FAULT ISOLATION METHOD AND APPARATUS IN ARTIFICIAL INTELLIGENCE BASED AIR DATA SYSTEMS FOR AIRCRAFT

摘要

A method providing fault isolation, in an air data system which uses artificial intelligence to generate an air data parameter, includes generating the air data parameter as a function of a plurality of measured values such as static pressures. Then, estimates of each of the plurality of measured values is generated as a function of the generated air data parameter. Each measured value ran then be compared to its corresponding estimate to determine if a difference between the measured value and its corresponding estimate exceeds a threshold and therefore indicates a fault in a device which provides the measured value. The artificial intelligence algorithms may be neural networks or support vector machines. The air data system may be a flush air data system (FADS). The air data parameter may be an angle of attack, an angle of sideslip, a Mach number, a global total pressure or a global static pressure for the air vehicle. The plurality of measured values may comprise a measured value indicative of a control surface position , a control surface loading, a force, a vehicle mass at take-off, a vehicle mass balance, a remaining fuel mass, an engine thrust, satellite information, an altitude, an air temperature, a vehicle acceleration, a vehicle attitude or a landing gear position.