将分解后对每个BLIMF所求取MWPE值作为该信号的多尺度特征量，其中

The decomposition is obtained for each value MWPE BLIMF as multiscale feature amount of the signal, wherein the number of samples MWPE value N, the scale factor s, delay embedding dimension m and about time t. <br>Considering the multi-scale weighted permutation entropy permutation entropy special form s = 1, and the multi-scale weighted permutation entropy calculation depends on the value of WPE, the WPE of the selected parameters discussed first. According to [8], the delay time is set herein to t = 1, N = 6873, for m = 2-8, s is the scale factor value, four oscillating signals of different conditions were VMD decomposition, Seven BLIMF, calculated for each BLIMF WPE values obtained, the specific trend can be seen in FIG. 7 to FIG. 7, when m = 7, 7 can well be obtained BLIMF WPE values to distinguish.

将分解后对每个BLIMF所求取MWPE值作为该信号的多尺度特征量，其中MWPE值与样本数量N、尺度因子s、嵌入维数m和延时时间t有关。<br>考虑到加权排列熵是多尺度排列熵s=1的特殊形式，且多尺度加权排列熵主要依赖于WPE值的计算，所以首先讨论WPE的参数选择。依据文献[8]所示，本文将延迟时间设为t=1，N=6873，对 m=2-8，尺度因子s为定值，对4种不同工况的振动信号分别进行VMD分解，得到7个BLIMF，对每个BLIMF计算得到WPE值，具体变化趋势如图7从图7中可以看出，当m=7时，可以很好的将7个BLIMF所得的WPE值进行区分。

The mwpe value obtained for each blimf after decomposition is taken as the multi-scale feature of the signal, where mwpe value is related to the number of samples n, scale factor s, embedding dimension m and delay time t.<br>Considering that the weighted permutation entropy is a special form of the multi-scale permutation entropy S = 1, and the multi-scale weighted permutation entropy mainly depends on the calculation of the WPE value, this paper first discusses the parameter selection of WPE. According to the literature [8], this paper sets the delay time as t = 1, n = 6873, M = 2-8, scale factor s as the fixed value, and performs VMD decomposition on the vibration signals of four different working conditions to obtain seven blimf, and calculates the WPE value of each blimf. The specific change trend is shown in Figure 7. When m = 7, the WPE value of seven blimf can be distinguished well.