Technique & Research On Fault Intelligent Diagnosis Based On ANNES

With the developing of modern technology, especially with the speedy developing of computer technology and network communication technology, industry system has been become bigger and bigger, and will further progress to integrated automation system mode, which is wide field interlinking, intelligent diagnosing and remote monitoring. These changes cause the structure of modern equipment become more complex, and the coupling of components become worse increasingly. So the factor of affecting system run has increased abruptly, and the probability of fault occurrence becomes higher and higher, the intelligent diagnosis technology emerges as the times require.Because the fault intelligent diagnosis technology involved in several fields' technique, there has no whole fault intelligent diagnosis theory so far. In the basis of discriminating and analyzing of existing diagnosis techniques, combining with the practicing of project, this paper researched into the fault intelligent diagnose technique based on ANNES.Firstly, paper briefly introduced the theory base of fault diagnoses technology, advanced the common structure of fault diagnoses technology and expounded the character of function module. Then author analyzed the acquirement and representation of knowledge in fault intelligent diagnosis system, advanced the common model of knowledge acquiring. After this, paper discussed each character and relationship between deep knowledge and shallow knowledge, presented the mixed representation model. Author also discussed the basic thinking and applications in diagnosis knowledge representing of oriented object technique. In order to exert both advantage fully and improve the knowledge representation performance of ANNES, according to function character of deep knowledge and shallow knowledge in the intelligent diagnose system, combining with oriented objecttechnique, author advanced the oriented object mix representation model of deep knowledge and shallow knowledge, and studied the object hierarchy, knowledge unit's fundamental structure and the realizing technique of the model. Furthermore, author delved the information's uncertainty of diagnosis system, analyzed the reason and studied the representation methods of uncertain knowledge and common model of uncertain ratiocination. Author also emphasized on fundamentals, algorithm and algorithm's amelioration of three-level forward BP neural network. After comparing and analyzing each peculiarity of neural network and expert system, author discussed the combination possibility of both, and then studied the combination's methods and approaches. In order to exert both advantages, author designed the structure of fault intelligent diagnose based on ANNES, studied the realization. Taking a example for the complex electronic equipment's fault intelligent diagnose, author discussed the application significance of ANNES. At last, author prospected the development of ANNES.