Study On Engine Fault Diagnosis Based On Multi-Information Fusion

Aero-engine not only has complex structure but also works in the special environment. These factors made the aero-engine difficult to work reliably. Therefore, it is particularly important to design effective methods and carry on fault diagnosis to aero-engine. In this paper, health diagnostic methods for aero-engine based on multi-source information fusion is proposed, which provides a new way for solving the problems caused by complex structure and the uncertain factor of the engine itself.Take gas path system of aero-engine as the research object, first of all, this paper discussed the basic principles of information fusion and fusion features of different levels, and demonstrated effectiveness of information fusion technology in field of fault diagnosis from the perspective of information theory. And then the paper studied the feature level fusion based on neural network technology, the decision-level fusion based on DS evidence theory and fuzzy integral in a deep level respectively. For the difficulties of determining the fuzzy integral measure, this paper proposed the idea of wtilizing particle swarm optimization algorithm to optimize fuzzy integral. Finally, according to the characteristics of each layer, the paper proposed the double-fusion model, which combined the method of neural network feature level fusion and decision level fusion. On this basis, this paper studied and analyzed the integration in different methods respectively in the aero-engine fault diagnosis. Simulation results verified the feasibility of the proposed method.