Abstract Aiming at the problem of impact signal detection under strong noise background, an adaptive stochastic resonance method for enhancement and extraction of gear weak impact Fault Signal is proposed. First, a new modified kurtosis index is constructed by using kurtosis index and correlation coefficient, which is applied as the measurement index of stochastic resonance for the detection of impact signals. Second, a data segmentation algorithm via sliding window is adopted to segment the impact signal with different impact amplitudes into multiple sub-signals with single impact component, which are used as the system input of stochastic resonance. And the genetic algorithm is employed to realize the adaptive selection of system parameters. Finally, the proposed method is applied to gearbox fault diagnosis of traveling unit of electric locomotive. The results show that this method can effectively extract the features of gear fault.
LI Ji-meng,ZHANG Yun-gang,ZHANG Jin-feng, et al. Enhancement and Extraction of Gear Weak Impact Fault Signal Based on an Adaptive Stochastic Resonance[J]. Acta Metrologica Sinica, 2017, 38(5): 602-606.
2017, Vol. 38 
