Method for detecting a failure in an aircraft and system configured for executing the method
AssigneeAirbus Operations GmbH
InventorsChristophe Vlacich, Stephan Roux, Thomas Massonneau, Driss El Maamar
Life of this patent
Estimating the term…
Abstract
A method for detecting and localizing a failure of an input sensor of a flight control computer of an aircraft including comparing a first piece of information representative of a probability of occurrence of the failure, delivered by a neural network, with a second piece of information representative of a probability of occurrence of the failure, delivered by a classifier trained to carry out a detection and a localization of the failure. It is thus possible to optimize a failure diagnosis and a localization of a failure of an input sensor used in the determination of flight controls.
Claims10
A method for detecting and localizing a failure of an input sensor of a flight control computer of an aircraft, the method being executed by a module for detecting and localizing failures comprising electronic circuitry configured for executing the method, which comprises:
- conditioning at least a first signal delivered to the sensor by the flight control computer thereby resulting in a first conditioned signal, and conditioning at least a second signal delivered to the flight control computer by the sensor thereby resulting in a second conditioned signal, according to a predetermined format;
- analyzing the first conditioned signal and second conditioned signal, by a neural network trained to carry out a detection and a localization of the failure using the first conditioned signal and the second conditioned signal representative of the first signal and the second signal, respectively, delivered by or to the flight control computer, the neural network delivering a piece of information representative of a probability of occurrence of a failure;
- comparison of a first piece of information representative of a first probability of occurrence of the failure, delivered by the neural network, with a second piece of information representative of a second probability of occurrence of the failure, delivered by a classifier trained to carry out a detection and a localization of the failure, using signals representative of the conditioned signals, in order to deliver a piece of information representative of a probability of occurrence of a failure;
- then,
- supplying at least a third piece of information representative of a third probability of occurrence of the failure based on the first and second pieces of information.
A device for assisting with detection and localization of a failure of an input sensor of a flight control computer of an aircraft, the device comprising a module for detecting and localizing a failure comprising electronic circuitry configured for carrying out:
- conditioning at least a first signal delivered to the sensor by the flight control computer thereby resulting in a first conditioned signal, and conditioning at least a second signal delivered to the computer by the sensor thereby resulting in a second conditioned signal, according to a predetermined format;
- analysis of the first conditioned signal and the second conditioned signal, by a neural network trained to carry out a detection and a localization of the failure using the first conditioned signal and the second conditioned signal which are representative of the first signal and the second signal, respectively, delivered by or to the flight control computer;
- comparison of a first piece of information representative of a first probability of occurrence of the failure, delivered by the neural network, with a second piece of information representative of a second probability of occurrence of the failure, delivered by a classifier trained to carry out a detection and a localization of the failure, using signals representative of the conditioned signals;
- then,
- supplying at least a third piece of information representative of a third probability of occurrence of the failure based on the first and second pieces of information.
Dependencies are inferred from the claim text.