Compressed Sensing High-accuracy Detection for Electric Power Interharmonics

Abstract

Interharmonics frequencies are not integer multiple of the fundamental frequency, and interharmonics amplitudes are far less than fundamental amplitude and harmonics amplitudes,  which mean high sensitivity to desynchronization problems, so it’s dificult to estimate interharmonics.  In this paper, a new method based on random sparse sampling and compressed sensing (CS) Bregman technique was proposed to estimate the interharmonics. The random sam pling has following advantages; alias-free、sampling frequency need not obey the Nyquist limit、and higher frequency resolution. So the random sampling can measure the signals which their frequencies component are close, and can implement the higher frequencies measurement with lower sampling frequency. However,the rangdom sampling exists the noise in spectrum analysis, so it’s difficult to estimate the low amplitude signals. Compressed sensing can work out this problem by designing observation matrix and with the sparsity reconstruction of the signal in the Fourier domain; in addition, the application of CS can estimate the amplitudes and phases of the signals exactly. The results ofexperiments show that the proposed method can estimate the interharmonics exactly,even if the interharmonics frequencies are close the fundam ental frequency and interharmonics amplitudes are far less than fundamental am plitude；and can measure high order interharmonics with lower sam pling frequency.