Ann-Driven Fault Detection Technique in HighVoltage Transmission

Abstract

This research investigates the application of Artificial Neural Networks (ANNs) for fault detection in 
high-voltage transmission lines, with a focus on the Onitsha to Enugu transmission power system in 
Nigeria. The primary objective is to enhance the accuracy and efficiency of fault detection, thereby 
improving the reliability and stability of power systems. The methodology involves modeling the 
transmission system using MATLAB/Simulink, generating pre-fault and fault data, and training the 
ANN using the Levenberg-Marquardt backpropagation algorithm. The ANN’s performance is 
evaluated through various metrics, including error histograms, regression plots, and performance 
validation plots. Key findings indicate that the ANN-based fault detection scheme can accurately 
identify and classify different fault types, such as line-to-ground (L-G), line-to-line (L-L), double lineto-ground (L-L-G), and three-phase faults (L-L-L and L-L-L-G). The results demonstrate the ANN’s 
ability to generalize well to new data and adapt to changing system conditions, making it a robust tool 
for power system protection. The implications of this research are significant, as it offers a viable 
alternative to conventional protection schemes, addressing their limitations and contributing to the 
advancement of intelligent fault detection technologies in power systems. This study highlights the 
potential of ANNs in enhancing power system protection and paves the way for future research to 
explore the integration of other machine learning techniques and real-time implementation in live 
power system environments. The findings are relevant to both experts in the field of power systems 
and general readers, highlighting the transformative impact of artificial intelligence on power system 
engineering.