| Scroll compressor is an important part of air conditioning equipment and plays a pivotal role in industry and life.Therefore,monitoring and diagnosing the real-time health status of scroll compressor is the basic guarantee for the permanent safety and stable operation of the equipment,and it is of great significance for the equipment to maintain a good working condition.The fault signals of scroll compressor are taken as the research object,fault type identification technology is studied,fault classification model is established by information fusion and machine learning,and fault diagnosis of scroll compressor is realized.The main contents of this paper are as follows:(1)Analyzing the failure reasons of scroll compressor and proposing solutions.The vibration mechanism of scroll compressor is studied and the internal and external factors are analyzed.The internal causes mainly include structural factors,processing factors and operating faults.External factors include pressure vibration caused by suction and exhaust and radial magnetic tension of the motor.According to the internal and external factors,the cause of failure is analyzed in detail,and the best solution is put forward.(2)Designing and constructing the test bench for scroll compressor.To obtain the fault signal of scroll compressor,the vibration testing platform is built.The test bench is divided into two parts,including scroll compressor prototype,piezoelectric sensor,signal conditioning circuit,data acquisition card,frequency converter and computer.Software system based on LabVIEW software development,including PID control module,data acquisition module,data storage module and data analysis module.According to the experimental scheme,the experimental equipment is tested.The vibration signal of the body is collected by the acceleration sensor on the scroll compressor housing and converted into an electrical signal,which is transmitted to the computer by the data acquisition card.Use software system to store data and conduct preliminary analysis and processing.(3)Studying the fault diagnosis of PSO-SVC scroll compressor based on information entropy fusion.Aiming at the instability of vibration signal of scroll compressor and the difficulty of obtaining a large number of fault samples,a fault diagnosis method for scroll compressor based on information entropy fusion and PSO-SVC is proposed to determine the fault type.The information entropy is calculated by extracting the characteristics of vibration signals in the time domain,frequency domain and time-frequency domain.The entropy value is fused into the feature vector of a single evaluation index by factor analysis,and the evaluation index is used as the input of PSO-SVC.The fault classification model of PSO-SVC scroll compressor is established by training.Experimental results show that the PSO-SVC vortex compressor fault method based on information entropy fusion can effectively identify four faults and achieve 94.5% accuracy. |
