摘要：Intrmittent wind power, photovoltaic and other renewable energies have been paralleled, whichmakes the phenomena of high-order harmonics and simple harmonics more and more serious in the powersystem, showing a wid-frequency trend. The existing measurement algorithms mainly aim at signals in mid-frequency and low-frequency. Besides, they are lack of a uniform high-precision algorithm for wide-frequency measurement. To solve this problem, we propose a high-precision algorithm based on Z-ADALINE.Firstly, Zoom FFT algorithm is used to analyze original sampled signals. This step enables the refinement ofits frequency spectrum, and obtains accurate frequency measurement results. At this time, the number offrequencies can also be determined. Secondly, the result of Zoom FFT is used as the input of the adaptivelinear neural network(ADALINE). ADALINE can estimate amplitude and phase with high precision. Thesimulation results show that the proposed algorithm can realize high-precision measurement of frequency,amplitude and phase of wide-frequency signal effectively. Among them, the frequency resolution can be upto 03 Hz The amplitude error is within IV. Phase error is less than 0.6°. The results may provide somesignificant references for practical wide-frequency signal measurement in power electronic power grid.
其他摘要：Intermittent wind power, photovoltaic and other renewable energies have been paralleled, which makes the phenomena of high-order harmonics and simple harmonics more and more serious in the power system, showing a wide-frequency trend. The existing measurement algorithms mainly aim at signals in midfrequency and low-frequency. Besides, they are lack of a uniform high-precision algorithm for widefrequency measurement. To solve this problem, we propose a high-precision algorithm based on Z-ADALINE. Firstly, Zoom FFT algorithm is used to analyze original sampled signals. This step enables the refinement of its frequency spectrum, and obtains accurate frequency measurement results. At this time, the number of frequencies can also be determined. Secondly, the result of Zoom FFT is used as the input of the adaptive linear neural network(ADALINE). ADALINE can estimate amplitude and phase with high precision. The simulation results show that the proposed algorithm can realize high-precision measurement of frequency, amplitude and phase of wide-frequency signal effectively. Among them, the frequency resolution can be up to 0.3 Hz. The amplitude error is within 1V. Phase error is less than 0.6°. The results may provide some significant references for practical wide-frequency signal measurement in power electronic power grid.
