To address the challenge of detecting early faults in rolling bearings, where weak fault features are occasionally obscured by background noise, an innovative early fault diagnosis method based on adaptive minimum noise amplitude deconvolution (MNAD) is introduced. Initially, a correlation Gini index function is defined to estimate the fault period, while leveraging the iterative advantages of MNAD to progressively approach the true fault period. This resolution effectively overcomes the requirement of prior knowledge regarding the fault period, thus reducing the number of input parameters. Subsequently, a composite index, combining square envelope Gini index and square envelope entropy, is constructed as the fitness function. The sand cat swarm optimization algorithm is employed to adaptively determine the optimal noise ratio and filter length for deconvolution, ensuring the acquisition of the finest filtered signal. Ultimately, envelope spectrum analysis is conducted on the filtered signal to extract fault features and facilitate early fault diagnosis. The effectiveness of the proposed method is validated through simulated and experimental data, highlighting its superior feature extraction capabilities and robustness compared to original MNAD and other sophisticated blind deconvolution methods.

中文翻译：

---

自适应最小噪声幅值反褶积及其在滚动轴承早期故障诊断中的应用

为了解决滚动轴承早期故障检测的挑战（其中微弱故障特征偶尔会被背景噪声掩盖），引入了一种基于自适应最小噪声幅度反卷积（MNAD）的创新早期故障诊断方法。首先定义相关基尼指数函数来估计故障周期，同时利用MNAD的迭代优势逐步逼近真实故障周期。该解决方案有效地克服了对故障周期先验知识的要求，从而减少了输入参数的数量。随后，结合方包络基尼指数和方包络熵构建复合指数作为适应度函数。采用沙猫群优化算法自适应确定反卷积的最佳噪声比和滤波器长度，确保获取最精细的滤波信号。最终对滤波后的信号进行包络谱分析，提取故障特征，便于早期故障诊断。通过模拟和实验数据验证了该方法的有效性，突显了与原始 MNAD 和其他复杂的盲反卷积方法相比，其优越的特征提取能力和鲁棒性。