DETECTION OF SHORT-TERM EVENTS (TRANSIENTS) IN THE ELECTRICAL NETWORK

Abstract

The paper proposes an approach to detecting short-term switching transients in single-phase voltage signals of an electrical network based on fast Fourier transform(FFT) with harmonic grouping in accordance with the IEEE 61000-4-7 standard. The method involves analysing the signal in sliding windows of variable duration (from 1 to 10 periods of the fundamental harmonic), calculating the spectral energy in harmonised groups, and forming an event indicator as the energy increase between neighbouring windows with an adaptive threshold. This approach allows for the effective detection of short switching transitions, which are not always noticeable in the time domain, and the timely identification of electromagnetic compatibility violations. Based on the data obtained using the ANTEZ-2 industrial recorder during three months of monitoring a section of the enterprise's network, 17 fast transients and 7 short-term voltage dips were recorded. Experimental results show that the use of short windows (1–2 periods) provides the best time localisation, increased resolution and higher sensitivity to impulse events, while extending the window leads to averaging of spectral energy and reduced detection accuracy. The proposed method is characterised by low computational complexity, which makes it suitable for implementation in real-time devices for power quality monitoring, transient analysis, and event logging in industrial and power networks. The algorithm can be adapted to multiphase systems, extended by combining it with machine learning methods for automatic classification of fault types, and implemented in modern digital power devices. The results confirm the practical feasibility of using FFT-index in intelligent power quality control systems and demonstrate the potential for further integration of the method into comprehensive analytical platforms for power monitoring and network status management.