A good prediction of the failure ratio of wind turbine (WT) components is pivotal to define a correct maintenance program and reduce the downtime periods. Even a small failure can lead to long downtime periods and high repairing costs. The installation sites, which generally have limited accessibility, and the necessity of special facilities to reach the components inside the nacelle, also play a key role in the correct management of WTs. In this study, a detailed survey on the failures occurred to the WTs managed by the Italian operator “e2i energie speciali” (more than 550 machines) over 16 years was performed and the results were analyzed in detail. Each failure was classified by considering the damaged component and the related downtime period. The analysis allowed the determination of several useful results such as the trend of failure occurrence with machine age and the identification of components and macrocomponents which are more critical in terms of both number of occurrences and downtime periods. The combination of component failure occurrences and related downtime periods was also computed to estimate which component is most critical for WT operation.
A Scalable Predictive Maintenance Model for Detecting Wind Turbine Component Failures Based on SCADA Data
