Symmetrical differential energy operator demodulation and energy operator demodulation are only applicable to single component AM / FM signals, while the signals obtained in the experiment are multi-component, so it is necessary to decompose the multi-component signals into several single component signals first, select the single component signal with the largest correlation through the maximum kurtosis, and demodulate the single component signal by symmetrical differential energy operator, so as to get the best result Vibration signal with obvious characteristic frequency.

Fig. 1 and Fig. 2 show the spectrum of a given original signal in normal state and fault state. It can be seen from the figure that the spectrum of the original signal is disordered and disordered. Whether it is normal state or fault state, we can not directly extract the required characteristics and laws from the spectrum, so we need to process the original signal.

Firstly, the original signal is demodulated by symmetric differential energy operator based on EMD, and its spectrum is obtained. Two groups of data under normal state and fault state are selected for analysis, and the results are shown in Fig. 6 and Fig. 7 It can be seen that the frequency performance of normal state and fault state data after processing is very messy, and it can not clearly reflect the obvious characteristics of the amplitude of the vibration signal of the planetary gear box at the meshing frequency and frequency doubling point, so it can not carry out the fault diagnosis work well. Therefore, the experimental analysis is carried out based on EEMD and symmetrical differential energy operator demodulation method.