Identification of a discrete-time piecewise affine model of a pitch-controlled wind turbine

In this work a dynamic discrete-time piecewise affine model of a pitch-controlled wind turbine is identified in the vicinity of the optimal wind turbine static characteristics. The aerodynamic rotor torque and thrust force of a 1 MW wind turbine are identified as piecewise affine static functions of wind speed, rotor speed and pitch angle using the clustering-based technique. These static functions are post-processed to rule out irregularities induced by the identification procedure and are then incorporated in a dynamic model of the wind turbine. The resulting model, that consists of 29 affine dynamics, is verified against the starting nonlinear model of the wind turbine and very good agreement with that model is obtained. The identified discrete-time piecewise affine model will be used in synthesis of optimal control laws for improving power production and reducing fatigue loads of the wind turbine.