Algorithm Realization For Fault Diagnosis Of Compressor Based On Graph Theory

Multistage reciprocating compressor is usually through the pulley drive to produce thecrankshaft to rotate by motor, the connecting rod drive piston to produce reciprocating motion.To reach the purpose of compressed gas, space volume which is made up of the cylinder,piston top surface and valve generate periodic change. Multistage compressor is a kind ofcomplex electromechanical system, it is very difficult to diagnose system fault because ofcylinder state correlative, mutual coupling.Fault diagnosis based on graph theory is an important branch in the field of faultdiagnosis, which has lots of applications and development prospects. In this dissertation, threereciprocating compressor is studied systematically and deeply, based on the graph theory.There are two kinds of fault diagnosis methods based on graph theory: failure propagationdirected graph and signed directed graph. It Points out the deficiencies of diagnosis methodwhich is based on the communication directed graph and elaborates the fault diagnosismethod which based on symbolic directed graph. Because of the low efficiency of originalalgorithm, it’s not suitable for the large scale and complicated system. The SDG and expertsystem are combined in this dissertation.The SDG model of compressor and the human-computer interaction interface are twomain aspects of this dissertation. In SDG modeling aspect, in view of the reciprocatingcompressor, a comprehensive SDG framework of fault diagnosis is proposed. The frameworkis based on empirical knowledge modeling method and involved classification SDG modelingthoughts. The C++software is used to realize the fault diagnosis system by the platform ofVisual Studio2005and SQL Server2000.There are a lot of advantages of combining fault diagnosis Based on SDG with expertsystem. it is a kind of prospecting fault diagnosis method.