Application of Neural Network in Restoring Metro ESP Algorithm

The distribution of DC surface potential (DCESP) is the key to the DC bias of power system transformers, and the stray current of rail transit traction power supply will affect the distribution of (DCESP). The distribution of DC surface potential (DCESP) is the key to the DC bias of power system transformers, and the stray current of rail transit traction power supply will affect the distribution of (DCESP). The distribution of DC surface potential is affected by the magnitude and location of the DC current into the ground, as well as soil structure and parameters. The topography between the two existing substations is complex, especially in cities, which are affected by factors such as highways and buildings. The number of locations for measuring ground potential data is limited and discontinuous. Based on electromagnetic theory, this paper derives the DCESP distribution calculation formula for the multilayer vertical geoelectric structure, and builds an accurate subway system model in CDEGS, and simulates the DCESP distribution data around the subway. Design a neural network to restore DCESP, thereby reducing the workload of ESP measurement and calculation. The research results of this paper provide a new idea for DC surface potential measurement, which can promote the development of DC bias field.