(1. Economic and Technology Research Institute of State Grid Hebei Electric Power Company, Shijiazhuang 050021, China; 2. School of Education, Tianjin University, Tianjin 300072, China; 3. School of Electrical Engineering and Automation, Tianjin University, Tianjin 300072, China)

Compressive sampling has become the development trend of real-time acquisition and dense detection of complex harmonic distortion signal in power systems. In view of the problems in harmonic and inter-harmonic detection, this paper proposes a new detective method with compressive sampling. By analyzing the related theorems and corollaries for obtaining sparsity degree, the multiple-extraction gradient pursuit(MEGP)algorithm is proposed to promote detective precision of harmonic components, especially the inter-harmonic components. Considering the spectrum leakage, a window sparse measure(WSM)matrix is created, which is generated simply and effective for all components. Results of experiments show that the MEGP algorithm is able to promote 0. 00053, 0. 011 2% and 0. 071° of detective precision for frequency, amplitude and phase, respectively, while the WSM matrix is able to promote the signal to noise ratio(SNR)about 30 dB. The detective errors of frequency, amplitude and phase by compressive-sampling detective method are within 0. 004 03 Hz, 0. 001 73% and 0. 037 221°, respectively, which totally meets the demand of the national standards.