数字内容技术与服务研究中心知识库

Nowadays periodic maintenance and routine check are adopted as the dominating maintenance mode of domestic aircraft. However, this mode requires a large number of experienced professionals, resulting in the waste of manpower and resources. In this paper, taking the flight data of a certain type of aircraft as the main data source, extracting the fault symptom information and the system failure mode, a fault diagnosis reasoning expert system based on failure mode is proposed. The main work of this paper is as following :

1) In this paper, the modified Back Propagation neural network is utilized to train the sample data. The trained neural network model is then saved in the knowledge base of the fault diagnosis expert system for future feature extraction of the new flight data; The Fault Tree Analysis (FTA) is adopted by the fault diagnosis expert system for fault detection of the aircraft. By combining production-rules and the minimum cut of the fault tree, the fault mode of the aircraft can be effectively extracted; An optimized reasoning engine is built based on the fault mode, which utilizes the forward reasoning pattern for logical inference. Combining the expert system with The Fault Tree Analysis, our system can be effective and efficiency benefit from the expert system. Besides, fault tree analysis can reduce the difficulty of diagnostic reasoning and knowledge acquisition.

2) A prototype fault diagnosis expert system, which is based on failure mode with Java by using MySQL as the database on the Windows 7 platform, is developed. This system has