Spectrum Fusion Feature Extraction Algorithm for Impulse Fault Signal
When the blind deconvolution method is used to extract the feature of fault signal in time domain,often appear too many signal results to confuse separation results.However,in the previous research,only the separation part is emphasized,and the separated signal is rarely processed further,which makes the practical application inconvenient.On the basis of blind deconvo-lution and spectrum fusion,this paper uses the kernel improved fuzzy c-means clustering algorithm,aiming at the impulse char-acteristics of mechanical fault signal,and proposes a practical algorithm for impulse fault signal processing.The effectiveness of the algorithm is verified by computer simulation.This algorithm optimizes the previous clustering and screening methods,which can effectively eliminate the interference of many useless signals after deconvolution,extract the features of fault pulse signals ac-curately,and improve the efficiency of fault diagnosis.
NETL
NSTL
万方数据
维普
