地震与爆破波形信号的EEMD多尺度分布熵提取和WOA-SOM神经网络识别研究

针对地震信号中存在大量环境噪声的问题，基于天然地震事件和人工爆破事件建立初始数据集，利用集合经验模态分解（EEMD）技术对波形信号进行分解、降噪，提取出较纯净的各阶本征模态函数（IMF）分量，然后对前10阶分量分别计算分布熵，即EEMD多尺度分布熵，建立神经网络输入矩阵。应用鲸鱼优化算法（WOA）对自组织特征映射（SOM）神经网络参数（竞争层维数、网络训练次数）进行优化，针对不同训练样本寻找对应的最优参数值以改善模式识别的稳定性，从而提高地震识别率。结果表明，EEMD多尺度分布熵结合WOA-SOM模型可有效识别天然地震和人工地震。

EEMD Multiscale Distribution Entropy Extraction and WOA-SOM Recognition between Seismic and Blast Waveform Signals

Due to the large amount of environmental noise in seismic signals, we establish the initial data set based on natural earthquake events and artificial blasting events, decompose and reduces the noise in waveform signals using the ensemble empirical mode decomposition(EEMD) technique, extract the purer intrinsic mode function(IMF) components of each order, and then calculate the distribution entropy for the first 10 order components separately to establish the neural network input matrix. The whale optimization algorithm(WOA) is applied to optimize the self-organizing feature mapping(SOM) neural network parameters(competitive layer dimensions and number of network training) to find the corresponding optimal parameter values for different training samples to improve the stability of pattern recognition, thus improving the seismic recognition rate. The results show that the EEMD multiscale distribution entropy combined with WOA-SOM model can effectively identify natural earthquakes and artificial blasting events.