Robust Control of Hydraulic Pitch System
Authors
Pedersen, Casper Strøm ; Fink, Magnus
Term
4. term
Education
Publication year
2024
Submitted on
2024-05-31
Pages
108
Abstract
Vindmøller er en central del af overgangen til vedvarende energi. Men vedligehold og driftstop bidrager væsentligt til energiprisen, og undersøgelser peger på, at pitchsystemet (der justerer vingernes vinkel) står for flest fejl og mest nedetid. Dette projekt undersøger, hvordan en intern lækagefejl i pitchsystemet påvirker driften, da denne fejltype er vurderet til at have den højeste risiko. En indført lækage forringer den almindelige (nominelle) positionsregulators ydeevne. Derfor udvikles en fejl-tolerant styringsstrategi, der er robust over for interne lækager. Designet omfatter en Unscented Kalman Filter (UKF), en beregningsmetode der estimerer skjulte størrelser ud fra målinger, til at bestemme både den interne lækage og lastmomentet (den modstand, der påvirker bevægelsen). Lækageestimatet bruges derefter i en Active Flow Feedforward-algoritme med lækagekompensation, som forudkommanderer den nødvendige tilførsel for at modvirke lækagen. Projektet viser, at UKF’en kan estimere intern lækage og lastmoment præcist og er robust over for usikkerheder i modellens parametre. Når UKF-lækageestimatet føres tilbage i den aktive feedforward, opnås bedre styreperformance end med den nominelle positionsregulator, der er designet til et fejlfrit pitchsystem.
Wind turbines are a key part of the shift to renewable energy. However, maintenance and downtime add significantly to the cost of energy, and studies indicate that the pitch system (which adjusts the blade angle) accounts for the most failures and downtime. This project examines how an internal leakage fault in the pitch system affects operation, since this fault type has the highest assessed risk. Introducing leakage degrades the performance of the standard (nominal) pitch position controller. To address this, a fault-tolerant control strategy is developed to remain robust under internal leakage. The design uses an Unscented Kalman Filter (UKF)—a computational method that infers hidden quantities from measurements—to estimate both the internal leakage and the load torque (the resisting force acting on motion). The leakage estimate is then used in an Active Flow Feedforward algorithm with leakage compensation, which pre-commands the needed input to counter the leakage. The project shows that the UKF can accurately estimate internal leakage and load torque and is robust to uncertainties in model parameters. Feeding the UKF leakage estimate into the active feedforward yields better controller performance than the nominal position controller designed for a healthy pitch system.
[This summary has been rewritten with the help of AI based on the project's original abstract]
Keywords
Documents