Advanced Data Mining Approach for Wind Turbines Fault Prediction

Wind turbine operation and maintenance costs depend on the reliability of its components. Thus, a critical task is to detect and isolate faults, as fast as possible, and restore optimal operating conditions in the shortest time. In this paper, a data mining approach is proposed for fault prediction by detecting the faults inception in the wind turbines, in particu lar pitch actuators. The role of the latter is to adjust the blade pitch by rotating it according to the current wind speed in order to optimize the wind turbine power production. The fault p rediction of pitch actuators is a challenging task because of the high variability of the wind speed, the confusion between faults and noise as well as outliers, the occurrence of pitch actuator faults in power optimization region in which the fault consequences are hidden and the actions of the control feedback which compensate the fault effects. To answer these challenges , the proposed approach monitors a drift from normal operating conditions towards failure condition. To ach ieve drift detection, two drift indicators are used. The first indicator detects the drift and the second indicator confirms it. Both indicators are based on the observation of changes in the characteristics of normal operating mode over time. A wind turbine simulator is used to validate the performance of the proposed approach.