Edge side power distribution network fault locating method and device based on digital twinning

Abstract

The embodiment of the application provides a kind of based on digital twinning's edge side distribution network fault location method and device, it is related to distribution internet of things technical field.The method comprises: obtaining the fault feature reported by edge gateway, the preliminary ranging result of fault is included in the fault feature;Using the digital twin model of the line where the fault is located, the fault parameters determined based on the fault feature are injected at the preliminary ranging result, obtain the simulation feature output by the digital twin model, calculate the deviation between the fault feature and the simulation feature;With the deviation, fault feature and line working condition as input feature, according to the mapping rule between input feature and ranging correction value obtained based on historical data, obtain ranging correction value, and obtain the corrected fault distance after the preliminary ranging result is corrected using the ranging correction value.The embodiment provided in the application finally realizes the rapid and accurate positioning of complex distribution network fault.

Description

Technical Field

[0001] This application relates to the field of power distribution Internet of Things (IoT) technology, specifically to a fault location method for edge-side power distribution networks based on digital twins, a fault location device for edge-side power distribution networks based on digital twins, an electronic device, a storage medium, and a computer program product. Background Technology

[0002] The power distribution network is the core end-point of the power system facing users, directly affecting the reliability of power supply. Faults in the distribution network (such as single-phase grounding, phase-to-phase short circuits, and open circuits) can easily lead to equipment damage, large-scale power outages, and other accidents, seriously affecting production, daily life, and the safe and stable operation of the power grid. With the widespread integration of distributed power sources, the increasing complexity of distribution network topologies, and the rise of complex fault scenarios such as mixed lines and high-resistance grounding, traditional fault location technologies are no longer adequate for the precise and rapid location requirements of smart distribution networks. Existing fault location technologies for distribution networks mainly fall into three categories: traditional measurement-based technologies (impedance method, traveling wave method) are easily affected by various factors; edge computing-based technologies offer fast response and moderate cost, but lack accuracy and have poor anti-interference capabilities; digital twin / cloud simulation-based technologies offer high location accuracy and adaptability to complex topologies, but suffer from large transmission delays, high computing power consumption, and data corruption.

[0003] Existing fault location technologies for power distribution networks can be mainly divided into three categories, all of which have significant drawbacks. Single edge computing-based ranging technology relies on low-power IoT sensors and local lightweight algorithms, offering fast response but insufficient sensor accuracy, difficulty adapting static parameters to topology changes, large ranging errors, and poor anti-interference and scenario adaptability. Single digital twin ranging technology achieves high-precision positioning through cloud simulation, but requires uploading massive amounts of data, resulting in significant transmission delays, high cloud computing power consumption, slow response speed, and easy functional failure when data is lost. Traditional impedance / traveling wave methods are mature but susceptible to factors such as transition resistance, load current, and line branches, leading to significant errors under complex faults. Dedicated devices are also costly, making them difficult to adapt to scenarios involving mixed lines and distributed power sources. None of these technologies can meet the comprehensive requirements of smart power distribution networks for rapid response, high precision, and adaptability to complex scenarios. Summary of the Invention

[0004] The purpose of this application is to provide a method and device for fault location in edge-side distribution networks based on digital twins. By constructing a collaborative architecture of "edge layer + cloud digital twin layer", fault data is collected locally on the edge side and preliminary ranging is completed through a lightweight algorithm. The cloud uses the virtual-real fusion of the digital twin model and machine learning to achieve error correction, and finally realizes rapid and accurate fault location in complex distribution networks, so as to at least solve some of the problems in the background technology.

[0005] To achieve the above objectives, this application provides a fault location method for edge-side distribution networks based on digital twins, comprising: acquiring fault features reported by the edge gateway, wherein the fault features include preliminary fault ranging results; using a digital twin model of the line where the fault is located, injecting fault parameters determined based on the fault features into the preliminary ranging results to obtain simulation features output by the digital twin model, and calculating the deviation between the fault features and the simulation features; using the deviation, fault features, and line operating conditions as input features, obtaining a ranging correction value according to the mapping law between the input features obtained based on historical data and the ranging correction value, and using the ranging correction value to correct the preliminary ranging results to obtain the corrected fault distance.

[0006] Optionally, after correcting the preliminary ranging result using the ranging correction value to obtain the corrected fault distance, the method further includes: determining whether there are other corrected fault distances on the faulty line, wherein the other corrected fault distances are obtained by processing the fault feature vectors and preliminary ranging results reported by other edge gateways on the faulty line using the digital twin model and the mapping rule; if there are other corrected fault distances, determining the corresponding fusion weight based on the node ranging variance of each corrected fault distance, and performing weighted fusion on all corrected fault distances to obtain the final fault distance.

[0007] Optionally, the preliminary ranging results of the fault reported by the edge gateway are obtained through the following steps: The arrival time of the traveling wave is defined as the moment when the first derivative of the current waveform changes abruptly after the fault, and the occurrence time of the fault is defined as the moment when the current changes abruptly. The difference between the arrival time of the traveling wave and the occurrence time of the fault is calculated. The first distance result is obtained based on the difference and the propagation speed determined based on the line type. The second distance result is obtained based on the relationship between the steady-state zero-sequence component and the zero-sequence impedance of the line after the fault. The preliminary ranging result is obtained by fusing the first distance result and the second distance result.

[0008] Optionally, fusing the first distance result and the second distance result to obtain the preliminary ranging result includes: using the maximum deviation ratio between the three-phase current and the average current as the unbalance degree; determining the fault type based on the relationship between the unbalance degree and a preset unbalance degree threshold; determining the fusion weight of the first distance result and the second distance result based on the fault type; and weighting and fusing the first distance result and the second distance result according to the fusion weight to obtain the preliminary ranging result.

[0009] Optionally, after determining the fusion weight of the first distance result and the second distance result according to the fault type, the method further includes: if the zero-sequence current in the steady-state zero-sequence component is less than a preset current threshold, then further reduce the fusion weight of the second distance result, and recalculate the fusion weight of the first distance result based on the sum of the fusion weights being 1.

[0010] Optionally, after obtaining the fault characteristics reported by the edge gateway and before adopting the digital twin model of the line where the fault occurs, the method further includes: performing a data alignment operation between the fault characteristics reported by the edge gateway and the digital twin model.

[0011] Optionally, the fault features include timestamp information; the data alignment operation of the fault features reported by the edge gateway with the digital twin model includes: aligning the fault features with the simulation timeline of the digital twin model using a unified reference clock.

[0012] Optionally, the fault characteristics include sensor location information; the data alignment operation of the fault characteristics reported by the edge gateway with the digital twin model includes: locating the installation location of the edge sensor in the digital twin model using the sensor location information and associating it with real-time topology information.

[0013] Optionally, the step of aligning the fault characteristics reported by the edge gateway with the digital twin model includes: correcting the preset basic zero-sequence impedance at the edge end according to the aging attenuation coefficient and the years of operation of the line, so as to obtain the zero-sequence impedance per unit length of the line after aging correction.

[0014] Optionally, the step of aligning the fault features reported by the edge gateway with the digital twin model includes: correcting the preliminary ranging result in the fault features reported by the edge gateway based on the rated output of the distributed power supply and the real-time output of the distributed power supply during the fault, to obtain a corrected preliminary ranging result; the corrected preliminary ranging result replaces the preliminary ranging result as the processing object of subsequent steps.

[0015] Optionally, the fault features reported by the edge gateway also include: the current value of the current change, the zero-sequence voltage, the zero-sequence current, the difference between the arrival time of the traveling wave and the time of the fault occurrence, and fault determination information; the current value of the current change, the zero-sequence voltage, the zero-sequence current, the difference between the arrival time of the traveling wave and the time of the fault occurrence, and the fault determination information are combined into a fault feature vector.

[0016] Optionally, the fault parameters include fault type and transition resistance; fault parameters determined based on the fault characteristics are injected into the preliminary ranging result, including: the transition resistance is calculated based on the zero-sequence impedance per unit length of the line, the preliminary ranging result, the zero-sequence voltage, and the zero-sequence current; the fault type determined by the fault determination information and the calculated transition resistance are injected into the preliminary ranging result.

[0017] Optionally, the mapping relationship between the input features obtained based on historical data and the ranging correction value is learned from training samples constructed based on historical data using a machine learning model.

[0018] Optionally, after obtaining the corrected fault distance or the final fault distance, the fault location and its impact range are visualized in the digital twin model.

[0019] This application also provides a fault location device for an edge-side distribution network based on digital twins. The device includes: a data acquisition module for acquiring fault features reported by the edge gateway, the fault features including preliminary fault ranging results; a simulation calculation module for using a digital twin model of the line where the fault is located, injecting fault parameters determined based on the fault features into the preliminary ranging results to obtain simulation features output by the digital twin model, and calculating the deviation between the fault features and the simulation features; and a deviation correction module for using the deviation, fault features, and line operating conditions as input features, obtaining a ranging correction value according to the mapping law between the input features obtained based on historical data and the ranging correction value, and using the ranging correction value to correct the preliminary ranging results to obtain the corrected fault distance.

[0020] This application also provides an electronic device, including: at least one processor; and a memory connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the at least one processor implements the aforementioned edge-side distribution network fault location method based on digital twin by executing the instructions stored in the memory.

[0021] This application also provides a machine-readable storage medium storing instructions that, when executed by a processor, configure the processor to perform the aforementioned digital twin-based edge-side distribution network fault location method.

[0022] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the aforementioned digital twin-based edge-side distribution network fault location method.

[0023] The above technical solution has the following beneficial effects: (1) A novel two-level collaborative fault location architecture combining edge computing and digital twins is proposed. In the architecture design, the dual advantages of local real-time processing on the edge side and global precise optimization in the cloud are maximized, which is conducive to improving the adaptability and response efficiency of fault location methods to complex power distribution networks (including hybrid lines and distributed power access).

[0024] (2) A simplified traveling wave-zero sequence impedance adaptive weighted fusion algorithm is proposed. This algorithm adds a fault type dynamic judgment mechanism on the basis of the traditional edge-side ranging algorithm, ensuring that the algorithm pays better attention to the core characteristics of different fault types, while improving the accuracy and lightweight adaptability of the edge-side preliminary ranging.

[0025] Other features and advantages of the embodiments of this application will be described in detail in the following detailed description section. Attached Figure Description

[0026] The accompanying drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the following detailed description to explain the embodiments of this application, but do not constitute a limitation on the embodiments of this application. In the drawings: Figure 1 This illustration schematically shows the steps of the edge-side distribution network fault location method based on digital twin according to an embodiment of this application; Figure 2 The illustration shows a schematic diagram of an implementation of the edge-side distribution network fault location method based on digital twins according to the embodiments of this application; Figure 3 This illustration schematically shows a structural diagram of an edge-side distribution network fault location device based on digital twins according to an embodiment of this application; Figure 4 The diagram schematically illustrates the internal structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0027] The specific embodiments of this application will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustration and explanation only and are not intended to limit the embodiments of this application.

[0028] Figure 1 The illustration shows a schematic diagram of the steps in the edge-side distribution network fault location method based on digital twins according to an embodiment of this application. For example... Figure 1 As shown, a fault location method for edge-side distribution networks based on digital twins is proposed, the method comprising: S01. Obtain the fault characteristics reported by the edge gateway, wherein the fault characteristics include the preliminary ranging results of the fault; S02. Using the digital twin model of the line where the fault is located, inject the fault parameters determined based on the fault characteristics into the preliminary ranging result to obtain the simulation characteristics output by the digital twin model, and calculate the deviation between the fault characteristics and the simulation characteristics. S03. Using the deviation, fault characteristics, and line conditions as input features, and based on the mapping relationship between the input features obtained from historical data and the ranging correction value, a ranging correction value is obtained. The ranging correction value is then used to correct the preliminary ranging result to obtain the corrected fault distance.

[0029] In this embodiment, it is preferably implemented based on a collaborative architecture of "edge layer + cloud digital twin layer". This method runs on the cloud digital twin layer. The fault characteristics reported by the edge gateway are generated by the edge layer through preset logic, which collects fault data locally. The fault characteristics include preliminary fault ranging results, which are obtained through a built-in lightweight algorithm. A preferred or optional implementation method will be provided later. The cloud utilizes the fusion of the virtual and real worlds of the digital twin model and machine learning to correct errors, ultimately achieving rapid and accurate fault location in complex distribution networks. This embodiment aims to solve the problem that existing technologies struggle to balance response speed, ranging accuracy, and adaptability to complex scenarios, providing an efficient fault ranging method suitable for hybrid lines and distributed power supply access scenarios.

[0030] In some embodiments of this application, after correcting the preliminary ranging result using the ranging correction value to obtain the corrected fault distance, the method further includes: determining whether there are other corrected fault distances on the faulty line, wherein the other corrected fault distances are obtained by processing the fault feature vectors and preliminary ranging results reported by other edge gateways on the faulty line using the digital twin model and the mapping rule; if there are other corrected fault distances, a corresponding fusion weight is determined based on the node ranging variance of each corrected fault distance, and all corrected fault distances are weighted and fused to obtain the final fault distance. This embodiment provides a multi-node fusion algorithm. If multiple edge sensors are deployed on the faulty line, the cloud performs weighted fusion of the multi-node correction results, with the weights inversely proportional to the ranging variance of each node, and finally outputs the final fault distance. The specific calculation process includes: 1. Calculation of variance in nodal distance measurement:

[0031] in, Let Variance be the distance measurement variance for the i-th node. Let j be the j-th corrected fault distance among the m corrected fault distances of the i-th node. Let m be the mean of the m corrected fault distances. m is the number of historical simulation samples for this node (greater than or equal to 1000). Let be the corrected distance for the i-th node in the j-th simulation. Let be the actual fault distance in the j-th simulation.

[0032] 2. Weighting of fusion:

[0033] in, Let n be the fusion weight of the i-th node, and n be the number of edge nodes participating in the fusion, satisfying the following condition: , Let Variance be the distance measurement variance of the k-th node out of a total of n nodes.

[0034] 3. Final fusion distance:

[0035] in, Let Variance be the distance measurement variance for the i-th node. The final fault distance, Correct the distance to the i-th node in the cloud.

[0036] In some embodiments of this application, the IoT sensors at the edge layer collect real-time power distribution network operation data. When a sudden current change (ΔI > 3 times the rated current) is detected, a fault is determined to have occurred, the fault time t0 is recorded, and the three-phase voltage / current waveforms from 0.1s before the fault to 0.3s after the fault are collected. The edge gateway then uses a weighted fusion of the "simplified traveling wave method + zero-sequence impedance method" to calculate the preliminary ranging result. Specifically, the preliminary ranging results of the fault reported by the edge gateway are obtained through the following steps: The simplified traveling wave method abandons the complex wavelet transform for extracting the traveling wave head and detects the arrival time of the traveling wave by using a threshold for abrupt changes in the first derivative of the current. Specifically, the arrival time of the traveling wave is defined as the moment when the first derivative of the current waveform changes abruptly after the fault. That is, the first difference of the current waveform after the fault is calculated, which replaces the derivative calculation here to reduce computational cost. When the absolute value of the difference exceeds a threshold... (When the preset value is 0.5 kA / ms), the time when the traveling wave arrives at the edge node is recorded. The fault occurrence time is determined by a sudden change in current, i.e., the triggering condition for a sudden change in current, for example... Determined as the time of the fault occurrence Calculate the difference between the arrival time of the traveling wave and the time of the fault occurrence, i.e.: .

[0037] The propagation speed determined by the line type is obtained through the following steps: Preset the propagation speed according to the line type (overhead line / cable line / mixed line). Among them, overhead lines Cable Mixed lines are calculated based on length proportions: +(1- )

[0038] in, The percentage of overhead line length in the hybrid line is preset in the edge gateway and comes from the pre-configured parameters of the digital twin model.

[0039] The first distance result is obtained based on the difference and the propagation speed determined based on the line type, namely:

[0040] Should This is a simplified calculation of the fault distance using the traveling wave method.

[0041] The second distance result is obtained based on the relationship between the steady-state zero-sequence component after the fault and the line zero-sequence impedance. Specifically, it includes:

[0042] in, Zero-sequence impedance per unit length of the line, unit: Pre-set in the edge gateway, from the digital twin model device parameter library; if In the case of high-resistance grounding, it is marked. For results deemed unreliable, their actual effective weight is automatically reduced during subsequent weighting. Finally, the first distance result and the second distance result are fused to obtain the preliminary ranging result.

[0043] In some embodiments of this application, the preliminary ranging result is obtained by fusing the first distance result and the second distance result, including: The imbalance is defined as the percentage of the maximum deviation between the three-phase current and the average current, used for dynamically determining the fault type.

[0044] in, = This is the average value of the three-phase current; This represents the degree of imbalance.

[0045] The fault type is determined based on the relationship between the imbalance degree and a preset imbalance threshold. For example, the preset imbalance threshold is 0.2, which can be adjusted according to the on-site scenario. When it is determined to be a single-phase ground fault, when The fault was determined to be a phase-to-phase short circuit.

[0046] Determining the fusion weights of the first and second distance results based on the fault type includes: dynamically allocating weights according to the fault type. and The weights satisfy the normalization condition. For example: when (Phase-to-phase short-circuit fault): Give the traveling wave method a higher weight. =0.8, ;when (Single-phase ground fault): Give higher weight to the zero-sequence impedance method. =0.3, .

[0047] The first distance result and the second distance result are weighted and fused according to the fusion weight to obtain the preliminary ranging result, that is: the final preliminary ranging result is: +

[0048] in, The initial ranging results output from the edge are used to upload to the cloud twin layer for further processing, i.e., subsequent precise correction.

[0049] In some embodiments of this application, after determining the fusion weights of the first distance result and the second distance result based on the fault type, the method further includes: if the zero-sequence current in the steady-state zero-sequence component is less than a preset current threshold, then further reducing the fusion weight of the second distance result, and recalculating the fusion weight of the first distance result based on the sum of the fusion weights being 1. This embodiment takes into account the aforementioned situation. This refers to a high-resistance grounding scenario, in which case it is marked as follows: If the result is of low reliability, its actual effective weight will be automatically reduced during subsequent weighting.

[0050] The aforementioned embodiments of this application innovatively propose a simplified traveling wave-zero sequence impedance adaptive weighted fusion algorithm. This algorithm adds a dynamic fault type determination mechanism to the traditional edge-side ranging algorithm, ensuring that the algorithm pays better attention to the core characteristics of different fault types, while improving the accuracy and lightweight adaptability of the initial edge-side ranging.

[0051] In some embodiments of this application, after obtaining the fault characteristics reported by the edge gateway and before using the digital twin model of the line where the fault occurs, the method further includes: performing a data alignment operation between the fault characteristics reported by the edge gateway and the digital twin model. The core of the digital twin layer in this application consists of a distribution network digital twin model, an electromagnetic transient simulation engine, and a lightweight random forest (RF) machine learning model, supported by a dynamic topology database, an equipment parameter database, and a historical fault database. This embodiment mainly achieves precise spatiotemporal alignment between edge data and the virtual model, which is more conducive to the accurate matching of fault characteristics reported by the edge gateway in the digital twin model, thereby obtaining a more accurate fault distance.

[0052] Furthermore, the aforementioned alignment operation includes time alignment, i.e., time synchronization. The fault features include timestamp information; the data alignment operation of the fault features reported by the edge gateway with the digital twin model includes: aligning the fault features with the simulation timeline of the digital twin model using a unified reference clock, such as GPS / BeiDou time stamps, with a synchronization accuracy controllable to no more than 1μs.

[0053] Furthermore, the aforementioned alignment operation also includes spatial alignment, i.e., spatial matching. The fault features include sensor location information; the data alignment operation of the fault features reported by the edge gateway with the digital twin model includes: locating the installation location of the edge sensor in the digital twin model using the sensor location information, and associating it with real-time topology information, which includes: line segments, branch nodes, and distributed power supply access locations. This step is crucial for the accuracy of the calculated final fault distance in corresponding to the real-world scenario.

[0054] Furthermore, the aforementioned alignment operation also includes parameter updates, that is, calling the dynamic device parameters of the digital twin model, such as the actual impedance after line aging. Real-time output of distributed power sources The replacement of preset parameters at the edge end can include: the data alignment operation between the fault characteristics reported by the edge gateway and the digital twin model, which includes: correcting the preset basic zero-sequence impedance at the edge end according to the aging attenuation coefficient and the service life of the line, to obtain the aging-corrected zero-sequence impedance per unit length of the line, i.e.:

[0055] in, This refers to the zero-sequence impedance per unit length of the line after aging correction. The baseline zero-sequence impedance is preset for the edge end, k is the aging attenuation coefficient, and T is the line's operational lifespan.

[0056] Furthermore, the data alignment operation between the fault features reported by the edge gateway and the digital twin model includes: correcting the preliminary ranging results in the fault features reported by the edge gateway based on the rated output of the distributed power supply and the real-time output of the distributed power supply during the fault, to obtain the corrected preliminary ranging results, i.e.:

[0057] in, The initial distance after adjusting for distributed power generation output. For the rated output of distributed power sources, To provide real-time power output for distributed power sources during faults.

[0058] Since an alignment step exists in some implementations, the corrected preliminary ranging result obtained based on the alignment step is used as the object of processing in subsequent steps instead of the preliminary ranging result. This will not be elaborated further below.

[0059] In some embodiments of this application, the fault characteristics reported by the edge gateway further include: a sudden current value ΔI and a zero-sequence voltage. Zero-sequence current The difference between the arrival time of the traveling wave and the time of the fault occurrence Imbalance (Since the fault type in this application is determined by the imbalance degree δ, the reported fault determination information is the imbalance degree.) The current value of the sudden current change, the zero-sequence voltage, the zero-sequence current, the difference between the arrival time of the traveling wave and the time of the fault occurrence, and the fault determination information are combined into a fault feature vector, denoted as fault feature vector X=[ΔI, , , t1-t0,δ].

[0060] Among them, the symmetrical component method is used to extract the zero-sequence voltage from the three-phase voltage and current data. and zero-sequence current This serves as the basis for calculations using the zero-sequence impedance method:

[0061]

[0062] in, , , The instantaneous values ​​of the three-phase voltage collected by the edge sensor. , , This represents the instantaneous value of the three-phase current.

[0063] In some embodiments of this application, the fault parameters include fault type and transition resistance; the fault parameters determined based on the fault characteristics are injected into the preliminary ranging result, including: calculating the transition resistance based on the zero-sequence impedance per unit length of the line, the preliminary ranging result, the zero-sequence voltage, and the zero-sequence current, i.e.:

[0064] in, The transition resistance at the fault point. The measured zero-sequence voltage is (kV). The measured zero-sequence current (kA) is given. Zero-sequence impedance per unit length of the line ( / km), The initial distance (km) after power output correction for distributed generation, in the absence of When adopted .

[0065] The fault type determined by the fault judgment information and the calculated transition resistance are injected into the preliminary distance measurement result. At the location, inject the fault type (short circuit / grounding) and transition resistance consistent with the actual measurement. .

[0066] Prior to this, the initialization of the digital twin model is also included, including loading the digital twin model of the faulty line, including line parameters, topology, and distributed power supply access information.

[0067] The final digital twin model outputs simulation features. To calculate the deviation between the fault features and the simulation features, the data structure of the simulation features needs to match the fault features. Preferably, the fault features adopt a fault feature vector X=[ΔI, , , t1-t0, The simulation features are set as simulation feature vectors. .in, These are the current values ​​of the current surge, zero-sequence voltage, zero-sequence current, the difference between the arrival time of the traveling wave and the time of the fault, and the unbalance degree obtained from the simulation of the digital twin model.

[0068] Accordingly, the deviation between the calculated fault characteristics and the simulated characteristics in step S02 includes:

[0069] Where X is the measured fault feature vector at the edge. The simulation feature vector is obtained from the simulation. The deviation is preferably a normalized deviation value.

[0070] In some embodiments of this application, the mapping relationship between the input features obtained based on historical data and the ranging correction value is learned from training samples constructed based on historical data using a machine learning model. For example, a lightweight random forest (RF) model is used for error correction, specifically as follows: Model training input and output definitions: Model inputs: measured feature vector X, virtual-to-real deviation Line type proportion Real-time output of distributed power sources Model output: Prediction error Model Training: Over 100,000 samples were generated based on a digital twin model, covering different fault distances, transition resistances, distributed power output, and line types. Sample inputs included edge feature vectors, deviations, line types, and real-time distributed power output; the output was the true error. A random forest model was trained with 50 decision trees and a maximum depth of 8 to obtain a distance correction value. The preliminary distance measurement results were then corrected using this correction value to obtain the corrected fault distance, including:

[0071] in, This is the corrected fault distance; when there is At that time, adopt Replace the above formula .

[0072] In some embodiments of this application, after obtaining the corrected fault distance or the final fault distance, the fault location and its impact range are visualized in the digital twin model. The digital twin model, as a high-fidelity virtual mapping of the physical power grid, includes the three-dimensional spatial coordinates and complete topological relationships of the topology, distributed power source access information, etc. Along the line route, the system automatically generates prominent fault markers at that distance, such as flashing red warning dots or three-dimensional exploded views, making the fault point immediately apparent. Simultaneously, combining switch tripping signals and topology coloring analysis, the system calculates in real time the power outage sections and related equipment affected by the fault, and prominently marks the power loss range and isolation areas in the model using semi-transparent red areas, highlighted boundaries, or dynamic particle flows. This visualization method allows control and repair personnel to intuitively and accurately grasp the exact location of the fault and its impact area, thereby effectively improving the efficiency of fault assessment and emergency response.

[0073] This application innovatively proposes a two-level collaborative fault location architecture that combines edge computing and digital twins. The architecture design maximizes the dual advantages of local real-time processing on the edge side and global precise optimization in the cloud, which helps to improve the adaptability and response efficiency of fault location methods to complex power distribution networks (including hybrid lines and distributed power source access).

[0074] Figure 2 The illustration shows a schematic diagram of an implementation of the edge-side distribution network fault location method based on digital twins according to an embodiment of this application. For example... Figure 2 As shown, the edge-side distribution network fault location method based on digital twins includes the following steps.

[0075] (1) Edge layer data acquisition. The IoT sensors at the edge layer acquire power distribution network operation data in real time. When a sudden change in current (ΔI>3 times the rated current) is detected, the fault is determined to have occurred, the fault time t0 is recorded, and the three-phase voltage / current waveforms from 0.1s before the fault to 0.3s after the fault are acquired.

[0076] (2) Preliminary fault distance measurement at the edge layer. The edge gateway uses a weighted fusion of the simplified traveling wave method and the zero-sequence impedance method to calculate the preliminary fault distance. .

[0077] (3) Edge layer feature upload. The edge gateway uploads the preliminary ranging results. Fault feature vector X=[ΔI, , The sensor location coordinates and timestamps [t1-t0, δ] are uploaded to the cloud twin layer through the communication layer, transmitting only key feature data to reduce communication bandwidth usage.

[0078] (4) Alignment of virtual and physical data in the cloud twin layer. After receiving data from the edge, the cloud twin layer performs the following operations: Time synchronization: Align edge data with the timeline of the twin model simulation using GPS / BeiDou time stamps (synchronization accuracy ≤1μs). Spatial matching: Locate the installation position of edge sensors in the digital twin model and associate them with the real-time topology (line segments, branch nodes, and distributed power supply access locations). Parameter update: Calls dynamic device parameters from the digital twin model, such as the actual impedance after line aging. The calculation formula is described above in the section on real-time output of distributed power sources. Replace the preset parameters at the edge and update the initial distance after the influence of distributed power sources. .

[0079] (5) Cloud-based twin model simulation. The fault scenario is reproduced based on the digital twin model, specifically as follows: Model initialization: Load the digital twin model of the faulty line, including line parameters, topology, and distributed power supply access information; Fault injection: In the initial distance... At the location, inject the fault type (short circuit / grounding) and transition resistance consistent with the actual measurement. (pass / (Estimation, calculation formulas are described above); Simulation output: Obtain the simulation feature vector. Deviation quantification: Calculate the relative deviation between measured features and simulated features.

[0080] (6) Machine learning error compensation. A lightweight random forest (RF) model is used to correct the error, as follows: Model training: 100,000+ samples are generated based on the digital twin model (covering different fault distances, transition resistances, distributed power output, and line types). The sample inputs are edge feature vectors, deviations, line types, and real-time output of distributed power sources, and the output is the true error. The random forest model is trained (number of decision trees = 50, maximum depth = 8) to obtain the error prediction model.

[0081] (7) Multi-node data fusion. If multiple edge sensors are deployed on the faulty line, the cloud performs weighted fusion of the multi-node correction results. The weight is inversely proportional to the ranging variance of each node, and finally outputs the accurate fault distance. The location and impact of the fault are visualized in the digital twin model.

[0082] As can be seen from the above implementation methods, the method in this application integrates the real-time performance of edge computing with the accuracy of digital twins, forming a closed-loop process of "measurement-preliminary calculation-cloud correction-fusion output". This not only solves the core contradiction of "precision and speed cannot be achieved at the same time" in the existing technology, but also improves the adaptability to complex power distribution network scenarios, which can effectively improve the efficiency of power distribution network fault repair and ensure power supply reliability.

[0083] Based on the same inventive concept, this application also provides an edge-side distribution network fault location device based on digital twins. Figure 3 A schematic diagram illustrating the structure of a digital twin-based edge-side distribution network fault location device according to an embodiment of this application is shown. Figure 3 As shown, the device includes: a data acquisition module for acquiring fault features reported by the edge gateway, the fault features including preliminary fault ranging results; a simulation calculation module for using a digital twin model of the line where the fault is located, injecting fault parameters determined based on the fault features into the preliminary ranging results to obtain simulation features output by the digital twin model, and calculating the deviation between the fault features and the simulation features; and a deviation correction module for using the deviation, fault features, and line operating conditions as input features, obtaining a ranging correction value according to the mapping rule between the input features obtained based on historical data and the ranging correction value, and using the ranging correction value to correct the preliminary ranging results to obtain the corrected fault distance.

[0084] The specific limitations of each functional module in the aforementioned digital twin-based edge-side distribution network fault location device can be found in the limitations of the digital twin-based edge-side distribution network fault location method described above, and will not be repeated here. Each module in the above system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module. This also achieves the beneficial effects of improving adaptability to complex distribution network scenarios, effectively improving the efficiency of distribution network fault repair, and ensuring power supply reliability.

[0085] In some embodiments of this application, an electronic device is also provided, comprising: at least one processor; and a memory connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, which executes the aforementioned edge-side distribution network fault location method based on digital twins. Its internal structure diagram can be shown as follows. Figure 4 As shown. Figure 4 The diagram schematically illustrates the internal structure of an electronic device according to an embodiment of this application. The electronic device includes a processor A01, a network interface A02, a memory (not shown), and a database (not shown) connected via a system bus. The processor A01 provides computing and control capabilities. The memory includes internal memory A03 and a non-volatile storage medium A04. The non-volatile storage medium A04 stores an operating system B01, a computer program B02, and a database (not shown). The internal memory A03 provides an environment for the operation of the operating system B01 and the computer program B02 stored in the non-volatile storage medium A04. The network interface A02 is used for communication with external terminals via a network connection. When the computer program B02 is executed by the processor A01, it implements a digital twin-based edge-side distribution network fault location method.

[0086] Those skilled in the art will understand that Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the electronic device to which the present application is applied. The specific electronic device may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0087] In one embodiment provided in this application, a machine-readable storage medium is provided, on which instructions are stored, which, when executed by a processor, cause the processor to be configured to perform the aforementioned edge-side distribution network fault location method based on digital twin.

[0088] In one embodiment provided in this application, a computer program product is provided, including a computer program that, when executed by a processor, implements the aforementioned edge-side distribution network fault location method based on digital twins.

[0089] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0090] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0091] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0092] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0093] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0094] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0095] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0096] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0097] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A fault location method for edge-side distribution networks based on digital twins, characterized in that, The method includes: Obtain the fault characteristics reported by the edge gateway, including the preliminary ranging results of the fault; Using the digital twin model of the line where the fault is located, fault parameters determined based on the fault characteristics are injected into the preliminary ranging results to obtain the simulation characteristics output by the digital twin model, and the deviation between the fault characteristics and the simulation characteristics is calculated. Using the aforementioned deviation, fault characteristics, and line operating conditions as input features, and based on the mapping relationship between the input features obtained from historical data and the ranging correction value, a ranging correction value is obtained. The ranging correction value is then used to correct the preliminary ranging result to obtain the corrected fault distance.

2. The method according to claim 1, characterized in that, After correcting the preliminary ranging result using the ranging correction value to obtain the corrected fault distance, the method further includes: Determine whether there are other corrected fault distances on the line where the fault is located. The other corrected fault distances are: the fault feature vectors reported by other edge gateways on the line where the fault is located and the preliminary distance measurement results of the fault, which are processed by the digital twin model and the mapping law. If other corrected fault distances exist, the corresponding fusion weight is determined based on the node ranging variance of each corrected fault distance. The final fault distance is obtained by weighted fusion of all corrected fault distances.

3. The method according to claim 1, characterized in that, The preliminary ranging results of the fault reported by the edge gateway are obtained through the following steps: The arrival time of the traveling wave is defined as the moment when the first derivative of the current waveform changes abruptly after the fault, and the fault occurrence time is defined as the moment when the current changes abruptly. The difference between the arrival time of the traveling wave and the fault occurrence time is calculated, and the first distance result is obtained based on the difference and the propagation speed determined based on the line type. The second distance result is obtained based on the relationship between the steady-state zero-sequence component and the zero-sequence impedance of the line after the fault. The preliminary distance measurement result is obtained by fusing the first distance result and the second distance result.

4. The method according to claim 3, characterized in that, The preliminary ranging result is obtained by fusing the first distance result and the second distance result, including: The unbalance degree is defined as the percentage of the maximum deviation between the three-phase current and the average current. The fault type is determined based on the relationship between the imbalance degree and the preset imbalance threshold. The fusion weights of the first and second distance results are determined based on the fault type. The first distance result and the second distance result are weighted and fused according to the fusion weight to obtain the preliminary distance measurement result.

5. The method according to claim 4, characterized in that, After determining the fusion weights of the first distance result and the second distance result based on the fault type, the method further includes: If the zero-sequence current in the steady-state zero-sequence component is less than a preset current threshold, the fusion weight of the second distance result is further reduced, and the fusion weight of the first distance result is recalculated based on the sum of the fusion weights being 1.

6. The method according to claim 1, characterized in that, After obtaining the fault characteristics reported by the edge gateway and before using the digital twin model of the line where the fault occurred, the method further includes: The fault characteristics reported by the edge gateway are aligned with the digital twin model.

7. The method according to claim 6, characterized in that, The fault characteristics include timestamp information; The step of aligning the fault characteristics reported by the edge gateway with the digital twin model includes: By using a unified reference clock, the fault characteristics are aligned with the simulation timeline of the digital twin model.

8. The method according to claim 6, characterized in that, The fault characteristics include sensor location information; The step of aligning the fault characteristics reported by the edge gateway with the digital twin model includes: In the digital twin model, the installation location of the edge sensor is located using the sensor location information and associated with real-time topology information.

9. The method according to claim 6, characterized in that, The step of aligning the fault characteristics reported by the edge gateway with the digital twin model includes: The baseline zero-sequence impedance at the edge end is corrected based on the aging attenuation coefficient and the number of years the line has been in operation, resulting in the zero-sequence impedance per unit length of the line after aging correction.

10. The method according to claim 6, characterized in that, The step of aligning the fault characteristics reported by the edge gateway with the digital twin model includes: The preliminary ranging results in the fault characteristics reported by the edge gateway are corrected based on the rated output of the distributed power source and the real-time output of the distributed power source during a fault, so as to obtain the corrected preliminary ranging results. The corrected preliminary ranging result replaces the initial ranging result as the object of processing in subsequent steps.

11. The method according to claim 1, characterized in that, The fault features reported by the edge gateway also include: the current value of the current change, the zero-sequence voltage, the zero-sequence current, the difference between the arrival time of the traveling wave and the time of the fault occurrence, and fault determination information; the current value of the current change, the zero-sequence voltage, the zero-sequence current, the difference between the arrival time of the traveling wave and the time of the fault occurrence, and the fault determination information are combined into a fault feature vector.

12. The method according to claim 11, characterized in that, The fault parameters include the fault type and transition resistance; Injecting fault parameters determined based on the fault characteristics into the preliminary ranging result, including: The transition resistance is calculated based on the zero-sequence impedance per unit length of the line, the preliminary distance measurement results, the zero-sequence voltage, and the zero-sequence current. The fault type determined by the fault determination information and the calculated transition resistance are injected into the preliminary ranging result.

13. The method according to claim 1, characterized in that, The mapping relationship between the input features obtained from historical data and the ranging correction value is learned by a machine learning model from training samples constructed based on historical data.

14. The method according to claim 2, characterized in that, After obtaining the corrected fault distance or the final fault distance, the fault location and its impact range are visualized in the digital twin model.

15. A fault location device for edge-side distribution networks based on digital twins, characterized in that, The device includes: The data acquisition module is used to acquire fault characteristics reported by the edge gateway, including preliminary ranging results of the fault. The simulation calculation module is used to inject fault parameters determined based on the fault characteristics into the preliminary ranging results using the digital twin model of the line where the fault is located, to obtain the simulation characteristics output by the digital twin model, and to calculate the deviation between the fault characteristics and the simulation characteristics. The deviation correction module is used to obtain a distance correction value by taking the deviation, fault characteristics and line conditions as input features, and according to the mapping law between the input features obtained based on historical data and the distance correction value. The distance correction value is then used to correct the preliminary distance measurement result to obtain the corrected fault distance.

16. An electronic device, characterized in that, include: At least one processor; A memory connected to the at least one processor; The memory stores instructions that can be executed by the at least one processor, and the at least one processor implements the edge-side distribution network fault location method based on digital twins as described in any one of claims 1 to 14 by executing the instructions stored in the memory.

17. A computer-readable storage medium having a computer program / instructions stored thereon, characterized in that, When the computer program / instruction is executed by the processor, it implements the edge-side distribution network fault location method based on digital twins as described in any one of claims 1 to 14.

18. A computer program product comprising a computer program / instructions, characterized in that, When the computer program / instruction is executed by the processor, it implements the edge-side distribution network fault location method based on digital twins as described in any one of claims 1 to 14.

Images