Electric power data analysis method and device based on graph calculation

Abstract

The invention relates to the technical field of electric power data analysis, in particular to an electric power data analysis method and device based on graph calculation, and the method comprises the steps: constructing an electric power network graph, carrying out the deep learning processing of multi-dimensional and heterogeneous data, and analyzing a dynamic change prediction trend; power dispatching is optimized in combination with reinforcement learning, an optimal power dispatching decision is matched with minimization of energy loss as a target, the technical problems that high-dimensional data and heterogeneous data are complex to process, the change trend of the power system cannot be accurately predicted and an effective dispatching decision is lacked in the operation process of the power system are solved, and the power system can be effectively dispatched according to the real-time power grid operation state and the prediction trend. According to the method, the aim of self-adaptive strategy optimization learning is achieved, power dispatching flexibility is greatly improved, energy loss minimization is taken as the aim, an optimal power dispatching decision is accurately matched through continuous iterative optimization, power system data analysis efficiency is improved, and strategies are adjusted to cope with continuously changing supply and demand conditions and power grid operation conditions; and the overall stability and reliability of the power grid are improved to the greatest extent.

Description

Technical Field

[0001] The present invention relates to the technical field related to power data analysis, and in particular to a power data analysis method and device based on graph computing. Background Art

[0002] Power data analysis methods usually use relational database technology and statistical methods to process data in power systems. They perform well when dealing with relatively simple and small-scale power grids, but they expose many shortcomings when facing modern complex and large-scale power grid systems. Modern power systems contain a large number of entities such as power plants, substations, and transmission lines. The logical connections between entities are complex and dynamically changing. The data generated during the operation of the power system has high-dimensional and heterogeneous characteristics, making it difficult to efficiently process highly concurrent complex data streams.

[0003] In summary, the existing technology has technical problems such as complex processing of high-dimensional data and heterogeneous data during the operation of the power system, inability to accurately predict the dynamic change trend of the power system, and lack of effective scheduling decision-making. Summary of the invention

[0004] This application provides a graph-based power data analysis device, aiming to solve the technical problems in the prior art of complex processing of high-dimensional data and heterogeneous data during power system operation, the inability to accurately predict the dynamic change trend of the power system, and the lack of effective scheduling decisions.

[0005] In view of the above problems, the technical solution to implement this application is: On the one hand, the present application provides a method for analyzing power data based on graph computing, wherein the method comprises: constructing a power network structure diagram, wherein nodes represent entities in the power system, including power plants, substations, and transmission lines, and edges represent logical connection relationships between entities; performing feature extraction on nodes and edges in the power network structure diagram to capture mutual influence relationships and interactive connection relationships between entities in the power system; based on the mutual influence relationships and interactive connection relationships, using a deep learning architecture to perform multimodal graph fusion calculations to process high-dimensional data in the power system and collect a first deep dynamic evolution set; based on the mutual influence relationships and interactive connection relationships, Influence relationships and interactive connection relationships are analyzed using a deep learning architecture to perform multimodal graph fusion calculations to process heterogeneous data in the power system and collect a second deep dynamic evolution set; taking the power grid operation status as the starting point, the dynamic changes of the power system in the time dimension and the space dimension are analyzed through the first deep dynamic evolution set and the second deep dynamic evolution set, and the perceived prediction trend is extracted; based on the power network structure diagram, a reinforcement learning algorithm is deployed to perform adaptive strategy optimization learning based on the real-time updated power grid operation status combined with the perceived prediction trend, and match the optimal power dispatch decision with the goal of minimizing energy loss.

[0006] On the other hand, the present application provides an electric power data analysis device based on graph computing, wherein the device includes: a structure graph construction module, which is used to construct an electric power network structure graph, wherein nodes represent entities in the electric power system, including power plants, substations, and transmission lines, and edges represent logical connection relationships between entities; a feature extraction module, which is used to extract features from nodes and edges in the electric power network structure graph to capture the mutual influence relationship and interactive connection relationship between entities in the electric power system; a first fusion calculation module, which is used to perform multimodal graph fusion calculation based on the mutual influence relationship and interactive connection relationship using a deep learning architecture to process high-dimensional data in the electric power system and collect a first deep dynamic evolution set; a second fusion calculation module A module is used to perform multimodal graph fusion calculations based on the mutual influence relationship and interactive connection relationship using a deep learning architecture to process heterogeneous data in the power system and collect a second deep dynamic evolution set; a dynamic change analysis module is used to take the power grid operation status as the starting point, analyze the dynamic changes of the power system in the time dimension and the space dimension through the first deep dynamic evolution set and the second deep dynamic evolution set, and extract the perceived prediction trend; a scheduling decision module is used to deploy a reinforcement learning algorithm based on the power network structure diagram, perform adaptive strategy optimization learning according to the real-time updated power grid operation status and the perceived prediction trend, and match the optimal power scheduling decision with the goal of minimizing energy loss.

[0007] In summary, one or more technical solutions provided in this application solve the technical problems of complex processing of high-dimensional data and heterogeneous data during the operation of the power system, the inability to accurately predict the dynamic change trend of the power system, and the lack of effective scheduling decisions. It achieves the goal of adaptive strategy optimization learning based on real-time power grid operation status and predicted trends, greatly improves the flexibility of power dispatching, and takes minimizing energy loss as the core goal. Through continuous iterative optimization, the optimal power dispatching decision is accurately matched, so as to improve the efficiency of power system data analysis, and automatically adjust the strategy according to the power grid status in different scenarios to cope with the ever-changing supply and demand conditions and power grid operating conditions, thereby maximizing the technical effect of improving the overall stability and reliability of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 A flow chart of a power data analysis method based on graph computing is provided for this application; Figure 2 A structural schematic diagram of a power data analysis device based on graph computing is provided for this application.

[0009] Explanation of the accompanying drawings: structure diagram construction module M100, feature extraction module M200, first fusion calculation module M300, second fusion calculation module M400, dynamic change analysis module M500, scheduling decision module M600. DETAILED DESCRIPTION

[0010] Embodiment 1

[0011] The present application is described in detail below with reference to the accompanying drawings. Figure 1 As shown, the present application provides a power data analysis method based on graph calculation, wherein the method includes: S1: Construct a power network structure diagram, in which nodes represent entities in the power system, including power plants, substations, and transmission lines, and edges represent logical connection relationships between entities; S2: Perform feature extraction on the nodes and edges in the power network structure diagram to capture the mutual influence and interactive connection relationships between entities in the power system.

[0012] Construct a power network structure diagram. Specifically, define nodes. Nodes represent various entities in the power system, including power plants, substations, transmission lines, etc. Each node represents a specific part of the power system and plays an important role in the operation of the power grid. Define edges. Edges represent the logical connection between entities. The connection relationship can be a physical connection (such as a transmission line connecting a power plant and a substation) or a logical dependency (such as a substation relying on multiple power plants for power supply). The existence of edges reflects the mutual influence and dependency between different parts of the power system. Connect the defined nodes and edges to form a power network structure diagram.

[0013] The nodes and edges in the power network structure diagram are extracted to capture the mutual influence and interactive connection between entities in the power system. Specifically, node features are extracted. For each node, a series of features are extracted to describe its attributes and status, such as the power generation of the power plant, the voltage level of the substation, the transmission capacity of the transmission line, etc. These features reflect the function and performance of the node in the power system; edge features are extracted. For the edges connecting different nodes, features such as transmission capacity, transmission loss, and transmission distance are extracted to reflect the strength and nature of the connection relationship represented by the edge. Clearly displaying the various parts of the power system and their interrelationships is of great significance for in-depth understanding of the power system, optimizing power dispatching, and ensuring the stability of the power grid, and provides strong support for subsequent data analysis and decision-making.

[0014] S3: Based on the mutual influence relationship and interactive connection relationship, a deep learning architecture is used to perform multimodal graph fusion calculations to process high-dimensional data in the power system and collect a first deep dynamic evolution set.

[0015] Based on the mutual influence and interactive connection relationships represented by nodes and edges in the power network structure diagram, a deep learning architecture is used for multimodal graph fusion calculation to process high-dimensional data in the power system. Furthermore, the power network structure diagram: contains nodes (power plants, substations, transmission lines, etc.) and edges (logical connection relationships); high-dimensional data refers to various data generated during the operation of the power system, such as power generation, voltage, current, load changes, etc., which are usually high-dimensional and complex.

[0016] Choose deep learning models suitable for processing graph structured data, such as Graph Neural Networks, Graph Convolutional Networks, etc. These models can capture the complex relationships in the graph structure, including the mutual influence between nodes and the interactive connection relationships between edges.

[0017] Since the data in the power system has multiple modes (such as numerical, text, image, etc.), multimodal fusion calculation is required. Multimodal fusion can be achieved by converting data of different modes into a unified representation and performing joint learning in a deep learning model.

[0018] The power system is a dynamically changing system whose operating status will change over time. Through the deep learning model, the nodes and edges in the power network structure diagram are dynamically evolved and analyzed to capture the changing patterns of the power system in the time dimension. After multimodal graph fusion calculation and dynamic evolution analysis, the collected data set is called the first deep dynamic evolution set, which contains the dynamic change information of the power system in the time dimension, as well as the mutual influence and interactive connection relationship between different entities.

[0019] Through multimodal graph fusion calculation based on mutual influence and interactive connection relationships, high-dimensional data in the power system can be effectively processed and useful dynamic change information can be extracted, which is of great significance for subsequent power system analysis, prediction and optimization decision-making. Using deep learning architecture for multimodal graph fusion calculation is a key step in processing high-dimensional data in the power system and capturing dynamic change laws, providing strong support for intelligent management and optimization of the power system.

[0020] S4: Based on the mutual influence relationship and interactive connection relationship, use a deep learning architecture to perform multimodal graph fusion calculation to process the heterogeneous data in the power system and collect a second deep dynamic evolution set.

[0021] Heterogeneous data refers to data from different sources with different structures and formats. In order to effectively process heterogeneous data, a deep learning architecture based on mutual influence and interactive connection relationships is used to perform multimodal graph fusion calculations to collect the second deep dynamic evolution set. Specifically, heterogeneous data from different sources are integrated, including sensor data, historical records, user feedback, etc. During the data integration process, it is necessary to solve problems such as inconsistent data formats, data redundancy, and data conflicts; the integrated heterogeneous data is cleaned to remove noise, missing values, and inconsistent data, and data preprocessing is performed, such as data standardization and normalization, for subsequent data analysis and processing; for heterogeneous data, features that can reflect the mutual influence and interactive connection relationships between power system entities are extracted, including data in multiple modes such as numerical, text, and image.

[0022] Based on the constructed power network structure diagram, the extracted features are integrated, where nodes and edges represent the mutual influence and interactive connection relationships between entities in the power system; a deep learning model suitable for processing heterogeneous data and graph structured data is selected. The deep learning model needs to have the ability to process multi-modal data and be able to capture the complex relationships in the graph structure; data of different modalities are integrated in the graph model. This can be achieved by converting data of different modalities into a unified representation and performing joint learning in the deep learning model. During the fusion process, the mutual influence and dependency between data of different modalities need to be fully considered.

[0023] Similar to processing high-dimensional data, the dynamic evolution analysis of nodes and edges in the power network structure diagram is performed; a deep learning model is used to capture the changing patterns of the power system in the time dimension, especially the complexity and uncertainty caused by heterogeneous data.

[0024] After multimodal graph fusion calculation and dynamic evolution analysis, the collected data set is called the second deep dynamic evolution set. The second deep dynamic evolution set not only contains the dynamic change information of the power system in the time dimension, but also integrates heterogeneous data from different sources, providing a more comprehensive description of the operating status of the power system.

[0025] Multimodal graph fusion calculation based on mutual influence and interactive connection relationships processes heterogeneous data in the power system, comprehensively captures the operating status and change laws of the power system, and collects the second deep dynamic evolution set to provide richer and more accurate data support for subsequent power system analysis, prediction and optimization decisions.

[0026] S5: Taking the grid operation status as the starting point, the dynamic changes of the power system in the time dimension and the space dimension are analyzed through the first deep dynamic evolution set and the second deep dynamic evolution set to extract the perception prediction trend.

[0027] Taking the grid operation status as the starting point, through the first deep dynamic evolution set and the second deep dynamic evolution set, the dynamic changes of the power system in the time dimension and the space dimension are deeply analyzed, and the perception prediction trend is extracted. Specifically, the current grid operation status is taken as the starting point of the analysis. The grid operation status includes key indicators such as power generation, load level, voltage stability, and frequency; various operation parameters of the grid are monitored in real time through sensor networks, monitoring systems, and data acquisition systems to ensure the accuracy and real-time nature of the data.

[0028] The first deep dynamic evolution set contains the dynamic change information obtained by processing high-dimensional data in the power system, reflecting the state changes of the power system at different time points and the mutual influence and interactive connection between entities; the second deep dynamic evolution set further integrates heterogeneous data from different sources, providing a more comprehensive description of the operating status of the power system, including numerical data, and also contains information in multiple modes such as text and images.

[0029] In the time dimension, the deep learning model is used to analyze the time series data in the first deep dynamic evolution set and the second deep dynamic evolution set to capture the dynamic change law of the power system in the time dimension, analyze the change trend of the power grid state at different time points, and predict future state changes. In the spatial dimension, combined with the power network structure diagram, the layout and connection relationship of the power grid in the spatial dimension are analyzed, the mutual influence and dependency between different regions and different nodes are evaluated, and potential vulnerabilities and fault propagation paths are identified.

[0030] By comprehensively analyzing the dynamic change information in the time dimension and space dimension, the key trends in the power system are identified, including load growth trends, failure probability, changes in energy supply and demand, etc.; based on the identified trends, a prediction model is established to predict the future state of the power system. The prediction model needs to consider a variety of influencing factors and uncertainty factors to provide accurate prediction results; the prediction results are compared and verified with the actual grid operation status, and the accuracy and reliability of the prediction model are continuously optimized; through real-time perception and feedback mechanisms, the power dispatch strategy is adjusted in time and the grid operation parameters are optimized to ensure the stability and reliability of the grid.

[0031] Taking the grid operation status as the starting point, through the analysis of the first deep dynamic evolution set and the second deep dynamic evolution set, the dynamic changes of the power system in the time dimension and the space dimension are comprehensively captured, and the perception prediction trend is extracted to provide strong support for the optimal scheduling, fault prediction and stable operation of the power system.

[0032] S6: Based on the power network structure diagram, a reinforcement learning algorithm is deployed to perform adaptive strategy optimization learning according to the real-time updated power grid operation status combined with the perceived prediction trend, and match the optimal power dispatch decision with the goal of minimizing energy loss.

[0033] In order to further improve the efficiency and stability of the power system, a reinforcement learning algorithm is deployed on the basis of the power network structure diagram to optimize the power dispatching decision. Specifically, the reinforcement learning algorithm is deployed. Furthermore, the power network structure diagram and its dynamic changes (including the first deep dynamic evolution set and the second deep dynamic evolution set) constitute the reinforcement learning environment, which contains various states (such as the load level of the power grid, voltage stability, power generation of power plants, etc.) and actions with a certain probability (such as adjusting generator output, switching transmission lines, starting or stopping backup power supplies, etc.).

[0034] In a feasible implementation, the real-time monitoring system continuously collects the latest data on the operation status of the power grid and feeds it back to the reinforcement learning algorithm; combined with the extracted perception prediction trend, the reinforcement learning algorithm can more accurately understand the current operation status of the power grid and future change trends; the reinforcement learning algorithm performs adaptive strategy optimization learning based on the current status and perception prediction trend. The goal of the reinforcement learning algorithm is to find a series of actions that can minimize energy loss and meet other operation constraints of the power grid (such as voltage stability, frequency control, etc.); according to the optimized strategy, the corresponding actions are executed to adjust the operation status of the power grid. The real-time monitoring system collects the results of the action execution and returns it to the reinforcement learning algorithm as a feedback signal; the reinforcement learning algorithm adjusts its internal parameters and strategies based on the feedback signal to continuously improve the future decision-making process.

[0035] The objective function of the reinforcement learning algorithm is set to minimize energy loss, which means that the algorithm will give priority to reducing unnecessary energy waste and improving energy efficiency when optimizing decisions. In addition to minimizing energy loss, the reinforcement learning algorithm can also consider other goals (such as improving the stability of the power grid, reducing the probability of failure, etc.) to achieve multi-objective optimization; by introducing weight factors in the objective function or adopting a hierarchical optimization method, a balance can be achieved between different goals. The reinforcement learning algorithm has the ability to adapt dynamically. As the operating status of the power grid continues to change, it can adjust its strategy in time to adapt to the changing supply and demand conditions and power grid operating conditions. Dynamic adaptability ensures the flexibility and effectiveness of power dispatch decisions, which helps to improve the overall performance and reliability of the power grid.

[0036] In summary, deploying reinforcement learning algorithms based on the power network structure diagram and performing adaptive strategy optimization learning based on the real-time updated power grid operation status and perceived prediction trends is an effective way to minimize energy loss and match optimal power dispatching decisions, making full use of the complex data and dynamic characteristics of the power system and improving the intelligent management level and operation efficiency of the power grid.

[0037] Furthermore, based on the mutual influence relationship and the interactive connection relationship, a deep learning architecture is used to perform multimodal graph fusion calculation to process high-dimensional data in the power system and collect a first deep dynamic evolution set. The method of the present application also includes: Obtain fault history records, mark them on the power network structure diagram, add fault type marks, the fault types corresponding to the fault type marks include short circuit, open circuit, overload, and equipment failure; based on the fault type marks, extract associated factors with the fault start time and the fault end time as constraints, and obtain a set of fault occurrence associated factors, the set of fault occurrence associated factors corresponds to weather conditions, load change factors, and equipment operating status factors during the fault period; construct a fault analysis model, take the set of fault occurrence associated factors as input, use machine learning algorithms for in-depth analysis, and identify external fault influences, internal fault influences, and interaction mechanisms that affect the occurrence of faults; establish a closed-loop fault management system based on the fault analysis model and the external fault influences, internal fault influences, and interaction mechanisms that affect the occurrence of faults.

[0038] Process high-dimensional data in the power system and build a closed-loop fault management system. Specifically, based on the mutual influence and interactive connection between entities in the power system, a deep learning architecture is used for multimodal graph fusion calculations, aiming to process high-dimensional data in the power system and collect the first deep dynamic evolution set, which reflects the dynamic change characteristics of the power system in the time dimension and space dimension. Obtain the fault history of the power system and mark the fault type on the power network structure diagram. These marks include different types of fault types, such as short circuit, open circuit, overload, equipment failure, etc. Through the fault type marking, the common fault modes and their distribution in the power system can be clearly identified.

[0039] Based on the fault type label, correlation factors are extracted with the fault start time and fault end time as constraints. This step aims to identify various factors closely related to the occurrence of the fault, such as weather conditions during the fault period, load change factors, equipment operating status factors, etc. By organizing these factors into a set of fault-related factors, rich data support is provided for subsequent fault analysis.

[0040] A fault analysis model is constructed. The fault analysis model takes a set of fault-related factors as input, uses a machine learning algorithm to perform in-depth analysis, and then identifies the external fault influences and internal fault influences that affect the occurrence of faults, as well as the interaction mechanism between them. This is of great significance for understanding the root causes of faults and formulating effective fault prevention measures.

[0041] Based on the fault analysis model and the various factors and mechanisms that affect the occurrence of faults identified by it, a closed-loop fault management system is established. The closed-loop fault management system includes key links such as quantitative evaluation of fault impact contribution, fault risk prediction, dynamic update of power network structure diagram, establishment of real-time feedback mechanism, and continuous adjustment of fault maintenance strategy, so as to achieve comprehensive monitoring, early warning and response to power system faults and ensure the stability and reliability of the power grid.

[0042] Furthermore, based on the fault analysis model and the external fault influences, internal fault influences, and interaction mechanisms that affect the occurrence of faults, a closed-loop fault management system is established. The method of the present application includes: A quantitative evaluation is performed based on the fault analysis model and the external fault influences and internal fault influences that affect the occurrence of the fault, and the contribution of the fault to the occurrence is determined; the contribution of the fault to the occurrence is used to predict the fault risk under the constraint of the interaction mechanism in combination with the real-time updated power grid operation status to obtain the fault prediction type and fault prediction probability; according to the fault prediction type and fault prediction probability, they are incorporated into the dynamic update of the power network structure diagram, and at the same time, a real-time feedback mechanism is established and the fault maintenance strategy is continuously adjusted to form a closed-loop fault management system.

[0043] Through the fault analysis model and the quantitative evaluation of external and internal fault effects, the prediction of fault risks and the adjustment of dynamic maintenance strategies are realized. Specifically, the collected fault history records are deeply analyzed using the fault analysis model. The fault analysis model can identify external factors (such as weather conditions and load changes) and internal factors (such as equipment operating status and maintenance level) that affect the occurrence of faults, and reveal the interaction mechanism between them. Through quantitative evaluation, the contribution of each external fault influence and internal fault influence to the occurrence of faults is determined. Quantifying the contribution of fault occurrence is crucial to understanding the root cause of the fault and identifying key influencing factors.

[0044] Combined with real-time updated grid operation status data, including key indicators such as power generation, load level, and voltage stability, real-time information is provided for fault risk prediction. Based on the known contribution of each influencing factor, the constraint conditions of the interaction mechanism are used to predict the fault risk, including predicting the fault types with a certain probability of occurrence (such as short circuit, open circuit, overload, etc.) and the corresponding fault prediction probability.

[0045] Incorporating information such as fault prediction type and fault prediction probability into the dynamic update of the power network structure diagram means marking potential fault areas and risk points on the power network structure diagram for real-time monitoring and rapid response. Based on the dynamically updated power network structure diagram, the operation strategy and dispatch plan of the power grid are optimized to reduce the probability of faults and reduce the impact of faults on the operation of the power grid.

[0046] Establishing a real-time feedback mechanism to compare and verify the real-time data of the power grid operation status with the fault prediction results will help to timely discover prediction errors and adjust the prediction model to improve the accuracy of the prediction. Based on the real-time feedback information, the fault maintenance strategy is continuously adjusted, including optimizing the maintenance plan, improving the equipment maintenance level, and strengthening the monitoring of key areas.

[0047] Through the cyclic execution of the above steps, a closed-loop fault management system is formed to achieve comprehensive monitoring, prediction and dynamic management of power grid faults, ensuring the stability and reliability of the power grid. With the continuous accumulation of power grid operation data and the continuous optimization of analysis models, the closed-loop fault management system will be able to continuously improve its prediction accuracy and management efficiency, and achieve accurate prediction and effective management of power grid faults.

[0048] Furthermore, the present application method also includes: Based on the fault type label, with directly affected nodes and indirectly affected nodes as constraints, correlation factors are extracted to obtain a set of fault distribution correlation factors, wherein the set of fault occurrence correlation factors corresponds to the fault propagation path factors affected by the spatial distribution of power, the load change factors of the affected area, and the backup power supply investment factors; using the set of fault distribution correlation factors as constraints, a graph neural network is used to simulate the fault propagation process in the power grid, and a fault propagation architecture is established, wherein the fault analysis model is trained based on the fault propagation architecture.

[0049] We have conducted in-depth research on the propagation of faults in the power grid and their impact. Furthermore, based on historical fault records and the current operating status of the power grid, we have marked the nodes in the power network structure diagram (such as power plants, substations, and transmission lines) with fault types. The fault types corresponding to the fault type markings include short circuit, open circuit, overload, equipment failure, etc.

[0050] Based on the fault type label, the correlation factors related to the fault distribution are further extracted. The correlation factors include constraints at two levels: directly affecting nodes and indirectly affecting nodes. Directly affecting nodes are directly related to the occurrence of faults, such as the transmission line or substation where the fault point is located; for directly affecting nodes, the correlation factors include equipment aging, load mutation, etc. Although the indirect affecting nodes do not directly fail, they are obviously affected by the fault, such as the adjacent substation connected to the fault point or the adjacent area where the load is transferred due to the fault; for indirectly affecting nodes, the correlation factors include the fault propagation path affected by the spatial distribution of power, the load change in the affected area, the input of backup power supply, etc. The above correlation factors are organized into a set of fault distribution correlation factors. The set of fault distribution correlation factors not only describes the direct impact area of ​​the fault, but also reveals the propagation path and potential impact range of the fault in the power grid.

[0051] Using the extracted set of fault distribution correlation factors as constraints, the Graph Neural Network is used to simulate the propagation process of faults in the power grid. The Graph Neural Network is good at processing graph structure data, can capture the complex relationship between nodes and predict its future change trend. Through the simulation calculation of the Graph Neural Network, a fault propagation architecture is established, which describes the entire process from the fault occurrence point to the diffusion path and then to the final impact range. It not only reflects the spatial distribution characteristics of the fault in the power grid, but also reveals the impact of the fault on the operation status and stability of the power grid. The established fault propagation architecture provides a model basis for the training of the fault analysis model; the fault analysis model uses this architecture to train and learn a large amount of historical fault data to improve its prediction and analysis capabilities for new faults. Overall, it helps to reduce the impact of faults on the operation of the power grid and improve the stability and reliability of the power grid.

[0052] Furthermore, the present application method also includes: Based on the fault propagation architecture, the grid vulnerability index defined by the fault type mark is evaluated by comparing the nodes and edges in the power network structure diagram; the potential impact range is combined with the grid vulnerability index defined by the fault type mark, the renewable energy power generation unit is connected, and the GAE encoder is used to extract the renewable energy power generation characteristics; through the renewable energy power generation characteristics, the grid energy combination and energy scheduling strategy are optimized, and returned to the renewable energy power generation unit, and the renewable energy power generation unit is bound to the closed-loop fault management system.

[0053] The evaluation of grid vulnerability and the access and optimization strategies of renewable energy generation units aim to improve the stability and reliability of the grid through the analysis of the fault propagation architecture. Furthermore, according to the established fault propagation architecture, a detailed vulnerability assessment is conducted against the nodes and edges in the power network structure diagram, taking into account various factors limited by the fault type label, such as the fault propagation path, the scope of influence, the criticality of nodes and edges, etc. The vulnerability index of each node and edge is calculated by comprehensively considering factors such as the probability of fault occurrence, the scope of fault influence, the criticality of nodes and edges, etc. The vulnerability index of each node and edge reflects the vulnerability of the grid under a specific fault type.

[0054] The potential impact range and the grid vulnerability index are combined for analysis to identify the most vulnerable areas and nodes in the grid with the widest potential impact range. Based on the results of the combined analysis, the priority of grid maintenance and transformation is determined, with priority given to nodes and areas with high vulnerability index and wide potential impact range.

[0055] Connect renewable energy generation units. Specifically, plan the connection locations of renewable energy generation units based on the results of vulnerability assessment and potential impact range analysis; give priority to those locations that can improve grid stability and reduce vulnerability. Integrate renewable energy generation units with the grid to ensure that they can operate stably and efficiently in the grid.

[0056] The GAE encoder is used to extract the renewable energy generation features. Specifically, the GAE encoder (GraphAutoencoder) is introduced as a feature extraction tool; the GAE encoder can capture the complex relationship between the renewable energy generation unit and the power grid and extract its key features. The GAE encoder is used to process the operating data of the renewable energy generation unit to extract its power generation features, stability features, and interaction features with other nodes.

[0057] Based on the characteristics of renewable energy generation, the energy mix of the power grid is optimized; the complementarity between different renewable energy generation units and the coordination between them and traditional energy generation units are considered to achieve efficient energy utilization and stable operation of the power grid. According to the optimized energy mix, a more reasonable energy dispatching strategy is formulated. The energy dispatching strategy should be able to fully consider the real-time operating status of the power grid, load changes, and the uncertainty of renewable energy generation, so as to ensure that the power grid can maintain stable operation under various working conditions.

[0058] The optimized energy mix and dispatch strategy are returned to the renewable energy generation unit and the power grid operation system for real-time application and adjustment. Based on the real-time data feedback of the power grid operation status and the renewable energy generation unit, a closed-loop control system is formed, and the energy mix and dispatch strategy are continuously adjusted and optimized, which significantly improves the stability and reliability of the power grid and promotes the effective use and development of renewable energy.

[0059] Furthermore, the application method includes: Grid safety indicators are introduced, including grid voltage qualification rate, grid stability index, N-1 safety verification index, N-2 safety verification index, line thermal stability limit and static stability limit. Graph embedding technology is used to integrate the grid safety indicators into the power network structure diagram to enhance the safety characteristics of nodes and edges. Through graph neural networks, risk assessments are performed on key nodes and key edges respectively to identify potential safety risks, and the closed-loop fault management system is adaptively optimized.

[0060] The introduction of power grid security indicators and the combination of graph embedding technology to enhance security features, as well as the use of graph neural networks for risk assessment and optimization, improve the security and stability of the power grid. Furthermore, in order to comprehensively evaluate the safety performance of the power grid, multiple power grid safety indicators are introduced, including but not limited to: power grid voltage qualification rate (a key indicator for measuring power grid voltage stability), power grid stability index (reflecting the ability of the power grid to maintain stable operation after being disturbed), N-1 safety verification index (evaluating whether the power grid can still maintain normal power supply in the event of a single component failure, single components are common such as transmission lines and transformers), N-2 safety verification index (further evaluating the power supply capacity and stability of the power grid when two components fail at the same time), line thermal stability limit (considering the line current carrying capacity and its stable operation limitations in high temperature environments), static stability limit (evaluating the stability boundary of the power grid under static conditions to prevent system collapse), so as to comprehensively evaluate the redundancy and risk resistance of the power grid and ensure that the power grid has higher reliability and stability.

[0061] Security features are enhanced using graph embedding technology, which can embed complex relationships and high-dimensional data of the power grid into a low-dimensional space while retaining important information in the original graph structure. By integrating the power grid security indicators into the power network structure diagram, the security features of nodes and edges are enhanced, which can be achieved by calculation in the embodiment of the present application, taking into account the original security features of the nodes, the dynamic weights of each security indicator, the security features of adjacent nodes, and the network topology relationship between them.

[0062] Using graph neural networks to conduct in-depth analysis of the enhanced power network structure diagram, GNN can capture the complex relationships between nodes and edges, as well as their mutual influence. By conducting risk assessment on nodes and edges, key nodes and edges in the power grid are identified; these nodes and edges play an important role in the operation of the power grid, and once a failure or attack occurs, it may have a serious impact on the power grid. Based on the risk assessment results, potential security risks in the power grid are identified, which are caused by various factors such as equipment aging, load mutations, and changes in the external environment.

[0063] According to the risk assessment results and the potential safety risks identified, formulate corresponding optimization strategies, including strengthening the monitoring and maintenance of key nodes, optimizing energy dispatch strategies, improving grid redundancy, etc. Apply the optimization strategies to grid operation, and continuously adjust and optimize the strategies through real-time feedback mechanisms. In this step, the closed-loop fault management system will continuously optimize adaptively according to the changes in grid operation status and safety performance, ensure the safe and stable operation of the grid, and provide strong guarantees for the reliable operation of the power system.

[0064] Furthermore, the grid security index is integrated into the power network structure graph using graph embedding technology to enhance the security features of nodes and edges. The method of this application includes: Grid security index fusion formula: ;in, is the updated security feature of node v, is the original security feature of node v, is the dynamic weight of the kth security indicator, is the value of the kth power grid security index corresponding to node v, N(v) is the set of adjacent nodes to node v, represents the network topology correlation coefficient between node v and adjacent node j, is the security feature of neighboring node j; is the weight function between node v and adjacent node j at time t, is the basic connection weight between node v and adjacent node j, is the decay rate, t is the current time, is the time of the last interaction. The grid security indicators are integrated into the power network structure diagram using graph embedding technology to enhance the security characteristics of nodes and edges. This involves the grid security indicator fusion formula, which details how to combine the original security characteristics of the node, the values ​​of each security indicator, the security characteristics of adjacent nodes, and the network topology and time relationship between them to calculate the updated security characteristics of the node.

[0065] By integrating the power grid security index into the power network structure diagram, the security features of nodes and edges are enhanced. This can be achieved by calculation in the embodiment of the present application. Further, the power grid security index fusion formula is configured: ,in, is the dynamic weight of the kth security indicator, reflecting the importance of different security indicators in evaluating the security characteristics of the node; is the value of the kth power grid safety index corresponding to node v, such as the grid voltage qualification rate, grid stability index, etc.; N(v) is the set of adjacent nodes to node v, indicating the connection relationship of node v in the network; It represents the network topological correlation coefficient between node v and adjacent node j, reflecting the closeness and importance of the connection between node v and adjacent node j. is the weight function between node v and adjacent node j at time t, which is used to consider the impact of time factors on the relationship between nodes; is the basic connection weight between node v and adjacent node j, indicating the static connection strength between them; is the decay rate, which reflects how fast the relationship between nodes changes over time; t is the current time, is the time of the last interaction, which is used to calculate the time difference and affect the value of the weight function.

[0066] As the original security feature of node v, it is retained in the updated security feature; by weighted summing the values ​​of each security index And multiply by the corresponding dynamic weight , integrating multiple power grid security indicators into the security characteristics of node v. Considering the mutual influence between nodes and the temporal dynamics of their relationships, the network topology correlation coefficient and time weight function , the security features of the adjacent node j Fusion into the security features of node v. Weight function The introduction of time factors enables the fusion of security features between nodes to reflect the real-time changes in the operation status of the power grid. and consider the last interaction time ,Precisely control the impact of time factors on the relationship between nodes.,Effectively integrate power grid security indicators into the power network structure diagram,,enhance the security feature representation capabilities of nodes and edges, and more accurately,evaluate the stability and reliability of the power grid.

[0067] In summary, the beneficial effects of the embodiments of the present application are: 1. Intuitively displaying the entities and their logical connections in the power system helps to understand the operating mechanism of the entire system, and process high-dimensional and heterogeneous data through deep learning technology, perform multi-modal graph fusion calculations, efficiently process high-dimensional and heterogeneous data in the power system, and improve the timeliness of data analysis.

[0068] 2. Based on the real-time update of the grid operation status and the perception and prediction trend, adaptively perform strategy optimization learning and match the optimal power dispatch decision, thereby minimizing energy loss and improving the overall performance of the system.

[0069] 3. By introducing grid security indicators and using graph embedding technology to enhance the security features of nodes and edges, analyzing the dynamic changes of the power system in time and space dimensions, identifying and evaluating potential security risks, and adaptively optimizing the closed-loop fault management system, faults in the power system can be effectively predicted and managed, achieving accurate prediction and prevention of faults and improving the stability and reliability of the power grid.

[0070] 4. Due to the adoption of the grid security index fusion formula: ;in, is the updated security feature of node v, is the original security feature of node v, is the dynamic weight of the kth security indicator, is the value of the kth power grid security index corresponding to node v, N(v) is the set of adjacent nodes to node v, represents the network topology correlation coefficient between node v and adjacent node j, is the security feature of neighboring node j; is the weight function between node v and adjacent node j at time t, is the basic connection weight between node v and adjacent node j, is the decay rate, t is the current time, is the time of the last interaction. Effectively integrate the grid security indicators into the power network structure diagram, enhance the security feature representation capability of nodes and edges, and more accurately evaluate the stability and reliability of the grid.

[0071] Embodiment 2 Based on the same inventive concept as the power data analysis method based on graph calculation in the aforementioned embodiment, Figure 2 As shown, an embodiment of the present application provides a power data analysis device based on graph calculation, wherein the device includes: The structure diagram construction module M100 is used to construct a power network structure diagram, in which nodes represent entities in the power system, including power plants, substations, and transmission lines, and edges represent logical connection relationships between entities; The feature extraction module M200 is used to extract features from nodes and edges in the power network structure diagram to capture the mutual influence relationship and interactive connection relationship between entities in the power system; A first fusion calculation module M300 is used to perform multimodal graph fusion calculation based on the mutual influence relationship and the interactive connection relationship using a deep learning architecture to process high-dimensional data in the power system and collect a first deep dynamic evolution set; The second fusion calculation module M400 is used to perform multimodal graph fusion calculation based on the mutual influence relationship and the interactive connection relationship using a deep learning architecture to process heterogeneous data in the power system and collect a second deep dynamic evolution set; The dynamic change analysis module M500 is used to analyze the dynamic changes of the power system in the time dimension and the space dimension by taking the power grid operation state as the starting point through the first deep dynamic evolution set and the second deep dynamic evolution set, and extract the perception prediction trend; The dispatching decision module M600 is used to deploy a reinforcement learning algorithm based on the power network structure diagram, perform adaptive strategy optimization learning according to the real-time updated power grid operation status and the perceived prediction trend, and match the optimal power dispatching decision with the goal of minimizing energy loss.

[0072] Furthermore, the first fusion calculation module M300 is used to perform the following method: Obtaining fault history records, marking them on the power network structure diagram, and adding fault type marks, wherein the fault types corresponding to the fault type marks include short circuit, open circuit, overload, and equipment failure; Based on the fault type tag, taking the fault time start point and the fault time end point as constraints, extracting correlation factors to obtain a set of fault occurrence correlation factors, wherein the set of fault occurrence correlation factors corresponds to weather condition factors, load change factors, and equipment operation status factors during the fault period; Construct a fault analysis model, take the set of fault-related factors as input, use machine learning algorithms to perform in-depth analysis, and identify external fault influences, internal fault influences, and interaction mechanisms that affect the occurrence of faults; Based on the fault analysis model and the external fault influences, internal fault influences, and interaction mechanisms that affect the occurrence of faults, a closed-loop fault management system is established.

[0073] Furthermore, the first fusion calculation module M300 is also used to execute the following method: Based on the fault analysis model, a quantitative evaluation is performed on the external fault influence and the internal fault influence that affect the occurrence of the fault, and the contribution degree that affects the occurrence of the fault is determined; Using the contribution degree affecting the occurrence of the fault, under the constraint of the interaction mechanism, combined with the real-time updated power grid operation status, the fault risk prediction is performed to obtain the fault prediction type and fault prediction probability; According to the fault prediction type and fault prediction probability, they are incorporated into the dynamic update of the power network structure diagram. At the same time, a real-time feedback mechanism is established and the fault maintenance strategy is continuously adjusted to form a closed-loop fault management system.

[0074] Furthermore, the first fusion calculation module M300 is also used to execute the following method: Based on the fault type label, with direct impact nodes and indirect impact nodes as constraints, correlation factors are extracted to obtain a set of fault distribution correlation factors, where the set of fault occurrence correlation factors corresponds to the fault propagation path factors affected by the power spatial distribution, the load change factors of the affected area, and the backup power supply investment factors; The set of fault distribution-related factors is used as constraint conditions, and a graph neural network is used to simulate the fault propagation process in the power grid to establish a fault propagation framework, wherein the fault analysis model is trained based on the fault propagation framework.

[0075] Furthermore, the first fusion calculation module M300 is also used to execute the following method: Based on the fault propagation architecture, comparing the nodes and edges in the power network structure diagram, evaluating the power grid vulnerability index under the fault type labeling limitation; The potential impact range is combined with the grid vulnerability index defined by the fault type mark, the renewable energy power generation unit is connected, and the renewable energy power generation characteristics are extracted using the GAE encoder; Through the renewable energy power generation characteristics, the power grid energy combination and energy dispatch strategy are optimized and returned to the renewable energy power generation unit, and the renewable energy power generation unit is bound to the closed-loop fault management system.

[0076] Furthermore, the first fusion calculation module M300 is also used to execute the following method: Introducing power grid safety indicators, which include power grid voltage qualification rate, power grid stability index, N-1 safety verification index, N-2 safety verification index, line thermal stability limit and static stability limit; Using graph embedding technology, the power grid security indicators are integrated into the power network structure graph to enhance the security characteristics of nodes and edges; Through graph neural networks, risk assessments are performed on key nodes and key edges respectively to identify potential safety risks, and the closed-loop fault management system is adaptively optimized.

[0077] Furthermore, the first fusion calculation module M300 is also used to execute the following method: Power grid security index fusion formula: ;in, is the updated security feature of node v, is the original security feature of node v, is the dynamic weight of the kth security indicator, is the value of the kth power grid security index corresponding to node v, N(v) is the set of adjacent nodes to node v, represents the network topology correlation coefficient between node v and adjacent node j, is the security feature of neighboring node j; is the weight function between node v and adjacent node j at time t, is the basic connection weight between node v and adjacent node j, is the decay rate, t is the current time, is the time of the last interaction. In summary, any step can be stored as a computer instruction or program in an unlimited computer memory, and can be called and recognized by an unlimited computer processor, and no unnecessary restrictions are made here.

[0078] Furthermore, the above-mentioned technical scheme only reflects the preferred technical scheme of the technical scheme of the embodiment of the present application. Some changes that may be made to some parts thereof by technicians in this technical field all reflect the novel principles of the embodiment of the present application. Obviously, technicians in this field can make various changes and modifications to the present application without departing from the scope of the present application.

Claims

1. A power data analysis method based on graph computing, characterized in that: The method comprises: Construct a power network structure diagram, where nodes represent entities in the power system, including power plants, substations, and transmission lines, and edges represent logical connections between entities; Extracting features from nodes and edges in the power network structure diagram to capture the mutual influence and interactive connection relationships between entities in the power system; Based on the mutual influence relationship and the interactive connection relationship, a deep learning architecture is used to perform multimodal graph fusion calculation to process high-dimensional data and heterogeneous data in the power system, and collect a first deep dynamic evolution set and a second deep dynamic evolution set; Taking the grid operation status as the starting point, the dynamic changes of the power system in the time dimension and the space dimension are analyzed through the first deep dynamic evolution set and the second deep dynamic evolution set, and the perception prediction trend is extracted; Based on the power network structure diagram, a reinforcement learning algorithm is deployed to perform adaptive strategy optimization learning according to the real-time updated power grid operation status combined with the perceived prediction trend, and match the optimal power dispatch decision with the goal of minimizing energy loss.

2. The power data analysis method based on graph computing according to claim 1, characterized in that: Based on the mutual influence relationship and the interactive connection relationship, a multimodal graph fusion calculation is performed using a deep learning architecture to process high-dimensional data in the power system and collect a first deep dynamic evolution set. The method also includes: Obtaining fault history records, marking them on the power network structure diagram, and adding fault type marks, wherein the fault types corresponding to the fault type marks include short circuit, open circuit, overload, and equipment failure; Based on the fault type tag, taking the fault time start point and the fault time end point as constraints, extracting correlation factors to obtain a set of fault occurrence correlation factors, wherein the set of fault occurrence correlation factors corresponds to weather condition factors, load change factors, and equipment operation status factors during the fault period; Construct a fault analysis model, take the set of fault-related factors as input, use machine learning algorithms to perform in-depth analysis, and identify external fault influences, internal fault influences, and interaction mechanisms that affect the occurrence of faults; Based on the fault analysis model and the external fault influences, internal fault influences, and interaction mechanisms that affect the occurrence of faults, a closed-loop fault management system is established.

3. The power data analysis method based on graph computing according to claim 2, characterized in that: Based on the fault analysis model and the external fault influences, internal fault influences, and interaction mechanisms that affect the occurrence of faults, a closed-loop fault management system is established, and the method includes: Based on the fault analysis model, a quantitative evaluation is performed on the external fault influence and the internal fault influence that affect the occurrence of the fault, and the contribution degree that affects the occurrence of the fault is determined; Using the contribution degree affecting the occurrence of the fault, under the constraint of the interaction mechanism, combined with the real-time updated power grid operation status, the fault risk prediction is performed to obtain the fault prediction type and fault prediction probability; According to the fault prediction type and fault prediction probability, they are incorporated into the dynamic update of the power network structure diagram. At the same time, a real-time feedback mechanism is established and the fault maintenance strategy is continuously adjusted to form a closed-loop fault management system.

4. The power data analysis method based on graph computing according to claim 3, characterized in that: The method further comprises: Based on the fault type label, with direct impact nodes and indirect impact nodes as constraints, correlation factors are extracted to obtain a set of fault distribution correlation factors, where the set of fault occurrence correlation factors corresponds to the fault propagation path factors affected by the power spatial distribution, the load change factors of the affected area, and the backup power supply investment factors; The set of fault distribution-related factors is used as constraint conditions, and a graph neural network is used to simulate the fault propagation process in the power grid to establish a fault propagation framework, wherein the fault analysis model is trained based on the fault propagation framework.

5. The power data analysis method based on graph computing according to claim 4, characterized in that: The method further comprises: Based on the fault propagation architecture, comparing the nodes and edges in the power network structure diagram, evaluating the power grid vulnerability index under the fault type label definition; The potential impact range is combined with the grid vulnerability index defined by the fault type mark, the renewable energy power generation unit is connected, and the renewable energy power generation characteristics are extracted using the GAE encoder; Through the renewable energy power generation characteristics, the power grid energy combination and energy dispatch strategy are optimized and returned to the renewable energy power generation unit, and the renewable energy power generation unit is bound to the closed-loop fault management system.

6. The power data analysis method based on graph computing according to claim 5, characterized in that: The method comprises: Introducing power grid safety indicators, which include power grid voltage qualification rate, power grid stability index, N-1 safety verification index, N-2 safety verification index, line thermal stability limit and static stability limit; Using graph embedding technology, the power grid security indicators are integrated into the power network structure graph to enhance the security characteristics of nodes and edges; Through graph neural networks, risk assessments are performed on key nodes and key edges respectively to identify potential safety risks, and the closed-loop fault management system is adaptively optimized.

7. The power data analysis method based on graph computing according to claim 6, characterized in that: Using graph embedding technology, the power grid security indicator is integrated into the power network structure graph to enhance the security characteristics of nodes and edges. The method includes: Grid security index fusion formula: ;in, is the updated security feature of node v, is the original security feature of node v, is the dynamic weight of the kth security indicator, is the value of the kth power grid security index corresponding to node v, N(v) is the set of adjacent nodes to node v, represents the network topology correlation coefficient between node v and adjacent node j, is the security feature of neighboring node j; is the weight function between node v and adjacent node j at time t, is the basic connection weight between node v and adjacent node j, is the decay rate, t is the current time, Is the time of the last interaction.

8. A power data analysis device based on graph computing, characterized in that: The device for implementing the power data analysis method based on graph calculation according to any one of claims 1 to 7 comprises: The structure diagram construction module is used to construct the power network structure diagram, where nodes represent entities in the power system, including power plants, substations, and transmission lines, and edges represent the logical connection relationship between entities; A feature extraction module is used to extract features from nodes and edges in the power network structure diagram to capture the mutual influence and interactive connection relationships between entities in the power system; A fusion calculation module is used to perform multimodal graph fusion calculation based on the mutual influence relationship and the interactive connection relationship using a deep learning architecture to process high-dimensional data and heterogeneous data in the power system and collect a first deep dynamic evolution set and a second deep dynamic evolution set; A dynamic change analysis module is used to analyze the dynamic changes of the power system in the time dimension and the space dimension by taking the power grid operation state as the starting point and extracting the perception prediction trend through the first deep dynamic evolution set and the second deep dynamic evolution set; The scheduling decision module is used to deploy a reinforcement learning algorithm based on the power network structure diagram, perform adaptive strategy optimization learning according to the real-time updated power grid operation status and the perceived prediction trend, and match the optimal power scheduling decision with the goal of minimizing energy loss.

Images