Aero-engine Fault Diagnosis Based On The Evidence Theory

Aero-engine,as a high-complexity and high-precision system,makes it very difficult to diagnose and manage its aeronautical problems.In addition,nowadays the international community pays more attention to the safety of spacecraft,making aeroengine fault diagnosis become a hot research topic.Aero-engine gas path failure’s ratio is probable about 80%of all kinds of failures,so doing research for gas path fault diagnosis become the most important thing.Information fusion technology,having been put forward and developed rapidly in recent years,has become a solid theoretical foundation for the fault diagnosis of aero-engine,and is promoting the improvement of the fault diagnosis technologies of various complex systems including aero-engine system.If we consider the engine as a complex information system,its failure modes and fault data characteristics become the sources of information fusion.However,due to the numerous failure modes,various uncertainties and the defects of information fusion algorithms,up to now,we don’t have systematic and feasible standards for fault diagnosis.In this paper,I focus on the research for aero-engine fault diagnosis,so the engine gas-path fault is taken as the importance,and the following work is studied:First of all,doing some research and discussion for aero-engine mathematical model.According to the status when aero-engine working,and some working points of a real engine in various working states,I completed the mathematic model of aeroengine carefully.And through the studying of the relevant parameters,I simulated the fault of aero-engine fans,compressors,combustion chambers,tail vents and other major gas path components,obtaining a series of engine data in various failure modes.Secondly,I deeply studied the information fusion algorithm-the Evidence theory.After it,I tried to introduce the information fusion based on the Evidence theory into the fault diagnosis of aero-engine gas path,established the effective basic framework and model of gas path fault diagnosis,and formed the fault eigenvectors in various fault modes of the aero-engine through handling the fault data of it.And a new method is used to directly calculate the basic reliability distribution required by the Evidence theory in the fault diagnosis process.finally,making data fusion by the Evidence theory fusion rules,I obtained a more satisfactory diagnosis.In the end,the problem of evidence conflict in the process of data fusion is analyzed.Comparing the various methods of conflict resolution,a method of reliability reallocation based on weight of evidence is put forward.Based on that,an effective fault information fusion framework with evidence conflict is established.And through the example of evidence-conflicted gas fault information,I proved the correctness,rationality and reliability of this framework.