Linear uncertain modelling of LIDAR systems for robust wind turbine control design.

Light detection and ranging (LIDAR) sensors measure the free wind ahead of the rotor, enabling the use of new feedforward control strategies. However, there exist some sources of error inherent to the measuring process that should be considered during the design of LIDAR-based controllers. Typically, the coherence function is used for that purpose, but it is not compatible with some robust design methodologies. This paper presents an analytic relation between the coherence function and a non-parametric uncertainty model of LIDAR sensors, suitable for the design of controllers via μ -synthesis or Quantitative Feedback Theory. Such a relation is applied to a realistic LIDAR simulator. First, the linear non-parametric uncertainty model is identified using simulation data obtained from the well-known NREL 5 MW wind turbine. Then, it is validated against the coherence model by comparing linear predictions of the simulation outputs. • The squared coherence function is the standard measurement of LIDAR reliability. • Squared coherence is inversely proportional to a nonparametric uncertainty model. • The uncertain model is useful for analysis and robust controller design purposes. • The new model is compatible with design tools such as QFT and H ∞ control toolbox. • The new model has been validated against the FAST nonlinear simulator. [ABSTRACT FROM AUTHOR]