Research On The Mechanical Fault Diagnosis Technology Based On Artificial Neural Network Theory

With the development of production, the importance of the mechanical fault diagnosis technology is more and more obvious. With regard to sophisticated mechanical equipment having much fault~ much process and sudden fault, the traditional diagnosis technology and theory method often show more limitations,so the mechanical fault diagnosis is found in some difficult position. In order to overcome these shortcomings, the thesis focuses on the application of artificial neural network(ANN)theory in the mechanical fault diagnosis technology.First, on the basis of systematically introducing the importance and the internal and external developing current situation~ trend, the thesis points out the application of ANN theory in mechanical fault diagnosis has great applying value and developing potential. At the same time, the thesis also indicates that this application is still imperfect and is in developing peroid.Next, the thesis carefully describes the definitions method and course of the mechanical fault diagnosis technology, and the structure~ algorithm~ trait and nature of ANN model. And it relates the theoretical meaning of the research on the mechanical fault diagnosis technology based on ANN theory. The thesis also points out that the peculiar structure and trait of ANN have poured into new vitality,so it is possible that the mechanical fault diagnosis technology extricates itself from predicament.Then, the thesis designs a program model of mechanical fault diagnosis neural network learning and diagnosis, and uses Turbo C2. 0 to achieve the total course of fault diagonisis on the gearbox in J?0 tractor based on ANN theory. The result indicated that the model not only overcomes the defects of traditional diagnosis method, but also raises the accuracy and promotes the efficiency of mechanical fault diagnosis greatly.Last, the thesis objectively analyzes the insufficiency and puts forward some improving measures.