A knowledge graph-based main and distribution network collaborative planning method, system, device and storage medium

Abstract

The application discloses a kind of based on knowledge graph's main distribution network collaborative planning method, system, equipment and storage medium, comprising: based on main distribution network physical topology structure, source load distribution information and planning rule, constructs main distribution network fusion knowledge graph;Based on knowledge graph, each load node and new energy access point are time sequence state prediction, generate the load power prediction sequence and new energy output prediction sequence mounted on graph node, form time sequence state graph;Based on time sequence state graph simulation extreme disturbance scene, deduce the node power state after disturbance, and according to power fluctuation amplitude before and after disturbance, filter out high-risk node set;Based on high-risk node set and time sequence state graph, combine the load state of power supply path and power balance constraint, calculate main distribution network collaborative optimization scheduling strategy, to adjust each node power distribution and path power flow, output final collaborative planning result.Can dynamically identify potential risk in power grid, reduce power grid instability risk.

Description

Technical Field

[0001] This invention relates to the technical field of power distribution planning, and in particular to a method, system, device, and storage medium for collaborative planning of main and distribution networks based on knowledge graphs. Background Technology

[0002] With the increasing complexity of power systems and the rapid integration of new energy sources, traditional power grid planning methods have gradually revealed many problems, especially in handling the coordinated planning of primary and distribution networks, load forecasting, and risk assessment. Current power grid planning mainly relies on manually set rules and simple mathematical models, which makes planning inefficient, optimization-capable, and unable to adapt to the needs of rapid changes and complex extreme events in the power system.

[0003] Furthermore, insufficient coordination between the distribution network and the main grid prevents comprehensive planning of loads and resources within the region, leading to resource waste and potential grid security risks. To address this challenge, a smart planning system capable of dynamically adapting to grid changes and effectively integrating source and load information is needed. This system should not only possess powerful data processing capabilities but also be able to combine the physical topology and actual operating status of the grid, based on time-series forecasting, extreme disturbances, and risk assessment, to provide globally optimal planning solutions. Summary of the Invention

[0004] In view of the aforementioned existing problems, this invention is proposed. Therefore, this invention provides a knowledge graph-based method, system, device, and storage medium for collaborative planning of primary and secondary power grids to address the current problems of low planning efficiency and weak optimization capabilities.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: In a first aspect, embodiments of the present invention provide a knowledge graph-based method for collaborative planning of a main distribution network, comprising: constructing a fusion knowledge graph of the main distribution network based on the physical topology of the main distribution network, source-load distribution information, and planning rules, wherein the knowledge graph includes a set of equipment nodes, a set of power supply edges, and a set of planning rules; Based on the knowledge graph, the time-series state prediction of each load node and new energy access point is performed. In the prediction process, power supply path smoothing constraints and planning rule penalty mechanisms are introduced to generate load power prediction sequences and new energy output prediction sequences attached to the graph nodes, forming a time-series state graph. Based on the time-series state map, extreme disturbance scenarios are simulated, the node power state after the disturbance is deduced, and a set of high-risk nodes is selected according to the power fluctuation amplitude before and after the disturbance. Based on the set of high-risk nodes and the time-series state map, combined with the load status and power balance constraints of the power supply path, a collaborative optimization scheduling strategy for the main and distribution networks is calculated to adjust the power allocation of each node and the power flow of the path, and the final collaborative planning result is output.

[0006] As a preferred embodiment of the knowledge graph-based main distribution network collaborative planning method described in this invention, the construction of the main distribution network fusion knowledge graph further includes: The set of equipment nodes includes main substations, distribution feeders, load areas and new energy access points. The basic attributes of the equipment nodes include rated capacity, operating status and spatial coordinates. The power supply edge set represents the direction of power flow. The power supply edge is established based on the actual connection relationship of the power grid. The power supply path length between the load node and the main substation is calculated by the shortest path algorithm to determine whether the power supply radius constraint is met. The set of planning rules is bound to the target node or power supply path.

[0007] As a preferred embodiment of the knowledge graph-based main and distribution network collaborative planning method described in this invention, the method includes: predicting the time-series status of each load node and new energy access point based on the knowledge graph, which includes: A load power time series prediction model is constructed by minimizing historical prediction errors, power supply path smoothing constraints and planning rule penalty mechanisms. The load power time series prediction model uses historical load data, meteorological data and social activity data as input features to solve the predicted power sequence of each load node. The power prediction sequences of each load node obtained by solving are combined with the power summary prediction sequence of the main substation, and then used as attribute data to be mounted on the corresponding nodes and power supply paths of the knowledge graph. The new energy output prediction is based on meteorological data and equipment ledger information. The output power of the new energy access point is predicted and calculated in a time series. The calculated output prediction sequence of each new energy access node is then attached to the corresponding new energy access node attribute in the knowledge graph. The knowledge graphs of the existing load power prediction sequence and the new energy output prediction sequence are integrated to form a time-series state graph.

[0008] As a preferred embodiment of the knowledge graph-based main and distribution network collaborative planning method described in this invention, the method includes: simulating extreme disturbance scenarios based on the time-series state graph and deducing the node power state after the disturbance, including: Based on historical power grid fault data, common equipment fault modes, and external environmental factors, various extreme disturbance scenarios are defined, including main substation faults, feeder disconnections, load surges, and voltage drops. For each extreme disturbance scenario, a corresponding disturbance amount is set, the predicted power value of each node under the undisturbed condition is obtained from the time-series state graph, the disturbance amount is applied to the corresponding node, and the node power value after disturbance is calculated. Based on the power grid topology and the electrical connection strength between nodes, the power flow intensity of each power supply path is calculated using the disturbed node power values.

[0009] As a preferred embodiment of the knowledge graph-based main and distribution network collaborative planning method described in this invention, the method of selecting a set of high-risk nodes based on the power fluctuation amplitude before and after the disturbance includes: calculating the risk level of each node under extreme disturbance, wherein the risk level is the ratio of the node power change amplitude to its rated capacity; Sort all nodes by risk level, select the top n nodes with the highest risk level to form a set of high-risk nodes, and output them as potential risk nodes of the power grid.

[0010] As a preferred embodiment of the knowledge graph-based main and distribution network collaborative planning method described in this invention, the main and distribution network collaborative optimization scheduling strategy is calculated based on the set of high-risk nodes and the time-series state graph, combined with the load status and power balance constraints of the power supply path, including: With the goal of minimizing node power scheduling deviation under extreme disturbances and the constraints of load status and power balance of power supply path, an objective function for the coordinated optimization of main and distribution networks is constructed. A mixed-integer programming algorithm is used to solve the problem by combining the topology of the time-series state graph, the set of high-risk nodes, and the objective function and constraints. By dynamically adjusting the power allocation of each load node, main transformer node, and new energy access node, and reconstructing the power flow direction of some power supply paths, the optimal main and distribution network collaborative optimization scheduling scheme is obtained.

[0011] As a preferred embodiment of the knowledge graph-based main and distribution network collaborative planning method described in this invention, the collaborative planning results include: optimized power scheduling values ​​for each node, adjusted power flow information for critical paths, and response strategies for risk nodes.

[0012] Secondly, the present invention provides a knowledge graph-based main distribution network collaborative planning system, comprising: The knowledge graph construction module is used to construct a fusion knowledge graph of the main and distribution networks based on the physical topology, source-load distribution information, and planning rules of the main and distribution networks. The knowledge graph includes a set of equipment nodes, a set of power supply edges, and a set of planning rules. The time-series state prediction module is used to predict the time-series state of each load node and new energy access point based on the knowledge graph. In the prediction process, a power supply path smoothing constraint and planning rule penalty mechanism are introduced to generate a load power prediction sequence and a new energy output prediction sequence attached to the graph nodes, forming a time-series state graph. The disturbance inference module is used to simulate extreme disturbance scenarios based on the time-series state map, infer the node power state after the disturbance, and screen out a set of high-risk nodes based on the power fluctuation amplitude before and after the disturbance. The strategy generation module is used to calculate the main and distribution network collaborative optimization scheduling strategy based on the set of high-risk nodes and the time-series state map, combined with the load status and power balance constraints of the power supply path, so as to adjust the power allocation of each node and the power flow of the path, and output the final collaborative planning result.

[0013] Thirdly, the present invention provides an electronic device, comprising: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, they implement the steps of the knowledge graph-based main distribution network collaborative planning method.

[0014] Fourthly, the present invention provides a computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the steps of the knowledge graph-based main distribution network collaborative planning method.

[0015] Compared with existing technologies, the beneficial effects of this invention are as follows: By constructing a knowledge graph that integrates power grid physical topology, planning rules, load, and renewable energy data, this invention provides a unified structured foundation for the power grid, ensuring that planning and operational decisions are based on real-time and accurate forecast data. Building upon load forecasting and renewable energy output forecasting, this invention further introduces power supply path smoothing terms and rule penalty terms, ensuring that the forecast results not only conform to the physical constraints of the power grid but also follow various planning rules, thereby improving the accuracy and stability of power grid dispatching. Based on this structured time-series state graph, through extreme disturbance simulation and risk node identification, this invention can dynamically identify potential risks in the power grid and propose targeted optimization strategies to minimize the risk of power grid instability. By employing a mixed-integer programming method, this invention can achieve coordinated optimization of power grid power dispatching and risk control, thereby providing safer and more reliable power grid operation guarantees in the face of extreme scenarios and enhancing the overall resilience of the power grid. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein: Figure 1 This is a schematic diagram of the process flow of a knowledge graph-based main and distribution network collaborative planning method according to an embodiment of the present invention. Detailed Implementation

[0017] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0018] Example 1, referring to Figure 1 As one embodiment of the present invention, this embodiment provides a knowledge graph-based main distribution network collaborative planning method, including: S100: Based on the physical topology of the main distribution network, source and load distribution information, and planning rules, a knowledge graph of the main distribution network is constructed. The knowledge graph includes a set of equipment nodes, a set of power supply edges, and a set of planning rules. S200: Based on the knowledge graph, the time-series state prediction of each load node and new energy access point is performed. In the prediction process, the power supply path smoothing constraint and planning rule penalty mechanism are introduced to generate the load power prediction sequence and new energy output prediction sequence attached to the graph node, forming a time-series state graph. S300: Simulates extreme disturbance scenarios based on time-series state graphs, infers the node power state after the disturbance, and selects a set of high-risk nodes based on the power fluctuation amplitude before and after the disturbance. S400: Based on the high-risk node set and time sequence state map, combined with the load state and power balance constraints of the power supply path, calculate the main and distribution network collaborative optimization scheduling strategy to adjust the power allocation of each node and the power flow of the path, and output the final collaborative planning result.

[0019] It should be noted that current power grid planning mainly relies on manually set rules and simple mathematical models, resulting in low planning efficiency, weak optimization capabilities, and difficulty in adapting to the needs of rapid changes and complex extreme events in the power system. Furthermore, insufficient coordination between the distribution network and the main grid prevents comprehensive planning of loads and resources within the region, leading to resource waste and potential power grid security risks. This invention constructs a knowledge graph integrating the physical topology of the main and distribution networks, source-load distribution, and planning rules, unifying the modeling of equipment nodes, power supply relationships, and constraint rules. Secondly, based on the graph, it predicts the temporal state of load and renewable energy output, introducing power supply path smoothing constraints and rule penalty mechanisms to ensure that the prediction results conform to the actual operation of the power grid. Thirdly, it simulates extreme disturbance scenarios to extrapolate the power grid's operating state and identify high-risk nodes. Finally, combining path load states and power balance constraints, it uses optimization algorithms to generate a coordinated scheduling strategy for the main and distribution networks, outputting structured planning results. This enables coordination between the main and distribution networks throughout the entire process of prediction, risk assessment, and optimization decision-making, improving the power grid's adaptability and overall resilience to renewable energy integration and extreme events.

[0020] In this embodiment of the invention, step S100, which involves constructing a knowledge graph for the convergence of the main and distribution networks, further includes: The set of equipment nodes includes main substations, distribution feeders, load areas, and new energy access points. The basic attributes of equipment nodes include rated capacity, operating status, and spatial coordinates. The power supply edge set represents the direction of power flow. The power supply edge is established based on the actual connection relationship of the power grid. The power supply path length between the load node and the main substation is calculated by the shortest path algorithm to determine whether the power supply radius constraint is met. The planning rule set is bound to the target node or power supply path.

[0021] Specifically, a knowledge graph capable of expressing the structure of the main and distribution networks, source-load distribution, and planning constraints is constructed to serve as a unified data and logical foundation for subsequent operational status prediction, extreme scenario simulation, and planning optimization. The knowledge graph needs to simultaneously depict the physical connections between the main grid, distribution network, loads, and distributed power sources, and embed planning rules to give it both structural and rule-based semantic capabilities.

[0022] Furthermore, regarding input data, the first step is to collect equipment ledgers and topology data for the main and distribution networks. This data originates from the power company's geographic information system and dispatch automation system, including the equipment attributes and connection relationships of main substations, transmission lines, distribution feeders, switches, busbars, and transformer substations. Secondly, source-load data needs to be collected. This data comes from the marketing management system and the new energy access database, mainly including load information for regional load centers, industrial parks, and large public buildings, as well as the spatial location and capacity of access points such as wind farms, photovoltaic power stations, and distributed energy storage. Finally, planning guidelines and technical specifications need to be collected. This content comes from the power company's internal regulations and national standard documents. Key clauses in the text need to be extracted manually or semi-automatically and converted into a rule base, such as power supply radius requirements, feeder length limits, equipment capacity utilization limits, and renewable energy access ratio constraints.

[0023] Furthermore, when constructing the knowledge graph, various objects in the power grid are abstracted into sets of nodes. The main substation is designated as... Distribution lines or feeders are denoted as The load area is denoted as The new energy access point is denoted as The power supply relationships between nodes constitute an edge set. , using directed edges Indicates that electrical energy flows from the node Flow to Node The basic attributes that come with each node include rated capacity. Operating status and spatial coordinates For example, when the data exported by the GIS system indicates that "a certain 220 kV substation supplies power to the 110 kV bus M1 via lines L1 and L2, and then supplies power to the industrial area I1 via feeder F1," it can be represented as a node in the map. , , , , and establish directed edges , , , This forms a complete power supply path.

[0024] Furthermore, after the set of nodes and edges is constructed, the planning rules need to be embedded to form a set of constraints. These rules are bound to nodes or paths and used for subsequent operational status analysis and optimization modeling. For example, the service radius constraint of the main substation can be expressed as: in, Indicates load node With the main transformer node The shortest power supply path length between nodes is determined by the node coordinates. and line edge collection Calculated using the shortest path algorithm; This indicates the maximum service radius specified in the planning regulations, such as 10 kilometers; Indicates the change from the primary node The set of load nodes supplying power. Using this constraint method, if the shortest path distance from a certain load node to the main substation exceeds... The map will then mark the path as not meeting the planning specifications.

[0025] In one feasible approach, in addition to the service radius, embodiments of the invention may also embed capacity utilization constraints, such as the actual load power of the main substation. It must not exceed its rated capacity. 90%, formally expressed as: in, This represents the active load power of the main transformer node, calculated by the load forecasting system. The rated capacity of the main transformer node is indicated by the equipment ledger data in the SCADA system. This constraint allows the graph to identify whether the main transformer is at risk of overload.

[0026] The resulting knowledge graph can be represented as: in, Represents the set of all nodes. Represents the set of power supply edges. Represents a set of rules.

[0027] It is worth noting that, unlike traditional network models that only describe the structure, this knowledge graph can reflect the physical topology of the power grid and also carry planning rules, providing a unified data carrier for subsequent state prediction, risk simulation and optimization decision-making.

[0028] It should be noted that the goal of step S200 is to utilize the constructed structural knowledge graph to attach the power grid load forecast and renewable energy output forecast to node attributes, forming a time-series state-based graph, which provides a foundation for subsequent extreme scenario simulation and planning optimization. This process combines the graph structure with the time-series forecast results through power supply path smoothing terms and rule penalty terms, ensuring that the predicted values ​​not only conform to the structural relationships of the power grid but also comply with the key constraints in the planning.

[0029] In this embodiment of the invention, step S200, which involves predicting the time-series status of each load node and renewable energy access point based on a knowledge graph, includes: A load power time series prediction model is constructed by minimizing historical prediction errors, power supply path smoothing constraints and planning rule penalty mechanisms. The load power time series prediction model uses historical load data, meteorological data and social activity data as input features to solve the predicted power sequence of each load node. The power prediction sequences of each load node obtained by solving are combined with the power summary prediction sequence of the main substation, and they are used as attribute data to be mounted on the corresponding nodes and power supply paths of the knowledge graph. Specifically, in the load forecasting section, this invention provides a method for each load node. A time-series load forecasting model is established. Load forecasting relies on historical load data, meteorological data (temperature, humidity, etc.), and social activity data (holidays, etc.), which are sourced from the marketing management system, meteorological data interface, and holiday database, respectively. A time-series forecasting method based on Long Short-Term Memory (LSTM) networks is employed to capture the periodic fluctuations and seasonal variations in load. To enhance the coordination between forecasting and the power grid structure, the forecasting objective function not only considers traditional forecasting errors but also incorporates a smoothing term for the power supply path and a penalty term for rules. The total loss function of the forecast is as follows: ; in, Indicates load node At any moment The predicted power; For load nodes At any moment The historical actual power output comes from marketing or distribution automation systems; This is the power supply path smoothing coefficient, which controls the difference in predictions between adjacent electrical nodes; The edge weight represents the strength of the power supply relationship between node pairs, taking into account the impedance and current capacity of the equipment during calculation; This is the coefficient for the rule penalty term, used to adjust the cost of violating planning rules; master variable node At any moment The total power This is the upper limit coefficient for capacity utilization. It is the sum of the power forecasts for all load nodes; The rated capacity of the main transformer node is provided by the scheduling system.

[0030] After load forecasting is completed, the power forecast sequence for each load node will be generated. Power aggregation prediction sequence with main substation These time-series data are combined and attached as attributes to the graph nodes and power supply paths.

[0031] It should be noted that the objective function has the following functions: the first term is the traditional load forecasting error, which minimizes the historical forecasting error; the second term is the path smoothing term, which ensures that the power flow of the power supply path conforms to the physical constraints of the power grid by constraining the differences in forecasts of adjacent nodes; and the third term is the rule penalty term, which is used to penalize forecasts that violate the main transformer capacity constraints to avoid overload situations.

[0032] Furthermore, in step S200, the power output prediction of new energy sources is based on meteorological data and equipment ledger information. The output power of the new energy access points is predicted and calculated in a time series. The calculated power output prediction sequence of each new energy access node is then attached to the corresponding new energy access node attribute in the knowledge graph. The knowledge graphs of the existing load power prediction sequence and the new energy output prediction sequence are integrated to form a time-series state graph.

[0033] Specifically, for new energy access nodes This invention employs a similar time-series forecasting method, utilizing meteorological data such as sunlight and wind speed, along with output capacity data from photovoltaic / wind power equipment, to predict the output power of new energy sources. Taking a photovoltaic power station as an example, its predicted power can be calculated using the following formula, which calculates the output power of the new energy access node based on factors such as sunlight intensity and module efficiency: in, Represents new energy nodes At any moment The predicted output; The efficiency of photovoltaic modules comes from the equipment ledger; For a moment The intensity of sunlight is provided by the meteorological system. The area of ​​the photovoltaic modules is from the equipment ledger.

[0034] It should be noted that, through the aforementioned load forecasting and renewable energy output forecasting models, the time-series forecast values ​​for each node are obtained and then attached to the graph nodes and paths. Finally, the time-series state graph is generated. It not only describes the static topology of the power grid, but also incorporates load changes and new energy output changes over a period of time, becoming a state-based graph.

[0035] Furthermore, the goal of step S300 is to generate a temporal state map. Extreme disturbance simulations, by simulating different types of extreme disturbances, extrapolate the operating state of the power grid under extreme conditions and identify potential risk nodes. Through simulations of various extreme events such as equipment failures, load surges, and natural disasters, the analysis identifies key nodes that could lead to grid instability or overload, providing a basis for subsequent planning optimization and safety enhancement. Extreme disturbance simulation is a crucial step in evaluating the performance of the power grid under extreme conditions, with the ultimate goal of ensuring the grid can maintain stable operation in the face of unforeseen events, thereby enhancing the grid's resilience.

[0036] In this embodiment of the invention, step S300, which simulates extreme disturbance scenarios based on time-series state maps and deduces the node power state after the disturbance, includes: Based on historical power grid fault data, common equipment fault modes, and external environmental factors, various extreme disturbance scenarios are defined, including main substation faults, feeder disconnections, load surges, and voltage drops. For each extreme disturbance scenario, a corresponding disturbance amount is set, the predicted power value of each node under the undisturbed condition is obtained from the time-series state graph, the disturbance amount is applied to the corresponding node, and the node power value after disturbance is calculated. Based on the power grid topology and the electrical connection strength between nodes, the power flow intensity of each power supply path is calculated using the disturbed node power values.

[0037] Specifically, the input in step S300 comes from the generated stateful graph. It contains time-series forecast data for each node in the power grid, including power forecast sequences for load nodes. Summary power prediction sequence of main transformer node And the output prediction sequence of new energy access nodes Furthermore, the power flow direction and intensity markings along the power supply path were also incorporated into the graph, providing essential information for the extreme disturbance simulation in this step. In particular, power prediction for load nodes was included. Combined power of the main substation All of these are derived from the time-series prediction model in step two, while the output prediction of the renewable energy access node... It is also based on the relevant prediction results in step S200.

[0038] Furthermore, in the extreme disturbance simulation process, it is first necessary to define possible extreme disturbance scenarios. These scenarios are constructed based on historical grid fault data, common equipment failure modes, and external factors such as weather. Common extreme events include main substation faults, feeder disconnections, load surges, and voltage dips. Each extreme event is simulated by applying different disturbance amounts. Predicted power to the spectral node The above is used to simulate the impact of the event on the power grid nodes. Disturbance quantity This represents the magnitude of the change in a node's state at a specific moment. Specifically, each node... At any moment Power after disturbance It can be represented as: in, Represents a node At any moment The power value after the disturbance This is the predicted power value under undisturbed conditions. This is the disturbance quantity. The settings are based on specific extreme scenarios.

[0039] For example, in the event of a main substation fault, the main transformer power output can be set to be momentarily lost. Let's assume... This simulates the power loss caused by the shutdown of the main substation.

[0040] Furthermore, after the node disturbances are applied, it is necessary to calculate the power flow of each power supply path using the power grid topology. , indicating from node To the node The power flow intensity. Since the power flow between nodes in a power grid is affected by various factors such as line impedance and equipment capacity, the electrical connection strength of each path needs to be considered when calculating the power flow. This reflects the load capacity of the path. The calculation is performed using known power supply paths in the graph. and the power value after disturbance Based on the physical properties of the electrical connection, the power flow variation along the path is estimated, and the power flow along the power supply path is calculated. The calculation formula is: in, Indicates time From node To the node The power flow intensity, It is with nodes Adjacent nodes The power value after the disturbance Path connectivity strength represents the capability of electrical connections between nodes.

[0041] In this embodiment of the invention, step S300, which involves selecting a set of high-risk nodes based on the power fluctuation amplitude before and after the disturbance, includes: calculating the risk level of each node under extreme disturbance, where the risk level is the ratio of the node's power change amplitude to its rated capacity. Sort all nodes by risk level, select the top n nodes with the highest risk level to form a set of high-risk nodes, and output them as potential risk nodes of the power grid.

[0042] Furthermore, by calculating the power flow after all disturbances, this embodiment of the invention further analyzes the stability of nodes, thereby identifying potential risk nodes in the power grid. To this end, this invention defines risk levels. This is used to quantify the security of nodes under extreme disturbances. Risk Level Measured the nodes The power variation range is related to its rated capacity The relationship between them. The specific formula is: in, For nodes The risk level, For nodes Rated capacity, This represents the difference between the power after the disturbance and the predicted power.

[0043] It is worth noting that this invention calculates the risk level of all nodes. This allows us to identify nodes that exhibit significant power fluctuations under extreme events; these nodes are considered potential risk nodes. By ranking the nodes by risk level, the nodes with the highest risk level are selected. These nodes are identified as risk nodes that require special attention.

[0044] Furthermore, the goal of step S400 is based on the output set of risk nodes. Combined with the time-series state map generated in step two This system generates a primary-distribution coordinated optimization strategy and provides a decision-making basis for power grid scheduling and planning through structured output. By identifying potential risk nodes and utilizing node attributes and power supply path information in the power grid graph, it optimizes the operation and scheduling of the power grid in the face of extreme events, minimizing the negative impact of extreme scenarios such as equipment failure or load surges on power grid stability. This strategy not only improves the responsiveness and stability of the power grid but also provides an optimization scheme for subsequent real-time operation and scheduling.

[0045] In this embodiment of the invention, step S400, based on the high-risk node set and time-series state map, and combined with the load state and power balance constraints of the power supply path, calculates the main and distribution network collaborative optimization scheduling strategy, including: With the goal of minimizing node power scheduling deviation under extreme disturbances and the constraints of load status and power balance of power supply path, an objective function for the coordinated optimization of main and distribution networks is constructed. A mixed-integer programming algorithm is adopted, which combines the topological structure of the time-series state graph, the set of high-risk nodes, and the objective function and constraints to solve the problem; By dynamically adjusting the power allocation of each load node, main transformer node, and new energy access node, and reconstructing the power flow direction of some power supply paths, the optimal main and distribution network collaborative optimization scheduling scheme is obtained.

[0046] Specifically, in this embodiment of the invention, the input for step S400 includes a set of risk nodes. It includes key nodes that could lead to power grid instability under extreme events, as well as time-series state graphs. The latter provides spatiotemporal data support for the time-varying structure of the power grid and the predicted power at nodes. Specifically, the predicted power values ​​at nodes... Total power value of the main substation Power output forecast of new energy nodes All of these originate from time-series prediction models, while the risk node set... These results are derived through extreme disturbance simulations and risk level calculations, and the power flow changes of these nodes under extreme conditions will be a key focus.

[0047] Based on this, the embodiments of the present invention first use current density To measure the load on the power supply path. Current density. This represents a path. The ratio of power flow to rated capacity. Calculating this value reveals the carrying capacity of certain paths in the power grid under high load conditions and identifies paths where overload may occur. The calculation formula is as follows: in, Indicates from node To the node At any moment power flow, This is the rated capacity of the path.

[0048] It should be noted that by calculating the current density of all power supply paths, it is possible to identify which paths have power flows close to or exceed their carrying capacity under extreme load conditions, and thus identify them as critical paths that require special attention.

[0049] In one feasible approach, besides current density, power balance constraints are another key element ensuring grid stability. In extreme scenarios, the power input and output of the grid must be balanced. This invention optimizes node power scheduling by minimizing the difference between the power after disturbance and the predicted power; the objective function is expressed as: in, For the nodes after the disturbance At any moment The power value, This represents the predicted power under normal conditions.

[0050] It should be noted that this objective function can optimize the power scheduling of each node, ensuring that the load change of the power grid does not exceed the capacity limit of the node when extreme events occur, thereby preventing the node from overloading or experiencing power imbalance.

[0051] Furthermore, based on current density and power balance constraints, this embodiment of the invention calculates the optimal power grid dispatch scheme using a mixed-integer programming algorithm. The mixed-integer programming algorithm considers the power grid topology and risk nodes during the solution process, and can dynamically adjust the power allocation of the power grid according to different extreme disturbances, minimizing the risk of power grid instability. By optimizing the dispatch strategy, it is possible to ensure that the power grid remains stable in the face of sudden events, and to take necessary dispatch measures in a timely manner at risk nodes to avoid overload.

[0052] Furthermore, based on current density and power balance constraints, this embodiment of the invention constructs a hybrid integer programming model for coordinated dispatching of main and distribution systems, using the set of risk nodes, the power flow changes of each power supply path, the rated capacity of nodes, the adjustable output range of main transformers, the operable status of tie switches, and the charge / discharge range of energy storage devices as joint inputs. Continuous decision quantities include main transformer output adjustment, energy storage charge / discharge power, and transferable load; discrete decision quantities include the on / off status of tie switches and the switching status of power supply paths. Optimization objectives include reducing the overload of risk nodes, minimizing the loss of unsupplied loads, especially critical loads, and controlling the dispatching costs associated with network reconfiguration and power adjustment. Constraints include node power balance constraints, line transmission capacity constraints, main transformer capacity constraints, energy storage operation boundary constraints, and network connectivity constraints. By solving the hybrid integer programming model, the optimal coordinated dispatching scheme for the corresponding extreme disturbance scenario is obtained.

[0053] In practice, after identifying different extreme disturbance scenarios, the first step is to lock down the failed power supply path based on the location of the faulty equipment and the distribution of risk nodes, and then re-select feasible transfer paths based on the topological connections in the graph. Subsequently, according to the risk level, the distribution of important loads and the remaining carrying capacity of the path, the output allocation of the main transformer, the opening and closing status of the tie switch, the energy storage support power and the load transfer sequence are adjusted in sequence to ensure that important loads are restored to power first and general loads are restored gradually within the allowable range of remaining capacity. For power supply paths that still have a local overload trend, the load transfer ratio of the corresponding branch is further reduced or the output of the adjacent supporting power source is increased, thereby suppressing the spread of overload to adjacent nodes and ensuring that the power grid maintains a feasible, stable and recoverable operating state under extreme events.

[0054] In this embodiment of the invention, the collaborative planning results in step S400 include: the optimized power scheduling value of each node, the adjusted power flow information of the critical path, and the response strategy for risk nodes.

[0055] It should be noted that the optimized grid collaborative dispatch strategy based on risk nodes, by adjusting the power dispatch of each node in the grid, can minimize grid instability caused by extreme events. Furthermore, the output structured planning results provide detailed power adjustment values ​​and power flow information for power supply paths for grid operation and dispatch. These results provide power companies with decision-making support, helping staff to effectively respond to extreme events, avoid grid instability or overload, and thus improve the overall resilience and security of the grid.

[0056] Example 2: The above example is an illustrative scheme of a knowledge graph-based main and distribution network collaborative planning method. It should be noted that the technical solution of this knowledge graph-based main and distribution network collaborative planning system belongs to the same concept as the technical solution of the knowledge graph-based main and distribution network collaborative planning method described above. Details not described in detail in this example can be found in the description of the knowledge graph-based main and distribution network collaborative planning method described above.

[0057] This embodiment presents a knowledge graph-based main distribution network collaborative planning system, comprising: The knowledge graph construction module is used to construct a fusion knowledge graph of the main and distribution networks based on the physical topology, source and load distribution information, and planning rules of the main and distribution networks. The knowledge graph includes a set of equipment nodes, a set of power supply edges, and a set of planning rules. The time-series state prediction module is used to predict the time-series state of each load node and new energy access point based on the knowledge graph. In the prediction process, the power supply path smoothing constraint and planning rule penalty mechanism are introduced to generate the load power prediction sequence and new energy output prediction sequence attached to the graph nodes, forming a time-series state graph. The disturbance inference module is used to simulate extreme disturbance scenarios based on time-series state maps, infer the node power state after the disturbance, and screen out a set of high-risk nodes based on the power fluctuation amplitude before and after the disturbance. The strategy generation module is used to calculate the main and distribution network collaborative optimization scheduling strategy based on the high-risk node set and time sequence state map, combined with the load state and power balance constraints of the power supply path, so as to adjust the power allocation of each node and the power flow of the path, and output the final collaborative planning result.

[0058] This embodiment also provides an electronic device applicable to the main distribution network collaborative planning method based on knowledge graphs, including: The system includes a memory and a processor. The memory stores computer-executable instructions, and the processor executes these instructions to implement the knowledge graph-based main and distribution network collaborative planning method proposed in the above embodiments.

[0059] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the knowledge graph-based main and distribution network collaborative planning method as proposed in the above embodiments.

[0060] The storage medium proposed in this embodiment and the knowledge graph-based main and distribution network collaborative planning method proposed in the above embodiments belong to the same inventive concept. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0061] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0062] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A knowledge graph-based collaborative planning method for main and distribution networks, characterized in that, include: Based on the physical topology of the main and distribution networks, source and load distribution information, and planning rules, a knowledge graph of the main and distribution networks is constructed. The knowledge graph includes a set of equipment nodes, a set of power supply edges, and a set of planning rules. Based on the knowledge graph, the time-series state prediction of each load node and new energy access point is performed. In the prediction process, power supply path smoothing constraints and planning rule penalty mechanisms are introduced to generate load power prediction sequences and new energy output prediction sequences attached to the graph nodes, forming a time-series state graph. Based on the time-series state map, extreme disturbance scenarios are simulated, the node power state after the disturbance is deduced, and a set of high-risk nodes is selected according to the power fluctuation amplitude before and after the disturbance. Based on the set of high-risk nodes and the time-series state map, combined with the load status and power balance constraints of the power supply path, a collaborative optimization scheduling strategy for the main and distribution networks is calculated to adjust the power allocation of each node and the power flow of the path, and the final collaborative planning result is output.

2. The knowledge graph-based main and distribution network collaborative planning method as described in claim 1, characterized in that, The construction of the integrated knowledge graph of the main and distribution networks also includes: The set of equipment nodes includes main substations, distribution feeders, load areas and new energy access points. The basic attributes of the equipment nodes include rated capacity, operating status and spatial coordinates. The power supply edge set represents the direction of power flow. The power supply edge is established based on the actual connection relationship of the power grid. The power supply path length between the load node and the main substation is calculated by the shortest path algorithm to determine whether the power supply radius constraint is met. The set of planning rules is bound to the target node or power supply path.

3. The knowledge graph-based main and distribution network collaborative planning method as described in claim 2, characterized in that, Based on the knowledge graph, time-series state prediction for each load node and renewable energy access point includes: A load power time series prediction model is constructed by minimizing historical prediction errors, power supply path smoothing constraints and planning rule penalty mechanisms. The load power time series prediction model uses historical load data, meteorological data and social activity data as input features to solve the predicted power sequence of each load node. The power prediction sequences of each load node obtained by solving are combined with the power summary prediction sequence of the main substation, and then used as attribute data to be mounted on the corresponding nodes and power supply paths of the knowledge graph. The new energy output prediction is based on meteorological data and equipment ledger information. The output power of the new energy access point is predicted and calculated in a time series. The calculated output prediction sequence of each new energy access node is then attached to the corresponding new energy access node attribute in the knowledge graph. The knowledge graphs of the existing load power prediction sequence and the new energy output prediction sequence are integrated to form a time-series state graph.

4. The knowledge graph-based main and distribution network collaborative planning method as described in claim 3, characterized in that, Based on the aforementioned time-series state map, extreme disturbance scenarios are simulated, and the node power state after the disturbance is deduced, including: Based on historical power grid fault data, common equipment fault modes, and external environmental factors, various extreme disturbance scenarios are defined, including main substation faults, feeder disconnections, load surges, and voltage drops. For each extreme disturbance scenario, a corresponding disturbance amount is set, the predicted power value of each node under the undisturbed condition is obtained from the time-series state graph, the disturbance amount is applied to the corresponding node, and the node power value after disturbance is calculated. Based on the power grid topology and the electrical connection strength between nodes, the power flow intensity of each power supply path is calculated using the disturbed node power values.

5. The knowledge graph-based main and distribution network collaborative planning method as described in claim 4, characterized in that, The selection of a high-risk node set based on the power fluctuation amplitude before and after the disturbance includes: calculating the risk level of each node under extreme disturbance, wherein the risk level is the ratio of the node power change amplitude to its rated capacity; Sort all nodes by risk level, select the top n nodes with the highest risk level to form a set of high-risk nodes, and output them as potential risk nodes of the power grid.

6. The knowledge graph-based main and distribution network collaborative planning method as described in claim 5, characterized in that, Based on the set of high-risk nodes and the time-series state map, and combined with the load status and power balance constraints of the power supply path, the main and distribution network collaborative optimization scheduling strategy is calculated as follows: With the goal of minimizing node power scheduling deviation under extreme disturbances and the constraints of load status and power balance of power supply path, an objective function for the coordinated optimization of main and distribution networks is constructed. A mixed-integer programming algorithm is used to solve the problem by combining the topology of the time-series state graph, the set of high-risk nodes, and the objective function and constraints. By dynamically adjusting the power allocation of each load node, main transformer node, and new energy access node, and reconstructing the power flow direction of some power supply paths, the optimal main and distribution network collaborative optimization scheduling scheme is obtained.

7. The knowledge graph-based main and distribution network collaborative planning method as described in claim 6, characterized in that, The collaborative planning results include: optimized power scheduling values ​​for each node, adjusted power flow information for the critical path, and response strategies for risky nodes.

8. A knowledge graph-based main distribution network collaborative planning system, applied to the method described in any one of claims 1-7, characterized in that, include: The knowledge graph construction module is used to construct a fusion knowledge graph of the main and distribution networks based on the physical topology, source-load distribution information, and planning rules of the main and distribution networks. The knowledge graph includes a set of equipment nodes, a set of power supply edges, and a set of planning rules. The time-series state prediction module is used to predict the time-series state of each load node and new energy access point based on the knowledge graph. In the prediction process, a power supply path smoothing constraint and planning rule penalty mechanism are introduced to generate a load power prediction sequence and a new energy output prediction sequence attached to the graph nodes, forming a time-series state graph. The disturbance inference module is used to simulate extreme disturbance scenarios based on the time-series state map, infer the node power state after the disturbance, and screen out a set of high-risk nodes based on the power fluctuation amplitude before and after the disturbance. The strategy generation module is used to calculate the main and distribution network collaborative optimization scheduling strategy based on the set of high-risk nodes and the time-series state map, combined with the load status and power balance constraints of the power supply path, so as to adjust the power allocation of each node and the power flow of the path, and output the final collaborative planning result.

9. An electronic device, comprising: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, they implement the steps of the knowledge graph-based main distribution network collaborative planning method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the steps of the knowledge graph-based main distribution network collaborative planning method according to any one of claims 1 to 7.

Images