Practical framework of Gini index in the application of machinery fault feature extraction

Gini index (GI) is an outstanding sparsity index that has high robustness for the interference of the random impulse noise. Yet, as a new index, the definition of GI in different domains is blurry which brings misdirection and restrictions for its application of machinery fault feature extraction. In view of this, this paper attempts to compensate for the loophole. With the mathematical deduction, a new GI evaluation frame, including the definition of new indexes based on GI, is firstly built. Based on this, enhancing signal processing methods via GI, such as spectrum kurtosis, decomposition methods, and multi-objective optimization algorithms, are designed. In addition, two blind deconvolution methods based on GI and its variants are originally proposed in this paper. Finally, the performance superiority of this application of GI is verified by the numerical simulation and real experimental cases compared with the most popular and the state-of-the-art methods in the field of machinery fault diagnosis.