Substation topology link self-healing method and system based on ant colony algorithm

Abstract

The invention discloses a substation topological link self-healing method and system based on an ant colony algorithm, and relates to the technical field of intelligent power grids. The method comprises the following steps: identifying a fault area and a fault node of a transformer substation, and generating a dynamically updated taboo table; searching a candidate path set under the constraint of a tabu table based on the fault area and the fault node; performing quantitative evaluation on the candidate path set through a preset network index evaluation function, and sorting and selecting an optimal path according to an evaluation result; and switching the fault path to the optimal path, and updating the circuit network topology. The method is used for solving the problems that a traditional method is high in fault recovery time delay, poor in complex topology adaptability and insufficient in global dynamic planning.

Description

Technical Field

[0001] This invention relates to the field of smart grid technology, and more specifically, to a method and system for self-healing substation topology links based on ant colony algorithm. Background Technology

[0002] With the continuous expansion of power system scale and the improvement of intelligence level, substations, as key hubs in the power network, have a significant impact on the stability of the entire power grid due to their operational reliability and self-healing capabilities. Existing power system fault tolerance and fault recovery methods mainly rely on pre-set redundancy designs and traditional path reconfiguration strategies.

[0003] However, traditional methods have the following shortcomings: fault modeling methods are mostly based on static or simplified network structures, which makes it difficult to fully reflect the combined effects of multiple factors such as nodes and links, resulting in limited prediction and evaluation accuracy; path search strategies often focus on a single indicator and rely on fixed rules or heuristic algorithms, making it difficult to take into account multiple needs in a large and complex topology environment; with the rapid changes in the power grid's operating status, existing methods also have insufficient dynamic adaptability, and path adjustment is often lagging, resulting in less than ideal recovery effects; multi-indicator comprehensive evaluation mechanisms are relatively simple, often using fixed weights or simple weighting methods, lacking flexibility and making it difficult to reflect the differences in the importance of each indicator under different scenarios.

[0004] To address the above problems, this invention proposes a solution. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a substation topology link self-healing method and system based on ant colony algorithm. By driving evaluation with fault propagation probability model, convergence of tabu list constraints, and adaptive fusion of multiple indicators, the optimal path is dynamically explored and screened, fault paths are quickly switched and topology is updated, and continuous operation of services is achieved. This solves the problems of high fault recovery latency, poor adaptability to complex topologies, and insufficient global dynamic planning in traditional methods.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a substation topology link self-healing method based on ant colony algorithm, comprising the following steps: identifying substation fault areas and fault nodes, and generating a dynamically updated tabu list; searching for a candidate path set based on the fault areas and fault nodes under the constraints of the tabu list; quantifying and evaluating the candidate path set through a preset network index evaluation function, sorting according to the evaluation results and selecting the optimal path; switching the fault path to the optimal path and updating the circuit network topology.

[0007] In a preferred embodiment, the identification of substation fault areas and fault nodes is performed using a fault propagation probability model. The specific method for constructing the fault propagation probability model is as follows: extracting node features and link features from the substation topology weighted graph; performing correlation modeling on the node features and link features to generate fault propagation rules; constructing a probability propagation model framework based on the node features, link features, and fault propagation rules; and training the fault propagation probability model framework using historical fault data and real-time operating data to obtain the fault propagation probability model.

[0008] In a preferred embodiment, the method for generating the fault propagation rules is as follows: constructing a probabilistic graphical model based on node features, link features, and preset propagation adjustment variables; performing fault reasoning through the probabilistic graphical model to determine the effective path of fault propagation; calculating the fault probability of associated nodes based on the effective path; and generating a set of fault propagation rules according to the conditional dependencies in the probabilistic graphical model and the fault reasoning results.

[0009] In a preferred embodiment, the probabilistic graphical model is a Bayesian network model, and the associated node refers to a node connected to the fault source node through the fault propagation path.

[0010] In a preferred embodiment, the step of identifying substation fault areas and fault nodes and generating a dynamically updated tabu list specifically involves: calculating the fault probability value of nodes using a fault propagation probability model, marking nodes exceeding a preset probability threshold as fault candidate nodes; generating fault propagation paths based on the fault propagation probability model to perform correlation clustering on the fault candidate nodes, forming preliminary fault areas; filtering the preliminary fault areas based on preset screening rules to determine the final fault areas; and incorporating the nodes within the final fault areas and their adjacent nodes into the initial tabu list, and dynamically updating the tabu list based on the predicted fault propagation trend.

[0011] In a preferred embodiment, the step of searching for a candidate path set based on the fault region and fault node under the constraints of a tabu list specifically involves: setting the fault node as the search starting point and initializing the search state within the fault region; obtaining the fault propagation probability parameters of the links adjacent to the starting point; constructing a comprehensive evaluation value based on the link pheromone concentration, heuristic function value, and fault propagation probability parameters; calculating the transition probability of each adjacent node based on the comprehensive evaluation value, and selecting the next node and adding it to the path in conjunction with the tabu list and fault region constraints; updating the link pheromone concentration based on the path search results, repeating the path search and pheromone update until the termination condition is met, and outputting the candidate path set.

[0012] In a preferred embodiment, the weight coefficients of each parameter in the comprehensive evaluation value are adaptively configured according to the risk level of the fault area; the adaptive configuration is set to increase the weight ratio of the fault propagation probability parameter in the comprehensive evaluation value in fault areas with higher operational risk levels.

[0013] In a preferred embodiment, an enhanced evaporation strategy is used to update pheromones for links located within the fault region or constrained by the taboo table.

[0014] In a preferred embodiment, the step of quantifying and evaluating the candidate path set using a preset network metric evaluation function, sorting the paths according to the evaluation results, and selecting the optimal path specifically involves: dynamically adjusting the weights corresponding to multiple network performance metrics based on the fault propagation probability; performing nonlinear fusion calculations on the candidate paths according to the adjusted weights to obtain a comprehensive score; sorting the candidate paths according to the comprehensive score, and using a multi-objective optimization strategy to determine the optimal path.

[0015] Secondly, this application provides a substation topology link self-healing system based on ant colony algorithm, including: a fault identification and taboo generation module, used to identify fault areas and fault nodes in the substation and generate a dynamically updated taboo table; a candidate path search module, used to search for a set of candidate paths based on the fault areas and fault nodes under the constraints of the taboo table; a path evaluation and ranking module, used to quantitatively evaluate the set of candidate paths through a preset network index evaluation function, rank them according to the evaluation results, and select the optimal path; and a path switching and topology update module, used to switch the faulty path to the optimal path and update the circuit network topology.

[0016] As can be seen from the above technical solutions, the present invention constructs constraints by dynamically updating the taboo table and combines optimized ant colony search and screening to evaluate and obtain the optimal path, and then quickly switches to the optimal path and updates the circuit network topology, effectively solving the problems of high fault recovery delay and poor adaptability to complex topologies, and improving the adaptability of global dynamic programming. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the self-healing method for substation topology links based on the ant colony algorithm of this invention.

[0018] Figure 2 This is a schematic diagram of the substation topology link self-healing system based on the ant colony algorithm of the present invention.

[0019] Figure 3 A flowchart illustrating the modeling process of the fault propagation probability model provided in this embodiment of the invention.

[0020] Figure 4The flowchart illustrates the ant colony algorithm for searching candidate paths provided in this embodiment of the invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0022] Example 1, Figure 1 This invention presents a substation topology link self-healing method based on ant colony algorithm, comprising the following steps: S1 identifies fault areas and fault nodes in the substation and generates a dynamically updated taboo table.

[0023] Obtain the initial data and construct a weighted topology graph of the substation.

[0024] In this embodiment, the first data includes device link relationship data, link parameter data, node attribute data, and topology structure data of the substation topology link network.

[0025] The construction of the substation topology weighted graph specifically includes: By parsing the device link relationships, a set of nodes and a set of edges are formed; By combining link parameters and node attributes, corresponding feature values ​​are assigned to edges and nodes, such as impedance, communication delay, reliability indicators, node degree, historical failure rate, etc. Based on this, a weighted adjacency matrix is ​​generated using a conventional weighting method, thereby obtaining a weighted substation topology graph representing the substation topology link network.

[0026] In this embodiment, the identification of substation fault areas and fault nodes is performed using a fault propagation probability model. The specific method for constructing the fault propagation probability model is as follows: Extract node and link features from the weighted topology graph of the substation; Correlation modeling is performed on node characteristics and link characteristics to generate fault propagation rules; Based on node characteristics, link characteristics, and fault propagation rules, a probabilistic propagation model framework is constructed. The fault propagation probability model framework is trained using historical fault data and real-time operational data to obtain the fault propagation probability model.

[0027] The construction process of the above fault propagation probability model can be summarized as follows: Figure 3 The flowchart shown; Specifically: Extract node features and link features from the weighted graph of the substation topology. The node features include node degree, criticality, historical failure rate, etc.; the link features include link weight, reliability index, and historical communication stability, etc. Node characteristics can be obtained through conventional graph theory analysis methods, such as counting the degree of nodes in the substation topology weighted graph and calculating historical failure rates by combining operation records. Link characteristics can be obtained through conventional calculations of parameters such as impedance, delay, and historical operational stability of the weighted edges; In this embodiment, the method for generating the fault propagation rules is as follows: A probabilistic graphical model is constructed based on node characteristics, link characteristics, and preset propagation adjustment variables; Fault reasoning is performed using probabilistic graphical models to determine effective fault propagation paths; Calculate the failure probability of associated nodes based on the effective path; Based on the conditional dependencies in the probabilistic graphical model and the results of fault inference, a set of fault propagation rules is generated.

[0028] Specifically: The probabilistic graphical model is a Bayesian network model, and the associated node refers to a node that is connected to the fault source node through the fault propagation path. First, obtain node features and link features, and set propagation adjustment variables to adjust the strength of conditional dependencies; The propagation adjustment variables consist of node criticality, link reliability, and link load; and are used to adjust the strength of conditional dependencies.

[0029] Among them, the propagation moderating variable is positively correlated with node criticality, that is, the higher the node criticality, the greater its propagation influence in the network; it is negatively correlated with link reliability, that is, the lower the link reliability, the more likely the link is to become a failure propagation path; and it is positively correlated with link load, that is, the higher the link load, the greater its failure propagation risk. Node features and link features are input into the structure learning algorithm, and a propagation adjustment variable is introduced into the scoring function.

[0030] For example, when using the K2 algorithm based on scoring, an adjustment factor can be added to the original likelihood scoring function to give nodes and links with high criticality, low reliability, or high load a higher weight in the scoring calculation. Through the above method, the directed acyclic graph structure of the Bayesian network is finally obtained; where the directed edges in the network structure represent conditional dependencies and are used to characterize the causal dependencies between different nodes.

[0031] The source node's fault state is input into the Bayesian network, and the posterior probabilities of other nodes' states are calculated using conditional probability, forming the fault probability distribution of the target node given the evidence. The set of posterior probabilities for all nodes constitutes the fault inference result. The formula for calculating the posterior probability of the state is:

[0032] in, The fault source node, For the target node; When the conditional probability on a certain path is greater than a preset threshold When this condition is met, the path is marked as a valid path. By traversing the candidate path set in the network, a set of all valid paths that meet the conditions is obtained. For a target node connected to the fault source node via a valid path, its fault probability can be calculated using the conditional probability chain along the path:

[0033] in, Indicates from the source point To the target node The effective path. Through cumulative calculation, the overall failure probability of the target node can be obtained; Based on the conditional dependencies of the Bayesian network model and the fault inference results, a set of fault propagation rules is generated. These rules can be expressed in IF-THEN form, for example: If node A fails AND link (A, B) reliability < rule condition threshold If so, then the probability of node B failing is greater than 0.7; IF Node C Fault AND Node D Criticality > Rule Condition Threshold If the probability of node E failing is greater than 0.6.

[0034] In this embodiment, based on node features, link features, and fault propagation rules, a three-layer probability propagation model framework comprising an input layer, an inference layer, and an output layer is constructed, specifically including: Construction of the input layer: First, the node features and link features are standardized or normalized respectively to eliminate the differences caused by different units; Then, the processed node features and link features are vectorized and organized into a unified feature input matrix. This matrix serves as the data foundation for the model input layer and is used for subsequent probability calculations and inferences.

[0035] Construction of the inference layer: Based on fault propagation rules, the dependencies between nodes and links are extracted; The dependencies are mapped to conditional probability tables (CPTs), and directed edge relationships are constructed.

[0036] For example: the rule "IF Node A failure AND Link (A, B) reliability < rule condition threshold" THEN node B failure probability = 0.7” will be represented as a directed edge A→B in the inference layer and the probability value will be recorded in CPT; When multiple rules act on the target node, a conditional probability distribution is obtained by weighted synthesis through propagation of adjustment variables.

[0037] Construction of the output layer: The output of the model is determined to be the failure probability of each target node; The conditional probability results of the inference layer are accumulated and calculated along the effective propagation path to obtain the final probability value of the target node; the function of the output layer is to map the probability calculation results of the inference layer into node-level fault risk indicators, providing support for subsequent prediction and operation and maintenance decisions.

[0038] Through the above methods, the input layer, inference layer, and output layer are gradually constructed to form a three-layer probability propagation model framework.

[0039] Based on the established three-layer probability propagation model framework, the model parameters are trained using historical fault data and real-time operation data of the substation topology link network to obtain a converged fault propagation probability model.

[0040] This step can be implemented using conventional parameter estimation methods in the field, such as maximum likelihood estimation or Bayesian estimation, and the specific implementation method is not limited.

[0041] In this embodiment, the process of identifying substation fault areas and fault nodes and generating a dynamically updated taboo table specifically involves: The failure probability value of a node is calculated using a failure propagation probability model, and nodes that exceed a preset probability threshold are marked as candidate failure nodes. Fault propagation paths are generated based on the fault propagation probability model, and candidate fault nodes are clustered to form preliminary fault regions. The initial fault areas are filtered based on the first preset filtering rules to determine the final fault areas; The nodes within the final fault region and their adjacent nodes are included in the initial tabu list, and the tabu list is dynamically updated according to the predicted fault propagation trend.

[0042] Specifically: The fault probability value of each node in the substation topology weighted graph is calculated using a trained fault propagation probability model. If the fault probability value of a certain node exceeds a preset probability threshold... If so, then mark the node as a candidate node for failure.

[0043] The probability threshold It can be set adaptively based on historical fault statistics, or it can be preset to a fixed value based on operational experience; Based on the substation topology weighted graph and fault propagation probability model, fault propagation paths between nodes are generated. Specifically, the weighted topology graph of a substation can be represented by a topology connection matrix, which describes the adjacency relationship and link weight between nodes. Combined with the conditional probability table (CPT) in the fault propagation probability model, the propagation probability between different candidate nodes can be calculated.

[0044] When the cumulative propagation probability of a path is greater than the path connectivity threshold When this happens, the path is marked as a fault propagation path; Based on the generated fault propagation paths, candidate fault nodes are clustered according to their correlations. In one alternative implementation, a connectivity-based graph partitioning method can be used. Specifically, when there is a propagation path between two candidate nodes, and the cumulative propagation probability is greater than the candidate node determination threshold. When the two are divided into the same subset, several subsets are obtained by recursive traversal, forming the initial fault region; In addition to the methods described above, other common clustering methods such as hierarchical clustering and spectral clustering can also be used, and this invention does not limit these methods. The initial fault areas are filtered using the first preset screening rule, eliminating areas that are too small or have an average fault probability lower than the regional average probability threshold. In the area, the main high-risk areas are retained to obtain the final failure area; The first preset filtering rules may include the number of nodes in the region, the link connectivity index in the region, etc. The nodes within the final fault region and their adjacent nodes are included in the initial tabu list. This tabu list is used to constrain the search space during subsequent path reconstruction or self-healing processes, preventing faulty nodes or nodes closely connected to faulty nodes from being selected repeatedly. The fault propagation probability model is used to predict the fault propagation trend, and nodes that may be affected in the future are dynamically added to the initial taboo list. Specifically, a probabilistic recursive method can be used to perform time-series deduction of node failure probabilities. For example, in one optional implementation, a state transition matrix can be constructed based on the conditional probability table in the failure propagation probability model, using a Markov chain recursive approach:

[0045] in, Indicates time Fault probability vector of each node, This represents the state transition matrix generated by the conditional probability table (CPT).

[0046] When the prediction results show that the probability of failure of a certain node in the future period exceeds the dynamic prediction threshold When the node is added to the initial tabu list, the final tabu list is obtained.

[0047] Furthermore, those skilled in the art will understand that time series prediction methods can also be used to fit the historical failure probability sequence to obtain the failure probability of future nodes, thereby achieving dynamic updates to the tabu list.

[0048] S2, based on the fault region and fault node, searches for a set of candidate paths under the constraints of the taboo list, specifically as follows: Set the faulty node as the search starting point and initialize the search state within the faulty area; Obtain the fault propagation probability parameters of the links adjacent to the starting point; A comprehensive evaluation value is constructed based on link pheromone concentration, heuristic function value, and fault propagation probability parameters; The transition probability of each adjacent node is calculated based on the comprehensive evaluation value, and the next node is selected and added to the path in combination with the taboo table and fault area constraints; Update the link pheromone concentration based on the path search results, repeat the path search and pheromone update until the termination condition is met, and output the candidate path set.

[0049] Figure 4 The detailed process of exploring candidate path sets under tabu list constraints using the ant colony algorithm is demonstrated.

[0050] First, initialize the ant colony parameters and set the faulty node as the initial position of the ants. Then, repeat the following steps in a loop: The algorithm obtains the failure propagation probability of adjacent links, calculates the comprehensive evaluation value and transition probability, selects the next hop node according to constraints, updates the ant path and makes a judgment, and decides whether to add the path to the candidate set after the individual termination condition is met.

[0051] In the comprehensive evaluation value, the weight coefficients of each parameter are adaptively configured according to the risk level of the fault area; The adaptive configuration is set to increase the weight of the fault propagation probability parameter in the comprehensive evaluation value in fault areas with higher operational risk levels. After all ants have completed their search, update the link pheromone concentration and check if the global termination condition is met. If it is, output the candidate path set; otherwise, continue iterating.

[0052] Specifically: The faulty node is set as the initial position of the ants, and the exploration state of the ant colony algorithm is initialized within the faulty area. The exploration state includes exploration parameters such as the number of ants, the initial pheromone concentration, and the heuristic function.

[0053] Set the initial pheromone value to:

[0054] in, Indicates link The pheromone concentration at the initial iteration step It is a constant.

[0055] Heuristic function is

[0056] in, For link The equivalent distance or relation weights are normalized to be dimensionless quantities.

[0057] The fault propagation probability parameters of the adjacent links at the initial position are obtained through a fault propagation probability model. Specifically, based on the conditional probability table (CPT) of a Bayesian network, the adjacent links can be obtained. Fault propagation probability :

[0058] in, , Representing nodes respectively , state, .

[0059] A comprehensive evaluation value is constructed based on the pheromone concentration of adjacent links, heuristic function values, and fault propagation probability parameters:

[0060] in, For the iteration step index, For link The overall evaluation value, , , They represent the risk index respectively. Relevant weighting coefficients; Unlike the fixed-weight method, the weight coefficients of each parameter in this embodiment can be determined based on the risk index of the fault area. Perform adaptive adjustment: when At higher levels, Increase the value to enhance the role of the fault propagation probability parameter in the overall evaluation value; when At lower levels, and The weights of pheromones and heuristic functions are relatively increased to highlight their guiding role. Risk Index It can be calculated based on the average failure probability of nodes in the fault area, the connectivity index of links in the area, or historical operating data.

[0061] Furthermore, the transition probability of each adjacent node is calculated based on the comprehensive evaluation value:

[0062] The Allowed set is the set of neighboring nodes that are still available after removing the taboo list and fault region constraints. Adjacent node set nodes; Subsequently, based on the transition probabilities of each adjacent node The next hop node is determined by a roulette wheel selection mechanism, which involves sampling according to the transition probability distribution. The higher the probability value of the adjacent node, the higher the probability of it being selected, and the next hop node is added to the ant's movement path. Ants gradually expand their path as they move, the first... An ant's path can be represented as a sequence. , This represents the node index in the path sequence, which is appended to the sequence when a new node is selected. Each ant repeatedly selects the next hop node until an individual termination condition is triggered, ending path generation. The individual termination condition includes reaching the target node, having no feasible adjacent nodes, or the path length reaching the upper limit. When path generation ends, if the obtained path satisfies the fault area constraint and link reliability judgment rules, the path is included in the candidate path set; otherwise, it is discarded. After all ants have completed their path search, the pheromones of the links are updated uniformly based on the quality of each path, and the next iteration begins. The specific update formula is as follows:

[0063] in, For link Pheromones are volatile factors; This represents the number of iterations. For the first Link during the next iteration Pheromones increase; You can set the path quality, for example:

[0064] in, This indicates the number of ants participating in the current iteration. Indicates an indicator function if and only if the first... Only one ant passes through the link The value is 1 if the condition is met, and 0 otherwise. Indicates the first The path length of an ant. This represents the pheromone release intensity constant.

[0065] In one improved approach, for links located within a fault region or constrained by a taboo list, an enhanced volatility strategy is employed, i.e., a larger [value] is set. The value is adjusted to cause the pheromone concentration to decay rapidly, thereby reducing the probability of this type of link being selected in subsequent iterations; for links in non-faulty areas, a conventional evaporation factor is used.

[0066] Repeat the above iterative process until the global termination condition is met; The global termination conditions include the number of iterations reaching a set threshold, or the optimal quality improvement of the candidate path in several consecutive iterations being lower than a preset threshold, thereby determining that the exploration has converged and outputting the final candidate path set.

[0067] S3, the candidate path set is quantitatively evaluated using a preset network metric evaluation function, and the optimal path is selected based on the evaluation results. Specifically: The weights of multiple network performance metrics are dynamically adjusted based on the probability of fault propagation. Based on the adjusted weights, a nonlinear fusion calculation is performed on the candidate paths to obtain a comprehensive score; Candidate paths are ranked based on comprehensive scores, and the optimal path is determined using a multi-objective optimization strategy.

[0068] Specifically: First, the fault propagation probability is obtained and used to dynamically adjust the weight parameters corresponding to multiple network performance indicators. The fault propagation probability is obtained from the output of the fault propagation probability model. During the weight adjustment process, when the probability of failure propagation of nodes or links in the candidate path is high, the weight of the "link reliability" indicator is increased accordingly; when the propagation probability is low, the weight of the "transmission timeliness" or "node load balancing" indicator is increased accordingly, thereby achieving adaptive adjustment of the importance of the indicators. Subsequently, multiple indicators of the candidate paths were quantitatively calculated.

[0069] The network performance metrics include: bandwidth redundancy, transmission timeliness, link reliability, and node load balancing.

[0070] After obtaining the values ​​of multiple indicators and their dynamic weights, a nonlinear fusion calculation is performed on the candidate paths to obtain a comprehensive score. This embodiment uses a weighted product model:

[0071] in, Indicates the first The overall score of the candidate paths, This indicates that the path is at the . Quantitative values ​​for each indicator The weight parameters are dynamically adjusted based on the fault propagation probability. This refers to the number of indicators.

[0072] Finally, the candidate paths are ranked according to the comprehensive score, and the optimal path is determined by combining the non-dominated solution screening strategy. The non-dominated solution selection adopts the Pareto criterion: if path A is not inferior to path B on all indicators, and is superior to path B on at least one indicator, then path A is considered to dominate path B.

[0073] S5 switches the faulty path to the optimal path and updates the circuit network topology.

[0074] In this embodiment, after obtaining the optimal path, the original faulty path is switched to the optimal path, and the circuit network topology is updated. Specifically, when a faulty path is detected as unable to continue carrying services, the system adjusts the routing table or forwarding table according to the selected optimal path to switch service traffic; at the same time, it updates the link and node connection relationships in the substation topology weighted graph to reflect the new topology status and provide basic data for subsequent path selection and fault recovery.

[0075] Example 2, Figure 2 This invention presents a substation topology link self-healing control system based on ant colony algorithm, comprising: The fault identification and taboo generation module is used to identify fault areas and fault nodes in substations and generate a dynamically updated taboo table. The candidate path search module is used to search for a set of candidate paths based on the fault region and fault node, under the constraints of the tabu list. The path evaluation and ranking module is used to quantitatively evaluate the candidate path set through a preset network metric evaluation function, and then rank and select the optimal path based on the evaluation results. The path switching and topology update module is used to switch faulty paths to the optimal paths and update the circuit network topology.

[0076] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0077] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0078] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0079] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0080] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0081] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A substation topology link self-healing method based on an ant colony algorithm, characterized in that, Includes the following steps: Identify fault areas and fault nodes in substations and generate a dynamically updated taboo list; Based on the fault region and fault node, a candidate path set is searched under the constraints of the taboo list; The candidate path set is quantitatively evaluated using a preset network metric evaluation function, and the optimal path is selected based on the evaluation results. Switch the faulty path to the optimal path and update the circuit network topology.

2. The substation topology link self-healing method based on ant colony algorithm according to claim 1, characterized in that, The identification of substation fault areas and fault nodes is achieved through a fault propagation probability model. The specific method for constructing the fault propagation probability model is as follows: Extract node and link features from the weighted topology graph of the substation; Correlation modeling is performed on node characteristics and link characteristics to generate fault propagation rules; Based on node characteristics, link characteristics, and fault propagation rules, a probabilistic propagation model framework is constructed. The fault propagation probability model framework is trained using historical fault data and real-time operational data to obtain the fault propagation probability model.

3. The substation topology link self-healing method based on ant colony algorithm according to claim 2, characterized in that, The method for generating the fault propagation rules is as follows: A probabilistic graphical model is constructed based on node characteristics, link characteristics, and preset propagation adjustment variables; Fault reasoning is performed using probabilistic graphical models to determine effective fault propagation paths; Calculate the failure probability of associated nodes based on the effective path; Based on the conditional dependencies in the probabilistic graphical model and the results of fault inference, a set of fault propagation rules is generated.

4. The substation topology link self-healing method based on ant colony algorithm according to claim 3, characterized in that, The probabilistic graphical model is a Bayesian network model, and the associated node refers to a node that is connected to the fault source node through the fault propagation path.

5. The substation topology link self-healing method based on ant colony algorithm according to claim 4, characterized in that, The process of identifying substation fault areas and fault nodes and generating a dynamically updated taboo table is as follows: The failure probability value of a node is calculated using a failure propagation probability model, and nodes that exceed a preset probability threshold are marked as candidate failure nodes. Fault propagation paths are generated based on the fault propagation probability model, and candidate fault nodes are clustered to form preliminary fault regions. The initial fault areas are filtered based on preset filtering rules to determine the final fault areas; The nodes within the final fault region and their adjacent nodes are included in the initial tabu list, and the tabu list is dynamically updated according to the predicted fault propagation trend.

6. The substation topology link self-healing method based on ant colony algorithm according to claim 5, characterized in that, The process of searching for candidate paths based on fault regions and fault nodes under tabu list constraints is as follows: Set the faulty node as the search starting point and initialize the search state within the faulty area; Obtain the fault propagation probability parameters of the links adjacent to the starting point; A comprehensive evaluation value is constructed based on link pheromone concentration, heuristic function value, and fault propagation probability parameters; The transition probability of each adjacent node is calculated based on the comprehensive evaluation value, and the next node is selected and added to the path in combination with the taboo table and fault area constraints; Update the link pheromone concentration based on the path search results, repeat the path search and pheromone update until the termination condition is met, and output the candidate path set.

7. The substation topology link self-healing method based on ant colony algorithm according to claim 6, characterized in that, In the comprehensive evaluation value, the weight coefficients of each parameter are adaptively configured according to the risk level of the fault area; the adaptive configuration is set to increase the weight ratio of the fault propagation probability parameter in the comprehensive evaluation value in fault areas with higher operational risk levels.

8. The substation topology link self-healing method based on ant colony algorithm according to claim 7, characterized in that, For links located within the fault region or constrained by the taboo list, an enhanced evaporation strategy is used for pheromone updates.

9. The substation topology link self-healing method based on ant colony algorithm according to claim 8, characterized in that, The process of quantifying and evaluating the candidate path set using a preset network metric evaluation function, sorting the paths based on the evaluation results, and selecting the optimal path is as follows: The weights of multiple network performance metrics are dynamically adjusted based on the probability of fault propagation. Based on the adjusted weights, a nonlinear fusion calculation is performed on the candidate paths to obtain a comprehensive score; Candidate paths are ranked based on comprehensive scores, and the optimal path is determined using a multi-objective optimization strategy.

10. A system for a substation topology link self-healing method based on ant colony algorithm, applied to the substation topology link self-healing method based on ant colony algorithm as described in any one of claims 1-9, characterized in that, include: The fault identification and taboo generation module is used to identify fault areas and fault nodes in substations and generate a dynamically updated taboo table. The candidate path search module is used to search for a set of candidate paths based on the fault region and fault node, under the constraints of the tabu list. The path evaluation and ranking module is used to quantitatively evaluate the candidate path set through a preset network metric evaluation function, and then rank and select the optimal path based on the evaluation results. The path switching and topology update module is used to switch faulty paths to the optimal paths and update the circuit network topology.

Images