An AI-based fault detection model using alarms and warnings from the SCADA system

Predictive maintenance is a key element for lowering Operation and Maintenance (O&M) costs of wind turbines. Predictive maintenance models are usually based on drivetrain vibration data or operational timeseries from the Supervisory Control And Data Acquisition (SCADA) system, while readily available alarms and warnings from the SCADA system are typically not utilized. In this work we present a novel Artificial Intelligence (AI) based approach for early fault detection of wind turbines using alarms and warnings from the SCADA system.