Research On Key Technology Of Condition Trend Prediction And Fault Diagnosis Expert System For Rotating Machinery

Rotating machinery is the core equipment of industrial sector. The rotating machinery is of large scale and intelligent. The structure is becoming increasingly complex, which made the operating status more attention has been paid. Condition prediction and fault diagnosis for rotating machinery can improve the safety and reliability of the rotating machinery, lessen maintenance crew time, improve crew utilization, reduce power generation costs and avoid catastrophic accidents. Moreover, it is the preliminary research on the development of fault prediction expert system.The full vector spectrum technology merges multi-channel data, overcome the weak point of incomplete single-channel data and extracted the vibration of the rotor accurately and comprehensively, which truly reflects the operational status of equipment. This paper researches the crucial technique for rotating machinery trend forecasting and fault diagnosis which is two modules in the rotating machinery online monitoring and fault diagnosis system,the vector spectrum technology was applied to the rotating machinery condition trend prediction and fault diagnosis, the fault diagnosis expert system has been exploited, as well. The main research is as follows:1.The basic theory, numerical algorithms, the compatibility traditional analysis techniques of vector spectrum technology is expounded. The instances of vibration data of steam turbine set are analyzed, which confirms the superiority of the vector spectrum technology in fault diagnosis.2.A new method to predict the spectrum of basing on full vector support vector regression(FVSVR) has been put forward. The operating data of the steam turbine unit has been case analyzed, which certificates the utility of the spectrum forecasting of full vector support vector regression(FVSVR). The picking of the relevant parameters has been investigated and the traditional parameter optimization method has been improved.3.Fault diagnosis approach basing on full vector fuzzy transformation and epistemology has been studied. The indication, extraction, treatment of the sign of the rotating machinery fault diagnosis expert system has been defined, as well as the representation and inference control strategy of knowledge. The full vector spectrum of amplitude frequencies has been extracted and blurerd as symptom vector. The failure assumptions set is obtained by the means of fault diagnosis method based on the fuzzy transformation. And in the same way, we can take advantage of knowledge-based diagnostic approach to backward reasoning for the malfunction of assuming collection so that the conclusions of failure can be obtained.4.The rotating machinery fault diagnosis expert system has been developed. We have designed the general structure of the system, the construction of the database and the form of diagnostic reports, as well as the method of establishing and maintain the failure libraries. Expert system development process using EXSYS in building expert system software, simplifies the process of the establishment of knowledge base and reasoning machine and maintenance.