Research On Interharmonics Estimation Using Kalman Filter Approach Based On AR Model

As the world's most widely used and most convenient energy in recent years,the electric power not only plays an important role in industrial modernization, but also is the key resources the whole society and people's lives rely on. So the quality of the electric power attracts more and more attention of scholars, and the pollution of harmonics and interharmonics is the most important problem need to be settled.The frequency of interharmonics in the power system is non-integer times of the fundamental frequency, and its spectrum is changeful with time. The classical FFT algorithm can detect the integer harmonics effectively, but hardly extract all the frequency components of interharmonics accurately, due to the impact of asynchronous sampling.This paper utilizes the method of modern spectral estimation, which has congenital advantages in the processing of stochastic signal, to analyze the power signal. It can describe frequency characteristics by using of statistics exactly. This paper discusses the principles of the autoregressive (AR) model spectrum and the interharmonics analysis based on the AR model. To overcome the disadvantages of the Burg algorithm, such as spectral splitting and peak shifting, an efficient method for the interharmonics detection of the measured power signal is proposed. To allow for more reliable frequency information in the signal and on-line processing, the proposed method uses the Kalman Filter algorithm to estimate the AR coefficients and obtains high-resolution power spectrum estimates. This paper describes the principles of the AR model spectrum estimation as well as the proposed procedures using Kalman Filter approach based on AR model. Compared with Burg algorithm, the method reduces the peak shifting level, and eliminates the spectral splitting significantly.Based on the accurate spectral estimation of the frequency components, a new method of the amplitude estimation using Kalman Filter algorithm and the spectral lines interpolation correction is proposed. The method utilizes the Kalman iterative operation to calculate the amplitude of the frequency components with known frequency value, and then the spectral lines interpolation algorithm to further amend the estimate of the frequency and amplitude, which can improve the accuracy of the interharmonics detection markedly.Finally, the verification experiments by Matlab simulation and the actual voltage waveform show that the proposed method not only improves spectral estimation performance and the noise immunity obviously, but also can be applied to the online measurement in the power system, due to its high computational efficiency and short time sampling.