Analysis of Subway Power Supply Compensation System Based on Neural Network Control

The subway provides a high-speed, safe and convenient solution to urban traffic problems and becomes the infrastructure of urban development in our country. However, the electronic devices in the subway power supply system produce a large number of harmonics, which will cause resonance and shorten the service life of the equipment, thus affecting the normal operation of the power system. Aiming at the requirement of reactive power compensation in subway, this paper analyzes the characteristics of magnetron reactor and SVG device, introduces neural network algorithm, designs PI controller based on BP neural network, and applies it to SVG control strategy. Through simulation experiment, the rationality of control scheme is verified. It provides a reference for similar subway projects to carry out reactive power compensation.