Hanning-Windowed Compressive Sensing with Interpolation for Power Harmonic Measurement
Abstract
To mitigate measurement errors caused by spectral leakage and the fence effect in power harmonic detection, while addressing the issues of excessive data volume and slow algorithm reconstruction in traditional methods, this paper proposes a Hanning-windowed compressed sensing and interpolation approach for power harmonic detection. First, the Hanning window is integrated into the compressed sampling process of compressed sensing to achieve windowed compressed sampling of signals. Subsequently, the Binary Search-based Sparse Adaptive Matching Pursuit (BSAMP) algorithm is employed to reconstruct the sparsely sampled vectors. The reconstructed sparse vectors are further refined using a three-spectral-line interpolation technique to obtain the final harmonic parameter estimates. Finally, an experimental platform is constructed to verify the theoretical correctness and practical feasibility of the proposed method. Experimental results demonstrate that, at a compression ratio of 50%, the maximum relative errors for frequency, amplitude, and phase measurements are 0.0074%, -0.33%, and 0.95%, respectively. This confirms the method’s capability to accurately detect complex harmonic signals with a reduced number of sampling points. In hardware experiments, the maximum absolute errors for frequency and amplitude measurements are 0.0153 Hz and 0.1627 V, respectively, further validating the effectiveness and feasibility of the proposed approach.