A New Solution for Frequency Response Measurement in Power Electronics Systems: A Compressed Sensing Approach

Efficiency is a crucial indicator for measuring the frequency response of power electronics systems. Traditional frequency response measurement methods are essentially Nyquist sampling, where they sample the frequency response characteristic curve in the frequency domain, either at a single point or multiple points. However, frequency domain sampling offers strong flexibility, and the frequency response characteristic curve may exhibit sparsity in its information. If these characteristics could be fully utilized, the efficiency of frequency response measurement might be further improved. Therefore, this paper proposes a random frequency response measurement method based on the compressed sensing theory, which leverages sparsity to achieve efficient sampling. This method aims to obtain the sparse features of frequency response characteristic curve using a limited number of random measurement points and subsequently reconstruct the curve based on these sparse features. Numerical test and experiment demonstrate the feasibility of this method, showing its capability to recover high-resolution data with a small number of sampling points. Additionally, the paper emphasizes the need for further research to improve the compression ratio and recovery stability.