Multi-agent Fault Diagnosis System Based On Information Fusion, And Spacecraft

As the development of technologies such as artificial intelligence and distribution theory, agent technology becomes the research focus gradually. Systems based on multi-agent have the characteristic of distributed, scalability as well as fault tolerance. And this system can form the more reliable intelligent system by combining with the data fusion technology. Multi-agent fault diagnosis system based on data fusion was proposed and it was applied to the fault diagnosis of RLV.The main research works of this paper are as follows:1. A system framework of multi-agent fault diagnosis based on data fusion was proposed.This system framework not only has the characteristic of distributed, scalability as well as fault tolerance, but also takes advantages of data fusion system such as fully using data source, and this enhanced the reliability of system.2. The thermal protection system of RLV was analyzed.? The thermal protection system of RLV was taken as the applied object. On the basis of analyzing its architecture and the material, its typical fault pattern, monitoring position, sensor collocations as well as the choice were researched deeply.3. The monitoring system of thermal protection system of RLV including both on-board monitoring diagnosis system, ground remote monitoring diagnosis system and ground manual maintenance system was proposed.This monitoring system is based on the analysis of foreign monitoring systems of RLV thermal protection system and the real-time health monitoring system of X-33. Hence, this system is reasonable.4. The design and development of the multi-agent on-board fault diagnosis system based on data fusion was fulfilled.The design and development includes of on-board fault monitoring agent, on-board fault recognition agent and on-board fault forecast agent. Specially developed the on-board fault forecast agent by VC++, and took the surface pressure of RLV wing leading edge as the reference data, forecasted its pressure by BP neural networks. The result is well. At last, this paper analyzed the multi-agent data fusion technology and fused the diagnosis results of multiple agents by D-S theory. It can enhance the reliability of diagnosis system.