Vibration signals are the information carrier of roller bearing work.In order to monitor and diagnose running status of inner rolling bearings

vibration acceleration signal was decomposed by using local wave.The parameters of approximate entropy and correlation were extracted.The normal and fault states were quickly distinguished through the approximate entropy and the trends of development based on the local wave decomposition.At the same time

the dominant mode of fault rolling bearing was detected and the fault feature frequency was accurately extracted according to the characteristics of correlation parameters change.As massive measured data are processed and analyzed

the method can accurately diagnose JZQ250 reducer different running states including the normal state

inner ring

outer ring and cage fault operations.The results shows that the method has certain project practical value.