Joint Actuator and Sensor Fault Diagnosis in Wind Turbine Systems

ABSTRACT This paper presents a fault diagnosis method for wind turbine systems that can simultaneously detect and estimate both actuator and sensor faults. Unlike previous studies that treat these faults separatelyand focus mainly on pitch angle‐related variables, our study recognizes the interdependence between actuator and sensor faults and expands the focus to also include variables of tower‐top velocity and generator speed. We propose a system transformation technique that makes sensor faults explicitly observable over timeby incorporating sensor faults into the state space model. This allows us to control the convergence rate of system states and fault parameters separately. The proposed technique was built on the implementation of an adaptive Kalman filter to handle time‐varying parameters. Through comprehensive simulation, we evaluated the proposed method under three distinct scenarios using real‐world input data from the Hywind Scotland project. The results demonstrate that our combined methods—the system transformation technique, prior to the adaptive Kalman filter—are able to accurately detect and estimate joint actuator and sensor faults in wind turbine systemswhile providing flexibility in controlling the convergence rate to specify the system requirements.