Research on new energy grid-connected power line communication denoising based on SVD decomposition and improved observation matrix

In new energy grid-connected broadband power line carrier communication, signals of different frequencies may affect signal stability and phase distortion due to the presence of noise when propagating in power lines. In order to reduce the interference of noise to broadband power line carrier communication in new energy grid connection, an improved observation matrix method based on SVD decomposition is proposed in this paper. In this algorithm, spectral decomposition is used to increase the column independence of observation matrix, and the optimal H Toeplitz observation matrix design method is used to reduce the cross-correlation between observation matrix and sparse basis. The experimental results show that the signal-to-noise ratio (SNR) optimized by the proposed method is $\mathbf{5. 4 5 d B}$ higher than that of the traditional Gaussian observation matrix, and the root-mean-square error (RMSE) is reduced by $32.67 \%$ on average. The good denoising effect obtained by the algorithm in this paper is helpful to promote the large-scale grid connection and utilization of new energy.