Fault identification in wind turbines: a data-centric machine learning approach

The last few years have been marked by the transition of the world energy matrix, predominantly with wind and solar sources considered clean energies. Wind turbines, responsible for the energy conversion process, are complex and expensive equipment susceptible to several failures due to multiple factors. Monitoring turbine components can assist in detecting failures before they occur, reducing equipment maintenance costs. This work compares machine learning techniques in a data-centric approach to wind turbine failure detection. Preliminary results demonstrate the importance of feature selection in this problem.This work has been supported by FCT - Fundação para a Ciência e a Tecnologia within the RD Units Project Scope Research Center in Digitalization and Intelligent Robotics (CeDRI) UIDB/05757/2020 and UIDP/05757/2020 and SusTEC (LA/P/0007/2021). The authors would like to thank the following Brazilian Agencies CAPES, CNPq, and FAPERJ.