Time-domain modeling of an equivalent aerodynamic model for wind turbines with six degrees of freedom and machine learning-based parameter prediction

For the wind turbines, the oscillatory motion of the rotor–nacelle assembly (RNA) leads to periodic fluctuations in aerodynamic thrust. These unsteady aerodynamic forces can be reformulated as frequency-dependent aerodynamic inertia and damping effects. In this study, a systematic framework is developed to evaluate the equivalent aerodynamic properties of wind turbines, explicitly incorporating six-degree-of-freedom (6-DOF) coupling and their frequency-dependent characteristics. First, within the generator torque control (GTR) and blade pitch regulation (BPR) regions, small perturbations are introduced and a first-order Taylor expansion is employed to linearize the relationship between aerodynamic loads and variations in inflow wind speed, rotor speed, and blade pitch angle. This procedure yields explicit analytical expressions for the equivalent aerodynamic mass and damping coefficients. Subsequently, a state-space-based time-domain model is established to represent aerodynamic inertia and damping effects, thereby improving the simulation fidelity of wind turbine dynamic responses under varying environmental conditions. In parallel, a machine learning–based surrogate model is developed to predict equivalent aerodynamic mass and damping under the operating conditions. Finally, a decoupled aeroelastic dynamic model is constructed based on the time-domain aerodynamic damping formulation. The predicted structural responses show close agreement with results obtained from fully coupled wind-wave-earthquake simulations.