Research On Steam Turbine Intelligent Fault Diagnosis And Maintenance Decision System

Turbine fault diagnosis is a complex system engineering with multi-discipline and multi-technology integration,which has the characteristics of variable operating conditions,coupling among parameters,and complex fault transmission and evolution law.In this paper,based on in-depth analysis of typical turbine fault modes,combined with knowledge engineering and deep learning theories,the key technologies of intelligent fault diagnosis for steam turbines are studied.Firstly,a method of fault knowledge analysis and acquisition was proposed to solve the problem of complex correlation between turbine faults.Based on the system engineering theory,the steam turbine equipment was divided into different levels.Based on the combination of fault tree analysis method and fault mode and influence analysis method,the knowledge related to common turbine faults were systematically analyzed and sorted out,and the unified digital coding was carried out.Secondly,in view of the lack of unified semantics of fault knowledge,the research on structured representation of fault knowledge based on ontology theory was carried out.On the basis of the acquired knowledge of turbine faults,the division principle of class,attribute and individual in fault knowledge ontology was put forward,and the construction of steam turbine fault semantic network was realized..At the same time.in order to further manage and store fault knowledge,a visual display technology of fault knowledge graph based on Neo4j was studied.Thirdly,in order to solve the coupling problem among state parameters of steam turbine,the research of fault early warning based on long short-term memory neural network and ontology semantic search was carried out.By selecting characteristic parameters and using the steady-state data samples after cleaning,the prediction model of long and short term memory neural network is established.By comparing the actual values with the predicted values of the model,the abnormal parameters were located.Furthermore,the semantic search based on fault knowledge ontology was combined to realize the fast matching from abnormal parameters to fault modes.Finally,based on the above theory research,driven by the demands of practical engineering,the turbine intelligent diagnosis system framework was designed,described the structure and function module of the system database analysis process,the steam turbine prototype intelligent fault diagnosis system was developed,promote the engineering application of the research on steam turbine intelligent operation and maintenance.