Fault Diagnosis Technology Of Turbine Generator Sets Based On Ontology And Signal Analysis

With the rapid development of electric power industry in China,the structure of turbine generator sets becomes more and more complex,and there are more failures in operation.In order to ensure the long-term safe,stable and efficient operation of the sets equipment,it is necessary to carry out accurate and reliable fault diagnosis.In view of the problem of multi-source and heterogeneous fault knowledge of turbine generator sets,which is easy to form "knowledge islands",ontology can be better applied to the knowledge representation in the field of fault diagnosis.At the same time,there are a lot of data containing the operation status information in the operation of the sets equipment.In view of this,it puts forward a fault diagnosis method based on ontology and signal analysis.and the method is verified and analyzed by an example.The main research contents and work of this thesis are summarized as follows:(1)The knowledge base of fault diagnosis for turbine generator sets based on ontology is constructed.In view of the problems such as multi-source and heterogeneous fault knowledge,difficulty in knowledge sharing and reuse,lack of reasoning ability and flexibility,the method based on ontology is considered to better represent the fault knowledge of turbine generator sets.The traditional "skeleton method" and "seven step method" are combined and improved.Based on this,the fault diagnosis ontology knowledge base of turbine generator sets is constructed to ensure the fault diagnosis the consistency of diagnosis ontology in knowledge structure is convenient for knowledge sharing and reuse,and it also lays a foundation for fault reasoning based on ontology.(2)A signal analysis method based on EEMD,permutation entropy and SVM is designed.In view of the complex and changeable working environment of the turbo-generator set,the interference noise is relatively large,and the vibration signals of the unit mostly have non-stationary,non-linear,strong noise characteristics etc,combining the EEMD and arrangement entropy methods to extract the fault characteristics of the vibration signal,according to the fault mechanism and correlation analysis of the turbine generator sets,the sensitive components in the fault characteristics are selected to form an effective frequency doubling component,and the remaining false components are reconstructed to calculate the arrangement entropy of each frequency band as the feature vector,supplemented by the PSO-SVM algorithm for fault identification,Improve the accuracy and identification effect of fault diagnosis.(3)A fault diagnosis method for turbine generator sets based on ontology and signal analysis is proposed.A semantic mapping is designed based on the concept name similarity algorithm,which correlates the results of the signal analysis with the instances in the ontology knowledge base,and infers the ontology instances obtained from the association,thereby obtaining the fault location,fault cause and corresponding Maintenance strategy,etc.A typical rotor fault of a turbine generator sets was simulated by a rotor simulation test bed.The results show that the whole process of turbine generator sets fault diagnosis in "data acquisition,feature extraction,fault identification,fault reasoning and fault resolution" is realized.(4)An intelligent fault diagnosis system for turbine generator sets based on ontology and signal analysis was developed.Protégé,SQL,MTALAB and C# are used to jointly develop the intelligent fault diagnosis system of the turbine generator sets,including the turbine generator sets condition monitoring module,fault knowledge processing module,fault signal analysis module,database management module and human-computer interaction interface.The example verification shows that the system can accurately and effectively realize the fault diagnosis of the turbine generator sets,and the fault diagnosis is more intelligent.