ALGORITHM FOR CHARACTERISTIC HARMONICS ANALYSIS BASED ON SPECTRAL REFINEMENT AND INTERPOLATION

摘要

The common method of characteristic harmonics analysis is fast Fourier transform (FFT). Because of the restraint of fence effect and spectrum leakage, the precision of FFT is not high. In order to improve the accuracy of characteristic harmonics analysis, an algorithm based on spectral refinement and interpolation is put forward. First, Zoom-FFT is adopted to refine the local frequency range of characteristic harmonics. Second, 6-term cosine windowed interpolation is used to correct the refined spectrum. Then the accurate estimations of parameters such as frequency, amplitude and phase can be obtained. Fast Fourier Transform, Zoom-FFT and the proposed method are adopted to process simulated characteristic harmonics and harmonics of 24-pluse rectification system. For the processing of simulated characteristic harmonics, the minimum relative errors of frequency, amplitude and phase are 2.3812 05×10~(-14)%, 5.3372 61×10~(-14)% and 2.7430 70×10~(-4)%. For the processing of harmonics of 24-pluse rectification system, the minimum errors of frequency, amplitude and phase are 1.8233×10~(-15) Hz, 1.0305 × 10~(-15) A and 2.4207 × 10~(-4). Results of the experiments demonstrate that the proposed method has the capability to improve the precision of parameter estimation and is a high-quality characteristic harmonics analysis algorithm.