Multiscale analysis of lightning return stroke

Ensemble empirical mode decomposition (EEMD) of the lightning negative return strokes was conducted and a set of almost complete and orthogonal components named intrinsic mode functions (IMFs) having physical meaning were obtained. The results show that IMF1 is the high frequency noise item and the residual item is mainly due to the electrostatic field. IMF2-IMF11 represent the discharge characteristics of the lightning channel in different scales, respectively. According to the calculation of the fractal dimension of different IMFs, the fractal dimension will increase with the increasing IMF order, indicating that the IMF with higher order corresponds to the discharge channel with increasing tortuosity. It is also found that EEMD-based Hilbert-Huang transform (HHT) spectrum can ameliorate mode mixing and it is more suitable for analyzing the lightning than EMD-based HHT analysis and wavelet analysis.