Signal noise reduction of clamp current Transformer used in distribution network based on EEMD-ICA-Wavelet

A new type of power system with new energy as the main body is constructed under the goal of ‘double carbon’. The integration of large-scale distributed energy into the distribution network will affect the stable operation of the power system. When the live calibration of the power transformer for distribution network is carried out, the output signal of the clamp current transformer contains a large amount of nonlinear noise and has strong randomness, and the output accuracy is easily affected by noise. Considering the advantages of EEMD, Wavelet and ICA, EEMD-Wavelet-ICA noise reduction model is established. Firstly, the low frequency and high frequency components of IMF are obtained by EEMD decomposition. And then through the Wavelet, ICA noise reduction processing, the noise reduction signal multi-scale reconstruction to obtain noise reduction signal. The noise reduction effect is evaluated by signal-to-noise ratio, mean square error, waveform similarity coefficient, signal energy and other indicators. The noise reduction results of simulation signal and experimental data show that : compared with ICA-Wavelet, FFT and DFT, the SNR, NCC and Ep are improved in different degrees, and the mean square error is reduced in different degrees.