Research On Graphic Recognition Technology In Fault Diagnosis System Under ICMMS

In the modern industrial manufacturing enterprise, the Control, Maintenance and Management isolated from one another. The poor information flows among them has lead to low efficiency and high cost ,so the whole enterprise's input-output ratio remains highly. It is obvious that this isolated mode is becoming more and more unsuitable for enterprise under this dog-eat-dog market condition. To solve this problem, many countries begin to do the industrially level research on Intelligent control-maintenance-technical management system (ICMMS). It realized the integration of control, maintenance and management, and promote the system's efficiency. As the most important and advanced way of maintenance in ICMMS, the predictive maintenance avoids the blindness and high cost of the maintenance activity. High performance of fault diagnosis system is the key to implement the predictive maintenance.Rotating Machinery is one of the common equipments in manufacturing,many rotating machinery like compressor,pump, etc,act as the key equipment of plants . It means the huge economic expense in case the fault occur .many rotating machines'fault symptom displayed as graphic. Different faults correspond to different graphic shapes. So the fault diagnosis can be regarded as the graphic shape recognition , and the recognition result directly infect diagnosis system's reliability.With increasing complexity of the manufacturing system, the large amounts of row data in the diagnosis, the effect of the noise and the random of the sensors are problems in fault diagnosis system. So the real-time and reliability of the fault diagnosis system in ICMMS has declined. Information fusion is a process to check, combine,correlate,and estimate the multi source data. The main idea is optimise and synthetically processing. To enhance the reliability and real-time of the fault diagnosis system,the application of the information fusion to fault diagnosis system is a main research area presently.To solve above problem, the main research in this thesis are listed as below:1,The concept,architecture and development of ICMMS are introduced. Also the necessary and importance of in fault diagnosis in ICMMS are included .Finally the application and the problems of information and pattern recognition technology to fault diagnosis is presented.2,The fault diagnosis,information fusion,pattern recognition technology is introduced,including four parts:(1)fault diagnosis technologies ,means. (2) information fusion theory and means. (3)pattern recognition theory and means.(4)graphic information fusion model of fault diagnosis system in ICMMS.3,Chain code,Fourier descriptors and moment invariant technology used in graphic recognition of fault diagnosis system in ICMMS is presented. A creative application of chain code,Fourier descriptors and moment invariant technology to shaft centre orbit recognition is discussed.4,A decision level fusion of fault diagnosis system in ICMMS is introduced .The decision-making method is based on Sugeno fuzzy integral .finally paper compared recognition result with the measures presented in chapter 3.