A method for measuring power grid frequency based on time-frequency atomic dictionary

Abstract

A method for measuring power grid frequency based on a time-frequency atomic dictionary, comprising the following steps: obtaining a distorted signal in a power grid, searching for a time-frequency atomic number and calculating a power grid frequency; wherein a power grid fundamental wave matching algorithm based on a time-frequency atomic dictionary is used to search for a time-frequency atomic number, and a power grid frequency measurement algorithm is used to calculate a power grid frequency; the method can project the obtained power grid distortion signal onto a time-frequency atomic dictionary that conforms to the characteristics of a power grid fundamental wave signal, and can search for a time-frequency atomic number and calculate a power grid frequency only by calculating a maximum matching value, and has the advantages of a simple algorithm structure, low computational complexity and high measurement accuracy; in addition, the method is more suitable for applications where voltage and current distortion are severe, and has the advantages of flexible and adjustable measurement range and resolution.

Description

Technical Field

[0001] The present invention belongs to the technical field of power grid frequency measurement, and in particular relates to a power grid frequency measurement method based on a time-frequency atomic dictionary. Background Art

[0002] The grid frequency can reflect the dynamic balance between power consumption and load. It is an important indicator for measuring the power quality of the power system and an important reference for regulating the normal operation of the power system.

[0003] The fluctuation range required for the frequency of my country's power grid is 50±1Hz, but due to the influence of various impacts, asymmetry and nonlinear loads, the power grid frequency often exceeds the fluctuation range. When the power grid frequency is too high or too low, it is easy to trigger a series of protection devices such as low-frequency load reduction or high-frequency machine cutting to malfunction. Therefore, the power grid frequency measurement method is a key technology in power quality analysis, power system control and safety assessment. At present, the commonly used power grid frequency measurement methods can be divided into zero-crossing method and transformation method. The zero-crossing method is mainly based on phase-locked loop technology, and the zero-crossing detection circuit converts the measured signal into a square wave of the same frequency and calculates the frequency value. The transformation law is mainly based on discrete Fourier transform (DFT), wavelet transform (WT), S transform (ST), Hilbert-Huang transform (HHT), etc. The sampled signal is projected into the frequency domain or time-frequency domain space through various transformations to obtain its frequency value.

[0004] However, as the problem of power quality disturbance pollution becomes more and more prominent, the voltage and current waveforms in the power grid are seriously distorted. For example, instantaneous pulses, voltage sags, harmonics, etc. These disturbances are superimposed on the power frequency sine wave, making the problem of power grid frequency measurement more complicated. In order to accurately measure the power grid frequency from the disturbance signal, the phase-locked loop technology based on the zero-crossing method has a high hardware development cost, so various transformation methods represented by DFT are more widely used. However, DFT is affected by factors such as sampling frequency, signal type, and noise. When processing non-stationary disturbance signals, spectrum leakage is prone to occur, resulting in a decrease in the accuracy of power grid frequency detection. Although WT can extract contour information and mutation information from non-stationary disturbance signals, it cannot directly obtain power grid frequency information. DFT is still required for the decomposed low-frequency signal in the later stage. ST and HHT have the advantages of both DFT and WT, but both methods have the problem of large computational complexity and it is difficult to meet the real-time requirements of the system. Summary of the invention

[0005] In order to overcome the shortcomings of the above-mentioned prior art, the purpose of the present invention is to provide a grid frequency measurement method based on a time-frequency atomic dictionary, which aims at how to quickly and accurately extract the fundamental signal from the distorted signal containing power quality disturbance pollution and calculate the grid frequency. Without adding hardware costs, the grid frequency can be quickly and accurately measured from the distorted signal containing power quality disturbance pollution.

[0006] To achieve the above object, the technical solution adopted by the present invention is:

[0007] A method for measuring power grid frequency based on a time-frequency atomic dictionary comprises the following steps:

[0008] Step 1: Obtain the distorted signal y(n) in the power grid

[0009] y(n)=s(n)+d(n) (1)

[0010] Among them, s(n) is the grid fundamental signal, d(n) is the power quality disturbance signal, and y(n) can be obtained through sensors and data acquisition cards installed at the grid frequency to be measured;

[0011] Step 2: Search for the time-frequency atomic number j

[0012] j=T1[y(n)] (2)

[0013] Among them, T1[·] is a grid fundamental wave matching algorithm based on the time-frequency atomic dictionary, which can be implemented in a measurement unit installed at the location where the grid frequency is to be measured;

[0014] Step 3: Calculate the grid frequency f

[0015] f=T2[j] (3)

[0016] Among them, T2[·] is a grid frequency calculation algorithm based on a time-frequency atomic dictionary, which can be implemented in a measurement unit installed at the location where the grid frequency is to be measured.

[0017] The digital angular frequency w of the grid fundamental signal s(n) is 2πf / f s is the fundamental angular frequency, f is the grid frequency, and f s is the sampling frequency.

[0018] The power quality disturbance signal d(n) has additive signal characteristics and has the following classifications, specifically:

[0019] d1(n)=a1(u(nN 1_1 )-u(nN 1_2 )) (4)

[0020] Among them, d1(n) is the instantaneous pulse sequence, a1 is the instantaneous pulse amplitude, N 1_1 is the instantaneous pulse starting time, N 1_2 is the instantaneous pulse termination moment; n is the sampling moment, u is the unit step sequence;

[0021] d2(n)=a2sin(w2n)(u(nN 2_1 )-u(nN 2_2 )) (5)

[0022] Where d2(n) is the voltage sag sequence, a2 is the voltage sag amplitude, w2=w is the power frequency angular frequency, N 2_1 is the starting time of voltage sag, N 2_2 is the voltage sag termination time;

[0023] d3(n)=a3sin(w3n) (6)

[0024] Wherein, d3(n) is the harmonic sequence, a3 is the harmonic amplitude, w3=iw is the harmonic angular frequency, and i is an integer greater than 1.

[0025] The grid fundamental wave matching algorithm T1[·] based on the time-frequency atomic dictionary comprises the following steps:

[0026] Step 1: Construct the time-frequency atom g j (n)

[0027] g j (n) = sin(w j n) (7)

[0028] Among them, w j =2πn(f down +j*f r ) / f s is the atomic frequency, f s is the sampling frequency, f r is the frequency resolution, j∈[0,1,…,C], C=(f up -f down ) / f r , f down is the lower limit of grid frequency fluctuation, f up is the upper limit of grid frequency fluctuation;

[0029] Step 2: Create a time-frequency atom dictionary A

[0030] A=[g1(n),g2(n),…,g C (n)] (8)

[0031] Among them, A is g j(n) composed of a time-frequency atom dictionary;

[0032] Step 3: Calculate the maximum matching value M

[0033] M=max|<y(n),A> | (9)

[0034] Among them, y(n) is the obtained power grid distortion signal, A is the time-frequency atomic dictionary;

[0035] Step 4: Search for the time-frequency atomic number j

[0036] j=pos(M) (10)

[0037] Among them, j is the searched time-frequency atomic number;

[0038] The grid frequency measurement algorithm T2[·] is specifically:

[0039] f=(f down +j*f r ) (11)

[0040] Among them, f is the calculated grid frequency, j is the searched time-frequency atomic number, and f r is the frequency resolution, f down It is the lower limit of grid frequency fluctuation.

[0041] The advantages and positive effects of the present invention are:

[0042] 1) The method proposed in the present invention can directly extract the grid fundamental wave from the distorted signal containing power quality disturbance and measure the grid frequency, which is suitable for occasions where voltage and current distortion are serious.

[0043] 2) The method proposed in the present invention can project the distorted signal containing power quality disturbance onto the time-frequency atom dictionary that meets the characteristics of the fundamental signal. The atom that matches the fundamental signal of the power grid can be determined by simply calculating the maximum projection value. Compared with the zero-crossing detection method, DFT, WT, ST and HHT, this method has the advantages of simple algorithm and low computational complexity. Therefore, the power grid frequency detection system based on the method of the present invention has the positive effects of low hardware development cost and good real-time performance;

[0044] 3) The power grid frequency detection range and accuracy of the method proposed in the present invention can be determined by the lower limit frequency, upper limit frequency and resolution of the time-frequency atom, so the measurement range and resolution of the method can be flexibly set through software. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 A block diagram of an implementation scheme of the method of the present invention.

[0046] FIG. 2( a ) is a flowchart of the implementation of algorithm T1[·] in the method of the present invention.

[0047] FIG2( b ) is a flowchart of the implementation of algorithm T2[·] in the method of the present invention.

[0048] Figure 3(a) is a diagram of a distorted signal containing instantaneous pulse disturbance and noise.

[0049] Figure 3(b) is the grid fundamental signal searched from the distorted signal.

[0050] FIG4( a ) is a diagram of a distorted signal including voltage sag and noise.

[0051] Figure 4(b) is the grid fundamental signal searched from the distorted signal.

[0052] Figure 5(a) is a diagram of a distorted signal containing harmonics and noise.

[0053] Figure 5(b) is the grid fundamental signal searched from the distorted signal.

[0054] Figure 6(a) is a diagram of a distorted signal containing composite disturbance and noise.

[0055] Figure 6(b) is the grid fundamental signal searched from the distorted signal. DETAILED DESCRIPTION

[0056] The present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. The content of the present invention provides a theoretical basis and guidance for the specific implementation of the method of the present invention.

[0057] like Figure 1 As shown, the implementation process of the present invention includes the following steps:

[0058] Step 1: Obtain the distorted signal y(n) in the power grid

[0059] y(n)=s(n)+d(n) (1)

[0060] Among them, s(n) is the grid fundamental signal, d(n) is the power quality disturbance signal, and y(n) can be obtained through sensors and data acquisition cards installed at the grid frequency to be measured;

[0061] Step 2: Search for the time-frequency atomic number j

[0062] j=T1[y(n)] (2)

[0063] Among them, T1[·] is a grid fundamental wave matching algorithm based on the time-frequency atomic dictionary, which can be implemented in a measurement unit installed at the location where the grid frequency is to be measured;

[0064] Step 3: Calculate the grid frequency f

[0065] f=T2[j] (3)

[0066] Among them, T2[·] is a grid frequency calculation algorithm based on a time-frequency atomic dictionary, which can be implemented in a measurement unit installed at the location where the grid frequency is to be measured.

[0067] As shown in FIG2 , the implementation process of the algorithm T1[·] in the method of the present invention is as follows:

[0068] Step 1, input the acquired power grid distortion signal y(n) and the constructed time-frequency atomic dictionary A;

[0069] Step 2: Calculate the maximum matching value M

[0070] M=max|<y(n),A> | (4)

[0071] Among them, y(n) is the acquired power grid distortion signal, A is the time-frequency atomic dictionary, and M is the maximum matching value;

[0072] Step 3: Search for the time-frequency atomic number j

[0073] j=pos(M) (5)

[0074] Among them, j is the searched time-frequency atomic number;

[0075] Step 4, output the time-frequency atomic number j.

[0076] As shown in FIG2 , the implementation process of the algorithm T2[·] in the method of the present invention is as follows:

[0077] Step 1: Enter the searched time-frequency atomic number j and frequency resolution f r And the frequency fluctuation lower limit f down ;

[0078] Step 2: Calculate the grid frequency f

[0079] f=(f down +j*f r ) (6)

[0080] Among them, f is the calculated grid frequency, j is the searched time-frequency atomic number, and f r is the frequency resolution, f down is the lower limit of grid frequency fluctuation;

[0081] Step 3: output the grid frequency f.

[0082] use Figure 1Four examples were set up according to the implementation scheme and process shown in Figure 2. All four examples were run on an Intel(R) Core(TM) i7-8565U CPU machine, the software version was Matlab 2016, and the sampling frequency was set to f s =10kHz, the sampling duration is set to t=0.2ms, and the number of sampling points is set to N=2000.

[0083] The grid fundamental signal s(n) in the embodiment described is:

[0084] s(n)=sin(wn) (7)

[0085] Where, the angular frequency of s(n) is w = 2πf / f s , f is the grid frequency, f s is the sampling frequency.

[0086] The power quality disturbance signal d(n) in the embodiment described has additive signal characteristics and has the following categories, specifically:

[0087] d1(n)=a1(u(nN 1_1 )-u(nN 1_2 )) (8)

[0088] Among them, d1(n) is the instantaneous pulse sequence, a1 is the instantaneous pulse amplitude, N 1_1 is the instantaneous pulse starting time, N 1_2 is the instantaneous pulse termination moment; n is the sampling moment, u is the unit step sequence;

[0089] d2(n)=a2sin(w2n)(u(nN 2_1 )-u(nN 2_2 )) (9)

[0090] Where d2(n) is the voltage sag sequence, a2 is the voltage sag amplitude, w2=w is the power frequency angular frequency, N 2_1 is the starting time of voltage sag, N 2_2 is the voltage sag termination time;

[0091] d3(n)=a3sin(w3n) (10)

[0092] Wherein, d3(n) is the harmonic sequence, a3 is the harmonic amplitude, w3=iw is the harmonic angular frequency, and i is an integer greater than 1.

[0093] The grid distortion signal y(n) in the embodiment described is:

[0094] y(n)=s(n)+d(n) (11)

[0095] Among them, s(n) is the grid fundamental signal, d(n) is the power quality disturbance signal;

[0096] Embodiment 1:

[0097] Formulas (7), (8) and (11) are used to generate a distorted signal containing a grid fundamental wave and an instantaneous pulse sequence, as shown in FIG3(a). The grid frequency calculated by the method of the present invention is f=49.5 Hz, and the output grid fundamental wave signal is shown in FIG3(b).

[0098] Embodiment 2:

[0099] The grid distortion signal is generated by using formulas (7), (9) and (11), as shown in FIG4(a). The grid frequency calculated by the method of the present invention is f=49.5 Hz, and the output grid fundamental wave signal is shown in FIG4(b).

[0100] Embodiment 3:

[0101] The grid distortion signal is generated by using formulas (7), (10) and (11), as shown in FIG5(a). The grid frequency calculated by the method of the present invention is f=49.5 Hz, and the output grid fundamental wave signal is shown in FIG5(b).

[0102] Embodiment 4:

[0103] Formulas (7), (9), (10) and (11) are used to generate a grid distortion signal, as shown in FIG6(a). The grid frequency calculated by the method of the present invention is f=49.5 Hz, and the output grid fundamental wave signal is shown in FIG6(b).

[0104] Table 1 shows the time-frequency atomic numbers j searched out by the method of the present invention under different noises, the calculated grid frequency f, and the comparison results with the DFT transformation method in the four embodiments.

[0105] Table 1

[0106]

[0107] It can be seen from Table 1 that the method of the present invention is not easily affected by noise, has good anti-noise robustness, and has better measurement accuracy than the DFT transformation method.

[0108] In summary, the method of the present invention can quickly and accurately measure the grid frequency from the distorted signal containing power quality disturbance. It not only has the advantages of low computational complexity, simple measurement structure, and flexible and adjustable measurement range, but also can effectively reduce the hardware development cost of the grid frequency detection device.

[0109] The contents not described in detail in the specification of the present invention belong to the prior art known to the professional and technical personnel in this field.

Claims

1. A method for measuring power grid frequency based on a time-frequency atomic dictionary, characterized in that: The following steps are involved: Step 1: Obtain the distorted signal y(n) in the power grid y(n)=s(n)+d(n) (1) Among them, s(n) is the grid fundamental signal, d(n) is the power quality disturbance signal, and y(n) can be obtained through sensors and data acquisition cards installed at the grid frequency to be measured; Step 2: Search for the time-frequency atomic number j j=T1[y(n)] (2) Among them, T1[·] is a grid fundamental wave matching algorithm based on the time-frequency atomic dictionary, which can be implemented in a measurement unit installed at the location where the grid frequency is to be measured; Step 3: Calculate the grid frequency f f=T2[j] (3) Wherein, T2[·] is a grid frequency value calculation algorithm, which is implemented in a measurement unit installed at a location where the grid frequency is to be measured.

2. A method for measuring power grid frequency based on a time-frequency atomic dictionary according to claim 1, characterized in that: The digital angular frequency of the grid fundamental signal s(n) is w=2πf / f s , where f is the grid frequency, f s is the sampling frequency.

3. The method for measuring power grid frequency based on a time-frequency atomic dictionary according to claim 1, characterized in that: The power quality disturbance signal d(n) has additive signal characteristics and has the following classifications, specifically: d1(n)=a1(u(nN 1_1 )-u(nN 1_2 )) (4) Among them, d1(n) is the instantaneous pulse sequence, a1 is the instantaneous pulse amplitude, N 1_1 is the instantaneous pulse starting time, N 1_2 is the instantaneous pulse termination moment; n is the sampling moment, u is the unit step sequence; d2(n)=a2sin(w2n)(u(nN 2_1 )–and(nN 2_2 )) (5) Where d2(n) is the voltage sag sequence, a2 is the voltage sag amplitude, w2=w is the power frequency angular frequency, N 2_1 is the starting time of voltage sag, N 2_2 is the voltage sag termination time; d3(n)=a3sin(w3n) (6) Wherein, d3(n) is the harmonic sequence, a3 is the harmonic amplitude, w3=iw is the harmonic angular frequency, and i is an integer greater than 1.

4. The method for measuring power grid frequency based on a time-frequency atomic dictionary according to claim 1, characterized in that: The grid fundamental wave matching algorithm T1[·] based on the time-frequency atomic dictionary comprises the following steps: Step 1: Construct the time-frequency atom g j (n) <h2 style=";text-align:left;direction:ltr">g<h2 style=";text-align:left;direction:ltr"> j <h2 style=";text-align:left;direction:ltr"> (n) = sin(w)<h2 style=";text-align:left;direction:ltr"> j <h2 style=";text-align:left;direction:ltr"> n) (7) Among them, w j =2πn(f down +j*f r ) / f s is the atomic frequency, f s is the sampling frequency, f r is the frequency resolution, j∈[0,1,…,C], C=(f up -f down ) / f r , f down is the lower limit of grid frequency fluctuation, f up is the upper limit of grid frequency fluctuation; Step 2: Create a time-frequency atom dictionary A A=[g1(n),g2(n),…,g C (n)] (8) Among them, A is composed of g j (n) composed of a time-frequency atom dictionary; Step 3: Calculate the maximum matching value M M=max|<y(n),A> | (9) Among them, y(n) is the obtained power grid distortion signal, A is the time-frequency atomic dictionary; Step 4: Search for the time-frequency atomic number j j=pos(M) (10) Among them, j is the searched time-frequency atomic number.

5. The method for measuring power grid frequency according to claim 1, characterized in that: The grid frequency measurement algorithm T2[·] is specifically: f=(f down +j*f r ) (11) Among them, f is the calculated grid frequency, j is the searched time-frequency atomic number, and f r is the frequency resolution, f down It is the lower limit of grid frequency fluctuation.

Images