Research On Marine Diesel Engine Fault Diagnosis Method Based On Information Fusion Technology And Application

Diesel engine fault diagnosis technology based on multi-sensor information fusion is the development certain because of the complexity of diesel engine architecture and difficulty of its fault diagnosis. In this paper, diesel engine fault diagnosis technology based on information fusion theory is deeply researched on the basis of former achievements linking to practical project.As a very important part of diesel engine fault diagnosis, feature extraction is directly related to the success for model recognition of diesel engine or not. Diesel engine feature extraction methods based on vibration signals are largely researched. But it is very difficult for diesel engine feature extraction based on vibration signal because diesel engine vibration signal is very complex with nonlinear and nonstationary characters. Hence, the author first made deeply research on diesel engine feature extraction based on vibration signals.The theory of vibration signals formed is very important for diesel engine fault diagnosis because it is basis of feature extraction. Hence, the theory of vibration signals formed is researched which can be useful for diesel engine feature extraction. The fluctuation attribute of diesel engine vibration signal is analyzed and method of its solution is investigated. Diesel engine feature extraction methods are researched on the basis of theory of vibration signals formed. In this paper, local wave method and wavelet K-L information distance of diesel engine feature extraction methods are investigated. The main purpose is to give a quantitative description of feature information. It is preparation for diesel engine fault diagnosis based on information fusion theory.A single cylinder, four-stroke diesel engine is used as exemplar to investigate on diesel engine simulation experiment. Eight type diesel engine working conditions are monitored. The data samples are used to diesel engine fault diagnosis technology research based on information fusion theory. According to the samples, the diesel engine fault diagnosis based on information fusion using neural networks is investigated. The fuzzy neural network as the center of fusion for fault diagnosis is deeply researched. At the same time, diesel engine fault diagnosis technology based on D-S theory in the decision level is also researched. These methods belong to basic fusion methods and are the preparation in deeply research of fusion theory and frame architecture design.The combination frame architecture of time-spatial information fusion system is put forward according to diesel engine fault diagnosis' complexity and signal fluctuation. This architecture is set up on the basis of neural networks and D-S theory. The superiority of this architecture is researched. The central question of using this architecture in diesel engine fault diagnosis is especially investigated. At the same time, the solution is given for the problem. The application process on diesel engine fault diagnosis is introduced by practical examplesanalysis. It can be concluded that the frame architecture is an effective method for diesel engine fault diagnosis.At last, the Zhenghe marine diesel engine fault diagnosis and intelligent system is explored and developed according to practical project. The system affords effective support for preventive maintenance of the marine diesel engine according to practical application. At the same time, it is first step for diesel engine fault diagnosis based on information fusion theory in practical application.