Article révisé par les pairs
Résumé : This work describes an autonomous condition monitoring framework to process and analyze data measured on wind turbine gearboxes. Industry 4.0 and the Industrial Internet of Things open the door for much more elaborate measurement and data analysis campaigns thanks to the reduction in cost of sensors and of processing power. This increase in data acquisition and handling potential is especially useful considering that most current state-of-the-art methods in signal processing often lead to large quantities of health indicators due to the multiple processing steps. Such large numbers of indicators become unfeasible to inspect manually when the data volume and the number of monitored turbines increases. Therefore, this paper illustrates a hybrid analysis approach that combines advanced signal processing methods with machine learning and anomaly detection. This approach is validated on an experimental wind turbine gearbox vibration data set.