The wind energy market is growing rapidly and the environmental awareness even more. High wind turbine performance is a very essential factor and is directly linked to global sustainability goals. This is the reason why more cost-effective and reliable operation of wind turbines is needed. This can be achieved by processing the supervisory control and data acquisition (SCADA) data using advanced data processing techniques. That way, the evaluation of the condition of each wind turbine subsystem can be accomplished, limiting unscheduled service time and allowing predictive maintenance. One of the most important subsystems in wind turbines is the pitch system, which controls the blade angles according to the operation strategy.
The aim of this research is to develop an online fault detection and identification system of the hydraulic pitch system in wind turbines and to analyse the most common failures, both tribological and hydraulics related, which occur at this system. The project consists of two parts. The first part focuses on implementing advanced signal processing techniques including machine learning and other artificial intelligence techniques in order to detect and predict failures in the pitch system. The second part includes the physics based (or hybrid) simulations of failures in order to investigate the root causes of the main failure types. Consequently, this knowledge will provide support for more accurate decisions and actions related to the operation of wind turbines.
Doctoral School of Industry Innovations (DSII)
Suomen Hyötytuuli Oy
Prof. Arto Lehtovaara
Dr. Matti Linjama
Dr. Jaakko Kleemola, Suomen Hyötytuuli Oy