Presentations

Abstract

The continuous increase in wind turbine rotor size, combined with advanced pitch and yaw control strategies, is reshaping the nature of operational risks in modern wind farms. Structural vibrations, aerodynamic imbalances and control-related effects are becoming more intertwined, making root-cause identification increasingly complex and often delayed. In many cases, abnormal behaviours remain undetected for months, or are only identified after significant performance losses or structural loading concerns have already materialised. The proposed approach fully exploits the diagnostic potential of standard SCADA data, aligning with the growing role of digitalisation and data-driven decision support in wind farm O&M. The analysed case study is a wind farm composed of 11 Vestas V117 3,3 MW wind turbines, owned and operated by ENGIE Italia. The methodology combines multiple SCADA measurement channels that are typically analysed in isolation, including 1P and 3P components of longitudinal and lateral tower vibrations, rotor torque indicators, blade pitch pressures, blade specific pitch distances and dual nacelle anemometers wind speed.  The diagnostic framework follows a sequential, fleet-based analysis aimed at progressively narrowing the set of plausible root causes: • Tower vibration analysis: fleet-level assessment of longitudinal and lateral tower vibrations using 1P- and 3P-based indicators, median metrics and turbine ranking to identify machines with systematically elevated vibration levels. • Torque analysis: analogous fleet-level analysis of rotor torque, considering absolute levels and 1P/3P-related indicators, enabling direct comparison with tower vibration behaviour. • Combined interpretation: classification of turbines into machines showing normal behaviour, anomalous tower vibrations without corresponding torque anomalies, anomalous torque behaviour without significant tower response, and anomalies affecting both torque and tower vibrations, suggesting different aerodynamic, yaw-related or control-related mechanisms. • Anemometers consistency analysis: fleet-level evaluation of dual nacelle anemometer measurements through bias, relative mismatch and regression-based indicators, with focus on turbines exhibiting elevated tower vibrations but normal torque behaviour, supporting the identification of possible yaw misalignment. • Pitch system analysis: assessment of blade pitch manifold pressures and blade-specific travelled distances to identify static and dynamic pitch asymmetries, particularly for turbines showing torque-related anomalies, and to distinguish pitch-related effects confined to the drivetrain from those propagating to tower loading. The proposed framework shows how a structured, multi-channel, SCADA-based fleet analysis can support robust root-cause identification in operating wind farms without relying on intrusive or high-frequency measurements. The approach enables early detection of emerging issues, reducing diagnostic times from months to days on a single turbine scale. The results provide actionable guidance for targeted maintenance actions, support informed interaction with maintainers and help design focused follow-up measurement campaigns, contributing to more proactive and cost-effective O&M strategies.