A Fault Mode Analysis Method for Low-Voltage Distribution Networks Based on Dynamic Topology

Abstract

The present invention belongs to the technical field of low-voltage distribution network fault diagnosis, and specifically relates to a method for analyzing low-voltage distribution network fault patterns based on dynamic topology. The method comprises: obtaining power line information and equipment information, establishing a distribution network data model and a low-voltage distribution network electronic map; obtaining node flow information, matching it according to the distribution network data model to obtain monitoring fault information; presetting a fault diagnosis database and defining the fault type based on the monitoring fault information, and establishing a fault tree model based on the fault diagnosis database to calculate the fault source probability of the fault event; obtaining feedback fault information and fault abnormality information and performing similarity matching to obtain fault root cause data; obtaining the equipment and power lines involved in the monitoring fault information and feedback fault information based on the fault source probability and the fault root cause data, and determining the fault area. The method of the present invention realizes the fault location of the low-voltage distribution network and improves the operation quality of the distribution network.

Description

Technical Field

[0001] The present invention belongs to the technical field of low-voltage distribution network fault diagnosis, and in particular relates to a low-voltage distribution network fault mode analysis method based on dynamic topology. Background Art

[0002] As the demand for electricity grows, so too does the demand for power supply quality and reliability. Due to the complexity of current distribution networks, the geographically dispersed and numerous electricity users hinder timely resolution and response to faults. Traditional fault location methods, due to the complex structure of distribution networks and the low level of automation of electrical equipment, rely solely on trial switches to pinpoint the fault area. This results in repeated activation of numerous switches, exceeding the load limit of the power lines. Summary of the Invention

[0003] To solve the problem existing in the prior art: due to the low degree of automation of electrical equipment in the distribution network, switches repeatedly operate, resulting in the load limit of the power line being exceeded; the present invention provides a low-voltage distribution network fault mode analysis method based on dynamic topology. The purpose of the present invention can be achieved through the following technical solutions:

[0004] Obtain power line information and equipment information, establish distribution network data model and low-voltage distribution network electronic map;

[0005] Obtain node flow information and match it with the distribution network data model to obtain monitoring fault information;

[0006] Presetting a fault diagnosis database and defining fault types based on monitored fault information, and establishing a fault tree model based on the fault diagnosis database to calculate the fault source probability of the fault event;

[0007] Obtain feedback fault information and fault exception information and perform similarity matching to obtain fault root cause data;

[0008] Based on the fault source probability and fault root cause data, the monitoring fault information and the equipment and power lines involved in the feedback fault information are obtained and the fault area is determined.

[0009] Specifically, the distribution network model includes coding, category, geometric attributes, association attributes, and management attributes; wherein the coding is the unique identification code of the device, the category includes power supply, line, switch, and transformer, the geometric attributes are the geographic spatial information of the device, the association attributes are the topological relationship between devices, and the management attributes are numerical parameters that characterize whether the device is in an operating state or a closed state.

[0010] Specifically, the method for determining the fault area is: obtaining a fault information sequence based on the power line, and synchronously saving the fault information sequence in the attribute database of the distribution network data model; obtaining the fault loop information and connection relationship based on the distribution network topology structure corresponding to the attribute data; marking all switching devices in the loop according to the obtained loop information, obtaining the area where the fault device is located according to the switch information, and highlighting the topological elements corresponding to the fault device on the low-voltage distribution network electronic map.

[0011] Specifically, the fault information sequence is generated by installing monitoring devices at line nodes to measure current load data. If the current load is greater than a preset threshold, the monitored fault information at the node is represented by a value of 0; if the current load is less than or equal to the preset threshold, the monitored fault information at the node is represented by a value of 1. The fault information from all nodes is then integrated into a fault information sequence. Fault information includes node failure and node normality, with 0 indicating a node failure and 1 indicating a node normality.

[0012] Specifically, the method for establishing the distribution network data model is as follows: first, the substation, transformer, knife switch, circuit breaker, and pole tower are transformed into point elements, the overhead line and the buried line are transformed into line elements, and the regional information is transformed into surface elements, and then the spatial data of the point elements, the line elements, and the surface elements are vectorized into a distribution network topology structure; the devices are associated with the attribute database and an attribute layer is established, and a distribution network data model is established according to the distribution network topology structure and the attribute layer, wherein the attribute database is a database storing a plurality of fault information sequences, and the fault information sequence is composed of monitoring fault information of a plurality of line nodes.

[0013] Specifically, a fault tree model is established based on the fault diagnosis database to calculate the fault source probability of the fault event, including:

[0014] Defining fault types, including phase-to-phase short circuit fault, low current ground fault, and high resistance fault, and setting the phase-to-phase short circuit fault type as a top-level event, the low current ground fault type as a middle-level event, and the high resistance fault as a bottom-level event;

[0015] Establishing a fault tree node, wherein the fault tree node structure includes a node number, a node name, a node type, a number of nodes, a parent node, a child node, a weight, and a description value, and is saved as a linked list and stored in the fault diagnosis database;

[0016] The linked list is traversed to calculate the minimum cut set of the fault tree, and the fault source probability of the corresponding top-level event fault type is calculated based on the minimum cut set. The calculation formula is:

[0017]

[0018] Where Q is the failure probability, T is the top event failure type, r is the number of minimum cut sets, i and j are the minimum cut set counts, p is the probability density function, k i is the number of middle-level events in the minimum cut set, k j is the number of underlying events of the minimum cut set.

[0019] Specifically, the similarity matching calculation method is: extracting the semantic features of the feedback fault information and the fault abnormality information to obtain a semantic feature vector and an abnormality vector, mapping the semantic feature vector and the abnormality vector to a high-dimensional space, and calculating the similarity matrix between the data points in the high-dimensional space. The calculation formula is:

[0020]

[0021] Among them, p j|i is the probability distribution of the proportion of abnormal vectors under the semantic representation vector, σ i is the standard deviation of the Gaussian distribution, x i is the high-dimensional representation of the semantic feature vector, x j is the high-dimensional representation of the anomaly vector, k is the traversal count of high-dimensional space data points, n is the total number of high-dimensional data points of the semantic feature vector, and m is the total number of high-dimensional data points of the anomaly vector.

[0022] The beneficial effects of the present invention are:

[0023] (1) By using a function model to reflect the topological structure information of the distribution network, and based on real-time tracking of the dynamic topological network, the network topology is divided into several areas for fault analysis, and the accurate location information of the distribution network fault is calculated, thereby improving the fault location accuracy;

[0024] (2) By collecting feedback fault information and fault abnormality information, the relationship and dependency between abnormal information are structured, the fault categories are divided and labeled using clustering algorithms, and the fault distribution and occurrence probability are determined by establishing a fault tree model. The fault location information is graphically displayed, and the faulty equipment is located and detected more intuitively, thereby improving the reliability and management level of the distribution system operation and simplifying the cost and workload of fault location and maintenance work. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] To facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0026] Figure 1 The present invention is a flowchart of a method for analyzing low-voltage distribution network failure modes based on dynamic topology. DETAILED DESCRIPTION

[0027] In order to further illustrate the technical means and effects adopted by the present invention to achieve the predetermined purpose of the invention, the specific implementation methods, structures, features and effects of the present invention are described in detail below in conjunction with the accompanying drawings and preferred embodiments.

[0028] See also Figure 1 ,A low-voltage distribution network fault mode analysis method based on dynamic topology,

[0029] S1: Obtain power line information and equipment information, establish distribution network data model and low-voltage distribution network electronic map;

[0030] In this embodiment, the distribution network model is established based on the power line information and the equipment information;

[0031] The electronic map of the low-voltage distribution network is obtained by topologically converting the distribution network model to obtain a topological dataset model, obtaining attribute data and spatial data of line equipment and substations, integrating the attribute data and spatial data into the topological dataset model to obtain a distribution network data model, and establishing a low-voltage distribution network electronic map based on the distribution network data model;

[0032] S2: Obtain node flow information and match it with the distribution network data model to obtain monitoring fault information;

[0033] The monitoring fault information acquisition method of this embodiment comprises: parsing the node flow information to obtain node space data and current load data, generating a fault information sequence based on the current load data, and matching the node space data with the distribution network data model based on the fault information sequence to obtain monitoring fault information;

[0034] S3: Preset a fault diagnosis database and define the fault type according to the monitored fault information, and establish a fault tree model based on the fault diagnosis database to calculate the fault source probability of the fault event;

[0035] Specifically, the fault source probability of a fault event is obtained as follows: a fault tree model is established based on the fault diagnosis database, the weight of the fault information in the fault tree model is obtained and a fault logic view is generated, and the fault source probability of the fault event is calculated according to the fault logic view;

[0036] S4: Obtain feedback fault information and fault abnormality information and perform similarity matching to obtain fault root cause data;

[0037] In this embodiment, the method for acquiring the root cause data of the fault includes: extracting the semantic features of the feedback fault information and the fault abnormality information to obtain a semantic feature vector and an abnormality vector, performing similarity matching on the semantic feature vector and the abnormality vector to obtain a similarity matrix, performing data dimensionality reduction on the similarity matrix to obtain a comparison value, judging whether the comparison value is greater than a preset threshold value, if so, performing abnormal matching with the fault abnormality template vector of the fault diagnosis database to obtain the root cause data of the fault, if not, performing entity recognition matching to extract component entities and semantic entities to form a semantic feature representation, calculating the similarity between the entities of the semantic feature representation to obtain an abnormality log sequence, and selecting the root cause data corresponding to the node with the highest similarity in the abnormality log sequence as the matching result;

[0038] S5: Obtain monitoring fault information and feedback the equipment and power lines involved in the fault information based on the fault source probability and fault root cause data, and determine the fault area;

[0039] Furthermore, the equipment and power lines involved in the monitoring fault information and the feedback fault information are obtained based on the fault source probability and the fault root cause data, and the fault area is determined by determining the fault area based on the equipment and the power lines, and is highlighted on the low-voltage distribution network electronic map by updating the distribution network data model.

[0040] By dividing the network topology into several areas for fault analysis, accurate location information of distribution network faults is calculated to improve fault location accuracy and overcome the problem of repeated switch action and exceeding the load limit of power lines due to the low degree of automation of distribution network electrical equipment.

[0041] In another embodiment, the distribution network model includes a code, a category, a geometric attribute, an association attribute, and a management attribute; wherein the code is a unique identification code of the device, the category includes power supply, line, switch, and transformer, the geometric attribute is the geographic spatial information of the device, the association attribute is the topological relationship between the devices, and the management attribute is a numerical parameter that characterizes whether the device is in an operating state or a closed state.

[0042] Specifically, the method for establishing the distribution network data model is as follows: transforming substations, transformers, switches, circuit breakers, and pole towers into point elements, transforming overhead lines and underground lines into line elements, and transforming regional information into surface elements, and vectorizing the spatial data of the point elements, the line elements, and the surface elements into a distribution network topology structure; associating the devices with an established attribute database and establishing an attribute layer, and establishing a distribution network data model according to the distribution network topology structure and the attribute layer, wherein the attribute database is a database storing a plurality of fault information sequences, and the fault information sequence is composed of monitoring fault information of a plurality of line nodes.

[0043] In this example, GIS-based data analysis of the low-voltage distribution network analyzes the connections between devices and between devices and lines. By correlating the spatial and attribute data of power devices in the distribution network GIS model, device model connectivity analysis is performed. Using a geographic electronic map as a backdrop, the data within the grid geometry is intuitively displayed and analyzed, enabling bidirectional query analysis from graph to attribute and vice versa.

[0044] Specifically, the method for generating the fault information sequence is as follows: installing a monitoring device at a line node to measure current load data; if the current load is greater than a preset threshold, using 0 to represent the monitoring fault information at the node; if the current load is less than or equal to the preset threshold, using 1 to represent the monitoring fault information at the node; integrating the fault information of all nodes into a fault information sequence; wherein the fault information includes node failure and node normality, 0 indicates node failure, and 1 indicates node normality.

[0045] Specifically, a fault tree model is established based on the fault diagnosis database to calculate the fault source probability of the fault event, including:

[0046] Defining fault types, including phase-to-phase short circuit fault, low-current ground fault, and high-resistance fault, and setting the phase-to-phase short circuit fault type as a top-level event, the low-current ground fault type as a middle-level event, and the high-resistance fault as a bottom-level event;

[0047] Establishing a fault tree node, wherein the fault tree node structure includes a node number, a node name, a node type, a number of nodes, a parent node, a child node, a weight, and a description value, and is saved as a linked list and stored in the fault diagnosis database;

[0048] The linked list is traversed to calculate the minimum cut set of the fault tree, and the fault source probability of the corresponding top-level event fault type is calculated based on the minimum cut set. The calculation formula is:

[0049]

[0050] Where Q is the failure probability, T is the top event failure type, r is the number of minimum cut sets, i and j are the minimum cut set counts, p is the probability density function, k i is the number of middle-level events in the minimum cut set, k j is the number of underlying events of the minimum cut set.

[0051] In another embodiment, a deductive modeling approach is employed. The established model is stored in a linked list structure, and the cut sets within the fault tree are analyzed using a descending method. The principle is to increase the capacity of the cut set when an AND gate is encountered during analysis. Starting with the top-level event, the analysis proceeds layer by layer, replacing the upper-level event with the lower-level event. When an AND gate is encountered, the events are output side by side. When an OR gate is encountered, the input events are output vertically. This process continues until all bottom-level events are replaced. All cut sets are obtained, and duplicate cut sets are removed to ultimately obtain the minimum cut set.

[0052] Specifically, the similarity matching calculation method is: extracting the semantic features of the feedback fault information and the fault abnormality information to obtain a semantic feature vector and an abnormality vector, mapping the semantic feature vector and the abnormality vector to a high-dimensional space, and calculating the similarity matrix between the data points in the high-dimensional space. The calculation formula is:

[0053]

[0054] Among them, p j|i is the probability distribution of the proportion of abnormal vectors under the semantic representation vector, σ i is the standard deviation of the Gaussian distribution, x i is the high-dimensional representation of the semantic feature vector, x j is the high-dimensional representation of the anomaly vector, k is the traversal count of high-dimensional space data points, n is the total number of high-dimensional data points of the semantic feature vector, and m is the total number of high-dimensional data points of the anomaly vector.

[0055] In another preferred embodiment, a coarse-grained text similarity matching module is used to perform root cause diagnosis of fault information template sequences with large text differences. When the text similarity matching module cannot meet the root cause diagnosis accuracy required by the model, the TF-IDF algorithm is used to extract keywords in the template text, and the importance of each word is quantified by considering the frequency of each word in the log template and the frequency of occurrence in the entire log template sequence. Combined with the component names contained in the current information, the system log knowledge graph (SLKG) is entity matched to perform fine-grained system fault root cause diagnosis. For the threshold setting of the text similarity matching module, the cross-validation method is used to determine the threshold of the text similarity matching module and the fault root cause diagnosis accuracy is used as the evaluation index.

[0056] Specifically, the method for determining the fault area is: obtaining a fault information sequence based on the power line, and synchronously saving the fault information sequence in the attribute database of the distribution network data model; obtaining fault loop information and connection relationship based on the distribution network topology structure corresponding to the attribute data; marking all switching devices in the loop according to the obtained loop information, obtaining the fault device according to the switch information, and highlighting the topological elements corresponding to the fault device on the low-voltage distribution network electronic map.

[0057] The computer storage medium of the embodiment of the present invention can adopt any combination of one or more computer-readable media. Computer-readable media can be computer-readable signal media or computer-readable storage media. Computer-readable storage media can be, for example, but not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or components, or any combination thereof. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by an instruction execution system, device or device or used in combination with it.

[0058] A computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0059] The program code that comprises on the computer-readable medium can be transmitted with any appropriate medium, including but not limited to wireless, electric wire, optical cable, RF etc., or above-mentioned any suitable combination.Can write the computer program code that is used to carry out the operation of the present invention with one or more programming languages or its combination, described programming language comprises object-oriented programming language--such as Java, Smalltalk, C++, also comprises conventional procedural programming language--such as " C " language or similar programming language.Program code can be executed on user's computer completely, partly on user's computer, execute as an independent software package, partly on user's computer and partly on remote computer, or execute completely on remote computer or server.In the situation that relates to remote computer, remote computer can be connected to user's computer by the network of any kind, including local area network (LAN) or wide area network (WAN), or be connected to external computer (for example utilizing Internet service provider to connect by Internet).

[0060] The above description is merely a preferred embodiment of the present invention and does not constitute any form of limitation to the present invention. Although the present invention has been disclosed as a preferred embodiment as above, it is not intended to limit the present invention. Any person skilled in the art can make some changes or modifications to equivalent embodiments using the technical contents disclosed above without departing from the scope of the technical solution of the present invention. However, any simple modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A method for analyzing low-voltage distribution network failure modes based on dynamic topology, characterized in that: include: Obtain power line information and equipment information and establish a distribution network data model; The method for establishing a distribution network data model comprises: first, transforming substations, transformers, switches, circuit breakers, and towers into point elements, transforming overhead lines and underground lines into line elements, and transforming regional information into surface elements; then, vectorizing the spatial data of the point elements, line elements, and surface elements into a distribution network topology structure; associating the devices through an attribute database and establishing an attribute layer; and establishing a distribution network data model based on the distribution network topology structure and the attribute layer, wherein the attribute database is a database storing a plurality of fault information sequences, and the fault information sequence is composed of monitoring fault information of a plurality of line nodes; Obtain node flow information and match it with the distribution network data model to obtain monitoring fault information; Presetting a fault diagnosis database and defining fault types based on monitored fault information, and establishing a fault tree model based on the fault diagnosis database to calculate the fault source probability of the fault event; Obtain feedback fault information and fault exception information and perform similarity matching to obtain fault root cause data; Based on the fault source probability and fault root cause data, the monitoring fault information and the equipment and power lines involved in the feedback fault information are obtained and the fault area is determined.

2. A method for analyzing low-voltage distribution network failure modes based on dynamic topology according to claim 1, characterized in that: The distribution network model includes a code, a category, geometric attributes, associated attributes, and management attributes; wherein the code is a unique identification code for the device, the category includes power supply, line, switch, and transformer, the geometric attribute is the geographic spatial information of the device, the associated attribute is the topological relationship between the devices, and the management attribute is a numerical parameter that characterizes whether the device is in an operating state or a closed state.

3. The method for analyzing low-voltage distribution network failure modes based on dynamic topology according to claim 1, characterized in that: The method for determining the fault area is: A fault information sequence is obtained based on the power line, and the fault information sequence is synchronously saved in an attribute database of a distribution network data model; fault loop information and connection relationships are obtained based on a distribution network topology structure corresponding to the attribute data; all switch devices in the loop are marked based on the obtained loop information, and a fault area where the fault device is located is obtained based on the switch information.

4. The method for analyzing low-voltage distribution network failure modes based on dynamic topology according to claim 3, characterized in that: The method for generating the fault information sequence is as follows: installing a monitoring device at a line node to measure current load data; if the current load is greater than a preset threshold, using 0 to represent the monitoring fault information at the node; if the current load is less than or equal to the preset threshold, using 1 to represent the monitoring fault information at the node; The fault information of all nodes is integrated into a fault information sequence.

5. The method for analyzing low-voltage distribution network failure modes based on dynamic topology according to claim 1, characterized in that: Establishing a fault tree model based on the fault diagnosis database to calculate the fault source probability of the fault event includes: Defining fault types, including phase-to-phase short circuit fault, low current ground fault, and high resistance fault, and setting the phase-to-phase short circuit fault type as a top-level event, the low current ground fault type as a middle-level event, and the high resistance fault as a bottom-level event; Establishing a fault tree node, wherein the fault tree node structure includes a node number, a node name, a node type, a number of nodes, a parent node, a child node, a weight, and a description value, and is saved as a linked list and stored in the fault diagnosis database; The linked list is traversed to calculate the minimum cut set of the fault tree, and the fault source probability of the corresponding top-level event fault type is calculated according to the minimum cut set.

6. The method for analyzing low-voltage distribution network failure modes based on dynamic topology according to claim 1, characterized in that: The similarity matching calculation method is as follows: extracting semantic features of the feedback fault information and fault anomaly information to obtain semantic feature vectors and anomaly vectors, mapping the semantic feature vectors and anomaly vectors to a high-dimensional space, and calculating the similarity matrix between data points in the high-dimensional space. The calculation formula is: ， in, is the probability distribution of the proportion of abnormal vectors under the semantic representation vector, is the standard deviation of the Gaussian distribution, is the high-dimensional representation of the semantic feature vector, is the high-dimensional representation of the abnormal vector, k Count the number of data points traversed in high-dimensional space, n is the total number of high-dimensional data points of semantic feature vectors, m is the total number of high-dimensional data points of abnormal vectors.

Images