Research And Realization Of The Frequency Intelligent Analysis System For Rotating Machinery

With the development of computer and fault diagnostic technology, all kinds of fault diagnostic software put into service. The Techniques of Digital Signal Processing are used in the system of equipment monitoring and fault diagnosis, but it is so hard for the customers and field monitoring persons that an intelligent analysis system can give the diagnosis results but only need some simple settings is urgent need for them. On the other hand there is plenty of information in the frequency spectrum. The frequencies and amplitudes of different faults are different. The amplitudes and proportion of the spectral eigenvector reflect corresponding faults. Because the frequency spectrum is the primary source of the rotating machinery's fault information, it is the data source to many analytical methods.Rotating machinery's usual faults and the characteristics are described. It will find that the spectrum image can be divided into the surrounded domain, the background domain, the spectrum boundary and the spectrum curve, the colors of which are so different that they contrast, so we can extract features directly from the spectrum take advantage of image processing techniques by means of these colors. Then the fault symptoms are got.Both expert system and neural network have their own relative merits, which can not meet the need of complex diagnostic system. To this question, a neural network fault diagnosis expert system combine expert system based on symbol reasoning with neural network based on numerical calculation is designed, which meet the intelligent need very well.Under the Windows XP system, taking advantage of object-oriented Java technology and powerful relational Oracle database, design and implement spectrum feature extraction, intelligent analysis and user interface. The system will give the analysis results that tell you if the equipment operating normally, which fault if it has any, possible causes and what methods should be taken.This system has application in frequency analysis of YanShan Petrol-Chemical Co.' monitoring reports, and it can make the quite accurate distinguish to the typical faults of rotating machineries. It has been proved practical and effective.