Experimental Validation of Leakage Detection in a Fluid Power Pitch System Using a State Augmented EKF-Approach

The pitch system is a crucial part of today's wind turbines and is used for power regulation when operating above rated wind speed. According to studies the pitch system is responsible for up to 20% of the total downtime of a wind turbine. As an attempt of increasing the reliability and availability attention has come to fault detection and condition monitoring of such systems. This paper presents a State Augmented Extended Kalman Filter, SAEKF, for detecting both internal and external rod leakage. Furthermore, an external load torque is included in the filter to improve the performance in the presence of external load. The SAEKF is tested experimentally for different levels of both internal and external rod leakage under working conditions similar to the ones experienced for a pitch system located in a wind turbine. The working conditions include unkown fast varying external loads. The experiments show that the SAEKF is capable of detecting internal and external leakage down to 0.10 l/min and 0.34 l/min, respectively. The internal leakage is detected with a estimation error of maximum 0.04 l/min while the external leakage estimation error is going up to 0.43 l/min for high levels of external leakage.