Power distribution network fault positioning method and system

Abstract

The invention discloses a power distribution network fault positioning method and system, relates to the technical field of power system fault positioning, and solves the problem of insufficient accuracy of power distribution network fault positioning in the prior art. According to the embodiment of the invention, the method comprises the steps: obtaining a fault traveling wave signal of a power distribution network, calculating an estimated fault distance based on the fault traveling wave signal, and carrying out the coarse positioning, and obtaining a candidate region where a fault possibly occurs in the power distribution network; and then fault traveling wave features of the extracted candidate areas are input into the decision tree integration model to perform accurate fault positioning, so that the problem of large troubleshooting workload caused by obtaining multiple candidate areas through traditional single-end traveling waves can be solved.

Description

Technical Field

[0001] This invention relates to the field of power system fault location technology, and in particular to a method and system for fault location in distribution networks. Background Technology

[0002] With the upgrading of power systems and the widespread access of distributed new energy sources, the topology of radial distribution networks is becoming increasingly complex, with more branch lines and parallel loads, making fault location more difficult. Meanwhile, the increased requirements for power supply reliability have made rapid and accurate fault location and shortening power outage time core operation and maintenance needs.

[0003] In traditional fault location methods, the method based on apparent impedance calculation is susceptible to interference from high impedance faults and fault initiation angle, resulting in poor location accuracy. Although single-ended traveling wave location has better adaptability, it is difficult to distinguish the traveling wave reflection path under multiple branches, often resulting in multiple candidate results, leading to low troubleshooting efficiency and slow power restoration. The method based on low-voltage area division requires the deployment of a large number of measuring devices and relies on the system impedance matrix, which is costly and has poor adaptability.

[0004] Therefore, a method and system for locating faults in power distribution networks are needed. Summary of the Invention

[0005] To address the problem of insufficient accuracy in fault location in existing power distribution networks, this invention provides a method and system for fault location in power distribution networks, which can improve the accuracy of fault location. The specific technical solution is as follows: In a first aspect, embodiments of this application provide a method for locating faults in a power distribution network, including: Acquire the fault traveling wave signal; based on the fault traveling wave signal, calculate the estimated distance between the fault point and the target meter; wherein the target meter is one of the two-terminal meters that detects the fault traveling wave signal; based on the estimated distance, determine candidate regions from multiple line regions of the distribution network; wherein the multiple regions of the distribution network are divided based on the meter deployment location and branch point location; extract fault traveling wave features based on the fault traveling wave signal of the candidate regions; input the fault traveling wave features into the corresponding decision tree ensemble model to obtain the fault region output by the decision tree ensemble model.

[0006] Preferably, the dual-ended meter includes a power supply side meter and a load side meter; the determination of a candidate region from multiple line areas of the distribution network based on the estimated distance includes: determining that the candidate region includes the line area between the dual-ended meters when the estimated distance is greater than 0 and less than the line length between the dual-ended meters; determining that the candidate region includes the line area upstream of the power supply side meter when the estimated distance is less than 0 and the target meter is a power supply side meter; and determining that the candidate region includes the line area downstream of the load side meter when the estimated distance is greater than the line length between the dual-ended meters and the target meter is a power supply side meter.

[0007] Preferably, the extraction of fault traveling wave features based on the fault traveling wave signal of the candidate region includes: extracting zero-sequence current signal and zero-sequence voltage signal from the fault traveling wave signal of the candidate region; acquiring the electrical signal of the 110kV bus of the distribution network, and calculating the superposition component based on the electrical signal; wherein the superposition component is the difference between the first instantaneous value of the electrical signal after the fault and the second instantaneous value of the electrical signal before the fault; and calculating the fault traveling wave features based on the zero-sequence current signal, the zero-sequence voltage signal, and the superposition component.

[0008] Preferably, the fault traveling wave characteristics include statistical features, informational features, amplitude features, and impedance features; the statistical features include harmonic mean, standard deviation, mean deviation, and kurtosis; the informational features include Shannon entropy and Raney entropy; the amplitude features include root mean square, peak value, and maximum-minimum window difference; and the impedance features include post-fault apparent impedance and superimposed impedance.

[0009] Preferably, the electrical signal includes a pre-fault reference signal for the second complete signal cycle before the fault occurs, and a post-fault analysis signal for the third complete signal cycle after the fault occurs; the first instantaneous value is obtained based on the post-fault analysis signal, and the second instantaneous value is obtained based on the pre-fault reference signal.

[0010] Preferably, before obtaining the fault region output by the decision tree ensemble model, the method further includes: determining the fault type based on the fault traveling wave signal; determining the corresponding target decision tree ensemble model based on the candidate region and the fault type; and inputting the fault traveling wave feature into the corresponding decision tree ensemble model to obtain the fault region output by the decision tree ensemble model, which includes: inputting the fault traveling wave feature into the target decision tree ensemble model to obtain the fault region output by the target decision tree ensemble model.

[0011] Preferably, after obtaining the fault region output by the decision tree ensemble model, the method further includes: calculating a multiple estimation reduction rate based on the total line length of the fault region and the total line length of the candidate region; wherein the multiple estimation reduction rate is used to indicate the reduction of invalid investigation scope.

[0012] Secondly, embodiments of this application provide a power distribution network fault location system, applied to the method described in the first aspect, the system comprising: The acquisition module is used to acquire fault traveling wave signals; The calculation module is used to calculate the estimated distance between the fault point and the target meter based on the fault traveling wave signal; wherein the target meter is one of the two-ended meters that detects the fault traveling wave signal. The determination module is used to determine candidate areas from multiple line areas of the distribution network based on the estimated distance; wherein the multiple areas of the distribution network are divided based on the location of electricity meter deployment and the location of branch points; The extraction module is used to extract fault traveling wave features based on the fault traveling wave signal of the candidate region; The localization module is used to input the fault traveling wave characteristics into the corresponding decision tree ensemble model to obtain the fault area output by the decision tree ensemble model.

[0013] Thirdly, embodiments of this application provide a computing device, including: a memory for storing a program; and a processor for loading the program to execute the method as described in the first aspect.

[0014] Fourthly, embodiments of this application provide a computer-readable storage medium including a stored program, wherein, when the program is executed, it controls the device where the computer-readable storage medium is located to perform the method described in the first aspect.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: by acquiring the fault traveling wave signal of the distribution network and calculating the estimated fault distance based on the fault traveling wave signal for coarse localization, candidate areas in the distribution network that may be faulted are obtained; then, the fault traveling wave features of the candidate areas are extracted and input into the decision tree ensemble model for accurate fault localization, which can solve the problem of large workload caused by obtaining multiple candidate areas from traditional single-end traveling wave. Attached Figure Description

[0016] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale.

[0017] Figure 1 A flowchart illustrating a power distribution network fault location method provided in an embodiment of this application; Figure 2 This application provides a schematic diagram of the division of a power distribution network area. Figure 3 This application provides a positioning error map for different fault distances in an embodiment of the present application. Figure 4 A multi-estimation mitigation result diagram provided in an embodiment of this application; Figure 5 A schematic diagram of a decision tree confusion matrix provided in an embodiment of this application; Figure 6 This is a schematic diagram of the structure of a power distribution network fault location system provided in an embodiment of this application; Figure 7 This is a schematic diagram of the structure of a computing device provided in an embodiment of this application. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0020] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0021] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0022] To address the problem of insufficient accuracy in fault location in traditional methods for distribution networks, this invention provides a method and system for fault location in distribution networks, which can improve the accuracy of fault location in distribution networks.

[0023] Please see Figure 1 , Figure 1 This application provides a flowchart illustrating a method for locating faults in a power distribution network, which is applied to a computing device. Figure 1 As shown, the method includes: Step 101: The computing device acquires the fault traveling wave signal.

[0024] The computing device can be a computing or control module deployed in the power distribution network, a server, or a smart terminal such as a personal computer or tablet directly operated by power system managers or maintenance personnel. This computing device can communicate with sensors deployed on the power distribution network lines via wired or wireless means, acquiring the traveling wave data of the power distribution network lines collected in real time by the sensors, thereby detecting whether a fault has occurred in the power distribution network lines in real time. When a fault is detected in a power distribution line, the computing device can acquire the fault traveling wave signal of that power distribution line.

[0025] Specifically, the fault traveling wave signal can be a traveling wave signal collected by a sensor within a preset time period after the fault occurs. For example, the fault traveling wave signal can be a fault traveling wave signal collected by a current transformer within 10ms after the fault occurs.

[0026] Specifically, computing devices can acquire electrical signals covering key nodes of the distribution network, quickly identify whether a fault has occurred in the distribution network, and accurately determine the fault type, providing prerequisites for subsequent targeted distance estimation and area positioning.

[0027] Based on the radial topology of the distribution network, core physical meters are deployed at the substation outlet on the power source side and at the end of the main line on the load side. Simultaneously, branch physical meters are installed at the end of each branch line (e.g., Figure 2 (M1-M6 in the table) to form a full node coverage of "main trunk + branches". Acquire and record the instantaneous voltage values ​​(unit: kV) and instantaneous current values ​​(unit: A) of phases A, B, and C in each meter.

[0028] After detecting an abnormal traveling wave signal, the computing device can obtain the traveling wave signals recorded by the two meters that first detected the abnormal traveling wave signal, and obtain the fault traveling wave signal.

[0029] Then, the computing device can use a feedforward neural network (FFNN) method based on the S-transform to perform time-frequency analysis on the signal, extract fault features, and input them into the neural network for classification. Alternatively, a machine learning method based on the Hilbert-Huang Transform (HHT) can be used to process non-stationary signals through Empirical Mode Decomposition (EMD) and combined with a machine learning model to complete the detection. These algorithms achieve a fault detection rate of up to 99.9%, and the classification accuracy is also at a high level, effectively distinguishing four common types of faults in the power grid: single-phase ground fault (LG), two-phase short-circuit fault (LL), two-phase ground fault (LLG), and three-phase short-circuit fault (LLL). Fault detection and classification provide a basis for subsequent calculations.

[0030] Step 102: The calculation device calculates the estimated distance between the fault point and the target meter based on the fault traveling wave signal.

[0031] The target meter is one of the two-ended meters used to detect the traveling wave signal of the fault. Specifically, based on the distance between the meter and the power supply and load ends, the computing device can divide the two-ended meter into a power supply side meter and a load side meter.

[0032] Since the power distribution network is a three-phase system, the three-phase voltage and current signals are coupled, and direct analysis would interfere with the extraction of traveling wave characteristics. Therefore, the computing device needs to decouple the three-phase signals into independent modal components through mode transformation. Specifically, the computing device can use the Carlson Transform to achieve mode decoupling, as shown in the following formula: ; in, This is the grounding mode voltage mode vector, which reflects the zero-sequence component of the three-phase signal and is mainly related to grounding faults; , This is the overhead mode voltage mode vector, reflecting the positive and negative sequence components of the three-phase signal. , , These are three-phase voltage signals.

[0033] in, Mode traveling waves are less affected by circuit parameters during propagation and have a stable propagation speed (approaching 90%-95% of the speed of light), therefore they are chosen. The analog signal is the core object of subsequent traveling wave analysis.

[0034] Wavelet transform possesses the dual advantages of "time-domain localization" and "frequency-domain localization," effectively capturing non-periodic traveling wave signals with pulse characteristics. It can be used to achieve accurate time identification. In practice, a suitable wavelet (such as the db4 wavelet) is selected as the base wavelet for... The analog voltage signal undergoes multi-level (e.g., 5-level) multi-resolution decomposition. Through iterative processes of "low-pass filtering + downsampling" and "high-pass filtering + downsampling," signal detail coefficients at different frequency scales are extracted. To eliminate noise interference, the maximum value of the multi-level detail coefficients during steady-state system operation is used as a benchmark, and a threshold is set to 1.1 times this maximum value (i.e., increased by 10%). When the signal detail coefficients first exceed the threshold, time detection is triggered. Subsequently, the signal peak values ​​within multiple sampling points after the threshold trigger are analyzed. The time corresponding to this peak value is the accurate time when the first traveling wave arrives at the meter, denoted as the traveling wave arrival time of the dual-ended meter. For example, here... As the arrival time of the traveling wave of the meter on the power supply side; This represents the arrival time of the traveling wave at the load-side meter.

[0035] Calculated based on the design parameters of the power distribution network lines The propagation speed of a traveling wave is calculated using the following formula: ; Where L is the inductance per unit length of the distribution network line, and C is the capacitance per unit length of the distribution network line.

[0036] Then, the computing device can calculate the time difference of arrival of the traveling wave at both ends of the meter. The calculation formula is: ; The absolute value is used in the formula to accommodate the different deployment locations of the two-terminal meters and to avoid negative time differences.

[0037] Finally, based on the actual line length l between the two-terminal meters, the estimated distance from the fault point to the power supply side meter is calculated using the two-terminal traveling wave localization formula. .

[0038] ; If the calculation result is positive and less than 1, the fault point is located on the main line between the two-terminal meters; if the result is negative or greater than 1, the fault point is located on a branch line, and further location needs to be determined by subsequent area pre-screening.

[0039] Among them, due to Taking the absolute value, therefore, in the calculation... When the value is greater than 1, it needs to be combined with and The order of priority is used to determine the estimated distance. Is it less than 0 or greater than 1?

[0040] exist Prior to In the case of a positive distance, it indicates that the fault point is closer to the power supply side meter. At this time, the fault point is not located on the main line between the two-terminal meters, so the distance between the fault point and the power supply side meter is negative, less than 0. Conversely, the fault point is closer to the load side meter, and the distance between the fault point and the load side meter is greater than 1.

[0041] In addition to the above process, the computing device can also use other improved two-end traveling wave localization methods to estimate the distance, such as wavefront recognition optimization combined with deep learning, and algorithms combined with time-invariant parameters; it can also use the single-end traveling wave localization method to estimate the location of the fault point.

[0042] Step 103: The computing device determines candidate regions from multiple line regions of the distribution network based on the estimated distance.

[0043] Once the estimated distance is obtained, the computing device can narrow down the possible location of the fault based on the estimated distance, reduce the amount of subsequent calculations, and improve the positioning efficiency.

[0044] The multiple regions of the power distribution network are divided based on the location of electricity meters and branch points. The network topology of the power distribution network and the division of these multiple regions can be preset in the computing device or obtained by the computing device from its database.

[0045] Preferably, the computing device can divide the distribution network topology into regions based on preset rules. The specific distribution network region division rules are as follows: Regardless of the number of branches and line lengths in the distribution network, the entire distribution network is divided into several non-overlapping fault analysis regions, with the "meter deployment point" and "branch junction point" as boundaries. Each region corresponds to a continuous line segment (including main line segments and branch line segments), and each region is assigned a unique number (e.g., S1, S2, S3, ..., Sn). After the region division is completed, the region information (including region number, start and end nodes, line length, line parameters, etc.) is entered into the distribution network topology database, forming a "region-line-parameter" relationship. For details, please refer to [link to relevant documentation]. Figure 2 , Figure 2 This is a schematic diagram illustrating the division of a power distribution network area according to an embodiment of this application.

[0046] Preferably, during the pre-screening process, the computing device can first retrieve the line length l between the two-terminal meters and the layout information of surrounding branch lines from the topology database, and combine this with the fault distance estimation results. Make a preliminary judgment: If the estimated distance is greater than 0 and less than the line length between the two-terminal meters, the computing device can determine that the candidate region includes the line area between the two-terminal meters; if the estimated distance is less than 0 and the target meter is a power supply side meter, the candidate region is determined to include the line area upstream of the power supply side meter; if the estimated distance is greater than the line length between the two-terminal meters and the target meter is a power supply side meter, the candidate region is determined to include the line area downstream of the load side meter.

[0047] Correspondingly, if the estimated distance is less than 0 and the target meter is a load-side meter, the candidate area is determined to include the line area downstream of the load-side meter; if the estimated distance is greater than the line length between the two-terminal meters and the target meter is a load-side meter, the candidate area is determined to include the line area upstream of the power-side meter.

[0048] It is understandable that when a fault occurs in a branch line within the line area between two-terminal meters, the fault traveling wave signal needs to propagate from the branch point to the two-terminal meters. In this case, it can be regarded as a fault point on the main line for coarse location, and the corresponding line area can also be located.

[0049] By using the above logic, multiple candidate fault regions can be identified. Then, the computing device can analyze only the signals corresponding to these candidate regions, reducing unnecessary calculations and improving computational efficiency.

[0050] For example, in Figure 2 In the specific example, line 1 is the physical starting point of S1: From a physical topology perspective, the area of ​​S1 starts from line 1 (the end connecting the transformer), extends through line 2 to line 3, therefore line 1 is the "physical starting point" of S1. Line 3, on the other hand, is the functional starting point of S1: From the functional logic of double-ended fault location, line 3 is the branch junction point where M1 supplies power to S3 and S4, and is defined as the "functional power supply starting point" (i.e., the zero-point reference for fault location).

[0051] This separation of "physical starting point" and "functional starting point" is to simultaneously satisfy the authenticity of physical topology and the logical efficiency of fault analysis—the physical starting point ensures the intuitiveness of line connections, while the functional starting point ensures that the numerical range of fault distance is completely aligned with the regional division.

[0052] Taking M1 and M4 as examples of pairs, the core function of the M1-M4 pair is "power transmission from M1 to M4", and the division of functional areas is based on the "power transmission direction": Among them, the functional power supply starting point is line 3, which is the branch point where M1 starts to supply power to S3 / S4; the line area upstream of the function (D<0) is the opposite direction of the power supply starting point, that is, the S1 area, which is the power supply input path of M1; the functional line area between the two-terminal meters (0≤D≤l) is the power supply path from the power supply starting point to M4, that is, the S3 area; the functional line area downstream (D>l) is the power supply extension path of M4, that is, the S4 area.

[0053] More specifically, S1 is the upstream branch of M1's power supply, physically including the installation location of M1, but functionally belonging to the "power input area of ​​the power supply side meter", and logically independent from the "power transmission backbone" of the M1-M4 pair; S3 is the power transmission backbone area of ​​the M1-M4 pair, covering the power transmission path from line 3 to S3; S4 is the downstream branch of M4's power transmission, belonging to the "power transmission extension area of ​​the load side meter", and its logical layering with the backbone of S3 is clearly defined.

[0054] Step 104: The computing device extracts fault traveling wave features based on the fault traveling wave signal of the candidate region.

[0055] After determining the candidate region, the computing device can obtain the fault traveling wave signal of the corresponding region from the measuring meters covering all nodes of the distribution network, and then extract the fault traveling wave features from the fault traveling wave signal.

[0056] In the event of a fault, the computing device can downsample the signal from the original sampling frequency (e.g., 12MHz) to 256 samples / cycle (system frequency is 50Hz). While preserving the transient characteristics of the fault, the amount of data is compressed to about 1 / 8 of the original, significantly reducing the computational burden.

[0057] Preferably, the computing device can extract zero-sequence current signal and zero-sequence voltage signal from the fault traveling wave signal in the candidate region; acquire the electrical signal of the 110kV bus of the distribution network, and calculate the superposition component based on the electrical signal; wherein, the superposition component is the difference between the first instantaneous value of the electrical signal after the fault and the second instantaneous value of the electrical signal before the fault; and calculate the fault traveling wave characteristics based on the zero-sequence current signal, the zero-sequence voltage signal and the superposition component.

[0058] Among them, Figure 2 In a specific example, a virtual meter is set up on the 110kV bus of the distribution network, which is displayed as... Figure 2The red dot on the bus; computing devices can obtain the electrical signals of the 110kV bus through this virtual meter. The virtual meter is located at the core of the 110kV bus and serves as the overall monitoring point on the power supply side. It can collect global electrical status data, providing a benchmark for superimposed component calculations; it can also reflect the overall impact of faults on the system, avoiding information loss from local meters; simultaneously, it can provide a unified time and amplitude benchmark for zero-sequence component and superimposed component calculations, ensuring accurate subsequent feature extraction.

[0059] Understandably, the 110kV bus is not within the regional division scope of this plan because it belongs to a higher voltage level power grid, and... Figure 2 The specific examples show that the 20kV distribution network levels in areas S1-S5 are different.

[0060] This superimposed component includes the current and voltage components of each of the three phases, specifically including... , , , , , .

[0061] Preferably, the electrical signal includes a pre-fault reference signal for the second complete signal cycle before the fault occurs, and a post-fault analysis signal for the third complete signal cycle after the fault occurs; the first instantaneous value is obtained based on the post-fault analysis signal, and the second instantaneous value is obtained based on the pre-fault reference signal.

[0062] After obtaining the zero-sequence component and the superposition component, the computing device can use them as multi-dimensional feature channels for subsequent key feature extraction. The zero-sequence component highlights the electrical characteristics of asymmetric faults such as ground faults, while the superposition component highlights the changes in the transient process of the fault (the difference before and after the fault). Together, they constitute a multi-dimensional feature set for subsequent algorithms to extract key fault information. No additional calculations or merging are required; they can be used directly as independent signal dimensions.

[0063] Preferably, the computing device can extract key features that effectively distinguish faults in different regions from the preprocessed signal, and construct fault traveling wave features as the input vector of the decision tree ensemble model. These fault traveling wave features encompass four dimensions: statistical features, informational features, amplitude features, and impedance features.

[0064] Specifically, statistical features include harmonic mean, which reflects the central tendency of the signal; standard deviation, which reflects the dispersion of the signal; mean deviation, which reflects the average level of the signal's deviation from the mean; and kurtosis, which reflects the steepness of the signal distribution.

[0065] Information-related features include Shannon entropy, used to quantify the information content of a signal; and Raney entropy, used to enhance sensitivity to local features of a signal.

[0066] Amplitude-related features include root mean square (RMS), which reflects the effective value of the signal; peak value, which reflects the maximum fluctuation amplitude of the signal; and maximum and minimum window difference, which reflects the amplitude of the signal change within the sliding window.

[0067] Impedance characteristics include apparent impedance after a fault, calculated by the phasor ratio of each phase voltage to current, used to reflect the change in system impedance after a fault; and superimposed impedance, calculated by the ratio of superimposed voltage to superimposed current, used to highlight the impedance increment caused by the fault.

[0068] To improve the discriminative power of features, computing devices can also split complex-form features into real and imaginary parts to form higher-dimensional feature vectors, ensuring that the decision tree ensemble model can fully capture the subtle differences in fault signals.

[0069] Step 105: The computing device inputs the fault traveling wave characteristics into the corresponding decision tree ensemble model to obtain the fault region output by the decision tree ensemble model.

[0070] The computing device employs an Extremely Randomized TreesClassifier (Extra-Tree-Classifier) ​​to construct an ensemble decision tree model. This model avoids overfitting by randomly selecting features and using random splitting thresholds to build multiple decision trees, and it also boasts high computational speed.

[0071] Specifically, the computing device can be configured with multiple decision tree estimators, such as 100; the maximum tree depth is set to "None," meaning it is automatically determined by the data; and the minimum number of sample splits is set to 2 to adapt to imbalanced datasets. Then, multiple fault cases covering different topologies, fault types, and fault locations are used for model training, generating corresponding feature vectors and region labels (e.g., "fault located in S2" is labeled 1, "fault not located in S2" is labeled 0). A 10-fold cross-validation method is then used for model optimization, randomly dividing the dataset into 10 parts: 9 parts for training and 1 part for validation. This process is repeated 10 times, and the average performance is taken. Simultaneously, 30% of the cases are reserved as a test set to verify the model's generalization ability. After training, independent decision tree ensemble models are constructed for each region and each possible fault type (LG, LL, LLG, LLL), with each model corresponding to one fault region.

[0072] Preferably, before inputting the fault traveling wave feature into the corresponding decision tree ensemble model to obtain the fault region output by the decision tree ensemble model, the computing device can determine the fault type based on the fault traveling wave signal; determine the corresponding target decision tree ensemble model based on the candidate region and the fault type; and then input the fault traveling wave feature into the target decision tree ensemble model to obtain the fault region output by the target decision tree ensemble model.

[0073] When identifying fault areas, the computing device inputs the feature vectors corresponding to the candidate areas into the corresponding decision tree ensemble model. The decision tree ensemble model outputs a binary classification result of "1" (indicating that there is a fault in the area) or "0" (indicating that there is no fault in the area).

[0074] At this point, the reliability of the results needs to be verified using the formula for the accuracy of fault area identification: ; Where TP is the number of samples in which the model correctly identifies fault regions, TN is the number of samples in which the model correctly excludes non-fault regions, FP is the number of samples in which the model misclassifies non-fault regions as fault regions, and FN is the number of samples in which the model misses fault regions.

[0075] when In this case, the computing device can determine that the candidate region is a faulty region, ensuring the accuracy of region identification.

[0076] After determining the fault area, the computing device can further locate the faulty line segment based on the fault traveling wave signal of each line segment within the area, thereby improving the positioning accuracy.

[0077] To comprehensively evaluate method performance, computing devices can quantify the mitigation of multiple estimations and the accuracy of fault distance estimation.

[0078] Preferably, after obtaining the fault region output by the decision tree ensemble model, the computing device can also calculate the multiple estimation reduction rate based on the total line length of the fault region and the total line length of the candidate region; wherein, the multiple estimation reduction rate is used to indicate the reduction of invalid investigation range.

[0079] The formulas for calculating the reduction rate through multiple estimations include: ; Where is the total line length of the finally determined fault area, and is the total line length of all candidate areas.

[0080] Then, the computing device can also verify the ranging accuracy using relative error, which is calculated using the following formula: Where D is the actual fault distance. To estimate the distance, l is the length of the line between the two meters.

[0081] Calculate R and using the above formulas Evaluating the effectiveness of the method and optimizing the feature selection or tree structure of the decision tree based on the results can ensure a continuous improvement in the ability to mitigate the multiple estimation problem and the accuracy of distance estimation.

[0082] In this embodiment, by acquiring the fault traveling wave signal of the distribution network and calculating the estimated fault distance based on the fault traveling wave signal for coarse localization, candidate areas in the distribution network that may be faulted are obtained; then, the fault traveling wave features of the candidate areas are extracted and input into the decision tree ensemble model for precise fault localization, which can solve the problem of large investigation workload caused by obtaining multiple candidate areas from traditional single-end traveling wave.

[0083] The following is combined with Figures 3 to 5 The technical effects of the embodiments of this application will be further explained.

[0084] Please see Figure 3 Regarding the low positioning accuracy of traditional apparent impedance methods, which are susceptible to high impedance and fault initiation angle interference leading to large errors, the method in this application integrates dual-end traveling wave positioning with mode transformation and wavelet analysis. Figure 3 As shown, under different combinations of double-ended meters and different fault distances (10%-90% of the line length), the maximum positioning error of the embodiment of this application does not exceed 2.7%, the fault error at the midpoint of the line is only 0.04%-0.36%, and the average error is as low as 0.79% (standard deviation 0.4%), which can accurately capture the subtle differences in the fault location.

[0085] Please see Figure 4 and Figure 5 To address the problem of multiple estimation in the single-ended traveling wave method, traditional single-ended localization is prone to misclassification of multiple regions in multi-branch networks. The method in this application identifies fault regions using decision trees, such as... Figure 4 As shown, the average accuracy rate for identifying the four types of fault areas exceeded 88.7%, with an accuracy rate of 92.6% for LLL faults, while the proportion of unidentified LG faults and misjudged multiple areas was only 12.4%. Figure 5 As shown in the (LG Fault Confusion Matrix), the TN identification accuracy in areas S1 and S5 reaches 100%, and the TP identification accuracy in areas S2-S4 is 87.3%-92.7%. This not only locks down the unique fault area, but also avoids relying on a large number of measuring devices, achieving a balance between positioning accuracy and mitigation of multiple estimations.

[0086] Regarding the poor adaptability of existing solutions, the traditional low-voltage area division method relies on the system impedance matrix and has weak adaptability to load fluctuations. The method in this application constructs a complete chain of "data acquisition - signal processing - area identification", based on the collaborative acquisition of virtual meters and physical meters, and combines superimposed components to cancel load interference, thus ensuring high reliability of fault location in complex topologies (such as radial networks containing distributed power sources).

[0087] This application implements a three-stage fusion process of "dual-end traveling wave localization—decision tree region identification—multiple estimation mitigation." First, voltage / current signals are synchronously acquired at both ends of the distribution network (physical meters M1-M6 at the feeder end and virtual meters in the substation). The three-phase signals are decomposed using mode transformation, and the traveling wave front time is detected using wavelet transform to construct a traveling wave fault localization model. Then, the signal is downsampled to extract zero-sequence components, superposition components, and various features. These are input into an extreme random tree binary classifier trained for multiple non-overlapping regions (such as S1-S5). The fault region is determined by confusion matrix and accuracy, solving the regional misjudgment caused by "multiple estimation" in traditional single-end traveling wave faults and ensuring the unique fault region is located. The entire solution constructs an innovative technology system for precise fault localization in radial distribution networks from three dimensions: "mode transformation and wavefront detection modeling for dual-end synchronous traveling wave localization," "multi-dimensional signal feature extraction and decision tree region classification mechanism," and "accuracy-based dynamic mitigation method for multiple estimation." This comprehensively overcomes the limitations of traditional methods in terms of localization accuracy, adaptability to complex topologies, and efficiency of multiple estimation mitigation.

[0088] The method provided in the embodiments of this application has been described above. The system provided in the embodiments of this application will be described below.

[0089] Please see Figure 6 , Figure 6 This is a schematic diagram of the structure of a power distribution network fault location system provided in an embodiment of this application, as shown below. Figure 6 As shown, the system 60 includes: Acquisition module 601 is used to acquire fault traveling wave signals; The calculation module 602 is used to calculate the estimated distance between the fault point and the target meter based on the fault traveling wave signal; wherein the target meter is one of the two-ended meters that detects the fault traveling wave signal. The determination module 603 is used to determine candidate areas from multiple line areas of the distribution network based on the estimated distance; wherein the multiple areas of the distribution network are divided based on the location of the electricity meter deployment and the location of the branch point; Extraction module 604 is used to extract fault traveling wave features based on the fault traveling wave signal of the candidate region; The positioning module 605 is used to input the fault traveling wave characteristics into the corresponding decision tree ensemble model to obtain the fault area output by the decision tree ensemble model.

[0090] Preferably, the dual-ended meter includes a power supply side meter and a load side meter; the determining module 603 is specifically used to determine a candidate region from multiple line regions of the distribution network based on the estimated distance, including: when the estimated distance is greater than 0 and less than the line length between the dual-ended meters, determining that the candidate region includes the line region between the dual-ended meters; when the estimated distance is less than 0 and the target meter is a power supply side meter, determining that the candidate region includes the line region upstream of the power supply side meter; when the estimated distance is greater than the line length between the dual-ended meters and the target meter is a power supply side meter, determining that the candidate region includes the line region downstream of the load side meter.

[0091] Preferably, the extraction module 604 is specifically used to extract the zero-sequence current signal and the zero-sequence voltage signal from the fault traveling wave signal of the candidate region; obtain the electrical signal of the 110kV bus of the distribution network, and calculate the superposition component based on the electrical signal; wherein the superposition component is the difference between the first instantaneous value of the electrical signal after the fault and the second instantaneous value of the electrical signal before the fault; and calculate the fault traveling wave characteristics based on the zero-sequence current signal, the zero-sequence voltage signal and the superposition component.

[0092] Preferably, the fault traveling wave characteristics include statistical features, informational features, amplitude features, and impedance features; the statistical features include harmonic mean, standard deviation, mean deviation, and kurtosis; the informational features include Shannon entropy and Raney entropy; the amplitude features include root mean square, peak value, and maximum-minimum window difference; and the impedance features include post-fault apparent impedance and superimposed impedance.

[0093] Preferably, the electrical signal includes a pre-fault reference signal for the second complete signal cycle before the fault occurs, and a post-fault analysis signal for the third complete signal cycle after the fault occurs; the first instantaneous value is obtained based on the post-fault analysis signal, and the second instantaneous value is obtained based on the pre-fault reference signal.

[0094] Preferably, the determining module 603 is further configured to determine the fault type based on the fault traveling wave signal; and to determine the corresponding target decision tree ensemble model based on the candidate region and the fault type; the positioning module 605 is specifically configured to input the fault traveling wave feature into the target decision tree ensemble model to obtain the fault region output by the target decision tree ensemble model.

[0095] Preferably, the calculation module 602 is further configured to calculate a multiple estimation reduction rate based on the total line length of the fault area and the total line length of the candidate area; wherein the multiple estimation reduction rate is used to indicate the reduction of invalid investigation range.

[0096] The power distribution network fault location system provided in this application can be understood by referring to the relevant content in the foregoing method embodiment section, and will not be repeated here.

[0097] like Figure 7 As shown, Figure 7 This is a schematic diagram of a possible logical structure of a computing device provided in an embodiment of this application. The computing device 70 includes a processor 701, a communication interface 702, a memory 703, and a bus 704. The processor 701, the communication interface 702, and the memory 703 are interconnected via the bus 704. In an embodiment of this application, the processor 701 is used to control and manage the operation of the computing device 70. For example, the processor 701 is used to execute... Figure 1 The steps in the embodiments and / or other processes used in the techniques described herein. Communication interface 702 is used to support communication by computing device 70. Memory 703 is used to store program code and data of computing device 70.

[0098] The processor 701 can be a central processing unit, a general-purpose processor, a digital signal processor, an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. The processor can also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a digital signal processor and a microprocessor, etc. The bus 704 can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 7 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0099] In another embodiment of this application, a computer-readable storage medium is also provided, the computer-readable storage medium including instructions that, when executed on a computer, cause the computer to perform the above-described... Figure 1 The method described in the embodiments.

[0100] Those skilled in the art will recognize that the units of the various examples described in connection with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of the invention.

[0101] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0102] In the embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual couplings, direct couplings, or communication connections may be through some interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

[0103] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0104] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0105] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0106] 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 foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be covered within the scope of the claims and specification of the present invention.

Claims

1. A method for fault location in a power distribution network, characterized by, The method comprises: acquiring a fault traveling wave signal; based on the fault traveling wave signal, calculating an estimated distance between a fault point and a target electric meter; wherein the target electric meter is one of the double-ended electric meters that detects the fault traveling wave signal; based on the estimated distance, determining a candidate region from a plurality of line regions of a power distribution network; wherein the plurality of regions of the power distribution network are divided based on the deployment location of the electric meter and the branch point location; extracting a fault traveling wave feature based on the fault traveling wave signal of the candidate region; inputting the fault traveling wave feature into the corresponding decision tree ensemble model to obtain the fault region output by the decision tree ensemble model.

2. The method of claim 1, wherein, The double-ended electric meter comprises a power supply side electric meter and a load side electric meter; the candidate region is determined based on the estimated distance from a plurality of line regions of a power distribution network, comprising: in the case where the estimated distance is greater than 0 and less than the line length between the double-ended electric meters, the candidate region comprises the line region between the double-ended electric meters; in the case where the estimated distance is less than 0 and the target electric meter is the power supply side electric meter, the candidate region comprises the line region upstream of the power supply side electric meter; in the case where the estimated distance is greater than the line length between the double-ended electric meters and the target electric meter is the power supply side electric meter, the candidate region comprises the line region downstream of the load side electric meter.

3. The method of claim 2, wherein, The fault traveling wave feature is extracted based on the fault traveling wave signal of the candidate region, comprising: extracting a zero sequence current signal and a zero sequence voltage signal from the fault traveling wave signal of the candidate region; acquiring an electrical signal of a 110kV bus of the power distribution network, and calculating a superimposed component based on the electrical signal; wherein the superimposed component is the difference between the first instantaneous value of the electrical signal after the fault and the second instantaneous value of the electrical signal before the fault; calculating the fault traveling wave feature based on the zero sequence current signal, the zero sequence voltage signal and the superimposed component.

4. The method of claim 3, wherein, The fault traveling wave feature comprises statistical features, information features, amplitude features and impedance features; the statistical features comprise harmonic mean, standard deviation, average deviation and kurtosis; the information features comprise Shannon entropy and Renyi entropy; the amplitude features comprise root mean square, peak value and maximum minimum window difference; the impedance features comprise apparent impedance after the fault and superimposed impedance.

5. The method of claim 3, wherein, The electrical signal comprises a pre-fault reference signal of the second complete signal period before the fault occurs, and a post-fault analysis signal of the third complete signal period after the fault occurs; the first instantaneous value is obtained based on the post-fault analysis signal, and the second instantaneous value is obtained based on the pre-fault reference signal.

6. The method according to any one of claims 1-5, characterized in that, Before the fault traveling wave feature is inputted into the corresponding decision tree ensemble model to obtain the fault region output by the decision tree ensemble model, the method further comprises: determining a fault type based on the fault traveling wave signal; determining a corresponding target decision tree ensemble model based on the candidate region and the fault type; the fault traveling wave feature is inputted into the corresponding decision tree ensemble model to obtain the fault region output by the decision tree ensemble model, comprising: input the fault traveling wave feature into the target decision tree ensemble model to obtain a fault area output by the target decision tree ensemble model.

7. The method according to any one of claims 1-5, characterized in that, After the inputting of the fault traveling wave feature into the corresponding decision tree ensemble model to obtain the fault area output by the decision tree ensemble model, the method further comprises: based on the total length of the line of the fault area and the total length of the line of the candidate area, calculating a multiple estimation reduction rate; wherein the multiple estimation reduction rate is used to indicate the reduction of the invalid investigation range.

8. A power distribution network fault location system characterized by, The system is applied to the method of any one of claims 1-7, and the system comprises: an acquisition module configured to acquire a fault traveling wave signal; a calculation module configured to calculate an estimated distance between a fault point and a target electric meter based on the fault traveling wave signal; wherein the target electric meter is one of double-end electric meters detecting the fault traveling wave signal; a determination module configured to determine a candidate area from a plurality of line areas of a power distribution network based on the estimated distance; wherein the plurality of areas of the power distribution network are divided based on electric meter deployment positions and branch point positions; an extraction module configured to extract a fault traveling wave feature based on a fault traveling wave signal of the candidate area; a positioning module configured to input the fault traveling wave feature into a corresponding decision tree ensemble model to obtain a fault area output by the decision tree ensemble model.

9. A computing device, comprising: comprise: a memory configured to store a program; a processor configured to load the program to execute the method of any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium comprises a stored program, wherein when the program runs, the device where the computer readable storage medium is located executes the method of any one of claims 1-7. The computer readable storage medium comprises a stored program, wherein when the program runs, the device where the computer readable storage medium is located executes the method of any one of claims 1-7.

Images