each mode,put the energy value as particle swarm optimization of least squares support vector regression machine sample data for training and learning,finally realizes the line fault location.The simulation results show that the fault location method based on the combination of the variational modal decomposition method and the particle swarm optimization least squares support vector regression is high precision and the maximum relative error is not more than 0.2%.The results proved the effectiveness of the method.