Bearing Fault Diagnosis Method Based on Adaptive CYCBD and DARTS
In order to solve the problem of non-obvious fault features of rolling bearing vibration signals due to strong background noise,a bearing fault diagnosis method combining adaptive maximum second-order cyclostationary blind deconvolution(CYCBD)with differentiable architecture search(DARTS)was proposed.Firstly,the fuzzy entropy was used as the fitness function of the whale optimization algorithm to optimize the length of CYCBD filter,and the combination of kurtosis and envelope spectrum peak was used as a step search index to search cycle frequency,so as to realized the adaptive noise reduction of CYCBD algorithm.Then,DARTS algorithm was introduced to achieve the self-construction of the rolling bearing fault recognition model.Finally,the effectiveness of the adaptive CYCBD-DARTS fault diagnosis method in multi-domain strong noise environment is verified by the published experimental dataset of rolling bearings.
万方数据
维普
