The fault diagnosis and prognosis research for Independent power-supply system(IPSS) was an important step to ensure the safety for the complex moving devices system. Current fault diagnosis system problems for IPSS were analyzed. The method of combining multi-sensor information fusion with multi-Agent was put forward to improve the diagnosis reliability and the system expansibility. A fault diagnosis system for IPSS using multi-Agent and information fusion was established based on Agent automation, distribution and collaboration. According to the fault symptom feature, Dempster-Shafer evidential theory was introduced into diagnosis result fusion. The multi-neural network local diagnosis Agent and the Dempster-Shafer evidential reasoning Agent implementation were disserted in detail. Finally, taking an aircraft electric power system as an example, the diagnostic process was simulated. The results indicate that the method can improve the diagnosis reliability.
Zhang Li, Yuan Haiwen, L�� Hong, Yuan Haibin.Fault diagnosis for independent power-supply system using multi-agent information fusion[J] JOURNAL OF BEIJING UNIVERSITY OF AERONAUTICS AND A, 2010,V36(8): 936-939
Teng Zhaosheng,Luo Longfu,Tong Tiaosheng.Intelligent detecting system and data fusion[M].Beijing:China Machine Industry Press,2000:246-247 (in Chinese)
Dempster A.Upper and lower probabilities induced by a multivalued mapping[J].Am Math Statist.1967,38:325-339
Zhang Li,Yuan Haiwen,Yuan Haibin,et al.Research and application of atml for aircraft electric power diagnosis and prognostic system[C]// Proceeding of 8th International Conference on Electronic Measurement and Instruments.Beijing:IEEE Press,2007:3615-3618