DETECTION, LOCALIZATION AND CLEARING OF FAULTS IN THREE PHASE 132KV TRANSMISSION SYSTEM USING OPTIMAL DEEP LEARNING TECHNIQUES

Summary

Faults in three-phase power transmission systems threaten system stability, equipment integrity, and power supply reliability. This study presents an integrated framework for fault detection, localization, and clearing using time-domain voltage and current signals in MATLAB/Simulink. Fault detection is performed using a Long Short-Term Memory (LSTM) autoencoder trained on healthy operating data. The reconstruction loss remained below 50 under normal conditions and increased above 500 during faults, enabling reliable detection within 10–20 ms. Fault localization was achieved through phase-wise analysis of voltage and current signals, accurately identifying line-to-ground (LG), line-to-line (LL), line-to-line-to-ground (LLG), and three-phase (LLL) faults. A logic-based protection scheme combined the detection and localization outputs to generate circuit breaker trip commands for fault isolation. Simulation results show that fault currents increased from below 200 A to approximately 1.5–3.9 kA, depending on the fault type, before rapidly decreasing to near zero after breaker operation. Voltage recovery occurred within 40–60 ms, with minimal post-fault oscillations and stable system operation. The proposed framework provides fast fault detection, accurate fault localization, and effective fault clearing, demonstrating its suitability for intelligent protection of modern three-phase power transmission systems.

Index Terms

Browse additional articles authored by one or more contributors to this paper.