Design and Tuning of State Feedback-Based Islanding Detection for Distributed Generators

Islands form in 3-phase distribution networks when a grid-following distributed generation (DG) source disconnects from the main grid, potentially causing safety risks if the DG continues to power local loads. Detecting such unintentional islands promptly is crucial to prevent hazardous situations. This study introduces a generalized framework for rapid islanding detection within a single fundamental cycle, employing state feedback. The methodology derives the feedback gains through dynamic phasor-based state-space modeling of the islanded system. The proposed approach utilizes the measurable system states as feedback to the system's current or power references. The feedback configuration is tuned to induce instability in the islanded network by placing a pole in the right half of the s-plane. The proposed tuning of the feedback gains ensures a zero non-detection zone (NDZ) for a range of system parameters while ensuring compatibility with low-voltage ride-through (LVRT) requirements. The islanded system's modeling accommodates local loads, facilitating island detection in scenarios with passive loads operating at varying power factors, in accordance with IEEE Std. 1547-2018. The versatile generalized framework provides the required feedback gains to suit DGs spanning a wide spectrum of power ratings. The proposed framework's design ensures heightened sensitivity of the detection parameters to the introduced feedback without the need for additional sensors or complex computations. The paper comprehensively delineates the analytical, simulation, and experimental validation employed to validate the efficacy of the proposed islanding detection approach.