Online detection and diagnosis of sensor faults for a non-linear system

Abstract

In systems, the fault is an internal occurrence. It becomes a failure if the defect is not detected and corrected. Sensors have been widely employed as a vital component of data collection systems, particularly in the industrial and agricultural sectors. Sensors are prone to failure due to their harsh operating environment. As a result, early detection of sensor faults is crucial for taking corrective action to reduce the impact. In this paper, faults in generator speed and wind turbine velocity have been investigated. The Extended Kalman Filter is utilized to identify the sensor faults in wind turbine model. The residual generation is used to detect the fault. The residual is the discrepancy between the real and estimated outputs. A Linear Quadratic Regulator controller is used for the stabilization of an unstable system.