Single-phase grounding fault line selection method for resonant grounding system based on spatial information entropy

Abstract

The present invention provides a single-phase grounding fault line selection method for a resonant grounding system based on spatial information entropy, which relates to the technical field of relay protection in power systems. Among them, the single-phase grounding fault line selection method for a resonant grounding system based on spatial information entropy includes recording waves and extracting the transient low-frequency zero-sequence current sequences of each distribution network line; converting the transient low-frequency zero-sequence current sequences between symbols and numbers on multiple time scales, generating a three-dimensional fault space through data verification and defining spatial information entropy, and completing fault line selection after comparing the entropy value difference. Compared with the prior art, the technical solution of the present invention can achieve correct line selection without increasing additional construction costs and operation control difficulties when a single-phase high-resistance grounding fault occurs, and can also achieve correct line selection in the case of a large amount of distortion or loss of sampling data; defining spatial information entropy unfolds the single-phase high-resistance grounding fault characteristics in a three-dimensional space domain, which not only makes the fault characteristics more three-dimensional and rich, but also makes the description of the symbol sequence difference more accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of relay protection in power systems, and in particular, to a single-phase grounding fault line selection method for a resonant grounding system based on spatial information entropy. Background Art

[0002] In the prior art, the neutral grounding methods of medium and low voltage distribution networks are ungrounded and resonant grounding. With the continuous expansion of the distribution network scale, the system capacitive current also increases accordingly. Therefore, the resonant grounding method is mostly used to reduce the short-circuit current. However, since this grounding method introduces the compensating current of the arc suppression coil, when a single-phase grounding fault occurs in the system, the amplitude of the fault information is small and it is not easy to distinguish the fault line. Moreover, there are many distribution network lines, the operation methods are flexible, and the fault characteristics are complex and changeable, such as high-resistance grounding and intermittent arc grounding faults. Due to the wide application of cable lines, in order to prevent phenomena such as cable trench fires caused by breakdown at the weak insulation point during asymmetric operation after a single-phase grounding fault, the prior art principle requires that the fault line be removed as soon as possible.

[0003] There are mainly two existing fault line selection methods. One is the method of processing various information collected after the fault and constructing a line selection criterion, and the other is the method of constructing a line selection criterion by adding system equipment or using the new information of the original equipment after the fault. Each of the two methods has its own advantages and disadvantages. In the former method, when the system has a single-phase high-resistance grounding, the collected fault characteristics may be too weak, beyond the measurement accuracy range of the current transformer and distorted, affecting the line selection; the latter method has clear and definite additional fault information, but it requires additional construction costs and operation control difficulties, and there is a possibility of affecting the safe operation of the system.

[0004] Therefore, the above technical problems need to be further solved. Summary of the Invention

[0005] The purpose of the embodiments of the present invention is to provide a single-phase grounding fault line selection method for a resonant grounding system based on spatial information entropy, so as to solve the problem that a large amount of data distortion affects the line selection accuracy rate when a single-phase grounding fault occurs in the resonant grounding system without increasing additional construction costs and operation control difficulties.

[0006] To solve the above technical problems, the embodiments of the present invention provide the following technical solutions:

[0007] The first aspect of the present invention provides a single-phase grounding fault line selection method for a resonant grounding system based on spatial information entropy, including:

[0008] Recording waves and extracting the transient low-frequency zero-sequence current sequences of each line in the distribution network;

[0009] Convert the transient low-frequency zero-sequence current sequence between symbols and numbers on multiple time scales, generate a three-dimensional fault space through data verification, define the spatial information entropy, and complete fault line selection by comparing the entropy value differences.

[0010] In some modified implementation manners of the first aspect of the present invention:

[0011] Optionally, in the recording of waveforms and extraction of the transient low-frequency zero-sequence current sequences of each line in the distribution network:

[0012] After a fault occurs, record the zero-sequence current of each line in the distribution network within eight cycles, and extract the transient low-frequency zero-sequence current of each line within the first three cycles to form a current sequence.

[0013] Optionally, in the conversion of the transient low-frequency zero-sequence current sequence between symbols and numbers on multiple time scales:

[0014] Adopt the first symbolic aggregation approximation method to convert the transient low-frequency zero-sequence current sequences of each line into two groups of symbol sequences on different time scales;

[0015] Among them, the symbols are set as a, b, c, d, e, the time scales are set as 20, 25, 30, 40, 50, and the numbers are set as 1, 2, 3, 4, 5.

[0016] Optionally, in the generation of a three-dimensional fault space through data verification and the definition of spatial information entropy:

[0017] Use the longest common subsequence to verify the transient low-frequency zero-sequence current symbol sequence with a time scale of 40 to obtain the longest common subsequence length matrix:

[0018]

[0019] In the formula, L is the total number of lines, and R lm is the longest common subsequence length value between line l and line m;

[0020] When all the element values in the longest common subsequence length matrix are equal, or only one element value in the row is equal to the length value of the transient low-frequency zero-sequence current symbol sequence and the other element values are equal, the data verification is passed.

[0021] Optionally, in the generation of a three-dimensional fault space through data verification and the definition of spatial information entropy:

[0022] If the data verification fails, use the steady-state zero-sequence current waveform from the fifth cycle to the eighth cycle after the fault as the data source for data verification. If this data verification still fails, mark the data as abnormal and proceed with the subsequent line selection process.

[0023] Optionally, in the three-dimensional fault space generated after the data verification:

[0024] When the symbol is a, it corresponds to the number 1; when the symbol is b, it corresponds to the number 2; when the symbol is c, it corresponds to the number 3; when the symbol is d, it corresponds to the number 4; when the symbol is e, it corresponds to the number 5;

[0025] When the time scale is 20, it corresponds to the number 1; when the time scale is 25, it corresponds to the number 2; when the time scale is 30, it corresponds to the number 3; when the time scale is 40, it corresponds to the number 4; when the time scale is 50, it corresponds to the number 5;

[0026] A three-dimensional coordinate is formed by each of the symbols, each of the time scales, and each of the numbers, generating a 5×5×5 fault space.

[0027] Optionally, in the process of generating a three-dimensional fault space and defining the spatial information entropy after the data verification, and completing fault line selection by comparing the entropy value difference:

[0028] Calculate the sum of the spatial information entropy and the entropy value difference at the time scale of each line;

[0029] The spatial information entropy is consistent among the sound lines, and the line with the largest entropy value difference is the fault line;

[0030] Among them, H(l) = ∑H w (l), w = 20, 25, 30, 40, 50;

[0031]

[0032] In the formula, H w (l) is the spatial information entropy at the time scale, H(l) is the sum of the spatial information entropy at the time scale, w is the time scale, h(l) is the entropy value difference, and H(m) is the spatial information entropy of line m.

[0033] Optionally, in the fault line selection:

[0034] The process of line selection includes high-resistance grounding faults and low-resistance grounding faults;

[0035] In the case of a high-resistance grounding fault, the phase of the power frequency zero-sequence current of the fault line leads the phase of the power frequency zero-sequence current of the sound line;

[0036] In the case of a low-resistance grounding fault, the phases of the low-frequency zero-sequence currents between the sound lines after the fault remain consistent and are different from the phase of the low-frequency zero-sequence current of the fault line.

[0037] Optionally, the first symbolic aggregation approximation method is:

[0038] Supplement the differential current on the premise that the waveform of the fault signal is known;

[0039] Assume that the original current sequence is y i , then the differential sequence y' i is expressed as:

[0040] In the formula, i is the length of the discrete sequence y.

[0041] Optionally, the spatial information entropy under the time scale is extended from the information entropy, specifically:

[0042] The digital sequences of the current and its differential current, is the spatial point to the Euclidean distance of the origin (0, 0, 0), is the probability of occurrence of this point.

[0043] Compared with the prior art, the single-phase grounding fault line selection method for the resonant grounding system provided in the first aspect of the present invention can achieve correct line selection without increasing additional construction costs and operation control difficulties when a single-phase high-resistance grounding fault occurs and in the case of a large amount of distortion or loss of sampling data.

[0044] Defining the spatial information entropy expands the single-phase high-resistance grounding fault characteristics in the three-dimensional space domain, which not only makes the fault characteristics more three-dimensional and rich, but also makes the description of the difference of the symbol sequence more accurate. Brief Description of the Drawings

[0045] By referring to the drawings and reading the following detailed description, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become easy to understand. In the drawings, several embodiments of the present invention are shown in an exemplary rather than restrictive manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0046] Figure 1 Schematically shows the simulation model diagram of the resonant grounding system used in the present invention;

[0047] Figure 2 Schematically shows the amplitude-frequency characteristic diagram of the FIR filter used in the present invention;

[0048] Figure 3 Schematically shows the schematic diagram of the transient low-frequency zero-sequence current symbol sequence of the faulty line L3 of the present invention;

[0049] Figure 4 Schematically shows the schematic diagram of the transient low-frequency zero-sequence current symbol sequence of the sound line L1 of the present invention;

[0050] Figure 5Schematically shows the schematic diagram of the symbol sequence of the transient low-frequency zero-sequence current of the healthy line L2 of the present invention;

[0051] Figure 6 Schematically shows the schematic diagram of the symbol sequence of the transient low-frequency zero-sequence current of the healthy line L4 of the present invention;

[0052] Figure 7 Schematically shows the schematic diagram of the symbol sequence of the transient low-frequency zero-sequence current of the healthy line L5 of the present invention. Detailed implementation manners

[0053] Hereinafter, the exemplary embodiments of the present disclosure will be described in more detail with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present disclosure can be more thoroughly understood and the scope of the present disclosure can be fully conveyed to those skilled in the art.

[0054] It should be noted that unless otherwise specified, the technical terms or scientific terms used in the present invention should have the ordinary meanings understood by those skilled in the art.

[0055] An embodiment provided by the present invention, a single-phase grounding fault line selection method for a resonant grounding system based on spatial information entropy, includes:

[0056] Recording waveforms and extracting the transient low-frequency zero-sequence current sequences of each line in the distribution network;

[0057] Converting the transient low-frequency zero-sequence current sequences between symbols and numbers on multiple time scales, generating a three-dimensional fault space through data verification and defining spatial information entropy, and completing fault line selection after comparing the entropy value difference.

[0058] When a single-phase high-resistance grounding fault occurs, it can achieve correct line selection without increasing additional construction costs and operation control difficulties in the case of a large amount of distortion or loss of sampling data.

[0059] Defining spatial information entropy unfolds the single-phase high-resistance grounding fault characteristics in the three-dimensional space domain, not only making the fault characteristics more three-dimensional and rich, but also making the description of the symbol sequence difference more accurate.

[0060] In recording waveforms and extracting the transient low-frequency zero-sequence current sequences of each line in the distribution network:

[0061] After a single-phase grounding fault occurs in the resonant grounding system, the zero-sequence current of each line in the distribution network is recorded for eight cycles, and the transient low-frequency zero-sequence current of each line in the first three cycles is extracted to form a current sequence. Among them, the low frequency is below 50 Hz.

[0062] Extract the transient low-frequency zero-sequence current using a low-pass filter such as an FIR filter that does not change the time delay between the input and output, ensuring that the phase difference of the transient low-frequency zero-sequence current between the faulty line and the healthy line does not shift. Among them, the FIR filter is an FIR digital low-pass filter designed by the window function method.

[0063] In the conversion between symbols and numbers of the transient low-frequency zero-sequence current sequence on multiple time scales:

[0064] Use the first symbolic aggregation approximation method to convert the transient low-frequency zero-sequence current sequences of each line into two groups of symbol sequences at different time scales;

[0065] Among them, the symbols are set as a, b, c, d, e, the time scales are set as 20, 25, 30, 40, 50, and the numbers are set as 1, 2, 3, 4, 5.

[0066] In order to be divisible by the number of samples conveniently, the time scales are set as 20, 25, 30, 40, 50, so it is convenient for calculation.

[0067] In generating a three-dimensional fault space and defining spatial information entropy after data verification:

[0068] Use the longest common subsequence to verify the transient low-frequency zero-sequence current symbol sequence at a time scale of 40 to obtain the longest common subsequence length matrix:

[0069]

[0070] In the formula, is the longest common subsequence length matrix, L is the total number of lines, and R lm is the longest common subsequence length value between line l and line m;

[0071] When all the element values in the longest common subsequence length matrix are equal, or the row only contains one element value equal to the length value of the transient low-frequency zero-sequence current symbol sequence and the other element values are equal, then the data verification passes.

[0072] When all the element values in the longest common subsequence length matrix are equal, it is determined as a bus fault.

[0073] When the row in the longest common subsequence length matrix only contains one element value equal to the length value of the transient low-frequency zero-sequence current symbol sequence and the other element values are equal, then enter the subsequent process.

[0074] In generating a three-dimensional fault space and defining spatial information entropy after data verification:

[0075] To reduce the subjectivity of simply introducing margin as a processing method, if the data verification fails, the steady-state zero-sequence current waveforms from the fifth cycle to the eighth cycle after the fault are used as the data source for data verification. If this data verification still fails, after marking the data as abnormal, the subsequent line selection process is carried out.

[0076] After passing the data verification, in the generated three-dimensional fault space:

[0077] When the symbol is a, it corresponds to the number 1; when the symbol is b, it corresponds to the number 2; when the symbol is c, it corresponds to the number 3; when the symbol is d, it corresponds to the number 4; when the symbol is e, it corresponds to the number 5.

[0078] When the time scale is 20, it corresponds to the number 1; when the time scale is 25, it corresponds to the number 2; when the time scale is 30, it corresponds to the number 3; when the time scale is 40, it corresponds to the number 4; when the time scale is 50, it corresponds to the number 5.

[0079] By using each symbol, each time scale, and each number to form three-dimensional coordinates, a 5×5×5 fault space is generated.

[0080] In this fault space, there are 165 coordinate points for each of the 5 time scales of each line, and a total of 825 coordinate points for 5 lines.

[0081] After passing the data verification, a three-dimensional fault space is generated and the spatial information entropy is defined. After comparing the entropy value difference degrees, in the fault line selection:

[0082] Calculate the sum of the spatial information entropies and the entropy value difference degrees for each line at each time scale;

[0083] Since the transient low-frequency zero-sequence current waveforms of healthy lines are consistent, the spatial information entropies of each healthy line are the same, and the line with the largest entropy value difference degree is the fault line;

[0084] Among them, H(l) = ∑H w (l), w = 20, 25, 30, 40, 50;

[0085]

[0086] In the formula, H w (l) is the spatial information entropy at the time scale, H(l) is the sum of the spatial information entropies at the time scale, w is the time scale, h(l) is the entropy value difference degree, and H(m) is the spatial information entropy of line m.

[0087] In the fault line selection:

[0088] Based on the phase difference of the transient low-frequency zero-sequence current between the fault line and the healthy lines, the line selection process includes high-resistance grounding faults and low-resistance grounding faults;

[0089] In case of a high-resistance grounding fault, the phase of the power-frequency zero-sequence current of the faulty line leads that of the healthy line. Specifically:

[0090] The power-frequency zero-sequence current i 0n,b of the faulty feeder and the power-frequency zero-sequence current i 0i,b(i≠n) of the healthy line are such that:

[0091]

[0092]

[0093] Wherein: is the power-frequency component of the zero-sequence current flowing through the transition resistance; c i = C 0i / C 0∑ , R 0i is the equivalent zero-sequence resistance to ground of line i;

[0094] The system compensation degree ν and the system damping ratio d are:

[0095]

[0096] d = 1 / ω0R ∑ C 0∑ > 0;

[0097] Wherein, ω0 is the angular frequency and L is the inductance of the arc suppression coil;

[0098] When c i ν + r R0i d 2 ≥ 0, since the range of the arctangent function is (-π / 2, π / 2), so Vθ n,sin > Vθ i,s . When c i ν + r R0i d 2 < 0, there is

[0099]

[0100] And

[0101] Wherein:

[0102]

[0103] Since

[0104] So there is That is, A > 0;

[0105] Since Therefore, B > 0;

[0106] From A > 0 and B > 0, it can be seen that Vθ n,sin -Vθ i,sin > 0, that is, Vθ n,sin > Vθ i,sin .

[0107] When the system has a single-phase high-resistance grounding fault, there are two states: underdamping and overdamping. Its transient process is composed of a decaying DC component and a power frequency component, and the resonance component of the arc suppression coil inductive current and the power frequency component. Since the transient main resonance component is basically concentrated near the power frequency, the transient low-frequency zero-sequence current mainly consists of the power frequency component. Therefore, when there is a high-resistance grounding, the transient low-frequency (below 50 Hz) zero-sequence current is basically the same as the transient power frequency zero-sequence current, that is, the phase of the low-frequency zero-sequence current of the faulty line is different from the phase of the low-frequency zero-sequence current of the sound line.

[0108] In the case of a low-resistance grounding fault, the resistive current of the arc suppression coil itself causes the phase of the inductive compensation current to shift. Also, since the compensation current only flows through the faulty line, after the fault, the phases of the low-frequency zero-sequence currents between the sound lines remain the same and are different from the phase of the low-frequency zero-sequence current of the faulty line.

[0109] The first symbolic aggregation approximation method in the present invention is as follows:

[0110] Supplement the differential current on the premise that the waveform of the fault signal is known. Among them, the known waveform of the fault signal is similar to the power frequency waveform and there is no dense fluctuation.

[0111] Assume that the original current sequence is y i , then the differential sequence y' i is expressed as:

[0112] In the formula, i is the length of the discrete sequence y, and when i equals 1, it refers to the first element of the discrete sequence y.

[0113] The first symbolic aggregation approximation method can solve the problem in the existing transient low-frequency zero-sequence current symbolic aggregation approximation method that only the mean information is retained, which may cause the loss of the original information and cannot accurately distinguish different signals with the same mean.

[0114] The transient low-frequency zero-sequence current is symbolized on multiple time scales to eliminate the influence of different time scales on the symbolization result. At the same time, the differential current symbolization is introduced as supplementary information to improve the accuracy of the description of the time series by the transient low-frequency zero-sequence current symbolic aggregation approximation method in the existing technology.

[0115] The transient low-frequency zero-sequence current symbolic aggregation approximation method in the existing technology is as follows:

[0116] First, convert the transient low-frequency zero-sequence current sequence I = {i1, i2, ..., i n} collected by the current transformer into a standard sequence I' with a mean of 0 and a standard deviation of 1.

[0117] Secondly, after completing the above process, reduce the dimension of the standard sequence according to the time scale to obtain the sequence

[0118]

[0119] k = 1, 2, …, w;

[0120] where w < n, w is the time scale that needs to be set, n is the length of the zero-sequence current sequence, and k is the length of the dimension-reduced sequence.

[0121] Finally, after setting the number of symbol types, find the symbol interval range according to the Gaussian distribution table, and convert the components inside into a discrete symbol sequence according to the corresponding unique symbol of the interval they belong to

[0122] The spatial information entropy at the time scale is extended from the information entropy, specifically:

[0123] The digital sequences of the current and its differential current, is the Euclidean distance from the spatial point to the origin (0, 0, 0), is the probability of the occurrence of this point.

[0124] An embodiment provided by the present invention is specifically applied as follows: As Figure 1 shown, a simulation model of a distribution network resonant grounding system is established on the PSCAD platform. The model includes cable lines, overhead lines, and hybrid lines, namely L1, L2, L3, L4, and L5, a total of five lines, and a constant power load is connected to the end of the lines. The total length of the cable lines is 32 km, and the total length of the overhead lines is 35 km. Among them, the length of a single line in the distribution network system model generally does not exceed 20 km.

[0125] The neutral point is led out from the bus by the grounding transformer and grounded through the arc suppression coil. Among them, the over-compensation degree of the arc suppression coil is 8%, and the active power loss is 4%.

[0126] Assume that a single-phase grounding fault occurs 5 km away from the bus on line L3, with a transition resistance of 300 Ω and an initial fault angle of 0 degrees. The zero-sequence current at the head of each line within 8 cycles after the fault occurs is recorded, and the transient low-frequency zero-sequence current sequence within 3 cycles after the fault is extracted using the FIR filter, with a total of 600 points. Among them, the sampling frequency is 10 kHz, and the amplitude-frequency characteristic of the FIR filter is as Figure 2 shown.

[0127] After calculating the difference sequence of the transient low-frequency zero-sequence current sequence, the symbolic aggregate approximation method is used to transform the transient low-frequency zero-sequence current sequence and its difference sequence into symbolic sequences at different time scales. The symbolization process of the transient low-frequency zero-sequence current of each line at time scale 30 is shown in Figure 3 , Figure 4 , Figure 5 , Figure 6 and Figure 7 . The symbolization process of the difference sequence is the same as this.

[0128] The longest common subsequence is the sequence with the longest length among all common subsequences between two different sequences. Since the common subsequence does not restrict the position of elements, only affected by the element order, the longest common subsequence may not be unique, but the length remains unchanged.

[0129] Using the longest common subsequence to verify the symbolic sequence of the transient low-frequency zero-sequence current at time scale 40, the longest common subsequence length matrix can be obtained as:

[0130]

[0131] In this longest common subsequence length matrix, only the third row contains only one element value equal to the length value of the symbolic sequence of the transient low-frequency zero-sequence current, and the verification passes.

[0132] When the symbol is a, it corresponds to the number 1, when the symbol is b, it corresponds to the number 2, when the symbol is c, it corresponds to the number 3, when the symbol is d, it corresponds to the number 4, and when the symbol is e, it corresponds to the number 5. When the time scale is 20, it corresponds to the number 1, when the time scale is 25, it corresponds to the number 2, when the time scale is 30, it corresponds to the number 3, when the time scale is 40, it corresponds to the number 4, and when the time scale is 50, it corresponds to the number 5. Through the three-dimensional coordinates formed by each symbol, each time scale, and each number, a 5×5×5 fault space is generated. There are 165 coordinate points for each of the 5 time scales of each line in the fault space, and a total of 825 coordinate points for 5 lines.

[0133] The calculated entropy value differences are 147.08, 147.68, 564.40, 147.43, 145.65. The entropy value difference corresponding to line 3 is the largest, which is 564.40. Therefore, it is determined that line 3 has a fault, and the line selection result is correct.

[0134] Assume that a single-phase ground fault occurs at the bus of the simulation model, the transition resistance is 0.1Ω, and the fault angle is 0 degrees. Using the longest common subsequence to verify the symbolic sequence of the transient low-frequency zero-sequence current at time scale 40, the longest common subsequence length matrix obtained is:

[0135]

[0136] The verification fails. Therefore, the zero-sequence current from the 5th cycle to the 8th cycle after the fault occurrence is used as the data source. The above steps are repeated to obtain the symbol sequence and perform verification. The longest common subsequence length matrix is as follows:

[0137]

[0138] It can be seen that the verification passes, the bus fault is determined, and the line selection is correct.

[0139] Suppose a single-phase grounding fault occurs at 11 km from the bus on line L3, with a transition resistance of 15000 Ω and a fault angle of 90 degrees. Assume that the current transformer is distorted when the zero-sequence current is less than 0.5 A. After discarding the distorted data segment, the effective data length only accounts for 4.5% of the original data length.

[0140] According to the foregoing method, the entropy value differences calculated successively are 151.187, 151.187, 604.749, 151.187, and 151.187. Therefore, it is determined that line 3 is the faulty line and the line selection is correct.

[0141] Therefore, the present invention can solve the problem of fault line selection when data is incomplete and overcome the limitation of the existing line selection technology that requires complete data.

[0142] The fault line selection simulation results under other different fault conditions are as follows:

[0143] Table 1 Fault line selection simulation results for different grounding resistances

[0144]

[0145] Table 2 Fault line selection simulation results for different fault distances

[0146]

[0147]

[0148] Table 3 Fault line selection simulation results for different fault angles

[0149]

[0150] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A single-phase grounding fault line selection method for resonant grounding systems based on spatial information entropy, characterized in that Including: Recording waves and extracting transient low-frequency zero-sequence current sequences of each line in the distribution network; Converting the transient low-frequency zero-sequence current sequences between symbols and numbers at multiple time scales, generating a three-dimensional fault space after data verification and defining spatial information entropy, and completing fault line selection after comparing the entropy value difference; In the conversion of the transient low-frequency zero-sequence current sequences between symbols and numbers at multiple time scales: Using the first symbolic aggregation approximation method to convert the transient low-frequency zero-sequence current sequences of each line into two groups of symbol sequences at different time scales; Among them, the symbols are set as a, b, c, d, e, the time scales are set as 20, 25, 30, 40, 50, and the numbers are set as 1, 2, 3, 4, 5; In the process of generating a three-dimensional fault space after data verification and defining spatial information entropy, and completing fault line selection after comparing the entropy value difference: Calculating the sum of the spatial information entropy and the entropy value difference of each line at the time scale; The spatial information entropy among healthy lines is consistent, and the line with the largest entropy value difference is the fault line; Among them, H(l) = ∑H w (l), where w = 20, 25, 30, 40, 50; where H w (l) is the spatial information entropy under the time scale, H(l) is the sum of the spatial information entropies under the time scale, w is the time scale, h(l) is the entropy value difference degree, and H(m) is the spatial information entropy of line m; The spatial information entropy at the time scale is deduced from the information entropy, specifically: Digital sequences of transient low-frequency zero-sequence current and its differential current, is the Euclidean distance from the spatial point to the origin (0, 0, 0), is the occurrence probability of this spatial point.

2. The single-phase grounding fault line selection method for a resonant grounding system based on spatial information entropy according to claim 1, characterized in that In the recording waves and extracting transient low-frequency zero-sequence current sequences of each line in the distribution network: After a fault occurs, recording the zero-sequence current of each line in the distribution network within eight cycles, and extracting the transient low-frequency zero-sequence current of each line in the first three cycles to form a current sequence.

3. The single-phase grounding fault line selection method for a resonant grounding system based on spatial information entropy according to claim 1, characterized in that In the process of generating a three-dimensional fault space after data verification and defining spatial information entropy: Using the longest common subsequence to verify the transient low-frequency zero-sequence current symbol sequence at the time scale of 40, and obtaining the longest common subsequence length matrix: In the formula, is the longest common subsequence length matrix, L is the total number of lines, and R lm is the longest common subsequence length value between lines l and m; When all the element values in the longest common subsequence length matrix are equal, or only one element value in the row is equal to the length value of the transient low-frequency zero-sequence current symbol sequence and other element values are equal, the data verification is passed.

4. The single-phase grounding fault line selection method for a resonant grounding system based on spatial information entropy according to claim 3, characterized in that, In the process of generating a three-dimensional fault space after data verification and defining spatial information entropy: If the data verification fails, using the steady-state zero-sequence current waveform from the fifth cycle to the eighth cycle after the fault as the data source for data verification. If this data verification still fails, mark the data as abnormal and then proceed with the subsequent line selection process.

5. The single-phase grounding fault line selection method for a resonant grounding system based on spatial information entropy according to claim 1, characterized in that, In the process of generating a three-dimensional fault space after data verification: When the symbol is a, it corresponds to the number 1; when the symbol is b, it corresponds to the number 2; when the symbol is c, it corresponds to the number 3; when the symbol is d, it corresponds to the number 4; when the symbol is e, it corresponds to the number 5; When the time scale is 20, it corresponds to the number 1; when the time scale is 25, it corresponds to the number 2; when the time scale is 30, it corresponds to the number 3; when the time scale is 40, it corresponds to the number 4; when the time scale is 50, it corresponds to the number 5; Generating a 5×5×5 fault space through each symbol, each time scale, and each number to form three-dimensional coordinates; 6. The single-phase grounding fault line selection method for a resonant grounding system based on spatial information entropy according to claim 1, characterized in that, In the fault line selection: The process of line selection includes high-resistance grounding faults and low-resistance grounding faults; In the case of high-resistance grounding faults, the phase of the power-frequency zero-sequence current of the fault line leads the phase of the power-frequency zero-sequence current of the healthy line; During the low-resistance grounding fault, the phases of the low-frequency zero-sequence currents between the healthy lines after the fault remain the same, and are different from the phase of the low-frequency zero-sequence current of the faulty line.

7. The single-phase grounding fault line selection method for a resonant grounding system based on spatial information entropy according to claim 1, characterized in that The first symbolic aggregation approximation method is as follows: Supplement the differential current on the premise that the waveform of the fault signal is known; Assume that the original current sequence is y i , then the difference sequence is expressed as: Wherein, i is the length of the discrete sequence y.

Images