Beyond Phasors: Synchro-Waveforms and Implicit Neural Representations for Power System Dynamics

Hamed Mohsenian-Rad – UC Riverside

Recent advancements in power system measurement technologies are enabling a shift in grid analytics from phasor-based representations to more authentic and granular time-domain waveform data. When combined with precise time synchronization across multiple locations, these measurements (referred to as synchro-waveforms) provide access to time-aligned, system-wide voltage and current signals at high sampling rates, capturing fast transients and dynamic behaviors that are not visible in traditional approaches, particularly in systems with increasing penetration of inverter-based resources. While these developments significantly increase data availability, they also introduce new challenges in how waveform data are represented, stored, and analyzed. This talk highlights emerging approaches to waveform-level analytics, with a focus on implicit neural representations (INRs) as a continuous and compact modeling framework for voltage and current waveforms. The presentation illustrates how INR-based models can capture both periodic structures and transient dynamics while enabling efficient representation of multi-phase and time-synchronized measurements. The talk also provides a perspective on how foundation models can be considered for learning fast power system dynamics from waveform data.