Failure Propagation Modeling And Diagnosis Early Warning Method Of Flue Gas Turbine

Flue gas turbine is the key equipment of catalytic cracking unit in refining and chemical enterprises.Due to the complexity of the flue gas turbine structure,severe operation condition,high complexity of its fault and strong coupling etc.Once a component or a subsystem breaks down,it will lead to chain reaction of other components or subsystems.And then it may result in failures of the entire unit.It is difficult for traditional fault diagnosis methods to eliminate the root causes of failure,because the coupling between components was not fully considered.This thesis focuses on the research of fault propagation and early warning modeling methods based on the analysis of correlation between flue gas turbine subsystems and components.It achieved the purpose of predicting fault propagation path and the degradation trend of components,and provided basis for equipment maintenance and safe operation.The main research contents are as follows.Flue gas turbine FMEA and FTA analysis were studied because the structure of flue gas turbine is complex and the types of parameters are variety,and there is a high degree of correlation and coupling between failures.The fault tree of flue gas turbine was established and all the fault sources of flue gas turbine were determined.A flue gas turbine modeling method based on Fuzzy Fault Petri net(FFPN)was proposed aiming at solving the problem of fault modeling of complex coupling system of mechanical equipment.The quatitative analysis of fault propagation path of flue gas turbine and the accurate positioning of fault source were realized based on forward and backward reasoning.In the case study,the fault propagation path obtained by forward reasoning was the catalyst fouling?the erosion/wear of the blade fatigue fracture?the blade fault?the rotor vibration was too large?the flue gas turbine fault.By backward reasoning,the cause of the flue gas turbine failure was the scaling of catalyst.The result is consistent with that of forward reasoning,which showed the effectiveness and practicability of this method.A fault diagnosis and early warning modeling method of flue gas turbine based on Dynamic Bayesian network(DBN)was proposed to solve the problem of correlation and coupling between flue gas turbine faults.Based on the FMEA and Petri net analysis,the degradation effect of flue gas turbine components was studied.The fault diagnosis reasoning was carried out and the degradation trend was predicted and the early warning function was realized.In the practical application,the abnormal state of flue gas turbine was analyzed,and the root cause factor of high temperature and amplitude of the bearing and high amplitude of the rotor was the reduction of bearing accuracy.The degradation trend of the rotor and the blade was predicted and the correlation between the degree of degradation and composition and the function of component was obtained.And the purpose of fault early warning was realized.