Aviation Piston Engine Wear Fault Diagnosis Expert System Design And Implementation

In general aviation area, there are a large number of planes using air piston engines as the power plant. During the usage process, the air piston engines will certainly encounter many sorts of fault and cannot work properly due to its complex structure, large number of parts and poor working conditions, while the wear failure is the most common failure mode for air piston engine. The diagnosis of air engine wear failure is an important way to ensure the flight safety. Maintenance expert will involve in the diagnosis because of the complexity and difficulty of diagnosis of aeroengine, but the number of aircraft maintenance experts is limited, therefore, in this paper, a specific problem domain expert system is proposed aiming at solving the problem of wear failure of small piston engines widely used in flight training and general aviation.On the background the real application of Lycoming engine, it mainly makes a reach on the four major friction pair conditions:cylinder barrel and piston ring, crank shaft and bearing pad, valve and valve seat, and cam and tappet,and analyzes the wear friction form and wear mechanism for each friction pair. Then it focuses on the analysis of friction pair of engine cam and tappet, thus obtaining knowledge. This paper also introduces related theory of several oil analysis technologies, it analyses the source of grain and the relationship between oil analysis and wear, illustrating the characteristics of several oil analyses, it focuses on describing the application of spectral analysis and iron spectral analysis in the fault diagnosis by some examples. This paper mainly introduced the development background of the subject, to complete the functions and development process,and described specifically the design ideas of air piston engines fault diagnosis expert system. The system uses the method of generating representation in diagnosis knowledge representation, which is intuitive and accords with thinking habits of human, thus facilitating developing the inference machine. The basic idea of reasoning is using positive reasoning strategy. The system conducts the analysis and study of wear fault diagnosis of air piston engines through oil analysis, and then establishing the knowledge base. Use the structure of separation between knowledge base and the inference machine to avoid affecting the normal operation of the inference machine when adding, deleting and expanding the knowledge base.The development of the wear fault diagnosis expert system prototype is through C++ Builder as the Front-end development tools, SQL2000 as the backend database. It Initially implemented functions in 4 aspects including fault diagnosis, knowledge-base system management, training and help. Based on the application of knowledge base and database techniques, the design objective of proposing maintenance suggestions through data extracted from input spectrum or iron spectral analysis is realized. The example is indicated that the system had the characteristic of easy operation and could simulate the fault diagnosis ability of expert.