Maximum multi-domain nonlinear Gini index deconvolution and its application to early fault diagnosis of planetary gearboxes

Blind deconvolution is an effective technique to recover fault-related periodic impulses within vibration signals. Early-stage faults in planetary gearboxes exhibit weak features that are easily obscured by background noise and other interferences. However, existing blind deconvolution methods relying on traditional sparsity measures struggle with performance degradation under complex interference conditions. To solve this problem, a maximum multi-domain nonlinear Gini index deconvolution method is proposed in this paper to precisely extract the weak fault features from the affected signals of the planetary gearboxes. Firstly, the advanced nonlinear Gini index is used to quantify both impulsivity and cyclostationarity in time and frequency domains, and a multi-domain nonlinear Gini index is constructed as the deconvolution objective function through geometric mean, without requiring prior period. Secondly, the sand cat swarm optimization algorithm combined with the generalized spherical coordinate transformation is employed to solve the optimal filter, sidestepping suboptimal solutions and securing the best possible filtered output. Finally, the filtered signal undergoes a squared envelope spectrum analysis to uncover the fault features, thereby achieving early fault diagnosis. The effectiveness of the proposed method is verified using simulated and experimental data, and the results show that the proposed method can achieve accurate diagnosis of early faults in planetary gearboxes, revealing superior performance in extracting weak fault features and suppressing interferences over other blind deconvolution methods.