Actuator Fault Detection by Aerodynamic Model Identification

Abstract

This paper presents a Fault Detection and Diagnosis (FDD) method for detecting faults in control surface actuators. Due to a lack of “local” measurements directly related to the actuator and control surface, detection of certain type of faults is very hard. The method introduced here makes use of the knowledge of the aircraft's dynamics to achieve fault detection and isolation. E.g., an aileron fault, the rolling moment of the aircraft is modeled. The parameters of the model are identified online, and if a fault occurs, a change in these parameters is observed. By evaluating these changes the faults can be detected and isolated. The advantage of this method is that it can detect different types of faults (sensor bias, jamming, disconnection) with only one method.