Optimal Control for Fatigue Reduction of a Ballast-Stabilized Floating Wind Turbine

Abbate, Giuseppe Battista

4. term

Control and Automation, Master

2013

2013-01-07

34

Den voksende efterspørgsel efter vedvarende energi driver udviklingen af vindteknologi, og moderne vindmøller bliver både større og mere kraftige. Traditionelt er møller placeret på land eller i lavt vand på faste fundamenter, men det kan give uønsket visuel påvirkning og støj for nærområdet. Længere fra kysten i dybere vand er vindforholdene også mere stabile, fordi der er færre forhindringer, der skaber turbulens. Den økonomisk mest fordelagtige løsning i dybt vand er at bruge flydende platforme. Erfaringer fra olie- og gasindustrien er her en vigtig baggrund. Flydende platforme bevæger sig dog med bølger og vind, hvilket stiller større krav til møllens styringssystem (kontrolsystem) for at holde konstruktionen stabil og begrænse belastninger. Det er vist, at den almindelige onshore "baseline"-regulator (en PI-bladpitch-regulator kombineret med en variabel hastigheds-generator-momentregulator) kan give negativ dæmpning på en flydende offshore-mølle – altså at reguleringen utilsigtet kan forstærke bevægelser. Rapportens første del beskriver indstilling (tuning) af baseline-regulatoren og udviklingen af en forenklet, styringsorienteret model af en flydende offshore-vindmølle på en ballast-stabiliseret platform. Modellen opstilles ud fra de vigtigste kræfter i systemet, parametre identificeres, og modellen valideres mod FAST, en detaljeret og frit tilgængelig vindturbinesimulator. Anden del præsenterer tre styringsmetoder baseret på både klassisk og avanceret regulering: en baseline PI-regulator (proportional-integral) og tre LQ-regulatorer (linear-quadratic) med forskellige mål. Formålet er at reducere træthed i tårnet – materialeskader, der opbygges over tid ved gentagne belastninger – ved at minimere enten variansen af tårnets afbøjning eller variansen af afbøjningshastigheden. Simulationer viser, at LQ-regulatorer, der reducerer tårnets afbøjningshastighed, kan mindske træthed ved tårnfoden i forhold til baseline PI. Afslutningsvis sammenlignes de tre LQ-regulatorer med baseline PI ved hjælp af skadeækvivalente laster og statistisk analyse. Hver regulator afprøves i FAST-simuleringer for at vurdere, om styringsloven faktisk reducerer træthed og dermed kan retfærdiggøre brugen af avanceret regulering på flydende vindmøller.

Growing demand for renewable energy is pushing wind technology forward, and modern turbines are becoming larger and more powerful. Traditionally, turbines are installed on land or in shallow water on fixed foundations, but this can cause unwanted visual impact and noise for nearby communities. Farther offshore in deeper water, visual and noise concerns are smaller, and winds are steadier because there are fewer obstacles that create turbulence. The most economically favorable way to install turbines in deep water is on floating platforms, drawing on experience from the oil and gas industry. However, floating platforms move with waves and wind, which makes the turbine’s control system critical for stability and for limiting structural loads. It has been shown that applying the standard onshore baseline controller (a PI blade pitch controller together with a variable-speed generator torque controller) to a floating turbine can introduce negative damping—control actions that inadvertently amplify platform motion. The first part of this thesis describes how the baseline controller is tuned and develops a simplified, control-oriented model of a floating offshore wind turbine on a ballast-stabilized platform. The model is derived from the main forces acting on the system, parameters are identified, and the model is validated against FAST, a detailed wind turbine simulator that is freely available. The second part presents three control approaches based on classic and advanced control theory: a baseline PI (proportional-integral) controller and three LQ (linear-quadratic) controllers with different objectives. The goal is to reduce tower fatigue—damage that accumulates under repeated loads—by minimizing either the variance of tower deflection or the variance of deflection velocity. Simulations indicate that LQ controllers designed to reduce deflection velocity can lower fatigue at the tower base compared with the baseline PI controller. Finally, the thesis compares the three LQ controllers with the baseline PI using Damage Equivalent Loads and statistical analysis. Each controller is tested in FAST simulations to assess whether the control law actually reduces fatigue and thus justifies the use of advanced control on floating wind turbines.

[This abstract was generated with the help of AI]

hywind ; fatigue ; control ; floating wind turbine ; LQ ; linear quadratic

Download
View record in AAU Student Projects