Using The Sequentially Elimination Method To Diagnose Turbine Blade Faults

Wind turbine blades play vital roles in wind turbine power generation and are exposed to variable environmental conditions and complex mechanical loads during their operation, which may cause surface damage that does not prevent wind turbine operation but reduces aerodynamic performance and energy generation. If defects are left to grow and develop, they result in structural damage. Therefore, early detection of faults in wind turbine blades is crucial for reducing maintenance costs and increasing power generation. However, the lack of blade measurements for early detection is a real problem. This paper presents a Local Outlier Factor approach to detect outliers and sequential elimination of components to diagnose blade faults early. This method uses SCADA data to analyze turbine performance and detect outliers in power production. When there is an anomaly, the data of the various components are analyzed using a local factor. If an outlier was found in the produced power, the component was examined to determine if this was the case. Is this component the cause of the loss? If not the cause of the loss, it is eliminated, and the remaining components are examined and eliminated sequentially until only the blades remain the cause of failure. Although this technique does not replace blade testing, it provides early fault detection and reduces blade testing times, which saves turbine downtime and helps reduce cascade damage caused by blade faults.