False Alarm Detection with K-NN Algorithm forWind Turbine Maintenance Management

The operation and maintenance management are critical for the efficiency and competitiveness of the wind farms. The detection of false alarms is a significant variable for wind turbine maintenance. This paper proposes a novel approach based on K-nearest neighbour algorithms that compares different k-fold cross-validation values. It is presented a real case study based on three wind turbines to test the methodology. The data obtained with supervisory control and data acquisition is analyzed and compared with the alarm log as a response variable. The behaviour of the three turbines has been demonstrated to be analogous, implying that the methodology is robust. It is demonstrated that the increase of the K-cross validation value does not increase the accuracy. The results achieved an accuracy of 98%, and the method can detect more than 22% of false alarms. These outcomes demonstrates that the suggested approach can detect false alarms in wind turbines.