Posters

Abstract

Wind turbines are complex machines that operate under harsh conditions and require regular maintenance to ensure optimal performance and reliability. However, conventional maintenance strategies based on fixed intervals or reactive actions can be costly and inefficient, leading to increased downtime and reduced energy production. Therefore, there is a need for data-driven approaches that can monitor the overall state of wind turbines and identify potential failure risks in advance, allowing for timely and preventive interventions. In this talk, we present our work on analyzing historic SCADA data from wind turbines to find patterns in sensor and operating data that indicate the onset of wear and damage in the wind turbine components, especially the gearbox. We demonstrate that through real-time monitoring of SCADA data, anomalies and deviations from normal behavior can be promptly detected. Additionally, we illustrate how this approach enables the estimation of the remaining lifetime of the components. We also discuss the limitations of SCADA data and how they can be overcome by combining them with a novel proprietary sensor that performs real-time wear particle characterization in the gearbox. This sensor can provide more accurate and specific information about the type and severity of the damage and enable more targeted and effective repairs.