A MILP optimization approach for optimizing primary frequency control of wind turbines in low-inertia islands

Deloading of wind turbines (WTs), i.e. operating them below their maximum power point, is a common strategy to reserve capacity for enabling WTs participate in the primary frequency control during power imbalances. Traditional approaches typically reserve a fixed power, based on the rated or available power of the WT. However, such static methods may misjudge actual reserve requirements, as they do not account for real-time grid conditions and electro-mechanical dynamics within the WT. To address this limitation, this paper introduces a two-stage approach as follows: first, the optimal droop parameters of WTs are determined through a Laplace-domain inertia frequency response model to ensure network’s frequency stability. Second, a dynamic Mixed Integer Linear Programming (MILP) optimization framework is proposed that determines the minimum necessary reserve level, based on the droop parameters specified in the previous stage, required to ensure WT’s rotor speed stability. Compared to conventional fixed-deloading strategies evaluated under steady-state conditions, assuming a 4 MW wind turbine operating with a constant 20% deloading, the proposed approach yields a steady-state-equivalent annual energy gain of 1.650 MWh (≈€99,000 per turbine).