Computing spatial estimates of the over-the-rotor wind distribution for advanced wind turbine active control

We formulate novel observers of the wind incident on the rotor of a wind turbine. Stochastic filtering processes reconstruct on-line optimal estimates of the tower and blade states, by using readings from accelerometers and strain gages placed along the structural members. From these reconstructed states, another filtering process estimates the wind states which, improving on the commonly used mean hub-height wind estimates or the point measurements available from on-board anemometers, also include information on the wind-over-the-rotor distribution. The proposed procedure is able to approximate vertical and horizontal wind shear, yawed flow and a vertical wind component. The instantaneous estimates of the flexible states of tower and blades and of the spatial wind distribution can be used for enabling sophisticated individual-blade fatigue and load alleviating control laws. The proposed procedures are demonstrated with the help of numerical experiments conducted using a high-fidelity aero-servo-elastic simulator.