Scheduling analysis method based on knowledge graph and greedy algorithm

Abstract

The invention discloses a scheduling analysis method based on a knowledge graph and a greedy algorithm, and belongs to the field of power scheduling. The method comprises the following steps of importing machine account information of a power grid access scheme by relying on an existing data medium platform and a D5000 platform of a power grid; the method comprises the following steps of: visualizing a previously accessed scheduling engineering scheme by relying on a graph of a power grid, and constructing an engineering scheme knowledge graph; analyzing a dispatching project power failure scheme, and performing power failure range analysis by using an improved greedy algorithm; and carrying out power station capacity analysis before and after power failure based on an algorithm optimization result. According to the method, power failure range analysis and capacity prediction are optimized, conflicts and resource waste are reduced, and therefore the intelligent analysis level of power grid dispatching is improved.

Description

Technical Field

[0001] The present invention relates to the field of scheduling analysis, and in particular to a scheduling analysis method based on knowledge graph and greedy algorithm. Background Art

[0002] With the increase of new energy sources connected to the grid and the diversification of dispatching engineering plans, the existing dispatching analysis methods have problems of low efficiency and frequent conflicts when dealing with multiple versions of plans, power outage plan conflict detection and capacity analysis. The conflicts between power outage plans and construction plans are frequent, which easily leads to waste of resources and dispatching incoordination. As a result, the traditional dispatching system is slow to respond to capacity overruns and load forecasting, and cannot provide timely warnings and management. Summary of the invention

[0003] The present invention aims to solve at least one of the technical problems existing in the prior art.

[0004] The technical solution of the present invention is: a scheduling analysis method based on knowledge graph and greedy algorithm, comprising the following steps:

[0005] Step 1): Relying on the existing data center of the power grid and the D5000 platform, import the ledger information of the power grid access plan;

[0006] Step 2): Relying on,visualizing the accessed scheduling engineering plan, and building the engineering plan knowledge graph;

[0007] Step 3): Analyze the power outage plan of the scheduling project and use the improved greedy algorithm to analyze the power outage range;

[0008] Step 4): Based on the algorithm optimization results, perform plant capacity analysis before and after the power outage.

[0009] Furthermore, in step 1), the user is allowed to import the ledger information of the dispatching power outage plan, or directly enter the plan information into the system, and automatically generate a plan list for the corresponding project for management personnel to review. Allow users to enter or import relevant information of the main substation, including name, number, rated capacity, connected capacity, location, etc., provide the function of checking the over-limit of new energy access, and judge whether it exceeds the tolerance according to the rated capacity and connected capacity of the main substation, so as to carry out management and early warning. And support the assignment of a unique identifier to each plan version of the project to distinguish different versions of the plan.

[0010] In step 2), a previously connected dispatching engineering plan is visualized based on a grid map. Based on the grid map platform, grid information such as sites and substations of various voltage levels can be displayed, substations or lines can be located in real time, the content displayed on the page can be controlled by layers, and historical wiring diagrams can be viewed.

[0011] After the scheduling engineering plan is visualized, the plan is learned and data is stored based on the knowledge graph.

[0012] The established knowledge graph contains distributed generation nodes, load nodes and power supply nodes, and mainly describes the connection relationship between these nodes, including the connection relationship between DG nodes and load nodes, and the connection relationship between load nodes and power supply nodes.

[0013] In step 3), users are allowed to enter or import the power outage information of the network construction project, including the power outage time, main substation, project, etc., and store it in the power outage information database. By analyzing the construction time of the project, the power outage plan of the main substation is calculated, the start and end time of the power outage and the scope of influence are determined, and the power outage plan of the project construction is used as the basic information of the construction power outage, and provided to the management personnel for reference and use. The specific analysis process of the scheduling project power outage plan is as follows:

[0014] The goal of power outage scope analysis is to use knowledge graphs and optimization algorithms to reasonably determine the affected area at the time of the power outage, and optimize the allocation of resources and recovery plans during the power outage. Power outage scope analysis not only involves capacity prediction and load scheduling before the power outage, but also combines the topology structure (knowledge graph) to accurately divide the area affected by the power outage to avoid unnecessary power outage impacts. The ultimate goal is to minimize the scope of the power outage and ensure the optimal overall operation of the power grid in the event of a power outage.

[0015] In order to effectively solve the problem of power outage scope, the improved greedy algorithm is combined with the knowledge graph, and the topological connection relationship in graph theory is used to assist in analysis. The process is as follows:

[0016] 3.1) Initialization: Input data

[0017] The knowledge graph G = (N, E), where N is the node set (substation, distributed power source, etc.) and E is the topological connection relationship between nodes (power lines, etc.).

[0018] Capacity information of each node Capacity(n i ); Load demand Load(n i ).

[0019] The set of nodes that need to be powered off Indicates the node for which power outage is planned.

[0020] Grid operating constraints during power outages, including capacity limits, load transfer restrictions, etc.

[0021] 3.2) Greedy Selection

[0022] 3.21) Select the power outage area: Starting from the power outage node set S, for each node n i∈S, according to the topological connection relationship TopoConnect(n i ,n j ), determine all nodes n connected to it j ∈N. The formula is expressed as:

[0023] N adj (n i )={n j ∈N|TopoConnect(n i .n j )=1}

[0024] Where N adj (n i ) indicates that n i A collection of adjacent nodes.

[0025] 3.22) Greedy selection of nodes: In the adjacent node set N adj (n i ), select the node with the least affected capacity, that is, according to the capacity formula:

[0026]

[0027] In the formula, Impact(n j ) represents node n j the extent to which the capacity of

[0028] Select the minimum Impact(n j ) nodes are prioritized for power outage.

[0029] 3.3) Capacity assessment and iteration

[0030] 3.31) Capacity assessment: For the selected power outage area, the system evaluates the capacity usage in the area, calculates the total load and total capacity of all nodes in the area, and ensures that the capacity limit will not be exceeded during the power outage. The formula is:

[0031]

[0032] In the formula, TotalLoad represents the total load of all nodes in the area, TotalCapacity represents the total capacity of all nodes in the area, and N adj Represents the set of all nodes directly or indirectly affected by the power outage;

[0033] If TotalLoad≤TotalCapacity, the current power outage plan is feasible;

[0034] 3.32) Iterative optimization: For each round of selected outage nodes, the outage scope is continued to be expanded through the greedy algorithm, and the above selection and capacity evaluation process is repeated until the goal of minimizing the impact of the outage is achieved or the system constraints are met.

[0035] 3.4) Stop Condition

[0036] Minimize the impact of power outages: Stop when the selected power outage range meets the minimization objective function (such as minimum impact, lowest load transfer cost, etc.).

[0037] Constraints: If system capacity is exceeded or load shifts are outside constraints, stop and make adjustments.

[0038] Improved greedy algorithm:

[0039] In order to improve the performance of the traditional greedy algorithm, the following improvements can be introduced:

[0040] Heuristic weight selection: When selecting adjacent nodes, not only the capacity impact is considered, but also the heuristic weight w(n j ), the weights can be based on the importance of the nodes in the power grid, the priority of load transmission, etc.

[0041]

[0042] In the formula, w(n j ) is the heuristic weight of the node.

[0043] In step 4), the capacity analysis before power outage includes:

[0044] Capacity data collection: Extract capacity data of relevant plants and substations from the grid access plan and D5000 system, including rated capacity, connected capacity and load data of main substations;

[0045] Load forecasting: Combine historical load data with current load conditions to predict load demand during power outages and assess whether there is a risk of insufficient capacity;

[0046] Capacity constraint verification: Based on the rated capacity and predicted load of the main substation, check whether there is a situation where the capacity limit is exceeded during the power outage, and issue an early warning; the specific calculation formula is as follows:

[0047]

[0048] Where CapacityUsagt(t) represents the capacity utilization rate of the node at the prediction time t; Load(t) is the load at the prediction time t; RatedCapacity is the rated capacity of the plant, if > 100%, an alarm is triggered;

[0049] Post-outage capacity restoration analysis includes:

[0050] Recovery capacity monitoring: Use sensors and monitoring equipment to collect capacity data after the plant is restored in real time to ensure that the plant load does not exceed its designed capacity after restoration;

[0051] Capacity balance analysis: Based on the actual load conditions of each plant and station before and after the power outage, compare whether the load distribution before and after the power outage is balanced, and evaluate whether load transfer or adjustment is needed; the specific calculation formula is as follows:

[0052] ΔC=CapacityPost(t)-CapacityPre(t)

[0053] Where △C represents the capacity change, LCapacityPost(t) and CapacityPre(t) represent the plant capacity at a certain time t after and before the power outage, respectively.

[0054] In the work, the present invention firstly integrates and visualizes the grid access scheme through knowledge graph technology, improves the transparency and traceability of data management, and effectively solves the problem of chaotic management of multi-version schemes. Secondly, the improved greedy algorithm is used to analyze the power outage plan, optimizes the prediction of the power outage range, reduces the power outage plan conflicts, and improves the efficiency of dispatch management. The system can also monitor and evaluate the capacity of the main substation in real time, and provide accurate early warning for the access of new energy. Overall, the present invention significantly improves the level of intelligence of power grid dispatching. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 The present invention is a flow chart of the method. DETAILED DESCRIPTION

[0056] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the examples described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0057] like Figure 1 As shown, the present invention provides a scheduling analysis method based on knowledge graph and greedy algorithm, comprising the following steps:

[0058] Step 1): Relying on the power grid data center and D5000 platform, import the ledger information of the power grid access plan;

[0059] Step 2): Visualize the accessed scheduling engineering plan and build the engineering plan knowledge graph;

[0060] Step 3): Analyze the power outage plan of the scheduling project and use the improved greedy algorithm to analyze the power outage range;

[0061] Step 4): Based on the algorithm optimization results, perform plant capacity analysis before and after the power outage.

[0062] The present invention further improves the intelligence and precision of the dispatch analysis method by integrating the operating data in the existing power grid, scientifically and rationally analyzes the power outage project plans in various areas of the power grid, warns and optimizes capacity overload problems, and improves the stability and dispatch efficiency of the power grid.

[0063] The specific process of the present invention is as follows:

[0064] Step 1): Relying on the existing data center and D5000 platform of the power grid, import the ledger information of the dispatching power outage project plan. During implementation, intelligent data import and dynamic management are used to ensure the integration and transparency of the dispatching power outage plan information. First, allow users to import the ledger information of the power outage plan with one click or directly enter it into the system, so as to automatically generate the corresponding project plan list, realize the rapid summary and display of the plan, provide managers with a clear and simple way to consult the plan, and greatly improve the efficiency of information acquisition. For the main substation information involved in the power outage plan, it supports detailed entry of its key attributes, including name, number, rated capacity, connected capacity and geographical location. Based on these entered data information, further perform over-limit verification operations: by comparing the rated capacity of the main substation with the current connected capacity, determine whether it reaches or exceeds the tolerance, so as to trigger an early warning when risks are found. This not only allows the capacity over-limit situation to be effectively identified and responded to in the early stage, but also avoids load distribution errors during power outages, laying the foundation for the safe operation of the power grid.

[0065] In addition, in the scheme management, a unique identifier is assigned to ensure the version distinguishability and traceability of each power outage plan. This process solves the management complexity of the scheduling plan under multiple version iterations, so that the plan versions of different time periods can always be accurately called, avoiding version confusion in the scheduling process. For example, in seasonal changes with large fluctuations in electricity demand (such as summer peaks and winter troughs), a unique identifier is used to ensure that each plan version can be clearly distinguished, so that the correct plan can be flexibly called under different load scenarios, thereby supporting more accurate scheduling decisions. Overall, this step fully supports the automated and intelligent management of power outage plans through data integration, capacity verification and version identification, and realizes the improvement of scheduling efficiency and the guarantee of operational stability.

[0066] Step 2): Relying on the One Grid Map, visualize the previously connected dispatching engineering schemes and build the engineering scheme knowledge graph. Based on the One Grid Map platform, you can display the grid information such as sites and substations of various voltage levels, locate substations or lines in real time, control the layers of the content displayed on the page, and view the historical wiring diagram. The One Grid Map platform is a conventional technology in the power system.

[0067] After the scheduling engineering plan is visualized, the plan is learned and data is stored based on the knowledge graph. In the scheduling analysis method, the engineering plan knowledge graph is an industry knowledge graph built for power scheduling scenarios, involving a large number of power equipment, such as transformers, voltage transformers, and interconnecting switches. In this patent, the established knowledge graph includes distributed power supply nodes, load nodes, and power supply nodes, and mainly describes the connection relationship between these nodes, including the connection relationship between DG nodes and load nodes, and the connection relationship between load nodes and power supply nodes. The connection between each node represents the attributes of the node, such as load capacity, power size, geographical location, load importance, etc.

[0068] The relationship between nodes does not contain specific attributes, but can be assigned values ​​according to specific analysis needs. Common assignments include: connection, disconnection, inclusion, and membership. For example, the connection between node A and node B can be described by a topological connection. In distribution network scheduling, the topological connection relationship is the most important relationship, especially the topological connection relationship between adjacent nodes is an important basis for building the overall topology of the power grid.

[0069] The knowledge graph construction method is applicable to distribution networks of different types and voltage levels. Although the operation modes of distribution networks of different voltage levels are different, in power outage analysis, the topological connection relationship can be simplified into two categories: "connected" or "unconnected". Therefore, the storage and derivation process based on the knowledge graph remains consistent under different operating conditions, and only the value of the topological relationship changes.

[0070] Assume that the knowledge graph is represented as a directed graph G = (N, R), where N is a set of nodes and R is a set of relationships. The topological connection relationship can be represented by a function:

[0071]

[0072] Define the capacity function Capacity(n) to represent the load capacity of node n, and define the power flow function PowerFlow(n i ,n j ) represents node n i to n j The power flow between.

[0073] In the power outage analysis, it is assumed that node n kA power outage occurs, and the set of nodes affected is N affected It can be expressed by the following formula:

[0074] N affected ={n j ∈N|TopoConnect(n k ,n j )=1}

[0075] The topological analysis method based on knowledge graph can more accurately and efficiently identify the relationship between different nodes in the power grid, optimize power grid dispatching management and support the efficient access and utilization of distribution network nodes.

[0076] Step 3): Analyze the power outage plan of the dispatching project, and use the improved greedy algorithm to analyze the power outage range. Allow users to enter or import the power outage information of the network project construction, including the power outage time, main substation, project, etc., and store it in the power outage information database. By analyzing the construction time of the project, calculate the power outage plan of the main substation, determine the start and end time and impact range of the power outage, and use the project construction power outage plan as the basic information for construction power outages, and provide it to management personnel for reference and use. The specific analysis process of the dispatching project power outage plan is as follows:

[0077] The goal of power outage scope analysis is to use knowledge graphs and optimization algorithms to reasonably determine the affected area at the time of the power outage, and optimize the allocation of resources and recovery plans during the power outage. Power outage scope analysis not only involves capacity prediction and load scheduling before the power outage, but also combines the topology structure (knowledge graph) to accurately divide the area affected by the power outage to avoid unnecessary power outage impacts. The ultimate goal is to minimize the scope of the power outage and ensure the optimal overall operation of the power grid in the event of a power outage.

[0078] In order to effectively solve the problem of power outage scope, the improved greedy algorithm can be combined with the knowledge graph and the topological connection relationship in graph theory can be used to assist in analysis. The process is as follows:

[0079] 3.1) Initialization: Input data

[0080] The knowledge graph G = (N, E), where N is the node set (substation, distributed power source, etc.) and E is the topological connection relationship between nodes (power lines, etc.).

[0081] Capacity information of each node Capacity(n i ); Load demand Load(n i ).

[0082] The set of nodes that need to be powered off Indicates the node for which power outage is planned.

[0083] Grid operating constraints during power outages, including capacity limits, load transfer restrictions, etc.

[0084] 3.2) Greedy Selection

[0085] 3.21) Select the power outage area: Starting from the power outage node set S, for each node n i ∈S, according to the topological connection relationship TopoConnect(n i ,n j ), determine all nodes n connected to it j ∈N. The formula is expressed as:

[0086] N adj (n i )={n j ∈N|TopoConnect(n i .n j )=1}

[0087] Where N adj (n i ) indicates that n i A collection of adjacent nodes.

[0088] 3.22) Greedy selection of nodes: In the adjacent node set N adj (n i ), select the node with the least affected capacity, that is, according to the capacity formula:

[0089]

[0090] In the formula, Impact(n j ) represents node n j the extent to which the capacity of

[0091] Select the minimum Impact(n j ) nodes are prioritized for power outage.

[0092] When Impact(n j ) is larger, indicating that the node load is close to its maximum capacity and may face overload risk; conversely, the smaller the value, the less the capacity of the node is affected, and the negative impact on the overall system during the scheduling process is also smaller. Therefore, when the greedy algorithm selects nodes, it usually gives priority to nodes with smaller values ​​to optimize resource allocation and reduce the impact on system stability.

[0093] 3.3) Capacity assessment and iteration

[0094] 3.31) Capacity assessment: For the selected power outage area, the system evaluates the capacity usage in the area, calculates the total load and total capacity of all nodes in the area, and ensures that the capacity limit will not be exceeded during the power outage. The formula is:

[0095]

[0096] In the formula, TotalLoad represents the total load of all nodes in the area, TotalCapacity represents the total capacity of all nodes in the area, and N adj Represents the set of all nodes directly or indirectly affected by the power outage, usually the set of nodes in the power outage area and their topologically adjacent nodes;

[0097] If TotalLoad≤TotalCapacity, the current power outage plan is feasible.

[0098] 3.32) Iterative optimization: For each round of selected outage nodes, the outage scope is continued to be expanded through the greedy algorithm, and the above selection and capacity evaluation process is repeated until the goal of minimizing the impact of the outage is achieved or the system constraints are met.

[0099] 3.4) Stop Condition

[0100] Minimize the impact of power outages: Stop when the selected power outage range meets the minimization objective function (such as minimum impact, lowest load transfer cost, etc.).

[0101] Constraints: If system capacity is exceeded or load shifts are outside constraints, stop and make adjustments.

[0102] Improved Greedy Algorithm,In order to improve the performance of the traditional greedy algorithm, the following improvements can be introduced:

[0103] Heuristic weight selection: When selecting adjacent nodes, not only the capacity impact is considered, but also the heuristic weight w(n j ), the weights can be based on the importance of the nodes in the power grid, the priority of load transmission, etc.

[0104]

[0105] In the formula, w(n j ) is the heuristic weight of the node.

[0106] Multi-round iterative optimization: After each round of greedy selection, a global evaluation is performed to further optimize the power outage range through subsequent iterations to reduce local optimal situations.

[0107] This paper uses a specific example to illustrate how to use the improved greedy algorithm and knowledge graph to solve the problem of power outage range analysis. Example background: Assume that the knowledge graph of a city power grid includes the following main nodes:

[0108] Substation A: rated capacity 100MW, current load 70MW.

[0109] Substation B: rated capacity 150MW, current load 120MW.

[0110] Distributed power generation C: rated capacity 50MW, current load 40MW, connected to substation A.

[0111] Load node D: connected to substation B, with a load of 50MW.

[0112] In addition, the nodes in the power grid are connected to each other through transmission lines to form a topological structure. In a planned power outage, substation A needs to be temporarily disconnected, which will affect other nodes connected to it.

[0113] 3.1) Initialization

[0114] The knowledge graph of the current power grid is input into the system, including the capacity and load data of each node (such as substation A, B, distributed power source C, load node D), as well as the topological connection relationship between these nodes.

[0115] Initialize the knowledge graph:

[0116] Node A (substation) is connected to node C (distributed generation).

[0117] Node B (substation) is connected to node D (load node).

[0118] Input formula: The capacity utilization rate of each node can be calculated by the formula: The capacity utilization rate of substation A is 70%, and the capacity utilization rate of substation B is 80%.

[0119] 3.2) Greedy selection of nodes

[0120] According to the topological connection relationship in the knowledge graph, the impact scope of the power outage is analyzed. When disconnecting substation A, the system needs to find the nodes connected to it, such as distributed power generation C, and evaluate its load transfer after the power outage.

[0121] In the greedy selection process, the system will give priority to nodes with less impact. For example, if the load of distributed power generation C can be temporarily transferred to other nodes, the scope of power outage can be reduced.

[0122] Select the standard formula and calculate the impact value of distributed power source C as 0.8

[0123] The system will compare the impact value of each adjacent node and prioritize nodes with smaller impact values ​​for power outages, ensuring that the operation of other parts of the power grid is not affected as much as possible.

[0124] 3.3) Capacity assessment and iteration

[0125] After each greedy selection, the system will evaluate the capacity usage within the current power outage range to ensure that the load does not exceed the capacity limit of the remaining nodes. For example, if the distributed power source C is out of power, the system will calculate whether the load can be successfully transferred to substation B. The calculation results are: TotalLoad(B)=160MW, CapacityUsage(B)=106.67%

[0126] Since the load of substation B exceeds the capacity, the power outage plan is not feasible, and the system will iteratively adjust and try other load transfer plans.

[0127] 3.4) Iterative optimization of power outage areas

[0128] Through the improved greedy algorithm, the system will continuously adjust the blackout area and load transfer to ensure that the capacity utilization rate of the entire power grid is maintained within a reasonable range. At this step, the system may choose to incompletely blackout or only partially transfer the load to ensure the smooth operation of the power grid.

[0129] 3.5) Stop Condition

[0130] When the system finds an optimal solution that minimizes the impact of power outages and does not exceed capacity constraints, it stops. At this point, the system generates an outage plan that lists which nodes need to be out of power, which nodes' loads will be transferred, and ensures the stable operation of the power grid during the outage.

[0131] Step 4): Capacity analysis before power outage: Before a power outage occurs, the capacity of the affected plants and stations needs to be analyzed to assess the impact of the power outage on the overall operation of the power grid.

[0132] The specific technical steps are as follows:

[0133] Capacity data collection: Extract capacity data of relevant plants and stations from the grid access plan and D5000 system, including the rated capacity, connected capacity and load data of the main substation.

[0134] Load forecasting: Combine historical load data with current load conditions and use load forecasting algorithms to predict load demand during power outages and assess whether there is a risk of insufficient capacity.

[0135] Quantity constraint verification: Based on the rated capacity and predicted load of the main substation, verify whether there is a situation of exceeding the capacity limit during the power outage, and issue an early warning. The specific calculation formula is as follows:

[0136]

[0137] Where CapacityUsagt(t) represents the capacity utilization rate of the node at the prediction time t; Load(t) is the load at the prediction time t; RatedCapacity is the rated capacity of the plant. If it is greater than 100%, an alarm is triggered.

[0138] Capacity recovery analysis after power outage: After the power outage ends, the system needs to monitor the capacity recovery of the plant in real time and conduct a detailed analysis.

[0139] The specific technical steps are as follows:

[0140] Recovery capacity monitoring: The system collects capacity data of the plant after recovery through sensors and monitoring equipment in real time to ensure that the plant load does not exceed its design capacity after recovery.

[0141] Capacity balance analysis: The system compares the load distribution before and after the power outage to see whether it is balanced, and evaluates whether load transfer or adjustment is needed, based on the actual load conditions of each plant and station before and after the power outage. The specific calculation formula is as follows:

[0142] ΔC=CapacityPost(t)-CapacityPre(t)

[0143] Where △C represents the capacity change, reflecting the impact of the power outage on the node capacity distribution or load distribution. CapacityPost(t) and CapacityPre(t) represent the plant capacity at a certain time t after and before the power outage, respectively.

[0144] Combined with specific examples, it can include two types of analysis: before and after the power outage: before the power outage, first collect the key capacity data of the relevant plants and stations, including the rated capacity, connected capacity and current load level of the main substation, as the basic data for subsequent analysis. Taking the main substation of an industrial park as an example, if the rated capacity is 500MVA and the connected new energy capacity is 450MVA, this data will be included in the capacity assessment process to lay the data foundation for the next step of load forecasting.

[0145] In the load forecasting process, the load forecasting algorithm is used to accurately predict the demand during the power outage by combining historical load data and current power demand. Assume that a significant load fluctuation is identified in the analysis on a certain day in the summer, and the load demand during the power outage is predicted to be close to 90% of the rated capacity. At this point, it is further evaluated whether the load will exceed the capacity limit under the power outage conditions. Once the predicted load demand exceeds the safety limit (such as 100% of the rated capacity), an early warning will be automatically triggered to provide the dispatcher with a risk warning of insufficient capacity so that load transfer or dispatch adjustment measures can be taken in advance.

[0146] Capacity recovery analysis after power outage: After the power outage ends, the recovery of the plant is continuously monitored to ensure that the capacity recovery is stable and meets the design limits. The recovery capacity monitoring function collects data in real time through sensors and monitoring equipment, and can detect the actual load changes of the plant after the power supply is restored. For example, after the power supply of a main substation is restored, its load gradually approaches 95% of the rated capacity. Comparative analysis based on real-time data ensures that the actual load remains balanced after restoration and verifies whether the restored capacity meets the design capacity standards.

[0147] In the capacity balance analysis phase, based on the actual load distribution before and after the power outage, the change trends of the two are compared. For example, before the power outage, the load of a plant is concentrated in the adjacent area, and after the power is restored, the load transitions to another plant. At this time, it is determined whether additional load transfer is needed based on the load balance before and after the power outage to avoid uneven load in the plant after the power is restored.

[0148] Overall, through load forecasting and capacity verification before power outages and recovery monitoring and capacity balance analysis after power outages, the risks of load imbalance and insufficient capacity can be effectively avoided, making the power grid's dispatching plan more adaptable and reliable.

[0149] The present invention, in application, includes:

[0150] Capacity assessment mechanism: First, the capacity of the power plants in the selected power outage area is assessed to verify whether their total load and total capacity during the power outage are within the safe range. This step ensures the capacity feasibility of the power outage plan, avoids scheduling failures due to overload, and lays the foundation for subsequent load transfer.

[0151] Optimize load transfer paths: After capacity assessment, the load in the blackout area is reasonably distributed to surrounding nodes. Based on the topological relationship in the knowledge graph, the nodes with the least load impact are preferentially selected for transfer to achieve efficient transfer of the blackout load. This process takes into account the remaining capacity and topological relationship of each node to ensure that the load distribution meets the capacity limit and reduces the overall fluctuation of the power grid.

[0152] Dynamic adjustment of topology: During the load transfer process, the topology of the nodes in the knowledge graph is dynamically adjusted to adapt to the changes in power flow during the power outage.

[0153] The node connection relationship and capacity information of the knowledge graph provide real-time support for scheduling, so that the topology structure remains optimal during power outages. Through continuous feedback and adjustment, the stability of power grid operation is ensured and the power loss on the load conduction path is reduced.

[0154] The present invention optimizes and improves the outage analysis method, improves the transparency and traceability of data management, optimizes the outage range prediction, reduces outage plan conflicts, and improves the efficiency of dispatch management. It can also monitor and evaluate the capacity of the main substation in real time, provide accurate early warning for the access of new energy, and alleviate the absorption pressure of the power grid. The present invention significantly improves the intelligent level of power grid dispatching and improves the accuracy and speed of dispatching work.

[0155] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, and these changes and improvements all fall within the scope of the present invention to be protected.

Claims

1. A scheduling analysis method based on knowledge graph and greedy algorithm, characterized in that: The following steps are involved: Step 1): Relying on the power grid data center and D5000 platform, import the ledger information of the power grid access plan; Step 2): Visualize the accessed scheduling engineering plan and build the engineering plan knowledge graph; Step 3): Analyze the power outage plan of the scheduling project and use the improved greedy algorithm to analyze the power outage range; Step 4): Based on the algorithm optimization results, perform plant capacity analysis before and after the power outage.

2. According to claim 1, a scheduling analysis method based on knowledge graph and greedy algorithm is characterized in that: In step 1), the user is allowed to import the ledger information of the dispatching power outage plan, or directly enter the plan information into the system to automatically generate a plan list for the corresponding project; the user is allowed to enter or import relevant information of the main substation, including name, number, rated capacity, connected capacity, and location.

3. The scheduling analysis method based on knowledge graph and greedy algorithm according to claim 1 is characterized in that: In step 2), the previously connected dispatching engineering scheme is visualized based on a grid map. Among them, based on the power grid one-picture platform, the power grid information of sites and substations of various voltage levels is displayed, substations or lines are located in real time, the layers of the content displayed on the page are controlled, and historical wiring diagrams can be viewed; After the scheduling engineering plan is visualized, the plan is learned and data is stored based on the knowledge graph.

4. According to claim 3, a scheduling analysis method based on knowledge graph and greedy algorithm is characterized in that: The established knowledge graph includes distributed power supply nodes, load nodes and power supply nodes, as well as the connection relationships between these nodes, including the connection relationship between distributed power supply nodes and load nodes, and the connection relationship between load nodes and power supply nodes.

5. The scheduling analysis method based on knowledge graph and greedy algorithm according to claim 1 is characterized in that: In step 3), the user is allowed to input or import the power outage information of the network access project construction, including the power outage time, main substation, and project, and store it in the power outage information database; By analyzing the construction time of the project, the power outage plan of the main substation is calculated, the start and end time and the impact range of the power outage are determined, and the construction power outage plan of the project is used as the basic information for construction power outages and provided to management personnel for reference and use.

6. The scheduling analysis method based on knowledge graph and greedy algorithm according to claim 5 is characterized in that: The power outage range analysis uses an improved greedy algorithm combined with a knowledge graph, including the following: 3.1) Initialization: Input data Knowledge graph G = (N, E), where N is the node set and E is the topological connection relationship between nodes; Capacity information of each node Capacity(n i ); Load demand Load(n i ); The set of nodes that need to be powered off Nodes representing planned power outages; Grid operating constraints during blackouts, including capacity caps and load shifting restrictions; 3.2) Greedy selection 3.21) Select the power outage area: Starting from the power outage node set S, for each node n i ∈S, according to the topological connection relationship TopoConnect(n i ,n j ), determine all nodes n connected to it j ∈N, the formula is expressed as: N adj (n i )={n j ∈N|TopoConnect(n i .n j )=1} Where N adj (n i ) indicates that i The set of adjacent nodes; 3.22) Greedy selection of nodes: In the adjacent node set N adj (n i ), select the node with the least affected capacity, according to the capacity formula: In the formula, Impact(n j ) represents node n j the extent to which the capacity of Select the minimum Impact(n j ) nodes as priority power outage targets; 3.3) Capacity assessment and iteration 3.31) Capacity assessment: For the selected power outage area, evaluate the capacity usage in the area, calculate the total load and total capacity of all nodes in the area, and ensure that the capacity limit will not be exceeded during the power outage. The formula is expressed as: In the formula, TotalLoad represents the total load of all nodes in the area, TotalCapacity represents the total capacity of all nodes in the area, and N adj Represents the set of all nodes directly or indirectly affected by the power outage; If TotalLoad≤TotalCapacity, the current power outage plan is feasible; 3.32) Iterative optimization: For each round of selected power outage nodes, the power outage range is continuously expanded through the greedy algorithm, and the above selection and capacity evaluation process is repeated until the goal of minimizing the impact of power outage is achieved or the system constraints are met; 34) Stop condition Minimize the impact of power outages: Stop when the selected power outage range meets the minimization objective function; Constraints: If system capacity is exceeded or load shifts are outside constraints, stop and make adjustments.

7. The scheduling analysis method based on knowledge graph and greedy algorithm according to claim 6 is characterized in that: The improved greedy algorithm also includes: Heuristic weight selection: When selecting adjacent nodes, not only the capacity impact is considered, but also the heuristic weight w(n j ), the weights are based on the importance of nodes in the power grid and the priority of load transmission, specifically: In the formula, w(n j ) is the heuristic weight of the node.

8. The scheduling analysis method based on knowledge graph and greedy algorithm according to claim 1 is characterized in that: In step 4), the capacity analysis before power outage includes: Capacity data collection: Extract capacity data of relevant plants and substations from the grid access plan and D5000 system, including rated capacity, connected capacity and load data of main substations; Load forecasting: Combine historical load data with current load conditions to predict load demand during power outages and assess whether there is a risk of insufficient capacity; Capacity constraint verification: Based on the rated capacity and predicted load of the main substation, check whether there is a situation where the capacity limit is exceeded during the power outage, and issue an early warning; the specific calculation formula is as follows: Where CapacityUsagt(t) represents the capacity utilization rate of the node at the prediction time t; Load(t) is the load at the prediction time t; RatedCapacity is the rated capacity of the plant, if > 100%, an alarm is triggered; Post-outage capacity restoration analysis includes: Recovery capacity monitoring: Use sensors and monitoring equipment to collect capacity data after the plant is restored in real time to ensure that the plant load does not exceed its designed capacity after restoration; Capacity balance analysis: Based on the actual load conditions of each plant and station before and after the power outage, compare whether the load distribution before and after the power outage is balanced, and evaluate whether load transfer or adjustment is needed; the specific calculation formula is as follows: ΔC=CapacityPost(t)-CapacityPre(t) Where △C represents the capacity change, LCapacityPost(t) and CapacityPre(t) represent the plant capacity at a certain time t after and before the power outage, respectively.

Images