Xu Zhenyang, Zhang Zuofu, Ren Fuqiang, Wang Xuesong, Liu Aobo, Guo Lianjun
School of Mining Engineering, University of Science and Technology LiaoNing, Anshan, 114051, China.
School of Civil Engineering, University of Science and Technology LiaoNing, Anshan, 114051, China.
Sci Rep. 2025 Mar 23;15(1):10035. doi: 10.1038/s41598-025-94411-5.
Blasting vibration signals are often contaminated by trend terms and noise, stemming from environments, and instrument errors. This contamination hinders subsequent signal processing and analysis. To obtain a pure blasting vibration signal, a parameter-adaptive variational mode decomposition (VMD) method based on the improved electric eel foraging optimization (EEFO) algorithm is proposed for preprocessing the blasting vibration signal. This method removes high-frequency noise and low-frequency trend components from the signal. Combining the abrupt and related characteristics of blasting vibration signals, a weighted multi-scale permutation entropy is constructed as the fitness function for parameter optimization. The EEFO algorithm, with strong global and local search capabilities, is employed to optimize the VMD decomposition parameters. This approach adaptively determines the optimal combination of decomposition modes K and the secondary penalty factor α. The analysis results of simulated vibration signals with different interference components and actual measured blasting vibration signals using this method show that, compared to traditional VMD, empirical wavelet transform, and empirical mode decomposition methods, EEFO-VMD has superior adaptability and anti-aliasing capabilities. Even in complex interference components, it can adaptively determine the optimal decomposition parameter combination, effectively removing interference components from the vibration signals. This method is suitable for preprocessing blasting vibration signals.
爆破振动信号常常受到源于环境和仪器误差的趋势项和噪声的污染。这种污染阻碍了后续的信号处理与分析。为了获得纯净的爆破振动信号，提出了一种基于改进的电鳗觅食优化（EEFO）算法的参数自适应变分模态分解（VMD）方法对爆破振动信号进行预处理。该方法去除信号中的高频噪声和低频趋势分量。结合爆破振动信号的突变性和相关性特点，构建加权多尺度排列熵作为参数优化的适应度函数。利用具有强大全局和局部搜索能力的EEFO算法对VMD分解参数进行优化。该方法自适应地确定分解模态数K和二次惩罚因子α的最优组合。使用该方法对具有不同干扰分量的模拟振动信号和实测爆破振动信号进行分析的结果表明，与传统的VMD、经验小波变换和经验模态分解方法相比，EEFO-VMD具有更强的适应性和抗混叠能力。即使在复杂的干扰分量情况下，它也能自适应地确定最优分解参数组合，有效去除振动信号中的干扰分量。该方法适用于爆破振动信号的预处理。
