断层(地质)
方位(导航)
约束(计算机辅助设计)
计算机科学
机械工程
地质学
人工智能
工程类
地震学
作者
J.M. Guo,Tianyao Zhang,Kunlin Xue,Jiehui Liu,jie Wu,Yadong Zhao
标识
DOI:10.1088/1361-6501/ad962d
摘要
Abstract Variational mode decomposition (VMD) is widely used in fault-bearing vibration-signal processing. Nonetheless, VMD remains a challenging task because of the difficulty in finding the optimal combination of parameters and excessive fault information in the residual term. The optimal parameter combination plays a balancing role in the optimization process, controlling the error between the reconstructed signal and the original signal while suppressing interference between modes. To address these defects, a parameter-adaptive re-constrained VMD method based on a subtraction average-based optimizer (SABO) is proposed. In this method, exponential functions are first used to build filters to implement a re-constrained VMD. Focusing on the fault information and minimizing it in the residuals. Then, SABO was employed to find the best parameter combination for subsequent signal processing. Finally, the signal is decomposed, and envelope spectral analysis is performed on each component to extract the fault frequencies, thereby identifying the specific fault type. Numerical simulations and real experimental data were used to demonstrate the effectiveness of the proposed method. In addition, the generalization ability of the proposed method was tested using 40 sets of sample data, and the average accuracy of this method reached 97.5%. Compared with other commonly used signal decomposition methods, the superiority of this method in rolling bearing fault feature extraction is proved.
科研通智能强力驱动
Strongly Powered by AbleSci AI