A power distribution network voltage anomaly diagnosis method and system based on a graph model

Abstract

The application discloses a power distribution network voltage anomaly diagnosis method and system based on a graph model, and the method comprises the following steps: constructing a power distribution network graph model according to the geographical topology of the power distribution network; detecting the voltage of a load node in the power distribution network graph model to obtain a voltage anomaly diagnosis candidate set; diagnosing a to-be-detected branch in the voltage anomaly diagnosis candidate set by using a hierarchical backtracking screening algorithm to obtain a voltage anomaly positioning result; and formulating a power distribution network maintenance plan according to the voltage anomaly positioning result. The application can accurately perform voltage anomaly diagnosis without relying on an intelligent algorithm, and reduces the occupation of computing resources.

Description

Technical Field

[0001] This application belongs to the field of power grid voltage monitoring technology, specifically relating to a distribution network voltage anomaly diagnosis method and system based on a graph model. Background Technology

[0002] The distribution network receives electrical energy from the transmission network or regional power plants and supplies power to various users locally or in stages according to voltage through distribution facilities. Abnormal voltage in the distribution network caused by load changes, equipment failures, external forces, etc., can lead to a decrease in the stability and reliability of the power supply system, damage to users' electrical equipment, and may even cause large-scale power outages. Therefore, it is necessary to diagnose abnormal voltage in the distribution network.

[0003] In existing technologies, most methods for diagnosing voltage anomalies in distribution networks rely on intelligent algorithms such as neural networks and machine learning. However, intelligent algorithms have high computational resource consumption, slow convergence, and require a large amount of high-quality data for training and learning. Moreover, intelligent algorithms are sensitive to the setting of initial or given parameters. When the parameter settings deviate to a certain extent, it may lead to incorrect results in the diagnosis of voltage anomalies in distribution networks, which is time-consuming and labor-intensive. Summary of the Invention

[0004] This application proposes a graph model-based method and system for diagnosing voltage anomalies in distribution networks. It can accurately diagnose voltage anomalies without relying on intelligent algorithms, thus reducing the consumption of computing resources.

[0005] The first aspect of this application provides a method for diagnosing voltage anomalies in distribution networks based on a graphical model, the method comprising:

[0006] Based on the geographical topology of the distribution network, construct a distribution network diagram model;

[0007] Detect the voltage of load nodes in the power distribution network diagram model to obtain a candidate set for voltage anomaly diagnosis;

[0008] The voltage anomaly location result is obtained by diagnosing the branches to be detected in the candidate set of voltage anomaly diagnosis through a hierarchical backtracking screening algorithm.

[0009] Based on the voltage anomaly location results, a power distribution network maintenance plan is formulated.

[0010] The above scheme first constructs a graphical model accurately describing the distribution network information based on its geographical topology. Then, it performs voltage detection on the load nodes of the distribution network based on the graphical model. By identifying load nodes with voltage anomalies, it determines the power supply branches that may have voltage anomalies and adds these branches to a voltage anomaly diagnosis candidate set, providing data support for subsequent branch detection. Then, a hierarchical backtracking filtering algorithm sequentially diagnoses voltage anomalies on the power supply branches that may have them. Accurate voltage anomaly location results can be obtained by detecting only certain power supply branches, eliminating the need for a neural network to detect all power supply branches, thus saving significant computational resources. Furthermore, no parameter settings are required during the detection process, reducing the risk of detection errors due to parameter settings issues.

[0011] In one possible implementation of the first aspect, the voltage of load nodes in the distribution network diagram model is detected to obtain a candidate set for voltage anomaly diagnosis, specifically:

[0012] Obtain the voltage of the load nodes in the power distribution network diagram model;

[0013] Based on a preset reference voltage range, the voltage of each load node is compared with the corresponding reference voltage range to determine the load nodes with abnormal voltage in the distribution network diagram model.

[0014] Based on the load nodes with voltage anomalies, the power supply branches of the distribution network diagram model are screened to obtain a candidate set for voltage anomaly diagnosis.

[0015] The above scheme locates voltage anomalies by detecting voltage at load nodes in the distribution network. The voltage at each load node is compared with its corresponding reference voltage to identify the load node with the anomaly. Then, the power supply branches containing and connected to the load nodes with anomalies are screened, and these branches are listed as potentially risky for voltage anomalies and added to the voltage anomaly diagnosis candidate set to provide data support for subsequent branch detection.

[0016] In one possible implementation of the first aspect, the power supply branches of the distribution network diagram model are screened based on the load nodes with voltage anomalies to obtain a candidate set for voltage anomaly diagnosis, specifically as follows:

[0017] Based on the distribution network diagram model, the power supply branch where each load node with voltage anomaly is located is determined and denoted as the first branch;

[0018] By using the load nodes with abnormal voltage, the power supply branches connected to the first branch in the distribution network diagram model are screened to obtain the branches to be detected.

[0019] Based on the branch to be tested, a set of candidates for voltage anomaly diagnosis is obtained.

[0020] In one possible implementation of the first aspect, a hierarchical backtracking filtering algorithm is used to diagnose the branches to be detected in the candidate set for voltage anomaly diagnosis, thereby obtaining the voltage anomaly location result, specifically as follows:

[0021] Select a starting point from the non-loaded nodes of the branch to be tested;

[0022] Starting from the aforementioned starting point and ending at the load node with abnormal voltage corresponding to the starting point, anomaly diagnosis is performed on the non-load nodes of all branches to be tested using a hierarchical backtracking screening algorithm to identify power supply branches with abnormal voltage.

[0023] Based on all power supply branches with voltage anomalies, determine the voltage anomaly location results.

[0024] In one possible implementation of the first aspect, an anomaly diagnosis is performed on the non-load nodes of all branches to be detected using a hierarchical backtracking screening algorithm to identify power supply branches with voltage anomalies, specifically:

[0025] Based on the historical failure rate of the branch, anomaly diagnosis is performed on the second branch where the starting point is located; among all the branches to be tested where the starting point is located, the second branch has the highest historical failure rate.

[0026] If there is a voltage abnormality in the second branch, stop the abnormality diagnosis and determine that the second branch is a power supply branch with a voltage abnormality.

[0027] If there is no voltage abnormality in the second branch, then perform anomaly diagnosis on the next level of the branch to be tested until a power supply branch with voltage abnormality is identified.

[0028] The above scheme first selects the branch with the highest historical failure rate to begin voltage diagnosis. Then, when the currently detected branch is normal, the next level of branches is tested to improve the speed of finding power supply branches with abnormal voltage. Moreover, only some power supply branches need to be tested to accurately obtain the voltage anomaly location result, eliminating the need for the neural network to test all power supply branches, thus saving a lot of computing resources.

[0029] The second aspect of this application provides a distribution network voltage anomaly diagnosis system based on a graph model, the system comprising: a graph model construction module, an anomaly diagnosis candidate set construction module, an anomaly location result acquisition module, and a distribution network maintenance module;

[0030] The graph model building module is used to build a graph model of the power distribution network based on the geographical topology of the power distribution network.

[0031] The anomaly diagnosis candidate set construction module is used to detect the voltage of load nodes in the distribution network diagram model and obtain the voltage anomaly diagnosis candidate set;

[0032] The anomaly location result acquisition module is used to diagnose the branches to be detected in the voltage anomaly diagnosis candidate set through a hierarchical backtracking filtering algorithm to obtain the voltage anomaly location result.

[0033] The power distribution network maintenance module is used to formulate a power distribution network maintenance plan based on the voltage anomaly location results.

[0034] In one possible implementation of the second aspect, the anomaly diagnosis candidate set construction module includes: a voltage anomaly detection unit;

[0035] The voltage anomaly detection unit is used to acquire the voltage of load nodes in the distribution network diagram model; based on a preset reference voltage range, the voltage of each load node is compared with the corresponding reference voltage range to determine the load nodes with voltage anomalies in the distribution network diagram model; and the power supply branches of the distribution network diagram model are screened according to the load nodes with voltage anomalies to obtain a candidate set for voltage anomaly diagnosis.

[0036] In one possible implementation of the second aspect, the anomaly diagnosis candidate set construction module includes: a branch acquisition unit to be detected;

[0037] The branch to be detected acquisition unit is used to determine the power supply branch where each load node with voltage anomaly is located based on the distribution network diagram model, denoted as the first branch; through the load node with voltage anomaly, the power supply branches connected to the first branch in the distribution network diagram model are screened to obtain the branch to be detected; and based on the branch to be detected, a voltage anomaly diagnosis candidate set is obtained.

[0038] In one possible implementation of the second aspect, the anomaly location result acquisition module includes: a voltage anomaly branch determination unit;

[0039] The voltage anomaly branch determination unit is used to select a starting point from the non-load nodes of the branch to be detected; taking the starting point as the starting point and the load node with voltage anomaly corresponding to the starting point as the ending point, it performs anomaly diagnosis on the non-load nodes of all branches to be detected through a hierarchical backtracking screening algorithm to determine the power supply branches with voltage anomalies; and determines the voltage anomaly location result based on all power supply branches with voltage anomalies.

[0040] In one possible implementation of the second aspect, the anomaly localization result acquisition module includes: a branch-level detection unit;

[0041] The branch-level detection unit is used to perform anomaly diagnosis on the second branch where the starting point is located based on the branch's historical failure rate. Among all the branches to be detected where the starting point is located, the second branch has the highest historical failure rate. If the second branch has a voltage anomaly, the anomaly diagnosis is stopped, and the second branch is determined to be a power supply branch with a voltage anomaly. If the second branch does not have a voltage anomaly, the anomaly diagnosis is performed on the next level of branches to be detected until a power supply branch with a voltage anomaly is determined. Attached Figure Description

[0042] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0043] Figure 1 This is a schematic flowchart of a distribution network voltage anomaly diagnosis method based on a graph model provided in a certain embodiment of this application;

[0044] Figure 2 This is a node-branch schematic diagram of a distribution network voltage anomaly diagnosis method based on a graph model provided in a certain embodiment of this application;

[0045] Figure 3 This is a specific structural diagram of a distribution network voltage anomaly diagnosis system based on a graph model provided in a certain embodiment of this application. Detailed Implementation

[0046] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0047] It should be understood that the step numbers used in the text are for ease of description only and are not intended to limit the order in which the steps are performed.

[0048] First Embodiment

[0049] Currently, neural networks are commonly used to sample and detect voltage data in power distribution networks to obtain voltage anomaly diagnosis results. However, these existing technologies relying on neural networks often consume a large amount of computing resources and have slow data processing speeds, requiring a large amount of high-quality data for training and learning. To address these issues, this application's embodiments investigate how to reduce the consumption of computing resources while ensuring the accuracy of voltage anomaly diagnosis.

[0050] like Figure 1 As shown, Figure 1 This application provides a schematic flowchart of a graph model-based distribution network voltage anomaly diagnosis method according to a certain embodiment. The graph model-based distribution network voltage anomaly diagnosis method of this embodiment includes steps S1 to S4, which are detailed below:

[0051] Step S1: Construct a distribution network diagram model based on the geographical topology of the distribution network.

[0052] In this embodiment of the application, a distribution network map model is first constructed based on the geographical topology of the distribution network in the target area. This construction process requires obtaining the geographical topology of the distribution transformers, overhead lines, cables, disconnect switches, and electrical loads within the distribution network.

[0053] To illustrate the power distribution network diagram model, Figure 2 A partial node-branch diagram of the graphical model is provided. As shown in the figure, node 1 is the power supply node, representing a distribution transformer of a higher voltage level; node 2 is an intermediate node, representing a substation transformer of an intermediate voltage level; node 4 is a high-voltage load node, directly powered by the substation transformer of node 2; node 5 is a primary load node, requiring two backup power supply paths. The first path powers the primary load of node 5 by passing through the first low-voltage transformer of node 3 from node 2, and the second path powers the primary load of node 5 by passing through the second low-voltage transformer of node 6 from node 2.

[0054] The distribution network diagram model construction method of this application embodiment is not only applicable to the distribution network topology of the target area, but also to other more complex distribution network topologies.

[0055] Step S2: Detect the voltage of the load nodes in the power distribution network diagram model to obtain a candidate set for voltage anomaly diagnosis.

[0056] In this embodiment of the application, based on the distribution network diagram model, the power supply nodes, intermediate nodes, load nodes and power supply branches of the distribution network can be determined.

[0057] For example, Figure 2The power supply node is node 1, the intermediate nodes are nodes 2, 3, and 6, the load nodes are nodes 4 and 5, and the power supply branch is the path between any two nodes. The power supply branch can represent overhead lines, underground cables, electrical cables, etc.; the load nodes actually correspond to pure load nodes in the distribution network, generator nodes with given active and reactive power outputs, and interconnection nodes where both injected active and reactive power are equal to zero.

[0058] The voltage of load nodes, i.e., the end voltage of the distribution network, is detected by sensors to obtain load node voltage data. Then, based on a preset reference voltage range, the voltage of each load node is compared with the corresponding reference voltage range to identify load nodes that exceed the reference voltage range, which are the load nodes with abnormal voltage in the distribution network diagram model.

[0059] Then, based on the distribution network diagram model, the power supply branch where each load node with voltage anomaly is located is determined, denoted as the first branch. Next, through the load nodes with voltage anomalies, the power supply branches connected to the first branch in the distribution network diagram model are screened to obtain the branches to be detected. Based on the branches to be detected, a candidate set for voltage anomaly diagnosis is obtained.

[0060] For example, in Figure 2 In the process, the voltages of load nodes 4 and 5 were detected and denoted as V4 and V5, respectively, and compared with their respective reference voltage ranges. The specific comparison results are as follows:

[0061] Scenario 1: If the comparison result shows that V4 is abnormal while V5 is normal, then the voltage anomaly diagnosis candidate set in this case is S = {L4}, that is, there is no voltage anomaly in the power supply branch from node 1 to node 5, and there may be a voltage anomaly in the power supply branch from node 1 to node 4.

[0062] Scenario 2: If the comparison result shows that V4 is normal while V5 is abnormal, then the voltage abnormality diagnosis candidate set in this case is S = {L2, L3, L5, L6}. That is, there may be one or more power supply branches from node 1 to node 5 with voltage abnormality, while there is no voltage abnormality in the power supply branches from node 1 to node 4.

[0063] Scenario 3: If the comparison results show that both V4 and V5 are normal, it is considered that there is no voltage abnormality in the distribution network. In this case, the voltage abnormality diagnosis candidate set is empty.

[0064] It should be noted that the reference voltage ranges for voltages V4 and V5 mentioned above are all set as reference voltage hysteresis ranges, rather than single voltage values. When voltage V4 or V5 exceeds its respective reference voltage hysteresis range, voltage V4 or V5 is considered abnormal. When voltage V4 or V5 is within its respective reference voltage hysteresis range, voltage V4 or V5 is considered normal.

[0065] Step S3: The branches to be detected in the voltage anomaly diagnosis candidate set are diagnosed by hierarchical backtracking screening algorithm to obtain the voltage anomaly location result.

[0066] In this embodiment, it is first determined whether the element data in the voltage anomaly diagnosis candidate set is greater than 1. If it is less than 1, the power supply branch where the voltage anomaly occurred can be directly identified, and the voltage anomaly location result can be directly obtained.

[0067] If the value is greater than 1, then the branch to be detected in the voltage anomaly diagnosis candidate set needs to be diagnosed by the hierarchical backtracking filtering algorithm to obtain the voltage anomaly location result.

[0068] exist Figure 2 Taking scenario two in step S2 as an example, since voltage V4 is normal while voltage V5 is abnormal, it can be determined that the direct path between voltage V4 and power node 1 is in normal working condition. However, the branch connecting this direct path to node 5 is abnormal. Therefore, node 2 should be determined as the starting point of the hierarchical backtracking filtering algorithm. Based on node 2, the detection path is selected. The branch with the higher historical failure rate among branches L2 and L3 is selected as the candidate element for priority diagnosis. For example, if the historical failure rate of branch L2 is greater than that of branch L3, then branch L2 is selected for priority voltage anomaly diagnosis. That is, the voltage of node 3 is detected and judged to be normal. If the voltage of node 3 is abnormal, then branch L2 is determined to be the location of voltage anomaly in the distribution network. If the voltage of node 3 is normal, then the diagnosis of the next level of branches continues. Specifically, starting from point 2 and ending at load node 5, the next level of anomaly diagnosis is performed using branch L2 as the path. This involves diagnosing voltage anomalies in branch L5, which is at a lower level than branch L2. If branch L2 has no lower-level branch and no other path leading back to the starting point via the undiagnosed branch, the process directly backtracks to point 2 and selects an undiagnosed branch for diagnosis. If branch L5 is normal, the process continues searching for a lower-level branch of branch L5 or another path leading back to the starting point to diagnose adjacent branches.

[0069] because Figure 2Branch L5 has no next-level branch, but it has a path (L6-L3) that starts from the load node V5 with voltage anomaly and ends at starting point 2, returning to the starting point via undiagnosed branches. Therefore, from the path that can trace back from branch L5 to node 2, the branch adjacent to L5 is selected as the next diagnostic branch, i.e., voltage anomaly diagnosis is performed on branch L6. Here, if there is more than one diagnostic branch among the branches adjacent to L5, the branch with a higher historical failure rate is selected as the priority diagnostic branch for anomaly diagnosis. If branch L6 is diagnosed as normal, the path that traces back to node 2 continues to search for branches with a higher historical failure probability as the next diagnostic branch, because... Figure 2 Only branch L3 was not diagnosed. Therefore, branch L3 was diagnosed for voltage anomaly until the location of the distribution network voltage anomaly was diagnosed, thus confirming the power supply branch with voltage anomaly.

[0070] Step S4: Based on the voltage anomaly location results, formulate a power distribution network maintenance plan.

[0071] In this embodiment of the application, a power distribution network maintenance plan is formulated based on the voltage anomaly location results of the power distribution network, so that the power distribution network anomalies can be resolved in a timely manner.

[0072] Implementing the embodiments of this application has the following beneficial effects:

[0073] This application first constructs a graph model accurately describing the distribution network information based on the geographical topology of the distribution network. Then, based on the graph model, voltage detection is performed on the load nodes of the distribution network. By identifying load nodes with voltage anomalies, potential power supply branches with voltage anomalies are determined and added to a voltage anomaly diagnosis candidate set, providing data support for subsequent branch detection. Then, a hierarchical backtracking filtering algorithm, supplemented by historical branch failure rates, is used to sequentially diagnose voltage anomalies on the potential power supply branches. Accurate voltage anomaly location results can be obtained by detecting only certain power supply branches, eliminating the need for a neural network to detect all power supply branches, saving significant computational resources. Furthermore, no parameter settings are required during the detection process, reducing the risk of detection errors due to parameter setting issues.

[0074] Second Embodiment

[0075] Furthermore, in order to implement the graph model-based distribution network voltage anomaly diagnosis system corresponding to the above method embodiments, and to achieve the corresponding functions and technical effects, Figure 3 A structural diagram of a distribution network voltage anomaly diagnosis system based on a graph model is provided. For ease of explanation, only the parts relevant to this embodiment are shown. The distribution network voltage anomaly diagnosis system based on a graph model provided in this application embodiment includes:

[0076] The graph model building module 201 is used to build a graph model of the distribution network based on the geographical topology of the distribution network.

[0077] In this embodiment of the application, a distribution network map model is first constructed based on the geographical topology of the distribution network in the target area. This construction process requires obtaining the geographical topology of the distribution transformers, overhead lines, cables, disconnect switches, and electrical loads within the distribution network.

[0078] The anomaly diagnosis candidate set construction module 202 is used to detect the voltage of load nodes in the distribution network diagram model and obtain the voltage anomaly diagnosis candidate set.

[0079] In this embodiment, the voltage of the load nodes in the power distribution network diagram model is obtained; based on a preset reference voltage range, the voltage of each load node is compared with the corresponding reference voltage range to determine the load nodes with abnormal voltage in the power distribution network diagram model; the power supply branches of the power distribution network diagram model are screened according to the load nodes with abnormal voltage to obtain a candidate set for voltage anomaly diagnosis.

[0080] The anomaly location result acquisition module 203 is used to diagnose the branches to be detected in the voltage anomaly diagnosis candidate set through a hierarchical backtracking filtering algorithm, and obtain the voltage anomaly location result.

[0081] In this embodiment, a starting point is selected from the non-load nodes of the branch to be tested; starting from the starting point, and taking the load node with abnormal voltage corresponding to the starting point as the endpoint, an abnormality diagnosis is performed on the non-load nodes of all branches to be tested using a hierarchical backtracking screening algorithm to determine the power supply branches with abnormal voltage; and the voltage abnormality location result is determined based on all power supply branches with abnormal voltage.

[0082] The distribution network maintenance module 204 is used to formulate a distribution network maintenance plan based on the voltage anomaly location results.

[0083] In this embodiment of the application, a power distribution network maintenance plan is formulated based on the voltage anomaly location results of the power distribution network, so that the power distribution network anomalies can be resolved in a timely manner.

[0084] In some embodiments, the anomaly diagnosis candidate set construction module 202 specifically comprises:

[0085] In this embodiment of the application, based on the distribution network diagram model, the power supply nodes, intermediate nodes, load nodes, and power supply branches of the distribution network can be determined. The power supply branches can represent overhead lines, underground cables, electrical cables, etc.; the load nodes actually correspond to pure load nodes, generator nodes with given active and reactive power outputs, and tie nodes where both injected active and reactive power are zero in the distribution network.

[0086] The voltage of load nodes, i.e., the end voltage of the distribution network, is detected by sensors to obtain load node voltage data. Then, based on a preset reference voltage range, the voltage of each load node is compared with the corresponding reference voltage range to identify load nodes that exceed the reference voltage range, which are the load nodes with abnormal voltage in the distribution network diagram model.

[0087] Then, based on the distribution network diagram model, the power supply branch where each load node with voltage anomaly is located is determined, denoted as the first branch. Next, through the load nodes with voltage anomalies, the power supply branches connected to the first branch in the distribution network diagram model are screened to obtain the branches to be detected. Based on the branches to be detected, a candidate set for voltage anomaly diagnosis is obtained.

[0088] For example, the voltages of load nodes 4 and 5 are detected and denoted as V4 and V5 respectively, and compared with the reference voltage range of their respective nodes. The specific comparison results are as follows:

[0089] Scenario 1: If the comparison result shows that V4 is abnormal while V5 is normal, then the voltage anomaly diagnosis candidate set in this case is all connected power supply branches from the non-load node to node 4, and these power supply branches may have voltage anomalies.

[0090] Scenario 2: If the comparison result shows that V4 is normal while V5 is abnormal, then the voltage anomaly diagnosis candidate set in this case is all connected power supply branches from the non-load node to node 5, and these power supply branches may have voltage anomalies.

[0091] Scenario 3: If the comparison results show that both V4 and V5 are normal, it is considered that there is no voltage abnormality in the distribution network. In this case, the voltage abnormality diagnosis candidate set is empty.

[0092] It should be noted that the reference voltage ranges for voltages V4 and V5 mentioned above are all set as reference voltage hysteresis ranges, rather than single voltage values. When voltage V4 or V5 exceeds its respective reference voltage hysteresis range, voltage V4 or V5 is considered abnormal. When voltage V4 or V5 is within its respective reference voltage hysteresis range, voltage V4 or V5 is considered normal.

[0093] In some embodiments, the anomaly location result acquisition module 203 specifically comprises:

[0094] In this embodiment, it is first determined whether the element data in the voltage anomaly diagnosis candidate set is greater than 1. If it is less than 1, the power supply branch where the voltage anomaly occurred can be directly identified, and the voltage anomaly location result can be directly obtained.

[0095] If the value is greater than 1, then the branches to be detected in the voltage anomaly diagnosis candidate set need to be diagnosed using a hierarchical backtracking filtering algorithm to obtain the voltage anomaly location result. The specific steps are as follows:

[0096] (1) Select a starting point from the non-load nodes of the branch to be tested; take the starting point as the starting point and the load node with abnormal voltage corresponding to the starting point as the ending point, select the second branch with the highest historical failure rate among the direct-connected branches of the starting point for abnormal diagnosis.

[0097] (2) If an anomaly is diagnosed, the voltage anomaly location result is directly determined; if no anomaly is diagnosed, it is determined whether the second branch has a next-level adjacent branch with the starting point as the starting point, the second branch as the path, and the load node with the voltage anomaly as the end point. If so, the branch with the higher historical failure rate among the next-level adjacent branches is used as the third branch for anomaly diagnosis until the direct branch of the load node with the voltage anomaly is diagnosed; if not, the load node with the voltage anomaly is used as the starting point and the starting point as the end point to determine whether there is an undiagnosed branch.

[0098] (3) If there are no undiagnosed branches, start from the starting point and end at the load node with the abnormal voltage. Select the branch with the highest historical failure rate among the undiagnosed branches as the fourth branch and continue the diagnosis. If there are undiagnosed branches, select the branch with the highest historical failure rate among the undiagnosed branches as the fifth branch and perform abnormal diagnosis.

[0099] (4) If an anomaly is diagnosed, the voltage anomaly location result is directly determined; if no anomaly is diagnosed, it is determined whether the fifth branch has an adjacent branch at the next higher level that starts from the load node with the voltage anomaly, passes through the fifth branch, and ends at the starting point; if so, the branch with the highest historical failure rate among the adjacent branches at the next higher level is used as the sixth branch for anomaly diagnosis, until the direct branch connected to the starting point is diagnosed; if not, the voltage anomaly location result is output.

[0100] Implementing the embodiments of this application has the following beneficial effects:

[0101] This application first constructs a graph model accurately describing the distribution network information based on the geographical topology of the distribution network. Then, voltage detection is performed on the load nodes of the distribution network based on the graph model. By identifying load nodes with voltage anomalies, potential power supply branches with voltage anomalies are determined and added to a voltage anomaly diagnosis candidate set, providing data support for subsequent branch detection. Then, a hierarchical backtracking filtering algorithm is used to sequentially diagnose voltage anomalies on the potential power supply branches. Accurate voltage anomaly location results can be obtained by detecting only certain power supply branches, eliminating the need for a neural network to detect all power supply branches, thus saving significant computational resources. Furthermore, no parameter settings are required during the detection process, reducing the risk of detection errors due to parameter setting issues.

[0102] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of this application. It should be understood that the above descriptions are merely specific embodiments of this application and are not intended to limit the scope of protection of this application. In particular, it should be noted that any modifications, equivalent substitutions, or improvements made by those skilled in the art within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A power distribution network voltage anomaly diagnosis method based on a graph model, characterized in that, The method comprises the following steps: According to the geographical topology of the power distribution network, a power distribution network graph model is constructed; The voltage of the load node in the power distribution network graph model is detected to obtain a voltage abnormality diagnosis candidate set; Through a hierarchical backtracking filtering algorithm, the branches to be detected in the voltage abnormality diagnosis candidate set are diagnosed to obtain a voltage abnormality positioning result, and the specific steps are as follows: a starting point is selected from the non-load nodes of the branch to be detected; taking the starting point as the starting point and the voltage abnormal load node corresponding to the starting point as the terminal point, the second branch with the highest historical failure rate in the directly connected branch of the starting point is selected for abnormality diagnosis; If the second branch is diagnosed as abnormal, the voltage abnormality positioning result is output; otherwise, it is determined whether the second branch has a next level adjacent branch; if it has, the third branch in the next level adjacent branch is diagnosed for abnormality until the directly connected branch of the voltage abnormal load node is diagnosed; if it does not have, it is determined whether there is an undiagnosed branch from the voltage abnormal load node as the starting point to the starting point; If there is no undiagnosed branch, the fourth branch with the highest historical failure rate in the directly connected branch of the starting point is selected as the fourth branch for abnormality diagnosis; if there is an undiagnosed branch, the fifth branch with the highest historical failure rate in the undiagnosed branch is selected as the fifth branch for abnormality diagnosis; If the fifth branch is diagnosed as abnormal, the voltage abnormality positioning result is output; if the fifth branch is not diagnosed as abnormal, it is determined whether the fifth branch has an adjacent branch of the previous level; if it has, the adjacent branch of the previous level is diagnosed for abnormality; if it does not have, the voltage abnormality positioning result is output; According to the voltage abnormality positioning result, a power distribution network maintenance plan is formulated.

2. The power distribution network voltage anomaly diagnosis method based on a graph model according to claim 1, characterized in that, The voltage of the load node in the power distribution network graph model is detected to obtain a voltage abnormality diagnosis candidate set, and the specific steps are as follows: The voltage of the load node in the power distribution network graph model is obtained; Based on the preset reference voltage interval, the voltage of each load node is compared with the corresponding reference voltage interval to determine the voltage abnormal load node in the power distribution network graph model; According to the voltage abnormal load node, the power supply branch of the power distribution network graph model is screened to obtain a voltage abnormality diagnosis candidate set.

3. The power distribution network voltage anomaly diagnosis method based on a graph model according to claim 2, characterized in that, According to the voltage abnormal load node, the power supply branch of the power distribution network graph model is screened to obtain a voltage abnormality diagnosis candidate set, and the specific steps are as follows: Based on the power distribution network graph model, the power supply branch where each voltage abnormal load node is located is determined, which is recorded as the first branch; According to the voltage abnormal load node, the power supply branch of the power distribution network graph model is screened to obtain a voltage abnormality diagnosis candidate set. The method comprises the following steps:

4. A power distribution network voltage anomaly diagnosis system based on a graph model, characterized in that, The graph model construction module, the abnormality diagnosis candidate set construction module, the abnormality positioning result acquisition module and the power distribution network maintenance module are used to construct a power distribution network graph model according to the geographical topology of the power distribution network; ​ ​ The abnormality diagnosis candidate set construction module is configured to detect the voltage of the load node in the power distribution network graph model to obtain a voltage abnormality diagnosis candidate set. The abnormality positioning result acquisition module is configured to diagnose the to-be-detected branch in the voltage abnormality diagnosis candidate set by using a hierarchical backtracking screening algorithm to obtain a voltage abnormality positioning result, and the specific steps are as follows: selecting a starting point from the non-load nodes of the to-be-detected branch; selecting a second branch with the highest historical failure rate in the directly connected branches of the starting point as a starting point for abnormality diagnosis, with the starting point as the starting point and the voltage abnormal load node corresponding to the starting point as the end point; if the second branch is diagnosed as abnormal, outputting the voltage abnormality positioning result; otherwise, determining whether the second branch has a next-level adjacent branch; if yes, diagnosing a third branch in the next-level adjacent branch for abnormality; if no, taking the voltage abnormal load node as the starting point and the starting point as the end point, and determining whether there is an undiagnosed branch; if there is no undiagnosed branch, reselecting the undiagnosed branch with the highest historical failure rate in the directly connected branches of the starting point as a fourth branch for abnormality diagnosis; if there is an undiagnosed branch, selecting the branch with the highest historical failure rate in the undiagnosed branch as a fifth branch for abnormality diagnosis; if the fifth branch is diagnosed as abnormal, outputting the voltage abnormality positioning result; if the fifth branch is not diagnosed as abnormal, determining whether the fifth branch has an adjacent branch at the previous level; if yes, diagnosing the adjacent branch at the previous level for abnormality; if no, outputting the voltage abnormality positioning result; The power distribution network maintenance module is configured to formulate a power distribution network maintenance plan according to the voltage abnormality positioning result.

5. The power distribution network voltage anomaly diagnosis system based on a graph model according to claim 4, characterized in that, The abnormality diagnosis candidate set construction module includes a voltage abnormality detection unit. The voltage abnormality detection unit is configured to obtain the voltage of the load node in the power distribution network graph model; compare the voltage of each load node with the corresponding reference voltage interval based on a preset reference voltage interval, determine the voltage abnormal load node in the power distribution network graph model, and screen the power supply branches of the power distribution network graph model according to the voltage abnormal load node to obtain the voltage abnormality diagnosis candidate set.

6. The power distribution network voltage abnormality diagnosis system based on a graph model according to claim 5, characterized by, The abnormality diagnosis candidate set construction module includes a to-be-detected branch acquisition unit. The to-be-detected branch acquisition unit is configured to determine the power supply branch in which each voltage abnormal load node is located based on the power distribution network graph model, denoted as a first branch; screen the power supply branches in the power distribution network graph model that are connected to the first branch through the voltage abnormal load node to obtain to-be-detected branches; and obtain the voltage abnormality diagnosis candidate set according to the to-be-detected branches.

Images