Development of a mathematical model for the diagnosis of short circuits in distribution networks based on a multilayer neural network

Abstract

Relevance: this article is based on the need to improve the reliability and efficiency of short circuit diagnostics in distribution networks, especially in the context of the active digital transformation of the Uzbek electric power industry. Traditional methods of damage detection are not effective enough in networks with a complex structure. The implementation of neural network models and digital logic allows real-time determination of the type and location of short circuits, and shortening the duration and time of interruptions. Given the large-scale modernization of power lines and substations in the country, the development of intelligent diagnostic algorithms is of practical importance and opens up opportunities to increase the sustainability of energy supply.

Aim: to develop a mathematical model based on a neural network for the automated determination of the type and location of short circuits.

Methods: symmetric component method, multilayer neural network (MLP), Boolean logic, simulation modeling in MATLAB, microprocessor programming are used.

Results: based on a neural network, a mathematical model was proposed to determine the type and location of a short circuit. The model was successfully implemented in a microcontroller, tested in the MATLAB environment and showed high accuracy in classifying 12 types of damage with the ability to transmit data over GSM/GPRS.