| Steam turbine generator sets,a large-scale rotating machine,as a key equipment in the power system,has high requirements for safety,stability and lifetime.It will cause personal injury and even serious economic losses when an unexpected failure occurred.Therefore,in order to improve the safety and reliability of the operation of the steam turbine,the accurate diagnosis and prediction of the failure of the steam turbine play an important role in engineering practice.The design and development of a condition monitoring and intelligent fault diagnosis system is the main means to ensure the smooth operation of the unit without shutdown,which is helpful for the technical personnel to monitor and maintain the faults of the unit.A health condition monitoring and intelligent fault diagnosis system for steam turbine generator sets is developed based on UML system modeling method,which multi-source heterogeneous ontology knowledge representation method and related case-based reasoning mechanism are proposed.The construction of system knowledge base and the design of reasoning engine are intensive studied in the thesis.The main contents and contributions are summarized as follows:(1)A condition monitoring and intelligent fault diagnosis system model based on UML systematic modeling method is designed.In view of the problems of multi functions,complex structure and long development cycle of fault diagnosis system for steam turbine generator sets,and considering that UML modeling method has strong expansibility,high universality and short development cycle,the condition monitoring and intelligent fault diagnosis system model of steam turbine generator sets is constructed,which included the overall system architecture,system use case model,function decomposition model,system static class model and system state model,system interaction model and component deployment model.(2)A multi-source and heterogeneous ontology modeling and knowledge fusion method for steam turbine generator sets is proposed.The global and local ontologies of steam turbine generator sets are constructed using Protégé software.The method and steps of modeling are explained in detail.The algorithm of knowledge fusion and the process of multi-source knowledge retrieval are improved.The feasibility of ontology knowledge model is verified through the retrieval process of multi-source knowledge.Finally,different faults of steam turbine generator rotor are simulated by MFS.The feasibility and effectiveness of knowledge fusion algorithm are proved.(3)The reasoning method is designed based on ontology and related case-based reasoning.The reasoning engine Fa CT ++ of Protégé software is used to carry out reasoning first,and then the case hierarchical retrieval is carried out according to the results of ontology reasoning,which solves the problems of incomplete ontology reasoning results and low reasoning efficiency.The inference engine of fault diagnosis system is designed combining the global similarity and local similarity algorithm of case retrieval.The accuracy and efficiency of the system diagnosis are ensured,and the reasonable fault solution is given.The whole process from “fault attributes input” to “solution output” is realized.(4)The software system of condition monitoring and intelligent fault diagnosis for steam turbine generator sets is developed.The construction of knowledge base and preliminary reasoning are completed by ontology editor.SQL Server is used to store the data and long-term accumulated fault cases from the unit.MATLAB is responsible for the encapsulation of relevant mapping algorithm,feature extraction algorithm and reasoning algorithm.The health condition monitoring and intelligent fault diagnosis system of steam turbine generator sets is developed combined with UML system model.The system provides a feasible solution for fault diagnosis of steam turbine generator sets by systematic performance debugging,which ensure healthy operation of the unit. |
PDF Full Text Request