Research And Application Of Harmonic Inter-harmonic Detection Method And Detection System Based On Prony

With the widespread application of nonlinear load in power system, the problem of power quality is getting worse and worse, which brings increasingly serious harm to the power users. And the problems of harmonic, inter-harmonic interference caused great concern to the society. In order to improve the power quality of power grid, it is very important to find an effective method to detect and analyze the power quality.The widely used three-phase instantaneous reactive power theory is studied in this paper, which includes the P-Q coordinate method, ip-iq harmonic detection method and d-q method. By constructing the signal source, the ip-iq method, d-q method is simulated under the condition of distortion and three-phase asymmetry in the power grid. It is verified that both of them can effectively detect the harmonic current in the case of voltage distortion. But the d-q method has a higher phase detection accuracy in the three-phase asymmetry condition.For the spectrum leakage and fence effect problems of discrete Fourier transform, this paper explains the reasons of these problems by analyzing the results of non-synchronous sampling process. Then, the effect of windowed algorithm is studied in this paper. Through the comparison of mainlobe width and sidelobe decay rate about various windowing function, this paper draws a conclusion that the Hanning window is the best function to restrain the spectral leakage effect. Besides, the effect of zero-padding method on FFT was studied in the end of this chapter.For the spectral leakage problem of traditional Fourier transform under non-synchronous sampling, and for the interference problems between the harmonic and inter-harmonic, this paper innovatively puts forward a method called quasi-synchronous- comb filter- Prony spectrum algorithm(QS-Prony). In this algorithm, the frequency shift method is used to realize the quasi-synchronization of asynchronous sampling signals. In order to avoid the mutual interference between harmonics and inter-harmonics, this quasi-synchronization sequence is further separated by the FIR notch filter. Then, the linear combination of exponential function is utilized by the Prony spectral analysis method to fit the sample data, and the frequency, amplitude and phase of the signal were extracted by using this process. By contrasting windowed interpolation FFT algorithm(WI-FFT), quasi-synchronous-comb filter separates- windowed interpolation FFT algorithm(QS-WI-FFT) and quasi-synchronous- comb filter separates- Prony spectrum analysis method under different noise environment, it shows that the QS-Prony method can effectively solve the problem of non-synchronization, avoiding the mutual interference problems between harmonic and inter-harmonic effectively. In addition to this, this algorithm has higher frequency resolution compared with the FFT algorithm.The lower computer system is designed based on FPGA chip, including current and potential transformer, analog signal input circuit, preamplifier circuit and analog to digital conversion circuit. ADS8364, a 16-bit 6 channel synchronous sampling chip, is employed as the digital-to-analogue conversion unit. Because the Prony algorithm is sensitive to noise in the host computer system, a kind of anti-noise circuit is innovatively optimized in the hardware design. The program of FPGA is designed, which contains three segments AD sampling program with one-hot code.The power quality detection system is built based on LabVIEW, which includes the real-time waveform display module, the effective value detection module, the power detection module and the harmonic detection module. The harmonic, inter-harmonic detection module uses the Prony based quasi-synchronous sampling algorithm, which can display the harmonic component within 30 times. Finally, through the field testing of power grid signal, this detection system have similar precision compared with the HIOKI3196 harmonic analyzer in terms of voltage, current and power measurement. Influenced by the insufficient of frequency resolution, it is difficult for HIOKI 3169 to distinguish inter-harmonic signal from the harmonic, while this detection system can pick out the inter-harmonic accurately.