A pilot protection method for active distribution network based on GFT

Abstract

The present invention discloses an active distribution network longitudinal protection method based on a road map GFT, belonging to the field of active distribution network line protection technology. The present invention utilizes the A-phase, B-phase, and C-phase current signals on both sides of the line in a time window before each sampling point to construct the road map of each phase on both sides, with each sampling point of each phase current as the vertex of the road map, and the instantaneous value of the current at each sampling point as the graph signal value of the vertex, and the adjacent vertices are sequentially connected to form the edge of the road map; the Laplace matrix of the road map of each phase on both sides is obtained, and the characteristic vector matrix is ​​obtained by orthogonal decomposition, and the graph Fourier transform GFT spectrum coefficient and the corresponding GFT spectrum coefficient smoothness are obtained, and the fault degree between the smoothness arrays of the same phase GFT spectrum coefficients on both sides is constructed, and the distribution network line longitudinal protection criterion is established. The present invention is not affected by the fault location, fault type, transition resistance, and noise, and is suitable for the case where the IIDG is connected to the downstream of the line. It can tolerate 100 ohm transition resistance and 20dB noise, and can reliably detect faulty lines.

Description

Technical Field

[0001] The present invention relates to the technical field of active distribution network line protection, and in particular to a longitudinal protection method for an active distribution network based on a GFT (Globally Coupled Futures Telecommunications) graph. Background Art

[0002] With the access of large-scale inverter-interfaced distributed generation (IIDG), the original single-power radial network structure of the distribution network has changed to a multi-terminal power supply system. Therefore, the power flow direction, operation mode and short-circuit current fault characteristics of the distribution network containing IIDG are complex and changeable. Conventional distribution network current protection will no longer be applicable, which is not conducive to the safe operation of the distribution network. The increase in IIDG penetration has also brought certain challenges to distribution network fault detection.

[0003] Wang Yingying, Zhang Jin, Li Yong (Longitudinal protection of distribution network containing IIDG based on dynamic matching of time series [J / OL]. Journal of Electric Power System and Automation, 1-10 [2024-04-18].) By analyzing the difference between the transient current waveforms of inverter-type distributed power sources and traditional power sources, and fully considering the influence of longitudinal synchronization error, in order to improve the ability of current differential protection to resist synchronization error of distribution network containing inverter-type distributed power sources, a longitudinal protection principle of distribution network containing inverter-type distributed power sources based on dynamic matching of time series is proposed. By dynamically matching the current sampling value sequence on both sides of the line to identify faults, the synchronization requirements of the sampling data on both sides of the line are reduced, and the average Euclidean distance of the time domain waveform is introduced to construct the protection criterion to improve the sensitivity. However, this method does not take into account the presence of noise. In the presence of certain noise interference, this method may fail.

[0004] Liao Xiaojun, Feng Xianzheng, Zhang Li, et al. (A method of characteristic vector centrality for power grid fault diagnosis based on recording start information [J]. Sichuan Electric Power Technology, 2021, 44(06): 1-6+44.) proposed a method for transmission grid fault diagnosis based on graph signal modeling based on fault recording start information and network node characteristic vector centrality algorithm. First, a fault start information network graph is constructed from the fault recording start information of the relay protection and fault information master station; secondly, the graph smoothness analysis method is used to determine whether the power grid has a fault and identify the fault type; finally, the fault component is identified based on the network node characteristic vector centrality algorithm and visualized. However, this method is mainly aimed at transmission lines, and does not consider the DG access situation.

[0005] The methods proposed in the above documents have their own shortcomings. None of them considers the access of distributed power sources or the influence of noise. When the distributed power sources are connected to the distribution network and there is noise interference, their algorithms may fail. In order to solve this problem, the present invention proposes an active distribution network longitudinal protection method based on the road map GFT. Summary of the invention

[0006] The purpose of the present invention is to propose an active distribution network longitudinal protection method based on a GFT graph to solve the problems raised in the background technology.

[0007] In order to achieve the above object, the present invention adopts the following technical solutions:

[0008] A method for longitudinal protection of an active distribution network based on a GFT diagram comprises the following steps:

[0009] S1, collect the currents of phase A, phase B, and phase C on the bus side of the line close to the power supply. When the three consecutive sampling values ​​of the sudden change of the current of any phase exceed the preset start threshold, the protection is started and the next step is entered; similarly, collect the currents of phase A, phase B, and phase C on the bus side of the line close to the load;

[0010] S2. Based on the A-phase, B-phase, and C-phase currents obtained in S1, a phase diagram on both sides of the line is constructed, with each sampling point of each phase as a vertex, the connection between each adjacent vertex as an edge, and each sampling value as an input signal of each vertex, to form an input signal array of each phase and an adjacency matrix of each phase diagram;

[0011] S3, for each road map of each phase on both sides of the line obtained in S2, calculate its Laplace matrix; then perform standard orthogonal decomposition on the obtained Laplace matrix to obtain the corresponding eigenvector matrix;

[0012] S4, based on the operations described in S2 to S3, for each road map, multiply its input signal array with the corresponding eigenvector matrix to obtain each GFT spectrum coefficient of the road map, form a corresponding GFT spectrum coefficient array, and then calculate the smoothness of the array;

[0013] S5, normalizing the smoothness of both sides obtained in S4 to obtain a normalized array of smoothness of both sides; calculating the absolute value of the difference between the normalized arrays of smoothness of each phase on both sides of the line to obtain a distance array between the normalized arrays of smoothness of each phase on both sides of the line;

[0014] S6. For each element in the distance array between the normalized smoothness arrays of each phase on both sides of the line obtained in S5, calculate the corresponding similarity coefficient, and then construct a similarity coefficient array corresponding to the distance array; use the obtained similarity coefficient array to calculate the fault degree of each phase of the line;

[0015] S7. Compare the fault degree of each phase of the line obtained in S6 with a preset protection threshold value to establish a line fault protection criterion.

[0016] Preferably, the S1 specifically includes the following contents:

[0017] Collect the current of phase A, phase B, and phase C on the bus m side of line mn close to the power supply When the current mutation of any phase exceeds 0.1 times the rated current for three consecutive sampling values, the protection starts; similarly, the A, B, and C currents of the bus n side of the line mn close to the load are collected.

[0018] Preferably, S2 specifically includes the following contents:

[0019] The A-phase current signal in the previous time window of each sampling point on the m side of the fault line mn obtained in S1 is used With each sampling point as a vertex, the connection between adjacent vertices as an edge, and each sampling value as the input signal of each vertex, construct the input signal array S ignal_ma ; Construct the current sampling point of phase A current on the m side of line mn. ma , the G ma Include p vertices; construct the adjacency matrix W ma , the W ma The diagonal elements are 0, for W ma For non-diagonal elements, the weight of the edge between two adjacent vertices is 1, and the weight of the edge between two non-adjacent vertices is 0;

[0020] Similarly, the B-phase and C-phase diagrams G of the m-side of line mn are constructed respectively. mb , G mc and the input signal array S ignal_mb , S ignal_mc And the n-side A phase, B phase, C phase diagram G na , G nb , G nc and the input signal array S ignal_na , S ignal_nb , S ignal_nc .

[0021] Preferably, S3 specifically includes the following contents:

[0022] S3.1. For each road map of the three-phase on both sides, calculate its Laplace matrix L. The specific calculation formula is:

[0023] L=DW

[0024] Where D represents the degree matrix; the diagonal elements of D are the degrees of node i It represents the sum of the weights of nodes connected to node i in the road graph, and the off-diagonal elements of D are 0;

[0025] S3.2. Perform standard orthogonal decomposition on the matrix L to obtain its eigenvector matrix V, where each column in V is an eigenvector.

[0026] Preferably, the S4 specifically includes the following contents:

[0027] S4.1. For the A phase diagram G of the m side of line mn in the cycle time window before each sampling time t after the line fault ma , the input signal array S ignal_ma Multiplying with the corresponding matrix V, we get the phase diagram G of side A at sampling time t ma The GFT spectral coefficient array f ma , the f ma Contains p elements; the lth spectral coefficient f ma (l) is the sum of the multiplication of each element in the input signal array with the corresponding element in the l-th column eigenvector in V, that is:

[0028]

[0029] S4.2. Calculate the phase A diagram G of the m side of line mn at sampling time t ma The GFT spectral coefficient array f ma Smoothness S ma (t), that is:

[0030] S ma (t) = f ma Lf ma T (9)

[0031] In the formula, the superscript T represents the transpose operation of the array;

[0032] S4.3, repeat the operations described in S4.1 to S4.2 above to obtain the smoothness S of the B phase and the C phase on the m side at the sampling time t mb (t), S mc (t), smoothness S of phase A, phase B, and phase C on the n side na (t), S nb (t), S nc (t);

[0033] S4.4. Using the smoothness of each moment in the cycle before the current moment, obtain the smoothness array S of phase A, phase B, and phase C on the m side of line mn ma , S mb , S mc , n-side A phase, B phase, C phase smoothness array S na , S nb, S nc .

[0034] Preferably, the S5 specifically includes the following contents:

[0035] S5.1 normalizes each smoothness and converts it to the interval [0,1]; for the smoothness array S of phase A on side m ma The i-th element S in ma (i), the normalized smoothness value is S ma_n (i)

[0036]

[0037] Through the above processing, we can get the normalized array S of the smoothness of phase A on the m side: ma_n ;

[0038] S5.2, repeat the operation described in S5.1, and obtain the normalized smoothness array S of the B phase and the C phase on the m side through normalization processing. mb_n , S mc_n ; The smoothness normalized array S of phase A, phase B, and phase C on the n side na_n , S nb_n , S nc_n ;

[0039] S5.3. Calculate the normalized array S of the smoothness of phase A on both sides of line mn ma_n , S na_n The i-th element S ma_n (i) and S na_n The absolute value of the difference between (i) and (i) is taken as the i-th smoothness distance d of phase A. a (i)

[0040] d a (i)=|S ma_n (i)-S na_n (i)| (11)

[0041] Through the above operation, the distance array d between the normalized arrays of phase A smoothness on both sides of line mn is obtained. a , referred to as the A phase distance array d of line mn a , the value range of each element is [0,1];

[0042] S5.4, repeat the operation described in S5.3 to further obtain the distance array d between the normalized smoothness arrays of phase B and phase C on both sides of line mn b ,d c , referred to as the B-phase and C-phase distance array d of line mn b ,d c .

[0043] Preferably, S6 specifically includes the following content:

[0044] S6.1. When the line mn is not faulty or a fault occurs outside the zone, each element in the distance arrays of each phase of the line mn is close to 0; conversely, when a fault occurs in the line mn, at least one element in the distance array of one phase has a value greater than 0. For the k-th element d a (k) of the A-phase distance array, calculate the corresponding similarity coefficient S da (k):

[0045]

[0046] In the formula, ε represents the segmentation threshold;

[0047] S6.2. When the smoothness distance d(k) ≤ ε, it indicates that the smoothness of the same phase on both sides is similar at this moment, and the corresponding similarity coefficient is 1; when ε < d(k) < 5ε, it is determined that the smoothness of the same phase on both sides is moderately similar at this moment, and the corresponding similarity coefficient is ε / d(i); when d(k) ≥ 5ε, it indicates that the smoothness of the same phase on both sides is not similar at this moment, and the corresponding similarity coefficient is 0;

[0048] Thus, the similarity coefficient array S da of the smoothness distance of the A-phase of the line within the previous cycle of the current moment is obtained, which is simply referred to as the A-phase similarity coefficient array S da ;

[0049] S.6.3. Repeat the operations described in S6.1 to S6.2 to further obtain the similarity coefficient arrays S db and S dc of the smoothness distances of the B-phase and C-phase within the previous cycle of the current moment, which are simply referred to as the B-phase and C-phase similarity coefficient arrays S db and S dc ;

[0050] S6.4. Using the A-phase similarity coefficient array S da , calculate the fault degree F a of the A-phase of the line at the current moment:

[0051]

[0052] In the formula, N is the number of elements in the similarity coefficient array;

[0053] S6.5. Repeat the above-mentioned in S6.4 to construct the fault degrees F b and F c of the B-phase and C-phase of the line;

[0054] S6.6. When the line mn operates normally or an out-of-zone fault occurs, the similarity coefficients in the similarity coefficient arrays of each phase on both sides of the line mn are close to 1, and the corresponding fault degree F is close to 0; conversely, when an in-zone fault occurs on the line mn, the similarity coefficients in the similarity coefficient arrays of at least one phase are close to 0, and the fault degree F of the phase is close to 1.

[0055] Preferably, the S7 specifically includes the following contents:

[0056] The fault degrees F of phase A, phase B and phase C of the line are respectively a 、F b 、F c With protection threshold F set By comparison, the line fault protection criterion is established:

[0057] (F a >F set )|| (F b >F set )|| (F c >F set ) (14)

[0058] In the formula, || represents the logical relationship "or";

[0059] If the fault degree of each phase of the line is F a 、F b 、F c If equation (7) is satisfied, it is judged that the line has an in-area fault, otherwise it is judged that the line is operating normally.

[0060] Compared with the prior art, the present invention provides an active distribution network longitudinal protection method based on a road map GFT, which has the following beneficial effects:

[0061] The present invention proposes a method for longitudinal protection of active distribution network based on road map GFT, which utilizes the current signals of each phase on both sides in a time window before each sampling point, takes each current sampling point as each vertex of the road map, and the graph signal value of each vertex is the instantaneous value of the current at the sampling point. The edges of the road map are connected in sequence between adjacent vertices to construct the road map of each phase on both sides; the Laplace matrix of the road map of each phase on both sides is obtained, and the characteristic vector matrix is ​​obtained by orthogonal decomposition, and the graph Fourier transform GFT spectrum coefficient is calculated, and then the corresponding GFT spectrum coefficient smoothness array is obtained, and the fault degree between the GFT spectrum coefficient smoothness arrays of the same phases on both sides is constructed to establish the line longitudinal protection criterion. The protection method proposed by the present invention is not affected by factors such as fault location, fault type, transition resistance, noise, etc., and can be adapted to the situation where an IIDG is connected to the downstream of the line, can tolerate a transition resistance of 100 ohms and a noise of 20dB, and can reliably detect the fault line. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 This is a flow chart of an active distribution network longitudinal protection method based on a road map GFT proposed by the present invention;

[0063] Figure 2 This is a topological diagram of the IIDG distribution network mentioned in Example 1 of the present invention. DETAILED DESCRIPTION

[0064] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0065] See also Figure 1 The present invention proposes an active distribution network longitudinal protection method based on a GFT, which specifically includes the following steps:

[0066] Step 1: Collect the current of phase A, phase B, and phase C on the bus m side of line mn close to the power supply When the current mutation of any phase exceeds 0.1 times the rated current for three consecutive sampling values, the protection starts; similarly, the A, B, and C currents of the bus n side of the line mn close to the load are collected.

[0067] Step 2: Use the A-phase current signal in the previous time window of each sampling point on the m side of the line mn obtained in step 1 after the fault With each sampling point as a vertex, the connection between adjacent vertices as an edge, and each sampling value as the input signal of the vertex, construct the input signal array S ignal_ma , construct the current sampling point of phase A current on the m side of line mn. ma , including p vertices, construct the adjacency matrix W ma , W ma The diagonal elements are 0, for W ma For non-diagonal elements, the weight of the edge between two adjacent vertices is 1, and the weight of the edge between two non-adjacent vertices is 0; similarly, the B-phase and C-phase graphs G of the m-side of line mn are constructed respectively. mb , G mc and the input signal array S ignal_mb , S ignal_mc And the n-side A phase, B phase, C phase diagram G na , G nb , G nc and the input signal array S ignal_na , S ignal_nb , S ignal_nc ;

[0068] Step 3: For each three-phase road map G on both sides, calculate its Laplace matrix L, L = DW, where D is the degree matrix and the diagonal elements of D are the degrees of node i represents the sum of the weights of the nodes connected to node i in the road graph, and the off-diagonal elements of D are 0; then, the matrix L is subjected to standard orthogonal decomposition to obtain its eigenvector matrix V, where each column is an eigenvector;

[0069] Step 4: for the A phase path diagram G of the m side of the line mn in the cycle time window before each sampling time t after the above line fault ma , the input signal array S ignal_ma Multiplying with the corresponding matrix V, we get the phase diagram G of side A at sampling time t ma The GFT spectral coefficient array f ma , contains p elements, the lth spectral coefficient f ma (l) is the sum of the multiplication of each element in the input signal array with the corresponding element in the l-th column eigenvector in V, that is:

[0070]

[0071] Calculate the phase A diagram G of the m side of line mn at sampling time t ma The GFT spectral coefficient array f ma Smoothness S ma (t), that is:

[0072] S ma (t) = f ma Lf ma T (16)

[0073] In the formula, the superscript T represents the transpose operation of the array;

[0074] Similarly, the smoothness S of phase B and phase C on side m at sampling time t is obtained mb (t), S mc (t), smoothness S of phase A, phase B, and phase C on the n side na (t), S nb (t), S nc (t);

[0075] Using the smoothness of each cycle before the current moment, we can get the smoothness array S of phase A, phase B, and phase C on the m side of line mn. ma , S mb , S mc , n-side A phase, B phase, C phase smoothness array S na , S nb , S nc ;

[0076] Step 5: The in-phase smoothness array S of both sides after the line fault m , S n The element values ​​of are very different, so each smoothness is normalized and the value is converted to [0,1]; for the smoothness array S of the A phase on the m side ma The i-th element S in ma (i), the normalized smoothness value is S ma_n (i)

[0077]

[0078] Get the normalized array S of the smoothness of phase A on side m ma_n ;

[0079] Similarly, through normalization, we get the normalized smoothness array S of phase B and phase C on the m side: mb_n , S mc_n ; The smoothness normalized array S of phase A, phase B, and phase C on the n side na_n , S nb_n , S nc_n ;

[0080] Calculate the normalized array S of the smoothness of phase A on both sides of line mn ma_n , S na_n The i-th element S ma_n (i) and S na_n The absolute value d of the difference between (i) a (i)

[0081] d a (i)=|S ma_n (i)-S na_n (i)| (18)

[0082] Thus, we can obtain the distance array d between the normalized arrays of phase A smoothness on both sides of line mn a , referred to as the A phase distance array d of line mn a , their values ​​range is [0,1];

[0083] Similarly, the distance array d between the normalized smoothness arrays of phase B and phase C on both sides of line mn is obtained. b ,d c , referred to as the B-phase and C-phase distance array d of line mn b ,d c ;

[0084] Step 6: When a fault occurs on line mn, the values ​​of the elements in the distance array of at least one phase are large. When line mn is not faulty or an out-of-area fault occurs, the values ​​of the elements in the distance arrays of each phase of line mn are close to 0.

[0085] For the k-th element d a (k) of the A-phase distance array, construct the corresponding similarity coefficient S da (k):

[0086]

[0087] In the formula, ε is the segmentation threshold, and take ε = 0.1;

[0088] When the distance d(k) ≤ ε, it indicates that the in-phase smoothness on both sides at this moment is similar, and the corresponding similarity coefficient is 1; when ε < d(k) < 5ε, it is judged that the in-phase smoothness on both sides at this moment is moderately similar, and the corresponding similarity coefficient is ε / d(i); when d(k) ≥ 5ε, it indicates that the in-phase smoothness on both sides at this moment is not similar, and the corresponding similarity coefficient is 0;

[0089] Thus, the similarity coefficient array S of the line A-phase smoothness distance within the previous cycle of the current moment is obtained da , simply referred to as the A-phase similarity coefficient array S da ;

[0090] Similarly, the similarity coefficient arrays S db and S dc of the line B-phase and C-phase smoothness distances within the previous cycle of the current moment are obtained, simply referred to as the B-phase and C-phase similarity coefficient arrays S db and S dc ;

[0091] Using the A-phase similarity coefficient array S da , construct the line A-phase fault degree F a at the current moment:

[0092]

[0093] In the formula, N is the number of elements in the similarity coefficient array;

[0094] Similarly, construct the line B-phase and C-phase fault degrees F b and F c ;

[0095] When a fault occurs in the protected zone of line mn, at least one phase has relatively large elements in the distance array of smoothness, the corresponding similarity coefficients in the similarity coefficient array are close to 0, and the corresponding fault degree F is close to 1; while during normal operation or external faults, the similarity coefficients in the similarity coefficient arrays of each phase on both sides of line mn are close to 1, and the corresponding fault degree F is close to 0;

[0096] Step 7. Respectively compare the line A-phase, B-phase, and C-phase fault degrees F a and F b and F c with the protection threshold F setBy comparison, the line fault protection criterion is established as follows:

[0097] (F a >F set )|| (F b >F set )|| (F c >F set ) (twenty one)

[0098] In the formula, || represents the logical relationship "or";

[0099] Considering the influence of noise and asynchronism, the protection threshold F is taken set =0.5;

[0100] If the fault degree of each phase of the line is F a 、F b 、F c If equation (7) is satisfied, it is judged that the line has an in-area fault, otherwise it is judged that the line is operating normally.

[0101] Based on the above content, an active distribution network longitudinal protection method based on a road map GFT proposed by the present invention is described below in combination with specific examples and drawings. The specific content is as follows.

[0102] Embodiment 1:

[0103] In order to verify the effectiveness of this protection method, the PSCAD / EMTDC simulation software is used to build the following Figure 2 The 10kV active distribution network model shown in the figure has a base voltage of 10.5kV, a transformer capacity of 50MVA, line parameters of (0.17+j0.34)Ω / km, a length of 4km, an IIDG capacity of 2.5MW, and load active and reactive powers of 0.8MW and 0.4MVA, respectively.

[0104] It is assumed that the line fails at 2.4s and the failure lasts for 0.5s.

[0105] 1) Short circuit fault test

[0106] use Figure 2 In the active distribution network model, when short-circuit faults in the AG, AB, ABG, and ABC zones occur at 5%, 50%, and 95% of the line B1C with 0.01Ω and 100Ω transition resistance, the corresponding simulation results are shown in Tables 1 and 2. AG, AB, ABG, and ABC represent phase A grounding fault, phase A and phase B short-circuit fault, phase A and phase B short-circuit grounding fault, and three-phase short-circuit grounding fault, respectively.

[0107] Table 1 Fault simulation results with 0.01Ω transition resistance at f2 in line B1C

[0108]

[0109]

[0110] Table 2 Fault simulation results for line B1C at f2 with 100Ω transition resistance

[0111]

[0112]

[0113] It can be seen from Tables 1 and 2 that when an internal short-circuit fault occurs with an IIDG connected downstream of f2, the fault type and fault location have little effect on the fault degree, and the fault degree is greater than the protection threshold of 0.5, so the protection can operate reliably.

[0114] 2) Test of the impact of CT saturation on protection

[0115] CT saturation of the current transformer will cause the current waveform on the secondary side to be distorted, which may affect the current differential protection. In order to verify the anti-CT saturation capability of the protection method proposed in this patent, when a phase A grounding short circuit fault and a phase A and B short circuit grounding fault occur at f1 and f2 respectively, and the phase A current of the current transformer is CT saturated, the simulation results are shown in Tables 3 and 4.

[0116] Table 3 Simulation results of normal line B1C when different faults occur at 0.01Ω at f1 in line AB1 and one side CT is saturated

[0117]

[0118] Table 4 Simulation results of line B1C when phase A fault occurs at 0.01Ω at f2 in line B1C and one side CT is saturated

[0119]

[0120] According to Table 3 and Table 4, it can be seen that compared with the case where CT saturation does not occur, the fault degree F of phase A of the fault line is a The value of has decreased, but it is still greater than the protection threshold. It can be seen that the protection algorithm is not affected by CT saturation. When CT saturation occurs in the fault zone and the fault zone outside the zone, the protection can reliably operate or not operate.

[0121] 3) Test the impact of data asynchrony on protection

[0122] In order to verify the ability of the protection method to resist the influence of data asynchrony, taking phase A as an example, different types of faults are set at f2, and the phase of the current of phase A at the DG side protection is used as a reference to adjust the phase of the current data of phase A at the power supply side protection, so that the current data of phase A on both sides are asynchronous. When there is a 1ms and 2ms asynchronous error in the data on both sides, the simulation results are shown in Tables 5 and 6.

[0123] Table 5 Simulation results when there is a 1ms error in the A phase current on both sides of the fault under a 0.01Ω transition resistance at f2 in line B1C

[0124]

[0125] Table 6 Simulation results when there is a 2ms error in the A phase current on both sides of the fault under a 0.01Ω transition resistance at f2 in line B1C

[0126]

[0127]

[0128] It can be seen from Tables 5 and 6 that when the fault current data in the fault zone is not synchronized, the fault degree of the line is reduced, but it is still greater than the protection threshold, and the fault line can be accurately identified; when the current data on both sides of the normal line are not synchronized, the fault degree of the normal line is increased, but it is still less than the protection threshold, and the protection will not malfunction.

[0129] 4) Noise resistance test

[0130] Four short-circuit faults are set at different positions of f1 in line AB1. At the same time, 30dB and 20dB white noises are added to the currents on both sides of line AB1 and line B1C respectively. The simulation results are shown in Tables 7 and 8. It can be seen that the noise added on both sides of the line increases from 30dB to 20dB, and the fault degree of the normal line B1C1 increases, but the maximum fault degree is 0.209, which is less than the protection threshold and will not be misjudged as a faulty line; the fault degree of the faulty line AB1 decreases, but is still much larger than the protection threshold, and the protection can act reliably. It can be seen that the protection method has a certain noise resistance ability and can withstand the influence of 20dB noise.

[0131] Table 7 Simulation results with 0.01Ω transition resistance at f1 in line AB1 and 30dB noise added to both ends of lines AB1 and B1C

[0132]

[0133]

[0134] Table 8 Simulation results of 0.01Ω transition resistance of f1 in line AB1 and 20dB noise added to both ends of lines AB1 and B1C

[0135]

[0136]

[0137] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for pilot protection of an active distribution network based on a GFT diagram, characterized in that: It includes the following steps: S1. Collect the currents of phase A, phase B, and phase C on the bus side near the power source of the line. When the sudden change of current in any phase exceeds the preset startup threshold for 3 consecutive sampling values, the protection starts and proceeds to the next step; Similarly, collect the currents of phase A, phase B, and phase C on the bus side near the load of the line; S2. Based on the currents of phase A, phase B, and phase C obtained in S1, construct the circuit diagrams of each phase on both sides of the line. Use each sampling point of each phase as a vertex, the connection between adjacent vertices as an edge, and each sampling value as the input signal of each vertex to form the input signal array of each phase and the adjacency matrix of each phase circuit diagram; S3. For each circuit diagram of each phase on both sides of the line obtained in S2, calculate its Laplacian matrix; then perform a standard orthogonal decomposition on the obtained Laplacian matrix to obtain the corresponding eigenvector matrix; S4. Based on the operations described in S2 - S3, for each circuit diagram, multiply its input signal array by the corresponding eigenvector matrix to obtain the GFT spectral coefficients of each circuit diagram, form the corresponding GFT spectral coefficient array, and then calculate the smoothness of this array; S5. Normalize the smoothness on both sides obtained in S4 to obtain the normalized smoothness array on both sides; calculate the absolute value of the difference between the normalized smoothness arrays of each phase on both sides to obtain the distance array between the normalized smoothness arrays of each phase on both sides; S6. For each element in the distance array between the normalized smoothness arrays of each phase on both sides obtained in S5, calculate the corresponding similarity coefficient, and then construct the similarity coefficient array corresponding to this distance array; calculate the fault degree of each phase of the line using the obtained similarity coefficient array; S7. Compare the fault degrees of each phase of the line obtained in S6 with the preset protection threshold to establish the line fault protection criterion.

2. According to claim 1, a method for active distribution network pilot protection based on a road map GFT is characterized in that: The specific content of S1 is as follows: Collect the current of phase A, phase B, and phase C on the bus m side of line mn close to the power supply When the current mutation of any phase exceeds 0.1 times the rated current for three consecutive sampling values, the protection starts; similarly, the A, B, and C currents of the bus n side of the line mn close to the load are collected.

3. According to claim 2, a method for active distribution network pilot protection based on a road map GFT is characterized in that: The specific content of S2 is as follows: The A-phase current signal in the previous time window of each sampling point on the m side of the fault line mn obtained in S1 is used With each sampling point as a vertex, the connection between adjacent vertices as an edge, and each sampling value as the input signal of each vertex, construct the input signal array S ignal_ma ; Construct the current sampling point of phase A current on the m side of line mn. ma , the G ma Include p vertices; construct the adjacency matrix W ma , the W ma The diagonal elements are 0, for W ma For non-diagonal elements, the weight of the edge between two adjacent vertices is 1, and the weight of the edge between two non-adjacent vertices is 0; Similarly, the B-phase and C-phase diagrams G of the m-side of line mn are constructed respectively. mb , G mc and the input signal array S ignal_mb , S ignal_mc And the n-side A phase, B phase, C phase diagram G na , G nb , G nc and the input signal array S ignal_na , S ignal_nb , S ignal_nc .

4. The active distribution network pilot protection method based on the GFT according to claim 3 is characterized in that: The specific content of S3 is as follows: S3.

1. For each circuit diagram of three phases on both sides, calculate its Laplacian matrix L, and the specific calculation formula is: L = D - W Where D represents the degree matrix; the diagonal elements of D are the degrees of node i It represents the sum of the weights of nodes connected to node i in the road graph, and the off-diagonal elements of D are 0; S3.

2. Perform a standard orthogonal decomposition on matrix L to obtain its eigenvector matrix V, and each column in V is an eigenvector.

5. The active distribution network pilot protection method based on the GFT diagram according to claim 4 is characterized in that: The specific content of S4 is as follows: S4.

1. For the A phase diagram G of the m side of line mn in the cycle time window before each sampling time t after the line fault ma , the input signal array S ignal_ma Multiplying with the corresponding matrix V, we get the phase diagram G of side A at sampling time t ma The GFT spectrum coefficient array f ma , the f ma Contains p elements; the lth spectral coefficient f ma (l) is the sum of the multiplication of each element in the input signal array with the corresponding element in the l-th column eigenvector in V, that is: S4.

2. Calculate the phase A diagram G of the m side of line mn at sampling time t ma The GFT spectrum coefficient array f ma Smoothness S ma (t), that is: S ma (t)=f ma Lf ma T (2) In the formula, the superscript T represents the transpose operation of the array; S4.3, repeat the operations described in S4.1 to S4.2 above to obtain the smoothness S of the B phase and the C phase on the m side at the sampling time t mb (t), S mc (t), smoothness S of phase A, phase B, and phase C on the n side na (t), S nb (t), S nc (t); S4.

4. Using the smoothness of each moment in the cycle before the current moment, obtain the smoothness array S of phase A, phase B, and phase C on the m side of line mn ma , S mb , S mc , n-side A phase, B phase, C phase smoothness array S na , S nb , S nc .

6. The active distribution network pilot protection method based on the GFT diagram according to claim 5 is characterized in that: The specific content of S5 is as follows: S5.1 normalizes each smoothness and converts it to the interval [0,1]; for the smoothness array S of phase A on side m ma The i-th element S in ma (i), the normalized smoothness value is S ma_n (i) Through the above processing, we can get the normalized array S of the smoothness of phase A on the m side: ma_n ; S5.2, repeat the operation described in S5.1, and obtain the normalized smoothness array S of the B phase and the C phase on the m side through normalization processing. mb_n , S mc_n ; The smoothness normalized array S of phase A, phase B, and phase C on the n side na_n , S nb_n , S nc_n ; S5.

3. Calculate the normalized array S of the smoothness of phase A on both sides of line mn ma_n , S na_n The i-th element S ma_n (i) and S na_n The absolute value of the difference between (i) and (i) is taken as the i-th smoothness distance d of phase A. a (i) d a (i)=|S ma_n (i)-S na_n (i)| (4) Through the above operation, the distance array d between the normalized arrays of phase A smoothness on both sides of line mn is obtained. a , referred to as the A phase distance array d of line mn a , the value range of each element is [0,1]; S5.4, repeat the operation described in S5.3 to further obtain the distance array d between the normalized smoothness arrays of phase B and phase C on both sides of line mn b ,d c , referred to as the B-phase and C-phase distance array d of line mn b ,d c .

7. The active distribution network pilot protection method based on the GFT diagram according to claim 6 is characterized in that: The specific content of S6 is as follows: S6.

1. When line mn is not faulty or has an out-of-zone fault, the elements in the distance arrays of each phase of line mn are close to 0; conversely, when line mn has a fault, the values ​​of the elements in the distance arrays of at least one phase are greater than 0. For the kth element d of the distance array of phase A, a (k), calculate the corresponding similarity coefficient S da (k): In the formula, ε represents the segmentation threshold; S6.

2. When the smoothness distance d(k) ≤ ε, it indicates that the smoothness of the same phase on both sides at this moment is similar, and the corresponding similarity coefficient is 1; when ε < d(k) < 5ε, it is judged that the smoothness of the same phase on both sides at this moment is moderately similar, and the corresponding similarity coefficient is ε / d(i); when d(k) ≥ 5ε, it indicates that the smoothness of the same phase on both sides at this moment is not similar, and the corresponding similarity coefficient is 0; Thus, we can get the similarity coefficient array S of the smoothness distance of phase A of the line within the previous cycle at the current moment. da , referred to as the A-phase similarity coefficient array S da ; S.6.3, repeat the operations described in S6.1 to S6.2, and further obtain the similarity coefficient array S of the smoothness distance of phase B and phase C in the previous cycle at the current moment db , S dc , referred to as the B-phase and C-phase similarity coefficient array S db , S dc ; S6.

4. Using the A-phase similarity coefficient array S da , calculate the fault degree F of phase A of line at the current moment a : In the formula, N is the number of elements in the similarity coefficient array; S6.5, repeat S6.4, construct the fault degree F of phase B and phase C of the line b 、F c ; S6.

6. When the line mn is operating normally or a fault occurs outside the zone, the similarity coefficients in the similarity coefficient arrays of each phase on both sides of the line mn are all close to 1, and the corresponding fault degree F is close to 0; On the contrary, when a fault occurs inside the line mn, at least one of the similarity coefficients in the similarity coefficient array of a certain phase is close to 0, and the fault degree F of this phase is close to 1.

8. The active distribution network pilot protection method based on the GFT diagram according to claim 7 is characterized in that: The S7 specifically includes the following contents: The fault degrees F of phase A, phase B and phase C of the line are respectively a 、F b 、F c With protection threshold F set By comparison, the line fault protection criterion is established: (F a >F set )||(F b >F set )||(F c >F set ) (7) In the formula, || represents the logical relationship "or"; If the fault degree of each phase of the line is F a 、F b 、F c If equation (7) is satisfied, it is judged that the line has an in-area fault, otherwise it is judged that the line is operating normally.

Images