How Decentral Smart Grid Control limits non-Gaussian power grid frequency fluctuations

TL;DR

This paper shows that power grid frequency fluctuations are inherently non-Gaussian, and proposes a decentralized control scheme that effectively reduces these fluctuations, enhancing grid stability without centralized data collection.

Contribution

It demonstrates the non-Gaussian nature of frequency fluctuations and derives a scaling law for decentralized control to mitigate these fluctuations.

Findings

Non-Gaussian frequency fluctuations are more likely than Gaussian models suggest.

Decentralized control reduces the magnitude of frequency excursions.

Theoretical predictions are validated using a test grid.

Abstract

Frequency fluctuations in power grids, caused by unpredictable renewable energy sources, consumer behavior and trading, need to be balanced to ensure stable grid operation. Standard smart grid solutions to mitigate large frequency excursions are based on centrally collecting data and give rise to security and privacy concerns. Furthermore, control of fluctuations is often tested by employing Gaussian perturbations. Here, we demonstrate that power grid frequency fluctuations are in general non-Gaussian, implying that large excursions are more likely than expected based on Gaussian modeling. We consider real power grid frequency measurements from Continental Europe and compare them to stochastic models and predictions based on Fokker-Planck equations. Furthermore, we review a decentral smart grid control scheme to limit these fluctuations. In particular, we derive a scaling law of how decentralized control actions reduce the magnitude of frequency fluctuations and demonstrate the power of these theoretical predictions using a test grid. Overall, we find that decentral smart grid control may reduce grid frequency excursions due to both Gaussian and non-Gaussian power fluctuations and thus offers an alternative pathway for mitigating fluctuation-induced risks.