Study On Fault Detection And Diagnosis System For Power Device Based On Ferrography And Image Recognition Technology

The ferrography, as a kind of non-disassembly fault detection and diagnosis method for wear abrasion, has been paid more attentions to the application by many experts and scholars. However, because of some demerits existed in the application of the traditional ferrography, the study of the automatic recognition system for debris based on ferrography and image recognition technology has played an important role in the process of the development and application of the ferrography.In this paper, the achievements in related field are introduced firstly, and a fault detection and diagnosis system for power device based on ferrography and image recognition technology is designed considering the characteristics of a given type of power device. Moreover, the related technologies and methods are studied deeply.The main contents of this paper are listed as follows.(1) Some general fault diagnosis methods of power device are compared; the design method of fault detection and diagnosis system in which the ferrography information is adopted as the information source is presented.(2) The supporting techniques and methods of the fault detection and diagnosis system are introduced; the method of extracting the debris's features based on the Principal Components Analysis (PCA) is put forward, and a debris recognition system based on the series model which combines the Principal Components Analysis with BP neural network is designed.(3) The advantages and disadvantages of the CLIPS using in the design of an expert system are analyzed; the design method of the expert system based on the technique of embedding the CLIPS in DELPHI6 is put forward.(4) A fault detection and diagnosis system for power device is designed and implemented meeting the requirement of the application, which combines debris recognition with diagnosis decision.(5) The validity of this fault detection and diagnosis system is demonstrated through two typical examples.In summary, some related means and approaches to solve the problem of fault diagnosis for power device is described in detail in this paper.