Creation of a learning microprocessor system for protection of contact network feeders using adaptive parametric identification methods

Abstract

Handling of high-weight trains, double trains, packet traffic schedule, recuperation on electric locomotives often leads to false protection relays, downtime of trains and increased wear of power switching and protective equipment of traction substations. The paper proposes methods to improve the selectivity of microprocessor relay protection of feeders of AC traction substations contact network by increasing the accuracy of extraction of the first current and voltage harmonic based on phase locked circuits, integration of learning subroutines by fixing the electric parameters and creating the sets of dynamic setpoints in accordance with the current train situation and adaptive parametric identification of dynamic assessment of current electric parameters of the traction network. It also analyzes structural diagrams of hardware and software of the proposed microprocessor protection system of contact network feeders.