Design Of Expert System For Fault Diagnosis Of Aero-engine Turbine Components Based On Knowledge Reduction

In the aviation gas turbine engine,the turbine components bear the greatest mechanical load,aerodynamic load and thermal load.Therefore,various failures and even safety accidents are most likely to occur during the operation of the engine.In order to solve the problem of difficulty in obtaining fault knowledge,reduce the cost of troubleshooting,and improve the health detection and fault diagnosis technology of aeroengines,this paper designs and implements a fault diagnosis expert system for aeroengine turbine components.First,the acquisition of fault knowledge is studied.The knowledge reduction algorithm is used for knowledge acquisition,the discretization matrix is used to perform attribute reduction and value reduction on the discretized expert knowledge base data,and then the rule detection is performed to obtain the fault rule for judgment.Secondly,the construction of the expert system knowledge base is studied.Three rule representation methods are used to represent the fault rules of the expert system for different rule usage scenarios.Then,according to the needs of data processing and inference engine,four relational data tables are designed,namely sample attribute table,decision attribute table,fault rule table and interpretation result table.Thirdly,the inference engine of the aero-engine turbine component fault diagnosis expert system is designed.First,the depth-first traversal matching algorithm based on the rule decision tree is studied to improve the reasoning efficiency of the inference engine;then the production rule fuzzy reasoning is studied,which improves the ability of the inference engine to deal with conditional uncertain facts;finally,the precise reasoning and the The hybrid reasoning algorithm combined with fuzzy reasoning takes into account both the ability to deal with fuzzy problems and the efficiency of reasoning.Finally,the aero-engine fault diagnosis software was developed.According to the previous theoretical research,the various functional modules were implemented using Visual Studio and SQL Server database,and the software was tested using aero-engine model simulation data and engine test data,which proved that the software has high practical value.