Fault Location in HVDC Systems Using Multilayer Perceptron and Wavelet Packet Transform

A Sustainable Approach for Renewable Energy Integration

Abstract

Objective: The study aims to develop and evaluate a model based on neural networks, specifically the Multilayer Perceptron (MLP) combined with the Wavelet Packet Transform (WPT), to improve fault location in High-Voltage Direct Current (HVDC) transmission systems, contributing to environmental sustainability by enhancing the efficiency and reliability of these systems.

Methodology: The research uses current data measured from the HVDC line as inputs to train the MLP model, evaluating its performance through metrics such as Mean Absolute Error (MAE) and Mean Squared Error (MSE). The study simulates different operational conditions to test the model's effectiveness.

Originality/relevance: This work addresses a theoretical gap concerning the need for more efficient and accurate techniques for fault location in HVDC systems. It highlights both the academic and practical relevance of solutions that minimize response time and maximize energy efficiency, thereby contributing to the reduction of environmental impact.

Results: The proposed model exhibits high accuracy in fault location, significantly reducing estimation errors and, consequently, transmission line downtime.

Theoretical/methodological contributions: The study contributes with an innovative approach combining MLP and WPT, offering a robust method for fault location in HVDC systems, with potential applications in other energy transmission systems.

Social and environmental contributions: The research highlights the environmental benefits of HVDC lines, including energy efficiency and reduced environmental impact, by proposing a solution that minimizes downtime and enhances system reliability, thereby promoting sustainability in power transmission.