An ANN-Wavelet Based Distribution Transformer Protection

Abstract

Transformers are crucial components in electrical power systems, serving as vital links for efficient energy transmission and distribution. With the increasing integration of renewable energy sources and the need for enhanced grid resilience, hybrid transformers have emerged as a promising solution. Combining the benefits of traditional transformers with modern technologies, hybrid transformers offer improved performance, reduced environmental impact, and advanced functionalities. However, this technological advancement also calls for equally advanced protection mechanisms to ensure their reliable and safe operation. In this paper, artificial neural network based approach is used to protect a two-winding transformer in laboratory environment. Feature vectors are differential currents and experimental studies cover energizing, sympathetic inrush currents, and internal faults. Simulation and real times experiments show that the proposed hybrid transformer protection scheme is able to detect normal and faulty conditions. © The Institution of Engineering & Technology 2024.