Robust fault detection and isolation of wind turbines using interval observers

In order to improve reliability and safety of wind turbines, it is important to detect and isolate faults as fast as possible. In this work, an approach based on interval observers is used for robust Fault Detection and Isolation (FDI) of wind turbines. Fault detection is addressed using interval observers and unknown but bounded descriptions of the noise and modeling errors, while fault isolation is based on a row-reasoning approach. The performance of the proposed robust FDI scheme is assessed using a set of fault scenarios considered in a wind turbine benchmark.