Wind turbine fault detection through principal component analysis and multivariate statistical inference
Document typeConference lecture
Rights accessOpen Access
This work addresses the problem of online fault detection of an advanced wind turbine benchmark under actuators (pitch and torque) and sensors (pitch angle measurement) faults of different type: fixed value, gain factor, offset and changed dynamics. The fault detection scheme starts by computing the baseline principal component analysis (PCA) model from the healthy or undamaged wind turbine. Subsequently, when the wind turbine is inspected or supervised, new measurements are obtained and projected into the baseline PCA model. When both sets of data—the baseline and the data from the current wind turbine— are compared, a multivariate statistical hypothesis testing is used to make a decision on whether or not the wind turbine presents some damage, fault or misbehavior. The effectiveness of the proposed fault- detection scheme is illustrated by numerical simulations on a well-known large offshore wind turbine in the presence of wind turbulence and realistic fault scenarios. The obtained results demonstrate that the proposed strategy provides an early fault detection, thereby giving the operators sufficient time to make more informed decisions regarding the maintenance of their machines.
CitationPozo, F., Vidal, Y., Acho, L. Wind turbine fault detection through principal component analysis and multivariate statistical inference. A: "8th European Workshop on Structural Health Monitoring": Bilbao, Spain, 5-8 July 2016.