| Centrifugal air compressor belongs to the air compressor,it is diffcult to rapidly and precisely find and recognize the fault by manual due to the intricate inner structure and plentiful fault forms.In this paper,a fault diagnosis algorithm based on support vector machine and type-2 fuzzy is proposed to resolve the inefficiency and low precision issue in fault diagnosis of centrifugal air compressor.Details of the research are as follows:(1)This paper introduces the structure and mechanism of air compressor,and compares the strengths and weaknesses of common fault diagnosis methods,outlines the usual faults of centrifugal air compressor and their reasons,devises the faults into modules,and establishes the corresponding fault tree model.(2)Using gray Wolf algorithm optimizes the parameters of support vector machine,and on this basis,an improved directed acyclic graph support vector machine algorithm(IDAG-SVM)based on modified consine similarity is devised to solve the systemlevel fault identification of centrifugal air compressors problem,and compares with other multiple SVM classification algorithm,through the emulational experimental result confirmed the justifiability of the algorithm.(3)Aiming at the problems of the traditional fuzzy reasoning system,such as the lack of fuzziness,the difficulty in designing membership function and the regular explosion caused by too high dimension,an interval two fuzzy fault diagnosis system is designed,and on this basis a centrifugal air compressor fault diagnosis module is built.(4)Aiming at the actual characteristics of centrifugal air compressor fault diagnosis,combining IDAG-SVM system and interval two-type fuzzy system,a kind of centrifugal air compressor fault diagnosis algorithm based on based on support vector machine and type-2 fuzzy is proposed.A ZH-type centrifugal air compressor was taken as the experimental object,and the feasibility of the designed system was verified by simulation. |
PDF Full Text Request