College of Railway Transportation, Hunan University of Technology, Zhuzhou 412007, China.
Hunan Provincial Key Laboratory of Health Maintenance for Mechanical Equipment, Hunan University of Science and Technology, Xiangtan 411201, China.
Sensors (Basel). 2022 Sep 1;22(17):6599. doi: 10.3390/s22176599.
Complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) effectively separates the fault vibration signals of rolling bearings and improves the diagnosis of rolling bearing faults. However, CEEMDAN has high memory requirements and low computational efficiency. In each iteration of CEEMDAN, fault vibration signals are added with noises, both the vibration signals added with noises and the added noises are decomposed with classical empirical mode decomposition (EMD). This paper proposes a rolling bearing fault diagnosis method that combines piecewise aggregate approximation (PAA) with CEEMDAN. PAA enables CEEMDAN to decompose long signals and to achieve enhanced diagnosis. In particular, the method first yields the vibration envelope using bandpass filtering and demodulation, then compresses the envelope using PAA, and finally decomposes the compressed signal with CEEMDAN. Test data verification results show that the proposed method is more effective and more efficient than CEEMDAN.
完全集合经验模态分解自适应噪声（CEEMDAN）有效地分离了滚动轴承的故障振动信号，并提高了滚动轴承故障的诊断能力。然而，CEEMDAN 具有较高的内存要求和较低的计算效率。在 CEEMDAN 的每次迭代中，故障振动信号都会添加噪声，添加噪声后的振动信号和添加的噪声都会用经典的经验模态分解（EMD）进行分解。本文提出了一种将分段聚合近似（PAA）与 CEEMDAN 相结合的滚动轴承故障诊断方法。PAA 使 CEEMDAN 能够分解长信号，从而实现增强的诊断。具体来说，该方法首先使用带通滤波和解调得到振动包络，然后使用 PAA 对包络进行压缩，最后用 CEEMDAN 对压缩信号进行分解。测试数据验证结果表明，与 CEEMDAN 相比，该方法更有效、更高效。
