In this paper the problem of detecting and isolating sensor faults on a general aviation aircraft, in the presence of external disturbances, is considered. The proposed approach consists of an extended Kalman observer applied to an augmented aircraft plant, where some integrators are added to the output variables subject to faults. The output of the integrators should be ideally zero in the absence of model uncertainties, external disturbances and sensor faults. A threshold based decision making system is adopted where the residuals are weighted with gains coming from the solution of an optimization problem. The proposed nonlinear observer has been tested both numerically on a large database of simulations in the presence of disturbances and model uncertainties and on input-output data recorded during real flights. In this case, the possibility of successfully applying the proposed technique to detect and isolate faults on inertial and air data sensors, modelled as step or ramp signals artificially added to the real measurements, is shown. © 2013 IEEE.
An SFDI observer-based scheme for a general aviation aircraft / Ariola, M.; Corraro, F.; Mattei, M.; Notaro, I.; Sollazzo, A.. - (2013), pp. 152-157. (Intervento presentato al convegno 2nd International Conference on Control and Fault-Tolerant Systems, SysTol 2013 tenutosi a Nice, fra nel 2013) [10.1109/SysTol.2013.6693930].
An SFDI observer-based scheme for a general aviation aircraft
Mattei M.;
2013
Abstract
In this paper the problem of detecting and isolating sensor faults on a general aviation aircraft, in the presence of external disturbances, is considered. The proposed approach consists of an extended Kalman observer applied to an augmented aircraft plant, where some integrators are added to the output variables subject to faults. The output of the integrators should be ideally zero in the absence of model uncertainties, external disturbances and sensor faults. A threshold based decision making system is adopted where the residuals are weighted with gains coming from the solution of an optimization problem. The proposed nonlinear observer has been tested both numerically on a large database of simulations in the presence of disturbances and model uncertainties and on input-output data recorded during real flights. In this case, the possibility of successfully applying the proposed technique to detect and isolate faults on inertial and air data sensors, modelled as step or ramp signals artificially added to the real measurements, is shown. © 2013 IEEE.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.