SSID-Kalman Telemetry for Dynamic Grid State Early Warning
Abstract
This capstone paper constitutes the fourth and final publication establishing the 5-Factor Decision Framework. Having isolated the physical drivers of voltage instability across the LV, MV, and Transmission tiers, this study shifts from forensic simulation to operational prevention. By integrating Subspace State-Space System Identification (SSID) with Kalman filtering, we propose a model-free telemetry pipeline capable of predicting network violations hours before physical onset.
Methodology
Content currently under peer review.
- Algorithmic Filtering: Application of Kalman filters to noisy, real-time SCADA and smart-meter data streams.
- State-Space Identification (SSID): Using model-free empirical estimation to reconstruct localized grid impedances dynamically, bypassing the need for perfect GIS topologies.
- Lead-Time Evaluation: Benchmarking Warning Lead Times (WLT) against False Alarm Rates (FAR) across distinct grid architectures.
Expected Outcomes
This final study establishes Factor 5 (Dynamic Telemetry) of the framework. It proves that grid operators can achieve reliable early warnings—ranging from 58.5 minutes on unpredictable LV solar feeders to over 249 minutes on transmission-level data centre corridors—entirely without massive batch power-flow simulations, ensuring sub-1% false positive rates.