Harmonic Analysis Based on Blackman-Harris Self-Multiplication Window

Asynchronous sampling and non-periodic truncation cause spectral leakage during Fast Fourier transform (FFT) analysis, resulting in low accuracy of harmonic parameter estimation. The windowed interpolation algorithm can effectively suppress the negative impact of the spectrum leakage and fence effects in the parameter estimation process. In this paper, a new cosine window with better side-lobe performance than some classical windows was introduced via second-order self-multiplication of the Blackman-Harris window. Based on this, the estimation algorithms for the frequency, amplitude and phase of the harmonics were also proposed by using triple-spectrum-line interpolation method. The simulation results verify effectiveness of the new method.