Event Detection and Characterization in Continuous Recording of Synchro-Waveforms: Field Experiments and Data Analytics

This article presents a novel practical and theoretical foundation for detecting and characterizing events in continuously recorded, time-synchronized voltage waveform measurements, a.k.a., synchro-waveforms. The proposed methods leverage an in-depth analysis of two waveform concepts through formalized formulations: per-cycle waveform distortions and differential waveforms based on varying cycle-shift parameters. Their mathematical and empirical relationships are established, leading to algorithms that effectively identify and characterize subtle disturbances in waveform data. The methods detect and classify diverse events, including local and nonlocal, subcycle and multicycle, single-phase and three-phase, oscillatory and nonoscillatory, and rare as well as recurring events, and periodic disturbances. Extensive case studies, based on nearly a terabyte of real-world synchro-waveform data, provide insightful observations for each scenario.