A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Research on rotor fault identification based on the integration of symbolic entropy and evidence theory
A rotor fault identification method integrating symbolic entropy and evidence theory was proposed for the problem of the difficulty to accurately identify the fault state caused by the nonlinearity and nonstationarity of the rotor vibration signal. Firstly, the fluctuation sequence was obtained by processing the recorded data of the rotor system by de-averaging and difference. Secondly the binary sequence was obtained by binary processing. Thirdly the symbol sequence was obtained by encoding to calculate the symbol entropy. In this paper, the code length in the calculation of symbol entropy was determined by an indicator of the relative growth rate of the mean value of symbol entropy. Through experimental analysis, the symbol entropy of the four channels of the rotor was used as the identification vector, and then compared with the identification standard vector obtained from the historical data to obtain the recognition probability of each state, and finally the results of two tests were fused through D-S evidence theory to obtain the final discriminatory result. Compared with the traditional fault identification method based on symbolic entropy, the proposed method can accurately identify the common states of the rotor system, providing a solution to the rotor fault state identification problem when the vibration signals are in nonlinear and nonstationary conditions.
Research on rotor fault identification based on the integration of symbolic entropy and evidence theory
A rotor fault identification method integrating symbolic entropy and evidence theory was proposed for the problem of the difficulty to accurately identify the fault state caused by the nonlinearity and nonstationarity of the rotor vibration signal. Firstly, the fluctuation sequence was obtained by processing the recorded data of the rotor system by de-averaging and difference. Secondly the binary sequence was obtained by binary processing. Thirdly the symbol sequence was obtained by encoding to calculate the symbol entropy. In this paper, the code length in the calculation of symbol entropy was determined by an indicator of the relative growth rate of the mean value of symbol entropy. Through experimental analysis, the symbol entropy of the four channels of the rotor was used as the identification vector, and then compared with the identification standard vector obtained from the historical data to obtain the recognition probability of each state, and finally the results of two tests were fused through D-S evidence theory to obtain the final discriminatory result. Compared with the traditional fault identification method based on symbolic entropy, the proposed method can accurately identify the common states of the rotor system, providing a solution to the rotor fault state identification problem when the vibration signals are in nonlinear and nonstationary conditions.
Research on rotor fault identification based on the integration of symbolic entropy and evidence theory
SHAN Song (author) / CHEN HongMing (author) / LIU JingXuan (author) / DENG ShuaiQin (author) / SU Yu (author)
2024
Article (Journal)
Electronic Resource
Unknown
Metadata by DOAJ is licensed under CC BY-SA 1.0
Hybrid Model Based Identification of Local Rubbing Fault in Rotor Systems
| British Library Online Contents | 2005
A non-parametric spatial independence test using symbolic entropy
| Online Contents | 2010
| DOAJ | 2016