The Research On Voltage Fluctuations And Flicker Detection

Voltage Flicker in the distribution network, as a major power quality disturbance, becomes more serious with the increasing large-capacity impact load in the power system. Therefore, it is necessary to detect and analyze them. Whether the voltage flicker envelope signal can be extracted efficiently and accurately is the key point in detection of voltage flicker. The current recommended method of flicker measurement of IEC can only give the statistical evaluation index of flicker, such as the severity of the short-term and long-term flickers, but it can not explain the specific parameter characteristics of flicker. To this end, the quick measurement and calculation method of voltage flicker has been proposed in this paper on the basis of combining the mathematical morphology filtering with the Hilbert transform.The extraction method of envelope curve based on Hilbert transform has the following advantages:it has the ability in accurately handling the flicker signals of inter-harmonics in high and low frequency bands; it needs only a small number of sampling points, easy to implement; it can process both the smooth and non-stationary signals; it also has good adaptation to the voltage flicker signals and can accurately show their variation trend so as to obtain the quantitative information about the severity of the voltage flicker.Mathematical morphological filtering method can effectively filter out the high frequency harmonics in the voltage flicker signals. Compared with the conventional digital filter, its computing is simple and it is convenient for the real-time implementation by using the low-cost microprocessor chip. Mathematical morphological filtering method has two important problems, that is, the selection of mathematical morphological filters and the design of structural operators.. In this paper, after selecting the mathematical morphological filter, a variable factor is introduced to establish a multi-scale structural operator, making the length of structural elements variable, so that the more subtle interference in signals can also be filtered.By combining the merits of the short-time Fourier transform and the wavelet transform, the time window width of the S transform changes inversely with frequency, i.e. the time window is wider at low frequency band, and thus a higher frequency resolution can be obtained, while the time window is narrow at the high frequency, so access to a very high temporal resolution can be obtained. The matrix obtained from modulo of all elements in S matrix is denoted as the S-mode matrix, with its row vector showing the time-frequency distribution of the signals at a certain frequency, and its column vector showing the amplitude-frequency characteristic. The voltage fluctuation and flicker are all steady-state power qualities, so it is enough to use the column vectors of the S-mode matrix in analyzing the spectrum of signals.