Urban charging stopping facility dispatching and power grid demand side control system for over-charging station

Abstract

The invention provides an overcharge station-oriented urban charging stopping facility scheduling and power grid demand side control system, and the system comprises a data sensing module which collects and preprocesses user charging preferences, vehicle battery states, overcharge station equipment states and power grid side data, extracts vehicle features, and classifies the vehicle features; the scheduling module generates a power distribution scheme through a dynamic priority strategy based on the classification result and the equipment state, and performs collaborative optimization on all the overcharge stations by using a collaborative model to generate a collaborative scheduling scheme; and the power grid demand module aggregates the states of all the overcharge stations to construct a virtual resource pool, analyzes power grid side regulation and control demands, and generates global regulation and control instructions issued to all the stations in combination with the collaborative scheduling scheme. According to the invention, intelligent scheduling and power grid demand side response of the overcharge station cluster are realized, and the charging service efficiency, the power grid operation stability and the resource utilization level are improved.

Description

Technical Field

[0001] This invention relates to the technical field of smart grid, coordinated scheduling of electric vehicle charging facilities and grid demand-side response, specifically to urban outage facility scheduling and grid demand-side control system for supercharging stations. Background Technology

[0002] With the rapid popularization of electric vehicles, urban supercharging stations, as important energy replenishment nodes, are facing increasingly prominent issues regarding large-scale operation and grid coordination. Existing charging facility scheduling is mostly limited to single-station optimization or simple queuing strategies, lacking comprehensive consideration of multi-station coordination, differences in user behavior, and the real-time status of the power grid. Uneven spatial and temporal distribution of charging loads can easily lead to congestion at some stations and low equipment utilization. Meanwhile, disorderly charging may exacerbate the peak-valley difference in the power grid, affecting regional power supply security and power quality. Furthermore, current systems generally lack an efficient linkage mechanism with the grid's demand-side response, making it difficult to effectively participate in grid peak shaving and alleviate local overload while ensuring a good user charging experience. This limits the regulatory potential of charging infrastructure as a flexible load resource.

[0003] On the other hand, the grid's capacity to accommodate and regulate large-scale, high-power charging loads remains insufficient. Existing dispatching models often passively adapt to charging demand, lacking proactive guidance and collaborative optimization capabilities, making it difficult to achieve dynamic matching between charging load, renewable energy output, and grid operating status. Simultaneously, the potential of electric vehicles with bidirectional charging and discharging capabilities as distributed energy storage resources has not been fully explored, and vehicle-to-grid (V2G) applications lack large-scale, safe, and controllable integrated dispatching solutions. Therefore, there is an urgent need to construct an intelligent dispatching and collaborative control system that integrates user needs, site status, and grid control objectives to improve charging service efficiency, enhance grid operational resilience, and promote the optimal allocation of energy resources. Summary of the Invention

[0004] To address the technical problems mentioned in the background section, this invention proposes a city-wide dispatching and grid demand-side control system for supercharging stations.

[0005] Therefore, the technical solution adopted by the present invention is as follows:

[0006] A city-wide dispatching and grid demand-side control system for supercharging stations, including:

[0007] Data perception module: Collects a dataset including user charging preference data, vehicle battery status data, supercharging station equipment status data, and grid-side data, and preprocesses the dataset; obtains vehicle features based on the charging preference data and battery status data, and classifies the vehicle based on a preset threshold and the vehicle features to obtain classification results;

[0008] Scheduling module: Based on the vehicle characteristics, classification results, and equipment status data, it generates a power allocation scheme through a dynamic priority strategy; based on the power allocation scheme, it performs collaborative analysis on all supercharging stations through a preset collaborative model to generate a collaborative scheduling scheme;

[0009] Grid demand module: Aggregates the current status data of all supercharging stations to construct a virtual resource pool; parses the grid-side data into specific control requirements; and generates and issues global control instructions to each supercharging station based on the collaborative scheduling scheme, control requirements, and the real-time status of the virtual resource pool.

[0010] Furthermore, the vehicle characteristics include charging urgency and dispatchability index;

[0011] The formula for calculating the charging urgency is as follows:

[0012]

[0013] in, Due to the urgency of charging, For the target state of charge, The current state of charge, For the battery's rated capacity, The maximum charging power of the battery. For battery health, For user urgency level, For charging efficiency;

[0014] The calculation steps for the scheduling index are as follows:

[0015] First step, calculate the charging time:

[0016]

[0017] in, This refers to charging time;

[0018] The second step is to determine the charging start time based on the charging time.

[0019] The third step is to calculate the scheduling time window based on the charging start time.

[0020] The fourth step is to normalize the scheduling time window into the scheduling index.

[0021] Furthermore, the classification results include fast charging and fast running type, fast charging and slow running type, slow charging and fast running type, and slow charging and slow running type;

[0022] If the slow-charging, slow-moving vehicle has bidirectional charging and discharging capabilities, it is marked as a candidate resource.

[0023] Furthermore, the power allocation scheme is generated as follows:

[0024] The first step is to assign initial priorities to different types of vehicles based on the classification results;

[0025] The second step is to calculate the dynamic priority of the vehicle based on the initial priority, using the following formula:

[0026]

[0027] in, For dynamic priority, As the initial priority, Due to time constraints, As for the power urgency factor, As a guiding factor for electricity prices, , and These are the weighting coefficients;

[0028] The third step involves sorting the vehicles based on the dynamic priority and applying different power allocation schemes to different types of vehicles. The specific operations are as follows:

[0029] For fast-charging and fast-moving vehicles, the charging power is the minimum value among the charging pile's rated power, the vehicle battery's maximum acceptable charging power, and the vehicle battery's maximum acceptable charging power after battery health status correction.

[0030] For vehicles that charge quickly but travel slowly, the charging power calculation formula is as follows:

[0031]

[0032] in, This refers to the charging power.

[0033] For vehicles that charge slowly and travel quickly, the charging power should be set to 60% to 80% of the charging pile's rated power.

[0034] For vehicles that charge slowly and travel slowly, the charging power should be set to 40% to 60% of the rated power of the charging station.

[0035] For vehicles marked as candidate resources, the supercharging station is discharged with the owner's authorization.

[0036] Furthermore, the collaborative model employs a distributed optimization framework based on multi-agent deep reinforcement learning;

[0037] The collaborative model models each supercharging station as an intelligent agent, and each intelligent agent includes a state space and an action space.

[0038] The reward function of the agent adopts a two-layer structure of individual reward and collective punishment;

[0039] The individual rewards include operational revenue and user satisfaction rewards;

[0040] The operational revenue reward function is as follows:

[0041]

[0042] in, As a reward for operating revenue, This represents the real-time charging power of the i-th charging pile at the supercharging station. For the time interval of revenue statistics, The unit price for charging at supercharging stations, The total power drawn by the supercharging station from the power grid. This refers to the real-time electricity price on the power grid.

[0043] The user satisfaction reward is quantified into a specific reward value based on average waiting time and charging completion rate.

[0044] The collective penalties include load volatility penalties and overload risk penalties;

[0045] The load volatility penalty function is as follows:

[0046]

[0047] in, The penalty value is the penalty for load volatility. The weighting coefficient for the load volatility penalty. The standard deviation of the total load of all supercharging stations in the region within a set time window. This represents the total load of all supercharging stations within the region.

[0048] The overload risk penalty function is as follows:

[0049]

[0050] in, The penalty value is the overload risk penalty. The weighting coefficient for overload risk penalty. The risk value for vulnerable nodes in the power grid. This is a preset risk threshold.

[0051] Furthermore, the vulnerable nodes are identified based on grid-side data;

[0052] The formula for calculating the risk value of the vulnerable node is as follows:

[0053]

[0054] in, This is a weighting factor for the node load rate. For the real-time load of vulnerable nodes, The rated load for vulnerable nodes, The health status of devices on vulnerable nodes. This is a weighting coefficient for the health status of the equipment. The weighting coefficient for the load growth trend. This represents the load increase rate of vulnerable nodes.

[0055] Furthermore, the current status data includes adjustable load resources and candidate resources;

[0056] The power grid-side data is parsed using conventional rule engines and data parsing techniques;

[0057] The global control command is generated using a weighted allocation algorithm;

[0058] The global control command includes execution information, control tasks, constraints and safety boundaries, feedback requirements, and auxiliary reference information.

[0059] Compared with the prior art, the advantages of the present invention are as follows:

[0060] 1. This invention integrates user charging preferences, vehicle battery status, supercharging station operation status, and real-time grid data to construct an integrated collaborative scheduling system for vehicles, stations, and the grid. This overcomes the limitations of traditional single-station scheduling or simple queuing strategies and effectively improves charging service efficiency and grid operation coordination.

[0061] 2. This invention classifies vehicles by characteristics and combines them with a dynamic priority strategy for power allocation. This allows for flexible response to different user needs and grid conditions, enabling refined load control and thus improving the overall utilization rate of charging facilities and energy allocation efficiency.

[0062] 3. By aggregating supercharging station clusters to build a virtual resource pool and integrating grid-side control commands to generate global optimization strategies, the system can proactively participate in grid peak shaving, alleviate local overloads, support orderly vehicle-grid interaction and access, and improve the grid's ability to accept high-power charging loads and its overall stability. Attached Figure Description

[0063] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0064] Figure 1 This is a schematic diagram of the system flow of the present invention;

[0065] Figure 2 This is a schematic diagram of the scheduling index calculation process of the present invention;

[0066] Figure 3 This is a schematic diagram of the classification results of the present invention. Detailed Implementation

[0067] To achieve the above objectives, this invention provides a city-wide dispatching and grid demand-side control system for supercharging stations. Please refer to [link to relevant documentation]. Figure 1 ,include:

[0068] Data perception module: Collects a dataset including user charging preference data, vehicle battery status data, supercharging station equipment status data, and grid-side data, and preprocesses the dataset; obtains vehicle features based on the charging preference data and battery status data, and classifies the vehicle based on a preset threshold and the vehicle features to obtain classification results;

[0069] The process of collecting charging preference data is as follows:

[0070] The system collects users' charging preference data through mobile applications or vehicle-to-everything (V2X) communication interfaces, including expected departure time, urgency of power demand, and preferred charging power. Users can actively input the above information, or the system can make intelligent predictions based on historical behavior data. For prediction scenarios, the system adopts a time-series prediction model based on long short-term memory networks, inputting features such as users' historical charging records, arrival time, and dwell time, and outputting the probability distribution of expected departure time.

[0071] Battery status data includes parameters such as current state of charge (SOC), battery capacity, maximum acceptable charging power, and state of health (SOH). This data is acquired through the on-board diagnostic system (OBD) interface or the vehicle-to-infrastructure (V2I) communication protocol, with a sampling frequency set to once per second to ensure real-time data accuracy.

[0072] Equipment status data includes real-time power output, availability, fault information, and upper and lower limits of charging and discharging power for each charging station.

[0073] The grid-side data includes real-time electricity price signals, grid dispatch instructions, load warning information, distribution network topology data, real-time load data of key nodes, line capacity limits, transformer operating status, etc. These data are obtained through the data interface with the grid dispatch system, and the update frequency is set to 1 minute to 15 minutes depending on the data type.

[0074] Preprocessing includes data cleaning, outlier detection and handling, missing value imputation, and data alignment and fusion;

[0075] Due to the heterogeneous data sources and asynchronous timestamps, this module maps multi-source data to the same time base and performs synchronization alignment. At the same time, it associates and merges discrete information belonging to the same vehicle or charging pile to generate standardized data records with complete descriptions, providing support for integrated analysis.

[0076] Vehicle characteristics include charging urgency and dispatchability index;

[0077] Charging urgency reflects a vehicle's time sensitivity to charging services and the intensity of its power demand. The calculation formula is as follows:

[0078]

[0079] in, Due to the urgency of charging, The target state of charge is derived from charging preference data, either input via a mobile application or predicted by the system based on historical data. The current state of charge, For the battery's rated capacity, The maximum charging power of the battery. For battery health, For user urgency level, To determine charging efficiency, parameters can be obtained from the equipment nameplate, through on-site measurements, or by adopting typical empirical values.

[0080] In this embodiment, the user urgency coefficient ranges from [0.3, 3.0], and is confirmed through direct acquisition and indirect inference. For example, the direct acquisition operation is as follows: on the charging reservation or start interface, provide options such as leave immediately after charging, normal charging, and long-term parking, and map them to different user urgency coefficients, such as 2.0, 1.0, and 0.5; the indirect inference operation is as follows:

[0081] Taxi, ride-hailing, and logistics vehicles are typically time-sensitive, and by default, their users' urgency level is higher than that of private cars.

[0082] The calculation steps for the scheduling index are as follows; please refer to [link / reference]. Figure 2 :

[0083] First step, calculate the charging time:

[0084]

[0085] in, The charging time is (0.5 + 0.5 × SOH), which is the correction factor for the charging power due to battery aging, reflecting the impact of battery health status on the actual acceptable charging power.

[0086] The second step is to determine the charging start time based on the charging time:

[0087]

[0088] in, The moment charging begins. For the expected departure time,

[0089] The third step is to calculate the scheduling time window based on the charging start time:

[0090]

[0091] in, For scheduling time windows, This refers to the time when the vehicle arrives at the supercharging station.

[0092] The fourth step is to normalize the scheduling time window into a scheduling index:

[0093]

[0094] in, The scheduling index ranges from 0 to 1. A scheduling index close to 1 indicates that the charging time window is very wide and the vehicle has a high degree of scheduling capability; a scheduling index close to 0 indicates that the time window is very tight and the scheduling capability is extremely low.

[0095] Thresholds include urgency thresholds and scheduling thresholds;

[0096] The classification results include fast charging and fast running, fast charging and slow running, slow charging and fast running, and slow charging and slow running. Please refer to [link / reference]. Figure 3 The specific classification process is as follows:

[0097] If the charging urgency is greater than or equal to the urgency threshold and the scheduling index is less than the scheduling threshold, the vehicle is marked as a fast-charging, fast-moving vehicle. These vehicles have extremely urgent charging needs and a narrow time window, and charging must be arranged immediately and the maximum power allocated.

[0098] If the charging urgency is greater than or equal to the urgency threshold and the dispatchability index is greater than or equal to the dispatchability threshold, the vehicle is marked as a fast-charging, slow-travel type. Although these vehicles have urgent charging needs, they have a certain degree of time flexibility. Under the premise of ensuring timely completion of charging, the charging time can be appropriately adjusted to cooperate with the grid load and electricity price optimization.

[0099] If the charging urgency is less than the urgency threshold and the scheduling index is less than the scheduling threshold, the vehicle is marked as a slow-charging-fast-moving vehicle. These vehicles do not have an urgent charging need but have a short dwell time and need to complete charging within a limited time window.

[0100] If the charging urgency is less than the urgency threshold and the dispatchability index is greater than or equal to the dispatchability threshold, the vehicle is marked as a slow-charging, slow-moving vehicle. These vehicles have no urgent charging needs and have ample dwell time, and have the highest dispatchability. They can be used as a system adjustment resource to participate in load balancing and demand response.

[0101] For vehicles that charge slowly and travel slowly, if they have bidirectional charging and discharging capabilities, their current state of charge is higher than the minimum reserve capacity set by the user, and the available scheduling time window is wide enough, they are marked as candidate resources that can participate in vehicle-to-grid interaction, and their discharge capacity is calculated.

[0102] In this embodiment, the urgency threshold and the scheduling threshold are 0.8 and 0.3, respectively;

[0103] The urgency threshold is set based on the following criteria:

[0104] The value of 0.8 can identify vehicles that truly need priority service, while also leaving the system with a certain margin for power adjustment. It also conforms to the engineering practice that electrical equipment should not be operated at full load for a long time.

[0105] The scheduling threshold is set based on the following criteria:

[0106] When the dispatch threshold reaches 0.3 or above, the system has sufficient time leeway to flexibly arrange charging periods based on factors such as the grid load curve and real-time electricity price, achieving goals such as peak shaving and valley filling or cost optimization. In actual operation, it can effectively distinguish between vehicles that must be charged immediately and those that can be flexibly dispatched.

[0107] Scheduling module: Based on the vehicle characteristics, classification results, and equipment status data, it generates a power allocation scheme through a dynamic priority strategy; based on the power allocation scheme, it performs collaborative analysis on all supercharging stations through a preset collaborative model to generate a collaborative scheduling scheme;

[0108] The power allocation scheme is generated as follows:

[0109] The first step is to assign initial priorities to different types of vehicles based on the classification results;

[0110] In this embodiment, for fast-charging, fast-travel vehicles, the initial priority is set to the highest value of 1.0 to ensure that these vehicles can obtain charging services immediately; for fast-charging, slow-travel vehicles, the initial priority is set to 0.8 to ensure their charging needs while reserving space for scheduling optimization; for slow-charging, fast-travel vehicles, the initial priority is set to 0.6, so that charging can be completed within the time window; and for slow-charging, slow-travel vehicles, the initial priority is set to 0.4, serving as the main resource for system load adjustment.

[0111] The second step is to calculate the dynamic priority of the vehicle based on the initial priority. The calculation formula is as follows:

[0112]

[0113] in, For dynamic priority, As the initial priority, Due to time constraints, As for the power urgency factor, As a guiding factor for electricity prices, , and These are weighting coefficients, which are the optimal configurations derived from actual operational experience and data verification. In this embodiment, they are 0.3, 0.4, and 0.2, respectively.

[0114] The formula for calculating the time urgency factor is as follows:

[0115]

[0116] in, For the current moment, The charging start time is obtained by subtracting the charging time from the expected departure time;

[0117] The formula for calculating the power urgency factor is as follows:

[0118]

[0119] The electricity price guidance factor takes a positive value when the real-time electricity price is low to encourage charging, and a negative value when the electricity price is high to suppress non-emergency charging. The specific value is determined based on the degree of deviation between the current electricity price and the average daily electricity price.

[0120] The third step is to sort the vehicles based on dynamic priority and apply different power allocation schemes to different types of vehicles. The specific operations are as follows:

[0121] For fast-charging and fast-moving vehicles, the maximum charging power is allocated. The maximum charging power is the minimum value among the rated power of the charging pile, the maximum acceptable charging power of the vehicle battery, and the maximum acceptable charging power of the vehicle battery after the battery health status is corrected.

[0122] For vehicles that charge quickly but travel slowly, the charging power calculation formula is as follows:

[0123]

[0124] in, This refers to the charging power.

[0125] For vehicles that charge slowly and travel quickly, a constant power charging strategy is adopted, with the charging power being 60% to 80% of the rated power of the charging pile. The specific value is dynamically determined based on the current total load in the station and the grid load status. The allocated power can be appropriately increased during the grid off-peak hours and reduced during peak hours.

[0126] For vehicles that charge slowly and travel slowly, the charging period should be scheduled during periods of low electricity prices or low grid load, and the charging power should be set to 40% to 60% of the rated power of the charging pile. The specific value should be dynamically determined based on the current energy storage charge status, photovoltaic power generation, and grid load forecast.

[0127] For vehicles marked as candidate resources, the supercharging station will discharge the battery with the owner's authorization. The discharge power is calculated using the following formula:

[0128]

[0129] in, For discharge power, For maximum discharge power, This refers to the vehicle's current state of charge, i.e., its current electrical charge level. The minimum retained charge state is defined as the minimum retained charge. This refers to the total capacity of the vehicle's power battery. The discharge time is limited by the user's expected departure time and the duration of grid demand response.

[0130] During the power allocation process, the system monitors the total power constraint within the station in real time to ensure that the total power allocation of all charging piles does not exceed the capacity limit of the power supply system within the station.

[0131] When power demand exceeds supply capacity, the system initiates a power reduction procedure, which follows these principles:

[0132] First, reduce the charging power of slow-charging, slow-travel vehicles to 20% to 40% of their original power, or even completely stop charging.

[0133] Secondly, the charging power of vehicles that charge slowly and travel quickly will be reduced to 50% to 70% of their original power.

[0134] Secondly, the charging power of fast-charging, slow-travel vehicles will be reduced to 70% to 85% of the original power;

[0135] Finally, the charging power of fast-charging vehicles will be reduced to 90% to 95% of the original power, but it is necessary to ensure that they can still be fully charged before leaving the site.

[0136] The collaborative model adopts a distributed optimization framework based on multi-agent deep reinforcement learning;

[0137] The collaborative model models each supercharging station in the region as an intelligent agent, and each intelligent agent includes a state space and an action space; in this embodiment, the region refers to the power supply area of ​​a distribution network.

[0138] The state space includes the current load of this station, information on the queue of vehicles waiting to be charged, the priority and charging demand of each vehicle, and the load status of adjacent stations.

[0139] The action space includes in-station power regulation and user guidance scheduling, etc.

[0140] The agent's reward function adopts a two-layer structure of individual reward and collective punishment;

[0141] Individual rewards include operational revenue and user satisfaction rewards;

[0142] The operating revenue reward function is as follows:

[0143]

[0144] in, As a reward for operating revenue, This represents the real-time charging power of the i-th charging pile at the supercharging station. The system has preset parameters for the time interval of revenue statistics, which can be set according to operational statistics needs and data update frequency, for example, 15 minutes. The unit price for charging at supercharging stations, The total power drawn by the supercharging station from the power grid. This refers to the real-time electricity price on the power grid.

[0145] User satisfaction rewards are quantified by average waiting time and charging completion rate. The shorter the waiting time and the higher the proportion of vehicles that complete charging on time, the greater the reward value, with a range of [0,1].

[0146] Collective penalties include load volatility penalties and overload risk penalties;

[0147] The load volatility penalty function is as follows:

[0148]

[0149] in, The penalty value is the penalty for load volatility. The weighting coefficient for load volatility penalty is determined based on regional power grid stability requirements and operational data verification, and its value ranges from [0.1, 0.6]. The standard deviation of the total load of all supercharging stations in the region within a set time window. This represents the total load of all supercharging stations within the region.

[0150] The overload risk penalty function is as follows:

[0151]

[0152] in, The penalty value is the overload risk penalty. The weighting coefficient for overload risk penalty is set based on the load-bearing capacity of power grid equipment and the importance of vulnerable nodes, with a value range of [0.3, 1.0]. Risk values ​​for vulnerable nodes in the power grid. The preset risk threshold is set based on power grid equipment design standards and distribution network operation specifications.

[0153] Vulnerable nodes in the power grid are identified based on power grid-side data. The specific process is as follows:

[0154] The system obtains core data such as the distribution network topology, real-time load of key nodes, line capacity limits, and transformer operating status through the power grid side data interface. It uses industry-standard indicators such as node load rate, voltage stability margin, line power flow congestion coefficient, and equipment aging degree as the core evaluation criteria. Each node is weighted and scored according to the above indicators. If the score exceeds the preset threshold, it is judged as a vulnerable node. At the same time, key hub nodes such as transformers, line connection points, and supercharging station access points are marked first.

[0155] The formula for calculating the risk value of a vulnerable node is as follows:

[0156]

[0157] in, The weighting coefficient for the node load factor is determined based on equipment type and power grid operation specifications, and is verified through parameter adjustment using actual operating data. Its value ranges from [0.4, 0.6]. For the real-time load of vulnerable nodes, The rated load for vulnerable nodes, This represents the health status of the vulnerable node's device, with a value ranging from [0,1]. The weighting coefficient for equipment health status is preset with parameters adjusted based on equipment type, service life, and maintenance records to highlight the risk sensitivity of older equipment. The value range is [0.2, 0.3]. This is a weighting coefficient for short-term load growth trends. It is adjusted based on load forecasting accuracy and the load fluctuation characteristics of supercharging station clusters to adapt to scenarios with short-term concentrated charging loads. Its value range is [0.1, 0.2]. For the short-term load increase of vulnerable nodes, such as within 15 minutes;

[0158] When the risk value of a vulnerable node exceeds the risk threshold, the power reduction rules of the scheduling module are used to reduce the charging power of the supercharging station associated with the vulnerable node.

[0159] The collaborative model is trained based on historical operational data and diverse data generated from simulation scenarios, covering edge operating conditions such as extreme loads and equipment failures. It employs a multi-agent near-end strategy optimization algorithm, combined with techniques such as experience replay and target networks as training techniques.

[0160] The training objective is to enable each agent to master collaborative decision-making logic that maximizes its own benefits without compromising power grid security.

[0161] The collaborative scheduling scheme includes:

[0162] Baseline load curve: This curve clarifies the expected charging load level of each supercharging station in the future, serving as the core reference for power allocation of charging piles within the station.

[0163] Adjustable power range: Informs each station of the power margin that can be adjusted upwards or downwards when receiving grid control commands;

[0164] Vehicle guidance suggestions: When the station is overloaded or the equipment is malfunctioning, the app will push recommendations for nearby stations to users. Taking into account factors such as driving distance, waiting time at the target station, and charging price, the system will guide users to divert traffic and avoid congestion at a single station.

[0165] Grid demand module: Aggregates the current status data of all supercharging stations to construct a virtual resource pool; parses the grid-side data into specific control requirements; and generates and issues global control instructions to each supercharging station based on the collaborative scheduling scheme, control requirements, and the real-time status of the virtual resource pool.

[0166] The current status data of the supercharging station includes adjustable load resources and candidate resources.

[0167] The power grid side data is parsed using a conventional rule engine and data parsing techniques, as follows:

[0168] First, it receives grid-side data from the grid dispatching system, including real-time electricity price signals, AGC frequency regulation commands, peak shaving or valley filling load adjustment commands, and load early warning information for vulnerable nodes in the grid.

[0169] Subsequently, the data is categorized and analyzed according to demand type. For example, electricity price signals are analyzed into periods that encourage charging and periods that suppress charging; dispatch instructions are analyzed into specific load adjustment ranges, ensuring that every piece of grid-side data is transformed into clear and quantifiable control targets.

[0170] The global control command is generated by a weighted allocation algorithm based on the collaborative scheduling scheme, the parsed power grid control requirements, and the real-time status of the virtual resource pool.

[0171] Global control instructions include execution information, control tasks, constraints and safety boundaries, feedback requirements, and auxiliary reference information;

[0172] Execution information includes:

[0173] Unique instruction identifier: A globally unified code used to trace the instruction execution trajectory and distinguish different control tasks;

[0174] Target supercharging station: Clearly define the recipient of the instruction;

[0175] Execution period: time window accurate to the minute, adapting to the real-time control needs of the power grid.

[0176] Control tasks include load adjustment tasks and candidate resource allocation tasks;

[0177] Load adjustment tasks include:

[0178] Adjustment direction: Clearly define load reduction, load increase, and maintain load stability;

[0179] Adjustment range: Quantify specific values ​​or proportions, such as reducing the load by 200kW or increasing the load to 1.2 times the current load;

[0180] Priority: Indicates the urgency of the task.

[0181] Constraints and safety boundaries include:

[0182] Power limit constraint: Define the threshold that the total power of the supercharging station must not exceed, and adapt to the upper limit of the power supply system capacity within the station;

[0183] User rights protection: Limiting load that cannot be reduced;

[0184] Equipment safety constraints: Restrictions based on the health status of the equipment.

[0185] Feedback requirements include:

[0186] Feedback content: Execution data that needs to be reported;

[0187] Feedback frequency: Regular control provides feedback once every 5 minutes, emergency control provides feedback once every 1 minute;

[0188] Abnormal reporting: Clearly define the reporting requirements for situations that fail to meet the standards.

[0189] Supplementary reference information includes:

[0190] Background of regulation: Briefly explain the basis for the directive, such as responding to the grid's peak shaving needs, mitigating the risk of overload at vulnerable nodes, and optimizing costs during periods of low electricity prices;

[0191] Reference data: Relevant auxiliary information, such as the current real-time electricity price of 1.2 yuan / kWh, etc., to help supercharging stations optimize their execution strategies.

[0192] The proposed urban charging facility scheduling and grid demand-side control system for supercharging stations addresses the issues of low charging efficiency and uneven resource utilization in supercharging station cluster operation by constructing a vehicle-station-grid collaborative optimization system that integrates real-time data from vehicles, stations, and the grid. It comprehensively utilizes dynamic priority scheduling and multi-agent collaborative decision-making. At the same time, it enhances the interaction between large-scale charging loads and the grid, effectively improving grid operation stability and renewable energy consumption while ensuring the user charging experience.

[0193] In summary, this invention establishes an integrated collaborative optimization and intelligent scheduling system for vehicles, stations, and networks, thereby achieving adaptive and precise allocation of charging resources and multi-station collaboration, which improves charging service efficiency and facility utilization.

[0194] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A city-wide dispatching and grid demand-side control system for supercharging stations, characterized in that, include: Data perception module: Collects a dataset including user charging preference data, vehicle battery status data, supercharging station equipment status data, and grid-side data, and preprocesses the dataset; obtains vehicle features based on the charging preference data and battery status data, and classifies the vehicle based on a preset threshold and the vehicle features to obtain classification results; Scheduling module: Based on the vehicle characteristics, classification results, and equipment status data, it generates a power allocation scheme through a dynamic priority strategy; based on the power allocation scheme, it performs collaborative analysis on all supercharging stations through a preset collaborative model to generate a collaborative scheduling scheme; Power grid demand module: Aggregates the current status data of all supercharging stations to build a virtual resource pool; The power grid-side data is parsed into specific control requirements; Based on the aforementioned collaborative scheduling scheme, control requirements, and the real-time status of the virtual resource pool, global control instructions are generated and issued to each supercharging station.

2. The system according to claim 1, characterized in that, The vehicle characteristics include charging urgency and dispatchability index; The formula for calculating the charging urgency is as follows: in, Due to the urgency of charging, For the target state of charge, The current state of charge, For the battery's rated capacity, The maximum charging power of the battery. For battery health, For user urgency level, For charging efficiency; The calculation steps for the scheduling index are as follows: First step, calculate the charging time: in, This refers to charging time; The second step is to determine the charging start time based on the charging time. The third step is to calculate the scheduling time window based on the charging start time. The fourth step is to normalize the scheduling time window into the scheduling index.

3. The system according to claim 2, characterized in that, The classification results include fast charging and fast running type, fast charging and slow running type, slow charging and fast running type, and slow charging and slow running type; If the slow-charging, slow-moving vehicle has bidirectional charging and discharging capabilities, it is marked as a candidate resource.

4. The system according to claim 3, characterized in that, The power allocation scheme is generated as follows: The first step is to assign initial priorities to different types of vehicles based on the classification results; The second step is to calculate the dynamic priority of the vehicle based on the initial priority, using the following formula: in, For dynamic priority, As the initial priority, Due to time constraints, As for the power urgency factor, As a guiding factor for electricity prices, , and These are the weighting coefficients; The third step involves sorting the vehicles based on the dynamic priority and applying different power allocation schemes to different types of vehicles. The specific operations are as follows: For fast-charging and fast-moving vehicles, the charging power is the minimum value among the charging pile's rated power, the vehicle battery's maximum acceptable charging power, and the vehicle battery's maximum acceptable charging power after battery health status correction. For vehicles that charge quickly but travel slowly, the charging power calculation formula is as follows: in, This refers to the charging power. For vehicles that charge slowly and travel quickly, the charging power should be set to 60% to 80% of the charging pile's rated power. For vehicles that charge slowly and travel slowly, the charging power should be set to 40% to 60% of the rated power of the charging station. For vehicles marked as candidate resources, the supercharging station is discharged with the owner's authorization.

5. The system according to claim 4, characterized in that, The collaborative model adopts a distributed optimization framework based on multi-agent deep reinforcement learning; The collaborative model models each supercharging station as an intelligent agent, and each intelligent agent includes a state space and an action space. The reward function of the agent adopts a two-layer structure of individual reward and collective punishment; The individual rewards include operational revenue and user satisfaction rewards; The operational revenue reward function is as follows: in, As a reward for operating revenue, This represents the real-time charging power of the i-th charging pile at the supercharging station. For the time interval of revenue statistics, The unit price for charging at supercharging stations, The total power drawn by the supercharging station from the power grid. This refers to the real-time electricity price on the power grid. The user satisfaction reward is quantified into a specific reward value based on average waiting time and charging completion rate. The collective penalties include load volatility penalties and overload risk penalties; The load volatility penalty function is as follows: in, The penalty value is the penalty for load volatility. The weighting coefficient for the load volatility penalty. The standard deviation of the total load of all supercharging stations in the region within a set time window. This represents the total load of all supercharging stations within the region. The overload risk penalty function is as follows: in, The penalty value is the overload risk penalty. The weighting coefficient for overload risk penalty. The risk value for vulnerable nodes in the power grid. This is a preset risk threshold.

6. The system according to claim 5, characterized in that, The vulnerable nodes are identified based on grid-side data. The formula for calculating the risk value of the vulnerable node is as follows: in, This is a weighting factor for the node load rate. For the real-time load of vulnerable nodes, The rated load for vulnerable nodes, The health status of devices on vulnerable nodes. This is a weighting coefficient for the health status of the equipment. The weighting coefficient for the load growth trend. This represents the load increase rate of vulnerable nodes.

7. The system according to claim 6, characterized in that, The current status data includes adjustable load resources and candidate resources; The power grid-side data is parsed using conventional rule engines and data parsing techniques; The global control command is generated using a weighted allocation algorithm; The global control command includes execution information, control tasks, constraints and safety boundaries, feedback requirements, and auxiliary reference information.

Images