Harmonic detection algorithm based on self-convolution window six-line interpolation and improved all-phase FFT

When using fast Fourier transform to analyze harmonics, it is difficult to realize high precision harmonic detection due to the spectrum leakage and fence effect caused by asynchronous sampling. In this paper, a combined optimization algorithm is proposed. By analyzing the characteristics of the self-convoluted window, the self-convolution window six-spectrum-line interpolation algorithm is proposed to estimate the amplitude and frequency parameters of harmonics. Meanwhile, an improved all-phase FFT algorithm is proposed to estimate harmonic phase parameters, considering the fact that the data before zero hour can not be collected. Finally, through simulation experiments, the proposed algorithm is compared with other algorithms in the case of classical harmonic signals and noise. The results show that the algorithm has higher harmonic detection accuracy and stronger anti-noise capability.