Single Pole-To-Earth Fault Detection and Location on the Tehran Railway System Using ICA and PSO Trained Neural Network

TL;DR

This paper presents a novel method combining ICA, PSO, and neural networks to accurately detect and locate single pole-to-earth faults in Tehran's railway power system, improving safety and reliability.

Contribution

It introduces an integrated approach using ICA and PSO to train neural networks for fault detection and location in railway systems, enhancing accuracy over traditional methods.

Findings

Fault location predicted with high accuracy

600Hz harmonic ripple effectively used for detection

Simulation results confirm method's effectiveness

Abstract

In a railroad feeding system, detecting a location of pole to earth faults is important for safe operation of the system. The goal of this paper is to use a combination of the evolutionary algorithm and neural networks to increase the accuracy of single pole-to-earth fault detection and location on Tehran railroad power supply system. Accordingly, Imperialist Competitive Algorithm (ICA) and Particle Swarm Optimization (PSO) are used to train the neural network for enhancing learning process accuracy and the convergence. Owing to the nonlinearity of system, the fault detection is an ideal application for the proposed method where 600 Hz harmonic ripple method is used in this paper for fault detection. The substations were simulated by considering various situations in feeding the circuit, the transformer and the silicon rectifier has been developed by typical Tehran metro parameters. Required data for the network learning the process have been gathered from simulation results. 600Hz components value will change with the change of the location of single pole to earth fault. Therefore, 600Hz components are used as inputs of the neural network when fault location is the output of the network system. The simulation results show that the fault location can be accurately predicted in proposed methods.