The Research Of Reciprocating Compressor Fault Diagnosis Expert System On Case-Based Reasoning

Reciprocating compressor is a kind of key equipment widely used inprocess industry, so fault diagnosis for reciprocating compressor is veryimportment. Nowadays, the development direction of equipment faultdiagnosis is that expert system based on AI will be used to diagnosefaults,so the reach for reciprocating compressor fault diagnosis expertsystem develops rapidly too. Case-based reasoning(CBR) is a emerging andhighly effective problem solving methods in artificial intelligence, and itcan simulate the thinking model that experts diagnose faults, so CBR canbe used to create fault diagnosis expert system, it is extremely meaningfulto improve the efficiency of the compressor, reduces the failure rate andreducing maintenance costs that CBR is applied to reciprocatingcompressor fault diagnosis expert system.In this thesis, the key technologis of reciprocating compressor faultdiagnosis expert system on CBR are reached based on the structurecharacteristics, working principle and fault types of reciprocatingcompressor, such as case representation based on object-orientedknowledge representation method, case hierarchical structure, weightcomputation method based modified AHP—3-scale AHP and optimizedweighting K-Nearest Neighbor computing method. Also the fusionmechanism and rule extracted from case are reached to fuse CBR and RBR.Finally, based on the research of knowledge representation,reasoningtechnology,functional structure and diagnosis process of reciprocatingcompressor fault diagnosis expert system are designed,and expert system isdevelopmented.