Intelligent line selection positioning method and system for single-phase broken line fault of small-current grounding system

Abstract

The invention discloses an intelligent line selection positioning method and system for a single-phase broken line fault of a small-current grounding system, and the method comprises the steps: obtaining the phase voltage, line voltage and zero-sequence current data of adjacent detection points at a load side after the single-phase broken line fault occurs, and eliminating the missing and disturbance data through data processing; inputting the phase voltage and the line voltage into an effective value change judgment network model, and identifying a line selection result of a single-phase broken line fault; and based on a line selection result, positioning a fault and determining a fault phase by using a positioning algorithm, and analyzing a single-phase earth fault condition after line breakage in combination with voltage change characteristics. According to the invention, single-phase disconnection and subsequent grounding states can be accurately identified, misjudgment is reduced, the fault positioning accuracy and reliability of the system are improved, and it is ensured that fault processing is more efficient.

Description

Technical Field

[0001] The present invention relates to the technical field of relay protection for distribution networks, and specifically to an intelligent line selection and fault location method and system for single-phase disconnection faults in a small current grounding system. Background Art

[0002] A small current grounding system refers to a three-phase system with an ungrounded neutral point, a neutral point grounded through an arc suppression coil, or a high impedance grounding, also known as a neutral point indirectly grounded system. A single-phase disconnection fault is a serious fault caused by adverse weather such as lightning strikes, strong winds, and snowstorms, external force damage, or line aging. In order to prevent serious safety accidents caused by disconnection faults, it is necessary to select the faulty line and determine the fault location to facilitate timely troubleshooting.

[0003] The most commonly used method for troubleshooting is manual line patrol. When the topology of the power grid is relatively complex, this method has a large workload and low efficiency. Therefore, a fault feature analysis algorithm is needed to achieve accurate line selection and fault location for disconnection faults. Usually, a complete power grid system is equipped with several devices to collect the phase voltage, phase current, zero-sequence voltage, and zero-sequence current at multiple detection points on the line. The algorithm needs to analyze whether the detection point is upstream or downstream of the fault point based on the characteristics collected at each detection point, and finally combine the location of the detection point to determine the faulty line and section where the disconnection fault occurs.

[0004] In the prior art, the principle of the method based on the effective value criterion of the faulty phase phase voltage is that the faulty phase phase voltage upstream of the break point is the same as that in normal operation, and the effective value of the faulty phase phase voltage downstream approaches zero, and the fault is located by detecting the phase voltage data. The principle of the method based on the line voltage criterion is that in a small current grounding system, the three-phase line voltage remains stable during a single-phase grounding fault. When a single-phase disconnection fault occurs, the effective value and phase angle of the upstream three-phase line voltage basically remain unchanged, and the effective value of the line voltage related to the downstream disconnection decreases, so as to judge the fault. The principle of the method based on the zero-sequence current criterion is that when a single-phase disconnection is not grounded, the load side is grounded, or the power supply side is grounded, the change of the zero-sequence current is regular, and the zero-sequence current can be used to judge the fault and subsequent grounding conditions.

[0005] However, all these three methods only use a single feature to identify single-phase disconnection faults, which is prone to misjudgment. There are also disadvantages such as failure to distinguish between single-phase disconnection and single-phase grounding faults, lack of an abnormal data identification mechanism, insufficient research on the single-phase grounding state after disconnection, failure to consider single-phase grounding automatic reclosing, and poor on-site adaptability. Summary of the Invention

[0006] To address the deficiencies in the existing technology, the present invention provides an intelligent line selection and fault location method and system for single-phase disconnection faults in a small current grounding system. By comprehensively utilizing phase voltage and line voltage, the steady-state effective value changes of the phase voltage and line voltage are used to identify the characteristics of single-phase disconnection and single-phase grounding after disconnection, and accurate fault location of single-phase disconnection is carried out.

[0007] The present invention adopts the following technical solutions.

[0008] The present invention proposes an intelligent line selection and fault location method for single-phase disconnection faults in a small current grounding system, including:

[0009] After a single-phase disconnection fault occurs in the small current grounding system, obtain the detection points adjacent to the load side in all branches; read the phase voltages of the detection points in the order from near to far from the load, generate the line voltages of the detection points, and simultaneously obtain the zero-sequence current;

[0010] Exclude abnormal data, input the line voltage into the effective value change judgment network model, and obtain the effective value change result;

[0011] Combine the effective value change result and the phase voltage to identify the single-phase disconnection line selection result and the fault phase; based on the single-phase disconnection line selection result, use the positioning algorithm to determine the single-phase disconnection fault location result;

[0012] Extract the fault phase phase voltages of the detection points adjacent to the power supply and the detection points adjacent to the load side of the line where the disconnection occurs, and identify the situation of single-phase grounding after single-phase disconnection.

[0013] Further, the abnormal data includes missing data and disturbance data, and among them, the disturbance data is identified in the following manner:

[0014] Sort the harmonic ratios from large to small, compare the sum of the top 5 with the harmonic ratio threshold, and exclude the cycle when the sum of the harmonic ratios is greater than the harmonic ratio threshold;

[0015] Introduce the average value of the absolute values of the height differences of all adjacent nodes to standardize the difference to obtain a value. If the absolute value of the value is greater than 0.4, then exclude the cycle.

[0016] Further, the network structure of the effective value change judgment network model is:

[0017] Input layer, the input format is [n, 256, 2]; where n represents the number of input samples; 256 represents 256 channels, and each channel represents a time node; 2 represents two cycles to be compared;

[0018] Convolution layer, one-dimensional convolution Conv1d, batch normalization BN1d, activation function ReLu;

[0019] Activation function layer, linear transformation Linear, activation function ReLu, followed by linear transformation Linear, activation function softmax;

[0020] Output layer, outputting [a, b, c], representing the probabilities of remaining basically unchanged, increasing, and decreasing respectively, and selecting the one with the highest probability among the three as the result of the effective value change.

[0021] Further, if the results of the effective value changes corresponding to the three line voltages are two decreases and one unchanged, there is a single-phase disconnection fault, and the faulty phase is selected.

[0022] Further, identify the single-phase disconnection line selection result and the faulty phase in combination with the result of the effective value change; based on the single-phase disconnection line selection result; the specific steps are as follows:

[0023] After determining the single-phase disconnection using the result of the effective value change, take out the outermost detection points (F1,..., F n ) of all the lines in the topological structure to detect whether they are downstream;

[0024] The method for judging whether it is downstream is: whether the detection point has the characteristic that the line voltage and phase voltage effective values of the faulty phase decrease simultaneously; if not, it means there is no disconnection fault; if so, the disconnection must be on this branch, and the line selection is completed;

[0025] Determine the faulty phase according to the line selection result.

[0026] Further, use the positioning algorithm to determine the single-phase disconnection positioning result on the identified branch; the specific steps are as follows:

[0027] The detection point adjacent to the power supply side (P point) is the upstream, and the detection point adjacent to the load side (F i point) is the downstream, where i = 1,..., n, and n represents the number of monitoring points;

[0028] Gradually check whether the detection point in the middle between the upstream and downstream detection points belongs to the upstream or downstream; if it is upstream, it means the disconnection is between the new upstream and the original downstream points; if it is downstream, it means the disconnection is between the original upstream and the new downstream;

[0029] And so on, until it is located that the real disconnection position is between two adjacent detection points.

[0030] Further, it is necessary to judge the single-phase grounding conditions that occur after a single-phase break, which are divided into power-side grounding, load-side grounding, and both-sided grounding; it is also necessary to judge whether the single-phase grounding situation after a single-phase break belongs to automatic recovery of single-phase grounding. The steady states after automatic recovery are divided into normal steady state, non-grounding steady state after break, power-side grounding steady state after break, load-side grounding steady state after break, and both-sided grounding steady state after break.

[0031] Further, the method for identifying single-phase grounding after a single-phase break is as follows:

[0032] Extract the phase voltages of the faulty phases at the detection point adjacent to the power source (Point P) and the detection point adjacent to the load side on the line where the break occurs (Point F);

[0033] If the steady-state effective values of the phase voltages of the faulty phases at the two detection points increase and decrease, the side with the decrease is grounded;

[0034] If the steady-state effective value of the phase voltage of the faulty phase at one detection point increases and the other remains basically unchanged, the side that remains basically unchanged is grounded;

[0035] If the steady-state effective value of the phase voltage of the faulty phase at one detection point remains basically unchanged and the other decreases, the side with the decrease is grounded;

[0036] If there is a simultaneous increase or decrease, it means that both sections A / B are grounded.

[0037] Further, the method for judging whether single-phase grounding is automatically restored is as follows:

[0038] When different stages are identified, the corresponding steady states need to be recorded for convenient comparison during subsequent automatic recovery;

[0039] When fluctuations occur in a local time period, combine the steady states of each stage recorded to judge whether the fluctuation belongs to automatic recovery of single-phase grounding or further single-phase grounding;

[0040] Among them, the method for identifying whether it belongs to automatic recovery of single-phase grounding is to check the line voltages and phase voltages of the faulty phases of the existing steady states at Point P and Point F and all the existing steady states that have been stored, and check whether they are basically the same. If they are all the same, it means automatic recovery of single-phase grounding.

[0041] The present invention also proposes a smart line selection and fault location system for single-phase break faults in a small current grounding system, including an electrical quantity acquisition module, an effective value analysis module, a fault location module, and a grounding analysis module, characterized in that:

[0042] The electrical quantity acquisition module, after a single-phase break fault occurs in the small current grounding system, obtains the detection points adjacent to the load side in all branches; in the order of the detection points, reads the phase voltages of the detection points, generates the line voltages of the detection points, and simultaneously obtains the zero-sequence current;

[0043] The effective value analysis module excludes abnormal data, inputs the line voltage into the effective value change judgment network model, and obtains the effective value change result;

[0044] The fault location module combines the effective value change result and the phase voltage to identify the single-phase disconnection line selection result and the fault phase; based on the single-phase disconnection line selection result, the single-phase disconnection location result is determined by using the location algorithm;

[0045] The grounding analysis module extracts the phase voltages of the fault phases at the detection points adjacent to the power supply and at the detection points adjacent to the load side of the line where the disconnection occurs, and identifies the single-phase grounding situation after single-phase disconnection.

[0046] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0047] 1. The present invention can fully combine the comprehensive characteristics of phase voltage and line voltage during the disconnection identification process. Through the comprehensive analysis of the changes in the effective values of phase voltage and line voltage at the upstream and downstream detection points, it effectively avoids misjudgment caused by the interference of a single characteristic, and significantly improves the accuracy of disconnection identification.

[0048] 2. The present invention can distinguish fault types to prevent misjudgment: by using the different change characteristics of line voltage in single-phase disconnection faults and single-phase grounding faults, these two types of faults can be clearly and accurately distinguished. In a single-phase disconnection fault, the effective value of the line voltage related to the fault phase at the downstream detection point decreases; while in a single-phase grounding fault, the line voltage at all detection points remains unchanged, thus preventing the confusion of the two types of faults and reducing the occurrence of misjudgment.

[0049] 3. The present invention can process abnormal data by combining a filtering mechanism and a convolutional neural network. The filtering mechanism can effectively identify and exclude data missing and perturbation states, and the convolutional neural network can accurately compare two segments of steady-state data, ensuring the reliability of the input data and providing a solid foundation for subsequent fault analysis.

[0050] 4. The present invention can distinguish characteristics according to different subsequent grounding states of the disconnection, deeply excavate the change characteristics of the effective values of the phase voltages of the fault phases at the upstream and downstream detection points, and achieve accurate distinction of different grounding states, providing a more targeted basis for fault handling.

[0051] 5. The present invention can automatically identify the situation of single-phase grounding automatic restoration. By recording the steady-state data at different stages and comparing it with the current steady state, it can accurately judge whether the fault is automatically restored or further grounded, avoiding misjudgment caused by local voltage fluctuations and improving the reliability of fault judgment. Description of the Drawings

[0052] Figure 1It is the overall flowchart of the present invention;

[0053] Figure 2 It is the method flowchart in the present invention;

[0054] Figure 3 It is the schematic diagram of data - missing samples in the present invention;

[0055] Figure 4 It is the waveform schematic diagram corresponding to different harmonic ratios in the present invention;

[0056] Figure 5 It is the comparison schematic diagram of rising - falling cycle and normal cycle in the present invention;

[0057] Figure 6 It is the network structure diagram of the effective - value change judgment network model in the present invention;

[0058] Figure 7 It is the flowchart of the positioning algorithm in the present invention;

[0059] Figure 8 It is the P / F point position diagram in the present invention. Detailed implementation manners

[0060] To make the objectives, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. The embodiments described in this application are only part of the embodiments of the present invention, rather than all embodiments. Based on the spirit of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the protection scope of the present invention.

[0061] The present invention proposes an intelligent line - selection and positioning method for single - phase disconnection faults in a small - current grounding system. Referring to Figure 1 , the overall process is as follows:

[0062] 1. Initialize the model and parameters.

[0063] 2. Read the topological structure and find the detection points closest to the load side of all branches.

[0064] 3. Read the phase voltages U a , U b , U c of a certain detection point to generate line U ab , U bc , U ac , and obtain the zero - sequence current I0.

[0065] 4. Detect U ab , U bc , U acThe first cycle and whether it belongs to the steady state: If not, intercept one more cycle backward and repeat this step until all are in the steady state, then continue to the next step; if it belongs to the steady state, then directly go to the next step.

[0066] 5. Detect U ab ,U bc ,U ac Whether the effective values between the three steady-state cycles of U

[0067] 6. Starting from the subsequent normal steady state, traverse with a half-cycle as the step size.

[0068] 7. Intercept one cycle of phase voltage and line voltage.

[0069] 8. Identify whether it is in the steady state. If not, repeat steps 7 - 8 until the traversal is complete. Otherwise, determine whether the effective values of the front and rear segments of the line voltage of all three phases are still basically equal: If equal, continue steps 7 - 8 until the traversal is complete; if not equal, go to step 9.

[0070] 9. Detect whether there is a decreasing trend in the phase voltage of the faulty phase on the load side: If it also has a decreasing trend, it is judged that a break has occurred. Record the time period when the break occurred, and at the same time save the steady state of the phase voltage of the subsequent faulty phase as the steady state when not grounded after the break, then continue to the next step; otherwise, repeat steps 7 - 9 until the traversal is complete.

[0071] 10. Continue to traverse the subsequent cycles with a half-cycle as the step size.

[0072] 11. Intercept one cycle of phase voltage.

[0073] 12. Identify the steady state of the intercepted cycle. If it is not in the steady state, continue steps 11 - 12; otherwise, continue to the next step.

[0074] 13. Check whether there is an increasing or decreasing trend in the phase voltage of the faulty phase. If there is, compare it with the saved normal steady state and the steady state when not grounded after the break, and check whether it belongs to fault recovery. If not, continue to the next step; otherwise, continue steps 11 - 13.

[0075] 14. Determine whether it belongs to section A grounding / section B grounding, continue steps 11 - 14 until the traversal is complete, and then continue steps 3 - 14.

[0076] 15. According to the results of the detection points closest to the load side of all branches, determine which line has a break, and use the dichotomy method to locate the branch with a break according to the previously saved break time period.

[0077] 16. Output result

[0078] Each step of the present invention will be described in detail below. As Figure 2 shown, the description content is as follows.

[0079] After a single-phase disconnection fault occurs in a small current grounding system, obtain the detection points adjacent to the load side in all branches; read the phase voltages of the detection points in the order of the detection points, generate the line voltages of the detection points, and simultaneously obtain the zero-sequence current.

[0080] Identify the three initial cycles corresponding to the three line voltages respectively to determine whether they belong to the steady state; use the initial steady state identification neural network, and its structure is as follows

[0081] Input layer, the input format is [n, 2, 256]; where n represents n waveforms; 2 represents 2 channels [original waveform, waveform FFT result]; 256 represents the default single-cycle length

[0082] Convolution layer, one-dimensional convolution Conv1d, batch normalization BN1d, activation function ReLu

[0083] Activation function layer, linear transformation Linear, activation function ReLu, followed by linear transformation Linear, activation function softmax

[0084] Output layer, output [a, b], where the output a represents the probability of being in the steady state, and b represents the probability of not being in the steady state

[0085] If one of the initial cycles is non-steady, continue to select the subsequent cycles to determine whether they are steady. When it is determined that all three cycles are steady, calculate whether the effective values corresponding to the three cycles are equal. If they are equal, proceed to the next step as a normal steady state; if they are not equal, it indicates that a disconnection fault has occurred, and it is necessary to reselect the detection points and repeat the above steps

[0086] Furthermore, perform steady state identification on the line voltage and phase voltage of a cycle corresponding to the normal steady state. If it is identified that it is not in the steady state, select the subsequent cycles and perform steady state identification again. If it is identified as being in the steady state, process the data of the phase voltage, the line voltage, and the zero-sequence current, excluding the missing and disturbed data; input the line voltages of the previous and subsequent two segments into the effective value change judgment network model to obtain the effective value change result

[0087] Specifically, process the data of the phase voltage and the line voltage, excluding the missing and disturbed data

[0088] The main manifestation of data missing is that there is an abnormal state of sudden large increase / decrease in a stable waveform, replacing the original normal stable state. For exampleFigure 3 As shown. Its recognition is relatively simple. It can detect whether the difference between two adjacent points in each zero-sequence current cycle exceeds a specified threshold (default value in the algorithm is 1000). If it exceeds, it is determined that there is data loss;

[0089] The method for recognizing the disturbance state includes whether the fundamental wave characteristics are obvious, whether there is an upward or downward trend, and whether it has the characteristics of a sine wave. Whether the fundamental wave characteristics are obvious is directly related to the harmonic ratio. For example, Figure 4 as shown, when the harmonic ratio is relatively large, the characteristics of the fundamental wave will be weakened by the harmonics. Therefore, this algorithm sorts the harmonic ratios from large to small, compares the sum of the top 5 with the harmonic ratio threshold, and filters the cycle when the sum of the harmonic ratios is greater than the harmonic ratio threshold; among them, the calculation formula for the harmonic ratio is:

[0090]

[0091] where γ represents the harmonic ratio; X m represents the effective value of the m-th harmonic; base represents the effective value of the fundamental wave; M represents the order of the harmonic.

[0092] Whether there is an upward or downward trend is related to the first node and the last node of the cycle, and the average value of the height differences between all adjacent nodes. For a sine cycle, the heights of the starting point and the ending point should be the same. If the offset between the two is too large, it will cause the overall waveform to have the characteristics of rising and falling, as Figure 5 shown.

[0093] However, it is still inappropriate to rely solely on the magnitude of the difference between the last point and the first point as the criterion for identifying whether the waveform has the characteristics of rising and falling, because the thresholds of different waveforms are different. For example, the waveform threshold of the zero-sequence voltage is much larger than that of the zero-sequence current. The average value of the absolute values of the height differences between all adjacent nodes is introduced to standardize the difference; the specific formula is as follows:

[0094]

[0095] where Y represents the rising and falling trend; Y d represents the height value of the d-th point in the cycle; D represents the total number of points.

[0096] By processing the phase voltage and line voltage data to exclude missing and disturbed data, the accuracy and reliability of fault diagnosis can be improved. The specific judgment method can effectively identify abnormal data, avoid the influence of data noise or unstable factors on the analysis results, thereby ensuring the accuracy of the line selection and positioning algorithm, making the entire fault detection system more accurate and stable.

[0097] Further, input the line voltage into the effective value change judgment network model to obtain the effective value change result; wherein, before the line voltage is input into the effective value change judgment network model, preprocessing is performed; the preprocessing process includes resampling and normalization. Resampling is to ensure that the number of nodes in a single cycle is a set value (default 256), and normalization is used to unify the thresholds of voltage and current to [0,1]. Fast Fourier Transform (FFT) is used to extract features of different frequencies. Resampling is carried out in two steps: 1. Unify the frequency. By means of linear interpolation, N points are supplemented between every two adjacent points for upsampling / take one point every N points for downsampling (N is the value obtained by rounding the quotient of the expected frequency divided by the actual frequency). 2. After unifying the frequency, there may still be a situation where the number of sampling points is inconsistent, especially when the actual frequency and the default frequency (12800Hz) are not in an integer multiple relationship. If it exceeds the set value, directly delete the nodes from the tail; if it is less than the set value, add them at the end. The adding method is to fit the a0·sin(a1·x + a2)+a3 function with the entire cycle data. Here, the fitting is through the nonlinear least squares method. After obtaining the optimal parameters, predict the subsequent points.

[0098] Nonlinear least squares method: There is a series of known point coordinates (x i , y i ), where i represents the i-th point. At the same time, there is also a functional relationship between x i and y i (y i = h(x i )). The nonlinear least squares method is to fit the parameters a of the most suitable function h(x). For example, in this algorithm, it is used to fit the steady state, and the steady state can be represented by the sine function h(x) = a0·sin(a1·x + a2)+a3. Use the existing data to fit the most suitable a = a0, a1, a2, a3. After having h(x), use the subsequent x to substitute into the function to calculate the predicted y. The calculation formula for fitting the optimal parameters is as follows:

[0099]

[0100] f(x i ) = y i - h(x i );

[0101] Among them, f(x i ) represents the difference between the simulated value and the true value of the i-th point.

[0102] By preprocessing the phase voltage and line voltage, including resampling, normalization, and discrete Fourier transform (FFT), the frequency characteristics in the voltage and current signals can be effectively extracted, enabling the subsequent effective value change judgment network model to more accurately judge the state change of the battery. The preprocessing process improves the data quality and the sensitivity of the model to the grid state, enhancing the effect of fault diagnosis.

[0103] Further, the network structure of the effective value change judgment network model is as Figure 6 shown:

[0104] Input layer, the input format is [n, 256, 2]; where n represents the number of input samples; 256 represents 256 channels, and each channel represents a time node; 2 represents two cycles to be compared.

[0105] Convolution layer, one-dimensional convolution Conv1d, batch normalization BN1d, activation function ReLu;

[0106] Activation function layer, linear transformation Linear, activation function ReLu, followed by linear transformation Linear, activation function softmax;

[0107] Output layer, the output is [a, b, c], representing the probability values of basically unchanged, increasing, and decreasing respectively. The one with the largest probability among the three is selected as the effective value change result.

[0108] Through multiple layers of convolution and activation functions, the model can process complex signal waveforms more precisely, thereby improving the accuracy of fault location and line selection. The adaptive ability of the model to complex data enables it to adapt to different types of grid faults.

[0109] Further, the single-phase open circuit line selection result and the fault phase are identified in combination with the effective value change result; based on the single-phase open circuit line selection result, the single-phase open circuit location result is determined using a location algorithm.

[0110] Specifically, if the three effective value change results corresponding to the three line voltages are two decreases and one unchanged, there is a single-phase open circuit fault, and the fault phase is selected in combination with the phase voltage.

[0111] Further, in combination with the single-phase open circuit line selection result, the single-phase open circuit location result is determined using a location algorithm, as Figure 7 shown; the specific steps are as follows:

[0112] After determining the single-phase open circuit using the effective value change result, the most end detection points (F1,..., F n ) of all the lines in the topological structure are taken out to detect whether they are downstream;

[0113] The method for judging whether it is downstream is as follows: whether the detection point has the characteristic that the effective value of a certain phase voltage and the two line voltages related to it decrease simultaneously; if not, it indicates that there is no disconnection fault; if so, the disconnection must be on this branch, and the line selection is completed.

[0114] This step reduces the links that require manual intervention and improves the automation level of fault handling. By accurately judging each branch, the location and isolation of power grid faults can be completed in a short time, greatly improving the speed and reliability of power grid fault recovery.

[0115] Further, on the branch where the disconnection is identified, use the positioning algorithm to determine the single-phase disconnection positioning result; the specific steps are as follows:

[0116] The detection point (P point) adjacent to the power supply side is the upstream, and the detection point (F i point) adjacent to the load side is the downstream, where i = 1,..., n, and n represents the number of monitoring points;

[0117] Gradually check whether the detection point in the middle between the upstream and downstream detection points belongs to the upstream or downstream; if it is upstream, it means that the disconnection is between the new upstream and the original downstream points; if it is downstream, it means that the disconnection is between the original upstream and the new downstream;

[0118] And so on, until it is located that the real disconnection position is between two adjacent detection points.

[0119] By positioning on the branch where the disconnection is identified, the fault point can be located more accurately. Gradually checking between different detection points helps to reduce misjudgment and missed judgment, making the final positioning result more accurate. By gradually analyzing the characteristics of upstream and downstream detection points, the positioning algorithm is optimized, improving the reliability and stability of the system.

[0120] After identifying a single-phase disconnection, it is also necessary to identify whether single-phase grounding occurs subsequently. The single-phase grounding that occurs after a single-phase disconnection is divided into power supply side grounding, load side grounding, and bilateral grounding. Bilateral grounding is actually the superposition state of the former two. The identification of single-phase grounding after a single-phase disconnection only identifies two types: A-section grounding and B-section grounding, where A-section grounding refers to grounding near the power supply side of the break point, and B-section grounding refers to grounding near the load side of the break point. When only A-section grounding occurs, it is power supply side grounding; when only B-section grounding occurs, it is load side grounding; when both A and B-section grounding occur, it is bilateral grounding.

[0121] A-section grounding: When A-section grounding occurs, the effective value of the fault phase voltage of the upstream detection point will decrease (the degree of decrease is related to system parameters), and the effective value of the fault phase voltage of the downstream detection point will increase (the degree of increase is related to system parameters). This characteristic can be used to identify A-section grounding.

[0122] Section B grounding: When a grounding fault occurs in Section B, the effective value of the phase voltage of the faulty phase at the upstream detection point will increase (the degree of increase is related to system parameters), and the phase voltage of the faulty phase at the downstream detection point will decrease (the degree of decrease is related to system parameters). This characteristic can be used to identify the grounding fault in Section B.

[0123] After identifying a single-phase open circuit, the faulty phase can be determined based on the line voltage (assuming that only the line voltage remains basically unchanged, then a fault occurs in Phase A). Then, the phase voltage of the faulty phase at the detection point closest to the power source (hereinafter referred to as Point P) and the detection point closest to the load side of the line where the open circuit occurs (hereinafter referred to as Point F) are extracted, as Figure 8 shown.

[0124] Specifically, the method for identifying single-phase grounding after a single-phase open circuit is as follows:

[0125] Extract the phase voltage of the faulty phase at the detection point adjacent to the power source (Point P) and the detection point adjacent to the load side of the line where the open circuit occurs (Point F);

[0126] If it is detected that the steady-state effective values of the phase voltages of the faulty phase at the two detection points increase and decrease, then the side with the decrease is grounded (for example, if the phase voltage of the faulty phase at Point P decreases and that at Point F increases, it indicates that a grounding fault has occurred in Section A);

[0127] If the steady-state effective values of the phase voltages of the faulty phase at the two detection points, one increases and the other remains basically unchanged, then the side that remains basically unchanged is grounded (for example, if the phase voltage of the faulty phase at Point P remains basically unchanged and that at Point F increases, it indicates that a grounding fault has occurred in Section A);

[0128] If the steady-state effective values of the phase voltages of the faulty phase at the two detection points, one remains basically unchanged and the other decreases, then the side with the decrease is grounded (for example, if the phase voltage of the faulty phase at Point P decreases and that at Point F remains basically unchanged, it indicates that a grounding fault has occurred in Section A);

[0129] If both increase or both decrease occur, then a simultaneous grounding fault in Section A / B has occurred.

[0130] By identifying different situations of single-phase grounding after a single-phase open circuit, the occurrence of grounding faults on the power source side, load side, and both sides can be accurately judged. This judgment method is based on the steady-state analysis of voltage changes, which helps to reduce errors caused by different electrical characteristics, ensure the accuracy of grounding judgment, and thus optimize the fault troubleshooting and repair process of the system.

[0131] After a single-phase line break occurs, it is generally impossible to return to normal. However, subsequent single-phase grounding may very likely return to normal (generally referred to as automatic restoration of single-phase grounding), such as complex situations like grounding restoration in section B after grounding restoration in section A after the line break. In the face of such situations, the algorithm should not simply identify the increasing or decreasing trend of the effective value of the phase voltage of the faulty phase in a local time period, but should divide the steady state into different stages: 1. Normal steady state; 2. Steady state without grounding after the line break; 3. Steady state with grounding on the power supply side after the line break; 4. Steady state with grounding on the load side after the line break; 5. Steady state with bilateral grounding after the line break. When the algorithm identifies different stages, the corresponding steady states need to be recorded. For example, after identifying that the first three line voltage effective values are basically equal at the beginning, the phase voltage and line voltage at this time need to be recorded as the normal steady state for subsequent comparison during automatic restoration. Similarly, the steady state without grounding after the line break, the steady state of the first grounding after the line break, etc. all need to be recorded.

[0132] When a fluctuation occurs in a local time period, it is judged by combining the steady states of each stage recorded. It is identified whether the fluctuation belongs to the automatic restoration of single-phase grounding or further single-phase grounding. The way to identify whether it belongs to the automatic restoration of single-phase grounding is to check the line voltage and the phase voltage of the faulty phase of the existing steady state at point P and point F and all the existing steady states that have been stored, and check whether they are basically the same. If they are all the same, it indicates the automatic restoration of single-phase grounding.

[0133] The determination of the automatic restoration of single-phase grounding can help detect and repair possible power grid faults. By tracking the steady states of different stages and comprehensively comparing and analyzing the phase voltages of the faulty phase upstream and downstream, the automatic restoration phenomenon can be quickly identified and judged, avoiding misjudgment. This mechanism enhances the system's ability to respond to complex fault situations and improves the stability and safety of power grid operation.

[0134] The present invention also proposes a smart line selection and fault location system for single-phase line break faults in a small current grounding system, including an electrical quantity acquisition module, an effective value analysis module, a fault location module, and a grounding analysis module, characterized in that:

[0135] The electrical quantity acquisition module, after a single-phase line break fault occurs in the small current grounding system, obtains the detection points adjacent to the load side in all branches; reads the phase voltages of the detection points in the order of the detection points, generates the line voltages of the detection points, and simultaneously obtains the zero-sequence current;

[0136] The effective value analysis module excludes abnormal data, inputs the line voltage into the effective value change judgment network model, and obtains the effective value change result;

[0137] The fault location module combines the effective value change result and the phase voltage to identify the single-phase line break line selection result and the faulty phase; based on the single-phase line break line selection result, uses the location algorithm to determine the single-phase line break location result;

[0138] The grounding analysis module extracts the phase voltages of the faulty phases at the detection points adjacent to the power source and at the detection points adjacent to the load side of the line where the broken line is located, and identifies the situation of single-phase grounding after single-phase line break.

[0139] The present disclosure may be a system, a method, and / or a computer program product. The computer program product may include a computer-readable storage medium having thereon computer-readable program instructions for causing a processor to implement various aspects of the present disclosure.

[0140] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: modifications or equivalent substitutions can still be made to the specific embodiments of the present invention, and any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention shall be covered by the protection scope of the claims of the present invention.

Claims

1. An intelligent line selection and positioning method for single-phase open circuit faults in a small current grounding system, characterized in that, Including: After a single-phase open-circuit fault occurs in a small current grounding system, obtain the detection points adjacent to the load side in all branches; Read the phase voltages of the detection points in the order of the distance from the detection points to the load from near to far, generate the line voltages of the detection points, and obtain the zero-sequence current at the same time; Exclude abnormal data, input the line voltage into the effective value change judgment network model, and obtain the effective value change result; Combine the effective value change result and the phase voltage to identify the single-phase open-circuit line selection result and the fault phase; based on the single-phase open-circuit line selection result, use the positioning algorithm to determine the single-phase open-circuit positioning result; Extract the fault phase phase voltages of the detection points adjacent to the power supply and the detection points adjacent to the load side of the line where the open circuit is located, and identify the single-phase grounding situation after the single-phase open circuit.

2. The intelligent line selection and positioning method for single-phase open-circuit faults in a small current grounding system according to claim 1, characterized in that: The abnormal data includes missing data and disturbance data, wherein the disturbance data is identified in the following manner: Sort the harmonic ratios from large to small, compare the sum of the top 5 with the harmonic ratio threshold, and exclude the cycle when the sum of the harmonic ratios is greater than the harmonic ratio threshold; Introduce the absolute value mean of the height differences of all adjacent nodes to standardize the difference to obtain a value. If the absolute value of the value is greater than 0.4, exclude the cycle.

3. The intelligent line selection and fault location method for single-phase disconnection fault in a small current grounding system according to claim 1, characterized in that, The network structure of the effective value change judgment network model is: Input layer, input format: [n, 256, 2]; where n represents the number of input samples; 256 represents 256 channels, and each channel represents a time node; 2 represents two cycles to be compared; Convolution layer, one-dimensional convolution Conv1d, batch normalization BN1d, activation function ReLu; Activation function layer, linear transformation Linear, activation function ReLu, followed by linear transformation Linear, activation function softmax; Output layer, output [a, b, c], representing the probability values of basically unchanged, increasing, and decreasing respectively, and select the one with the largest probability among the three as the effective value change result.

4. The intelligent line selection and positioning method for single-phase open-circuit faults in a small current grounding system according to claim 1, characterized in that: If the three effective value change results corresponding to the three line voltages are two decreases and one unchanged, there is a single-phase open-circuit fault, and the fault phase is selected.

5. The intelligent line selection and fault location method for single-phase open circuit fault in a small current grounding system according to claim 1, characterized in that Combine the effective value change result to identify the single-phase open-circuit line selection result and the fault phase; based on the single-phase open-circuit line selection result; the specific steps are: After determining single-phase disconnection using the change result of the effective value, take out the detection points (F1,..., F n ) at the very ends of all lines in the topological structure to detect whether they are downstream; The judgment method of whether it is downstream is: whether the detection point has the characteristics that the line voltage and the effective value of the phase voltage of the fault phase decrease at the same time; if not, it means that it is not an open-circuit fault; if so, the open circuit must be on this branch, and the line selection is completed; Determine the fault phase according to the line selection result.

6. The intelligent line selection and fault location method for single-phase disconnection fault in a small current grounding system according to claim 5, characterized in that, Use the positioning algorithm to determine the single-phase open-circuit positioning result on the branch where the open circuit is identified; the specific steps are: The detection point (point P) adjacent to the power supply side is the upstream, and the detection point (point F i point) adjacent to the load side is the downstream, where i = 1,..., n, and n represents the number of monitoring points; Gradually check whether the detection points in the middle between the upstream and downstream detection points belong to the upstream or downstream; if it is upstream, it means that the open circuit is located between the new upstream and the original downstream points; if it is downstream, it means that the open circuit is located between the original upstream and the new downstream. And so on, until it is located that the real broken line position is between two adjacent detection points.

7. The intelligent line selection and fault location method for single-phase broken line fault in a small current grounding system according to claim 1, characterized in that: Judge the single-phase grounding situation occurring after the single-phase broken line, which is divided into power supply side grounding, load side grounding and both sides grounding; it is also necessary to judge whether the single-phase grounding situation after the single-phase broken line belongs to automatic restoration of single-phase grounding, and the steady state after automatic restoration is divided into normal steady state, non-grounding steady state after broken line, power supply side grounding steady state after broken line, load side grounding steady state after broken line and both sides grounding steady state after broken line.

8. The intelligent line selection and fault location method for single-phase broken line fault in a small current grounding system according to claim 7, characterized in that: The method for identifying single-phase grounding after single-phase broken line is: Extract the phase voltages of the faulty phases at the detection point (P point) adjacent to the power supply and the detection point (F point) adjacent to the load side on the line where the broken line is located; If the steady-state effective values of the phase voltages of the faulty phases at the two detection points increase and decrease, the side with the decrease is grounded; If the steady-state effective value of the phase voltage of the faulty phase at one detection point increases and the other remains basically unchanged, the side with the basically unchanged value is grounded; If the steady-state effective value of the phase voltage of the faulty phase at one detection point remains basically unchanged and the other decreases, the side with the decrease is grounded; If there is an increase or decrease at the same time, it means that there is simultaneous grounding in section A / B.

9. The intelligent line selection and fault location method for single-phase broken line fault in a small current grounding system according to claim 8, characterized in that: The judgment method for whether single-phase grounding is automatically restored is: When different stages are identified, the corresponding steady states need to be recorded to facilitate comparison during subsequent automatic restoration; When fluctuations occur in a local time period, combine the steady states of each stage recorded to determine whether the fluctuation belongs to automatic restoration of single-phase grounding or further single-phase grounding; Among them, the way to identify whether it belongs to automatic restoration of single-phase grounding is to check the line voltages and phase voltages of the faulty phases of the existing steady states of point P and point F and all the existing steady states, and check whether they are basically the same. If they are all the same, it means automatic restoration of single-phase grounding.

10. A small current grounding system single-phase broken line fault intelligent line selection and fault location system based on the method according to any one of claims 1-9, including an electrical quantity acquisition module, an effective value analysis module, a fault location module and a grounding analysis module, characterized in that: The electrical quantity acquisition module, after a single-phase broken line fault occurs in the small current grounding system, obtains the detection points adjacent to the load side in all branches; reads the phase voltages of the detection points in the order of the detection points, generates the line voltages of the detection points, and obtains the zero-sequence current at the same time; The effective value analysis module eliminates abnormal data, inputs the line voltage into the effective value change judgment network model, and obtains the effective value change result; The fault location module combines the effective value change result and the phase voltage to identify the single-phase broken line line selection result and the faulty phase; based on the single-phase broken line line selection result, uses the location algorithm to determine the single-phase broken line location result; The grounding analysis module extracts the phase voltages of the faulty phases at the detection point adjacent to the power supply and the detection point adjacent to the load side on the line where the broken line is located, and identifies the single-phase grounding situation after the single-phase broken line.

Images