| Centrifugal air compressor,using air as working medium,is a common equipment which continuously produces compressed air by its impeller to drive gas rotating at high speed.Due to the advantages of stable,reliable,long life and high gas supply quality and so on,the compressor has been widely applied in the engineering field.In view of the present problems of low accuracy in fault diagnosis and inaccurate classification in intelligent fault diagnosis of centrifugal air compressor,the research contents in this paper are as follows:(1)The basic structure of 2MCL centrifugal air compressor is introduced,and the centrifugal air compressor unit is further analyzed.Secondly,the operation of the centrifugal air compressor unit is described in detail,and the common faults of the centrifugal air compressor are summarized according to the operation status of the unit.Finally,the intelligence of the centrifugal air compressor is designed.The overall framework of the diagnostic system and the functions of each layer of the system are introduced in detail.(2)The design outline of fault diagnosis system for centrifugal air compressor is given,including main contents and technical requirements.The hardware design of the fault diagnosis system is carried out.The overall hardware structure,signal pickup system and alarm subsystem of the system are designed in detail,and the hardware installation layout of the air compressor is given.Then,the overall design of the software function for the system is carried out.The function modules of the software are designed in detail.(3)The dynamic model of the centrifugal air compressor rotor is established,and the fault characteristics of the centrifugal air compressor are extracted.Secondly,the fault diagnosis algorithm of the air compressor is studied.BP neural network algorithm,weighted fuzzy C-means clustering algorithm(WFCMA)and improved particle swarm optimization algorithm(IPOS)are designed respectively.Finally,an improved particle swarm—weighted fuzzy C-means clustering algorithm(IPSO-WFCMA)is proposed that used for fault clustering analysis of the air compressor.(4)The experimental research on fault diagnosis system of centrifugal air compressor is carried out,including simulation experiment and case experiment study.The results show that the IPSO-WFCMA has better performance in the accuracy,convergence speed and consumption time of fault diagnosis,shows stronger clustering effect,IPSO-WFCMA can effectively and accurately identify and diagnose unit faults,and can meet the requirements of centrifugal air compressor fault diagnosis. |
