Power distribution network line fault positioning method, device and equipment

Abstract

The application discloses a power distribution network line fault positioning method, device and equipment, relates to the technical field of power distribution network line fault positioning, and comprises the following steps: acquiring operation data and lightning ground flash data; performing feature extraction on the operation data, performing space-time feature mapping on the lightning ground flash data, and generating multi-source feature parameters; performing probability reasoning on the input multi-source feature parameters through a distributedly arranged section intelligent agent and a Bayesian network submodule thereof, and outputting real-time fault probability values of each independent section of the power distribution network; and taking the real-time fault probability values of each independent section as input, dynamically coordinating output results of each intelligent agent through a Nash equilibrium strategy, and generating a globally consistent fault positioning decision. The application is used to solve the problems that a traditional line fault positioning method is difficult to effectively fuse heterogeneous data and difficult to coordinate local judgment conflicts.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power distribution network line fault location, more particularly, the present application relates to a power distribution network line fault location method, device and equipment. BACKGROUND

[0002] The power distribution network is a key link for the power system to supply power to the end user, and its operation reliability directly affects the power supply quality and the normal development of social and economic activities. With the advancement of new power system construction, the structure of the power distribution network is becoming increasingly complex, large-scale distributed power supply access, diversified load types and frequent extreme weather events, making line fault location face new technical challenges. The traditional fault location method mainly relies on electrical quantity measurement data and protection device action information, which has obvious limitations in practical application: it is difficult to effectively integrate multi-source heterogeneous data and lack of coordination mechanism for distributed decision conflict.

[0003] For example, the invention patent CN118797527A discloses a DC power distribution network line fault diagnosis method, relating to the technical field of DC power distribution network, including the following steps: obtaining mycological model abnormal data before locating line fault, obtaining network information fusion data when diagnosing line fault type; comparing the mycological model abnormal data, comparing the network information fusion data to obtain the first comparison result; combining the first comparison result with the constraint conditions of each diagnosis effect category to analyze and obtain the actual diagnosis effect category; optimizing the diagnosis scheme according to the actual diagnosis effect category; establishing an optimization effect evaluation model through machine learning according to the optimized mycological model abnormal data and network information fusion data to generate an optimization effect evaluation coefficient; analyzing to obtain the power grid line fault diagnosis effect; predicting the effect of the DC power distribution network fault diagnosis method to improve the accuracy and stability of the DC power distribution network line fault diagnosis.

[0004] For example, the invention patent CN120354254A discloses a kind of active power distribution network fault location and identification method and system based on space-time graph network, relating to active power distribution network fault location and identification technical field, for: obtaining each node data of active power distribution network after fault occurs, build structured graph data, and mark fault node and type, construct sample data set;With sample data set as source domain data, pre-train fault diagnosis model based on space-time graph network;In each iteration process, generate adversarial sample data using multi-scale adversarial perturbation addition method based on gradient optimization, through data class dynamic balance mechanism and confidence evaluation mechanism, screen data and add to data set, train model using updated data set;Introduce target domain data, adopt transfer learning strategy based on dynamic kernel maximum mean difference, fine-tune pre-training model;The node data actually acquired is input into the model to realize the accurate positioning of fault node and the accurate identification of fault type.

[0005] The existing scheme mainly relies on network information fusion data, and the multi-source data fusion dimension is insufficient, it is difficult to cope with the failure caused by lightning and other natural disasters, and lacks the design of power distribution network section intelligent agent collaborative decision-making, and the centralized diagnosis model cannot adapt to the distributed power supply access scene of large-scale power distribution network; although some schemes use space-time graph network to process node data, the heterogeneous data collaboration ability is weak, the weighted fusion design of operating steady-state characteristics and lightning space-time characteristics is lacking, and the centralized transfer learning model is used for fault positioning, which cannot solve the local judgment conflict in the ring network or multi-power supply scene. In summary, the existing technical solutions either because the multi-source heterogeneous data fusion is insufficient, or because the power distribution network section collaborative decision-making design is lacking, neither of which can effectively solve the problems of fusion of heterogeneous data and local judgment conflict.

[0006] In view of the above problems, the present application provides a solution. SUMMARY

[0007] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present application provide a power distribution network line fault positioning method, device and equipment, which extracts multi-source feature parameters, establishes a multi-agent Bayesian sub-network, and fuses the Bayesian sub-network through Nash equilibrium strategy, thereby solving the problems that the traditional line fault positioning method is difficult to effectively fuse heterogeneous data and difficult to coordinate local judgment conflict.

[0008] To achieve the above object, the present application provides the following technical scheme: a power distribution network line fault positioning method, characterized in that it comprises the following steps: obtaining operating data and lightning ground flash data; performing feature extraction on the operating data and time-space feature mapping on the lightning ground flash data to generate multi-source feature parameters; performing probability reasoning on the input multi-source feature parameters through the distributed deployment of section agents and their Bayesian network sub-modules, and outputting real-time fault probability values of each independent section of the power distribution network; taking the real-time fault probability values of each independent section as input, dynamically coordinating the output results of each agent through Nash equilibrium strategy, and generating a globally consistent fault positioning decision.

[0009] In a preferred embodiment, the operating data includes binary encoding values of isolated section states, action time sequence characteristic values and feeder effective values, the ground flash data includes spatial correlation coefficients of line sections and ground flash fault time difference coupling degrees, and the multi-source feature parameters include operating steady-state feature parameters and lightning space-time feature parameters.

[0010] In a preferred embodiment, the operating data is feature extracted, and the lightning ground flash data is spatio-temporal feature mapped to generate multi-source feature parameters, specifically: the operating data is standardized, and operating steady-state feature parameters are generated by feature vector splicing; a power distribution network topology structure is constructed, and lightning spatio-temporal feature parameters are generated by spatial fusion of the ground flash data and the topology structure; and the operating steady-state feature parameters and the lightning spatio-temporal feature parameters are weighted and fused to output the multi-source feature parameters.

[0011] In a preferred embodiment, the multi-source feature parameters are input to the distributed segment agents and their Bayesian network sub-modules for probability inference, and real-time fault probability values of each independent segment of the power distribution network are output, specifically: the lines are divided into a plurality of physical segments according to the power distribution network topology structure, and each physical segment is associated with an agent, and a Bayesian network sub-module is constructed for each agent; the multi-source feature parameters are input to the corresponding agent as observation data of the Bayesian network sub-module; a conditional probability table of the Bayesian network is set based on historical data and expert experience; the posterior probability of each segment fault is calculated by a Bayesian probability propagation algorithm according to the observation data and the conditional probability table; and the real-time fault probability value of each segment is determined by a posterior probability estimation algorithm based on the posterior probability.

[0012] In a preferred embodiment, the power distribution network topology structure is constructed, specifically: line model data and power grid spatial coordinate information are obtained; node distribution and device connection relationship are extracted based on the line model data and the power grid spatial coordinate information; a topology verification algorithm is used to perform connectivity verification on the node distribution and the device connection relationship to generate verified topology information; and the verified topology information is integrated to construct a power distribution network topology structure containing physical connection relationship.

[0013] In a preferred embodiment, the real-time fault probability values of each independent segment are input to dynamically coordinate the output results of each agent by a Nash equilibrium strategy to generate a globally consistent fault location decision, specifically: the Bayesian network sub-module of each agent is defined as a game participant, and the strategy space parameter of each agent is the real-time fault probability value of the corresponding segment; a benefit function is constructed for each agent, and the function value is a weighted combination of the fault location accuracy index and the adjacent segment fault correlation degree index; the strategy space parameter of each agent is dynamically adjusted by an iterative optimization algorithm until the benefit function of all agents reaches a preset threshold and the strategy space parameter combination meets the Nash equilibrium strategy; and the strategy space parameter meeting the Nash equilibrium strategy is output as the globally consistent fault location decision.

[0014] In a preferred embodiment, the strategy space parameters of each agent are dynamically adjusted by an iterative optimization algorithm, specifically: taking the strategy space parameters of each agent as optimization variables, calculating the updated values of the strategy space parameters of each agent in the current iteration round based on a preset profit function expression; taking the updated values of the strategy space parameters as the initial values of the next iteration, and repeatedly performing the strategy space parameter optimization calculation for all agents; continuously iterating until the convergence condition is met, the convergence condition including that the change amount of the strategy space parameters of all agents in the adjacent two iterations is less than a preset convergence threshold, and the profit function value of all agents reaches a preset profit function threshold state.

[0015] In a preferred embodiment, the line is divided into several physical sections according to the power distribution network topology, and each physical section is associated with an agent, specifically: according to the power distribution network topology, the line is divided into several physical sections with the protection device as the boundary; an agent is created for each physical section, and the agent is constructed based on the collected adjacent section connection relationship, the type parameters of the section power equipment, the initial fault probability distribution generated by the historical fault statistics, and the fault propagation weight coefficient generated based on the topological distance.

[0016] A power distribution network line fault location device, characterized by comprising a data acquisition module, a feature extraction and mapping module, a probability reasoning module, and a global decision module, and there is a connection between the modules: the data acquisition module acquires operation data and lightning ground flash data; the feature extraction and mapping module extracts features from the operation data and maps the time and space features of the lightning ground flash data to generate multi-source feature parameters; the probability reasoning module performs probability reasoning on the input multi-source feature parameters through the distributed section agents and their Bayesian network submodules, and outputs the real-time fault probability values of each independent section of the power distribution network; the global decision module takes the real-time fault probability values of each independent section as input, dynamically coordinates the output results of each agent through the Nash equilibrium strategy, and generates a globally consistent fault location decision.

[0017] A power distribution network line fault location device, characterized by comprising at least one processor; and an input / output interface in communication connection with the at least one processor; a memory in communication connection with the at least one processor; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to execute the power distribution network line fault location method.

[0018] The technical effects and advantages of the power distribution network line fault location method, device and equipment of the application are as follows:

[0019] The power distribution network line fault positioning method, device and equipment provided by the application successfully break through the barrier between power distribution network operation data and lightning monitoring data through multi-source data fusion technology. The power distribution network operation data contains key contents such as isolated section state and switch action information, and the lightning monitoring data involves important parameters such as ground flash position and lightning intensity. These heterogeneous data are originally difficult to work together. The application extracts the steady-state characteristics of the operation data and maps the time and space characteristics of the ground flash data, effectively integrates the two into unified feature parameters, provides a comprehensive and reliable data basis for subsequent fault positioning, and solves the pain point that heterogeneous data are difficult to be applied cooperatively in the traditional line fault positioning method.

[0020] In the local fault judgment, the multi-agent Bayesian network submodule is used to realize accurate calculation of the fault probability of each section. The power distribution network is divided into multiple sections, each section corresponds to an agent and is configured with a Bayesian network submodule, and the conditional probability table set by combining historical data and expert experience enables each agent to accurately calculate the fault probability of the section according to the input feature parameters, greatly improving the reliability of local fault judgment.

[0021] In the global decision, the Nash equilibrium strategy is used to coordinate each agent, effectively overcoming the problem of local judgment conflict. The Bayesian network submodule of each agent as a game participant defines a reasonable benefit function and iteratively calculates the strategy space parameters, and finally reaches a globally consistent fault positioning result. This process ensures that the judgments of each section can be coordinated with each other to form a unified conclusion, avoiding positioning confusion caused by local judgment differences.

[0022] The problems that the traditional line fault positioning method is difficult to effectively fuse heterogeneous data and difficult to coordinate local judgment conflicts are effectively solved. BRIEF DESCRIPTION OF DRAWINGS

[0023] Figure 1 The power distribution network line fault positioning method flowchart provided for the embodiment of the application.

[0024] Figure 2 The power distribution network line fault positioning device structure diagram provided for the embodiment of the application.

[0025] Figure 3 The power distribution network line fault positioning device structure diagram provided for the embodiment of the application. DETAILED DESCRIPTION

[0026] With reference to the accompanying drawings, the technical solutions in the embodiments of the present application will be described clearly and completely. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all the other embodiments obtained by a person of ordinary skill in the art without creative work are within the protection scope of the present application.

[0027] It should be noted that, in this document, the terms such as first and second are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Also, the terms "comprises", "comprising", or any other variations thereof are intended to cover non-exclusive inclusions, so that a process, method, article, or apparatus including a series of elements includes not only those elements, but also other elements not explicitly listed, or other elements inherent to such a process, method, article, or apparatus. Without more limitations, the elements defined by the statement "comprise" do not exclude the presence of other identical elements in the process, method, article, or apparatus including the elements.

[0028] Embodiment 1, Figure 1 A power distribution network line fault positioning method is given, comprising the following steps:

[0029] S1, obtaining operation data and lightning ground flash data;

[0030] S2, performing feature extraction on the operation data, and performing space-time feature mapping on the lightning ground flash data to generate multi-source feature parameters;

[0031] S3, performing probability reasoning on the input multi-source feature parameters through the distributedly deployed section intelligent agent and its Bayesian network submodule, and outputting real-time fault probability values of each independent section of the power distribution network;

[0032] S4, taking the real-time fault probability values of each independent section as input, dynamically coordinating the output results of each intelligent agent through a Nash equilibrium strategy, and generating a globally consistent fault positioning decision.

[0033] The power distribution network line fault positioning method, device and equipment provided by the application successfully break the barrier between power distribution network operation data and lightning monitoring data through multi-source data fusion technology. The power distribution network operation data contains key contents such as isolated section state and switch action information, and the lightning monitoring data involves important parameters such as ground flash position and lightning intensity. These heterogeneous data are originally difficult to work together. The application extracts steady-state features from operation data and maps time-space features of lightning ground flash data, effectively integrates the two into unified feature parameters, provides a comprehensive and reliable data basis for subsequent fault positioning, and solves the pain point of difficult application of heterogeneous data in traditional line fault positioning methods.

[0034] In the local fault judgment, the multi-agent Bayesian network submodule is used to realize accurate calculation of the fault probability of each section. The power distribution network is divided into multiple sections, each section corresponds to an agent and is configured with a Bayesian network submodule, and the conditional probability table set by combining historical data and expert experience enables each agent to accurately calculate the fault probability of the section according to the input feature parameters, greatly improving the reliability of local fault judgment.

[0035] In the global decision, Nash equilibrium strategy is used to coordinate each agent, effectively overcoming the problem of local judgment conflict. The Bayesian network submodule of each agent as a game participant defines a reasonable benefit function and iteratively calculates the strategy space parameters, and finally reaches a globally consistent fault positioning result. This process ensures that the judgments of each section can be coordinated with each other to form a unified conclusion, avoiding positioning confusion caused by local judgment differences.

[0036] The problems of difficult effective fusion of heterogeneous data and difficult coordination of local judgment conflict in traditional line fault positioning methods are effectively solved.

[0037] S1, obtaining operation data and lightning ground flash data.

[0038] In this embodiment, the operation data includes binary coded values of isolated section states, action time sequence feature values and feeder effective values, the ground flash data includes spatial correlation coefficients of line sections and ground flash fault time difference coupling degrees, and the multi-source feature parameters include operation steady-state feature parameters and lightning time-space feature parameters.

[0039] S2, feature extraction is performed on the operation data, and time-space feature mapping is performed on the lightning ground flash data to generate multi-source feature parameters.

[0040] In this embodiment, the feature extraction is performed on the operation data, and the time-space feature mapping is performed on the lightning ground flash data to generate multi-source feature parameters, specifically:

[0041] The operation data is standardized, and the operation steady-state characteristic parameters are generated by splicing feature vectors;

[0042] The power distribution network topology structure is constructed, the spatial fusion is performed according to the ground flash data and the topology structure, and the lightning space-time characteristic parameters are generated;

[0043] The operation steady-state characteristic parameters and the lightning space-time characteristic parameters are weighted and fused, and the multi-source characteristic parameters are output.

[0044] The operation steady-state characteristic parameters of the operation data are obtained according to the isolation section state, the switch action information and the feeder short-circuit current, the isolation section state is obtained by binary coding to obtain a binary coding value, the switch action information is obtained by time sequence coding to obtain an action time sequence characteristic value, the feeder short-circuit current is obtained by effective value extraction to obtain a feeder effective value, and the binary coding value, the action time sequence characteristic value and the feeder effective value are standardized and spliced with feature vectors to obtain the operation steady-state characteristic parameters of the operation data;

[0045] The power distribution network topology is acquired, and the specific steps for acquiring the lightning space-time characteristic parameters are as follows: first, digital modeling of the power distribution network topology is carried out. The line basic data in the power distribution network geographic information system is collected, covering the specific coordinates of the towers, the type of the line adopted, the position information of each branch node and the connection relationship between the nodes. The line topology matrix is constructed according to this, the connection state between different line sections is identified by the matrix elements, and the line space distribution vector composed of the spatial coordinate set of each line section is generated. Second, the spatial correlation coefficient is calculated. The position coordinates, occurrence time and lightning current intensity of ground flashes are obtained from the ground flash monitoring system. For each line section, the vertical distance from the ground flash position to the line section is calculated. The spatial attenuation coefficient is set, and the spatial correlation coefficient is calculated by the exponential function based on the vertical distance and the attenuation coefficient. The larger the coefficient value is, the closer the spatial correlation between the ground flash and the line section is. Third, the lightning intensity tolerance value calculation is completed. The lightning withstand level parameters of each line section are extracted, which comprehensively considers the tower grounding resistance, insulator withstand voltage and other factors. The ratio of the lightning current intensity to the lightning withstand level of the line section is calculated, and the ratio is mapped to the lightning intensity tolerance value by using the Sigmoid function. When the ratio is greater than or equal to 1, the tolerance value tends to 1, indicating that the lightning intensity may exceed the line withstand capacity. Fourth, the time coupling degree calculation is implemented. The specific time when the line fails is determined, and the time difference between the occurrence time of the ground flash and the failure time of the line is calculated. The time window threshold is set, which is determined with reference to the line protection action time. When the time difference is within the threshold range, the time coupling degree decreases linearly with the increase of the time difference; when the time difference exceeds the threshold range, the time coupling degree is 0. Fifth, the lightning space-time characteristic parameters are obtained by weighted fusion. The weight coefficients of the spatial correlation coefficient, the lightning intensity tolerance value and the time coupling degree are determined by using the analytic hierarchy process. The preliminary characteristic values of each line section are obtained by weighted sum according to the weights of the three parameters. The preliminary characteristic values of all line sections are normalized to finally form the lightning space-time characteristic vector.

[0046] The running steady-state characteristic parameters and the lightning space-time characteristic parameters are weighted and fused to obtain the multi-source characteristic parameters.

[0047] The running steady-state characteristic parameter calculation formula is:

[0048]

[0049] wherein, represents the running steady-state characteristic parameter of the i-th line section, , , , is the running steady-state weight coefficient, is the binary encoding value of the isolation section state, is the action time sequence characteristic value of the switch action information after normalization, the effective value of the feeder of the mth line section.

[0050] The calculation formula of the lightning space-time characteristic parameter is:

[0051]

[0052] wherein, denotes the lightning space-time characteristic parameter of the mth line section, , , is a lightning space-time weight coefficient, is a space correlation coefficient of the mth line section, is a lightning space-time lightning stroke intensity tolerance value, is a ground flash failure time difference coupling degree. The calculation formula of the multi-source characteristic parameter is:

[0053]

[0054]

[0055] wherein, denotes a multi-source characteristic parameter set of the mth line section, is a characteristic parameter fusion weight coefficient. S3, by means of the distributed deployment of the section agent and its Bayesian network submodule, carries out probability inference on the input multi-source characteristic parameters, and outputs the real-time failure probability value of each independent section of the distribution network.

[0056] In the embodiment, the probability inference on the input multi-source characteristic parameters by means of the distributed deployment of the section agent and its Bayesian network submodule, and the output of the real-time failure probability value of each independent section of the distribution network are specifically:

[0057] According to the topology structure of the distribution network, the line is divided into a plurality of physical sections, each physical section is associated with an agent, and a Bayesian network submodule is constructed for each agent;

[0058] The multi-source characteristic parameters are input into the corresponding agent as the observation data of the Bayesian network submodule;

[0059] The conditional probability table of the Bayesian network is set based on historical data and expert experience;

[0060] According to the observation data and the conditional probability table, the posterior probability of each section failure is calculated by means of the Bayesian probability propagation algorithm;

[0061]

[0062] ​​​​Based on the posterior probability, a real-time fault probability value of each section is determined by a posterior probability estimation algorithm.

[0063] According to the distribution network topology, the lines are divided into several physical sections, each of which is an agent. The distribution network topology contains key information such as the connection mode of the line, the distribution of nodes, the layout of equipment, etc. Based on this, the physical sections are divided, which can ensure that each section has relatively independent electrical characteristics and fault performance. When dividing, the protection device is taken as the clear boundary, because the setting of the protection device is usually related to the importance of the line, the load distribution and other factors, so that the fault detection and isolation of each physical section can be more targeted. After each physical section is given the attribute of the agent, it has the ability of autonomous perception, analysis and decision-making. In order to make the agent more accurately carry out the work, the detailed attribute information of each physical section will be collected, including the connection relationship of adjacent sections (clear association with other sections, which is convenient for information exchange and collaborative judgment), section device type parameters (such as line material, length, tower type, etc., which will affect the fault probability and characteristics of the section), initial fault probability distribution (based on historical fault data statistics, as the initial reference for fault probability calculation) and fault propagation weight coefficient (reflecting the possibility of fault propagation from the current section to the adjacent section), and the corresponding agent is created according to these attributes, so that each agent can accurately represent the characteristics of the section.

[0064] The conditional probability table is the key to the function of the Bayesian network submodule, which reflects the probability dependence relationship between different variables. Based on statistical analysis of historical data (including a large number of past fault cases, corresponding feature parameters and actual fault results), the objective probability law between variables can be summarized; at the same time, combined with expert experience (experts have a deep understanding of the fault mechanism of distribution network and the influence degree of each factor), the statistical results are corrected and improved, so that a reasonable and accurate conditional probability table is set, ensuring that the reasoning of the Bayesian network submodule has high reliability and accuracy.

[0065] The calculation formula of the input evidence is:

[0066]

[0067] Wherein is the input evidence of the th line section, and is the set of input evidence of each Bayesian network submodule.

[0068] The posterior probability calculation formula of each section fault is:

[0069]

[0070] Wherein, is the posterior probability of failure in each section, For section fault status, is the prior probability of section failure (initial distribution), is the conditional probability of evidence appearing under a given fault condition (from the conditional probability table), is the marginal probability of the evidence.

[0071] The calculation formula for the probability of failure in each section is:

[0072]

[0073] in, is the probability of failure in each section, is the posterior probability weight, is the number of devices in the segment, For the section The historical failure probability of the equipment, For the Type weight of class device, is the historical failure impact coefficient.

[0074] In this embodiment, the construction of the distribution network topology is specifically as follows:

[0075] Obtain line model data and grid spatial coordinate information;

[0076] Extract node distribution and equipment connection relationships based on line model data and grid spatial coordinate information;

[0077] A topology verification algorithm is used to perform connectivity verification on the node distribution and device connection relationship to generate verified topology information;

[0078] Integrate the verified topology information and construct the distribution network topology structure including the physical connection relationship.

[0079] First, the collection and integration of basic data are carried out. Through the connection of the power distribution network dispatching system, geographic information system (GIS) and equipment management platform, line model data and power grid spatial coordinate information are synchronously obtained. The line model data covers the physical parameters of the line (such as conductor type, cross-sectional area, impedance value), equipment attributes (such as transformer capacity, switch type, transformer ratio) and hierarchical relationship (such as the subordinate relationship of main line and branch line); the power grid spatial coordinate information includes the latitude and longitude coordinates, altitude and relative position relationship (such as the connection point coordinates of the tower and the line) of each device. Based on the above data, the node distribution information of the distribution network (including transformer bus nodes, switch station nodes, branch box nodes and user access nodes) is extracted through spatial topology analysis algorithm, and the connection relationship between devices is sorted out according to the physical connection characteristics (such as the connection between cable head and switch and the fixed relationship between conductor and tower), forming an initial node-connection relationship atlas.

[0080] Secondly, the initial information is verified by using a topology verification algorithm. The core verification algorithm includes connectivity verification based on adjacency matrix: the nodes of the distribution network are taken as matrix elements, an adjacency matrix is constructed (the element value in the matrix is 1 indicating that two nodes are directly connected, and 0 indicating that they are not connected), and the matrix is traversed by depth-first search (DFS) or breadth-first search (BFS) to verify whether each node is in a connected state and to check isolated nodes (such as branch boxes not connected to the main network) or false connections (such as lines recorded in the database as connected but actually physically disconnected). In addition, device attribute consistency verification (such as verifying whether the rated current of the switch matches the carrying capacity of the connected line to avoid topology logical errors caused by parameter conflicts) and spatial position rationality verification (such as calculating the line length between two nodes through GIS coordinates, comparing it with the actual line length to verify the authenticity of the connection relationship) are supplemented. Through multi-dimensional verification, the error information is eliminated, the missing association is supplemented, and the verified node distribution and device connection relationship dataset is formed.

[0081] Finally, the verified information is integrated and processed. The spatial coordinates of the nodes, the physical parameters of the devices and the verified connection relationship are associated and mapped to construct a topology model containing three-dimensional information of "node position-device attribute-connection mode". The physical location of each node (such as the specific address of the transformer substation and the coordinates of the tower), the technical parameters of the device (such as the voltage level of the transformer and the operation mode of the switch) and the physical form of the connection (such as overhead line or cable connection and the specific position of the connection point) are clearly marked in the model. Through data standardization processing (such as unified coordinate system and standardized device naming rules), a structured distribution network topology structure data is formed, which provides a basis for subsequent power grid load flow calculation, fault location and network reconstruction and other applications.

[0082] S4, taking the real-time fault probability value of each independent section as input, dynamically coordinating the output results of each agent through Nash equilibrium strategy, generating a global consistent fault location decision.

[0083] In the embodiment, the taking the real-time fault probability value of each independent section as input, dynamically coordinating the output results of each agent through Nash equilibrium strategy, generating a global consistent fault location decision, specifically:

[0084] The Bayesian network submodule of each agent is defined as a game player, and the strategy space parameter of each agent is the real-time fault probability value of the corresponding section.

[0085] An income function is constructed for each agent, and the function value is the weighted combination of the fault location accuracy index and the adjacent section fault correlation index.

[0086] The strategy space parameters of each agent are dynamically adjusted through an iterative optimization algorithm until the income function of all agents reaches a preset threshold and the strategy space parameter combination meets the Nash equilibrium strategy.

[0087] The strategy space parameter meeting the Nash equilibrium strategy is output as the global consistent fault location decision.

[0088] The preset threshold of the income function is a benchmark value considering the fault location accuracy and the adjacent section fault correlation.

[0089] The Bayesian network submodule of each agent is defined as a game player, and the strategy space parameter of each agent is the probability of each section fault output by each agent, that is, the strategy of agent is , , is the total number of sections, represents the probability of agent judging that section has a fault.

[0090] The fault location accuracy calculation formula is:

[0091]

[0092] wherein, is the fault location accuracy, is the actual fault state of section , and represents the probability of agent judging that section has a fault.

[0093] The adjacent section fault correlation calculation formula is:

[0094]

[0095] in, is the fault correlation degree of adjacent sections, For intelligent agents The number of neighboring agents, is the set of neighboring agents, Representing an agent Judgment section The probability of failure.

[0096] The profit function expression is:

[0097]

[0098] in, For intelligent agents The function return, is the weight parameter of the profit function.

[0099] The calculation formula for global consistency fault location results is:

[0100]

[0101] in, is the global consistency fault location result, For intelligent agents The weight of Representing an agent Judgment section The probability of failure.

[0102] In this embodiment, the strategy space parameters of each agent are dynamically adjusted by the iterative optimization algorithm, specifically:

[0103] Taking the policy space parameters of each agent as the optimization variable, the updated value of the policy space parameters of each agent in the current iteration round is calculated based on the preset profit function expression;

[0104] Use the updated policy space parameter values ​​as the initial values ​​for the next iteration, and repeat the policy space parameter optimization calculation for all agents.

[0105] The iteration continues until the convergence conditions are met. The convergence conditions include that the changes in the strategy space parameters of all agents in two adjacent iterations are less than the preset convergence threshold, and the profit function values ​​of all agents reach the preset profit function threshold state.

[0106] In each iteration process, each agent is selected in turn as an iteration agent for policy optimization. For the currently selected iteration agent, the system fixes the policy space parameters of all other agents at the current iteration round, and these fixed parameters will serve as a reference benchmark for the iteration agent to optimize its own policy. At this time, the iteration agent regards its own policy space parameters as optimization variables that need to be adjusted, and its core goal is to change these parameters so that its own revenue function value can be improved towards approaching the preset threshold.

[0107] Based on the revenue function expression, which comprehensively considers factors such as fault location accuracy and correlation with adjacent agent policies, the iteration agent uses the gradient ascent method to calculate new policy space parameters. The gradient ascent method can adjust the parameters in the direction of the fastest growth of the revenue function. By continuously calculating the partial derivatives of the revenue function with respect to each policy parameter, the adjustment amplitude and direction of the parameters are determined, so that new policy space parameters that can improve the revenue function value are quickly found.

[0108] When an iteration agent completes the optimization and update of its own policy space parameters, the system temporarily stores the updated parameters, then selects the next agent as a new iteration agent, and repeats the process of fixing the parameters of other agents and optimizing its own parameters. Until all agents complete the update of their policy space parameters in this iteration, the iteration is over.

[0109] Subsequently, the system uses the updated policy space parameters of all agents in this iteration as the initial values for the next iteration, and starts a new iteration process. This cycle repeats, with the operation of fixing the parameters of other agents and each agent optimizing its own parameters in turn.

[0110] After each iteration is completed, the system calculates the change in the policy space parameters of all agents after the update in the current iteration, compared to the policy space parameters at the end of the previous iteration. When the change in the policy space parameters of all agents in adjacent iterations is less than a pre-set convergence threshold, it means that the policy space parameters of each agent have stabilized, and further iteration has little effect on the improvement of the revenue function value. At this time, the policy space of each agent reaches the threshold state of the revenue function, and the iteration process is terminated.

[0111] In this embodiment, the line is divided into several physical sections according to the topology of the power distribution network, and each physical section is associated with an agent, specifically:

[0112] According to the topology of the power distribution network, the line is divided into several physical sections with protection devices as boundaries;

[0113] An agent is created for each physical section, which is constructed based on the collected adjacent section connection relationship, the type parameters of the section power equipment, the initial fault probability distribution generated by historical fault statistics, and the fault propagation weight coefficient generated based on the topological distance.

[0114] According to the power distribution network topology, the line is divided into several physical sections with the protection device as the boundary. The power distribution network topology clearly presents the trend of the line, the distribution of nodes, the connection relationship of equipment, and other key information, which is an important basis for dividing sections. The protection device, such as a circuit breaker and a fuse, has the main function of rapidly acting when a fault occurs in the line, cutting off the fault current, and preventing the spread of the fault. Dividing the physical section with the protection device as the boundary can ensure that the protection device can independently function when a fault occurs in each section, realize the rapid isolation of the fault, and also make the range of each section relatively clear and have electrical independence, laying a foundation for subsequent fault location and analysis. For example, the line part between two adjacent circuit breakers in a feeder line can be regarded as an independent physical section.

[0115] The adjacent section connection relationship, the section equipment type parameters, the initial fault probability distribution, and the fault propagation weight coefficient of each physical section are collected, and the corresponding agent is created according to the above attributes. The adjacent section connection relationship refers to the connection mode and connection position of the current section and other sections, and the clear relationship helps the agent to consider the possible influence of adjacent sections when judging the fault, so as to realize the interaction and cooperation of information. The section equipment type parameters cover the model, specification, material and other information of the devices such as lines, transformers and switches contained in the section. Different types of devices have different fault characteristics, and these parameters are important basis for the agent to analyze the fault probability of the section. The initial fault probability distribution is obtained based on the past fault data statistics of the section, which reflects the possibility of the fault of the section under normal operation, and provides a reference benchmark for the initial judgment of the agent. The fault propagation weight coefficient reflects the possibility of the fault propagation from the current section to the adjacent section, and the greater the coefficient, the higher the probability of fault propagation.

[0116] After collecting these attribute information, it is set as the core attribute of the agent, so that an agent corresponding to each physical section is created. Each agent can accurately represent the characteristics of the physical section it is in, independently carry out fault monitoring, analysis and judgment, and also provide data support for the collaborative operation of agents.

[0117] Embodiment 2, Figure 2 A device for a power distribution network line fault location method is given, which includes a data acquisition module, a feature extraction and mapping module, a probability reasoning module, and a global decision module, and there is a connection between the modules:

[0118] a data acquisition module, configured to acquire operation data and lightning ground flash data;

[0119] a feature extraction and mapping module, configured to perform feature extraction on the operation data and to perform space-time feature mapping on the lightning ground flash data, to generate multi-source feature parameters;

[0120] a probability reasoning module, configured to perform probability reasoning on the input multi-source feature parameters through a distributedly deployed section intelligent agent and a Bayesian network submodule thereof, and to output real-time fault probability values of each independent section of the power distribution network;

[0121] a global decision module, configured to take the real-time fault probability values of each independent section as input, to dynamically coordinate output results of each intelligent agent through a Nash equilibrium strategy, and to generate a globally consistent fault location decision.

[0122] A power distribution network line fault location device, characterized in that it comprises an electronic device, characterized in that the electronic device comprises:

[0123] Embodiment 3, Figure 3 The application provides a device for a power distribution network line fault location method, which comprises at least one processor, an input / output interface in communication connection with the at least one processor, and a memory in communication connection with the at least one processor, wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to execute the power distribution network line fault location method.

[0124] The above formulas are all dimensionless numerical calculations, and the formulas are obtained by software simulation of a large amount of data to obtain a formula of the nearest real situation, and the preset parameters in the formula are set by a person skilled in the art according to the actual situation.

[0125] The above embodiments can be realized wholly or partially by software, hardware, firmware or any combination thereof. When realized by software, the above embodiments can be realized in the form of a computer program product wholly or partially.

[0126] Those skilled in the art can realize that the modules and algorithm steps of the examples described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized by hardware or software depends on the specific application and design constraints of the technical solutions. Those skilled in the art can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0127] In addition, each function module in each embodiment of the present application can be integrated in one processing module, or each module can be physically present alone, or two or more modules can be integrated in one module.

[0128] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0129] Finally: the above is only a preferred embodiment of the present application and is not used to limit the present application, any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. A method for locating a fault in a distribution network, characterized in that: The following steps are involved: Obtain operation data and lightning ground flash data; Feature extraction is performed on the operating data, and spatiotemporal feature mapping is performed on the lightning ground flash data to generate multi-source feature parameters. Specifically, the operating data is standardized and the operating steady-state feature parameters are generated by splicing feature vectors; the distribution network topology is constructed, and the ground flash data and the topology are spatially integrated to generate lightning spatiotemporal feature parameters; the operating steady-state feature parameters and the lightning spatiotemporal feature parameters are weightedly integrated to output the multi-source feature parameters; Through the distributed deployment of the segment agent and its Bayesian network submodule, probabilistic reasoning is performed on the input multi-source characteristic parameters to output the real-time fault probability value of each independent segment of the distribution network; Taking the real-time fault probability of each independent segment as input, the Nash equilibrium strategy dynamically coordinates the output results of each intelligent agent to generate a globally consistent fault location decision. The operation data includes the binary code value of the isolation section status, the action timing characteristic value and the effective value of the feeder. The ground lightning data includes the spatial correlation coefficient of the line section and the ground lightning fault time difference coupling degree. The multi-source characteristic parameters include the operation steady-state characteristic parameters and the lightning time and space characteristic parameters.

2. The method for locating a distribution network line fault according to claim 1, wherein: The distributed deployed segment agent and its Bayesian network submodule perform probabilistic reasoning on the input multi-source feature parameters and output the real-time fault probability value of each independent segment of the distribution network, specifically: According to the topology of the distribution network, the line is divided into several physical sections. Each physical section is associated with an intelligent agent, and a Bayesian network submodule is constructed for each intelligent agent. Input the multi-source feature parameters into the corresponding agent as the observation data of the Bayesian network submodule; Set the conditional probability table of the Bayesian network based on historical data and expert experience; Based on the observed data and conditional probability table, the posterior probability of each section fault is calculated using the Bayesian probability propagation algorithm; Based on the posterior probability, the real-time fault probability value of each section is determined by a posterior probability estimation algorithm.

3. The method for locating a distribution network line fault according to claim 2, wherein: The construction of the distribution network topology structure is specifically as follows: Obtain line model data and grid spatial coordinate information; Extract node distribution and equipment connection relationships based on line model data and grid spatial coordinate information; A topology verification algorithm is used to perform connectivity verification on the node distribution and device connection relationship to generate verified topology information; Integrate the verified topology information and construct the distribution network topology structure including the physical connection relationship.

4. The method for locating a fault in a distribution network according to claim 3, wherein: The real-time fault probability value of each independent segment is used as input, and the output results of each intelligent agent are dynamically coordinated through the Nash equilibrium strategy to generate a globally consistent fault location decision. Specifically: The Bayesian network submodule of each agent is defined as a game participant, and the strategy space parameter of each agent is the real-time failure probability value of the corresponding segment; A profit function is constructed for each intelligent agent, whose function value is a weighted combination of the fault location accuracy index and the adjacent section fault correlation index; Dynamically adjust the policy space parameters of each agent through an iterative optimization algorithm until the payoff function of all agents reaches the preset threshold and the policy space parameter combination conforms to the Nash equilibrium strategy; The policy space parameters that satisfy the Nash equilibrium strategy are output as the globally consistent fault location decision.

5. The method for locating a distribution network line fault according to claim 4, characterized in that: The dynamic adjustment of the strategy space parameters of each agent through the iterative optimization algorithm is specifically as follows: Taking the policy space parameters of each agent as the optimization variable, the updated value of the policy space parameters of each agent in the current iteration round is calculated based on the preset profit function expression; Use the updated policy space parameter values ​​as the initial values ​​for the next iteration, and repeat the policy space parameter optimization calculation for all agents. The iteration continues until the convergence conditions are met. The convergence conditions include that the changes in the strategy space parameters of all agents in two adjacent iterations are less than the preset convergence threshold, and the profit function values ​​of all agents reach the preset profit function threshold state.

6. The method for locating a distribution network line fault according to claim 5, characterized in that: The line is divided into several physical sections according to the distribution network topology, and each physical section is associated with an intelligent agent, specifically: According to the topology of the distribution network, the line is divided into several physical sections with the protection device as the boundary; An intelligent agent is created for each physical segment, which is constructed based on the collected connection relationships between adjacent segments, type parameters of segment power equipment, initial fault probability distribution generated by historical fault statistics, and fault propagation weight coefficient generated based on topological distance.

7. A device using the distribution network line fault location method according to any one of claims 1 to 6, characterized in that: It includes data acquisition module, feature extraction and mapping module, probability reasoning module and global decision module. There are connections between modules: Data acquisition module, to obtain operation data and lightning ground flash data; The feature extraction and mapping module extracts features from operational data and performs spatiotemporal feature mapping on lightning-to-ground flash data to generate multi-source feature parameters. The probabilistic reasoning module uses distributed segment agents and their Bayesian network submodules to perform probabilistic reasoning on the input multi-source feature parameters and output the real-time fault probability value of each independent segment of the distribution network; The global decision-making module takes the real-time fault probability value of each independent segment as input, dynamically coordinates the output results of each intelligent agent through the Nash equilibrium strategy, and generates a globally consistent fault location decision.

8. An electronic device, characterized in that: The electronic device comprises: At least one processor; and an input / output interface communicatively connected to the at least one processor; a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the distribution network line fault location method according to any one of claims 1 to 6.

Images