Load balancing based branch topology automatic switching control method and system

Abstract

This invention discloses an automatic switching control method and system for branch topology based on load balancing, belonging to the field of automatic switching control technology. The method includes the following steps: acquiring operating condition data and constructing a joint graph; based on the joint graph, extracting power transmission paths from the backbone node to each load node to form a set of high-load segments; for the high-load transmission segments, constructing a multi-objective switching feasibility set and generating a coupling relationship mapping graph; based on the coupling relationship mapping graph, constructing a path search model and using a reinforced iterative algorithm to solve for the optimal topology switching path; generating a topology switching sequence according to the optimal topology switching path and implementing power transfer regulation on the high-load transmission segments. This invention, by constructing a joint graph and generating the optimal topology switching path, effectively alleviates local overload, reduces power fluctuations, and improves the overall load balancing level and active regulation capability of the system.

Description

Technical Field

[0001] This invention relates to the field of automatic switching control technology, and more specifically, to an automatic switching control method and system for branch topology based on load balancing. Background Technology

[0002] In the current integrated power grid system, with the rapid increase in the number of loads connected and the continuous improvement of power electronics, the load distribution of the power grid exhibits high dynamism and local clustering. Against this backdrop, achieving flexible control of the power grid topology and dynamic load balancing has become crucial for ensuring the safe and stable operation of the system. However, existing topology switching control methods still rely heavily on static configuration and empirical rules, depending heavily on manual judgment and local information collection by maintenance personnel, making it difficult to adapt to the control challenges brought about by high-frequency disturbances and complex interactions.

[0003] Specifically, existing systems typically use alarm signals such as node current exceeding limits and voltage exceeding limits as criteria for load imbalance, ignoring the energy coupling relationships between paths at each level in the primary and secondary distribution networks and their impact on the overall system operation. This approach of localized perception and isolated analysis makes it difficult for the system to detect potential high-risk transmission paths in a timely manner before load imbalance problems are exposed, and it also fails to effectively assess the power flow disturbances and electrical stability impacts brought about by topology switching, thus limiting the practicality and robustness of automated switching strategies.

[0004] The above-disclosed technical solutions have at least the following technical problems: existing systems usually use simple overcurrent alarms or node voltage judgments as the basis for load imbalance, and lack a branch load assessment mechanism from a global perspective. Specifically, they cannot identify the "high load transmission segment" in the branch path; and they cannot assess the coupling degree and response sensitivity of the load transmission chain between nodes.

[0005] To address the above problems, this invention proposes a solution. Summary of the Invention

[0006] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide an automatic switching control method and system for branch topology based on load balancing. By constructing a state-structure joint graph, high-load transmission segments are identified, and the optimal topology switching path is generated based on the coupling mapping relationship. This solves the problems in the prior art, such as the inability to accurately identify key areas of load imbalance, the lack of a systematic evaluation mechanism for topology switching paths, and the high risk of disturbance during the switching execution process.

[0007] To achieve the above objectives, the present invention provides the following technical solution:

[0008] On the one hand, the automatic switching control method for branch topology based on load balancing includes the following steps: collecting operating condition data of electrical nodes and constructing a joint graph of node operating status and node topology; based on the joint graph, extracting the power transmission path from the backbone node to the load node through a weighted path tracing algorithm, and marking the key segments in the load transmission path to form a high-load segment set; for the high-load segment set, generating a multi-objective switching feasibility set that meets safety conditions by combining the physical constraints and scheduling capabilities of the power grid, and establishing a coupling relationship mapping graph between the switching path and electrical parameters; taking the minimum load deviation as the optimization objective, constructing a path search model based on the coupling relationship mapping graph, and using an enhanced iterative algorithm to solve for the optimal topology switching path; generating a topology switching sequence based on the optimal topology switching path, and implementing power transfer regulation on the high-load transmission segment.

[0009] In a preferred embodiment, the step of collecting the operating condition data of electrical nodes and constructing a joint graph of node operating status and node topology specifically involves: collecting the operating condition data of each electrical node through intelligent acquisition terminals deployed in the integrated main and distribution power grid, wherein the operating condition data includes node voltage, current, active power, topology connection relationships, and switch status; synchronizing the asynchronous sampling data of the main grid and distribution network; based on the original topology connection matrix of the power grid, encoding the synchronized node operating status and topology structure into a multi-dimensional state vector through tensor embedding method; stacking the multi-dimensional state vectors according to the time series to generate a state-structure joint graph tensor, and outputting the joint graph.

[0010] In a preferred embodiment, the step of extracting the power conduction path from the backbone node to the load node based on the joint graph and using a weighted path tracing algorithm specifically involves: identifying a set of backbone nodes based on the connectivity of nodes and the rated capacity of branches in the joint graph, wherein the backbone nodes include the main transformer node, the feeder outlet node, and the main power supply point of the ring network; starting from each backbone node, traversing all downstream paths using a weighted depth-first search algorithm, recording the path sequence to each end load node, and accumulating the total weighted value of the paths as the path cost; for scenarios where multiple paths can reach the same load node, selecting the unique path with the lowest path cost as the main power conduction path, and establishing a directed main power conduction linked list from the backbone node to the load node.

[0011] In a preferred embodiment, marking key segments in the load conduction path to form a high-load segment set specifically involves: calculating the unit power flow gradient between adjacent node pairs for each power conduction path; selecting all node pairs on the path, constructing historical power fluctuation sequences for each node pair, and calculating the cooperative coupling coefficient between any two node pairs; if multiple consecutive node pairs simultaneously meet the judgment criteria, they are merged into a high-load clustering path segment, and the average power density and concentration index of the path segment are calculated; the judgment criteria include cooperative coupling coefficient ≥ preset first threshold, unit power flow gradient ≥ preset second threshold, and cumulative path length ≤ preset clustering radius R; sorting according to the weighted values ​​of the average power density and concentration index, selecting the top K path segments as priority processing objects; marking all path segments that meet the judgment criteria as key segments to form a high-load segment set.

[0012] In a preferred embodiment, the step of generating a multi-objective switching feasibility set that meets safety conditions for the high-load segment set, combined with the power grid physical constraints and scheduling capabilities, specifically involves: identifying a candidate path set based on joint graph identification and the existence of structural bypasses or potential electrical channel connections between the high-load transmission segments;

[0013] For the candidate path set, a multi-objective path evaluation model is constructed, which includes power flow change deviation, voltage stability index, and node scheduling cost. During the path evaluation process, a set of nonlinear constraint functions is constructed, and a feasibility region mapping algorithm is applied to generate a branch-node switching reachability matrix. The reachability matrix is ​​then corrected for topology consistency: path combinations that lead to topology closure are eliminated; paths that cause power flow reversal are excluded; a subset of switchable paths that meet the physical constraints of power grid operation is retained; and all paths that simultaneously satisfy structural reachability, electrical safety, and scheduling responsiveness, along with their objective evaluation results, are included in the multi-objective switching feasibility set.

[0014] In a preferred embodiment, establishing the coupling relationship mapping map between switching paths and electrical parameters specifically involves: extracting the set of electrical nodes affected by each feasible path in the multi-objective switching feasibility set and their corresponding power change data, and establishing a mapping relationship between paths and nodes; calculating the power influence weight of the path on the nodes based on the active power change value caused by the path to the relevant nodes under the switching state, and forming a weighted connection relationship between paths and nodes; and organizing the mapping relationship and the weighted connection relationship into a graph to construct a path-node-power ternary coupling relationship mapping map containing a path set, a node set, and a power offset weight.

[0015] In a preferred embodiment, the optimization objective of minimizing load deviation, constructing a path search model based on a coupling relationship mapping graph, and solving for the optimal topology switching path using a reinforced iterative algorithm, specifically involves: based on the coupling relationship between paths, nodes, and power offset in the coupling relationship mapping graph, performing dimensionality reduction using a 3D graph embedding algorithm to generate low-dimensional dense vectors for each path in the topology switching space; combining the current system operating state, constructing a set of state transition trajectories using the low-dimensional dense vectors, and converting it into a training sample set in the form of state-action-reward triples; inputting the path search model based on the training sample set, training it using a deep policy gradient algorithm, and outputting the optimal topology switching path.

[0016] In a preferred embodiment, generating a topology switching sequence based on the optimal topology switching path specifically involves: analyzing the topological positional relationship and electrical coupling direction of each node in the path based on the optimal topology switching path output by the path search model; generating a time-ordered switching instruction set based on the electrical coupling direction and branch action priority rules; and using a sequential activation strategy to control the access order of candidate branches to form the final topology switching sequence.

[0017] In a preferred embodiment, the power transfer control of the high-load transmission segment specifically involves: extracting the set of power injection nodes and load-bearing nodes corresponding to the high-load transmission segment based on the optimal topology switching path; predicting power flow disturbances by combining power flow direction, voltage level, and path redundancy; decomposing the switching path into multiple micro-switching segments and constructing a release priority table according to the conduction path depth, node coupling strength, and voltage sensitivity coefficient; loading control segment by segment according to the priority table; setting dynamic buffer delay and voltage disturbance threshold for each segment; monitoring the voltage change rate and frequency fluctuation trend of adjacent nodes in real time; pausing the switching operation if the monitored value is close to the threshold and continuing after the system stabilizes; and guiding the activation of backup branches to divert part of the load to avoid concentrated power backflow or local undersupply.

[0018] On the other hand, the automatic switching control system for branch topology based on load balancing includes the following modules: a joint graph construction module, used to collect operating data of electrical nodes and construct a joint graph of node operating status and node topology; a power path extraction module, used to extract power transmission paths from backbone nodes to load nodes based on the joint graph using a weighted path tracing algorithm, and mark key segments in the load transmission paths to form a set of high-load segments; a switching feasibility analysis module, used to generate a multi-objective switching feasibility set that meets safety conditions for the set of high-load segments, combined with the physical constraints and scheduling capabilities of the power grid, and establish a coupling relationship mapping graph between switching paths and electrical parameters; a path optimization decision module, used to construct a path search model based on the coupling relationship mapping graph with the minimum load deviation as the optimization objective, and use a reinforced iterative algorithm to solve for the optimal topology switching path; and a topology switching and control module, used to generate a topology switching sequence based on the optimal topology switching path and implement power transfer control for high-load transmission segments.

[0019] The technical effects and advantages of the automatic switching control method and system for branch topology based on load balancing of this invention are as follows:

[0020] 1. This invention achieves accurate identification and quantitative evaluation of high-load transmission segments in integrated power grids by constructing a state-structure joint graph and introducing refined power path modeling and key segment identification mechanisms. Combined with multi-objective switching feasibility analysis and reinforcement learning-driven path optimization model, it can intelligently generate topology switching paths with minimal disturbance and optimal response, effectively alleviating local overload, reducing power fluctuations, and significantly improving the overall load balancing level and active control capability of the system.

[0021] 2. This invention constructs a micro-switching control mechanism based on slow-release priority, and adopts dynamic delay, voltage disturbance threshold, and backup path-guided power transfer strategies to effectively suppress disturbances such as voltage fluctuations and frequency offsets during topology reconfiguration. At the same time, it supports post-switching state self-feedback and iterative correction of path sets, realizing adaptive closed-loop optimization of topology strategies, effectively reducing the risk of misoperation and cascading trips, and ensuring scheduling safety and stable system operation during topology changes. Attached Figure Description

[0022] Figure 1 This is a flowchart illustrating the automatic switching control method for branch topology based on load balancing according to the present invention.

[0023] Figure 2 This is a schematic diagram of the automatic switching control system for branch topology based on load balancing according to the present invention. Detailed Implementation

[0024] 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.

[0025] Example 1, Figure 1 The present invention provides an automatic switching control method for branch topology based on load balancing, comprising the following steps:

[0026] S1, collect the operating condition data of electrical nodes and construct a joint graph of node operating status and node topology;

[0027] The operating data includes node voltage, current, active power, topology connections, and switch status; the joint graph simultaneously describes the node load status and its topological path location in the power grid; the node operating status includes voltage, current, and active power information; and the node topology structure includes static connections and dynamic edge weight attributes.

[0028] The fusion refers to semantically associating the node's operating state with the topology to form a joint "state-location" representation, including node state, topological location, and dynamic association: the node state vector and the topological path vector are coupled through tensor embedding technology to form a fused feature vector.

[0029] In this embodiment, the step of collecting the operating condition data of electrical nodes and constructing a joint graph of node operating status and node topology specifically involves:

[0030] The system collects operating data of each electrical node, including node voltage, current, active power, topology connection relationship and switch status, through intelligent acquisition terminals (including PMU, AMI, edge controller, etc.) deployed in the main and distribution integrated power grid.

[0031] By synchronously sampling the main grid and distribution network, the operating condition data is asynchronously aligned using a sliding time window mechanism, and the missing time point data is interpolated using a bilateral exponential interpolation algorithm.

[0032] Based on the original topology connection matrix of the power grid, the data after asynchronous alignment is structurally fused. The tensor embedding method is used to encode the node operating state and topology structure into a multi-dimensional state vector, and stack them in time sequence to form a state-structure joint graph tensor. The output is a joint graph that fuses the node operating state and topology structure. In the graph, the nodes represent electrical equipment entities, the edges represent topology connection relationships, and the edge weights represent real-time power flow and impedance parameters between nodes.

[0033] S2, based on the joint graph, extracts the power conduction path from the backbone node to the load node through the weighted path tracing algorithm, and marks the key segments in the load conduction path to form a set of high load segments;

[0034] Backbone nodes: refer to the core nodes that connect to the upper-level power grid or main power source (such as substation bus nodes, the starting nodes of main lines), serving as the starting point set of power transmission paths;

[0035] Load node: refers to the terminal node directly connected to the user load (such as the low-voltage side node of the distribution transformer, the user access point), which serves as the endpoint set for path tracing.

[0036] The key paragraphs are those with concentrated loads and significant coupling sensitivity;

[0037] In this embodiment, the extraction of the power conduction path from the backbone node to the load node based on the joint graph and using a weighted path tracing algorithm specifically involves:

[0038] Based on the connectivity of nodes and the rated capacity of branches in the joint graph, the set of backbone nodes is identified. The main nodes include main transformer nodes, feeder outlet nodes, and ring network main power supply points;

[0039] From each backbone node Starting from the beginning, the weighted depth-first search algorithm is used to traverse all downstream paths connected to it, record the path sequence to each end load node l, and accumulate the total weighted value of the path as the path cost.

[0040] In scenarios where multiple paths can reach the same load node, the path with the lowest cost is selected as the main power conduction path, and a main power conduction linked list is established from the backbone node to each load node based on this path.

[0041] The path cost is specifically as follows:

[0042]

[0043] in, Let be the instantaneous power change from node i to i+1. The maximum rated power of the node in the system. The impedance between nodes. The reference impedance value, The voltage phase angle difference between adjacent nodes. , , Each has its own weighting factor (set based on historical experience). This represents the total number of nodes on the path.

[0044] In this embodiment, the step of marking key segments in the load transmission path to form a high-load segment set specifically involves:

[0045] For each obtained power conduction path Calculate the unit power flow gradient between adjacent node pairs. ,in Let be the instantaneous power change from node i to i+1. This represents the topological path length.

[0046] Select all node pairs on the path Construct historical power fluctuation sequences for node pairs And calculate the cooperative coupling coefficient. ,in For covariance, Standard deviation;

[0047] If multiple consecutive node pairs meet the judgment conditions and their cumulative path length is less than the preset aggregation radius, they are merged into a high-load aggregation path segment, and the average power density and concentration index of the segment are calculated. The top K path segments with the largest values ​​are then sorted and selected as priority processing objects.

[0048] All path segments that meet the above criteria are marked as key segments and grouped into a set. This refers to the set of high-load segments.

[0049] In this embodiment, the determination condition is specifically as follows:

[0050] When the unit power flow gradient is greater than a preset first threshold and the cooperative coupling coefficient is greater than a preset second threshold.

[0051] The average power density is specifically:

[0052]

[0053]

[0054] The concentration is specifically defined as follows:

[0055]

[0056] in, For average power density, This represents the sum of topological distances. Let be the instantaneous power change from node i to i+1. Let be the length of the topological path from node i to i+1. The number of nodes For concentration, This represents the number of load nodes in the path segment.

[0057] S3, for the high-load segment set, combines the power grid physical constraints and dispatch capabilities to generate a multi-objective switching feasibility set that meets safety conditions, and establishes a mapping graph of the coupling relationship between switching paths and electrical parameters;

[0058] The coupling relationship mapping map is a model used in high-load power grid dispatching to intuitively present the relationship between switching paths, affected electrical nodes, and power offset weights. It can clearly reflect the impact of path switching on node electrical parameters and the limiting logic of parameter constraints on path selection, providing a quantitative and visual basis for safe switching decisions.

[0059] In this embodiment, the step of generating a multi-objective switching feasibility set that meets safety conditions for the high-load segment set, combined with the power grid physical constraints and scheduling capabilities, specifically involves:

[0060] Based on the high-load transmission segment, a set of candidate paths that have structural bypass or potential electrical channel connections with the segment are identified in the state-structure joint map;

[0061] Based on the aforementioned path candidate set, a multi-objective path evaluation model is constructed that simultaneously includes power flow change deviation, voltage stability index, and node scheduling cost.

[0062] During the path evaluation process, a family of nonlinear constraint functions is constructed, which includes line power limits, voltage upper and lower boundaries, phase angle difference range, and grid-connected equipment topology feasibility. A branch-node switching reachability matrix is ​​generated based on the feasibility region mapping algorithm.

[0063] The reachability matrix is ​​corrected for topology consistency, eliminating topology loops or power flow reversal caused by path combinations, and retaining the set of switchable paths that meet the physical constraints of operation.

[0064] All paths that simultaneously satisfy structural reachability, electrical safety, and scheduling responsiveness, along with their target evaluation results (prioritizing the path with the lowest evaluation result), are included in the multi-target switching feasibility set.

[0065] The multi-objective path evaluation model is specifically as follows:

[0066]

[0067] The family of nonlinear constraint functions specifically includes:

[0068]

[0069] in, This represents the change in system power flow after path switching. This represents the voltage offset of the affected node. The overall response cost of the scheduling nodes required for path switching (such as adjustment delay, power adjustment margin consumption, etc.). For the i-th candidate path, For the active power flow on branch (i,j), For maximum permissible power transmission capacity, The voltage magnitude at node k in the path. Let be the voltage phase angle difference between nodes i and j.

[0070] The establishment of the coupling relationship mapping between the switching path and electrical parameters specifically involves:

[0071] Extract the set of electrical nodes affected by each feasible path in the multi-objective switching feasibility set and their corresponding power change data, and construct the mapping relationship between paths and nodes;

[0072] Based on the active power change values ​​of each path to the relevant nodes under the switching state, the power influence weight of the path on the node is obtained, and a weighted connection relationship between the path and the node is formed.

[0073] The mapping relationship and weighted connection relationship between paths and nodes are organized into a graph, and a path-node-power ternary coupling relationship mapping graph containing path set, node set and power offset weight is constructed.

[0074] In the aforementioned graph, the response of nodes with high power change sensitivity (greater than a preset change threshold) is measured to obtain the switching response intensity index of each node.

[0075] Clustering nodes based on response intensity identifies a subset of sensitive regions with common response characteristics, which supports assigning guiding weights to key nodes in subsequent topology optimization.

[0076] S4, with the minimum load deviation as the optimization objective, constructs a path search model based on the coupling relationship mapping graph, and uses an enhanced iterative algorithm to solve for the optimal topology switching path;

[0077] In this embodiment, the optimization objective is to minimize load deviation. A path search model is constructed based on the coupling relationship mapping graph, and a reinforced iterative algorithm is used to solve for the optimal topology switching path. Specifically:

[0078] Based on the coupling information of path, node and power relationship in the coupling relationship mapping graph, the coupling information is embedded and dimensionality reduced by a 3D graph embedding algorithm to obtain low-dimensional dense vectors of each path in the topology switching space.

[0079] A set of state transition trajectories is constructed based on low-dimensional dense vectors and the current system operating state (including node voltage, load changes, reserve capacity, etc.), and a training sample set is generated in the form of state-action-reward. The state represents the current node and path structure features, the action is the path selection operation, and the reward is calculated by a weighted function of load power variance and node voltage deviation.

[0080] Based on the above training sample set, a path search model is constructed and trained using deep policy gradient to obtain the optimal topology switching path.

[0081] The path search model is specifically as follows:

[0082]

[0083] in, This is the vector of load power changes caused by the current path switching. For power fluctuation variance, Let be the reference voltage value for node j. Let be the voltage amplitude at node j at time t. The total number of nodes. , These are the power deviation penalty coefficient and the voltage deviation penalty coefficient, respectively (obtained through correlation analysis).

[0084] S5. Generate a topology switching sequence based on the optimal topology switching path, and implement power transfer control on the high-load transmission segment.

[0085] In this embodiment, the topology switching sequence is generated based on the optimal topology switching path, specifically as follows:

[0086] Based on the optimal topology switching path output by the path search model, and combined with the topology location, electrical coupling direction and branch action priority of each node in the path, a time-ordered switching instruction set is constructed. A sequential activation strategy is adopted to control the access order of each candidate branch, thus obtaining the topology switching sequence.

[0087] The sequential activation strategy specifically involves prioritizing switching operations that have a smaller impact on system load and are located in terminal branches or low-voltage branches, gradually advancing to the main trunk section, and ensuring that power fluctuations and voltage disturbances during the entire topology transformation process remain within the system's allowable range.

[0088] The priority of branch actions is as follows:

[0089] Power disturbance amplitude priority: The smaller the power change before and after the branch is connected, the higher the priority;

[0090] Priority of node voltage safety margin: Branches with smaller differences from the reference voltage have priority;

[0091] Branch location priority: edge / terminal branches take precedence over trunk sections;

[0092] Backup capacity margin priority: Branches with greater remaining power carrying capacity should be connected first;

[0093] Power flow direction consistency priority: branches consistent with the original power flow direction are given priority;

[0094] Historical stability priority: ranked according to stability in historical operation or simulation;

[0095] Prioritize control response speed: prioritize switching devices with short operation time and fast communication response.

[0096] The power transfer regulation implemented on the high-load transmission segment specifically includes:

[0097] Based on the high-load transmission segment identified in the optimal topology switching path, extract the corresponding power injection node and load node set, and combine the current power flow direction, voltage level and path redundancy to perform power flow disturbance prediction analysis.

[0098] Based on the prediction results, the original switching path is decomposed into multiple micro-switching segments, and a release priority table is constructed according to the conduction path depth, inter-node coupling strength and voltage sensitivity coefficient.

[0099] According to the priority table, segment-by-segment loading control is implemented, and dynamic buffer delay and voltage disturbance threshold are set for each segment. Before execution, the voltage change rate and frequency fluctuation trend of adjacent nodes are continuously monitored.

[0100] When the detected value approaches the threshold, subsequent micro-switching operations are paused and resumed after the system stabilizes, forming a disturbance-mitigated topology switching closed-loop path.

[0101] Meanwhile, during the power transfer process, the backup branch is activated through the guided adjustment path, so that some loads are diverted or temporarily transferred in advance, avoiding concentrated power backflow or local undersupply during the instantaneous switching process, and improving the overall switching stability and controllability.

[0102] The power transfer control for high-load transmission segments further includes: after switching, re-evaluating the load coupling strength and response stability in the load transmission path based on the updated state-structure graph; if there are still abnormal clustering segments, returning to step S3 to dynamically adjust the set of switchable paths to achieve closed-loop optimization of the strategy.

[0103] Furthermore, the automatic switching control method for branch topology based on load balancing also includes introducing a protection adaptation mechanism for boundary nodes during the switching control process, specifically including:

[0104] Before path switching, the protection settings, action logic and main / backup switching configuration of the area covered by the expected switching path are intelligently adapted and verified to avoid protection mismatch or over-level tripping caused by topology reconstruction.

[0105] Example 2, Figure 2 The present invention provides an automatic switching control system for branch topology based on load balancing, comprising the following modules:

[0106] Joint graph construction module: used to collect operating condition data of electrical nodes and construct a joint graph of node operating status and node topology;

[0107] Joint graph construction module: used to collect operating condition data of electrical nodes and construct a joint graph that integrates state and topology;

[0108] Power path extraction module: Based on the joint graph, it extracts the power conduction path from the trunk node to the load node through a weighted path tracing algorithm, and marks the key segments in the load conduction path to form a set of high load segments;

[0109] The switching feasibility analysis module is used to generate a set of multi-objective switching feasibility studies that meet safety conditions for a set of high-load segments, combining the physical constraints of the power grid and the scheduling capabilities, and to establish a mapping diagram of the coupling relationship between switching paths and electrical parameters.

[0110] The path optimization decision module is used to construct a path search model based on the coupling relationship mapping graph with the minimum load deviation as the optimization objective, and to solve the optimal topology switching path using a reinforced iterative algorithm.

[0111] Topology switching and control module: used to generate a topology switching sequence based on the optimal topology switching path and implement power transfer control on high-load transmission segments.

[0112] 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.

[0113] 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.

[0114] 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.

[0115] 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.

[0116] 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.

[0117] 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. An automatic switching control method for branch topology based on load balancing, characterized in that, Includes the following steps: Collect operating data of electrical nodes and construct a joint graph of node operating status and node topology; Based on the joint graph, a weighted path tracing algorithm is used to extract the power conduction path from the backbone node to the load node. Specifically, the backbone node set is identified according to the connectivity of the nodes and the rated capacity of the branches in the joint graph. Starting from the backbone node, a weighted depth-first search algorithm is used to traverse all downstream paths, record the path sequence to the end load node, and accumulate the total weighted value of the path as the path cost. For scenarios where multiple paths lead to the same load node, the unique path with the lowest path cost is selected as the main power conduction path, and a directed main power conduction linked list is established from the backbone node to the load node. Key segments in the load conduction path are marked to form a high-load segment set. Specifically: for each power conduction path, the unit power flow gradient between adjacent node pairs is calculated; all node pairs on the path are selected, historical power fluctuation sequences of the node pairs are constructed, and the cooperative coupling coefficient between any two node pairs is calculated; if multiple consecutive node pairs simultaneously meet the judgment criteria, they are merged into a high-load clustering path segment, and the average power density and concentration index of this path segment are calculated; the top K path segments are selected as priority processing objects based on the weighted values ​​of the average power density and concentration index; all path segments that meet the judgment criteria are marked as key segments to form a high-load segment set; the judgment criteria include cooperative coupling coefficient ≥ preset first threshold, unit power flow gradient ≥ preset second threshold, and cumulative path length ≤ preset clustering radius R. For high-load segments, a set of multi-objective switching feasibility that meets safety conditions is generated by combining the physical constraints and scheduling capabilities of the power grid, and a mapping graph of the coupling relationship between switching paths and electrical parameters is established. With minimum load deviation as the optimization objective, a path search model is constructed based on the coupling relationship mapping graph, and an enhanced iterative algorithm is used to solve for the optimal topology switching path. A topology switching sequence is generated based on the optimal topology switching path, and power transfer control is implemented on the high-load transmission segment.

2. The automatic switching control method for branch topology based on load balancing according to claim 1, characterized in that, The process of collecting operating data from electrical nodes and constructing a joint graph of node operating status and node topology involves the following steps: The operating data of each electrical node is collected by intelligent acquisition terminals deployed in the main and distribution power grid. The operating data includes node voltage, current, active power, topology connection relationship and switch status. Synchronize the asynchronous sampling data between the main grid and the distribution network; Based on the original topology connection matrix of the power grid, the synchronized node operating state and topology structure are jointly encoded into a multi-dimensional state vector through tensor embedding method. Stack multidimensional state vectors according to time series to generate a state-structure joint graph tensor, and output the joint graph.

3. The automatic switching control method for branch topology based on load balancing according to claim 2, characterized in that, The main nodes include the main transformer node, the feeder outlet node, and the main power supply point of the ring network.

4. The automatic switching control method for branch topology based on load balancing according to claim 3, characterized in that, Specifically, for the set of high-load segments, a multi-objective switching feasibility set that meets safety conditions is generated by combining the physical constraints of the power grid and the scheduling capabilities. Based on joint graph identification, a set of candidate paths with structural bypasses or potential electrical channel connections to high-load transmission segments is identified. For the candidate path set, a multi-objective path evaluation model is constructed, which includes power flow change deviation, voltage stability index and node scheduling cost; During the path evaluation process, a set of nonlinear constraint functions is constructed, and a feasibility region mapping algorithm is applied to generate a branch-node switching reachability matrix. Perform topological consistency correction on the reachability matrix: eliminate path combinations that lead to topological loops; Eliminate paths that could cause a reversal in the trend direction; Retain a subset of switchable paths that comply with the physical constraints of power grid operation; All paths that simultaneously satisfy structural accessibility, electrical safety, and scheduling responsiveness, along with their target evaluation results, are included in the multi-target switching feasibility set.

5. The automatic switching control method for branch topology based on load balancing according to claim 4, characterized in that, The establishment of the coupling relationship mapping between the switching path and electrical parameters specifically involves: Extract the set of electrical nodes affected by each feasible path in the multi-objective switching feasibility set and their corresponding power change data, and establish the mapping relationship between paths and nodes; Based on the changes in active power caused to relevant nodes by the path in the switching state, the power impact weight of the path on the node is calculated to form a weighted connection relationship between the path and the node. The mapping relationships and weighted connection relationships are organized into a graph to construct a path-node-power ternary coupling relationship mapping graph containing path sets, node sets, and power offset weights.

6. The automatic switching control method for branch topology based on load balancing according to claim 5, characterized in that, The optimization objective is to minimize load deviation. A path search model is constructed based on the coupling relationship mapping graph, and an enhanced iterative algorithm is used to solve for the optimal topology switching path. Specifically: Based on the coupling relationship between paths, nodes and power offset in the coupling relationship mapping graph, a 3D graph embedding algorithm is used to perform dimensionality reduction to generate low-dimensional dense vectors for each path in the topology switching space. Based on the current system operating state, a set of state transition trajectories is constructed using the low-dimensional dense vectors, and then transformed into a training sample set in the form of state-action-reward triples; Based on the training sample set, the input path search model is trained using a deep policy gradient algorithm to output the optimal topology switching path.

7. The automatic switching control method for branch topology based on load balancing according to claim 6, characterized in that, The process of generating a topology switching sequence based on the optimal topology switching path specifically involves: Based on the optimal topology switching path output by the path search model, the topological positional relationship and electrical coupling direction of each node in the path are analyzed. Based on the electrical coupling direction and branch action priority rules, a time-ordered switching instruction set is generated; A sequential activation strategy is adopted to control the access order of candidate branches, forming the final topology switching sequence.

8. The automatic switching control method for branch topology based on load balancing according to claim 7, characterized in that, The power transfer regulation implemented on the high-load transmission segment specifically includes: Based on the optimal topology switching path, the set of power injection nodes and load nodes corresponding to the high-load transmission segment is extracted, and power flow disturbance is predicted by combining power flow direction, voltage level and path redundancy. The switching path is decomposed into multiple micro-switching segments, and a release priority table is constructed according to the conduction path depth, node coupling strength and voltage sensitivity coefficient. Load control is implemented segment by segment according to the priority table: dynamic buffer delay and voltage disturbance threshold are set for each segment; the voltage change rate and frequency fluctuation trend of adjacent nodes are monitored in real time; if the monitored value is close to the threshold, the switching operation is paused and resumed after the system stabilizes. Guide the activation of backup branches to divert some loads, avoiding concentrated power backflow or localized power shortages.

9. The system of the automatic switching control method for branch topology based on load balancing according to claim 8, characterized in that, Includes the following modules: Joint graph construction module: used to collect operating condition data of electrical nodes and construct a joint graph of node operating status and node topology; Power path extraction module: Based on the joint graph, it extracts the power conduction path from the trunk node to the load node through a weighted path tracing algorithm, and marks the key segments in the load conduction path to form a set of high load segments; The switching feasibility analysis module is used to generate a set of multi-objective switching feasibility studies that meet safety conditions for a set of high-load segments, combining the physical constraints of the power grid and the scheduling capabilities, and to establish a mapping diagram of the coupling relationship between switching paths and electrical parameters. The path optimization decision module is used to construct a path search model based on the coupling relationship mapping graph with the minimum load deviation as the optimization objective, and to solve the optimal topology switching path using a reinforced iterative algorithm. Topology switching and control module: used to generate a topology switching sequence based on the optimal topology switching path and implement power transfer control on high-load transmission segments.

Images