Study Of Rough Set-based Knowledge Acquisition Model Of Remote Fault Diagnosis For Aircraft

Along with the development of China's civil aviation industry, how to protect civilian aircraft in flight safety has become increasingly important issue. The key to solve this problem is to establish an aircraft fault diagnosis system in which the aircraft faults can be timely and accurate diagnosed and its position can be found out. However, there are a huge number of civilian aircrafts and routes in China, it is impossible to set up technical support systems for each airline in every airport. After1990s, as the rapid development of Internet, it is becoming possible to set up remote fault diagnosis system for aircraft through Internet technology.Because the knowledge base is an important part of the fault diagnosis system, how to acquire knowledge becomes very urgent. However, the existing models and theories of knowledge acquisition can not grant the demand of the aircraft data. Rough set theory issued by Poland academics Z.Pawlak in 1980s can discovery the inherent law directly from the given description set. It provides a feasible method of knowledge acquisition for aircraft fault diagnosis system.Some exploration on knowledge acquisition technology based on rough set for the aircraft remote fault diagnosis is made in this paper in the following areas:[1] Remote fault diagnosis based on Internet technology is selected after some fault diagnosis technologies are introduced.[2] Equal interval width and fuzzy clustering discretization method and attribute-oriented data reduction method are mainly discussed according to the characteristics of aircraft remote diagnosis after some discretization methods and data reduction methods based on rough set are introduced. Then algorithms for them are made up.[3] Parameters of aircraft system are researched and a knowledge acquisition model based on rough set is established on some engine parameters.[4] A framework for remote aircraft fault diagnosis system which is based on 3-Tie C/S structure, developed by Borland C++ Builder 6.0 and supported by Microsoft SQL Server 2000 is founded after the requirements of remote aircraft fault diagnosis system are analyzed.[5] The knowledge acquisition model and remote aircraft fault diagnosis system are tested and proved to be right by data from military aircraft engine and civilian aircraft engine through the existing campus network.