Compound fault feature extraction of rolling bearing based on parameters adaptive CYCBD
In view of the difficulty to accurately extract and separate the features of the early fault signals of rolling bearings,a compound fault feature extraction method of rolling bearing based on parameters adaptive maximum second-order cyclostationarity blind deconvolution(CYCBD)was proposed.Based on different fault types,the harmonics energy ratio index was used as the fitness function,and the sparrow search algorithm was used to adaptively obtain the optimal filter length and cycle frequency of deconvolution.The obtained optimal parameters combination was used to extract the fault components in the original signal one by one,and the envelope spectrum analysis of the deconvolution signal was carried out to realize the diagnosis of compound fault of the bearing.The analysis results showed that the proposed method can clearly and accurately separate 1-4 times of the inner ring characteristic frequency and 1-6 times harmonic component of the outer ring fault from the measured signal of bearing fault under the background of strong noise,while other common methods can only extract a few fault frequencies with low resolution.The proposed method has obvious diagnostic effect,higher application value and promotion performance.
NETL
NSTL
万方数据
