Study of fault diagnosis of rolling bearings based on the largest Lyapunov exponent

The largest Lyapunov exponent was proposed in order to describe quantitatively the non-linear behavior of rolling bearings,and to study the criterion of fault impairment grade.Noise reduction processing was done for the original signal using envelope demodulation method in 0～1?000,2?000～3?000 and 3?000～4?500 Hz frequency bands respectively,the relationship between Lyapunov exponent and bearing's fault impairment grade was analyzed,and the effect of signal noise on Lyapunov exponent was expounded.The experimental results show that in 0～1?000 Hz frequency band,Lyapunov exponent is 0.612 under the state of incipient fault of the bearing.After running five months,the Lyapunov exponent is up to 0.810?5.Moreover,in 2?000～3?000 Hz and 3?000～4?500 Hz frequency bands,Lyapunov exponent increases gradually.So the largest Lyapunov exponent can be used as the characteristics for recognizing rolling bearing's fault impairment grade.