Vibration Fault Diagnosis Of Large Steam Turbine Generator Based On Expert System

The development of electrical power industry in China is good during the eleventh five-year plan period. In recent years, lots of large steam turbine generators have been put into operation. They are supporting the people's production and living as well as meet the demand of energy saving and emission reduction. However, vibration always exists as turbine-generator is a rotation device. In particular, with the trend of high capacity and parameter of generating set, the structure of generation unit will more complicated which means the number of shafting span and the length of span will larger. It can be predicted that vibration problem will more frequent and vibration fault will more complicated. Therefore the research on genesic mechanism of vibration fault and the development of vibration fault diagnosis system has becomes an important topic in diagnosing generating devices of power plant.This dissertation focuses on analysis of common vibration faults in turbo-generating sets as well as the expert system which diagnosis these faults. The information of common vibration faults is obtained through investigates and survey in several power plants. The knowledge and experience of diagnosis are learned from specialists and workers on the site. Based on this, the information of vibration faults is classified and analyzed with the considering of failure mechanism. The reason and the solution of the failure can be understood, the features of failure as well as diagnosis rules can be formed. This provides the knowledge of vibration faults diagnosis expert system. With the depth analysis of knowledge base, the organization structure, data structure and management function of the knowledge base in the expert system are designed. Reasoning module is designed by introducing confidence theory and the use of knowledge representation based on rules and hybrid reasoning control strategy. The diagnosis reasoning produces are introduced in detail in this dissertation. An expert system for fault diagnosis is developed by using Delphi 7 and SQL Server 2000. The interface, knowledge base and module for diagnostic reasoning of expert system are developed. Finally, the expert system is tested through the experiment of rub-impact fault and steam flow excited vibration. The diagnosis result from the system is compared with opinions of experts from central laboratory institute. It is proved that correct diagnosis has obtained, expected result has achieved.