No abstract

The generalized logarithm sparse regularization method (G-log) for fault diagnosis of rotating devices can effectively reconstruct repetitive transient shocks from noise-disturbed signals, but its reconstruction accuracy frequently becomes inferior due to unsuitable regularization parameters. Moreover, the conventional sparse regularization methods perform nothing on the input signals to guarantee that the impulse characteristics remain constant during the entire iteration process, which exacerbates the influence of noise on the reconstruction accuracy. To overcome these challenges, an adaptive generalized logarithm sparse regularization method (AG-log) based on the second order cyclostationary indicator (ICS2) and the improved maximum correlation Pearson correlation coefficient deconvolution method (IMCPCCD) is proposed in this paper. Firstly, the optimal threshold parameter k for each iteration of AG-log is determined based on the ICS2 criterion to ensure the optimal reconstruction accuracy, while the optimal combination of iterations number N and k is established. Secondly, the original signal and the IMCPCCD filtered signal are alternately used as the input signal of AG-log according to the parity of the iterative steps to reduce the interference of noise. Finally, the application on simulated and two engineering case signals demonstrates AG-log has better reconstruction accuracy compared with the conventional nonconvex sparse regularization methods.

Blind deconvolution (BD) methods applied to bearing fault detection often cause inferior performance due to inaccurate input parameters. Moreover, the optimal parameters of BD vary for different speeds and fault types of bearings, which seriously undermines the applicability of BD in practical industries. In this scenario, this paper proposes a parameter-adaptive BD method (MOBD) based on the multi-objective adaptive guided differential evaluation algorithm (MOAGDE). Firstly, based on the linear discriminant analysis, the quotient of inter-class distance and intra-class distance is used to determine the superiority of common bearing fault characteristic indicators to establish the multi-objective function of MOAGDE. Then, the optimal parameters of BD are searched by MOAGDE improved by dynamic switched crowding method (DSC-MOAGDE). Finally, the bearing is judged whether or what kind of fault has occurred according to the fault type locating index proposed in this paper. The main advantage of MOBD is that only bearing speed and type priories are required to achieve online detection of bearing faults. The results of simulation and experimental signals demonstrate that MOBD significantly outperforms the traditional BD method.