Data-driven control for the next generation of wind turbines

An ongoing trend in wind turbine development is the increase in rotor size. These large rotors inherently become more flexible and introduce more interaction between blades and tower. Several new technologies have been conceived to reduce loads on these future wind turbines. Among these are individual pitch control (IPC) and the smart rotor, where distributed trailing edge flaps are used to influence the lift locally. To allow effective control of fatigue and extreme loads, new measurement techniques are being developed such as LIDAR and local flow measurements. Such measurement techniques may allow a degree of feed-forward control by anticipating wind disturbances.

In this research programme we investigate data-driven methodologies to design controllers for future wind turbines. These include, for example, novel system identification techniques, online disturbance modelling and efficient model predictive control. The basis for all methods is formed by numerically reliable and robust algorithms from the fields of linear algebra and convex optimisation.

Project members: S.T. Navalkar, MSc (Sachin), J.W. van Wingerden (Jan-Willem)

Learning and adaptive control, Model predictive control, Identification and estimation, Robotics and mechatronics, Wind energy

Sponsored by: