An Improvement Of Interpolating Window Fft Algorithm And Study On Phase Differende Correction Method

With many kinds of non-linear loads with power electronic devices in particular widely used in power system, power system harmonics pollution have become more serious. The impact of power quality of harmonics pollution has become a major public hazard, and extremely affects the security and economics of power system. Therefore, it is very necessary to measure the harmonic content of the grid in real-time, grasp exactly the actual situation of the power system harmonic, to prevent harmonic hazards, to maintain the safe operation of power grids.Discrete Fourier Transform (DFT) algorithm, fast Fourier transform (FFT) algorithm in particular , is usually used as the primary means of harmonic analysis, for it easily computer. But the frequency of the power system is not the constant value all the time, it will fluctuate at a small range including the rated frequency. It is not certain that the real-time frequency is integer multiple of sampling frequency. And then synchronous sampling can not be realized. In that the fence effect and the phenomenon of spectrum leakage give rise to measurement errors. Interpolation algorithm can reduce the error caused by the fence effect [3～7], the error which is caused by the spectral leakage can be reduced by adding window function. Rife-Vincent (III) Window function have less spectral leakage, the paper use adding Rife-Vincent (III) window.In this paper, the windowingFFT interpolation algorithm and the phase correction method which are commonly used are improved. the coefficient formula of correction amplitude is too complicated , and it has large amount of computation . This paper utilized cubic spline function to approximate to the correction function of frequency correction factor of seven polynomial and complex amplitude ,and used the nested (effective) form of cubic spline interpolation function to calculate the frequency correction factor and the complex amplitude of the correction factor. This need only three times of multiplication and four times of addition . The formula is simple and has less amount of calculation ,easier achievement of procedures, better instantaneity. furthermore , it is continuous at the piecewise point where it has a precise value ? The simulation computed result shows that the Rife—Vincent(III) window interpolation algorithm by using cubic spline function has high precision .In this paper, phase correction is realized by using recursive windowing DFT algorithm. using recursive windowing DFT algorithm which is time-domain recurrence, frequency-domain windowing to compute FFT transform twicely , obtaining difference value of phase of each-frequency harmonics after a sampling cycle of changes in the amount of change calculated then using the difference value to work out the corresponding frequency difference value, and then using it to amend the amplitude spectrum. This method is fast, track the change of signal fast, and it does not require computing of trigonometric functions, so the amount of computation is small, and the method is easy to achieve by Assemble Language. These resolve the shortcomings of phase correction method which is described in the literature [54].Based on the new algorithm, the paper use MATLAB to program, utilize the harmonic detection of the simulation analysis in the new given algorithm. The results show that the new algorithm is isolated each-frequency harmonics better, provides a powerful tool for power system harmonic detection.