An electric vehicle charging load power grid operation real-time scheduling optimization method and system

Abstract

The application discloses a real-time scheduling optimization method and system for electric vehicle charging load power grid operation, belongs to the technical field of electric vehicles, and solves the problems of insufficient coordination between new energy consumption and network congestion regulation and control and weak real-time dynamic response capability of real-time scheduling optimization for electric vehicle charging load power grid operation; the method comprises the following steps: constructing a vehicle-road-network coupling three-domain knowledge graph fusion architecture, bidirectionally coupling dynamic sensing of electric vehicle charging and discharging behaviors and sensing of charging pile load, outputting a final optimization scheme, and pushing the optimization scheme to a user side, a power grid side and a road network side; the application considers the coupling relationship among vehicles, a road network and a power grid, solves the power grid congestion peak regulation problem by coupling of three kinds of knowledge graphs and electric vehicle charging and discharging characteristics, enhances the coordination between new energy consumption and network congestion regulation and control, improves the real-time dynamic response capability of electric vehicles, and provides a technical reference for power grid operation and management personnel.

Description

Technical Field

[0001] This invention belongs to the field of electric vehicle technology and relates to a method and system for real-time scheduling and optimization of electric vehicle charging load grid operation. Background Technology

[0002] Currently, with the rapid popularization and application of new energy vehicles, electric vehicles are playing an increasingly important role in power grid dispatching and operation management. As the impact of electric vehicle (EV) charging loads on the power grid continues to grow, the power grid needs to ensure safe, stable, and economical operation through real-time monitoring, optimization, and dispatching. At present, the real-time dispatching of EV charging loads in the power grid has not fully considered the quantitative modeling of the bounded rationality behavior of EV users, resulting in insufficient coordination between renewable energy consumption and network congestion control, and weak real-time dynamic response capabilities, which are having an increasingly profound impact on the safe and stable operation of the power grid.

[0003] Knowledge graph technology possesses the ability to analyze multi-label big data for reasoning and is currently widely used in the power system field. For example, invention patent application CN118674175A discloses a knowledge graph-based method for modeling electric vehicle charging load characteristics. This method achieves refined mining of electric vehicle charging load characteristics and establishes a multi-dimensional, autonomous, and controllable model of typical electric vehicle charging load characteristics. However, existing technologies do not consider the coupling relationship between vehicles, networks, and the power grid, making it difficult to utilize the charging and discharging characteristics of electric vehicles to solve the problem of grid congestion and peak shaving. Real-time scheduling of electric vehicle charging load operation based on the aforementioned coupling relationship involves multiple knowledge graphs related to vehicles, networks, and the power grid, and also involves path planning for electric vehicles in different scenarios, resulting in a massive computational load and diverse scenarios. Summary of the Invention

[0004] The technical solution of this invention is used to solve the problems of insufficient coordination between new energy consumption and network congestion control, and weak real-time dynamic response capability in the real-time scheduling optimization of electric vehicle charging load grid operation.

[0005] The present invention solves the above-mentioned technical problems through the following technical solutions:

[0006] A method for real-time scheduling and optimization of electric vehicle charging load grid operation includes the following steps:

[0007] S1. Construct a vehicle-road-network coupled three-domain knowledge graph fusion architecture, including vehicle feature graph, road network resource graph and power grid topology graph, and store knowledge graph data through a graph database and update it in real time;

[0008] S2, a dynamic perception layer is constructed based on a vehicle-road-network coupled three-domain knowledge graph, including:

[0009] To construct a dynamic perception system for the charging and discharging behavior of electric vehicles, which can be used to predict the spatiotemporal distribution of electric vehicle charging demand;

[0010] Build a charging pile load sensing system to predict the charging and discharging power of the charging pile;

[0011] The dynamic perception of electric vehicle charging and discharging behavior is bidirectionally coupled with the perception of charging pile load to improve the accuracy of dynamic perception of electric vehicle charging behavior.

[0012] S3 constructs a multi-objective optimization layer based on a vehicle-road-network coupled three-domain knowledge graph, builds a multi-objective optimization function, and uses a hybrid adaptive particle swarm optimization to optimize the multi-objective optimization function, outputting the final optimization scheme;

[0013] S4 constructs a strategy generation layer based on a vehicle-road-network coupled three-domain knowledge graph, and pushes the optimization scheme to the user side, power grid side and road network side.

[0014] Furthermore, the entity of the vehicle feature map mentioned in S1 is the vehicle itself, and the categories of the vehicle feature map include basic vehicle information, user behavior profiles, and real-time data of the vehicle BMS; the attributes of the basic vehicle information include brand, model, fuel type, and rated power; the attributes of the user behavior profile include location preference, time preference, and bounded rationality threshold; the attributes of the real-time data of the vehicle BMS include charging status, health status, charging rate, and location data.

[0015] The entities in the power grid topology map are electric vehicle charging piles. The categories of the power grid topology map include basic power grid information, power grid information, and real-time status. The attributes of the basic power grid information include user number, capacity, number of parking spaces, charging type, address, and energy consumption category. The attributes of the power grid information include grid-connected lines and grid-connected substations. The attributes of the real-time status include available parking spaces, power margin, and real-time electricity price.

[0016] The road network resource map is categorized into traffic flow and dynamic road congestion; the attributes of traffic flow include path time and congestion level, while the attributes of dynamic road congestion include weather conditions and parking space prediction.

[0017] Furthermore, the storage and real-time updating of knowledge graph data through a graph database as described in S1 specifically refers to:

[0018] The graph database uses Neo4j graph database, which uses vehicle terminal data to dynamically update vehicle feature maps in real time, uses power grid SCADA and distribution network OMS data to establish power grid topology maps, and uses Gaode API to construct road network resource maps.

[0019] Further, S2 includes the following steps:

[0020] S21, construct dynamic perception of electric vehicle charging and discharging behavior, specifically construct a dynamic perception model of user charging demand, dynamically update vehicle feature map in real time through vehicle terminal, obtain historical SOC and SOH data of electric vehicle on the vehicle itself, use bidirectional long short-term memory network to predict electric vehicle charging and discharging power and judge electric vehicle charging and discharging behavior, and couple vehicle itself user behavior profile.

[0021] S22, Construct charging pile load perception, specifically by taking the charging pile itself as a unit, acquiring historical charging load data of the charging pile, and after data preprocessing, using a bidirectional long short-term memory network to predict the charging and discharging power of the charging pile.

[0022] S23 simultaneously couples the dynamic perception of electric vehicle charging and discharging behavior with the load perception of charging piles in both directions, taking into account the multi-temporal and spatial load characteristics, thereby improving the accuracy of dynamic perception of electric vehicle charging behavior.

[0023] Furthermore, the user charging demand dynamic perception model described in S21 is specifically an edge computing model based on incremental updates;

[0024] First, set incremental update trigger conditions. When the vehicle's charging status changes by more than 1% or more, its health status changes by more than 1% or more, and the time interval is more than or equal to 1 minute, the vehicle feature map is incrementally updated. Only the changed data is updated, and the unchanged data is not uploaded and updated.

[0025] Secondly, by adopting an edge computing model, abnormal data changes are processed only on the vehicle terminal side, filtering out instantaneous SOC and SOH changes.

[0026] Furthermore, the method described in S21, which uses a bidirectional long short-term memory network to predict the charging and discharging power of an electric vehicle and determine its charging and discharging behavior, coupled with the vehicle's user behavior profile, specifically involves:

[0027] First, the Long Short-Term Memory Network outputs the predicted charging and discharging power of the vehicle based on the historical SOC and SOH data of the individual vehicle, and analyzes the vehicle's SOC. When the SOC data increases, it is considered that the vehicle is charging, and vice versa.

[0028] Secondly, based on the predicted vehicle charging and discharging power and the determined charging and discharging behavior, the user location preference and time preference data in the vehicle feature map are coupled to obtain the real-time location and time preference results for electric vehicle charging and discharging.

[0029] In step S21, the Long Short-Term Memory Network prediction model is trained through a combination of incremental training and full training. Specifically, the prediction model is first trained incrementally, and the prediction model is dynamically iterated and fine-tuned using historical SOC and SOH data from the previous 7 days to continuously update and optimize the prediction results. The average prediction error of the prediction model is calculated, and when the average prediction error calculated for 3 consecutive days exceeds a preset threshold, the prediction model is trained in full.

[0030] Furthermore, the bidirectional coupling described in S23 specifically refers to:

[0031] By aggregating the charging and discharging power of each electric vehicle, the aggregated charging and discharging power of the charging pile at this moment is obtained:

[0032]

[0033] In the formula, express The charging power of charging piles aggregated at any given time. This indicates the number of electric vehicles aggregated. express At this moment The charging power of an electric vehicle ;

[0034] The calculation is based on a two-way coupling of user preference coefficients and historical feature coefficients, using the following logical representation:

[0035]

[0036] In the formula, Indicates bidirectional coupling Current charging power of charging stations Represents the user preference coefficient. Represents historical characteristic coefficients. express At any given time, a bidirectional long short-term memory network is used to predict the charging and discharging power of charging piles; among which, the user preference coefficient and historical characteristic coefficient First, solve the system of equations based on the data from the first two historical moments. Then, continuously update the system dynamically, using the following logic:

[0037]

[0038] In the formula, , These represent the results after bidirectional coupling. time, The charging power of the charging station at any given time. , They represent time, The charging and discharging power of the charging pile is predicted in real time using a bidirectional long short-term memory network. , They represent time, The charging power of charging piles at any given moment.

[0039] Further, S3 includes the following steps:

[0040] S31, on the grid side, design the objective function for renewable energy absorption rate, and add constraints such as line congestion peak shaving, power balance, charging pile capacity, charging pile parking space, SOC, and SOH.

[0041] S32, on the user side, design the user charging cost objective function, and add the maximum and minimum constraints of electricity price and user satisfaction constraints;

[0042] S33, on the road network side, design the objective function for optimizing the electric vehicle path;

[0043] S34. Based on the above objective function, construct a multi-objective optimization function, using the following logical representation:

[0044]

[0045] In the formula, Indicates the charging cost target. express Time-of-use electricity pricing at any given moment express The instantaneous charging power of electric vehicles at any given time. This indicates the total charging cost for the electric vehicle over the entire charging period. This indicates the time when the electric vehicle begins charging. Indicates the time when the electric vehicle finishes charging. This indicates the total charging time for the electric vehicle; This indicates the target for the rate of new energy consumption. P represents the amount of new energy consumed. EV This indicates the projected power generation from new energy sources. This represents the amount of new energy consumed during the period from the start to the end of a multi-objective optimization function. This represents the predicted power generation from new energy sources during the period from the start to the end of a multi-objective optimization function. This represents the entire optimization cycle of the multi-objective optimization function. Indicates the renewable energy consumption rate. , The goal is to optimize the path for electric vehicles. Path time for charging electric vehicles express , and Solve the following problem while satisfying the constraints S31 to S32 respectively. , and All achieve the smallest global optimal solution set;

[0046] S35 employs a hybrid adaptive particle swarm optimization multi-objective optimization function.

[0047] Furthermore, the multi-objective optimization function using hybrid adaptive particle swarm optimization described in S35 is specifically as follows:

[0048] S351, First, generate a particle swarm, where each particle represents a charging power allocation scheme;

[0049] Secondly, dynamic weight adjustment is performed; constraint processing is carried out according to the constraints of S31~S32, and the constraint processing specifically involves using a repair strategy to handle out-of-bounds particles;

[0050] S352, calculate the multi-objective fitness of each particle, and screen the non-dominated optimal solution set based on the Pareto front;

[0051] S353, adjust user preferences based on user satisfaction, specifically by performing a satisfaction-weighted sorting of the Pareto solution set and outputting the final optimized solution.

[0052] Furthermore, S4 specifically includes:

[0053] S41, On the user side, based on the above final optimization scheme, a minute-level rolling real-time optimization mechanism is established to push the user-side path selection to the electric vehicle terminal in real time;

[0054] S42, on the grid side, a real-time response curve for electric vehicle charging price is generated based on the real-time electricity price in the grid topology map, and line congestion and hot spot maps for renewable energy consumption are obtained based on the power flow information and renewable energy power in the grid topology map.

[0055] S43, on the road network side, generates an electric vehicle path flow map based on traffic flow in the road network knowledge graph, and obtains a charging pile parking space hotspot map based on the real-time status of charging piles in the power grid topology graph.

[0056] This invention also provides a real-time scheduling and optimization system for electric vehicle charging load grid operation, comprising:

[0057] The data storage module is used to construct a vehicle-road-network coupled three-domain knowledge graph fusion architecture, including vehicle feature graph, road network resource graph and power grid topology graph. The knowledge graph data is stored and updated in real time through a graph database.

[0058] The dynamic perception module, based on a vehicle-road-network coupled three-domain knowledge graph, constructs a dynamic perception layer, including:

[0059] To construct a dynamic perception system for the charging and discharging behavior of electric vehicles, which can be used to predict the spatiotemporal distribution of electric vehicle charging demand;

[0060] Build a charging pile load sensing system to predict the charging and discharging power of the charging pile;

[0061] The dynamic perception of electric vehicle charging and discharging behavior is bidirectionally coupled with the perception of charging pile load to improve the accuracy of dynamic perception of electric vehicle charging behavior.

[0062] The multi-objective optimization module constructs a multi-objective optimization layer based on a vehicle-road-network coupled three-domain knowledge graph, builds a multi-objective optimization function, and uses a hybrid adaptive particle swarm optimization to optimize the multi-objective optimization function, outputting the final optimization scheme.

[0063] The strategy generation module constructs a strategy generation layer based on a vehicle-road-network coupled three-domain knowledge graph, and pushes the optimization scheme to the user side, power grid side and road network side.

[0064] The advantages of this invention are:

[0065] (1) This invention takes into account the coupling relationship between vehicles, networks and power grids, integrates multi-dimensional data of vehicles, networks and power grids through the coupling of three knowledge graphs, and maps entities across domains based on graph databases to enhance the synergy between new energy consumption and network congestion control.

[0066] (2) The present invention constructs a dynamic perception layer based on a three-domain knowledge graph and updates the vehicle feature graph in real time through the vehicle terminal based on an incremental update edge computing model. While improving the quality of vehicle feature graph data, it effectively reduces the amount of electric vehicle update data, alleviates the difficulty of local data processing, and optimizes resource allocation.

[0067] This invention uses a bidirectional long short-term memory network to predict the charging and discharging power of each electric vehicle and determine its charging and discharging behavior. It couples the user-side charging and discharging behavior of electric vehicles with the user behavior profile of the vehicle itself, and aggregates the charging and discharging behavior perception results of each electric vehicle on a per-charging-pile basis. On the one hand, the charging pile uses a bidirectional long short-term memory network to predict the charging and discharging power; on the other hand, it uses user preference coefficients and historical feature coefficients to bidirectionally couple the predicted charging and discharging power with the charging and discharging behavior perception results of electric vehicles, effectively improving the accuracy of dynamic perception of electric vehicle charging behavior and predicting the spatiotemporal distribution of electric vehicle charging demand.

[0068] (3) Based on the topology analysis capability of knowledge graph, this invention constructs a multi-objective optimization function and designs constraints from three perspectives: the power grid side, the user side, and the road network side. It provides a quantitative model for the bounded rationality behavior of electric vehicle users, uses HAPSO to solve the multi-objective optimization function, and re-weights and sorts the output optimal solution set according to user satisfaction to generate an optimal charging scheme that better meets user preferences. It uses the charging and discharging characteristics of electric vehicles to solve the problem of grid congestion and peak shaving, enhances the synergy between new energy consumption and network congestion control, improves the real-time dynamic response capability of electric vehicles, and provides technical reference for power grid operation and management personnel. Attached Figure Description

[0069] Figure 1 This is a flowchart of a real-time scheduling and optimization method for electric vehicle charging load power grid operation according to Embodiment 1 of the present invention;

[0070] Figure 2 This is a schematic diagram of the structure of the vehicle-road-network coupled three-domain knowledge graph of Embodiment 1 of the present invention;

[0071] Figure 3 This is a schematic diagram of the structure of the dynamic sensing layer in Embodiment 1 of the present invention;

[0072] Figure 4 This is a schematic diagram of the structure of the multi-objective optimization layer in Embodiment 1 of the present invention;

[0073] Figure 5 This is a schematic diagram of the structure generated by the strategy in Embodiment 1 of the present invention. Detailed Implementation

[0074] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0075] The technical solution of the present invention will be further described below with reference to the accompanying drawings and specific embodiments:

[0076] Example 1

[0077] Knowledge graph technology, capable of analyzing multi-labeled big data for reasoning, is widely used in the power system field. Real-time scheduling of electric vehicle charging loads in the power grid involves multiple knowledge graphs and path planning for electric vehicles under different scenarios, resulting in massive computational demands and diverse scenarios. Knowledge graph technology, capable of establishing real-time optimized scheduling models for electric vehicles using graph models, is more suitable for the real-time scheduling method for electric vehicle charging loads based on multi-layered vehicle-road-grid coupled knowledge graphs provided in this invention.

[0078] Neo4j is an open-source graph database storage software. Its extensive practical applications both domestically and internationally have led to new developments in the software. Neo4j can also be seen as a high-performance graph engine with all the features of a mature database, providing a foundation for the implementation of the real-time scheduling method for electric vehicle charging load grid operation based on a multi-layer vehicle-road-network coupled knowledge graph provided in this invention.

[0079] To provide a better understanding of the structural features and effects achieved by the present invention, a detailed description is provided below, accompanied by preferred embodiments and accompanying drawings:

[0080] like Figure 1 As shown, specifically, a real-time scheduling optimization method for electric vehicle charging load grid operation based on a multi-layer vehicle-road-grid coupled knowledge graph is disclosed, including the following steps:

[0081] S1 constructs a vehicle-road-network coupled three-domain knowledge graph fusion architecture, and stores knowledge graph data through a graph database and updates it in real time.

[0082] Currently, the power grid ledger contains a large amount of historical data on electric vehicle charging loads and power grid topology data. This data is used to construct a vehicle-road-grid coupled three-domain knowledge graph fusion architecture. This architecture mainly comprises three knowledge graph subgraphs: a vehicle feature graph (VKG), a road network resource graph (RKG), and a power grid topology graph (GKG). The vehicle feature graph describes the current operating status of electric vehicles and owner preferences; the road network resource graph describes the road traffic status of vehicles; and the power grid topology graph describes the power grid operating status. This VKG-road-grid coupled three-domain knowledge graph fusion architecture achieves multi-layer data coupling between vehicles, roads, and the grid, providing a data foundation for real-time scheduling and optimization of electric vehicle charging load power grid operation.

[0083] Specifically, such as Figure 2As shown, the vehicle-road-network coupled three-domain knowledge graph fusion architecture in this embodiment includes three knowledge graph subgraphs: Vehicle Feature Graph (VKG), Road Network Resource Graph (RKG), and Power Grid Topology Graph (GKG). Each knowledge graph subgraph is modeled in the form of triples.

[0084] The vehicle feature map employs dynamic ontology modeling technology, with the vehicle itself as the entity. It includes three categories: basic vehicle information, user behavior profiles, and real-time data from the vehicle's BMS. The basic vehicle information includes four attributes: brand, model, fuel type, and rated power. The user behavior profile includes three attributes: location preference, time preference, and bounded rationality threshold. The real-time data from the vehicle's BMS includes four attributes: state of charge (SOC), state of health (SOH), charging rate, and location data, and is updated in real-time based on the vehicle's real-time status.

[0085] The power grid topology map uses electric vehicle charging stations as entities and includes three main categories: basic power grid information, power grid information, and real-time status. The basic power grid information includes six attributes: user number, capacity, number of parking spaces, charging type, address, and energy consumption category. The power grid information includes two attributes: grid-connected lines and grid-connected substations. Grid-connected lines include four sub-attributes: power flow information, renewable energy power, thermal stability limit, and energy storage information. The real-time status includes three attributes: available parking spaces, power margin, and real-time electricity price.

[0086] The road network resource map includes two categories: traffic flow and dynamic road resistance. Traffic flow includes two attributes: path time and congestion level, while dynamic road resistance includes two attributes: weather conditions and parking space prediction.

[0087] In this embodiment, the graph database uses Neo4j to store vehicle-road-network coupled three-domain knowledge graph data, achieving vehicle-road-network coupled three-domain knowledge graph fusion. Vehicle feature graphs are dynamically updated in real time using vehicle terminal data, and a power grid topology graph is established using power grid SCADA and distribution network OMS data. Road network resource graphs are automatically constructed using the Gaode API. Through the fusion of full-domain data from vehicle terminals, roadside units, and the power grid SCADA system, a three-domain knowledge graph with spatiotemporal correlation characteristics is constructed, enabling real-time capture of electric vehicle movement trajectories, traffic network status, and power grid operating conditions.

[0088] S2, a dynamic perception layer is constructed based on a vehicle-road-network coupled three-domain knowledge graph, including:

[0089] To construct a dynamic perception system for the charging and discharging behavior of electric vehicles, which can be used to predict the spatiotemporal distribution of electric vehicle charging demand;

[0090] Build a charging pile load sensing system to predict the charging and discharging power of the charging pile;

[0091] The dynamic perception of electric vehicle charging and discharging behavior is bidirectionally coupled with the perception of charging pile load to improve the accuracy of dynamic perception of electric vehicle charging behavior.

[0092] like Figure 3 As shown, this embodiment constructs a dynamic perception layer based on a vehicle-road-network coupled three-domain knowledge graph fusion architecture, specifically including the following steps:

[0093] S21, construct dynamic perception of electric vehicle charging and discharging behavior, specifically by constructing a dynamic perception model of user charging demand. The vehicle feature map is updated in real time through the vehicle terminal. Taking the vehicle itself as the unit, the historical SOC and SOH data of the electric vehicle are obtained. The historical data of individual vehicles are used to predict the charging and discharging power of the electric vehicle and judge the charging and discharging behavior of the electric vehicle, and coupled with the user behavior profile of the vehicle itself.

[0094] In this embodiment, the real-time dynamic update of the vehicle feature map via the vehicle terminal specifically involves the vehicle terminal reading vehicle data via the OBD-II interface at a high frequency of 1Hz using the CAN bus, and dynamically updating the real-time data of the electric vehicle using local caching and compressed uploading.

[0095] Specifically, due to the massive amount of data from electric vehicles, this embodiment constructs a dynamic perception model for user charging demand based on an incremental update edge computing model to dynamically update real-time electric vehicle data. First, incremental update trigger conditions are set: when the vehicle's charging status changes by more than 1% or more, its health status changes by more than 1% or more, and the time interval is more than or equal to 1 minute, the vehicle feature map is incrementally updated. Only the changed data is updated; unchanged data is not uploaded or updated, effectively reducing the amount of updated electric vehicle data. Second, to improve the quality of vehicle feature map data and reduce the difficulty of local data processing, this embodiment adopts an edge computing mode, processing abnormal jump data only on the vehicle terminal side and filtering instantaneous SOC and SOH jumps, thereby improving the quality and efficiency of vehicle feature map data updates.

[0096] In this embodiment, the incremental update triggering condition is set using the following logical representation:

[0097]

[0098]

[0099]

[0100] In the formula, This represents the vehicle charging status data at the previous time point (time k). This represents the vehicle charging status data at the current time (time k+1). This represents the vehicle health status data at the previous time point (time k). This represents the vehicle health status data at the current time (time k+1). It represents the time interval between the previous moment and the current moment.

[0101] In this embodiment, historical SOC and SOH data of individual vehicles are obtained through vehicle feature maps. A bidirectional long short-term memory network (BiLSTM) is used to predict the charging and discharging behavior of electric vehicles. This prediction is then coupled with the user behavior profile of the vehicle to achieve the perception of electric vehicle charging and discharging behavior. First, the long short-term memory network outputs the predicted vehicle charging and discharging power based on the historical SOC and SOH data of individual vehicles, and analyzes the vehicle's state of charge (SOC). When the SOC data increases, it is considered that the vehicle is charging; otherwise, it is considered that the vehicle is discharging. Second, based on the predicted vehicle charging and discharging power and the determined charging and discharging behavior, the user location preference and time preference data in the vehicle feature map are coupled to obtain the real-time location and time preference results of electric vehicle charging and discharging.

[0102] In this embodiment, the Long Short-Term Memory (LSTM) network prediction model is trained through a combination of incremental and full training. Specifically, the prediction model is first incrementally trained by dynamically iterating and fine-tuning the model using historical SOC and SOH data from the previous 7 days, continuously updating and optimizing the prediction results. The average prediction error of the prediction model is then calculated. When the average prediction error calculated over 3 consecutive days exceeds a preset threshold, the prediction model is then fully trained. The average prediction error of the prediction model is calculated using the following logic:

[0103]

[0104] In the formula, Indicates the average prediction error. This indicates the time involved in calculating the average prediction error, expressed in 15-minute increments. Divided into multiple time periods, Indicates the first The prediction error at time t, i.e., the calculation of the t... A 15-minute prediction error.

[0105] Furthermore, by dynamically sensing the charging and discharging behavior of electric vehicles, regional charging hotspot identification can be established in 1-minute increments, providing a reference for power grid operation and management personnel and electric vehicle owners regarding the charging and discharging behavior of electric vehicles.

[0106] In this embodiment, by aggregating the charging location preference and time preference of each electric vehicle on a per-charging-pile basis using the above-mentioned dynamic perception results of electric vehicle charging and discharging behavior, the predicted number of electric vehicles charged for each charging pile can be obtained, and regional-level charging hotspot identification on a per-charging-pile basis can be formed based on the predicted data.

[0107] S22, Construct a charging pile load perception system. Specifically, take the charging pile itself as a unit, obtain historical charging load data of the charging pile, and after data preprocessing, use a bidirectional long short-term memory network (BiLSTM) to predict the charging and discharging power of the charging pile.

[0108] In this embodiment, the historical charging load data of the charging pile can be obtained from the power flow information in the power grid topology map.

[0109] S23 simultaneously couples the dynamic perception of electric vehicle charging and discharging behavior with the load perception of charging piles in both directions, taking into account the multi-temporal and spatial load characteristics, thereby improving the accuracy of dynamic perception of electric vehicle charging behavior.

[0110] In this embodiment, the charging and discharging power of each electric vehicle is aggregated to obtain the aggregated charging and discharging power of the charging pile at this moment:

[0111]

[0112] In the formula, express The charging power of charging piles aggregated at any given time. This indicates the number of electric vehicles aggregated. express At this moment The charging power of an electric vehicle can be obtained from the rated power in the vehicle characteristic map. .

[0113] From the user's perspective, the aggregated charging power of charging stations can reflect vehicle preference information, such as... Electric vehicles prefer to charge at the designated charging stations compared to other charging stations. From the perspective of the power grid, the predicted charging and discharging power data of the charging stations reflects grid-side characteristic data. This embodiment couples these two pieces of information to improve the accuracy of dynamic perception of electric vehicle charging behavior based on multi-temporal and spatial load characteristics. Specifically, this embodiment calculates bidirectionally based on user preference coefficients and historical characteristic coefficients, using the following logical representation:

[0114]

[0115] In the formula, Indicates bidirectional coupling Current charging power of charging stations Represents the user preference coefficient. Represents historical characteristic coefficients. express At any given time, a bidirectional long short-term memory network is used to predict the charging and discharging power of the charging pile. Among these factors, the user preference coefficient... and historical characteristic coefficient First, solve the system of equations based on the data from the first two historical moments, then follow the time... , as well as The changes are constantly and dynamically updated, which can be represented by the following logic:

[0116]

[0117] In the formula, , These represent the results after bidirectional coupling. time, The charging power of the charging station at any given time. , They represent time, The charging and discharging power of the charging pile is predicted in real time using a bidirectional long short-term memory network. , They represent time, The charging power of charging piles at any given moment.

[0118] Current electric vehicle (EV)-based grid dispatching models often only consider the real-time operating status of the power grid and road network, neglecting the spatiotemporal distribution of charging demand due to EV mobility characteristics, as well as the load transfer of charging stations caused by road network congestion propagation. This embodiment constructs a dynamic perception layer based on a vehicle-road-network coupled three-domain knowledge graph. First, it builds a dynamic perception of EV charging and discharging behavior, predicting the charging and discharging power of individual EVs and judging their charging and discharging behavior, coupled with a user behavior profile of the vehicle itself. Second, based on multiple user behavior profiles of the vehicle itself, it aggregates the charging power of individual charging piles and bidirectionally couples this with the predicted charging power output by the charging pile based on the prediction model. This establishes a bidirectional coupling model for EV charging and discharging power prediction and charging pile charging and discharging power prediction based on a three-domain knowledge graph, enabling dynamic perception of the multi-temporal and spatiotemporal load characteristics of EV charging and effectively improving the accuracy of dynamic perception of EV charging behavior.

[0119] S3 constructs a multi-objective optimization layer based on a vehicle-road-network coupled three-domain knowledge graph, builds a multi-objective optimization function, and uses a hybrid adaptive particle swarm optimization to optimize the multi-objective optimization function, outputting the final optimization scheme.

[0120] like Figure 4As shown, this embodiment, based on the aforementioned vehicle-road-network coupled three-domain knowledge graph, constructs a multi-objective optimization layer through the aforementioned dynamic perception layer based on the vehicle-road-network coupled three-domain knowledge graph. Specifically, it includes the following steps:

[0121] S31, On the grid side, design the objective function for renewable energy absorption rate, adding constraints such as line congestion peak-shaving, power balance, charging pile capacity, charging pile parking space, SOC, and SOH. The objective function for renewable energy absorption rate is expressed using the following logic:

[0122]

[0123] in, This indicates the target for the rate of new energy consumption. P represents the amount of renewable energy absorbed, specifically the total amount of renewable energy that can be absorbed by the power grid. EV This indicates the predicted power generation from new energy sources, specifically the predicted power generation from new energy sources on the grid side.

[0124] In this embodiment, to comprehensively consider the coupling between the vehicle, road, and grid, constraints are designed from the perspective of the distribution network, including the following grid-side constraints:

[0125] Line congestion peak shaving constraints are represented using the following logic:

[0126]

[0127] In the formula, The active power of the line. This is the minimum load power of the line. The above data, representing the maximum load power of the line, is derived from power flow information in the power grid topology map.

[0128] Power balance constraints are represented using the following logic:

[0129]

[0130] In the formula, Indicates bidirectional coupling The charging power of the charging pile at any given time is calculated based on S2 and stored in the power margin in the power grid topology map and dynamically updated. Indicates other loads, This represents the amount of electricity generated from energy storage, and the data comes from energy storage information in the power grid topology map.

[0131] The charging pile capacity constraint is represented by the following logic:

[0132]

[0133] In the formula, , These represent the minimum and maximum power of the charging pile, respectively, with the data sourced from the power margin in the power grid topology map.

[0134] The parking space constraints for charging stations are represented using the following logic:

[0135]

[0136] In the formula, express The charging stations are currently occupied. , They represent The minimum and maximum number of parking spaces for charging stations at any given time are calculated from the available parking spaces in the power grid topology map.

[0137] SOC constraints are represented using the following logic:

[0138]

[0139] In the formula, express The charging status of an electric vehicle at any given time. , They represent Minimum charging time for electric vehicles ,maximum .

[0140] SOH constraints are represented using the following logic:

[0141]

[0142] In the formula, express The current health status of electric vehicles being charged. , They represent Minimum charging time for electric vehicles ,maximum .

[0143] S32, on the user side, by adding a charging cost-time game model, a dynamic game equilibrium between individual charging demand and group electricity consumption behavior is achieved. Specifically, the user charging cost objective function is designed, and constraints on the maximum and minimum electricity prices and user satisfaction are added. The user charging cost objective function is expressed using the following logic:

[0144]

[0145] In the formula, Indicates the charging cost target. express The time-of-use electricity price at any given moment is based on real-time electricity prices from the power grid topology map. express The instantaneous charging power of electric vehicles at any given time. This indicates the time when the electric vehicle begins charging. This indicates the time when the electric vehicle finishes charging.

[0146] In this embodiment, the user-side constraints include:

[0147] The maximum and minimum constraints on electricity prices are represented using the following logic:

[0148]

[0149] In the formula, express The current electricity price for charging electric vehicles. , They represent The minimum and maximum electricity prices for charging electric vehicles at any given time are obtained from the real-time electricity prices in the power grid topology map.

[0150] User satisfaction constraints are represented logically as follows:

[0151]

[0152] In the formula, express Current user satisfaction , They represent The minimum and maximum user satisfaction at any given time are determined by data derived from user behavior profiles in the vehicle feature map.

[0153] S33, on the road network side, design the objective function for optimizing the electric vehicle path, specifically:

[0154]

[0155] in, The goal is to optimize the path for electric vehicles. The path time for charging electric vehicles is derived from path time data in the road network resource map.

[0156] S34. Based on the above objective function, construct a multi-objective optimization function, using the following logical representation:

[0157]

[0158] In the formula, This indicates the total charging cost for the electric vehicle over the entire charging period. This indicates the total charging time for the electric vehicle. This represents the amount of new energy consumed during the period from the start to the end of a multi-objective optimization function. This represents the predicted power generation from new energy sources during the period from the start to the end of a multi-objective optimization function. This represents the entire optimization cycle of the multi-objective optimization function. Indicates the renewable energy consumption rate. , express , and Solve the following problem while satisfying the constraints S31 to S32 respectively. , and All of them reach the minimum global optimal solution set.

[0159] S35 employs a hybrid adaptive particle swarm optimization multi-objective optimization function.

[0160] This embodiment uses Hybrid Adaptive Particle Swarm Optimization (HAPSO) to solve the above optimization algorithm, specifically as follows:

[0161] S351, First, generate a particle swarm, where each particle represents a charging power allocation scheme; the charging power allocation scheme is specifically the charging pile and charging time scheme selected by the electric vehicle.

[0162] Secondly, dynamic weight adjustment is performed; constraint processing is carried out according to the constraints of S31~S32. The constraint processing specifically involves using a repair strategy to handle out-of-bounds particles, such as adjusting the charging power to the boundary value to ensure that the charging power does not exceed the maximum or minimum charging power, and constraining the charging pile parking space, etc.

[0163] S352, calculate the multi-objective fitness of each particle, and screen the non-dominated optimal solution set based on the Pareto front. Specifically, use the multi-objective optimization function to calculate the target fitness of each particle on the road network side, power grid side, and user side respectively, and screen out the optimal solution set.

[0164] The Pareto front is one of the methods used in multi-objective optimization to describe the set of optimal solutions. Since multi-objective optimization problems often have multiple objectives, and these objectives may conflict, the Pareto optimal solution set represents the solutions that can be further optimized without sacrificing other objectives.

[0165] In this embodiment, the multi-objective fitness of each particle is calculated. Specifically, a multi-objective optimization function is used to evaluate the target fitness of each particle's solution on the road network side, power grid side, and user side, respectively. By comparing the multi-objective fitness of different particles, the optimal solution set is selected.

[0166] S353, adjust user preferences based on user satisfaction, specifically by performing a satisfaction-weighted sorting of the Pareto solution set and outputting the final optimized solution.

[0167] In this embodiment, considering that users may have different preferences for certain objectives, such as charging time, real-time electricity price, or location preference, this embodiment performs a satisfaction-weighted adjustment on the fitness of each solution in the Pareto solution set, so that the priority of different objectives will be adjusted according to user preferences, in order to generate an optimal charging scheme that better meets user preferences.

[0168] In step S3, a multi-objective optimization layer is constructed based on a vehicle-road-network coupled three-domain knowledge graph. On the grid side, based on the topology analysis capability of the knowledge graph, a new energy absorption rate objective function considering the dynamic thermal stability limit of the line is constructed, and constraints are designed from the perspective of the distribution network to achieve spatiotemporal matching between peak-shaving resources and fluctuating power sources. On the user side, a charging cost-time game model is designed, and a user charging cost objective function is established to achieve dynamic game equilibrium between individual charging demand and group electricity consumption behavior. On the road network side, a joint optimization of charging guidance and traffic management is developed, coupling charging station capacity constraints and traffic flow. The optimal charging navigation strategy is generated using the path reasoning capability of the knowledge graph. Finally, a multi-objective optimization function from three perspectives is constructed, and HAPSO is used to solve the multi-objective optimization function. The optimal solution set output is re-weighted and sorted according to user satisfaction to generate an optimal charging scheme that better meets user preferences.

[0169] S4 constructs a strategy generation layer based on a vehicle-road-network coupled three-domain knowledge graph, and pushes the optimization scheme to the user side, power grid side and road network side.

[0170] like Figure 5 As shown, this embodiment establishes a strategy generation layer for a vehicle-road-network coupled three-domain knowledge graph based on the above steps S1~S3, specifically including the following steps:

[0171] S41, On the user side, based on the above final optimization scheme, a minute-level rolling real-time optimization mechanism is established to push the user-side path selection to the electric vehicle terminal in real time, specifically as follows:

[0172] Based on the location data in the vehicle feature knowledge graph, the real-time location information of the vehicle is updated every minute. Since different vehicles have different locations, the paths to the surrounding charging stations are also different. Based on the real-time location information of the vehicle, the final optimization plan output by S3 (specifically the path from the current vehicle location to the charging station) is pushed to the user side in real time to ensure the effectiveness of the optimization results.

[0173] S42, on the grid side, a real-time response curve for electric vehicle charging price is generated based on the real-time electricity price in the grid topology map. Line congestion and hot spot maps for renewable energy consumption are obtained based on the power flow information and renewable energy power in the grid topology map for reference by power grid operation and management personnel.

[0174] S43, on the road network side, generates an electric vehicle path flow map based on traffic flow in the road network knowledge graph, and obtains a charging pile parking space hotspot map from the real-time status of charging piles in the power grid topology graph for reference.

[0175] In step S4, on the user side, this embodiment accurately and in real-time pushes the user-side path selection to the electric vehicle terminal based on the optimization results output by the optimization layer, and establishes a minute-level rolling real-time optimization mechanism to push the real-time optimization results to the user side to ensure the real-time nature of the optimization results. Through a strategy generation layer based on a vehicle-road-network coupled three-domain knowledge graph, real-time scheduling optimization of electric vehicle charging load grid operation can be achieved, thereby improving the solutions to problems such as insufficient quantitative modeling of bounded rationality behavior of electric vehicle users, inadequate coordination between renewable energy consumption and network congestion control, and weak real-time dynamic response capabilities.

[0176] This invention also provides a real-time scheduling and optimization system for electric vehicle charging load grid operation, comprising:

[0177] The data storage module is used to construct a vehicle-road-network coupled three-domain knowledge graph fusion architecture, including vehicle feature graph, road network resource graph and power grid topology graph. The knowledge graph data is stored and updated in real time through a graph database.

[0178] In this embodiment, the vehicle-road-network coupled three-domain knowledge graph fusion architecture includes three knowledge graph subgraphs: vehicle feature graph (VKG), road network resource graph (RKG), and power grid topology graph (GKG). Each knowledge graph subgraph is modeled in the form of triples.

[0179] The vehicle feature map employs dynamic ontology modeling technology, with the vehicle itself as the entity. It includes three categories: basic vehicle information, user behavior profiles, and real-time data from the vehicle's BMS. The basic vehicle information includes four attributes: brand, model, fuel type, and rated power. The user behavior profile includes three attributes: location preference, time preference, and bounded rationality threshold. The real-time data from the vehicle's BMS includes four attributes: state of charge (SOC), state of health (SOH), charging rate, and location data, and is updated in real-time based on the vehicle's real-time status.

[0180] The power grid topology map uses electric vehicle charging stations as entities and includes three main categories: basic power grid information, power grid information, and real-time status. The basic power grid information includes six attributes: user number, capacity, number of parking spaces, charging type, address, and energy consumption category. The power grid information includes two attributes: grid-connected lines and grid-connected substations. Grid-connected lines include four sub-attributes: power flow information, renewable energy power, thermal stability limit, and energy storage information. The real-time status includes three attributes: available parking spaces, power margin, and real-time electricity price.

[0181] The road network resource map includes two categories: traffic flow and dynamic road resistance. Traffic flow includes two attributes: path time and congestion level, while dynamic road resistance includes two attributes: weather conditions and parking space prediction.

[0182] In this embodiment, the graph database uses Neo4j to store vehicle-road-network coupled three-domain knowledge graph data, achieving vehicle-road-network coupled three-domain knowledge graph fusion. Vehicle feature graphs are dynamically updated in real time using vehicle terminal data, and a power grid topology graph is established using power grid SCADA and distribution network OMS data. Road network resource graphs are automatically constructed using the Gaode API. Through the fusion of full-domain data from vehicle terminals, roadside units, and the power grid SCADA system, a three-domain knowledge graph with spatiotemporal correlation characteristics is constructed, enabling real-time capture of electric vehicle movement trajectories, traffic network status, and power grid operating conditions.

[0183] The dynamic perception module, based on a vehicle-road-network coupled three-domain knowledge graph, constructs a dynamic perception layer and includes the following units:

[0184] The dynamic perception unit is used to build a dynamic perception model of user charging demand. It updates the vehicle feature map in real time through the vehicle terminal, obtains the historical SOC and SOH data of electric vehicles on the basis of the vehicle itself, and uses a bidirectional long short-term memory network (BiLSTM) to predict the charging and discharging power of electric vehicles and judge the charging and discharging behavior of electric vehicles, and couples the user behavior profile of the vehicle itself.

[0185] In this embodiment, the real-time dynamic update of the vehicle feature map via the vehicle terminal specifically involves the vehicle terminal reading vehicle data via the OBD-II interface at a high frequency of 1Hz using the CAN bus, and dynamically updating the real-time data of the electric vehicle using local caching and compressed uploading.

[0186] Specifically, due to the massive amount of data from electric vehicles, this embodiment constructs a dynamic perception model for user charging demand based on an incremental update edge computing model to dynamically update real-time electric vehicle data. First, incremental update trigger conditions are set: when the vehicle's charging status changes by more than 1% or more, its health status changes by more than 1% or more, and the time interval is more than or equal to 1 minute, the vehicle feature map is incrementally updated. Only the changed data is updated; unchanged data is not uploaded or updated, effectively reducing the amount of updated electric vehicle data. Second, to improve the quality of vehicle feature map data and reduce the difficulty of local data processing, this embodiment adopts an edge computing mode, processing abnormal jump data only on the vehicle terminal side and filtering instantaneous SOC and SOH jumps, thereby improving the quality and efficiency of vehicle feature map data updates.

[0187] In this embodiment, the incremental update triggering condition is set using the following logical representation:

[0188]

[0189]

[0190]

[0191] In the formula, This represents the vehicle charging status data at the previous time point (time k). This represents the vehicle charging status data at the current time (time k+1). This represents the vehicle health status data at the previous time point (time k). This represents the vehicle health status data at the current time (time k+1). It represents the time interval between the previous moment and the current moment.

[0192] In this embodiment, historical SOC and SOH data of individual vehicles are obtained through vehicle feature maps. A bidirectional long short-term memory network (BiLSTM) is used to predict the charging and discharging behavior of electric vehicles. This prediction is then coupled with the user behavior profile of the vehicle to achieve the perception of electric vehicle charging and discharging behavior. First, the long short-term memory network outputs the predicted vehicle charging and discharging power based on the historical SOC and SOH data of individual vehicles, and analyzes the vehicle's state of charge (SOC). When the SOC data increases, it is considered that the vehicle is charging; otherwise, it is considered that the vehicle is discharging. Second, based on the predicted vehicle charging and discharging power and the determined charging and discharging behavior, the user location preference and time preference data in the vehicle feature map are coupled to obtain the real-time location and time preference results of electric vehicle charging and discharging.

[0193] In this embodiment, the Long Short-Term Memory (LSTM) network prediction model is trained through a combination of incremental and full training. Specifically, the prediction model is first incrementally trained by dynamically iterating and fine-tuning the model using historical SOC and SOH data from the previous 7 days, continuously updating and optimizing the prediction results. The average prediction error of the prediction model is then calculated. When the average prediction error calculated over 3 consecutive days exceeds a preset threshold, the prediction model is then fully trained. The average prediction error of the prediction model is calculated using the following logic:

[0194]

[0195] In the formula, Indicates the average prediction error. This indicates the time involved in calculating the average prediction error, expressed in 15-minute increments. Divided into multiple time periods, Indicates the first The prediction error at time t, i.e., the calculation of the t... A 15-minute prediction error.

[0196] The load sensing unit is used to acquire historical charging load data of the charging pile itself. After data preprocessing, the charging and discharging power of the charging pile is predicted using a bidirectional long short-term memory network (BiLSTM).

[0197] In this embodiment, the historical charging load data of the charging pile can be obtained from the power flow information in the power grid topology map.

[0198] The bidirectional coupling unit is used to simultaneously couple the dynamic perception of electric vehicle charging and discharging behavior with the perception of charging pile load, taking into account multi-temporal and spatial load characteristics, thereby improving the accuracy of dynamic perception of electric vehicle charging behavior.

[0199] In this embodiment, the charging and discharging power of each electric vehicle is aggregated to obtain the aggregated charging and discharging power of the charging pile at this moment:

[0200]

[0201] In the formula, express The charging power of charging piles aggregated at any given time. This indicates the number of electric vehicles aggregated. express At this moment The charging power of an electric vehicle can be obtained from the rated power in the vehicle characteristic map. .

[0202] From the user's perspective, the aggregated charging power of charging stations can reflect vehicle preference information, such as... Electric vehicles prefer to charge at the designated charging stations compared to other charging stations. From the perspective of the power grid, the predicted charging and discharging power data of the charging stations reflects grid-side characteristic data. This embodiment couples these two pieces of information to improve the accuracy of dynamic perception of electric vehicle charging behavior based on multi-temporal and spatial load characteristics. Specifically, this embodiment calculates bidirectionally based on user preference coefficients and historical characteristic coefficients, using the following logical representation:

[0203]

[0204] In the formula, Indicates bidirectional coupling Current charging power of charging stations Represents the user preference coefficient. Represents historical characteristic coefficients. express At any given time, a bidirectional long short-term memory network is used to predict the charging and discharging power of the charging pile. Among these factors, the user preference coefficient... and historical characteristic coefficient First, solve the system of equations based on the data from the first two historical moments, then follow the time... , as well as The changes are constantly and dynamically updated, which can be represented by the following logic:

[0205]

[0206] In the formula, , These represent the results after bidirectional coupling. time, The charging power of the charging station at any given time. , They represent time, The charging and discharging power of the charging pile is predicted in real time using a bidirectional long short-term memory network. , They represent time, The charging power of charging piles at any given moment.

[0207] The multi-objective optimization module constructs a multi-objective optimization layer based on a vehicle-road-network coupled three-domain knowledge graph, builds a multi-objective optimization function, and employs a hybrid adaptive particle swarm optimization approach to optimize the multi-objective optimization function, outputting the final optimized solution. Specifically, the multi-objective optimization module includes the following units:

[0208] The grid-side unit is used to design the objective function for renewable energy absorption rate. It incorporates constraints such as line congestion and peak shaving, power balance, charging pile capacity, charging pile parking space, State of Charge (SOC), and State of Hypothesis (SOH). The objective function for renewable energy absorption rate is expressed using the following logic:

[0209]

[0210] in, This indicates the target for the rate of new energy consumption. P represents the amount of renewable energy absorbed, specifically the total amount of renewable energy that can be absorbed by the power grid. EV This indicates the predicted power generation from new energy sources, specifically the predicted power generation from new energy sources on the grid side.

[0211] In this embodiment, to comprehensively consider the coupling between the vehicle, road, and grid, constraints are designed from the perspective of the distribution network, including the following grid-side constraints:

[0212] Line congestion peak shaving constraints are represented using the following logic:

[0213]

[0214] In the formula, The active power of the line. This is the minimum load power of the line. The above data, representing the maximum load power of the line, is derived from power flow information in the power grid topology map.

[0215] Power balance constraints are represented using the following logic:

[0216]

[0217] In the formula, Indicates bidirectional coupling The charging power of the charging pile at any time is dynamically updated based on the power margin calculated by the dynamic sensing module and stored in the power grid topology map. Indicates other loads, This represents the amount of electricity generated from energy storage, and the data comes from energy storage information in the power grid topology map.

[0218] The charging pile capacity constraint is represented by the following logic:

[0219]

[0220] In the formula, , These represent the minimum and maximum power of the charging pile, respectively, with the data sourced from the power margin in the power grid topology map.

[0221] The parking space constraints for charging stations are represented using the following logic:

[0222]

[0223] In the formula, express The charging stations are currently occupied. , They represent The minimum and maximum number of parking spaces for charging stations at any given time are calculated from the available parking spaces in the power grid topology map.

[0224] SOC constraints are represented using the following logic:

[0225]

[0226] In the formula, express The charging status of an electric vehicle at any given time. , They represent Minimum charging time for electric vehicles ,maximum .

[0227] SOH constraints are represented using the following logic:

[0228]

[0229] In the formula, express The current health status of electric vehicles being charged. , They represent Minimum charging time for electric vehicles ,maximum .

[0230] The user-side unit is used to achieve a dynamic game equilibrium between individual charging demand and group electricity consumption behavior by adding a charging cost-time game model. Specifically, it designs a user charging cost objective function, adds maximum and minimum electricity price constraints and user satisfaction constraints, and uses the following logic to represent the user charging cost objective function:

[0231]

[0232] In the formula, Indicates the charging cost target. express The time-of-use electricity price at any given moment is based on real-time electricity prices from the power grid topology map. express The instantaneous charging power of electric vehicles at any given time. This indicates the time when the electric vehicle begins charging. This indicates the time when the electric vehicle finishes charging.

[0233] In this embodiment, the user-side constraints include:

[0234] The maximum and minimum constraints on electricity prices are represented using the following logic:

[0235]

[0236] In the formula, express The current electricity price for charging electric vehicles. , They represent The minimum and maximum electricity prices for charging electric vehicles at any given time are obtained from the real-time electricity prices in the power grid topology map.

[0237] User satisfaction constraints are represented logically as follows:

[0238]

[0239] In the formula, express Current user satisfaction , They represent The minimum and maximum user satisfaction at any given time are determined by data derived from user behavior profiles in the vehicle feature map.

[0240] The road network side unit is used to design the objective function for optimizing the electric vehicle path, specifically:

[0241]

[0242] in, The goal is to optimize the path for electric vehicles. The path time for charging electric vehicles is derived from path time data in the road network resource map.

[0243] The multi-objective optimization unit is used to construct a multi-objective optimization function based on the above objective function, and is represented by the following logic:

[0244]

[0245] In the formula, This indicates the total charging cost for the electric vehicle over the entire charging period. This indicates the total charging time for the electric vehicle. This represents the amount of new energy consumed during the period from the start to the end of a multi-objective optimization function. This represents the predicted power generation from new energy sources during the period from the start to the end of a multi-objective optimization function. This represents the entire optimization cycle of the multi-objective optimization function. Indicates the renewable energy consumption rate. , express , and Solve the following problem while satisfying the constraints S31 to S32 respectively. , and All of them reach the minimum global optimal solution set.

[0246] The particle swarm optimization unit is used to optimize multi-objective optimization functions using hybrid adaptive particle swarm optimization.

[0247] This embodiment employs Hybrid Adaptive Particle Swarm Optimization (HAPSO) to solve the aforementioned optimization algorithm, specifically including:

[0248] First, a particle swarm is generated, with each particle representing a charging power allocation scheme; the charging power allocation scheme is specifically the charging pile and charging time scheme selected by the electric vehicle.

[0249] Secondly, dynamic weight adjustment is performed; constraint processing is carried out according to the constraints from the grid-side unit to the user-side unit. The constraint processing specifically involves using a repair strategy to handle out-of-bounds particles, such as adjusting the charging power to the boundary value to ensure that the charging power does not exceed the maximum or minimum charging power, and constraints on charging pile parking spaces, etc.

[0250] The multi-objective fitness of each particle is calculated, and the non-dominated optimal solution set is screened based on the Pareto front. Specifically, the multi-objective optimization function is used to calculate the target fitness of each particle on the road network side, power grid side, and user side respectively, and the optimal solution set is screened out.

[0251] The Pareto front is one of the methods used in multi-objective optimization to describe the set of optimal solutions. Since multi-objective optimization problems often have multiple objectives, and these objectives may conflict, the Pareto optimal solution set represents the solutions that can be further optimized without sacrificing other objectives.

[0252] In this embodiment, the multi-objective fitness of each particle is calculated. Specifically, a multi-objective optimization function is used to evaluate the target fitness of each particle's solution on the road network side, power grid side, and user side, respectively. By comparing the multi-objective fitness of different particles, the optimal solution set is selected.

[0253] User preferences are adjusted based on user satisfaction, specifically by weighting the Pareto solution set according to satisfaction and outputting the final optimized solution.

[0254] In this embodiment, considering that users may have different preferences for certain objectives, such as charging time, real-time electricity price, or location preference, this embodiment performs a satisfaction-weighted adjustment on the fitness of each solution in the Pareto solution set, so that the priority of different objectives will be adjusted according to user preferences, in order to generate an optimal charging scheme that better meets user preferences.

[0255] The strategy generation module constructs a strategy generation layer based on a vehicle-road-network coupled three-domain knowledge graph, and pushes optimized solutions to the user side, power grid side, and road network side. Specifically, the strategy generation module includes the following units:

[0256] The push unit is used to establish a minute-level rolling real-time optimization mechanism on the user side, based on the aforementioned final optimization scheme, and to push the user-side path selection to the electric vehicle terminal in real time. Specifically:

[0257] Based on the location data in the vehicle feature knowledge graph, the real-time location information of the vehicle is updated every minute. Since different vehicles have different locations, the paths to the surrounding charging stations are also different. Based on the real-time location information of the vehicle, the final optimization scheme output by the multi-objective optimization module (specifically the path from the current vehicle location to the charging station) is pushed to the user in real time to ensure the effectiveness of the optimization results.

[0258] The power grid hotspot map unit is used on the power grid side to generate a real-time response curve for electric vehicle charging prices based on real-time electricity prices in the power grid topology map, and to obtain line congestion and renewable energy consumption hotspot maps based on power flow information and renewable energy power in the power grid topology map, for reference by power grid operation and management personnel.

[0259] The road network hotspot map unit is used to generate electric vehicle path flow maps based on traffic flow in the road network knowledge graph and to obtain charging pile parking space hotspot maps for reference based on the real-time status of charging piles in the power grid topology graph.

[0260] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for real-time scheduling and optimization of electric vehicle charging load grid operation, characterized in that, Includes the following steps: S1. Construct a vehicle-road-network coupled three-domain knowledge graph fusion architecture, including vehicle feature graph, road network resource graph and power grid topology graph, and store knowledge graph data through a graph database and update it in real time; S2, a dynamic perception layer is constructed based on a vehicle-road-network coupled three-domain knowledge graph, including: S21, construct a dynamic perception of electric vehicle charging and discharging behavior to predict the spatiotemporal distribution of electric vehicle charging demand; specifically, construct a dynamic perception model of user charging demand, update the vehicle feature map in real time through the vehicle terminal, obtain the historical SOC and SOH data of electric vehicles on the vehicle itself, use the bidirectional long short-term memory network to predict the charging and discharging power of electric vehicles and judge the charging and discharging behavior of electric vehicles, and couple the user behavior profile of the vehicle itself. S22, Construct a charging pile load perception system to predict the charging and discharging power of the charging pile; Specifically, take the charging pile itself as a unit, obtain historical charging load data of the charging pile, and after data preprocessing, use a bidirectional long short-term memory network to predict the charging and discharging power of the charging pile. S23, dynamically sensing the charging and discharging behavior of electric vehicles and sensing the load of charging piles are coupled bidirectionally to improve the accuracy of dynamically sensing the charging behavior of electric vehicles; the bidirectional coupling specifically refers to: The charging and discharging power of each electric vehicle is aggregated to obtain the aggregated charging and discharging power of the charging pile at this moment. The aggregated charging power of charging piles can reflect vehicle preference information, and the charging and discharging power prediction data of charging piles reflects the characteristic data of the power grid side. The two information are coupled to improve the accuracy of dynamic perception of electric vehicle charging behavior based on multi-temporal and spatial load characteristics. S3 constructs a multi-objective optimization layer based on a vehicle-road-network coupled three-domain knowledge graph, builds a multi-objective optimization function, and uses a hybrid adaptive particle swarm optimization to optimize the multi-objective optimization function, outputting the final optimization scheme; S4 constructs a strategy generation layer based on a vehicle-road-network coupled three-domain knowledge graph, and pushes the optimization scheme to the user side, power grid side and road network side.

2. The method for real-time scheduling and optimization of electric vehicle charging load grid operation according to claim 1, characterized in that, The entity of the vehicle feature map described in S1 is the vehicle itself. The categories of the vehicle feature map include basic vehicle information, user behavior profiles, and real-time data from the vehicle BMS. The attributes of the basic vehicle information include brand, model, fuel type, and rated power. The attributes of the user behavior profile include location preference, time preference, and bounded rationality threshold. The attributes of real-time data from the vehicle's BMS include charging status, health status, charging rate, and location data. The entities in the power grid topology map are electric vehicle charging piles. The categories of the power grid topology map include basic power grid information, power grid information, and real-time status. The attributes of the basic power grid information include user number, capacity, number of parking spaces, charging type, address, and energy consumption category. The attributes of the power grid information include grid-connected lines and grid-connected substations. The attributes of the real-time status include available parking spaces, power margin, and real-time electricity price. The road network resource map is categorized into traffic flow and dynamic road congestion; the attributes of traffic flow include path time and congestion level, while the attributes of dynamic road congestion include weather conditions and parking space prediction.

3. The method for real-time scheduling and optimization of electric vehicle charging load grid operation according to claim 1, characterized in that, The user charging demand dynamic perception model described in S21 is specifically an edge computing model based on incremental updates. First, set incremental update trigger conditions. When the vehicle's charging status changes by more than 1% or more, its health status changes by more than 1% or more, and the time interval is more than or equal to 1 minute, the vehicle feature map is incrementally updated. Only the changed data is updated, and the unchanged data is not uploaded and updated. Secondly, by adopting an edge computing model, abnormal data changes are processed only on the vehicle terminal side, filtering out instantaneous SOC and SOH changes.

4. The method for real-time scheduling and optimization of electric vehicle charging load grid operation according to claim 3, characterized in that, S21 describes the use of a bidirectional long short-term memory network to predict the charging and discharging power of electric vehicles and determine their charging and discharging behavior, coupled with the vehicle's user behavior profile. First, the Long Short-Term Memory Network outputs the predicted charging and discharging power of the vehicle based on the historical SOC and SOH data of the individual vehicle, and analyzes the vehicle's SOC. When the SOC data increases, it is considered that the vehicle is charging, and vice versa. Secondly, based on the predicted vehicle charging and discharging power and the determined charging and discharging behavior, user location preference and time preference data in the vehicle feature map are coupled to obtain the real-time location and time preference results for electric vehicle charging and discharging. In step S21, the Long Short-Term Memory Network prediction model is trained through a combination of incremental training and full training. Specifically, the prediction model is first trained incrementally, and the prediction model is dynamically iterated and fine-tuned using historical SOC and SOH data from the previous 7 days to continuously update and optimize the prediction results. The average prediction error of the prediction model is calculated, and when the average prediction error calculated for 3 consecutive days exceeds a preset threshold, the prediction model is trained in full.

5. The method for real-time scheduling and optimization of electric vehicle charging load grid operation according to claim 4, characterized in that, The bidirectional coupling described in S23 specifically refers to: By aggregating the charging and discharging power of each electric vehicle, the aggregated charging and discharging power of the charging pile at this moment is obtained: In the formula, express The charging power of charging piles aggregated at any given time. This indicates the number of electric vehicles aggregated. express At this moment The charging power of an electric vehicle ; The calculation is based on a two-way coupling of user preference coefficients and historical feature coefficients, using the following logical representation: In the formula, Indicates bidirectional coupling Current charging power of charging stations Represents the user preference coefficient. Represents historical characteristic coefficients. express At any given time, a bidirectional long short-term memory network is used to predict the charging and discharging power of charging piles; among which, the user preference coefficient and historical characteristic coefficient First, solve the system of equations based on the data from the first two historical moments. Then, continuously update the system dynamically, using the following logic: In the formula, , These represent the results after bidirectional coupling. time, The charging power of the charging station at any given time. , They represent time, The charging and discharging power of the charging pile is predicted in real time using a bidirectional long short-term memory network. , They represent time, The charging power of charging piles at any given moment.

6. The method for real-time scheduling and optimization of electric vehicle charging load grid operation according to claim 1, characterized in that, S3 includes the following steps: S31, on the grid side, design the objective function for renewable energy absorption rate, and add constraints such as line congestion peak shaving, power balance, charging pile capacity, charging pile parking space, SOC, and SOH. S32, on the user side, design the user charging cost objective function, and add the maximum and minimum constraints of electricity price and user satisfaction constraints; S33, on the road network side, design the objective function for optimizing the electric vehicle path; S34. Based on the above objective function, construct a multi-objective optimization function, using the following logical representation: In the formula, Indicates the charging cost target. express Time-of-use electricity pricing at any given moment express The instantaneous charging power of electric vehicles at any given time. This indicates the total charging cost for the electric vehicle over the entire charging period. This indicates the time when the electric vehicle begins charging. Indicates the time when the electric vehicle finishes charging. This indicates the total charging time for the electric vehicle; This indicates the target for the rate of new energy consumption. P represents the amount of new energy consumed. EV This indicates the projected power generation from new energy sources. This represents the amount of new energy consumed during the period from the start to the end of a multi-objective optimization function. This represents the predicted power generation from new energy sources during the period from the start to the end of a multi-objective optimization function. This represents the entire optimization cycle of the multi-objective optimization function. Indicates the renewable energy consumption rate. , The goal is to optimize the path for electric vehicles. Path time for charging electric vehicles express , and Solve the following problem while satisfying the constraints S31 to S32 respectively. , and All achieve the smallest global optimal solution set; S35 employs a hybrid adaptive particle swarm optimization multi-objective optimization function.

7. The method for real-time scheduling and optimization of electric vehicle charging load grid operation according to claim 6, characterized in that, The specific implementation of the hybrid adaptive particle swarm optimization multi-objective optimization function described in S35 is as follows: S351, First, generate a particle swarm, where each particle represents a charging power allocation scheme; Secondly, dynamic weight adjustment is performed; constraint processing is carried out according to the constraints of S31~S32, and the constraint processing specifically involves using a repair strategy to handle out-of-bounds particles; S352, calculate the multi-objective fitness of each particle, and screen the non-dominated optimal solution set based on the Pareto front; S353, adjust user preferences based on user satisfaction, specifically by performing a satisfaction-weighted sorting of the Pareto solution set and outputting the final optimized solution.

8. The method for real-time scheduling and optimization of electric vehicle charging load grid operation according to claim 1, characterized in that, Specifically, S4 is: S41, On the user side, based on the above final optimization scheme, a minute-level rolling real-time optimization mechanism is established to push the user-side path selection to the electric vehicle terminal in real time; S42, on the grid side, a real-time response curve for electric vehicle charging price is generated based on the real-time electricity price in the grid topology map, and line congestion and hot spot maps for renewable energy consumption are obtained based on the power flow information and renewable energy power in the grid topology map. S43, on the road network side, generates an electric vehicle path flow map based on traffic flow in the road network knowledge graph, and obtains a charging pile parking space hotspot map based on the real-time status of charging piles in the power grid topology graph.

9. A real-time scheduling and optimization system for electric vehicle charging load power grid operation, characterized in that, include: The data storage module is used to construct a vehicle-road-network coupled three-domain knowledge graph fusion architecture, including vehicle feature graph, road network resource graph and power grid topology graph. The knowledge graph data is stored and updated in real time through a graph database. The dynamic perception module, based on a vehicle-road-network coupled three-domain knowledge graph, constructs a dynamic perception layer, including: To construct a dynamic perception model for electric vehicle charging and discharging behavior, which is used to predict the spatiotemporal distribution of electric vehicle charging demand; specifically, to construct a dynamic perception model for user charging demand, and to dynamically update the vehicle feature map in real time through the vehicle terminal, to obtain historical SOC and SOH data of electric vehicles on the basis of the vehicle itself, and to use a bidirectional long short-term memory network to predict the charging and discharging power of electric vehicles and judge the charging and discharging behavior of electric vehicles, and to couple the user behavior profile of the vehicle itself. A charging pile load perception system is constructed to predict the charging and discharging power of the charging pile. Specifically, the historical charging load data of the charging pile is obtained on a per-unit basis. After data preprocessing, a bidirectional long short-term memory network is used to predict the charging and discharging power of the charging pile. The dynamic sensing of electric vehicle charging and discharging behavior is bidirectionally coupled with the load sensing of charging piles to improve the accuracy of dynamic sensing of electric vehicle charging behavior; the bidirectional coupling specifically involves: The charging and discharging power of each electric vehicle is aggregated to obtain the aggregated charging and discharging power of the charging pile at this moment. The aggregated charging power of charging piles can reflect vehicle preference information, and the charging and discharging power prediction data of charging piles reflects the characteristic data of the power grid side. The two information are coupled to improve the accuracy of dynamic perception of electric vehicle charging behavior based on multi-temporal and spatial load characteristics. The multi-objective optimization module constructs a multi-objective optimization layer based on a vehicle-road-network coupled three-domain knowledge graph, builds a multi-objective optimization function, and uses a hybrid adaptive particle swarm optimization to optimize the multi-objective optimization function, outputting the final optimization scheme. The strategy generation module constructs a strategy generation layer based on a vehicle-road-network coupled three-domain knowledge graph, and pushes the optimization scheme to the user side, power grid side and road network side.

Images