Work

A Conceptual Data Fusion Framework for Oscillation Risk Assessment in Renewable Power Systems

Year: 2026
Type: article
Abstract: Oscillation assessment in renewable power systems must distinguish measurement confidence, source evidence, present condition, and future warning. This paper proposes a conceptual data fusion framework whose methodological contribution is an explicit set of dependency and evidence constraints: prediction is excluded from its current-risk target, propagation coefficients require valid physical connections, and unavailable observations remain distinguishable from zero risk. These constraints make assessment assumptions testable without claiming a new physical source locator. The numerical illustration audits a simplified 500 MVA offshore wind and HVDC benchmark with prescribed disturbances, power-deviation proxies, heuristic state scores, and a dependent four-level hierarchy. It does not implement signed branch-energy attribution, calibrated multiband filtering, estimated Markov transitions, or independent physical device-graph propagation. Equation replay reproduces the archived S01 and S12 outputs within floating-point precision. The imposed 25 Hz WF1 case reaches an internal source score of 0.963 and a legacy system score of 0.563. Controlled changes separate forcing reduction, power reduction, and the receiving-grid term. A 1 s restoration ramp lowers the combined-case full-record maximum from 0.567 to 0.341; removing only the WF1 forcing lowers the late-window legacy score to 0.0630 while residual grid effects persist. A prediction-free retrospective target exposes false-warning and missed-warning tradeoffs; the historical 0.861 s crossing separation is not predictive validation. Executable algebraic checks and a self-contained numerical supplement support verification of the stated scope. Independent physical and statistical validation remains necessary.
Source: Energies
Language: English
Open Access status: gold
Estimated APC paid: $3,137
OpenAlex ID: W7235869198
FWCI:
Cites: 22
Cited by:
Related to: 10