Research On Online Intelligent Fault Diagnosis System

With the development and improvement of science and technology, modernmanufacture become more and more automated, at the same time equipment is moreand more complex, the industrial production process also constantly turn into largescale，complicated and intelligent. When the production process is out of order, that willtrigger not only serious disastrous accidents and economic loss, but also perpetualecological environment pollution. Base on the study of current technology, the thesisstudied online intelligent fault diagnosis method by integrating fault diagnosis methodand expert system. This thesis realized online intelligent fault diagnosis system withfault detection, fault diagnosis, alarm indicator, history query, and other functionsdeveloped by the object-oriented C#.NET language development tools based onthe.NET framework and the C/S mode, and also by SQL Server2005and MATLAB.The design improved the performance of various modules in the system, such as:speeding up the error, improving the accuracy of fault diagnosis, and clearing thedisplay of charts etc.In order to verify the feasibility of the designed system, online intelligent faultdiagnosis system is applied to the fault diagnosis of TE process. The system of TEprocess consists of five major units: reactor, condenser, compressor, separator, andstripper. There are8kinds of ingredients A, B, C, D, E, F, G and H,41measured and12manipulated variables in the TE process. As online intelligent fault diagnosis system isstarted, firstly, the user needs to login the system and the system will opencorresponding function modules to the user according to user level. System collects thedata of TE process through the data collecting module. Then according to thecharacteristics of the system, as far as possible, the data processing time of method isshorter, the result accuracy is higher, the output is simpler to evaluate result. The systemdetect fault with the collected data by C#program based on DPCA and fuzzy c-meansclustering algorithm. And import the detecting information to the knowledge base ofexpert system. According to the information in the knowledge base, the inferencemachine of expert system will diagnose the fault of TE process. After diagnosing thefault of TE process, the system identifies the fault types using tree traversal method. Atlast, the system records the fault diagnosis information to history database，and provides diagnostic information and advice for users. The thesis proves the feasibility of thesystem by the design of online intelligent fault diagnosis system applied to faultdiagnosis of TE process.