The Research Of Remote Fault Diagnosis Based On The Mechanical Vibration

The remote fault diagnosis system based on internet is introduced in this paper. According to the researching and analysing of the mechanism and character of typical faults for Rotating Machinery, and combining with expert system technology and computer technology, we analyze it's feasibility and superiority,introduce the browser/server structure,summarize it's frame and main function module,finally indicate the capability of system's self-study.Especially, the rough set can optimize the structure of BP neural network. In fault diagnosis area, the study quantity of neural network model is often great in the engineering diagnosis, because some faults have all kinds of fault types and each of the faults has all kinds of signs. The reduction of attribute in rough set make the nerve cell of input layer of neural network reduce, which simplify the structure of neural network, with the result that increase the efficiency and accuracy of diagnosis. Final output use fuzzy theory to gain the fault type within the extent of limited value. The model will be well application prospect in fault diagnosis.Visual Studio.Net is used in system as developing environment, Visual C#.Net,XML and ASP was united together as developing tools, and developing efficiency and running efficiency perfectly combine on the basis of the technology of DOM,SAX and COM.