In view of the common signal interference in the fault analysis of sludge return pump

the MEEMD and ICA methods are proposed for fault diagnosis and analysis. The vibration signal is decomposed by MEEMD to obtain a series of IMF components. By calculating the correlation coefficient between each component and the original signal

the corresponding components are selected for reconstruction. The corresponding components are formed into a matrix with the original signal for ICA calculation. By comparing the characteristic frequency of normal operation and failure operation before and after MEEMD and ICA treatment

the fault analysis and diagnosis of sludge reflux pump are carried out. The experimental results show that the MEEMD and ICA methods can reduce the influence of noise and effectively carry out fault diagnosis.