The Study Of Diesel Engine Fault Diagnosis Based On Information Fusion

As a power equipment, diesel engine plays an important role in the current many areas, which has a huge social and economic benefits. In order to reduce or even avoid losses caused by the failure, the appropriate fault diagnosis methods, become the focus of attention.In order to improve the accuracy of fault diagnosis, this paper puts forward the study of diesel engine fault diagnosis based on information fusion. By introducing a new kind of swarm intelligence optimization algorithm, combined the neural network method, which is the mainstream of information fusion technology, this article proposes a new method of fault diagnosis, namely gravitational search-BP neural network method, and apply it to a vibration diagnosis of PA6 diesel engine wear fault. On the vibration signal data acquisition, its characteristic parameters as sample data of information fusion, make the simulation test, verify the effectiveness and feasibility of the proposed method, which is applied to the diesel engine fault diagnosis. Besides that, the corresponding software is designed.First, information fusion technology is systematically researched. The three levels of information fusion model, that is the data layer, feature layer, decision layer are analyzed, and compared the advantages and disadvantages of them. At the same time, information fusion methods for fault diagnosis are studied, to verify the feasibility of information fusion technology that is applied to fault diagnosis, which provides a theoretical basis on choosing feature layer fusion and neural network method.Second, the failure mechanism of diesel engine is studied fully. Through the analysis of the common faults of diesel engine model and the main fault features, put the cylinder head vibration signal that the most can reflect the fault feature as the research object, research the related characteristics of vibration signals, foreshadowing for follow-up study.Finally, the application of BP neural network in fault diagnosis is insufficient, gravitational search algorithm is used to optimize the initial weights and threshold of BP neural network in this article. Moreover, an intelligent fault diagnosis method based on the combination of gravitational search algorithm and BP neural network is proposed. It is applied to the vibration diagnosis of the diesel engine wear fault. With the example of diesel engine, the results show that the proposed method has higher accuracy than BP neural network. Besides, it is effective in classification and diagnosis of the faults of diesel engine. At the same time, a software for diesel engine fault diagnosis is developed, and the diagnostic results are visually presented.