A gear fault diagnosis method based on maximum CGGI deconvolution filtering

Under complex conditions such as varying operating conditions and strong noise interference, fault features in gear transmission systems are highly susceptible to being masked by noise, posing significant challenges to fault diagnosis techniques. To address the limitations of traditional blind deconvolution objective functions (OF), this paper proposes a novel blind deconvolution method based on maximizing the Cycle-embedded Generalized Gini Index (CGGI). The core innovation of this method lies in constructing an OF grounded in a new sparsity measure index, CGGI. By introducing a Generalized Weighted Envelope Spectrum (GWE), it effectively integrates the feature enhancement effects of signals under different transformation parameters. Furthermore, it inherits the superior capability of the Generalized Gini Index in quantifying signal impulsivity and cyclostationarity, thereby significantly improving the robustness of fault feature identification under complex operating conditions. Numerical analysis demonstrates that CGGI more effectively suppresses the influence of interference components compared to traditional indicators. Validations conducted on a public dataset of a two-stage parallel-shaft gearbox and a fault test bench for a high-speed train’s body-suspended transmission system show that the proposed method significantly outperforms traditional methods in terms of the Fault Characteristic Ratio (FCR) across various fault scenarios. This proves the effectiveness and superiority of the proposed method for condition monitoring and fault diagnosis of complex transmission systems in practical engineering applications.