Fault detection of the rotating machines through optimized orthogonal matching pursuit by golden jackal optimization

Ensuring the reliability and functionality of industrial machinery necessitates an effective fault diagnosis. However, the identification of fault signatures is complicated by significant environmental noise and other interference. This paper presents a technique employing the golden jackal optimization (GJO) algorithm to optimize the parameters of the orthogonal matching pursuit (OMP) method, which operates in conjunction with a Gabor dictionary to extract fault signatures from signals. Initially, a dynamic Gabor atom dictionary was developed to extract sparsity from the signal by adapting to the residual signals. Subsequently, GJO was integrated with OMP to fine-tune its parameters alongside the Gabor dictionary atoms, with the objective of approximating the original signal and enhancing the signal sparsity. The proposed method is applied to detect defects in faulty bearings and the two-stage gearbox of a belt conveyor drive. Comparisons with existing methods demonstrated their superiority in extracting fault features.