Load intelligent dynamic regulation and control method and system for intelligent power grid

Abstract

The embodiment of the invention provides an intelligent dynamic load regulation and control method and system for an intelligent power grid. Wherein power parameters of key nodes are collected in real time through time sequence analysis and a deep learning algorithm, future power demand fluctuation is predicted, and a power grid topological structure model is constructed to identify the key nodes and a potential influence range; a genetic algorithm, particle swarm optimization and other multi-objective optimization algorithms are adopted to generate specific operation instructions for different power utilization areas or devices, and optimization objectives comprise minimization of energy cost, maximization of renewable energy utilization rate and guarantee of power supply stability; converting the strategies into guidance suggestions, pushing the guidance suggestions to a terminal user, collecting feedback information, generating a preference model by applying a reinforcement learning algorithm, and selecting an optimal strategy for implementation based on user acceptance and feasibility evaluation; risks and abnormal conditions are detected through the Bayesian network, and the strategy is adjusted in time. According to the technical scheme provided by the embodiment of the invention, the efficient distribution of power resources and the energy utilization efficiency in the smart power grid are improved.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of smart grid technology, and in particular, to a method and system for intelligent dynamic load control for a smart grid. Background Art

[0002] The load regulation of smart grids needs to cope with complex and changing power demand and supply environments. With the increase in the proportion of renewable energy and the widespread use of distributed energy resources (DERs), the dynamics and uncertainty of the power system have increased significantly. In order to ensure the efficient and stable operation of the power system, it is necessary to achieve real-time collection and analysis of power parameters at key nodes to accurately predict power demand fluctuations and build an accurate grid topology model to identify key nodes in the power transmission path and their potential impact range.

[0003] At present, load regulation in smart grids mainly relies on traditional shortest path algorithms and static optimization models.

[0004] Traditional load regulation schemes have several shortcomings. First, they have difficulty handling complex power demand forecasting problems, especially when faced with rapidly changing market conditions and unstable renewable energy inputs. Second, existing grid topology analysis methods are not sophisticated enough to fully reveal the key nodes of the power transmission path and their potential impact range, which limits the in-depth understanding of the characteristics of the grid structure. Furthermore, single-objective optimization strategies cannot take into account multiple mutually constrained objectives, which can easily lead to energy waste or sacrifice certain aspects of performance. Finally, the lack of an effective user interaction mechanism means that the feedback and preferences of end users are not fully utilized, which weakens the feasibility and acceptance of regulation measures. Summary of the invention

[0005] The embodiments of the present application provide a method and system for intelligent dynamic load control for a smart grid, which are used to solve the problems of uneven distribution of power resources and low energy utilization efficiency in the prior art.

[0006] In a first aspect, an embodiment of the present application provides a method for intelligent dynamic control of loads for a smart grid, comprising:

[0007] Using time series analysis and deep learning algorithms to collect real-time power parameters of key nodes in the smart grid, obtain collected data, and predict power demand fluctuations within a preset time based on the collected data to obtain prediction results, and using graph theory algorithms to build a power grid topology model, identify key nodes and potential impact ranges of power transmission paths in the smart grid, and obtain topology information;

[0008] According to the prediction results and the topological structure information, a multi-objective optimization algorithm is used in combination with the optimization objectives pre-set in the smart grid to generate a load adjustment strategy for specific operation instructions for different power consumption areas or equipment, wherein the multi-objective optimization algorithm includes a genetic algorithm and a particle swarm optimization, and the optimization objectives include minimizing energy costs, maximizing renewable energy utilization, and ensuring power supply stability;

[0009] The load adjustment strategy is converted into guidance suggestions and pushed to the end user, feedback information from the end user is collected, a reinforcement learning algorithm is applied to learn the long-term behavior patterns and preferences of the end user, a preference model including the end user's acceptance, feasibility evaluation results and personalized consulting solutions is generated, and based on the feedback information and the preference model, a corresponding load adjustment strategy is selected and implemented to obtain an execution effect;

[0010] The dynamic control system of the smart grid is used to monitor the execution effect and the operating status of the smart grid, and the risk of the execution effect and the abnormality of the operating status are detected in combination with the Bayesian network, the deviation from the optimization target is identified, and the execution effect is adjusted or switched to the backup load adjustment strategy.

[0011] Optionally, the use of a graph theory algorithm to construct a power grid topology model, identify key nodes and potential impact ranges of power transmission paths in the smart grid, and obtain topology information includes:

[0012] Based on the key nodes and connection relationships in the smart grid, geographic information system data is introduced, and a directed weighted graph is constructed using complex network theory to obtain an initial structural model of the smart grid. The connection probability distribution of the key nodes is initialized in the initial structural model, the uncertainty and randomness of the grid are simulated, and a time-varying grid topology with a time dimension attribute is generated;

[0013] Analyzing connectivity and redundancy of the smart grid according to the time-varying grid topology to obtain risk analysis results;

[0014] Based on the risk analysis results, the key nodes are redefined using a page ranking algorithm and a hyperlink induced topic search algorithm, and target nodes whose behavior patterns deviate from normal behavior patterns are identified according to an anomaly detection method, and the historical data of the target nodes are compared with the problem points predicted by the current situation to obtain evaluation results of the key nodes and the target nodes, wherein the anomaly detection method includes an isolation forest and a local anomaly factor;

[0015] Based on the evaluation results, Monte Carlo simulation combined with a propagation dynamics model is used to simulate the fault diffusion process of different node or path failure scenarios of the smart grid, estimate the impact range and recovery time of the fault, formulate an emergency response strategy, and use the emergency response strategy to integrate a deep reinforcement learning framework to generate solutions for specific types of faults;

[0016] Based on the solution, the entities, relationships and attributes in the smart grid are organized into a semantic information network, and the information network is visualized in the form of a map. Blockchain technology is deployed according to the map form, the status of the smart grid after each update is recorded, and the topological structure information of the smart grid is obtained.

[0017] Optionally, the analyzing the connectivity and redundancy of the smart grid further includes:

[0018] The maximum flow and minimum cut theorem is used to determine the global and local maximum flow and minimum cut set of the smart grid, evaluate the maximum carrying capacity and the most vulnerable point of the smart grid, and obtain the global and local flow distribution and key node information of the smart grid;

[0019] Utilizing the flow distribution and key node information, the shortest path between the key nodes and the average path length between all node pairs are calculated to generate a power grid performance evaluation report;

[0020] According to the power grid performance evaluation report, the connectivity analysis is extended using hypergraph theory, and the subsystems of the smart grid are identified through a community discovery algorithm.

[0021] Optionally, based on the risk analysis result, the key nodes are redefined using a page ranking algorithm and a hyperlink induced topic search algorithm, and a target node whose behavior pattern deviates from a normal behavior pattern is identified according to an anomaly detection method, and the historical data of the target node is compared with the problem points predicted by the current situation to obtain the evaluation results of the key nodes and the target node, including:

[0022] Based on the risk analysis results, the importance of each of the key nodes in the time-varying power grid topology is evaluated using a page ranking algorithm and a hyperlink-induced topic search algorithm to obtain a key node score;

[0023] According to the key node scores, the target nodes whose behavior patterns deviate from the normal behavior patterns are identified in combination with the anomaly detection method in machine learning, and a target node list is obtained;

[0024] Compare the historical data of the target node with the current situation, analyze the target node list, and generate evaluation results of the key node and the target node.

[0025] Optionally, the flow distribution and key node information are used to calculate the shortest path between the key nodes and the average path length between all node pairs to generate a power grid performance evaluation report, including:

[0026] Based on the traffic distribution and key node information of the smart grid, the connectivity between the key nodes is analyzed, and the maximum carrying capacity and the most vulnerable point of the smart grid are determined by the maximum flow minimum cut theorem to obtain a preliminary evaluation result;

[0027] Based on the preliminary evaluation results, the traffic distribution and key node information are used as inputs, and a graph theory algorithm is used to calculate the shortest path between the key nodes;

[0028] Based on the shortest path, calculating the average path length between all node pairs to obtain the average connection distance between any two nodes in the smart grid;

[0029] The shortest path data and the average connection distance are combined to generate a power grid performance evaluation report.

[0030] Optionally, the use of the dynamic control system of the smart grid to monitor the execution effect and the operating status of the smart grid, and combining the Bayesian network to detect the risk of the execution effect and the abnormality of the operating status, identify the deviation from the optimization target, and adjust the execution effect or switch to the standby load adjustment strategy, including:

[0031] Using the dynamic control system of the smart grid to monitor the execution effect and the operating status to obtain monitoring data, wherein the monitoring data includes the operating parameters of the smart grid, and the operating parameters of the smart grid include voltage, current and power;

[0032] Building a Bayesian network model based on the monitoring data, detecting the risk of the execution effect and the abnormality of the operation status, and generating a risk assessment result and an abnormality detection report, wherein the smart grid operation parameters serve as input variables of the Bayesian network model, and the risks and the abnormalities serve as output variables of the Bayesian network model;

[0033] According to the risk assessment result and the abnormality detection report, combined with the optimization goal, the deviation is identified to obtain a deviation judgment conclusion;

[0034] Based on the deviation judgment conclusion, the execution effect is adjusted to obtain an adjusted execution effect, and after the adjustment fails, switching to a standby load adjustment strategy.

[0035] Optionally, the multi-objective optimization algorithm is used in combination with the optimization objectives preset in the smart grid to generate a load adjustment strategy for specific operation instructions for different power consumption areas or equipment, including:

[0036] Selecting a multi-objective optimization algorithm based on the characteristics of the smart grid and a preset optimization goal, and initializing and processing parameters of the target optimization algorithm to obtain an optimization algorithm configuration;

[0037] Determine input information according to the prediction result, the topology information and the operation status of the smart grid, wherein the input information is key data generated by the multi-objective optimization algorithm for the current status and future trend of the smart grid;

[0038] Utilizing the optimization target, configuring the optimization algorithm and the input information to formulate a preliminary load adjustment strategy for different power consumption areas or equipment operation instructions;

[0039] The preliminary load adjustment strategy is optimized based on a simulated annealing algorithm and an iterative optimization algorithm to generate the load adjustment strategy.

[0040] Optionally, the using the traffic distribution and key node information to calculate the shortest path between the key nodes and the average path length between all node pairs includes:

[0041] Using the traffic distribution and key node information as input, a graph theory algorithm is used to calculate the shortest path between the key nodes, where the shortest path d ij (t) is calculated by the following formula:

[0042]

[0043] Among them, d ij (t) represents the shortest path length from key node i to key node j, represents the set of all paths from i to j, c(u,v) represents the transmission cost connecting nodes u and v, l(u,v) represents the transmission loss, t uv represents the transmission time from node u to node v, λ is the adjustment parameter, which reflects the impact of transmission time on weight, γ is the decay parameter, which reflects the change of node importance over time, t represents the current time point, u represents a starting node or intermediate node on the path, v represents the next node directly connected from node u, and P represents a specific path from a starting node i to a target node j;

[0044] Based on the shortest path data, the average path length between all node pairs is calculated, wherein the average path length Calculated by the following formula:

[0045]

[0046] Where N represents the total number of nodes in the power grid, represents the average path length, represents the number of combinations of selecting two different nodes from N nodes, r(u,v) represents the reliability of connecting nodes u and v, represents the number of different paths from i to j.

[0047] In a second aspect, an embodiment of the present application provides a load intelligent dynamic control system for a smart grid, including:

[0048] A collection module is used to collect real-time power parameters of key nodes in the smart grid using time series analysis and deep learning algorithms to obtain collected data, and predict power demand fluctuations within a preset time based on the collected data to obtain prediction results, and use graph theory algorithms to build a power grid topology model to identify key nodes and potential impact ranges of power transmission paths in the smart grid to obtain topology information;

[0049] A generation module, configured to generate a load adjustment strategy for specific operation instructions for different power consumption areas or devices by using a multi-objective optimization algorithm in combination with the optimization objectives pre-set in the smart grid according to the prediction results and the topological structure information, wherein the multi-objective optimization algorithm includes a genetic algorithm and a particle swarm optimization, and the optimization objectives include minimizing energy costs, maximizing renewable energy utilization, and ensuring power supply stability;

[0050] A feedback module is used to convert the load adjustment strategy into guidance suggestions and push them to the end user, collect feedback information from the end user, apply a reinforcement learning algorithm to learn the long-term behavior patterns and preferences of the end user, generate a preference model including the end user's acceptance, feasibility evaluation results and personalized consulting solutions, and select and implement the corresponding load adjustment strategy based on the feedback information and the preference model to obtain the execution effect;

[0051] The detection module utilizes the dynamic control system of the smart grid to monitor the execution effect and the operating status of the smart grid, and combines the Bayesian network to detect the risks of the execution effect and the abnormal conditions of the operating status, identify the deviation from the optimization target, and adjust the execution effect or switch to the backup load adjustment strategy.

[0052] In a third aspect, an embodiment of the present application provides a computing device, comprising a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a method for intelligent dynamic load control for a smart grid as described in the first aspect above.

[0053] In a fourth aspect, an embodiment of the present application provides a computer storage medium storing a computer program, which, when executed by a computer, implements a method for intelligent dynamic load control for a smart grid as described in the first aspect.

[0054] In an embodiment of the present application, time series analysis and deep learning algorithms are used to collect real-time power parameters of key nodes in a smart grid to obtain collected data, and power demand fluctuations within a preset time are predicted based on the collected data to obtain prediction results. A grid topology model is constructed using a graph theory algorithm to identify key nodes and potential impact ranges of power transmission paths in the smart grid to obtain topology information. Based on the prediction results and the topology information, a multi-objective optimization algorithm is used in combination with the optimization goals pre-set in the smart grid to generate a load adjustment strategy for specific operation instructions for different power consumption areas or equipment, wherein the multi-objective optimization algorithm includes a genetic algorithm and a particle swarm optimization, and the optimization goals include minimizing energy costs, maximizing renewable Energy utilization rate and ensuring power supply stability; converting the load adjustment strategy into guidance suggestions and pushing them to end users, collecting feedback information from the end users, applying reinforcement learning algorithms to learn the long-term behavior patterns and preferences of the end users, generating a preference model that includes end-user acceptance, feasibility assessment results and personalized consulting solutions, and selecting and implementing corresponding load adjustment strategies based on the feedback information and the preference model to obtain execution effects; using the dynamic control system of the smart grid to monitor the execution effects and the operating status of the smart grid, and combining the Bayesian network to detect the risks of the execution effects and the abnormalities of the operating status, identifying deviations from the optimization goals, and adjusting the execution effects or switching to backup load adjustment strategies.

[0055] The technical solution of this application has the following beneficial effects:

[0056] By accurately predicting power demand and optimizing dispatch strategies, unnecessary energy waste is reduced and overall energy efficiency is improved. Combining multi-objective optimization algorithms and real-time monitoring, the stability and reliability of power supply are ensured, and the risk of power outages and other failures is reduced. The optimization strategy specifically considers maximizing the utilization of renewable energy, which helps promote the application and development of green energy. The introduction of reinforcement learning algorithms to learn user behavior patterns provides personalized consulting services and enhances user acceptance and satisfaction with load adjustment measures. Dynamic monitoring of execution effects and risk assessment combined with Bayesian networks can quickly identify and respond to deviations from optimization goals to ensure the safe operation of the power grid. The system can flexibly adjust the load strategy according to actual conditions, and has good adaptability and scalability to meet changing needs.

[0057] Furthermore, by introducing geographic information system data and using complex network theory to construct a directed weighted graph, the initial structural model of the smart grid is formed, and the connection probability distribution of key nodes is initialized to simulate the uncertainty and randomness of the power grid, generating a time-varying power grid topology with time dimension attributes. Based on this topology, the connectivity and redundancy are analyzed to obtain the risk analysis results; then the key nodes are redefined using the PageRank algorithm and the Hyperlink Induced Topic Search algorithm (HITS), and the target nodes whose behavior patterns deviate from the normal are identified by combining anomaly detection methods such as isolated forests and local anomaly factors, and the historical data and current situation of these nodes are evaluated to predict potential problem points. According to the evaluation results, the Monte Carlo simulation combined with the propagation dynamics model is used to simulate the fault diffusion process under different failure scenarios, estimate the impact range and recovery time, formulate emergency response strategies, and integrate the deep reinforcement learning framework to generate solutions for specific types of faults. Finally, the entities, relationships and attributes are organized into a semantic information network, and the status after each update is recorded through blockchain technology, visualized as a map, and the topological structure information of the smart grid is completed.

[0058] This method significantly improves the risk prediction capability and emergency response efficiency of smart grids, and ensures the stable operation of the power system through precise modeling and dynamic simulation. It can not only monitor and predict potential risks in the power transmission path in real time, but also quickly identify abnormal behaviors and provide targeted solutions, enhancing the robustness and adaptability of the system. In addition, the use of blockchain technology to record and manage changes in the state of the power grid ensures the transparency and immutability of the data, providing reliable decision-making support for power dispatching and maintenance. Overall, this method has achieved a comprehensive improvement in the security, reliability and intelligent management level of smart grids.

[0059] These and other aspects of the present application will become more clearly understood in the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0061] Figure 1 A flow chart of a method for intelligent dynamic control of loads for a smart grid provided by the present application is shown;

[0062] Figure 2 A schematic diagram of the structure of a load intelligent dynamic control system for a smart grid provided by the present application is shown;

[0063] Figure 3 A schematic diagram of the structure of a computing device provided by the present application is shown. DETAILED DESCRIPTION

[0064] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0065] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0066] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0067] Figure 1 A flow chart of a method for intelligent dynamic load control in a smart grid is provided for an embodiment of the present application, such as Figure 1 As shown, the method includes:

[0068] 101. Use time series analysis and deep learning algorithms to collect real-time power parameters of key nodes in the smart grid to obtain collected data, and predict power demand fluctuations within a preset time based on the collected data to obtain prediction results, use graph theory algorithms to build a power grid topology model, identify key nodes and potential impact ranges of power transmission paths in the smart grid, and obtain topology information;

[0069] Among them, time series analysis is a statistical method used to process data points arranged in time order, aiming to identify trends, periodicity and other patterns in the data; in smart grids, it includes data such as historical electricity consumption records to predict future electricity demand;

[0070] Deep learning algorithms are a type of machine learning technology that automatically learns features and patterns from large amounts of data through multi-layer neural networks; in this step, these algorithms are used to perform real-time analysis of power parameters at key nodes to improve prediction accuracy;

[0071] Key nodes refer to nodes with important functions or locations in the smart grid, such as substations, power plants, and large load centers; they are key components of the power transmission path, and their status directly affects the overall performance of the grid;

[0072] Collected data refers to real-time power parameters obtained from key nodes, such as voltage, current, power, etc. These data are used for subsequent analysis and prediction;

[0073] Electricity demand fluctuations refer to the dynamic changes in electricity demand over time and in the environment, including changes in peak and trough periods; accurately predicting these fluctuations is critical to optimizing grid operation;

[0074] Graph theory is a branch of mathematics that studies the structure of graphs consisting of vertices (nodes) and edges (connections). In this step, graph theory is used to build a topological model of the power grid to help understand the power transmission path and its potential impact range.

[0075] The grid topology model is an abstract representation of the physical connection relationship of the grid, describing the connection between nodes; it is used to simulate the flow of power and identify which nodes play a vital role in the power transmission process;

[0076] The power transmission path is the physical path that electricity takes from the generating station to the user, involving multiple intermediate nodes and transmission lines; understanding these paths helps optimize power distribution and management;

[0077] The potential impact range refers to the area that may be affected when a node or path fails, which is very important for risk assessment and emergency response;

[0078] Topological information is a detailed description of the structural characteristics of the power grid, including the location of nodes, connection methods, and the importance of each node, providing a basis for subsequent analysis.

[0079] In actual operation, firstly, time series analysis and deep learning algorithms are used to collect real-time power parameters of key nodes in the smart grid to obtain collected data including voltage, current, power and other indicators. Based on these data, the system predicts the fluctuation of power demand within a preset time and generates prediction results.

[0080] Then, a graph theory algorithm is used in combination with geographic information system (GIS) data to construct a directed weighted graph of the power grid, form an initial structural model, and initialize the connection probability distribution of key nodes to simulate the uncertainty and randomness of the power grid.

[0081] Finally, a time-varying power grid topology with time dimension attributes is generated, and the connectivity and redundancy are analyzed, the key nodes in the power transmission path and their potential impact range are identified, and detailed topological structure information is obtained.

[0082] For example, suppose that in a city-level smart grid, multiple sensors are deployed to monitor power parameters at substations, distribution rooms, and other key nodes. Through time series analysis, it is found that the power demand increases significantly around 5 pm every day, which may be due to residents starting to use home appliances after getting home from get off work. Using deep learning algorithms, the specific magnitude and duration of this demand fluctuation can be predicted more accurately.

[0083] Subsequently, the urban power grid topology model constructed by graph theory algorithm showed that certain substations are key nodes, and any failure of them may cause large-scale power outages.

[0084] Therefore, measures can be taken in advance to ensure the safe operation of these nodes and plan alternative paths to reduce the potential impact range.

[0085] 102. Based on the prediction result and the topological structure information, a multi-objective optimization algorithm is used in combination with the optimization objectives pre-set in the smart grid to generate a load adjustment strategy for specific operation instructions for different power consumption areas or equipment, wherein the multi-objective optimization algorithm includes a genetic algorithm and a particle swarm optimization, and the optimization objectives include minimizing energy costs, maximizing renewable energy utilization, and ensuring power supply stability;

[0086] Among them, multi-objective optimization algorithm is a method that considers multiple mutually constrained objective functions for optimization at the same time; here, genetic algorithm and particle swarm optimization are used to find the best solution to minimize energy costs, maximize renewable energy utilization and ensure power supply stability;

[0087] Optimization goals refer to the specific goals that we want to achieve. In smart grids, they mainly include reducing energy costs, improving the utilization of renewable energy, and maintaining power supply stability. There may be conflicts between these goals, which need to be considered comprehensively.

[0088] Operation instructions are specific action guidelines for different power consumption areas or equipment, such as adjusting power generation, changing transmission line configuration, or controlling user-side loads;

[0089] The load adjustment strategy is an operation plan formulated based on the prediction results and topology information, aiming to optimize power distribution and ensure efficient and stable operation of the power grid.

[0090] In actual operation, based on the prediction results and topological information obtained in the previous step, the system uses multi-objective optimization algorithms (such as genetic algorithms and particle swarm optimization) and combines pre-set optimization goals to generate specific operation instructions for different power consumption areas or equipment to form load adjustment strategies. These strategies aim to minimize energy costs, maximize the utilization of renewable energy, and ensure power supply stability, thereby ensuring the optimal configuration of power grid resources.

[0091] For example, continuing with the previous embodiment, based on the predicted peak electricity demand at 5 pm, the system analyzes the key nodes and potential impact range in the power grid topology. In response to this situation, the system uses genetic algorithms and particle swarm optimization to develop a load adjustment strategy, which recommends transferring part of the load to off-peak hours before the peak hours, starting backup generators, and giving priority to renewable energy sources such as wind and solar energy. In addition, it is recommended to optimize the configuration of transmission lines to avoid overload. These strategies are sent to relevant equipment for execution through the automated control system, ensuring the efficient operation of the power grid and the rational use of resources.

[0092] 103. Convert the load adjustment strategy into guidance suggestions and push them to the end user, collect feedback information from the end user, apply the reinforcement learning algorithm to learn the long-term behavior patterns and preferences of the end user, generate a preference model including the end user's acceptance, feasibility evaluation results and personalized consulting solutions, and select and implement the corresponding load adjustment strategy based on the feedback information and the preference model to obtain the execution effect;

[0093] Among them, guidance suggestions are to convert load adjustment strategies into information in easy-to-understand language or chart form and push them to end users so that they can take corresponding actions.

[0094] End users refer to individuals or businesses that directly use electricity services, and their behavior patterns and preferences are crucial to the successful implementation of load adjustment strategies.

[0095] Reinforcement learning algorithm is a machine learning method that learns the best behavior strategy by interacting with the environment. Here, it is used to analyze the long-term behavior patterns and preferences of end users to generate personalized consulting solutions.

[0096] The preference model is a mathematical model based on user behavior patterns and feedback information. It is used to predict user acceptance and feasibility evaluation results, and provide a basis for personalized consulting services.

[0097] In actual operation, the above load adjustment strategies are converted into guidance suggestions and pushed to end users through mobile applications or other communication channels. The system collects user feedback information, applies reinforcement learning algorithms to learn users' long-term behavior patterns and preferences, and generates a preference model that includes user acceptance, feasibility evaluation results, and personalized consulting solutions. Based on this model, the load adjustment strategy that best suits the user's actual situation is selected and implemented to improve user satisfaction and the effectiveness of the strategy.

[0098] For example, continuing with the above embodiment, when the system predicts that the peak electricity demand will occur at 5 pm, it will convert the load adjustment strategy into guidance suggestions and notify the user through a smartphone application. For example, it is recommended that users use washing machines or air conditioners during off-peak hours. The system collects user feedback on these suggestions and analyzes user behavior patterns through reinforcement learning algorithms, and learns that some users tend to follow the suggestions, while others prefer to maintain their original habits. Based on this information, the system generates personalized consulting plans, providing incentives such as electricity discounts for users who are willing to cooperate, and more flexible options for users who are less willing to change their habits. This not only improves user participation, but also makes the load adjustment strategy closer to actual needs.

[0099] 104. Utilize the dynamic control system of the smart grid to monitor the execution effect and the operating status of the smart grid, and use a Bayesian network to detect the risk of the execution effect and the abnormality of the operating status, identify the deviation from the optimization target, and adjust the execution effect or switch to a backup load adjustment strategy.

[0100] Among them, the dynamic control system is an integrated software platform used to monitor the operation status of the smart grid in real time to ensure that all indicators are within the normal range.

[0101] A Bayesian network is a probabilistic graphical model that represents conditional dependencies between variables. In this step, it is used to detect risks in execution performance and anomalies in operating conditions.

[0102] The fault diffusion process refers to how a problem spreads in the power grid and affects other parts after a node or path fails. Through simulation, the impact range and recovery time of the fault can be estimated.

[0103] Emergency response strategies are rapid reaction measures taken when deviations from optimization targets are detected, aiming to minimize negative impacts and restore normal operations.

[0104] Blockchain technology is a distributed ledger technology that is used to record the status of smart grids after each update to ensure data transparency and immutability.

[0105] In actual operation, the dynamic control system of the smart grid monitors the execution effect of the load adjustment strategy and the operation status of the power grid in real time, and combines the Bayesian network to detect the risks of the execution effect and the abnormal operation status. Once a deviation from the optimization target is identified, the system will immediately adjust the execution effect or switch to the backup load adjustment strategy. In addition, the system will use Monte Carlo simulation combined with the propagation dynamics model to simulate the fault diffusion process, estimate the impact range and recovery time, and formulate an emergency response strategy. Finally, all updated smart grid states are recorded through blockchain technology to ensure data integrity and security.

[0106] For example, continuing with the above embodiment, suppose that during the implementation of a certain load adjustment strategy, the dynamic control system detects that a key node has an abnormally high load and there is a risk of overload. Bayesian network analysis shows that this situation may lead to a larger power outage. The system took prompt action, using Monte Carlo simulation combined with a propagation dynamics model to estimate the possibility and scope of fault spread, and decided to immediately switch to a backup load adjustment strategy to redistribute electricity, avoiding potential large-scale power outages. At the same time, all these adjustments and state changes are recorded on the blockchain, ensuring the transparency and immutability of the data, and providing a reliable basis for subsequent audits and improvements.

[0107] Through the implementation of steps 101 to 104, efficient and stable allocation of power resources in the smart grid is achieved, and energy utilization efficiency and service quality are improved. Unnecessary energy waste is reduced by accurately predicting power demand and optimizing dispatch strategies; the introduction of multi-objective optimization algorithms ensures the stability and reliability of power supply; personalized consulting services enhance users' acceptance and satisfaction with load adjustment measures; real-time monitoring and risk assessment mechanisms ensure the safe operation of the system and timely respond to emergencies; data records supported by blockchain technology improve the transparency and credibility of decision-making. Overall, this method significantly enhances the intelligent management level of smart grids and provides a solid scientific basis and technical support for power dispatch and maintenance.

[0108] In order to solve the problem of identifying key nodes of the power transmission path and evaluating the potential impact range in the smart grid, and further improve the safety, stability and efficiency of the power grid operation, in some embodiments, the graph theory algorithm is used to construct a power grid topology model in step 101, identify the key nodes and potential impact range of the power transmission path in the smart grid, and obtain topology information, including:

[0109] Based on the key nodes and connection relationships in the smart grid, geographic information system data is introduced, and a directed weighted graph is constructed using complex network theory to obtain the initial structural model of the smart grid. The connection probability distribution of the key nodes is initialized in the initial structural model to simulate the uncertainty and randomness of the grid and generate a time-varying grid topology with time dimension attributes. According to the time-varying grid topology, the connectivity and redundancy of the smart grid are analyzed to obtain risk analysis results. Based on the risk analysis results, the key nodes are redefined using a page ranking algorithm and a hyperlink-induced topic search algorithm, and the target nodes whose behavior patterns deviate from the normal behavior patterns are identified according to the anomaly detection method. The historical data of the target nodes are compared with the problem points predicted by the current situation to obtain the The evaluation results of the key nodes and the target nodes are obtained, wherein the anomaly detection method includes an isolation forest and a local anomaly factor; based on the evaluation results, a Monte Carlo simulation is used in combination with a propagation dynamics model to simulate the fault diffusion process of different nodes or paths of the smart grid under failure scenarios, the impact range and recovery time of the fault are estimated, an emergency response strategy is formulated, and the emergency response strategy is used to integrate a deep reinforcement learning framework to generate solutions to specific types of faults; based on the solution, the entities, relationships and attributes in the smart grid are organized into a semantic information network, and the information network is visualized in a map form, and blockchain technology is deployed according to the map form to record the status of the smart grid after each update, and obtain the topological structure information of the smart grid.

[0110] In this embodiment, graph theory algorithm is a branch of mathematics that studies the graph structure composed of vertices (nodes) and edges (connections); in smart grids, it is used to build a topological structure model of the grid to help understand the power transmission path and its potential impact range;

[0111] Geographic Information System (GIS) data refers to data sets containing geographic location information, usually including maps, terrain, buildings and other geographic features. When building a power grid model, GIS data provides spatial distribution information of key nodes and connection relationships.

[0112] Complex network theory is a method for studying network structure and dynamic characteristics, suitable for describing and analyzing systems with complex connection patterns; it helps to reveal the interactions between nodes in the power grid and the behavior of the overall network;

[0113] A directed weighted graph is a graphical representation in which each node has a direction and each edge has a weight value representing different factors such as distance, cost or time. For smart grids, these weights can reflect the cost, loss or reliability of power transmission.

[0114] The initial structural model is a preliminary abstract representation of the physical connection relationship of the power grid, describing the basic connection mode between nodes; it is the basis for further analysis and optimization;

[0115] Connection probability distribution refers to the possible distribution of connections between nodes, simulating the uncertainty and randomness in power grid operation; this helps to more realistically reflect the actual operating environment;

[0116] Time-varying grid topology is a dynamic grid structure formed after considering the time dimension, reflecting the grid status that changes with time and conditions; it provides an important spatiotemporal context for prediction and response;

[0117] Connectivity and redundancy refer to the ability to connect any two points in the grid directly or indirectly, and the extent to which other paths can continue to supply power when some parts fail; they are crucial to assessing the stability and reliability of the grid.

[0118] The PageRank algorithm was originally used for search engine ranking, which determines the order of search results by calculating the importance of web pages; here, it is used to redefine the key nodes in the power grid and identify those nodes that play an important role in the network;

[0119] Hyperlink Induced Topic Search (HITS) is also a web page ranking algorithm that, in addition to evaluating the importance of the web page itself, also considers the quality of the links pointing to the web page; when applied to the power grid, it can assist in identifying highly influential nodes;

[0120] Anomaly detection methods are a class of techniques used to identify abnormal behaviors in data sets, such as Isolation Forest and Local Outlier Factor (LOF), which can find target nodes that deviate from normal patterns and are very important for fault prevention and maintenance;

[0121] Monte Carlo simulation is a numerical calculation method based on a large number of random samplings, which is often used to estimate the uncertainty of complex systems. Combined with the propagation dynamics model, it can simulate the fault diffusion process and help estimate the impact range and recovery time.

[0122] Emergency response strategies are quick reaction measures taken when deviations from optimization targets are detected, aiming to minimize negative impacts and restore normal operations;

[0123] Blockchain technology is a distributed ledger technology used to record the status of smart grids after each update, ensuring data transparency and immutability;

[0124] The semantic information network organizes the entities, relationships and attributes in the power grid into a structure that is easy to understand and use, facilitating management and decision support.

[0125] In an embodiment of the present application, firstly, a directed weighted graph is constructed as the initial structural model of the smart grid by using a graph theory algorithm combined with GIS data and complex network theory, and the connection probability distribution of key nodes is initialized to simulate the uncertainty and randomness of the power grid and generate a time-varying power grid topology with time dimension attributes.

[0126] Next, the connectivity and redundancy of the power grid are analyzed based on this topology to obtain the risk analysis results.

[0127] Then, we use PageRank and HITS algorithms to redefine key nodes, identify target nodes whose behavior patterns deviate from normal, and compare their historical and current data to predict problem points. Based on the evaluation results, we use Monte Carlo simulation combined with the propagation dynamics model to simulate the fault diffusion process, estimate the impact range and recovery time, formulate emergency response strategies, and integrate the deep reinforcement learning framework to generate solutions.

[0128] Finally, all information is organized into a semantic information network and deployed through a visual map supported by blockchain technology to record the status of the power grid after each update and form complete topological structure information.

[0129] Here is a specific example:

[0130] In a city-level smart grid project, in order to better understand and manage the grid structure, engineers first collected the locations and connections of all substations, power plants, and major transmission lines in the city, and introduced detailed GIS data. Using complex network theory, they constructed a directed weighted graph as the initial structural model of the grid, where each node represents a substation or power station, and the weight of the edge reflects the cost and loss of power transmission. In order to simulate the uncertainty and randomness of the grid, they initialized the connection probability distribution of key nodes in the model and created a time-varying grid topology with time dimension attributes.

[0131] Next, engineers analyzed the connectivity and redundancy of the power grid and found several key nodes that are prone to cascading failures. Using PageRank and HITS algorithms, they redefined the importance of these key nodes and identified some target nodes with abnormal behavior patterns through anomaly detection methods such as isolation forest and local anomaly factor. By comparing the historical data of these nodes with the current situation, they predicted possible problem points.

[0132] To prepare for possible failure scenarios, engineers used Monte Carlo simulation combined with propagation dynamics models to estimate the impact range and recovery time when different nodes or paths fail. Based on this information, they developed detailed emergency response strategies and used deep reinforcement learning frameworks to generate solutions for specific types of failures. For example, if a substation fails, the system will automatically switch to backup power and notify relevant technicians for emergency repairs.

[0133] Finally, engineers organized all the information into a semantic information network and deployed it through a visual map supported by blockchain technology to ensure that the status of the power grid is accurately recorded every time it is updated. This not only improves the transparency of power grid management, but also provides a reliable data foundation for future planning and decision-making. In this way, the city has achieved a comprehensive grasp of the smart grid structure and significantly improved the security and stability of the power grid.

[0134] In order to solve the problem of accurate evaluation of smart grid connectivity and redundancy and further improve the maximum carrying capacity and robustness of the grid, in some embodiments, the analysis of the connectivity and redundancy of the smart grid further includes:

[0135] The maximum flow and minimum cut theorem is used to determine the global and local maximum flow and minimum cut set of the smart grid, evaluate the maximum carrying capacity and the most vulnerable point of the smart grid, and obtain the global and local flow distribution and key node information of the smart grid; using the flow distribution and key node information, the shortest path between the key nodes and the average path length between all node pairs are calculated to generate a power grid performance evaluation report; based on the power grid performance evaluation report, the hypergraph theory is used to expand the connectivity analysis, and the subsystems of the smart grid are identified through the community discovery algorithm.

[0136] In this embodiment, the maximum flow minimum cut theorem is a basic theorem in graph theory, which is used to determine the maximum flow from a source node to a sink node in a network, and the minimum cut set that restricts the flow; in smart grids, it helps evaluate the maximum carrying capacity of power transmission and identify the most vulnerable points;

[0137] Maximum flow refers to the maximum amount of electricity that can be transmitted from the source (such as a power station) to the end point (such as the user end) in the network; this helps to understand the performance of the power grid when it is running at full load;

[0138] The minimum cut set is the set of edges that, when cut, interrupt any path from source to destination in the network. For smart grids, the minimum cut set identifies the critical paths or nodes that may cause system failure.

[0139] Global and local maximum carrying capacity refer to the maximum power transmission capacity of the entire grid and a specific area respectively; understanding this is crucial for planning and optimization;

[0140] The most vulnerable points refer to those nodes or paths that may cause large-scale power outages or other serious consequences if they fail. Identifying the most vulnerable points can help prevent potential risks.

[0141] Flow distribution and key node information include the power flow between nodes in the power grid and their importance; this information is critical for evaluating power grid performance and developing optimization strategies;

[0142] The shortest path is the path with the lowest cost or distance among all possible paths from one node to another. In power transmission, the shortest path means the most efficient way to deliver power.

[0143] Average path length is an indicator of grid connectivity and efficiency, indicating the average connection distance between any two nodes; shorter average path length generally means more efficient power transmission;

[0144] Hypergraph theory is a method that extends traditional graph theory by allowing edges to connect multiple vertices instead of just two; it can be used to better describe many-to-many relationships in complex systems and to enhance connectivity analysis in smart grids;

[0145] Community discovery algorithms are a class of techniques used to detect closely connected subgroups in a network; when applied to power grids, they can identify functionally independent subsystems, which is important for management and optimization.

[0146] In the embodiment of the present application, firstly, by applying the maximum flow minimum cut theorem, the global and local maximum flows and minimum cut sets of the smart grid are evaluated to determine its maximum carrying capacity and most vulnerable point, and obtain detailed flow distribution and key node information.

[0147] This information is then used to calculate the shortest paths between key nodes and the average path length between all node pairs, generating a comprehensive grid performance assessment report.

[0148] Finally, based on this report, the connectivity analysis is extended using hypergraph theory, and the subsystems in the smart grid are identified through community discovery algorithms to further refine and optimize the management strategy.

[0149] Here is a specific example:

[0150] According to the above embodiment, in a large-scale urban smart grid project, in order to ensure the security and stability of power supply, engineers first apply the maximum flow minimum cut theorem to analyze the global and local characteristics of the power grid. By simulating the maximum flow of different substations and transmission lines, they determine the maximum carrying capacity of the entire power grid and find some of the most vulnerable points - that is, key nodes or paths that may cause large-scale power outages if a failure occurs.

[0151] Next, engineers used the traffic distribution and key node information to calculate the shortest path between key nodes and the average path length between all node pairs. This not only revealed the most efficient path for power transmission, but also provided important insights into the overall connectivity and efficiency of the power grid. Based on these data, they generated a detailed power grid performance evaluation report, providing a solid foundation for subsequent optimization.

[0152] To gain a deeper understanding of the grid structure, engineers further adopted hypergraph theory to expand connectivity analysis. This approach enables them to capture complex relationships that are difficult to express in traditional graph theory, such as situations where multiple substations supply power together. In addition, by implementing a community discovery algorithm, they successfully identified several functionally independent subsystems, such as the power supply networks of commercial, residential, and industrial areas. This discovery helps to achieve more refined management and scheduling strategies, such as developing personalized load adjustment plans for different subsystems, or giving priority to protecting critical areas in emergencies.

[0153] Finally, by combining the maximum flow and minimum cut theorem, shortest path calculation, hypergraph theory, and community discovery algorithms, engineers not only improved their understanding of the city's smart grid, but also developed practical optimization measures that significantly enhanced the safety and reliability of the grid. The application of this method provides valuable experience and technical support for the development of future smart grids.

[0154] In order to solve the importance of key nodes and abnormal behavior identification problems in smart grids and further improve the risk management and response efficiency of power grids, in some embodiments, based on the risk analysis results, the key nodes are redefined using a page ranking algorithm and a hyperlink induced topic search algorithm, and target nodes whose behavior patterns deviate from normal behavior patterns are identified according to an anomaly detection method, and the historical data of the target nodes are compared with the problem points predicted by the current situation to obtain evaluation results of the key nodes and the target nodes, including:

[0155] Based on the risk analysis results, the importance of each of the key nodes in the time-varying power grid topology is evaluated using a page ranking algorithm and a hyperlink-induced topic search algorithm to obtain a key node score; based on the key node score, the target nodes whose behavior patterns deviate from normal behavior patterns are identified in combination with anomaly detection methods in machine learning to obtain a target node list; the historical data of the target nodes are compared with the current situation, the target node list is analyzed, and evaluation results of the key nodes and the target nodes are generated.

[0156] In this embodiment, the risk analysis results are based on the statistics and prediction of the grid operation data to evaluate the possible risk factors and their impact in the grid; these results provide a basis for subsequent optimization and improvement;

[0157] The PageRank algorithm was originally used for search engine sorting, which determines the order of search results by calculating the importance of web pages; here, it is used to redefine key nodes in the power grid and identify those nodes that play an important role in the network, that is, nodes with high power transmission importance;

[0158] Hyperlink Induced Topic Search (HITS) is also a web page ranking algorithm that, in addition to evaluating the importance of the web page itself, also considers the quality of the links pointing to the web page; when applied to power grids, it can help identify high-influence nodes, especially those key points that are not only important in themselves but also connect to other important nodes;

[0159] The key node score refers to the importance and influence score of each key node calculated by the page ranking algorithm and the hyperlink-induced topic search algorithm; this score reflects the relative importance of the node in the network;

[0160] Anomaly detection methods are a class of techniques used to identify abnormal behaviors in data sets, such as Isolation Forest and Local Outlier Factor (LOF). They can find target nodes that deviate from normal patterns, which is very important for fault prevention and maintenance.

[0161] The target node list is a list of nodes that need special attention, which may show abnormal behavior or potential risks, generated based on key node scores and anomaly detection methods;

[0162] Comparison of historical data with current conditions refers to comparing the historical operating data of the target node with its current status to identify any significant changes or trends, which helps predict possible future problems.

[0163] The evaluation results are conclusions drawn after comprehensive analysis of key nodes and target nodes, including their importance, abnormal behaviors, and predictions of their impact on future grid performance.

[0164] In the embodiment of the present application, firstly, based on the previous risk analysis results, the importance of each key node in the time-varying power grid topology is evaluated using the page ranking algorithm (PageRank) and the hyperlink induced topic search algorithm (HITS) to generate a key node score.

[0165] Then, combined with anomaly detection methods in machine learning, such as isolation forest and local anomaly factor, target nodes whose behavior patterns deviate from normal behavior patterns are identified to form a target node list.

[0166] Finally, compare the historical data of these target nodes with the current situation, analyze their changing trends, predict possible problem points, and generate detailed evaluation results about key nodes and target nodes.

[0167] Here is a specific example:

[0168] According to the above embodiment, in a smart grid project, in order to ensure the safe and stable operation of the power system, engineers first conducted a comprehensive risk analysis to evaluate the potential risk factors of the power grid under different conditions, such as extreme weather events, equipment aging, etc. Based on these risk analysis results, engineers applied the PageRank algorithm and the Hyperlink Induced Topic Search algorithm (HITS) to redefine the key nodes in the power grid, evaluate the relative importance of each node in the power transmission network, and obtain the key node score.

[0169] Next, engineers combined the anomaly detection methods in machine learning - isolation forest and local anomaly factor - to identify some target nodes whose behavior patterns deviated from the normal ones. For example, some substations or transmission lines showed load fluctuations or temperature changes that were different from the past, which might be caused by equipment failure or external environment. In this way, they generated a list of target nodes, listing all the nodes that need special attention.

[0170] To gain a deeper understanding of the conditions at these target nodes, engineers compared their historical data with their current operating conditions. For example, the average load at a particular substation has remained stable over the past year, but has experienced irregular peak loads in recent weeks. Through this comparative analysis, engineers can predict possible problem points, such as potential equipment overload or a surge in power demand in a local area.

[0171] Finally, the engineers generated a detailed assessment report that included the importance of key nodes and target nodes, abnormal behaviors, and their predicted impact on future grid performance. This report not only helped them better understand the current status of the grid, but also provided a scientific basis for formulating preventive measures and optimization strategies. For example, for those target nodes that showed abnormal behaviors, engineers arranged inspections and maintenance work in advance to ensure the safe and stable operation of the grid. The application of this method significantly improved the reliability and response efficiency of the grid, laying a solid foundation for future intelligent management and optimization.

[0172] In order to solve the problem of accurately evaluating the connectivity and overall performance between key nodes in the smart grid and further improve the maximum carrying capacity and robustness of the power grid, in some embodiments, the flow distribution and key node information are used to calculate the shortest path between the key nodes and the average path length between all node pairs to generate a power grid performance evaluation report, including:

[0173] Based on the traffic distribution and key node information of the smart grid, the connectivity between the key nodes is analyzed, and the maximum carrying capacity and the most vulnerable point of the smart grid are determined by the maximum flow minimum cut theorem to obtain a preliminary evaluation result; based on the preliminary evaluation result, the traffic distribution and key node information are used as input, and a graph theory algorithm is used to calculate the shortest path between the key nodes; based on the shortest path, the average path length between all node pairs is calculated to obtain the average connection distance between any two nodes in the smart grid; the shortest path data and the average connection distance are combined to generate a power grid performance evaluation report.

[0174] In this embodiment, the flow distribution and key node information include the power flow between various nodes in the power grid and their importance; this information is critical for evaluating power grid performance and formulating optimization strategies, and is used to analyze connectivity and determine maximum carrying capacity;

[0175] The shortest path is the path with the lowest cost or distance among all possible paths from one node to another. In power transmission, the shortest path means the most efficient way to deliver power, reducing losses and improving efficiency.

[0176] Average path length is an indicator of grid connectivity and efficiency, indicating the average connection distance between any two nodes; shorter average path length generally means more efficient power transmission;

[0177] The maximum flow minimum cut theorem is a fundamental theorem in graph theory that is used to determine the maximum flow from a source node to a sink node in a network, as well as the minimum cut set that constrains that flow; it helps assess the maximum carrying capacity of power transmission and identify the most vulnerable points;

[0178] The preliminary evaluation results refer to the initial conclusions on the global and local characteristics of the smart grid (such as the maximum carrying capacity and the most vulnerable point) obtained by applying the maximum flow minimum cut theorem; this is the basis for further calculations and analysis;

[0179] Graph theory is a branch of mathematics that studies the structure of graphs consisting of vertices (nodes) and edges (connections). In smart grids, it is used to build a topological model of the grid to help understand the power transmission path and its potential impact range.

[0180] The Grid Performance Assessment Report is a comprehensive document that summarizes a comprehensive analysis of grid performance, including indicators such as connectivity, redundancy, shortest path, average path length, etc., to provide decision support for management and optimization.

[0181] In the embodiment of the present application, firstly, based on the flow distribution and key node information of the smart grid, the connectivity between the key nodes is analyzed, and the maximum carrying capacity and the most vulnerable point of the power grid are determined using the maximum flow minimum cut theorem to obtain a preliminary evaluation result.

[0182] Then, based on these preliminary evaluation results, graph theory algorithms are used to calculate the shortest paths between key nodes.

[0183] Next, based on the shortest path data, the average path length between all node pairs is calculated to obtain the average connection distance between any two nodes in the smart grid.

[0184] Finally, the shortest path data and average connection distance are combined to generate a detailed power grid performance evaluation report, providing a scientific basis for the management, maintenance and optimization of the power grid.

[0185] Here is a specific example:

[0186] According to the above embodiment, the flow distribution data of substations, power plants and main transmission lines in the whole city were first collected, and the location and function of key nodes were marked. Using this information, engineers analyzed the connectivity between key nodes and applied the maximum flow minimum cut theorem to determine the maximum carrying capacity and the most vulnerable point of the entire power grid.

[0187] For example, through the calculation of the maximum flow and minimum cut theorem, engineers found that the transmission line in a certain area is the most vulnerable link in the entire power grid, and any failure may cause a large-scale power outage. This discovery became part of the preliminary assessment results and provided direction for subsequent optimization.

[0188] Next, based on these preliminary assessments, engineers used graph theory algorithms to calculate the shortest paths between key nodes. This step not only revealed the most efficient paths for power transmission, but also provided specific guidance for optimizing power distribution. For example, the shortest paths between certain nodes were confirmed as optimal transmission routes, while other paths showed high transmission costs or losses.

[0189] The engineers then calculated the average path length between all node pairs based on the shortest path data, and derived the average connection distance between any two nodes in the smart grid. A shorter average path length indicates that the overall connectivity and efficiency of the grid are higher, while a longer distance indicates that the power transmission network in some areas needs to be improved.

[0190] Finally, combining the shortest path data and the average connection distance, engineers generated a detailed grid performance assessment report. This report not only summarizes the key performance indicators of the grid, but also puts forward specific improvement suggestions. For example, for the most vulnerable points, engineers recommend adding backup transmission lines or strengthening the maintenance of existing lines; for areas with longer average path lengths, they propose to re-plan the power transmission path or add new substations.

[0191] In order to solve the problem of real-time monitoring and abnormality detection of the execution effect and operation status of the smart grid and further improve the safety and response efficiency of the power grid, in some embodiments, the step 104 uses the dynamic control system of the smart grid to monitor the execution effect and the operation status of the smart grid, and combines the Bayesian network to detect the risk of the execution effect and the abnormality of the operation status, identify the deviation from the optimization target, and adjust the execution effect or switch to the standby load adjustment strategy, including:

[0192] The execution effect and the operating status are monitored by the dynamic control system of the smart grid to obtain monitoring data, wherein the monitoring data includes the operating parameters of the smart grid, and the operating parameters of the smart grid include voltage, current and power; a Bayesian network model is constructed based on the monitoring data to detect the risk of the execution effect and the abnormality of the operating status, and a risk assessment result and an abnormality detection report are generated, wherein the operating parameters of the smart grid are used as input variables of the Bayesian network model, and the risk and the abnormality are used as output variables of the Bayesian network model; according to the risk assessment result and the abnormality detection report, in combination with the optimization target, the deviation is identified to obtain a deviation judgment conclusion; based on the deviation judgment conclusion, the execution effect is adjusted to obtain an adjusted execution effect, and after the adjustment fails, switching to a standby load adjustment strategy.

[0193] In this embodiment, the dynamic control system is an integrated software platform for real-time monitoring of the operation of the smart grid to ensure that various indicators are within normal ranges. It collects and processes data from various nodes of the grid, such as operating parameters such as voltage, current and power, to evaluate the health of the grid and guide optimization measures.

[0194] Monitoring data refers to information collected by the dynamic control system about the operation of the smart grid, including but not limited to key parameters such as voltage level, current intensity and power consumption. This data is the basis for analyzing grid performance, detecting potential problems and supporting decision making.

[0195] The Bayesian network model is a probabilistic graphical model used to represent the conditional dependencies between variables. In this step, it is used to detect the risks of execution effects and abnormalities in operating conditions, and predict the probability distribution of output variables (such as risks and abnormalities) through input variables (such as power grid operating parameters).

[0196] The risk assessment results are conclusions output from the Bayesian network model about the risks that the current operation of the power grid may face. This includes factors such as potential failure points in power transmission, security and stability threats, and provides a basis for taking preventive measures.

[0197] The anomaly detection report is a document generated based on the Bayesian network model, which records all abnormal situations detected in the operation of the power grid and their detailed descriptions. This report helps to quickly locate the problem and provide guidance for subsequent repair work.

[0198] Optimization goals refer to specific goals that we want to achieve. In smart grids, they mainly include reducing energy costs, increasing the utilization of renewable energy, and maintaining power supply stability. These goals may conflict with each other and need to be considered comprehensively.

[0199] The deviation judgment conclusion is the final judgment on whether the current execution effect deviates from the optimization goal after combining the risk assessment results and the anomaly detection report. This conclusion directly determines the next action to be taken.

[0200] Backup load adjustment strategies are alternative solutions used when the main load adjustment strategy cannot effectively respond to grid changes or problems. These strategies are usually pre-set so that they can be quickly switched when necessary to ensure stable operation of the grid.

[0201] In the embodiment of the present application, firstly, the dynamic control system of the smart grid is used to monitor the execution effect and the operating status of the power grid in real time, and the operating parameters including voltage, current and power are collected as monitoring data.

[0202] Then, based on these monitoring data, a Bayesian network model is built to detect the risks of execution effects and abnormalities in operation status, and generate risk assessment results and abnormality detection reports. Based on these assessment results and reports, combined with the pre-set optimization goals, any deviation from expectations is identified and a deviation judgment conclusion is formed.

[0203] Finally, based on the deviation judgment conclusion, necessary adjustments are made to the current execution effect; if the adjustment fails, the pre-prepared backup load adjustment strategy is immediately switched to ensure the safe and stable operation of the power grid.

[0204] Here is a specific example:

[0205] Continuing with the above example, an advanced dynamic control system is first deployed, which can collect real-time operating parameters of substations, transmission lines and other key nodes across the city, such as voltage, current and power. These data not only reflect the instantaneous status of the power grid, but also provide valuable information for subsequent analysis.

[0206] Based on the collected monitoring data, engineers built a Bayesian network model to assess the risks of power grid operation and detect abnormal situations. For example, the model can identify that the voltage fluctuations of certain substations are beyond the normal range, or that the current intensity of a specific transmission line is close to its carrying limit. This information is summarized into risk assessment results and anomaly detection reports to help engineers understand possible problems in the power grid in a timely manner.

[0207] Based on the risk assessment results and anomaly detection reports, combined with the pre-set optimization goals - minimizing energy costs, maximizing renewable energy utilization, and ensuring power supply stability - engineers identified deviations. Suppose that a sudden surge in power demand in a certain area causes excessive loads on some transmission lines, which obviously deviates from the optimization goal. At this time, engineers adjusted the existing load adjustment strategy based on the deviation judgment conclusion, such as reallocating power resources or starting backup generators to alleviate the high load pressure.

[0208] However, if the adjustment fails to effectively solve the problem, such as an unforeseen failure of a key transmission line, engineers will immediately switch to alternative load adjustment strategies. These alternative strategies may be pre-designed, including measures such as activating backup transmission lines and adjusting power supply to other areas to ensure that the basic functions and service quality of the power grid can be maintained even in the worst case.

[0209] In order to solve the optimal load allocation problem for different power consumption areas or devices in the smart grid and further improve the efficiency of power resource utilization and power supply stability, in some embodiments, the multi-objective optimization algorithm used in step 102 is combined with the optimization goals pre-set in the smart grid to generate a load adjustment strategy for specific operation instructions for different power consumption areas or devices, including:

[0210] A multi-objective optimization algorithm is selected based on the characteristics of the smart grid and a pre-set optimization goal, and the parameters of the target optimization algorithm are initialized to obtain the optimization algorithm configuration; input information is determined based on the prediction results, the topology structure information and the operating status of the smart grid, and the input information is key data generated by the multi-objective optimization algorithm for the current status and future trends of the smart grid; using the optimization goal, a preliminary load adjustment strategy for different power consumption areas or equipment operation instructions is generated through the optimization algorithm configuration and the input information; the preliminary load adjustment strategy is optimized based on a simulated annealing algorithm and an iterative optimization algorithm to generate the load adjustment strategy.

[0211] In this embodiment, the multi-objective optimization algorithm is a method for optimizing by simultaneously considering multiple mutually constrained objective functions; in smart grids, commonly used techniques include genetic algorithms (GAs), particle swarm optimization (PSOs), etc., which aim to find a solution that can balance different optimization objectives;

[0212] Optimization goals refer to the specific goals that we want to achieve. In smart grids, they mainly include minimizing energy costs, maximizing the utilization of renewable energy, and maintaining power supply stability. There may be conflicts between these goals, which need to be considered comprehensively.

[0213] Optimization algorithm configuration refers to the process of setting the initial parameters for the selected multi-objective optimization algorithm; this includes selecting the appropriate algorithm type, defining the variable range, setting convergence conditions, etc. to ensure that the algorithm can run effectively and produce meaningful results;

[0214] The forecast results are estimates of future electricity demand fluctuations based on time series analysis and deep learning algorithms; these forecasts help determine the time frame and expected effect of load adjustment strategies;

[0215] Topological information is an abstract representation of the physical connection relationship of the power grid, describing how the nodes are connected; it is used to simulate the flow of power and identify which nodes play a vital role in the power transmission process;

[0216] Input information refers to the key data required by the multi-objective optimization algorithm, which reflects the current status and future trends of the smart grid, such as prediction results, topology information, and current operating status;

[0217] The initial load adjustment strategy is the first round of operation instructions generated by the multi-objective optimization algorithm based on the input information; it is a preliminary plan for how to distribute power to different power consumption areas or equipment;

[0218] The simulated annealing algorithm is a global optimization technique that simulates the atomic arrangement changes during the cooling process of metals and gradually approaches the optimal solution. In smart grids, it can be used to improve load adjustment strategies and avoid falling into local optimality.

[0219] Iterative optimization algorithms are a class of optimization methods that improve the quality of a solution by repeatedly iterating; in each iteration, the algorithm tries to find a better solution until a certain stopping condition is met;

[0220] The load adjustment strategy is the specific operating instructions ultimately formed for different power consumption areas or equipment, aiming to optimize power distribution and ensure efficient and stable operation of the power grid.

[0221] In the embodiment of the present application, firstly, based on the characteristics of the smart grid and the pre-set optimization objectives, an appropriate multi-objective optimization algorithm is selected, and the parameters of the algorithm are initialized to obtain the optimization algorithm configuration.

[0222] Then, based on the prediction results, topological structure information and the operating status of the smart grid, the key data as input information of the multi-objective optimization algorithm are determined.

[0223] Then, using these input information and optimization objectives, preliminary load adjustment strategies for different power consumption areas or equipment are generated through optimization algorithm configuration.

[0224] Finally, the preliminary load adjustment strategy is optimized based on the simulated annealing algorithm and iterative optimization algorithm to generate the final load adjustment strategy to ensure that it can effectively achieve the predetermined optimization goals.

[0225] Here is a specific example:

[0226] According to the above embodiment, the characteristics of the entire power grid are first evaluated, including the distribution of power stations, the layout of transmission lines and the user demand pattern. Combined with the pre-set optimization goals - namely, minimizing energy costs, maximizing the utilization of renewable energy and ensuring power supply stability - particle swarm optimization (PSO) is selected as the multi-objective optimization algorithm, and it is initialized and configured, and relevant parameters such as population size, number of iterations and fitness function are set.

[0227] Next, engineers integrated data from multiple sources, including historical electricity consumption records, weather forecasts, holiday schedules, etc., to predict future electricity demand fluctuations. At the same time, they used the previously constructed grid topology model to describe in detail the connection relationship and importance of each key node. All of this information together constitutes the input information of the multi-objective optimization algorithm, reflecting the current status and future trends of the smart grid.

[0228] Based on these input information and optimization goals, engineers used the PSO algorithm to generate preliminary load adjustment strategies for different power consumption areas or equipment. For example, during certain peak hours, it is recommended to prioritize renewable energy such as wind and solar energy to reduce the use of fossil fuels; during off-peak hours, the power generation can be appropriately reduced to save energy costs.

[0229] To further optimize these preliminary strategies, engineers introduced simulated annealing and iterative optimization algorithms. The simulated annealing algorithm helps avoid possible local optimal solutions and ensures that the global optimal solution is found; while the iterative optimization algorithm continuously improves the quality of the solution through multiple iterations until the best effect is achieved. After this series of optimization processes, a detailed load adjustment strategy was finally formed, which not only takes into account energy costs and renewable energy utilization, but also ensures the stability of power supply.

[0230] This application takes into account that the complexity and dynamics of smart grids make it difficult for traditional shortest path and average path length calculation methods to accurately reflect actual operating conditions. Specifically, traditional methods often ignore the impact of time factors on transmission costs and losses, as well as the changes in node importance over time. In addition, for large-scale power grid networks, how to efficiently calculate the average path length between all node pairs and consider path reliability and diversity is also a challenge, so a new optional solution is proposed, which includes:

[0231] The using the traffic distribution and key node information to calculate the shortest path between the key nodes and the average path length between all node pairs includes:

[0232] Using the traffic distribution and key node information as input, a graph theory algorithm is used to calculate the shortest path between the key nodes, where the shortest path d ij (t) is calculated by the following formula:

[0233]

[0234] Among them, d ij (t) represents the shortest path length from key node i to key node j, represents the set of all paths from i to j, c(u,v) represents the transmission cost connecting nodes u and v, l(u,v) represents the transmission loss, t uv represents the transmission time from node u to node v, λ is the adjustment parameter, which reflects the impact of transmission time on weight, γ is the decay parameter, which reflects the change of node importance over time, t represents the current time point, u represents a starting node or intermediate node on the path, v represents the next node directly connected from node u, and P represents a specific path from a starting node i to a target node j;

[0235] The following is a detailed explanation of each parameter:

[0236] d ij (t) represents the shortest path length from key node i to key node j, which changes with time. This parameter is calculated by graph theory algorithms (such as Dijkstra or Floyd-Warshall), taking into account factors such as transmission cost, loss, transmission time and changes in node importance over time, and reflects the optimal connection cost between two nodes at a specific time point.

[0237] P ij Represents the set of all paths from i to j; these paths consist of a series of adjacent nodes, and each path represents a possible connection method; this set can be obtained by traversing and analyzing the power grid topology to ensure that all potential path choices are covered, providing basic data for subsequent calculations;

[0238] c(u,v) represents the transmission cost of connecting nodes u and v, which usually reflects the amount of resources required in the power transmission process, such as the construction cost and maintenance cost of transmission lines. This parameter is based on the actual investment in power grid infrastructure and the statistics of operating costs, and is one of the key indicators for evaluating the economic efficiency of power grids.

[0239] l(u,v) represents the transmission loss between nodes u and v, which mainly includes the energy loss during power transmission, such as the power consumption caused by resistance heating. This value is obtained through physical model simulation or historical data analysis, which is crucial for optimizing power transmission efficiency.

[0240] t uv It represents the transmission time from node u to node v, reflecting the time interval required for the power signal to propagate between two points. This parameter is determined according to the physical distance of the actual power grid, the characteristics of the transmission medium, and the delay of the communication system, and affects the real-time response capability and system stability.

[0241] λ is an adjustment parameter that reflects the impact of transmission time on weight. It is used to quantify the relationship between transmission time and transmission cost and loss. It is usually set through experimental calibration or expert experience to ensure that the formula can accurately reflect the actual situation.

[0242] γ is a decay parameter that reflects the change of node importance over time. This parameter takes into account the importance of the node's position in the network and its dynamic changes over time. It is usually obtained through network analysis tools or machine learning model predictions, which helps to simulate the behavior of the power grid more accurately.

[0243] t represents the current time point and is a time variable used to track the development of the power grid status over time; it can be an absolute time (such as seconds, minutes) or a relative time (such as the time since a certain starting event), depending on the needs of the specific application scenario;

[0244] u represents a starting node or an intermediate node on the path, and participates in the calculation as part of the path in graph theory; each node is selected based on the topological structure of the power grid, and its position in the network is determined by a traversal algorithm;

[0245] v represents the next node directly connected from node u, which is also calculated as part of the path; these two nodes together constitute an edge in the power grid, which is used to describe the specific path of power transmission;

[0246] P represents a specific path from a starting node i to a target node j, consisting of a series of adjacent nodes; the path is searched and selected through a graph theory algorithm and is used to calculate the shortest path length from i to j.

[0247] The reason for the overall design of the formula is to solve the complexity and dynamic challenges faced by load regulation in smart grids. Traditional shortest path calculation methods often ignore the impact of time factors on transmission costs and losses, as well as the change of node importance over time. Therefore, this formula introduces a time decay factor α(t) = e -γ·t and edge weight adjustment coefficient This makes the calculation results closer to the actual operating conditions. In addition, the formula not only takes into account the transmission cost and loss, but also adds the important factor of transmission time, thus providing a more comprehensive and accurate evaluation framework. This helps the smart grid management system make more scientific and reasonable decisions, ensuring the safe and stable operation of the power grid and the efficient use of resources.

[0248] Here is a specific example:

[0249] Input data:

[0250] Transmission cost c(u,v): c(u 1 ,u 2 )=5;c(u2 ,u 3 )=7;c(u 3 ,u 4 )=6;

[0251] Transmission loss l(u,v): l(u 1 ,u 2 )=2;l(u 2 ,u 3 )=3;l(u 3 ,u 4 )=4;

[0252] Transmission time t uv :

[0253] Tuning parameter λ: 0.1;

[0254] Attenuation parameter γ: 0.05;

[0255] Current time point t: 1.0;

[0256] Path set P ij :The path from i to j is P ij =(u 1 ,u 2 ,u 3 ,u 4 )

[0257] Enter the value:

[0258]

[0259] Specific calculation:

[0260] d ij (1.0) = (5 + 2) e -0.1·0.5 ·e -0.05·1.0 +(7+3)·e -0.1·0.7 ·e -0.05·1.0 +(6+4)·e -0.1·0.6

[0261] ·e -0.0 5·1.0

[0262] d ij (1.0) = 7·e -0.1 +10·e -0.12 +10·e -0.11

[0263] d ij (1.0)≈7 0.9048+10 0.8869+10 0.8951

[0264] dij (1.0)≈24.1536

[0265] Through the above calculation, the shortest path length d from key node i to key node j is obtained ij (1.0)≈24.1536. This shows that at a given time point and transmission conditions, the cost and loss of the path are adjusted to about 24.1536 units. This result reflects the actual cost of the optimal path between two nodes taking into account changes in time and transmission conditions, which helps to more accurately evaluate the connectivity and efficiency of the power grid.

[0266] Based on the shortest path data, the average path length between all node pairs is calculated, wherein the average path length Calculated by the following formula:

[0267]

[0268] Where N represents the total number of nodes in the power grid, represents the average path length, represents the number of combinations of selecting two different nodes from N nodes, r(u,v) represents the reliability of connecting nodes u and v, represents the number of different paths from i to j.

[0269] The following is a detailed explanation of each parameter:

[0270] N represents the total number of nodes in the power grid, that is, the number of all nodes involved in the calculation; this parameter is directly obtained through the power grid topology diagram or node list, reflecting the size of the power grid system; it is one of the basic data for evaluating the overall connectivity and efficiency of the power grid;

[0271] Represents the weighted average path length between all node pairs, which changes over time. This parameter is calculated by comprehensively considering factors such as the shortest path length, path reliability, and path diversity. It is used to evaluate the overall connectivity and efficiency of the power grid and provide key information about the structural characteristics of the power grid.

[0272] Represents the number of combinations of selecting two different nodes from N nodes, mathematically equal to This parameter is a basic concept in combinatorial mathematics, which is used to determine the total number of paths between any two nodes and ensure that the formula covers all possible node pair combinations during statistical analysis;

[0273] r(u,v) represents the reliability of the connection nodes u and v, reflecting the stability of the path under different conditions. This parameter is usually estimated based on historical operation data, fault records or reliability test results, and is crucial for evaluating the safety and stability of the power grid.

[0274] |P ij | represents the number of different paths from node i to node j, reflecting the choice and redundancy of the path; this parameter is obtained by traversing the grid topology using a depth-first search (DFS) or breadth-first search (BFS) algorithm, which helps to measure the robustness and flexibility of the grid.

[0275] The reason for the design of the formula is to provide a comprehensive and accurate method for evaluating the overall connectivity and efficiency of the smart grid. The traditional average path length calculation method often ignores the impact of path reliability and diversity, and cannot accurately reflect the actual performance of the power grid. Therefore, this formula introduces path reliability r(u,v) and path diversity |P ij |, so that the calculation results not only consider the shortest path length, but also take into account the stability and choice of the path. In addition, the formula comprehensively considers the path lengths between all node pairs by weighted average, ensuring that the evaluation results can truly reflect the overall characteristics of the power grid. This helps power dispatchers better understand the power grid structure, formulate more reasonable optimization strategies, improve the stability and reliability of the system, and thus achieve efficient management of smart grids and optimal resource utilization.

[0276] Here is a specific example:

[0277] Input data:

[0278] The shortest path length d ij (t):d 12 (1.0) = 24.1536; d 13 (1.0) = 30.0; d 14 (1.0) = 35.0;

[0279] Path reliability r(u,v): r(u 1 ,u 2 )=0.95;r(u 2 ,u 3 )=0.9;r(u 3 ,u 4 )=0.92;

[0280] Number of different paths |P ij |:|P 12 |=1;|P 13 |=2;|P 14 |=3;

[0281] Total number of nodes N = 4;

[0282] Substituting the values:

[0283]

[0284] Specific calculation:

[0285]

[0286] Through the above calculation, the average connection distance between any two nodes in the entire power grid is obtained: This shows that the average connection distance between any two nodes in the power grid is about 7.50312 units, taking into account the shortest path length, path reliability and path diversity. This result provides important information about the overall connectivity and efficiency of the power grid, helping decision makers to better understand the structural characteristics of the power grid and develop corresponding optimization strategies to improve the stability and reliability of the system.

[0287] Figure 2 A schematic diagram of a load intelligent dynamic control device (or system) for a smart grid is provided for an embodiment of the present application, such as Figure 2 As shown, the device comprises:

[0288] The acquisition module 21 is used to collect real-time power parameters of key nodes in the smart grid using time series analysis and deep learning algorithms to obtain collected data, and predict power demand fluctuations within a preset time based on the collected data to obtain prediction results, and use graph theory algorithms to build a power grid topology model, identify key nodes and potential impact ranges of power transmission paths in the smart grid, and obtain topology information;

[0289] A generation module 22, configured to generate a load adjustment strategy for specific operation instructions for different power consumption areas or devices by using a multi-objective optimization algorithm in combination with a pre-set optimization goal in the smart grid according to the prediction result and the topology information, wherein the multi-objective optimization algorithm includes a genetic algorithm and a particle swarm optimization, and the optimization goals include minimizing energy costs, maximizing renewable energy utilization, and ensuring power supply stability;

[0290] Feedback module 23, used to convert the load adjustment strategy into guidance suggestions and push them to the end user, collect feedback information from the end user, apply reinforcement learning algorithm to learn the long-term behavior pattern and preference of the end user, generate a preference model including the end user's acceptance, feasibility evaluation results and personalized consulting solutions, and select and implement the corresponding load adjustment strategy based on the feedback information and the preference model to obtain the execution effect;

[0291] The detection module 24 utilizes the dynamic control system of the smart grid to monitor the execution effect and the operating status of the smart grid, and combines the Bayesian network to detect the risks of the execution effect and the abnormal conditions of the operating status, identify the deviation from the optimization target, and adjust the execution effect or switch to the backup load adjustment strategy.

[0292] Figure 2 The intelligent dynamic load control device for smart grid can execute Figure 1 The implementation principle and technical effect of the intelligent dynamic control method for loads in smart grids described in the illustrated embodiment will not be described in detail. The specific manner in which each module and unit performs operations in the intelligent dynamic control device for loads in smart grids in the above embodiment has been described in detail in the embodiment of the method, and will not be described in detail here.

[0293] In one possible design, Figure 2 The load intelligent dynamic control device for a smart grid in the embodiment shown can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0294] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0295] The processing component 32 is used for the above Figure 1 The embodiment provides a method for intelligent dynamic control of loads in a smart grid.

[0296] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to perform the above method.

[0297] The storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0298] Of course, the computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0299] The input / output interface provides an interface between the processing component and the peripheral interface module, which may be an output device, an input device, etc.

[0300] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0301] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0302] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 The illustrated embodiment is a method for intelligent dynamic load control in a smart grid.

[0303] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0304] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative work.

[0305] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0306] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for intelligent dynamic load control in a smart grid, characterized in that: include: Using time series analysis and deep learning algorithms to collect real-time power parameters of key nodes in the smart grid, obtain collected data, and predict power demand fluctuations within a preset time based on the collected data to obtain prediction results, and using graph theory algorithms to build a power grid topology model, identify key nodes and potential impact ranges of power transmission paths in the smart grid, and obtain topology information; According to the prediction results and the topological structure information, a multi-objective optimization algorithm is used in combination with the optimization objectives pre-set in the smart grid to generate a load adjustment strategy for specific operation instructions for different power consumption areas or equipment, wherein the multi-objective optimization algorithm includes a genetic algorithm and a particle swarm optimization, and the optimization objectives include minimizing energy costs, maximizing renewable energy utilization, and ensuring power supply stability; The load adjustment strategy is converted into guidance suggestions and pushed to the end user, feedback information from the end user is collected, a reinforcement learning algorithm is applied to learn the long-term behavior patterns and preferences of the end user, a preference model including the end user's acceptance, feasibility evaluation results and personalized consulting solutions is generated, and based on the feedback information and the preference model, a corresponding load adjustment strategy is selected and implemented to obtain an execution effect; The dynamic control system of the smart grid is used to monitor the execution effect and the operating status of the smart grid, and the risk of the execution effect and the abnormality of the operating status are detected in combination with the Bayesian network, the deviation from the optimization target is identified, and the execution effect is adjusted or switched to the backup load adjustment strategy.

2. The method according to claim 1, characterized in that The method of constructing a power grid topology model using a graph theory algorithm, identifying key nodes and potential impact ranges of power transmission paths in the smart grid, and obtaining topology information includes: Based on the key nodes and connection relationships in the smart grid, geographic information system data is introduced, and a directed weighted graph is constructed using complex network theory to obtain an initial structural model of the smart grid. The connection probability distribution of the key nodes is initialized in the initial structural model, the uncertainty and randomness of the grid are simulated, and a time-varying grid topology with a time dimension attribute is generated; Analyzing connectivity and redundancy of the smart grid according to the time-varying grid topology to obtain risk analysis results; Based on the risk analysis results, the key nodes are redefined using a page ranking algorithm and a hyperlink induced topic search algorithm, and target nodes whose behavior patterns deviate from normal behavior patterns are identified according to an anomaly detection method, and the historical data of the target nodes are compared with the problem points predicted by the current situation to obtain evaluation results of the key nodes and the target nodes, wherein the anomaly detection method includes an isolation forest and a local anomaly factor; Based on the evaluation results, Monte Carlo simulation combined with a propagation dynamics model is used to simulate the fault diffusion process of different node or path failure scenarios of the smart grid, estimate the impact range and recovery time of the fault, formulate an emergency response strategy, and use the emergency response strategy to integrate a deep reinforcement learning framework to generate solutions for specific types of faults; Based on the solution, the entities, relationships and attributes in the smart grid are organized into a semantic information network, and the information network is visualized in the form of a map. Blockchain technology is deployed according to the map form, the status of the smart grid after each update is recorded, and the topological structure information of the smart grid is obtained.

3. The method according to claim 2, characterized in that The analyzing the connectivity and redundancy of the smart grid further includes: The maximum flow and minimum cut theorem is used to determine the global and local maximum flow and minimum cut set of the smart grid, evaluate the maximum carrying capacity and the most vulnerable point of the smart grid, and obtain the global and local flow distribution and key node information of the smart grid; Utilizing the flow distribution and key node information, the shortest path between the key nodes and the average path length between all node pairs are calculated to generate a power grid performance evaluation report; According to the power grid performance evaluation report, the connectivity analysis is extended using hypergraph theory, and the subsystems of the smart grid are identified through a community discovery algorithm.

4. The method according to claim 2, characterized in that: Based on the risk analysis results, the key nodes are redefined using a page ranking algorithm and a hyperlink induced topic search algorithm, and the target nodes whose behavior patterns deviate from the normal behavior patterns are identified according to an anomaly detection method, and the historical data of the target nodes are compared with the problem points predicted by the current situation to obtain the evaluation results of the key nodes and the target nodes, including: Based on the risk analysis results, the importance of each of the key nodes in the time-varying power grid topology is evaluated using a page ranking algorithm and a hyperlink-induced topic search algorithm to obtain a key node score; According to the key node scores, the target nodes whose behavior patterns deviate from the normal behavior patterns are identified in combination with anomaly detection methods in machine learning to obtain a target node list; Compare the historical data of the target node with the current situation, analyze the target node list, and generate evaluation results of the key node and the target node.

5. The method according to claim 3, characterized in that: The flow distribution and key node information are used to calculate the shortest path between the key nodes and the average path length between all node pairs, and generate a power grid performance evaluation report, including: Based on the traffic distribution and key node information of the smart grid, the connectivity between the key nodes is analyzed, and the maximum carrying capacity and the most vulnerable point of the smart grid are determined by the maximum flow minimum cut theorem to obtain a preliminary evaluation result; Based on the preliminary evaluation results, the traffic distribution and key node information are used as inputs, and a graph theory algorithm is used to calculate the shortest path between the key nodes; Based on the shortest path, calculating the average path length between all node pairs to obtain the average connection distance between any two nodes in the smart grid; The shortest path data and the average connection distance are combined to generate a power grid performance evaluation report.

6. The method according to claim 1, characterized in that The method of using the dynamic control system of the smart grid to monitor the execution effect and the operation status of the smart grid, and combining the Bayesian network to detect the risk of the execution effect and the abnormality of the operation status, identifying the deviation from the optimization target, and adjusting the execution effect or switching to the standby load adjustment strategy includes: Using the dynamic control system of the smart grid to monitor the execution effect and the operating status to obtain monitoring data, wherein the monitoring data includes the operating parameters of the smart grid, and the operating parameters of the smart grid include voltage, current and power; Building a Bayesian network model based on the monitoring data, detecting the risk of the execution effect and the abnormality of the operation status, and generating a risk assessment result and an abnormality detection report, wherein the smart grid operation parameters serve as input variables of the Bayesian network model, and the risks and the abnormalities serve as output variables of the Bayesian network model; According to the risk assessment result and the abnormality detection report, combined with the optimization goal, the deviation is identified to obtain a deviation judgment conclusion; Based on the deviation judgment conclusion, the execution effect is adjusted to obtain an adjusted execution effect, and after the adjustment fails, switching to a standby load adjustment strategy.

7. The method according to claim 1, characterized in that The multi-objective optimization algorithm is combined with the optimization objectives pre-set in the smart grid to generate a load adjustment strategy for specific operation instructions for different power consumption areas or equipment, including: Selecting a multi-objective optimization algorithm based on the characteristics of the smart grid and a preset optimization goal, and initializing and processing parameters of the target optimization algorithm to obtain an optimization algorithm configuration; Determine input information according to the prediction result, the topology information and the operation status of the smart grid, wherein the input information is key data generated by the multi-objective optimization algorithm for the current status and future trend of the smart grid; Utilizing the optimization target, configuring the optimization algorithm and the input information to formulate a preliminary load adjustment strategy for different power consumption areas or equipment operation instructions; The preliminary load adjustment strategy is optimized based on a simulated annealing algorithm and an iterative optimization algorithm to generate the load adjustment strategy.

8. A load intelligent dynamic control system for smart grid, characterized in that: include: A collection module is used to collect real-time power parameters of key nodes in the smart grid using time series analysis and deep learning algorithms to obtain collected data, and predict power demand fluctuations within a preset time based on the collected data to obtain prediction results, and use graph theory algorithms to build a power grid topology model to identify key nodes and potential impact ranges of power transmission paths in the smart grid to obtain topology information; A generation module, configured to generate a load adjustment strategy for specific operation instructions for different power consumption areas or devices by using a multi-objective optimization algorithm in combination with the optimization objectives pre-set in the smart grid according to the prediction results and the topological structure information, wherein the multi-objective optimization algorithm includes a genetic algorithm and a particle swarm optimization, and the optimization objectives include minimizing energy costs, maximizing renewable energy utilization, and ensuring power supply stability; A feedback module is used to convert the load adjustment strategy into guidance suggestions and push them to the end user, collect feedback information from the end user, apply a reinforcement learning algorithm to learn the long-term behavior patterns and preferences of the end user, generate a preference model including the end user's acceptance, feasibility evaluation results and personalized consulting solutions, and select and implement the corresponding load adjustment strategy based on the feedback information and the preference model to obtain the execution effect; The detection module utilizes the dynamic control system of the smart grid to monitor the execution effect and the operating status of the smart grid, and combines the Bayesian network to detect the risks of the execution effect and the abnormal conditions of the operating status, identify the deviation from the optimization target, and adjust the execution effect or switch to the backup load adjustment strategy.

9. A computing device, characterized in that It comprises a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a load intelligent dynamic control method for a smart grid as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a computer, a method for intelligent dynamic control of loads for a smart grid as described in any one of claims 1 to 7 is implemented.

Images