Iterative Feedback Tuning of an LPV Feedforward Controller for Wind Turbine Load Alleviation

Abstract

Modern wind turbines incorporate active control techniques such as Individual Pitch Control (IPC) to reduce lifetime dynamic loads. However, system identification and controller design are typically difficult for turbine load control on account of the LPV nature of the system and disturbance. This challenge is addressed by extending the data-based Iterative Feedback Tuning (IFT) technique to systems with LPV output matrices, controlled with a parameterised feedforward controller LPV in output matrices. In order to compensate for the parameter-varying nature of the plant, controller and disturbance, and to estimate the IFT cost gradients in an unbiased manner, an increased number of experiments is required. Such an LPV controller, tuned and tested in a high-fidelity simulation environment, is able to show enhanced load reductions as compared to an LTI controller.