Model Predictive Wind Turbine Control with Move-Blocking Strategy for Load Alleviation and Power Leveling

The Science of Making Torque from Wind

Jassmann, U.^1 ; Dickler, S.^1 ; Zierath, J.^2 ; Hakenberg, M.^1 ; Abel, D.^1

Institute of Automatic Control, Center for Wind Power Drives, RWTH Aachen University, Aachen

52056, Germany^1

W2E Wind to Energy GmbH, Strandstrasse 96, Rostock

18055, Germany^2

关键词: Computational burden;    Electrical output;    Load alleviation;    Mechanical loads;    Model predictive controllers;    Multibody simulations;    Turbine operation;    Wind turbine control;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/753/5/052021/pdf
DOI  :  10.1088/1742-6596/753/5/052021

来源: IOP

【 摘 要 】

This contribution presents a Model Predictive Controller (MPC) with moveblocking strategy for combined power leveling and load alleviation in wind turbine operation with a focus on extreme loads. The controller is designed for a 3 MW wind turbine developed by W2E Wind to Energy GmbH and compared to a baseline controller, using a classic control scheme, which currently operates the wind turbine. All simulations are carried out with a detailed multibody simulation turbine model implemented in Alaska/Wind. The performance of the two different controllers is compared using a 50-year Extreme Operation Gust event, since it is one of the main design drivers for the wind turbine considered in this work. The implemented MPC is able to level electrical output power and reduce mechanical loads at the same time. Without de-rating the achieved control results, a move-blocking strategy is utilized and allowed to reduce the computational burden of the MPC by more than 50% compared to a baseline MPC implementation. This even allows to run the MPC on a state of the art Programmable Logic Controller.