Intelligent Diagnosis Method Research On Airplane Fuel System

The aircraft fuel system, which affects the airplane performances, the flight safety and the collection of seaworthy evidence, is a crucial system on the aircraft. Along with fast development of the aviation technology, the aircraft fuel system becomes more and more complex day by day, and people set higher requests regarding the aircraft fuel system reliability. Thus the failure detection and the diagnosis technology become the essential methods to enhance the system reliability.This article roots in a fund project. The primary mission is to do researches of some diagnosis methods which are based on the signal processing and the neural network principles. Main work as follows:Aimed at a certain aircraft fuel system, the fault types of appendixes, measuring elements, the pipelines and the fuel tanks are analyzed. Then the breakdown patterns and each kind of breakdown indications are analyzed.Finally, the breakdown types and patterns of the system are summarized.Various diagnosis methods are studyed and analysed.Then appropriate diagnosis methods and a diagnosis hierarchical fault diagnosis scheme are proposed according to characteristics of the fuel system.Wavelet analysis method is studied. Then the wavelet analysis is proposed to be applied to fault diagnosis of the sensors on the airplane fuel system according to the characteristics of it. Lastly, the validity of the fault diagnosis method for the sensor of oil level is proved in the environment of MATLAB and C++Builder by the real-time data collected in the semi-physical simulation.To solve the problem that the failure of sensor fault diagnosis on a certain airplane fuel system, a new method of sensor fault diagnosis based on wavelet analysis combined with RBF neural network is presented. Lastly, the validity of the fault diagnosis method for the sensor of oil level is proved in the environment of MATLAB and C++Builder by the real-time data collected in the semi-physical simulation.To solve the diagnosis problems of the pumps, the fuel tanks and the pipelines, a diagnosis method based on the fuzzy theory and the nerve network theory is proposed, the validity of the method is proved in the environment of MATLAB.