Ensemble local mean decomposition has been gradually introduced into mechanical vibration signal processing due to its excellent performance in electroencephalogram signal analysis. However, an unsatisfactory problem is that ensemble local mean decomposition cannot effectively process vibration signals of complex mechanical system due to the constraints of moving average. The process of moving average is time-consuming and inaccurate in complex signal analysis. Therefore, an improved ensemble local mean decomposition method called C-ELMD with modified envelope algorithm based on cubic trigonometric cardinal spline interpolation is proposed in this article. First, the shortcomings in sifting process of ensemble local mean decomposition is discussed and, furthermore, advantages and disadvantages of the common interpolation methods adopted to improve ensemble local mean decomposition are compared. Then, cubic trigonometric cardinal spline interpolation is employed to construct the local mean and envelope curves in a more precise way. In addition, the influence of shape-controlling parameter on envelope estimation accuracy in cubic trigonometric cardinal spline interpolation is also discussed in detail to select an optimal shape-controlling parameter. The effectiveness of cubic trigonometric cardinal spline interpolation for improving the accuracy of ensemble local mean decomposition is demonstrated using a synthetic signal. Finally, the proposed cubic trigonometric cardinal spline interpolation is tested to be effective in gear and bearing fault detection and diagnosis.

集合局部均值分解由于其在脑电图信号分析中的出色表现，已逐渐引入机械振动信号处理。然而，一个不令人满意的问题是，由于移动平均的约束，整体局部均值分解不能有效地处理复杂机械系统的振动信号。在复杂的信号分析中，移动平均的过程既耗时又不准确。因此，本文提出了一种改进的整体均值分解方法，即基于三次三角基数样条插值的改进包络算法C-ELMD。首先，讨论了整体局部均值分解的筛选过程中的缺点，此外，比较了用于改进整体局部均值分解的常用插值方法的优缺点。然后，使用三次三角基数样条插值以更精确的方式构造局部均值和包络曲线。此外，还详细讨论了形状控制参数对三次三角基数样条插值中包络线估计精度的影响，以选择最佳形状控制参数。使用合成信号证明了三次三角基数样条插值对提高整体局部均值分解精度的有效性。最后，对所提出的三次三角基数样条插值进行了测试，以有效地检测齿轮和轴承故障。