Study On The Muti-harmonics Analysis

The advancement of high-power semiconductor devices and their excellentcontrol capabilities have resulted in extensive use of power-electronics-basednonlinear equipment in commercial and industrial sectors for various applications.These devices inject harmonics and inter-harmonics into the system, which causemany severe problems to the system and equipment. Thus，harmonics should becontrolled without delay.It is a prerequisite to estimate harmonic parametersaccurately for the power harmonic control. Fast Fourier transform (FFT) is the mostused widely computation algorithm for harmonic measurement. However, because ofaliasing effect, leakage effect and picket-fence effect caused by the FFT applied inasynchronously sampling, the signal parameters are not precisely estimated, which aremore serious for inter-harmonic analysis. In addition, FFT’s low resolution isunfavorable for inter-harmonic analysis, Thereupon it matters a lot to break theresolution limit of the FFT. Aiming at these problems, this article studied how toimprove effectively the accuracy and reliability of the estimation.(1) A new method of electric power harmonic parameters based on energycentrobaric that considering the leakage effect was presented in this paper. To reducethe error brought about by parameter estimation of the spectrum interference signal inthe energy centrobaric correction method, the iterative process is introduced to modifythe spectrum amplitude. The iterative initial value can be obtained by the energycentrobaric correction method and the iteration continuously subtracts the spectrumleakage value of positive frequencies and negative frequencies. The energycentrobaric correction is invoked repeatedly and the termination of the iteration is upto the iterative error defined in the time domain. Simulation results demonstrate thatthe algorithm is practical, which presents satisfactory estimation accuracy for weaksignals and needs small number of iteration times and less computation.(2) A new accurate estimation method of electric power harmonic parametersbased on information theory and MUSIC method is presented in this paper. Firstly,based on information theory (AIC,MDL) and CCT method, the model order of theelectric power is calculated.Through comparing the simulation result, we estimate themodel order by MDL norm in the subsequent simulation.After get the correct modelorder, the spatial spectrum estimation MUSIC algorithm is applied to estimateharmonic parameters. Result that it compares with Burg algorithm by computer simulation shows that it has higher veracity.(3) In this paper,the estimation of signal parameters via rotational invariancetechnique (ESPRIT)-based method is proposed with an accurate model order (thenumber of frequency components)estimate for power system harmonic andinterharmonics detection.Firstly, the proposed method uses the RD curve to estimatethe model order, and then the frequency,amplitude and phase are estimated byESPRIT algorithm.Finally, simulation shows that the proposed method performsbetter in term of precision and resolution, and robustness as well.