Diagnosis Method for Sparse Optimization Based on Multi-step Grid Search of Bearing Coupling Faults

Abstract

Under the unbalance-bearing coupling fault of the rotor system, the unbalance fault features and bearing faults modulate each other, and the weak fault features are easily submerged by the strong fault features and strong noise components. Thus, a multi-step grid search optimization sparse diagnosis method is proposed to improve the anti-interference ability of sparse representation algorithm by selecting the optimal parameters of sparse representation to reduce the sparse reconstruction error. Firstly, based on the multi-step grid search theory, a multi-step grid search optimization sparse representation method model is established to improve the calculation accuracy. Secondly, the signal reconstruction based on adaptive Gabor atom dictionary orthogonal matching tracking algorithm is realized by optimizing the displacement factor and frequency factor. Finally, the coupling simulation fault data with different SNR are analyzed. The results show that the anti-interference ability of the proposed method increases with the increase of noise interference and different experimental cases further verify the accuracy and applicability of the proposed method.

Keywords

• Rotor system;Coupling fault diagnosis;Sparse representation;Grid search;Parameter optimization