Contingency-Aware Reliability of Microgrids via Neural Lyapunov based Stability Screening

Microgrids (MGs) that are dominated by converters and have low inertia may meet steady-state adequacy but can fail when faced with credible disturbances. Therefore, system reliability assessment must include not just adequacy but also dynamic ride-through capability. An integrated adequacy-stability assessment is introduced in this paper, which combines steadystate dispatch with a rapid, certificate-based transient screen. The operating durations, which are adequacy-success and include a contingency, are assessed for transient stability. An anglepreserving linearized AC model is used to establish the operating point, and a Neural-Lyapunov (NL) certificate is employed to assess the region of attraction and verify stability along faulton trajectories under contingencies. The WSCC system and a modified IEEE 33-bus network serve as demonstration cases for the proposed method. The results show that the power system can experience transient instability during the contingency and adequacy-success periods. This results in a measurable change in composite reliability indices and reveals the importance of incorporating stability screening during reliability assessment in low-inertia, converter-rich MGs.