Early Fault Diagnosis of Gearbox Based on Autoregression and EWMA
Aiming at the problem that traditional gearbox monitoring is difficult to find fault in time,the early abnormal detec-tion of gearbox based on autoregressive model and statistical process control is studied.Time synchronous averaging(TSA)algo-rithm is used to remove the noise in the original vibration signal,then the autoregressive model(AR)is established and the re-siduals are obtained.The standard deviation,kurtosis and root mean square of the residuals are extracted to preliminarily judge the early fault of the gear.Based on the statistical analysis of residual standard deviation,the mean control chart,single value moving range difference control chart and exponential weighted moving average(EWMA)control chart are established respec-tively.The result shows that EWMA control chart detected the 62nd file began to exceed the control limit,and the other two con-trol charts began to appear abnormal from the 65th file.The EWMA control chart detected abnormal points earlier,indicating that the combination of AR model and EWMA control chart can judge the early abnormal points more effectively.
Time Synchronous AveragingGearboxEarly Fault DetectionAutoregressive ModelStatistical Process Control
NETL
NSTL
万方数据
