Spectrogram-Based Joint Detection, Localization, and Classification of Events in Continuously Recorded IBR Waveforms

2026-07-22

AI Summary

This study addresses the challenge of automatically identifying dynamic power system events in continuous waveforms from inverter-based resources (IBRs) by proposing a spectrogram-based temporal object detection method. The approach constructs a multi-channel spectrogram tensor via short-time Fourier transform and unifies event detection, localization, and classification into a joint task performed directly on the spectrogram, explicitly capturing transient and harmonic characteristics. To the best of our knowledge, this is the first work to employ spectrograms as the input representation for IBR waveform event analysis. Evaluated on both single-phase disturbance and three-phase fault datasets, the method significantly outperforms baseline approaches that operate directly on raw waveforms, demonstrating superior performance in detection accuracy, temporal localization precision, and classification capability.

Abstract

Continuously recorded high-resolution waveform measurements provide rich information about fast power system dynamics. However, they require automated methods to identify events. This problem is addressed by developing a spectrogram-based framework to jointly detect, localize, and classify events in real-world continuously recorded waveforms at the terminal of an Inverter-Based Resource. We recast this problem as a temporal object detection problem on spectrogram images, as they capture the transient and harmonic signatures more explicitly than in raw waveform data. Each time-series waveform is transformed using the short-time Fourier transform, and the resulting per-channel spectrograms are stacked as a tensor for event detection. We benchmark this method against a detector operating directly on raw time-series measurements. Experiments on single-phase disturbances and three-phase faults demonstrate that the proposed spectrogram method consistently improves event detection, localization, and classification over the raw waveform baseline.

Problem

• event detection
• event localization
• event classification
• spectrogram
• inverter-based resource

Innovation

• spectrogram
• temporal object detection
• Inverter-Based Resource (IBR)
• short-time Fourier transform
• event classification