Aeroelastic model field validation and performance state estimation of wind turbine active flaps

This PhD thesis summarizes research findings on the field validation of aeroelastic models of active trailing edge flaps for wind turbines and on the estimation of the flap performance state with machine learning.
As wind energy plays a fundamental role in achieving a climate-resilient future, the wind power sector must advance wind turbine technology to cope with the escalating demand. A promising technology is active trailing edge flaps, devices placed at the trailing edge of the wind turbine blades to control the local aerodynamic forces acting on the blades. They have the potential to control and reduce wind turbine loads while enhancing overall performance. As the aeroelastic modeling has a fundamental role in the design of wind turbines, the accuracy and reliability of active flap aeroelastic models are paramount for the design soundness of commercial wind turbines equipped with active flaps. The active flap aeroelastic models have been only limitedly validated, mainly due to the scarcity of full-scale experimental data. Therefore, more validation is needed to fully integrate active flaps in the design of commercial wind turbines. This integration also requires the development of systems to detect, monitor, and quantify any potential fault or performance degradation of the flap system to avoid jeopardizing the wind turbine’s safety and performance.

The aims of this industrial PhD thesis are the validation of the aeroelastic models of active trailing edge flap with full-scale field measurements and the development of an application to estimate the performance state of the active flap system. To achieve them, two research questions are investigated: 1) Do the aeroelastic models of wind turbines equipped with active trailing edge flaps estimate accurately and reliably the loads and behavior of an actual commercial-scale wind turbine? and 2) Can a machine learning model estimate the flap performance states of a wind turbine to improve the reliability of the active flap system design? 

Regarding the first question, a major outcome of this industrial PhD thesis is the validation of the flap models implemented in the two cutting-edge aeroelastic codes BHawC and HAWC2 with full-scale measurement data from a commercial wind turbine (4.3 MW rated power, 120 m diameter, 20 m flap length). The validation shows that the flap models accurately and reliably model the impact of the flap actuation on the wind turbine’s aerodynamic, loading, and operational parameters. The study also highlights that modeling and tuning the flap actuator and the flap aerodynamic properties are crucial elements for correctly modeling active flaps in wind turbine aeroelastic codes. Finally, the study suggests that the capability of the active flap model to estimate the unsteady effect on the aerodynamic forces due to the flap motion would become relevant at high flap actuation frequencies.

Regarding the second question, an application has been developed for the first time to estimate the flap performance states for asymmetric faults. The application has remarkable precision, higher than 90%, and is based on a simple and well proven machine learning technique. Furthermore, it requires only input signals commonly available on commercial wind turbines, simplifying its possible implementation in future wind turbines.
The findings of this PhD research project have significantly contributed to increasing the accuracy and reliability of the active flap aeroelastic tools and the capability of detecting active flap system faults. These contributions are essential milestones to enable the safe and reliable design of future wind turbines equipped with active flaps.