An Interaction Sensitivity Framework in Perturbation Space for the Converter-Driven Stability of Power Systems

Abstract

As the integration of renewables accelerates the decarbonization process in the energy sector, power systems are becoming increasingly dominated by converter-interfaced generation. However, this results in the emergence of multiple oscillations, which can trigger cascading disconnections and potentially damage generators. To address these critical issues, modern power systems commonly employ classic participation factor (PF) analyses, which relate oscillatory modes (eigenvalues) to controller state variables. While these methods are effective, a critical yet previously unexplored challenge involves capturing the interaction reconfigurations observed after a perturbation occurs. Because PF analyses are limited to fixed operating points, they are structurally unable to analyze these event-dependent mode--state interactions under perturbations. Here, we establish an interaction sensitivity framework that is formulated in perturbation space, uncovering the mechanisms by which modes and states respond to perturbations. This framework provides analytic expressions and physically interpretable mechanisms that explain why specific control adjustments enhance or reduce stability, revealing causal relationships beyond conventional stability analysis. More broadly, this formulation provides a general analytical basis for understanding perturbation-driven interaction reconfigurations in high-dimensional dynamical networks described by state-space equations, and supports advanced controller design in complex systems.