Analysis of hydro-turbine non-stationary vibration signals based on CEEMDAN and Wigner-Ville Distribution

This paper proposes a time-frequency method based on complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) and Winger-Ville distribution (WVD) for hydro-turbine non-stationary vibration signals analysis. Unlike the traditional EMD and EEMD, CEEMDAN can efficiently reduce the mode mixing and has a better convergence. Base on a series of intrinsic mode functions (IMFs) obtained by using CEEMDAN, WVD can effectively suppress the interference of cross-terms and accurately depict the time-frequency characteristics of original vibration signals. The proposed method can be employed to analyze the hydro-turbine non-stationary vibration signals successfully and shows excellent performance.