In-flight Aircraft Monitoring using a Frequency Domain Parameter Identification

To improve A/C safety and autonomy, the early detection of unexpected events is of primary importance for handling the Flight Control System and providing enhanced cueing of the envelope limits to the pilot. The approach proposed for achieving this on-line monitoring relies on a real time parameter identification technique. Owing to its robustness to measurement noises, an output error method in the frequency domain is developed to estimate the stability/control derivatives of a linearized state-space model. The process includes also pre and post-processing stages, to prevent from and to filter out inaccurate estimations. Evaluation results involve actuator Fault Detection, Isolation and Estimation as well as icing detection, and are compared to a time domain batch algorithm used as a reference method.