Identification of Railway Power Supply Voltage Sag Source Using Optical Fiber Voltage Sensor

Abstract

With the construction of the smart grid and the continuous improvement of voltage levels, the traditional voltage sensor can no longer meet the development needs of the modern power system. There is an urgent need for a new high-quality voltage sensor to replace it. Aiming at the problems of complex structure, difficult adjustment, poor temperature stability, large optical power loss and inconvenient voltage introduction of the existing optical fiber voltage sensor, an optical fiber voltage sensor based on full polarization state detection without polarizer and analyzer is proposed. The sensor only contains three main components: Grin Lens, BGO crystal and total reflection mirror. It has few devices, simple structure and easy alignment. Moreover, the polarization state evolution and high voltage on the transmission line of Pockels effect of BGO crystal are analyzed. The voltage sag simulation model of the electrified railway is established. The sample data of different disturbance sources are obtained with the assistance of the designed optical fiber voltage sensor. S-transformation and Radial Basis Function (RBF) neural networks are introduced and combined with the prototype of electrified railway power quality detection and analysis device to identify voltage sag sources. The designed optical fiber voltage sensor is suspended. There is no electrode or grounding on the sensor, which saves expensive insulators and increases the sensor’s measuring range. The sensor has a good linear relationship in the power frequency AC voltage range of 0~10 kV at room temperature. It is proved that the design is effective. The working condition analysis of the actual railway power supply reveals that it can detect the voltage sag amplitude and disturbance time, and identify the type of voltage sag disturbance source.