Research On Composite Fault Diagnosis Based On K-S Matching And Evidence Theory

With the development of modern industrial system and manufacturing equipment gradually toward large-scale,complicated and precise direction,large rotating machinery occupies a pivotal position in the modern industry,and their good performance is directly related to the normal safety of the whole plant production and operation.Increasing scholars are committed to the study of fault diagnosis of large rotating machinery,therefore compound fault has become the focus of fault diagnosis at the present time.Because of the multiple faults,their forms of expression are various,and the characteristics of different faults are mixed,showing the characteristics of multi coupling and ambiguity,which brings great challenge to fault diagnosis.With the gradual development of information technology and artificial intelligence,new technologies continue to be transplanted and applied to mechanical fault diagnosis,which enriches the theory and technology of fault diagnosis,and promotes the development of fault diagnosis.Most of the present research work at home and abroad is focused on fault prediction,fault model design,fault target tracking and so on.On the basis of reading a large number of literature,an integrated diagnosis method based on K-S and evidence theory is used,which is combined with the relevant research background.The main work is as follows:(1)Faults of rotating machinery are usually compound faults in practical conditions,which existing diagnosis methods are difficult to deal with.In this paper,K-S matching calculates similarity of faults,and then the conflict is processed by evidence theory.The result shows that the method can quickly judge common compound faults in the unit.(2)In view of K-S has the advantage in dealing with two sample data,in this paper cumulative distribution function curves of dimensionless data samples are tested by K-S,which can be used to identify the fault.(3)According to the complexity and uncertainty of rotating machinery composite fault,this paper makes full use of advantages that the evidence theory deals with uncertain information.Evidence theory is used to fuse the results of K-S matching.Finally multiple experiment in large units verified that K-S and evidence theory data fusion method can identificate compound fault of large units quickly and accurately.