Research On Detection Methods Of Voltage Fluctuation And Flicker

In recent years, along with the power system load becoming more complicated, large capacity non-linear and impact loads are unceasing increased, causes the power quality problems increasingly serious, and the utilities and users have paid much more attention to it. Among with the research objects, the voltage fluctuation and flicker is a currently important issue in power quality, taking effective measurements of voltage flicker, are important significance to the governance of the power quality problems. According to the development status of the power quality detection technology, having analyzed and researched the widely used detection methods in power quality disturbance signal. And then from two parts to analyze and research the voltage fluctuation and flicker, one is to extract the signal envelope, the other is to select the correlation spectrum estimation methods to estimate the modulation parameter, and then accurately calculate the short term flicker value Pst. Firstly, this paper has analyzed the common detection methods of the original voltage fluctuation signal, such as the IEC recommendation, wavelet transform and Hilbert transform; then by the analysis of the spectral estimation methods of modulation parameters, namely FFT algorithm, Pisarenko Harmonic Decomposition method and the improved Prony spectral estimation method. The paper focuses on the study of Prony method, on the basis of original algorithm, uses a kind of improved Prony method, to improve the fitting accuracy, overcome the actual order identification, the fitting ability for the non-stationary signal, sensitivity to noise, and other deficiencies. An evaluating method based on Hilbert transform and improved Prony method is proposed to evaluate the modulating wave, the method can precisely extract the envelope of the voltage fluctuation, and detect the modulation amplitude, frequency and other parameters, finally the short term flicker value Pst are calculated by the statistical ranking method. The simulation data by MATLAB show the accuracy of the theory, i.e..