[0030] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings.
[0031] like figure 1 As shown, it is a method for detecting a SAG signal in a DVR device provided by an embodiment of the present invention, and the method includes:
[0032] Step S101, according to the preset sampling frequency and the power frequency period, obtain the same number of sampling points of the power grid signal in each power frequency period, and set a circular buffer with a length equal to the number of the sampling points;
[0033] The specific process is that the sampling frequency of the power grid signal can be set to 6.4KHz, the power frequency period is set to 20ms, or the sampling frequency can be set to 12.8KHz, and the power frequency period is set to 20ms and other special values, so that the sampling within 20ms of each power frequency period The number of points is 128 or 256 to the power of 2. Once the preset sampling frequency and power frequency period are fixed, the number of sampling points obtained by the power grid signal in the power frequency period is the same, and a circular buffer with a length equal to the number of sampling points in a single period is constructed, such as a circular buffer The length of the zone is 128 or 256.
[0034] Step S102, after obtaining the sampling value corresponding to each sampling point on a power frequency period, the obtained sampling value is stored in the circular buffer according to the sampling time sequence, forming a sliding analysis window;
[0035] The specific process is to extract the sampling value corresponding to each sampling point on a power frequency period of 20ms, which can be the current power frequency period or the previous power frequency period, and put these sampling values (128 or 256) in the cycle The buffer is stored in the order of sampling time, thereby forming an analysis window that slides on the sequence of sampling values. It can be understood that the sliding analysis window is in a first-in, first-out manner in computer program processing, which improves the processing efficiency of the computer and saves processing time.
[0036] Step S103: When the sampling time of the next sampling point arrives, the sampling value of the next sampling point is obtained and stored in the analysis window, and after discarding the sampling value with the earliest sampling time in the analysis window, update all the sampling values. the analysis window;
[0037] The specific process is that, since the analysis window is currently data to be processed, if new data is input, the original data will be overwritten, thereby forming a circular buffer. Therefore, when the sampling time of the next sampling point arrives, that is, for each new sampling point for the analysis window, the sampling value obtained from the new sampling point will be stored in the analysis window, and the earliest sampling time in the analysis window will be discarded. After the sampled value of , update the analysis window. It can be understood that the sampling values of the new sampling points are also stored in the order of sampling time.
[0038] Step S104, performing wavelet analysis on each sampling value in the updated analysis window, to obtain the modulus maximum value corresponding to the next sampling point at its sampling time;
[0039] The specific process is to perform wavelet analysis on each sampled value in the updated analysis window to obtain the modulo maximum value corresponding to the next sampling point at the sampling time. Since only one data is updated each time in the analysis window, the wavelet analysis result of the data in the entire circular buffer can be obtained as long as the local update is performed according to the data dependency in the wavelet analysis process. Since other data points are not updated, there is no need to recalculate, so that the amount of calculation can be saved and analysis results can be obtained at high speed. Similarly, the inverse transformation process of wavelet analysis adopts the same local update strategy, and only needs to update and calculate the affected numerical points to complete the wavelet inverse transformation. The saving of calculation amount makes it possible to execute the algorithm on devices with limited computing resources such as DSP.
[0040] The key of wavelet transform is the determination of mother wavelet Ψ(t) (also called basic wavelet), Ψ(t)∈L 2 (R), R is the set of real numbers, L 2 (R) is the real number domain space, and the wavelet transform formula of Ψ(t) is:
[0041] ψ a , b ( t ) = | a | 1 2 ψ ( t - b a ) - - - ( 1 )
[0042] In formula (1), a, b∈R, and a≠0, Ψ a,b(t) is a wavelet sequence, a is a scaling factor, and b is a translation factor. When the input function is set as f(t), the wavelet transform expression is:
[0043] WT f ( a , t ) = 1 a ∫ R f ( t ) ψ ( t - b a ) dt - - - ( 2 )
[0044] Daubechies wavelet is a wavelet function constructed by the world-famous French wavelet analyst Inrid Daubechies, abbreviated as dbN, where N is the order of the wavelet. The support region in the wavelet Ψ(t) and the scaling function Φ(t) is (2N-1), and the vanishing moment of Ψ(t) is N. db wavelet mathematical expression:
[0045] p ( y ) = Σ k = 0 N - 1 C k N - 1 + k y k - - - ( 3 )
[0046] In formula (3), is the binomial coefficient, then:
[0047] | m 0 ( ω ) | 2 = ( cos 2 ω 2 ) p ( sin 2 ω 2 ) - - - ( 4 )
[0048] In formula (4), m 0 ( ω ) = 1 2 Σ k = 0 2 N - 1 h k e - jkω .
[0049] Daubechies wavelets have the following characteristics:
[0050] 1) It is finitely supported in the time domain, that is, the length of Ψ(t) is limited.
[0051] 2) In the frequency domain Ψ(ω) has an Nth order zero at ω=0.
[0052] 3) Ψ(t) and its integer displacement are orthonormal, ie ∫Ψ(t)Ψ(t-k)dt=δk.
[0053] 4) The wavelet function Ψ(t) can be obtained from the so-called "scale function" Φ(t). The scaling function Φ(t) is a low-pass function with limited length, and the support domain is in the range of t=0~(2N-1).
[0054] The essence of wavelet transform is the convolution of the signal and the wavelet base. The time-frequency characteristics of the signal are obtained by projecting the signal on different wavelet scales. However, using different wavelet bases will have different time-frequency characteristics. Optional.
[0055] Therefore, according to the principle of wavelet transform modulus maximum, the wavelet transform of the input signal will have a wavelet coefficient modulus maximum value at the singular point, and the wavelet transform modulus maximum value at each scale will be selected first in the calculation.
[0056] Step S105, determine whether the obtained modulo maximum value is greater than 0; if so, execute the next step S106; if not, return to the step S103;
[0057] The specific process is that, since the length of the historical circular buffer is selected as a power frequency period, when the power grid signal does not change abruptly, the newly entered sampling point data has a great correlation with the adjacent data. The energy components of the frequency bands do not change massively. In the same principle, if the power grid signal is distorted, the newly entered sampling point will lose the correlation with the adjacent data, and the result after wavelet transform will have obvious component disturbance immediately. Therefore, in the detection process, it will be found that the detected voltage is a standard sine wave, the corresponding modulus maximum value is zero, and the modulus maximum corresponding to the sine wave containing the voltage sag start point (ie, the appearance of the SAG signal) is The value will show a jump at the starting point of the voltage sag, that is, the maximum value of the modulo value obtained at the starting point of the voltage sag is > 0. If the maximum value of the modulo corresponding to the sine wave jumps, perform step S106, if the sine wave If the corresponding modulus maximum value does not jump, then return to step S103, and re-detect whether the modulus maximum value corresponding to the sine wave jumps.
[0058] Step S106 , determining that the SAG signal appears in the power grid, and determining that the sampling time of the next sampling point is the start moment when the SAG signal appears.
[0059] The specific process is that when the modulo maximum value solved by the new sampling point is greater than 0, the result after wavelet transformation of the newly entered sampling point will immediately appear obvious component disturbance, and this disturbance can be used as the criterion that the power grid parameters have been distorted. The SAG signal is reported in real time, so that it can be determined that the SAG signal appears in the power grid, and the sampling time of the new sampling point is obtained as the starting moment of the SAG signal appearing.
[0060] When the disturbance occurs, the sampling value is compared with the wavelet inverse transform result sent from the historical data, and the positive and negative sequence components of the asymmetric grid voltage are used to separate the negative sequence components to eliminate the disturbance of the voltage sag detection, and then the grid parameter changes can be obtained. As the control parameter of the DVR compensation device, the method further includes: obtaining the current voltage amplitude of the power grid at the starting moment when the SAG signal appears and the original voltage amplitude corresponding to the previous sampling point at the starting moment value, and the difference between the obtained current voltage amplitude and the original voltage amplitude is used as a control parameter for voltage compensation of the DVR device.
[0061] like figure 2 As shown, it is a system for detecting a SAG signal in a DVR device provided by an embodiment of the present invention, and the system includes:
[0062] A circular buffer unit 110 is set for obtaining the same number of sampling points of the power grid signal in each power frequency period according to a preset sampling frequency and a power frequency period, and setting a loop with a length equal to the number of the sampling points buffer;
[0063] The sampling value storage and sorting unit 120 is used for obtaining the sampling value corresponding to each sampling point on a power frequency period, and storing the obtained sampling values in the circular buffer in the order of sampling time to form a sliding analysis window;
[0064] The sampling value updating unit 130 is configured to obtain the sampling value of the next sampling point and store it in the analysis window when the sampling time of the next sampling point arrives, and discard the sample with the earliest sampling time in the analysis window After the value, update the analysis window;
[0065] A wavelet analysis unit 140, configured to perform wavelet analysis on each sampled value in the updated analysis window to obtain a modulus maximum value corresponding to the next sampling point at its sampling time;
[0066] Judging unit 150, configured to judge whether the obtained modulo maximum value is greater than 0;
[0067] The SAG signal determining unit 160 is configured to determine that the SAG signal appears in the power grid, and determine the sampling time of the next sampling point as the starting moment of the SAG signal appearing.
[0068] Wherein, the system further includes a voltage sag amplitude unit 170, and the voltage sag amplitude unit 170 is configured to obtain the current voltage amplitude of the power grid at the starting moment when the SAG signal occurs and the previous sample at the starting moment The original voltage amplitude corresponding to the point, and the difference between the obtained current voltage amplitude and the original voltage amplitude is used as a control parameter for voltage compensation of the DVR device.
[0069] Wherein, the sampling frequency is 6.4KHz, and the power frequency period is 20ms; or the sampling frequency is 12.8KHz, and the power frequency period is 20ms.
[0070] Implementing the embodiment of the present invention has the following beneficial effects:
[0071] In this embodiment of the present invention, since a circular buffer with a fixed length for storing sample values is established, when the sampling time of each new sampling point arrives, only one data is updated in the circular buffer and wavelet is performed at a time. Analysis and detection of the starting time and amplitude of the SAG signal, and the other sampling values existing in the circular buffer do not need to be recalculated, which can save calculation time and make real-time detection based on this incremental wavelet analysis method. The starting time and amplitude of the SAG signal achieve the purpose of reducing the time delay of SAG fault detection in the DVR device.
[0072] It is worth noting that, in the above system embodiment, each system unit included is only divided according to functional logic, but is not limited to the above division, as long as the corresponding function can be realized; in addition, the specific The names are only for the convenience of distinguishing from each other, and are not used to limit the protection scope of the present invention.
[0073] Those of ordinary skill in the art can understand that all or part of the steps in the method of the above embodiments can be completed by instructing the relevant hardware through a program, and the program can be stored in a computer-readable storage medium, and the storage Media such as ROM/RAM, magnetic disk, optical disk, etc.
[0074] The above disclosures are only preferred embodiments of the present invention, and of course, the scope of the rights of the present invention cannot be limited by this. Therefore, equivalent changes made according to the claims of the present invention are still within the scope of the present invention.
