Power systems are increasingly dominated by power-electronic converters, and this shift has brought a category of stability problems that conventional tools were not designed to handle: sub-synchronous oscillations and weak-grid interactions driven by the coupling between convert
...
Power systems are increasingly dominated by power-electronic converters, and this shift has brought a category of stability problems that conventional tools were not designed to handle: sub-synchronous oscillations and weak-grid interactions driven by the coupling between converters and the surrounding network. Real-time simulators such as RTDS/RSCAD support impedance scanning, but their built-in tools offer little beyond a frequency sweep. External analysis frameworks go further analytically, but cannot run natively inside a real-time environment. Neither approach assesses how stability changes as grid conditions and control settings vary, which is a significant gap when the operating space of an offshore wind farm spans a wide range of short-circuit ratios and controller configurations.
This thesis addresses that gap with an RSCAD-native, Python-driven pipeline that handles admittance extraction, MIMO stability assessment, and parameter-dependent robustness evaluation within a single automated workflow. Admittance matrices at one or more points of common coupling are reconstructed in the dq-domain using a two-run D/Q injection procedure. Before any frequency point enters the analysis, it is screened by a four-pillar verification framework that checks numerical conditioning, spectral purity, stationarity, and signal-to-noise ratio. Stability is then assessed by applying the Generalised Nyquist Criterion and eigenvalue decomposition of the closed-loop impedance together, yielding both a pass/fail verdict and a spatial-spectral picture of the dominant interaction.
Two diagnostic extensions build on this baseline. Parametric Stability Space Mapping (PSSM) re-runs the full RTDS scan across combinations of short-circuit ratio, X/R, and control-gain values, mapping how stability evolves across operating conditions rather than at a single point. The Robust Stability Margin Metric (RSMM) converts the binary Nyquist verdict into a continuous scalar, which a mitigation workflow then maximises over a detected critical frequency band to identify the controller setting that most improves the worst-case margin.
The pipeline is first verified against the RSCAD Frequency Scan Analysis Tool on a passive benchmark. It is then applied to two MMC-HVDC offshore wind-park models. Both are classified as stable at the nominal operating point, but the extended diagnostics reveal substantially different robustness profiles, dominant participating PCCs, and critical frequency bands which are located in the sub-synchronous range. The PSSM-RSMM results reproduce the expected coupling between grid strength and controller tuning and point to concrete, model-specific mitigation actions: PLL gain retuning for the multi-terminal HVDC model and inner current-loop gain reduction for the single-converter model.
Taken together, the results show that the proposed workflow is numerically reliable, scales naturally from single- to multi-PCC configurations, and produces diagnostics such as the critical frequency, dominant PCC, robustness margin, and recommended controller setting that are directly actionable, rather than stopping at an isolated stability verdict.
