Jixie chuandong (Jan 2018)
Bearing Fault Diagnosis Method based on Self-adaptive Stochastic Resonance of Genetic Algorithm and VMD
Abstract
The stochastic resonance（ SR） needs to meet the approximative condition when dealing with the measured bearing fault signal,it needs to meet small signal parameters. That is to meet the small parameter signal（ signal amplitude,signal frequency,noise intensity is far less than 1）. The detection of measured vibration signal is greatly restricted by this problem. Aiming at this phenomenon,the self-adaptive re-scaling stochastic resonance based on genetic algorithm（ GA） and the Variational Mode Decomposition（ VMD） is proposed. Firstly,the compression ratio R is set to compress the measured signal appropriately and it needs to meet the small parameter condition. And then,the signal-to-noise ratio of output signal is defined as the objective function. By using the genetic algorithm,the a and b that are structure parameter of re-scaling stochastic resonance are optimized,and the optimal value is selected to drag-in the re-scaling stochastic resonance,the de-noising of measured signal is carried out,Finally,the de-noising signal is decomposed by VMD,from the component spectrum diagram of each IMF,the bearing fault characteristic frequency can be found. Based on the experimental data,the results show that the method can effectively improve the accuracy of bearing fault diagnosis.
