Research On Fault Diagnosis Of Information Exergy Of Rotating Machinery Based On Information Fusion Of Process

This work was supported by the National Natural Science Foundation of China (NSFC), Project No.50775083. This paper pointed out the shortcoming of the state-based fault diagnosis methods and proposed a new diagnosis idea based on information fusion of process. On this basis, a new fault diagnosis system of information exergy of rotating machinery based on information fusion of process was first established. In this system., the systemic research on fault identification, fault orientation and fault intensity of vibration signals were carried out.Firstly, according to the needs of research, the fault simulation test project of rotor system was designed, and the vibration signals of eight kinds of typical faults during speed-up were collected from the rotor test platform. These data provided strong support for research in this paper. Through comparing with existing faults symptom in literature, not only the correctness of the simulated fault signals was verified, but also some new faults symptom and change rules of some faults during speed-up were found.Secondly, the shortcoming of the state-based fault diagnosis methods was pointed out and a new diagnosis idea based on information fusion of process was proposed. A dual-channel multi-speed information entropy matrix was constructed by a single shaft vibration signal and a single bearing vibration signal which can best reflect the process change rule of each fault vibration signal. Then the information entropy sample matrix of eight kinds of typical faults in time domain, frequency domain and time-frequency domain were calculated respectively.Thirdly, a new concept of information exergy based on information fusion of process was first proposed and defined. The process change rule was quantitatively characterized through establishing the fault diagnosis method of information exergy of rotating machinery vibration. Then the information exergy sample matrix of eight kinds of typical faults in time domain, frequency domain and time-frequency domain were calculated respectively. In the calculation example, for the shortcoming of the diagnosis method, the fault diagnosis method of information exergy based on space close degree of information exergy was established.Fourthly, from another perspective, the process change rule between two corresponding states in the difference process was also defined as a new information exergy. Because the amplitude spectrum of fault signals during speed-up can best reflect the process change rule, a new concept of space-time feature spectrum in frequency domain was defined and then it was used as eigenvalue. On this basis, the fault diagnosis method of information exergy based on space-time feature spectrum in frequency domain was established and the information exergy sample matrix of eight kinds of typical faults based on space-time feature spectrum in frequency domain was constructed respectively. The results show that the method had strong ability of fault classification, and is an effective new diagnosis method of rotating machinery based on information fusion of process.Finally, a new fault diagnosis system of information exergy of rotating machinery based on information fusion of process was first established. In this system, the systemic research on fault identification, fault orientation and fault intensity of vibration signals were carried out. Its diagnosis approach was:First, the fault type of vibration signal was identified by fault diagnosis method of information exergy of dual-sensor based on space-time feature spectrum in frequency domain. Then the fault location of vibration signal was confirmed by fault orientation method of information exergy of multi-sensor based on space-time feature spectrum in frequency domain. Finally, the fault intensity of vibration signal was confirmed by the original waveform data during speed-up. Analysis showed that the diagnosis system was a powerful, high diagnostic accuracy, strong adaptability new model of fault diagnosis of rotating machinery.