New method uses SVD to cluster IBR operating regions for faster oscillation analysis

Sub-synchronous oscillations (SSO) in grids dominated by inverter-based resources (IBRs) are typically analyzed using frequency-scan estimates from black-box models at selected operating points. However, IBRs can operate across a wide range of conditions, and frequency responses from a few points may not capture the full dynamic behavior needed for system-level stability analysis. This heterogeneity makes accurate parameterization challenging, as dynamics can change discontinuously or non-smoothly across the operating space. The paper analytically characterizes these conditions and proposes a geometric representation via singular value decomposition (SVD) to identify coherent operating regions—zones where the IBR dynamics behave consistently.

Within each coherent region, the operating-point dependence of the IBR frequency response is captured using simple linear regression, avoiding the need for repeated full-system frequency scans. The framework was validated on a modified IEEE 39-bus test system. Results show that the parameterized frequency responses accurately reconstruct system-level dynamics at the prevailing operating condition, enabling both frequency-response and modal analysis without rescanning. This approach significantly reduces computational overhead for grid operators studying SSO stability in high-IBR grids.