Adaptive Inertia‐Damping Coordination for Sending‐End VSG Control in a Wind‐PV MMC‐HVDC Collection System

ABSTRACT

A DQN‐based adaptive inertia‐damping coordination strategy is proposed for the sending‐end MMC‐VSG in a wind‐PV MMC‐HVDC collection system. A physically interpretable coordination law is first constructed according to frequency deviation and RoCoF, so that virtual inertia and damping can be adjusted at different frequency dynamic stages. The key parameters of the coordination law are discretized into bounded candidate modes, and the DQN selects the suitable parameter mode online, while the final inertia and damping are continuously generated by the coordination law and smoothing process. Comparative electromagnetic transient simulations show that the proposed method reduces the maximum frequency deviation by up to 39.7% and the maximum RoCoF by up to 35.8% compared with fixed J ‐ D control. Under ten random wind‐PV disturbance cases, the proposed method achieves the highest mean and worst‐case frequency nadirs of 49.851 and 49.746 Hz, respectively. Non‐ideal‐condition tests and input‐state drift analysis further verify its robustness under practical uncertainties.