ICM-Wind: semantics-empowered fluid condition monitoring of wind turbines

We present the first system, called ICM-Wind, for semantics-empowered fluid condition monitoring (FCM) in wind turbines. It monitors the condition of fluids in the wind turbine gearbox, recognizes actual and the onset of failures of FCM sensors and components installed on the gearbox, and provides knowledge-based failure diagnosis support to non-experts. For this purpose, the ICM-Wind system performs semantic sensor data analysis by applying semantic technologies for interpreting the state of turbine parts and answering questions related to their maintenance. Domain knowledge is encoded in OWL2 and with SPIN rules. Fault detection and diagnosis queries are answered by use of the semantic reasoners Fact++, STAR, TopSPIN rule engine, and SwiftOWLIM store. The system prototype was successfully tested in cooperation with the HYDAC Filter Systems GmbH based on given selected samples of a two-year recording of FCM multi-sensor and operational data for two wind turbines of a regional on-shore wind farm operated by the ABO Wind AG.